UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Time inconsistent commercialization competitions Zubchenko, Ganna 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2012_fall_zubchenko_ganna.pdf [ 1.78MB ]
Metadata
JSON: 24-1.0072776.json
JSON-LD: 24-1.0072776-ld.json
RDF/XML (Pretty): 24-1.0072776-rdf.xml
RDF/JSON: 24-1.0072776-rdf.json
Turtle: 24-1.0072776-turtle.txt
N-Triples: 24-1.0072776-rdf-ntriples.txt
Original Record: 24-1.0072776-source.json
Full Text
24-1.0072776-fulltext.txt
Citation
24-1.0072776.ris

Full Text

TIME INCONSISTENT COMMERCIALIZATION COMPETITIONS  by Ganna Zubchenko BSc., Dnipropetrovsk State Agrarian University, 1996 MSc., Dnipropetrovsk State Agrarian University, 2000   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in  THE FACULTY OF GRADUATE STUDIES (Agricultural Economics)      THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   April 2012   © Ganna Zubchenko, 2012  ii Abstract Business innovation is a powerful source of productivity increase and economic growth.  Lack of financing at commercialization stages of innovation development often dooms socially valuable innovations to failure. To address pre-commercialization market failure, public agencies assist entrepreneurs with financial resources through commercialization competition mechanisms. Public agencies administering commercialization competitions may have dual objectives. Ex ante, they aim to induce entrepreneurs to invest an efficient amount of resources, accounting for resource costs and the associated expected social benefits. Ex post, commercialization awards should be allocated so that the expected social welfare aggregated across all projects is maximized. This would likely entail maximizing the number of socially valuable projects which are successfully commercialized. The objective of this thesis is to demonstrate that the ex ante and ex post objectives of a public agency may be in conflict. The ex ante award allocation criteria announced when the competition is launched may differ from the ex post award criteria retained by the agency as private information. Thus, agency decision-making may be time inconsistent.  To achieve this objective, first, a theoretical model of commercialization competition was developed based on a real-world prototype, New Ventures Competition, administered by the British Columbia Innovation Council. It demonstrates that the competition causes underinvestment on the entrepreneurs’ side and allows the agency to deviate from ex ante announced award allocation criteria.  Second, since time inconsistency is related to competition fairness, an experiment was conducted to determine whether respondents trade the competition fairness for “greater  iii social good”. It was established that when respondents are aware of both entrepreneurs’ financial contributions to projects and their financial need, and when they are put into situation where maximization of social welfare is in conflict with maintaining ethical values, respondents tend to deviate from an allocation decision based on ex ante announced criteria.  Third, it was established that when respondents face a particularly strong ethical trade-off, the higher the grades they received on average in their last year at school, the less they tend to change their original rankings to assist in the commercialization of comparatively lower quality projects at the expense of competition fairness.  iv Preface  Experimental research conducted in this thesis was approved by the University of British Columbia Behavioral Research Ethics Board. Ethics Approval Certificate number is H11-01078.  I credit my supervisor Dr. James Vercammen for developing an idea for this research and theoretical model of commercialization competition. He came up with the framework of the social welfare versus fairness model presented in this thesis. I credit my external reviewer Dr. Nathan Schiff for the idea of the ex ante public agency’s objective function being different from the ex post one, and also for bringing the notion of individual fairness to my attention as the one being involved in the time inconsistency problem. Methodology of the conducted experiment and the questionnaires were developed in close collaboration with the supervisor Dr. James Vercammen, and members of the Supervisory Committee Dr. Carol McAusland and Dr. Chloe Tergiman. Experimental sessions were conducted by the master student. Experimental data was shared with the group of four Stat 450 students to be used in their course project and subsequent consultations with the students and Dr. John Petkau were carried out on the appropriate statistical procedures to use for analysis of the experimental data. Analysis of data itself was performed by the master student using R and Excel. Manuscript was prepared by the master student.    v Table of Contents  Abstract .................................................................................................................................... ii Preface ..................................................................................................................................... iv Table of Contents .................................................................................................................... v List of Tables .......................................................................................................................... ix List of Figures .......................................................................................................................... x Acknowledgements ................................................................................................................ xi Dedication ............................................................................................................................. xiii Chapter  1: Introduction ........................................................................................................ 1 1.1 Role of Commercialization Competitions in Stimulating Business Innovation in Canada….…………………………………………………………………………………..1     1.2 Problem Statement ……………………………………………………………………5     1.3   Research Objective.…………………………………………………………………...5     1.4  Methods………………………………………………………………………………..7     1.5 Outline of the Thesis……………………………………………………………………9 Chapter  2: Overview of Government Support for Business Innovation ........................ 10 2.1 Business Innovation as a Prerequisite for Productivity Growth in Canada  .............. 10 2.2 Expenditure of Business Research and Development in Canada  ............................. 11 2.2.1 Private Sector Expenditure on Research and Development ............................... 11 2.2.2    Federal and Provincial Expenditure in Support of Business Research and Development …………………………………………………………………………...15   vi 2.3  Overview of Federal Initiatives Designed to Support Business Research and Development ………………………….…………………………………………………..16 2.4. Federal Programs Designed to Support Businesses Specifically at Commercialization Stage of Innovation Process………………………….……………………………………24 2.5   Commercialization Programs Administered by the British Columbia Innovation Council ……………………………………………………………………………………32 2.5.1 BCIC New Ventures Competition ...................................................................... 32 2.5.2 BCIC CAT Competition ..................................................................................... 35 Chapter  3: Analytical Framework: Game-Theoretic Model of Commercialization Comeptition ........................................................................................................................... 38 3.1 Review of Relevant Research Tournament Literature ............................................... 38 3.2 Time Inconsistency in Economic Literature .............................................................. 42 3.3  Model of Commercialization Competition………………………………………….45 3.3.1 A Public Agency Problem .................................................................................. 46 3.3.2 Entrepreneur's Optimal Effort when Commercializing Innovation without Government Assistance  ................................................................................................. 48 3.3.3 Entrepreneur's Optimal Effort if Participating in Commercialization Competition .................................................................................................................... 49 3.3.4 Time Inconsistency ............................................................................................. 54 Chapter  4: Experimental Study of Fairness versus the Greater Good in Winner Selection Process of Commercialization Competitions ..................................................... 58 4.1 Motivation behind the Experiment ............................................................................ 58  vii 4.2 Social Welfare Function and Individual Fairness Concept with Relation to Commercialization Competition…………………..……………………………………....60 4.3 Economic Experiments in the Literature……………………………………………64 4.4    Description of the Experiment……………………………………………………....68 4.4.1 Questionnaire Instruments .................................................................................. 68         4.4.2.   Hypothesis ……………………………………………………………………...73         4.4.3.   Size of the Sample……………………………………………………………...74 Chapter  5: Data Analysis………………………………………………………………….76 5.1 Relevant Literature..................................................................................................... 76 5.2 Description of the Data .............................................................................................. 78 5.2.1 Ranking Data ...................................................................................................... 79 5.2.2 Summary of Socio-Demographic Information Provided by the Respondents ... 82 5.3 Testing the Agreement of Respondents' Rankings and Re-rankings Within Each Group ......................................................................................................................... 83 5.4 Estimation of Intergroup Concordance ...................................................................... 85 5.4.1 U-Statistic Estimator of Intergroup Concordance ............................................... 85 5.4.2 Coefficient of Structural Concordance ............................................................... 84 5.5 Socio-Demographic Determinants of Respondents' Re-ranking Decisions............... 88 Chapter  6: Conclusions and Discussion ………………………………………………….97 6.1 Summary and Conclusions  ....................................................................................... 97 6.2 Restrictions and Recommendations ......................................................................... 101 References ............................................................................................................................ 103 Appendix A    Federal Programs Designed to Support Commercialization of  viii Business Innovation  ........................................................................................................... 114 Appendix B    Overview of Several Business Commercialization Support Programs Implemented at the Provincial Level ................................................................................ 123 Appendix C    Comparison of Socially Desired and Privately Optimal Levels of Innovation Effort………………………………………………………………………….130 Appendix D    Survey Questionnaire (Group A) .............................................................. 132 Appendix E    Survey Questionnaire (Group B) .............................................................. 138 Appendix F    Socio-Demographic Characteristics of the Respondents ........................ 144 Appendix G    Bootstrapping Results ................................................................................ 145    ix List of Tables  Table 1:  Criteria for Classification of Existing Federal Business R&D                Support Programs………………………………………………………..….20 Table 2:  Largest Federal Direct Expenditure Programs Designed to Support                Commercialization of Business Innovation…………………………….…..27 Table 3:  Examples of Projects Described by Three Quality Characteristics ………...70 Table 4:  Probability of Successful Commercialization and Personal    Contribution by Entrepreneur …………………………………….………..71 Table 5:  Example of Ranking and Re-ranking in Each of Two Groups……….…….73 Table 6:  Descriptive Statistics of Ranks Assigned to Projects in Both Groups ….….79 Table 7:  Kendall’s W Values of Intragroup Concordance………………….………..84 Table 8:  U-statistics as a Measure of Concordance between Rankings and Re-rankings across Two Groups of Respondents……………………..…….87 Table 9:  U-statistics as a Measure of Concordance of Rankings and Re-rankings                within Each of the Two Groups of Respondents .....………………………..88 Table 10:  Coefficient of Structural Concordance as a Measure of Agreement                  between the Two Groups of Respondents……………………..……………91 Table 11:  Regression Results for Factors Determining Re-ranking Decisions …..…...94 Table 12:    Socio-Demographic Characteristics of the Respondents ………………….144     x List of Figures  Figure 1:  Labor Productivity Annual Growth Rate in Some OECD Countries, % …..11 Figure 2:  BERD Funded by Business Sector of Canadian Economy, $ million……....13 Figure 3:  Business Expenditures on R&D in Some OECD Countries, % of GDP…...14 Figure 4:  Business Enterprise Research and Development Expenditures in                 Canadian Agriculture, $ million .…………………………………………...14 Figure 5:  Federal and Provincial Expenditure on Business R&D, Canada,                  $ million……………………….…………………………………………….16 Figure 6: Major Groups of Federal Business R&D Programs by the Form                 of Support, 2010-2011 fiscal year ………………………………………… 21 Figure 7:     Federal Funding of High Education R&D (HERD) and Government                    Performed R&D (GOVERD), $ million…………………………………….25 Figure 8:    Relationship between Prize Size and Firms’ Best Response Functions….....54 Figure 9:    Effects of Time Inconsistency on Firm’s Innovation Efforts ……………....57 Figure 10:  Trade-off between Competition Fairness and Efficiency Gain…….……….63 Figure 11:  An Individual Judge Trade-off between Social Welfare Gain and Fairness of the Competition………..………………………………..….64 Figure 12:  Comparison of Mean Ranks Assigned to Projects at Original                    Ranking and Re-ranking by Two Groups of Respondents………………….80 Figure 13:  Gross Domestic Expenditures on R&D Funded by Provincial Government Sector in Canada, total, million $ …………………………………………..123    xi Acknowledgements  Special thanks to members of my supervisory committee: Professors James Vercammen, Carol McAusland  and Chloe Tergiman. First and foremost, I extend my sincere gratitude to Dr. Vercammen. As my supervisor, Dr. Vercammen has provided excellent guidance and unfailing encouragement throughout the course of preparing for and conducting this research project. I sincerely and greatly appreciate his outstanding professional advice, critical feedback and emotional support that I will never forget. I would also like to thank Dr. McAusland and Dr. Tergiman whose steadfast support of this project, especially at the stages of designing and conducting the experiment was greatly needed and highly appreciated. I send my sincere gratitude to the external reviewer of my thesis Dr. Nathan Schiff. I highly appreciate his availability and promptness, as well as completeness of the review he provided.  His highly valuable comments helped me to increase the quality of this thesis. I thank Professor John Petkau from the Department of Statistics, UBC, and undergraduate students from Stat 450, Jonathan Baik, Vanessa Goh, Mujtaba Badat and Yujia Zhu with whom I cooperated under the Case Studies in Statistics project. Their advice on how to statistically approach the quite unusual type of data obtained in the experiment conducted in this thesis is highly appreciated.  I appreciate useful comments to my thesis provided by Dr. Richard Barichello. I would like to thank Dr. Sumeet Gulati, Faculty of Land and Food Systems, UBC, for his help with accommodating one of the experimental sessions.  Sincere thanks to Larry Bomford, my career mentor whom I met in Victoria, BC, for encouraging me to enter the graduate program in Agricultural Economics at UBC.  xii  I would also like to acknowledge Richard Hallman, Director of Life Sciences and Clean Technologies at BCIC, as well as his colleagues Arlene Fernandez and Jennifer Whelan, for their help in understanding the organizational details of the Commercialization of Agricultural Technologies Competition, BCIC.  xiii Dedication  This work is dedicated to my parents, Liudmyla S. Zubchenko and Gennady I. Zubchenko, without whose caring support and unconditional love it would not have been possible, and to my husband and our son who inspired and encouraged me during this journey.   1 Chapter  1: Introduction  1.1 Role of Commercialization Competitions in Stimulating Business Innovation in Canada Business innovation is viewed by many world governments as a powerful vehicle to accumulate a competitive advantage which, along with innovation produced at research institutions and universities, leads to a progressive productivity increase and economic growth. Based on the latest Organization for Economic Cooperation and Development (OECD) data, the value of innovations produced by Canadian businesses constitutes slightly more than 50% of total R&D value. Based on this parameter, Canada lags behind some other highly developed countries. In addition, R&D spillovers that are often viewed by economists as a distortion to the amount of research undertaken by business sector, contribute to further underinvestment into business innovative activities. Paired with the lack of financial resources experienced by many businesses when trying to bring their innovations to the market, this leads to situations where some socially desirable innovations are commercialized either slowly or not commercialized at all, which may result in significant loss of economic value for society. Thus, economists have shown that government intervention to support business R&D activities is necessary. The Canadian government is aware of the importance of its continuous support of business sector R&D. The way in which this support should be delivered has been the cause of debate among policy makers. As an Expert Panel which conducted a review of federal support to research and development in 2011 estimates, currently almost 55% of federal  2 expenditures in support of most specifically business R&D is allocated in the form of tax credits (Review of Federal Support to Research and Development - Expert Panel Report, 2012). At the same time, the share of government initiatives delivered to businesses as direct funding lies below 10%. Thus, the mix of tools used by the government to support and encourage business innovation is highly skewed towards indirect support. Conversely, according to OECD reports, many countries that are world leaders in innovation today rely greatly on direct support of innovation in the form of various highly targeted programs. In fact, one of the recommendations made by the above-mentioned Expert Panel was to redeploy funds from tax credits to a re-tailored mix of direct support initiatives that would allocate resources where market forces are unlikely to operate effectively. The commercialization stage of the innovation process is one of these areas. Economists have shown that due to limited public financial resources and firms’ unobservable innovation efforts and outputs, it is efficient for governments to use competition or tournament formats to determine the most socially promising innovations, and to support them. This line of research has a long tradition in economics and was formally originated by the ground-breaking work of Lazear and Rosen (1981) on tournaments based on rank order with application to labor economics. The authors compared three compensation schemes: a linear price rate, compensation based on comparison against a fixed standard, and a tournament. They modeled a situation in such a way that the output of workers depends on their own efforts and on an additive stochastic shock parameter that is common to all workers. It was shown that when workers are risk averse and the variance of random shock is large, a tournament performs better than the other remuneration methods.  The work by Lazear and Rosen (1981) was extended by other economists who looked mostly at optimal  3 design of tournaments. Examples of this include works by Green and Stokey (1983), who looked at the efficiency of tournaments under large numbers of contestants and various levels of shock, and by Nalebuff and Stiglitz (1983), who focused on the use of prizes and incentives versus punishment. There is a relatively small group of papers in the economics of tournaments that have studied research and development competitions. Taylor (1995) was the first who noticed that a research tournament where contestants compete to develop the innovation of the highest value to win the competition award is similar to labor tournaments in terms of environment of unobservable research efforts and unverifiable research outputs. Fullerton and McAfee (1999) studied efficient design of research tournaments with heterogeneous contestants and a fixed prize. They showed that research competitions result in underinvestment in R&D on contestants’ part.   The authors went further and demonstrated that this problem can be resolved by limiting the number of contestants to two. Since it is costly for the public agency to investigate which two contestants deserve to participate, Fullerton and McAfee also proposed an auction type mechanism to pre-select the most qualified contestants. In a more recent paper, Che and Gale (2003) departed from the idea of fixed-prize tournament and showed that an auction format with the two most qualifying firms bidding for their prizes is optimal. Thus, the fixed-prize tournaments and all-paid auctions are the two most efficient and popular mechanisms used to select the recipients of public assistance for innovation. Currently, federal and provincial governments deliver a number of programs to support business R&D. Recent estimates show that the number of these initiatives at the federal level alone exceeds sixty (Review of Federal Support to Research and Development - Expert Panel Report, 2012). Some of them such as the Industrial Research Assistance  4 Program (IRAP) address innovation process facilitating innovation from idea development to marketing. Others help entrepreneurs by addressing problems arising at specific stages of innovation development, such as commercialization. This thesis focuses predominantly on a subclass of programs designed specifically to assist entrepreneurs at commercialization stages of the innovation process. Commercialization is often considered as a core process by which an idea is transferred into a successfully marketed product. Lack of venture capital at this stage is a major reason of failure for many socially valuable innovations. Our analysis indicates that providing assistance to Canadian businesses struggling with commercialization of their innovations became one of the federal policy priorities in the business innovation area. Government incrementally increases funding of commercialization programs on a yearly basis and re-adjusts their mix in order to better meet entrepreneurs’ needs. Government also constantly audits current programs and launches new initiatives. Several commercialization initiatives administered by British Columbia Innovation Council (BCIC) over the last several years are a major focus of this thesis. In 2009-2010, BCIC held a Commercialization of Agricultural Technology Competition (CAT). This fixed-prize two-stage competition awarded prizes ranging from $150,000 to $250,000 to four firms with innovations in agriculture, food and bio-products sectors. Currently, BCIC administers the New Ventures Competition which targets a wider range of industries and delivers a slightly smaller set of prizes, but still maintains a competition format. The way these programs are designed generally echoes the research tournament model suggested by Fullerton and McAfee (1999) and incorporates an entry fee, fixed set of prizes, multi-stage setting and heterogeneous contestants. At the same time, it was noticed that final prize allocation decision in these  5 competitions is made by the public agency alone based on expert panel recommendations. This decision is highly subjective. The way it is reached is generally not documented.  1.2  Problem Statement  Public agencies that host commercialization competitions may have dual objectives. On one hand, their ex ante goal is to induce participating firms to invest an efficient amount of resources into their projects, accounting for the cost and the expected social benefits resulting in successful commercialization of the innovation. On the other hand, from an ex post perspective, the ultimate goal of commercialization competition is to maximize the expected social value aggregated across all participating projects, which may entail maximizing the number of socially valuable projects that are successfully commercialized. This discrepancy between the ex ante and ex post objectives may cause the final prize allocation decision to be made based on criteria that is different from those announced when the competition was launched. In this thesis, this problem is refereed as a time inconsistency problem.  1.3 Research Objective This thesis studies the group of publicly funded competitions, which through allocation of pre-determined fixed prizes, assist entrepreneurs in commercialization of their business innovations. The winner-selection rules state that the prizes are awarded to the most potentially successful innovation projects with the highest quality. This thesis focuses on the potential time inconsistency of the winner-selection process. A public agency sponsoring the competition can be modeled as a monopolist buyer that devises this competition as a credible  6 method to commit itself not to use its market power and induce higher innovative effort of participating firms. On the other hand, due to subjectivity of the winner selection process it still has a potential to exercise its market power in order to extract maximum ex post surplus. This may result in ex post prize allocation decisions to be inconsistent with ex ante announced criteria. If participating firms can anticipate that this is the case, their ex ante incentives become distorted which results in lower innovative efforts and potentially lower social welfare. An informal discussion with BCIC officials was the prime motivation for this research. The final prize allocation decision in the commercialization competitions studied in this thesis is made by a jury which faces strong ethical trade-off between deviation from the originally announced winner selection rules and social welfare maximization. Judges realize that instead of allocating the prize to the most deserving project which at the same time can be commercialized by the entrepreneur without the prize, the public money could bring greater benefit being allocated to the projects with slightly lower quality and requiring financial support to reach the market. In other words, the jury chooses between the competition fairness and the greater good. It was noticed that this dilemma constitutes the basis of time inconsistency problem in this type of competitions, and deserves special attention in this thesis.  Thus, the research objective includes following tasks: - review the current level of business innovation assistance in Canada and establish the place of commercialization support in the mix of innovation assistance initiatives implemented by the government;  7 - develop a research tournament model with one risk-neutral principal and two risk- neutral homogeneous contestants, and show how the optimal level of innovation investments changes in presence of time inconsistency; - develop a “competition fairness versus greater good” theoretical model and conduct an economic experiment to reproduce commercialization competition environment and experimentally test the outcomes of a trade-off between competition fairness and the greater good during the winner selection process; - determine which socio-demographic characteristics of experiment participants had significant impact on their award allocation decisions. This research is expected to contribute to the research tournament literature line started by Taylor (1995) and Fullerton and McAfee (1999).  1.4 Methods First of all, this thesis uses a game-theoretic approach of studying strategic decision- making to build a theoretical model of simplified commercialization competition with two participants and a public agency. This model is constructed after a real world prototype, New Ventures Competition, administered by the British Columbia Innovation Council (BCIC). The Nash equilibrium set of strategies are derived as a solution for this sequential moves model. The model shows how the New Ventures Competition in current format can cause R&D underinvestment and distortion of the winner selection process. It demonstrates that if the public agency’s objective is to maximize the number of successfully commercialized projects, than the winner selection process becomes time inconsistent. The implications of inconsistent awarding process for competition efficiency are also derived.  8  Also, a structural model of competition fairness versus greater good maximization was build. The model describes the ethical trade-off that the judges face in commercialization competitions. It postulates that the level of utility of individual judge is determined by a complex mix of factors, and some degree of unfairness which assumes deviation from ex ante competition rules as a compensation for a greater good, is one of them. The model demonstrates that each judge’s decision about the final award allocation is determined by interaction of two phenomena: his individual indifference curve that reflects his preferences for social welfare gain versus competition fairness loss, and his “social welfare gain versus fairness” possibility frontier. Method of economic experiment is used in this thesis in order to investigate the issue of trade-off between competition fairness and the greater good as the one that is closely related to the potential time inconsistency of the competitions outcome. Data in the form of rankings collected from two groups of respondents in the experiment were used to statistically estimate the potential for public agency to deviate from the ex ante announced decision rules when the goal is to maximize the number of successfully commercialized projects. There was no one perfect statistical procedure, nor a statistical package, to carry out an appropriate analysis with the type of data on hands. Thus, several non-parametric statistical approaches supplementing each other were used to test the hypothesis of this study. All of them are extensions of Kendall’s rank correlation coefficient. Together they provided valuable insights into the question of the research.  Empirical investigation of the causal relationship between the socio-demographic characteristics of the respondents and their award allocation decisions were executed by estimating the parameter of linear regression model applying Ordinary Least Squares.  9 1.5 Outline of the Thesis The thesis is organized into six chapters. Chapter 2 provides a descriptive summary of major facts about business innovation in Canada and levels of federal support towards business R&D. It also offers a review of existing federal and provincial business R&D and commercialization competitions. In Chapter 3, the theoretical model of innovation competition with an agency and two firms is presented. The Nash equilibrium strategies are derived; distortions and inconsistencies that can potentially arise under current format of some commercialization competitions are shown. Chapter 4 presents a theoretical model of the ethical trade-off between the competition fairness and greater good. It also describes the economic experiment conducted to recreate the commercialization competition environment and test for the outcome of the competition fairness versus the greater good maximization trade-off that the competition judges face. In the same chapter, hypothesis and experimental procedures are presented. Chapter 5 delivers description and analysis of experimental data. It also offers a discussion of the obtained results. Chapter 6 provides conclusions and suggestions for further research.          10 Chapter  2: Overview of Government Support for Business Innovation This descriptive chapter provides analysis of business innovation in Canada and examines the level of government support of innovation in the private sector. Section 2.1 describes the current productivity level in Canada since it depends on the innovation intensity. Next, in Section 2.2, the level of business investments into R&D is compared with the level of federal and provincial funding directed to support business innovation.  Review of federal initiatives designed to support business innovation is offered in Section 2.3., followed by analysis of major programs addressing difficulties businesses might have specifically at the commercialization stage of innovation development in Section 2.4. Finally, the description of two commercialization programs that became real-life prototypes for the theoretical model in this thesis can be found in Section 2.5.  2.1 Business Innovation as a Prerequisite for Productivity Growth in Canada The issue of how innovation relates to productivity growth has been studied extensively for at least forty years. A number of studies conducted at micro and macro levels found evidence that R&D has a positive impact on labor productivity, which is usually expressed as amount of output per labor input (Criliches & Mairesse, 1982). Thus, economists agree that high rates of business innovation translate into increasing productivity growth and, consequently, economic growth. Recent analysis conducted by the Organization for Economic Co-operation and Development (OECD) shows that firms’ spending on new knowledge explains a good portion of multifactor productivity growth in many OECD countries (OECD, 2010b). Canada is a founding member of OECD, which was established in  11 1986. The following analysis shows how labor productivity in Canada compares to the level its levels in other OECD countries. Labor productivity is expressed as GDP per hour worked in Canada. As can be observed from Figure 1, Canada’s average productivity growth for the period of 2001-2010 was 57% of the average productivity growth for all OECD countries. Figure 1: Labor Productivity Annual Growth Rate in Some OECD Countries, %  Source of data: OECD statistics, 2011, Table “Labor Productivity Growth in the Total Economy” Although the absolute value of labor productivity in Canada is higher than Finland and Germany, and is slightly lower than in the United States, the rate of productivity increase in Canada is low. By this parameter in 2010, the country was much far behind such flagmen as Korea and Finland, and was not able to reach the OECD average level of annual productivity growth.  2.2 Expenditure on Business Research and Development in Canada 2.2.1. Private Sector Expenditure on Research and Development Scientists confirm that investments into research and development (R&D) along with investments into education are major contributors into labor productivity growth (Verbic, 0 1 2 3 4 5 6 7 2001 2010  12 Majcen, Ivanova, & Cok, 2011). At this point, a clarification of several terms appears necessary. Although the terms R&D and innovation are used sometimes interchangeably, it is important to realize the difference between spending on innovation and R&D spending. Research and development is defined as development of a new product or service and, in this sense, has the same meaning as invention. On the other hand, many analysts nowadays think of innovation as a much broader multilevel concept with commercialization being a crucial component of the process. Most innovation results from commercialization of a developed idea. Although many international organizations as well as Canadian agencies collect data on business R&D, it appears that data characterizing the level of innovation is not that widely available. Most of the innovation data come in the form of Statistics Canada Surveys on Innovation (manufacturing industry related: the first survey came in 1987 and the most recent one in 2007), as well as interviews. Thus, since this thesis focuses mostly on the government initiatives related to commercialization of business innovation, for the purpose of the next section and analysis of Canada’s competitive position and policy interest, R&D expenditure and innovation expenditure are refereed interchangeably. Traditionally, there are two major sources of R&D funding. These are business and government expenditures on R&D. In OEDC and Statistics Canada classifications, business sector’s spending on R&D is referred as BERD. Despite the competitive need and desire for Canadian businesses to invest more into innovation, BERD in Canada remained essentially static over the last decade (Figure 2).     13 Figure 2: BERD Funded by Business Sector of Canadian Economy, $ million  Source of data: Statistics Canada, CANSIM, 2011, Table 358-0024 Private sector investments in R&D are approximately 30 times higher than the federal government innovation assistance to the sector. It can also be observed from the figure that in those years when federal government funding for R&D went down, business responded with extra investment and vice versa. Statistics Canada projected that in 2011 businesses will spend $15.6 billion, or 5% more than in 2010. Despite this optimistic prognosis, and still in comparison with countries that are innovation leaders in the world, Canada performs at a moderate level. The next chart represents BERD in Canada as a percentage of GDP in comparison with such world productivity growth leaders as Finland, Korea and the United States. During the period of 2001-2010, R&D expenditures of Canadian businesses were always lower than the average for all OECD countries (Figure 3). Considering that about 15% of business R&D expenditures in Canada are financed abroad, the domestic private sector’s efforts on innovation are quite weak. 0 50 100 150 200 250 300 350 400 450 500 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 Business Enterprise Sector Federal Government Sector  14 Figure 3: Business Expenditures on R&D in Some OECD Countries, % of GDP  Source of data: OECD Statistics, 2011, Table “Main Science and Technology Indicators” After consistent growth during 2002-2007 by almost 20%, research and development expenditures of businesses decreased in 2010 in a majority of the industries, including manufacturing, agriculture, forestry, fishing and hunting (Figure 4).  The year of 2011 brought hope for recovery of R&D spending. Total BERD in all industries increased by 5%, and agricultural enterprises invested 6% more. Figure 4: Business Enterprise Research and Development Expenditures in Canadian Agriculture, $ million  Source of data: Statistics Canada, CANSIM, 2011, Table 358-0024 0 0.5 1 1.5 2 2.5 3 3.5 4 2001 2009 0 20 40 60 80 100 120 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 Total All Industries Agriculture  15 2.2.2. Federal and Provincial Expenditure in Support of Business Research and Development No matter how high private investments in R&D are, economists are convinced that government support is a must when it comes to business innovation. As Schmookler shown, due to a variety of reasons such as uncertainty of the outcome and high cost, private enterprises are unable to perform socially desirable levels of research and innovation (Schmookler, 1959). In addition, theory predicts that private profit-maximizing entities tend to underinvest in R&D from a socially optimal point of view due to spillovers that cause market failure. These are probably the most important reasons why the Canadian government, along with other governments in the world, has continued to support industry R&D for decades through various targeted programs and tax preferences. Thus, government support of business innovation is a second crucial factor after business investments into R&D that determine productivity growth in Canada. After a significant decrease in federal funding directed to support research and development in the business sector during the period of 2002-2007, a recovery trend was observed during the last five years (Figure 5).  Provincial governments play an increasingly important role in assisting entrepreneurs in innovation quest. In 2011, provinces spent 30 times more than in 2011 to help develop and commercialize new goods and services. Budget 2010 set up a call for review of existing federal support for business innovations with aim to develop ways of increasing their effectiveness. The Expert Panel on Federal Support to Research and Development (2011) conducted a Review of federal initiatives in this area and submitted a Report on how the Canadian government can make even better use of the funding it provides to enhance private innovations.  16 Figure 5: Federal and Provincial Expenditure on Business R&D, Canada, $ million  Source of data: Statistics Canada, CANSIM, 2011, Table 358-0024 This document provides the most recent and the most exhaustive analysis of business R&D support to date and is frequently to in this thesis. Based on thorough analysis of innovation in Canada, the Expert Panel made a conclusion that “Canada has a business innovation problem” and called for action (Review of Federal Support to Research and Development - Expert Panel Report, 2012).  2.3 Overview of Federal Initiatives Designed to Support Business Research and Development The major interest of this thesis lies within the area of financial instruments used by the Canadian government to support innovation. Whereas the federal government funds the majority of fundamental research, this section reviews a variety of tools used by the federal government to support and promote private innovations. The next section discusses federal 0 50 100 150 200 250 300 350 400 450 500 Federal Government Sector Provincial Governments Sector  17 initiatives designed to support entrepreneurs specifically during the commercialization stage of innovation development. Governments use various tools to stimulate research and development in the private sector. Among the most widely used are fiscal incentives and direct funding in the form of grants. Traditionally, fiscal tools include R&D tax credits, R&D allowances, accelerated depreciation of R&D capital, etc. Among the direct funding measures the most widely used are grants, loans and procurement. On the other hand, tax credits are considered an indirect measure usually used to encourage innovation. It is traditionally considered that tax support stimulates short-term applied research, whereas direct funding promotes long-term research (OECD, 2010b).  Such countries as USA and Spain, for example, predominantly use the direct assistance approach. On the other hand, out of 34 member countries of OECD, 22 (64%) countries, including Canada, widely provide fiscal incentives to improve the business innovation performance. It is estimated that in 2010-2011, the Canadian government offered $3.53 billion of indirect R&D assistance to businesses, which constituted almost 55% of total expenditure in support of business R&D. On the other hand, a recent OECD study involving 21 countries showed that firms receiving direct public support for innovation tend to invest 40-70% more than those that do not (OECD, 2010b). As was estimated by the Expert Panel on Federal Support to Research and Development (2011), the direct support to business R&D in 2010-2011 fiscal year was almost 10% of total federal expenditures, or 630 million dollars. In 2008, Canada ranked second among the Organization for Economic Co-operation and Development countries in total value of support and incentives for business innovations as a percentage of GDP, but ranked sixteenth in terms of business investments into R&D as a percentage of GDP (Review of Federal Support to Research and Development, 2011). Many  18 analysts agree that Canada provides generous R&D funding, but the question is whether it works as expected. For example, a recent study of effectiveness of the Canadian Scientific Research and Experimental Development (SR&ED) tax credits shows that in Canada each dollar of tax revenue forgone generates $1.30 of additional R&D, 2.28 times less than in the US. (PRO INNO Europe, 2009). Seemingly large level of all kinds of investments into R&D and comparatively low levels of resulting productivity growth makes many analysts stipulate that there is something about innovation that is not captured completely by the current policy. The Expert Panel in the Report of Federal Support to Research and Development offered their perspective on this concern. Contrary to more traditional views of innovation as a product of research and development only, the Panel sees innovation as an innovation ecosystem, including not only development of an idea and its prototyping, but also commercialization of the innovation and collaboration with the key stakeholders such as customers, competitors, suppliers and universities. As an illustration of this point, an analysis conducted by OECD on firms’ collaboration on innovation in 2004-2006 revealed interesting data (OECD, 2010a). Countries possessing the highest level of labor productivity and BERD seem to collaborate more on both the national and international level. In one of the world innovation leaders, Finland, 57% of all innovating firms collaborate, with 24% collaborating only with national partners. Similar level of collaboration on innovation is reached in Chile. In Canada, slightly more than 20% of all innovating firms looked to cooperate, out of which approximately 15% chose to collaborate with international partners. In Chile, it is indicative that 78% of the firms with a high level of R&D investments collaborate on innovation, as well as almost 35% of  19 firms that do not spend on R&D at all. In Canada, only about 10% of the firms with no R&D spending were able to access help from the government and/or research institutions. Federal government currently delivers dozens of publicly funded programs to stimulate and support businesses at all stages of the innovation process. The Expert Panel on Federal Support to Research and Development made a first unprecedented step to conceptualize this extremely diverse federal business R&D support suite. In its analysis, the Panel focused on a database comprised of 60 major programs delivered by 17 federal entities, focusing on how effective the structure of the portfolio is and the programs themselves. In this thesis, major interest is in different aspects of these federal programs, particularly their format and funding allocation procedures. Thus, the following review is intended to complement the Expert Panel on Federal Support to Research and Development findings. An ambitious task of reviewing all of the existing business innovation programs is outside the scope of this thesis. The thesis focuses on several major federal initiatives in the area of business innovation support and details their size, format, application and funding allocation processes. Since innovation itself is a multi-level activity, current program mix forms a whole universe in itself.  Table 1 offers a list of criteria that can be helpful in understanding and visualizing this set of federal initiatives currently available to support entrepreneurs through various stages of innovation process.     20 Table 1: Criteria for Classification of Existing Federal Business R&D Support Programs Criteria Type of programs Input supported  Ideas and knowledge  Talented and educated entrepreneurial people  Networks, collaborations  Capital and financing Type of activity supported  Basic research  Applied research  Experimental development and commercialization Form of support  Tax incentives  Repayable and non-repayable grants  Provision of services  Procurement of research Eligible recipient  Business  Other organization providing commercially oriented R&D Scope  National  Sectoral  Regional Note: the table is generated by the author based on classification presented in Innovation Canada: Call for Action, Review of Federal Support to Research and Development – Expert Panel Report, Industry Canada, (2011).  Based solely on the form of the R&D support, the majority of federal initiatives supporting business R&D and commercially oriented innovation specifically can be categorized into three main groups (Figure 6).  21 Figure 6: Major Groups of Federal Business R&D Programs by the Form of Support, 2010-2011 fiscal year   Note: the figure is generated by the author based on data presented in Innovation Canada: Call for Action, Review of Federal Support to Research and Development – Expert Panel Report, Industry Canada, (2011).  The spectrum of federal R&D support mechanisms is very broad and virtually impossible to cover in this thesis. Further, this section presents a brief overview of some largest federal initiatives from each of the three groups presented in the figure above. Particularly, a closer look at the Scientific Research and Experimental Development Tax Credit Program (SR&ED) as an indirect support vehicle, Industrial Research Assistance Program (IRAP) as the largest direct support instrument targeting business innovation, and FPInnovations as a pure research institution and recipient of a significant chunk of federal Federal Programs to support Busienss Innovation $6.44 billion Indirect support SR&ED tax credit $3.53 billion Direct Support to public/non-profit commercialily relevant R&D $0.98 billion Direct Support to Business R&D $0.63 billion  22 R&D funding is taken.   The suite of federal initiatives to support business innovation is heavily skewed towards indirect support in the form of R&D tax credits. Scientific Research and Experimental Development Tax Credit Program (SR&ED) is one of the most significant indirect measures currently used on the federal level to support business innovation responsible for almost 55% of federal business R&D expenditures. It is entitled to $3.5 billion in annual reduction of R&D costs for almost 24,000 businesses across the country. In 2010-2012, it was responsible for about 70% of all federal funding designated to support business and commercially oriented R&D (Review of Federal Support to Research and Development - Expert Panel Report, 2012). The program reduces the amount of taxed revenues on the amount of eligible R&D expenses that promotes investments in R&D in Canada. Despite the obvious recognition of the program by business and government communities, there exists concern that the substantial amount of funding may cause the program to transition into a subsidiary type of instrument compromising the whole idea of stimulating development of innovations by firms (Review of Federal Support to Research and Development - Expert Panel Report, 2012). In 2010-2011, another significant cluster of programs responsible for more than 15% of all federal expenditures encouraging R&D, targets research in public and non-for-profit organizations such as research institutes.  This money is distributed by agencies such as the Natural Sciences and Engineering Research Council of Canada (NCERC) in the form of researcher driven inquiry grants, scholarships and fellowships, and funding of Canada Research Chairs and industry-academic R&D partnerships, etc. Keeping in mind that this thesis is mostly concerned with direct support for business R&D type of programs, particularly those addressing commercialization stages of the innovation process, the  23 FPInnovations programs is reviewed here as the largest in this group of programs. FPInnovations is the world’s largest private, non-for-profit forest research institute. It is formed of four forest research institutions that consolidated their efforts in order to significantly strengthen forest research potential in Canada. The Canadian Wood Fiber Centre, formed by the federal agency Natural Resources Canada in 2006, is one of them. FPInnovations leads industry innovation through development and transfer of new technologies and finding new economic market opportunities for the forest sector. Funding of non-repayable grants and contributions through FPInnovations reached $78.3 million in 2010-2011, making it responsible for 5.3% of direct federal funding of R&D.  Among the most recent research initiatives, there is forest inventory incorporating fiber attributes and research targeted at increasing productivity and competitiveness of the Canadian forest sector. As was estimated by the Expert Panel in the Review of Federal Support to Research and Development (2011), 47% of federal direct expenditure programs are targeted to small and large businesses. The remaining 53% is allocated to support research at universities, Canadian non-profit institutions and other federally performed R&D. The Industrial Research Assistance Program (IRAP) is the largest of all currently delivered direct expenditure programs offering financial support and a variety of services to businesses developing new technologies.  The total actual spending under this program in the 2010-2011 financial year was above $286 million (Treasury Board of Canada Secretariat, 2011).  The 2011 federal budget announced $80 million in spending over three years for IRAP.  Although this amount is going to be a significant support to innovating Canadian businesses, the value declined noticeably in comparison to the previous levels of funding. IRAP and other similar  24 programs are geared towards helping Canadian small and medium size enterprises through all stages of the technological innovation process, starting with development of an idea and ending with its commercialization. IRAP is an example of government using at least two basic policy instruments simultaneously: financial and information-based instruments. It is considered highly successful and received wide support and recognition in the society (PRO INNO Europe, 2009). Administered by the National Research Council (NRC) in 1961, IRAP is one of oldest government programs in the area of technological innovations that is still in practice. NRC delivers this service through five regional offices across Canada with thirty- three Industrial Technology Advisors working in the Pacific office where British Columbia and Yukon belong. Assistance is provided in the form of technology advice and customized services, as well as cost-shared financial support. The major benefit to Canadians is considered an increase of wealth through building an innovation capacity of around 10,000 small and medium size enterprises annually (National Research Council Canada, 2009). The decision about financial assistance is made based on commercial capabilities of the project and financial capabilities of the firm.  2.4 Federal Programs Designed to Support Businesses Specifically at Commercialization Stage of Innovation Process Many stakeholders consider that the low capacity of Canadian businesses to translate innovation into successful market products significantly contributes to Canada’s comparatively week productivity (Final Report of the Expert Panel on Commercialization, 2006). The previous section shows that federal support of business R&D, although moderate comparing with some other countries in the world, is nevertheless historically high and stable  25 over the last decade. At the same time, Canada made tremendous progress in accumulating significant innovation potential at the idea development stage by increasing funding of university research and federal government sector research over the last decade (Figure 7). Figure 7: Federal Funding of High Education R&D (HERD) and Government Performed R&D (GOVERD), $ million  Source of data: Statistics Canada, CANSIM, 2011, Table 358-0001 Thus, there must be other factors that do not allow Canada to use this vast potential. As analysis conducted by two consecutive Expert Panels on Commercialization in 2006 and 2011indicates, Canada struggles with the demand side of innovation. Particularly, the problem lies with the businesses being sometimes unable to commercialize available innovations. Commercialization often implies the means required to turn an idea or prototype into the marketable product. The major means is capital. There has been a long-standing concern at the level of federal and provincial policy makers that Canadian entrepreneurs 1,500 1,700 1,900 2,100 2,300 2,500 2,700 2,900 3,100 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 HERD GOVERD  26 investing into innovations are in fact handicapped when it comes to commercialization of their ideas. In fact, it is a reality that a certain percentage of innovative ideas never reach the market due to lack of funding at commercialization stage. Despite agreement on the necessity to bolster commercialization of business innovation, there is no consensus on how this should be done. Three major components of commercialization – people, idea and capital – have to perform together for the innovation to be successfully launched. In 2007, and for the first time, Statistics Canada conducted a nationwide survey testing the ability of Canadian firms to commercialize. Among the other findings of this study, the lack of financial support in the post-development phases of the innovation was the most-often named reason of innovation failure (Rosa & Rose, 2007). Respondents of the survey indicated that there is enough financial support from public institutions at the stage of concept development, whereas financial assistance from the government is limited at the following stages of development and commercialization of the new product. Particularly, there is a gap in financial support during the final stage of research where the new product or service moves from a laboratory into the market. This challenge is especially intense for smaller and medium size businesses. In 2010-2011, federal programs designed to help businesses to commercialize their innovations accounted for almost 26% of direct support for business R&D and 2.5% of all federal funding in support of business innovation.  Over the last decade, government introduced new programs that help cover pre-commercialization gap and was consistently increasing funding of already existing programs. Table 2 shows the six largest federal initiatives designed to provide either direct non-repayable or repayable funding to help  27 innovators bring their ideas to the market. This list is far from being exhaustive and is intended to provide a general idea of the size and nature of the major federal initiatives. Table 2: Largest Federal Direct Expenditure Programs Designed to Support Commercialization of Business Innovation Name of the Program Agency      , $million Repayment requirements Western Diversification Program (WDP) Western Economic Diversification Canada (WD) 73.3 non-repayable basis Centres of Excellence for Commercialization and Research (CECR) Tri-Council (the three granting councils: NSERC, SSHRC, CIHR) 49.8 non-repayable basis Business and Regional Growth Program (BRGP) Canada Economic Development for Quebec Regions (CED-Q) 51.2 non-repayable basis Networks of Centres of Excellence (NCE) Tri-Council (the three granting councils: NSERC, SSHRC, CIHR) 78.4 non-repayable basis Business Development Program (BDP), innovation element Atlantic Canada Opportunities Agency (ACOA) 13.4** repayable basis Agricultural Innovation Program (AIP), commercialization stream Agriculture and Agri-Food Canada 20.0 repayable basis Note: the table is generated by the author based on data presented in Innovation Canada: Call for Action, Review of Federal Support to Research and Development – Expert Panel Report, Industry Canada, (2011).   Author’s estimation, ** Total funding  Three out of six above mentioned programs are designed to help commercialize technology in particular regions of Canada. The oldest and the largest in terms of delivered support is the Western Development Program (WDP) covering British Columbia, Alberta, Saskatchewan and Manitoba. Business and Regional Growth Program (BRGP) targets businesses that through commercialization of innovative ideas contribute to sustainable  28 development of Qu ́bec. Finally, the Business Development Program (BDP) specifically helps entrepreneurs in Atlantic Canada. With exception of WDP, all the programs target small and medium size businesses in a wide variety of sectors. For example, CECR delivers support to entrepreneurs in environmental science, natural resources, energy and health science, whereas Agricultural Innovation Program (AIP) deals with commercialization of agricultural technologies and bioproducts. As was suggested by the Expert Panel on Commercialization (2011), government constantly increases the amount of direct support at the stage of commercialization. Thus, four out of six programs under consideration deliver funding in the form of grants and contributions, in many cases to cover specified type of costs. Two programs, BDP and AIP, provide access to interest-free repayable loans with favorable terms of repayment. A detailed description and review of these and some other commercialization programs is provided in Appendix A. All these programs maintain an open application process, which assumes that applications can be filed at any time until the budget is completely allocated. In all these programs, the application and recipient selection process is usually multistage. Applicants are asked to submit a Proposal detailing the innovation project and a Funding Request in the form of Business Plan. They usually show the amount of own and borrowed funds that have been already committed to the project. Applicants must also reveal their current financial standing by submitting their financial statements for current and previous years. An assessment and funding allocation process is at least two-stage. It usually involves assessment by a panel of external experts that provides recommendations to an agency decision making committee. Final funding decision is made solely by this committee. A  29 study of the recipient selection procedures shows that for a majority of the programs under consideration, program selection criteria used for short-listing the applicants are usually well articulated and available for applicants to consult on the Web. At the same time, there was no a single example found when the final allocation decision process would be explained. Thus, in many cases, despite of the very transparent assessment process, the most important stage when the final decision is made remains a mystery for the involved businesses. Two examples that follow illustrate this point. Depending on duration of the program, audits and evaluations of federal programs take place either on a regular basis or after the program closes. The last evaluation of WDP conducted in October 2008 was reviewed. Evaluation was focused on estimating relevance, success, cost-effectiveness, as well as design and delivery of the program. The variety of methods used in evaluation included data and file reviews, informant interviews, recipient and non-recipient surveys, case studies and focus groups. In terms of design and delivery of the program, the general conclusion based on recipients’ survey was that WDP achieved the expected results and was effective. At the same time, suggested improvements primarily dealt with increasing clarity and streamlining of the approval process. All non-funded applicants who participated in the survey indicated that they were unsure why they did not receive the funding. Among the issues mentioned by key participants were the need to improve transparency and regularity of the funding process. Out of eight suggested modifications to design and delivery of the program, one was streamlining of the internal assessment and decision-making process. In order to address these concerns, it was recommended to improve databases in order to collect more data about projects that do not  30 receive funding and why they do not receive funding (Western Economic Diversification Canada, 2012). Agri-Opportunities Program (AOP) was implemented by Agriculture and Agri-Food Canada during the period of January 2007—March 2011 and was a predecessor of Agricultural Innovation Program (AIP) currently implemented by the agency. Although AOP was terminated in 2011, its inclusion in this review was caused by availability of audit results shedding light on how in fact the assessment and selection of funding recipients was conducted. The Program guide clearly stated that intention of the program is to “contribute to as many eligible projects as possible” with priority given to projects that increase demand for primary agricultural projects. The most recent AOP audit took place in June 2011. The Office of Audit and Evaluations provided valuable findings and recommendations that confirm the validity of questions raised by this thesis. In particular, the audit examined how funded projects met three major project assessment criteria announced by the programs and how this was documented. It was established that there was no evidence from the assessment documents for how 7 out of 27 funded projects met the criteria of increasing market demand for Canadian agricultural products. The recommendation to reinforce documentation of the assessment process as to how all the criteria are met was issued. The Industry Review Committee (IRC) played an advisory role in project assessment. It was called to provide professional industry advice to Agriculture and Agri-Food Canada on approval or rejection of each project. In accordance with regulation documents, IRC was supposed to submit a written summary analysis for each project. They were not prepared (Agriculture and Agri- Food Canada, 2011).  31 With relation to this, availability of data on implementation of R&D programs in Canada, and particularly on the funding allocation process, is a big issue. This has been pointed out by several authors who studied the effectiveness of the programs’ implementation. For example, Cozzarin emphasizes the lack of quality data on outputs of Canadian R&D programs and, as a consequence, the lack of systematic analysis and examination of their effectiveness on the policy-makers’ side (Cozzarin, 2008). In addition, despite the wide range of data on innovation, research and development collected by Statistics Canada over the years, the current statistical system does not accumulate data on the commercialization process. In fact, there was no a single systematic study of this issue on the federal level until 2006 when the first Expert Panel on Commercialization was created. So far, only federal initiatives on assisting businesses in commercialization of their ideas were discussed in this chapter. Besides, provinces implement numerous programs in this area as well. Their format is often similar to those implemented at the federal level. Due to the scope limitation of this thesis, the reader is referred to Appendix B for more detailed description of the most noticeable programs at the provincial level. Thus, federal government implements a wide range of programs with focus on supporting business innovation at the commercialization stage. Funding of these programs increases, and along with traditional programs like WDP, new initiatives such as AIP are regularly introduced. Direct commercialization support is delivered in a variety of forms, particularly through non-repayable grants and contributions and re-payable interest-free loans. The application and applicants’ assessment process is usually multistage and involves external experts’ evaluation. The final decision is discretionary and lies with an agency administering the program. As audit of some programs revealed, the final funding allocation  32 decision is not always transparent and well documented, causing concerns of applicants who did not receive the funds.  2.5 Commercialization Programs Administered by the British Columbia Innovation Council British Columbia Innovation Council (BCIC) is a British Columbia (BC) Crown Corporation. This means that the only shareholder of this company is the BC Government. It was created in 1979 as one of the numerous entities serving the purpose of delivering government initiatives in the areas where government intervention is required. Its mission is promotion of entrepreneurs’ development and commercialization of technologies. It delivers programs in the following three areas: Entrepreneurial talent development and retention, Commercialization of technology and Technology Awareness. It is located in downtown Vancouver and has 20 employees. BCIC currently delivers government support in a variety of forms. The most noticeable are mentorship programs, scholarships, targeted workshops, certificates, and a network of incubators inside educational institutions, awards and competitions.  2.5.1. BCIC New Ventures Competition In order to stimulate commercialization of technologies in the province, during the last ten years the BCIC administers BCIC-New Ventures competition, an annual business- idea competition providing entrepreneurs with all sorts of resources, including prize money they might need to successfully commercialize their ideas. It is considered one of the North America’s largest technology business-idea competitions. In 2012, the BCIC-New Ventures  33 competition offers a total of $350,000 in prizes to the winning business ideas from a range of industries including life sciences. This money is awarded in the form of three major prize packages ($100,000, $55,000 and $35,000) and several smaller targeted prizes. The competition requires a $200 entry fee that also covers the cost of educational seminars and networking events (British Columbia Innovation Council, 2011). BCIC-New Ventures is targeted towards privately held non-incorporated BC businesses that are currently permanent residents in BC looking for public support in order to commercialize innovative technologies. The format of the New Ventures competition includes four rounds. In the first round the participants register and submit a general description of their ideas. In the following round, a feasibility test, they must present a more detailed five-page description of their projects. This submission is evaluated by five judges according to five criteria: intellectual property, market, distribution, competition and financial. According to each criterion, the projects are graded on a scale of 1 to 5: 1=poor, 2=fair, 3=good, 4=very good, 5=excellent.  Twenty five best ideas will advance to round three. In the third round, the 25 preselected participants receive training in business plan composition and will have to present their eight-page business plans to a jury. Besides a description of technology, market, distribution channels, management and competitors, the business plan must also state a financial position of the venture. In particular, entrepreneurs have to reveal how much money has been spent on the project, how much money has been raised, how much money is required, as well as provide cash flow statements and pro-formas for two years. Ten successful competitors will advance to the final stage of the competition. Here, they present their business plans to the jury panel consisting of industry experts and specialists. The jury is appointed by New Venture BC Board. Jury members are provided  34 with the following instructions: “Rank the submissions in the order they deem to be most likely to succeed as a business venture in the current market” (BCIC New Ventures Competition, 2012). The final winner decision is subject to approval by New Venture BC Board (BCIC New Ventures Competition, 2011).  In this type of competition the prizes are awarded based on subjective opinions of the jury panel. Even though the criteria for selection is announced as “the most potentially successful project”, in a sense “the best” project, the total welfare is not usually maximized when the best commercially viable ideas with the highest value are awarded the final prizes. The reason for this is the possibility of some of the “best” business plans to be commercialized even without the prize. In situations where judges are aware of the financial situation of the participants, ex post they are tempted to consider not only the quality of the business ideas but also a financial need of the participants which they are fully aware of based on submitted financial documents. As a result, a business idea of slightly lower than the “best” quality but in need of external financing may receive the prize, whereas the “best” business plan that can be commercialized even without the prize money may end up not receiving an award. At the same time, if BCIC announced dual criteria, the quality and financial need of the participant to commercialize the innovation, the participants would realize the possibility of inconsistency of winner selection and consciously lower their innovation efforts. This would result in the loss of the welfare as well. In 2011, as a part of BCIC-New Ventures Competition, BCIC also awards two Shildroth Agritech Innovation prizes, $30,000 and $20,000 each. This new initiative was created to support business innovation in BC agriculture and is financed together with the BC  35 Ministry of Agriculture. The BCIC New Venture participant who belongs to either the primary agricultural production industry, the food or bioenergy/bioproducts industries, along with the regular set of BCIC New Venture Competition prizes ranging from $130,000 to $20,000, may also compete for two additional prizes of $30,000 and $20,000. This initiative is a sort of competition within a competition. The Agritech participant must express their intention to participate during round 1 of the BCIC New Ventures competition. The jury will evaluate the projects and choose the top two companies. It appears that selection of Agritech Innovation Award recipients is similar to BCIC New Ventures, but exact assessment and selection process as well as participation requirements are not clear from information available in the public domain. It is possible that the entrepreneur may win the Agritech Innovation prize in addition to major BCIC New Ventures competition prizes. In the case when none of the BCIC New Ventures competition winners deserve the Shildroth Agritech Innovation award, it is not awarded.  2.5.2. BCIC CAT Competition In 2009-2010, the British Columbia Innovation Council (BCIC) held the Commercialization of Agricultural Technology Competition (CAT), which was intended to help commercialization of new technologies particularly in BC’s agricultural sector. CAT was announced as an annual competition specifically designed for entrepreneurs in BC whose innovation dealt with agriculture, food and bio products. It was organized in two stages. According to BCIC Commercialization Agricultural Technology (CAT) Competition Guidelines, the first stage is Expression of Interest (EOI) which results in allocation of up to 25 $10,000 vouchers to be used to expand EOI into a full-length Business Plan. The eligible  36 innovations must be based on science and technology; they need to be commercialized within two years starting at the end of the competition, and developed to improve environmental and economic sustainability.  In the next stage, the Proof of Concept Competition, the top 10 business plan applicants will make a presentation to a panel of evaluators and four applicants will win Proof of Concept awards ranging from $100,000 to $250,000. The winners of this competition have to match an award by their own contribution. Out of the 60 projects initially entered CAT in 2009, 22 were awarded $10,000 (interim) prizes to be used for business plan development. Eventually, at the beginning of 2010, BCIC awarded four prizes in the announced size to four BC based firms who won the 2009 CAT competition. The winning projects dealt with non-diary slice technology, innovative juice processors, fruit packaging and sunlight-based plant micro-propagation facility development. In a commercialization competition such as CAT, the public agency is obligated to pay the announced award and cannot change the size of the prize. For the 2009-2010 CAT, the funding was provided by the Ministry of Agriculture and Lands and The Ministry of Small Business, Technology and Economic Development.  Similar to the BCIC New Ventures competition currently administered by BCIC, selection of the final CAT awards recipients is a subjective and secretive process. The BC Innovation Council CAT competition Guidelines (British Columbia Innovation Council, Life Sciences, 2009) clearly articulates that BCIC will appoint an evaluation panel for the Expression of Interest stage of competition, which is the first stage of the competition resulting in $10,000 voucher allocation. This panel makes recommendations to the BCIC Agriculture, Food and Bioproducts Innovation Fund Advisory Group, who will determine the voucher winners. As for the final stage of the competition, the Guidelines provide no information about how the  37 decision of the major prizes recipients will be made. From a conversation with top BCIC executives, it was learnt that the final decision in the 2009 competition was made by a confidential panel of experts who were government employees and specialists from the industry, none of whom were from BCIC. We were not able to obtain specific documented information about how winners were selected at the final stage. The CAT Guidelines indicate that the “exact value of each award will be based on information provided by the award recipient and on the amount matched by the recipient.” In situations when the expert panel is aware of financial resources available to each potential winner, as well as financial commitment to the project, this creates the possibility of time inconsistent decision. Thus, it was shown that most federal and provincial business R&D support programs, in the form of direct assistance, carry similar features. Particularly, the project assessment process is often a multi-stage one, and application assessment/approval mechanisms as well as distribution of funds is made based on subjective decisions by designated individuals (experts, clerks, etc.) who are totally aware of the participant’s financial situation. Sometimes the decision process is not clear or not explained at all. Several commercialization programs administered in British Columbia in unique competition formats were discussed.         38 Chapter  3: Analytical Framework: Game-Theoretic Model of Commercialization Competition This chapter is dedicated to construction of the commercialization competition model. A formal theoretical model encompassing four areas of analysis, including public agency problem, firm problem, firms’ problem under commercialization competition, and distortions potentially arising in this type of competitions was built. Section 3.1 presents a review of relevant literature on research tournaments and the time inconsistency problem. Section 3.2 introduces a two-stage game. When public agency can commit to an announced winner selection process, the situation can be modeled as a simple dynamic game of complete information which is presented in section 3.3. The stages of the game are: (1) public agency announces and commits to contest rules including size of the prize and criteria for winner selection; (2) the firms form expectations of their profits based on the size of the prize and stochastic shock variable and make innovation investment choices. The game is solved by backward induction. Finally, as a result of the firms recognizing potentially time inconsistent winner-selection process, the reduced innovation effort of the firms is derived.  3.1 Review of Relevant Research Tournament Literature  Tournaments and competitions are very popular tools used in situations when a principal is not able to estimate actual absolute efforts of contestants, and instead uses their relative performance as a basis for making decisions about the winner. Broad analysis of situations when government might employ tournaments for policy implementation is delivered by Prendergast (Prendergast, 1999), who also points out that tournaments preclude agency from reneging on paying the announced reward.  39 The tournament theory was pioneered by Lazear and Rosen (Lazear & Rosen, 1981) and views rank-order payments as an alternative to payments secured by labor contracts. They studied how the relative performance of workers can result in wage differences. The authors considered a group of workers competing for some set of rewards. To win these pre- specified rewards, the workers exert efforts and effort level determines the probability of winning the highest reward. Consequently, the compensation of individual output of the workers can be very well approximated by a less costly alternative based on their relative performance. Thus, the wages were viewed as prizes. The authors introduced heterogeneity of participants and showed that tournaments elicit effort response. Particularly, workers’ efforts increase with the spread between winning and losing prizes. Further research in this direction testifies that the workers’ effort is increasing in size of prize and in the efficiency of monitoring. The first relation was widely tested empirically using non-experimental data and reported by Ehrenberg and Bognanno (Ehrenberg & Bognanno, 1990). The research that followed the famous work by Lazear and Rosen (1981) mostly addressed the way heterogeneity in contestants and outcomes can be modeled. Green and Stokey (Green & Stokey, 1983) and Nalebuff and Stiglitz (Nalebuff & Stiglitz, 1983) generalized and extended findings of Lazer and Rosen (1981). The assumption in these works is that the contestants are identical in their abilities and uncertainty stems from possible error when it comes to measuring their output. It means that each contestant will choose the same level of efforts but will receive different levels of reward due to uncertainty. On the other hand, Glazer and Hassin (Glazer & Hassin, 1988) do not allow uncertainty when ranking the contestants’ output. Instead, assuming contestants are unaware of what level of output their competition had chosen, the equilibrium level of efforts of contestants  40 will vary. They also studied tournaments where contestants have different levels of abilities and demonstrated that under such a setting only a few of the prizes should be greater than the reservation wage.  Since Lazear and Rosen (1981), the mechanism of fixed-prize tournaments was widely explored with application to labor economics and international trade (Coughlan & Schmidt, 1985; Stein, 1997) Within the extensive literature on contests and tournaments there is a relatively limited set of economic papers that applied tournament mechanisms to area of R&D. Taylor (Taylor, 1995) studied the optimal design of research tournaments by applying a mechanism of a two-stage game with imperfect information. He showed that unique subgame-perfect equilibrium exists in this game. This paper focuses on the competition among symmetric innovating firms: the agency benefits from buyer power as the innovating firms are competing against each other. Taylor’s main finding is that the contest suffers from underinvestment in firms’ efforts. To address this issue he advocates restriction of the pool of contest participants, offering a prize of an optimal size and implementation of an entry fee that will allow the buyer to extract all expected surplus. Fullerton and McAfee (Fullerton & McAfee, 1999) developed these findings further, showing that the optimal number of contestants in fixed-prize research tournaments is actually two. They developed a tournament model with heterogeneous contestants, which allows the selection of best qualified contestants for innovation competition. They questioned the possibility of setting optimal entry fees in real life and emphasized limitations and drawbacks of entry fees. The authors came up with an alternative way of selecting the best qualified participants – the contestant-selection auction, which is a variant of all-pay auction.  41 As was discussed in Section 2.5, the notion of entry fee is incorporated in the format of the New Ventures Competition and CAT Competition administered by the BCIC. To be able to compete in the New Ventures Competition, entrepreneurs pay a $200 entry fee. Fullerton and MacAfee (1999) emphasize that in order to collect the entry fee public agency might choose to avoid monetary bids, but rather have participants to incur the sunk cost of entering the competition. In the case of CAT, the competition is organized in two stages in order to increase the selection efficiency. At the first stage, the potential participants express their interest through submission of summaries of their innovations. Next, the winners of this stage will get a monetary interim prize intended to be used for the business plan development. Finally, the contestants selected at the previous stage compete for the primary tournament prize (usually a set of several prizes) by developing and presenting the innovations to an independent panel of judges. Thus, the investment made by participants into development of the Expression of Interest or summary of innovation serves as an entry fee in the case of the CAT competition. In this way, entry is always weakly preferred to non- entry for every potential contestant.     Terwiesch and Xu (Terwiesch & Xu, 2008) show that the decrease in contestants’ efforts resulting from contest inefficiency can be mitigated by restructuring the fixed-prize tournament into a performance-contingent one. As a response to Fullerton and McAfee (1999) findings, they show that a large pool of contest participants can actually be beneficial to the contest organizers since they get a wider range of innovations, and obtained benefits may outweigh the losses from inefficient equilibrium of R&D efforts. Fu and Lu  (Fu & Lu, 2009) support this finding and note that an equilibrium effort does not necessarily decrease  42 when the number of contestants is low. Rather, it is always optimal to shortlist them when the pool of contestants is large. Che and Gale (Che & Gale, 2003) and Fullerton et al. (Fullerton, Linster, McKee, & Slate, 2002) wrote two widely cited papers on innovation contests demonstrating that first- price auction performs better than fixed-prize tournaments when identical innovators follow symmetrical strategies. Che and Gale used deterministic innovation technologies, whereas Fullerton et al. explored a stochastic one. In contrast to their results, Schoettner (Schoettner, 2008) shows that under different assumptions on the firm’s innovation process the procurer may actually prefer the innovation contest without entry fee to be organized as a fixed-prize tournament. Much of the existing literature on innovation competition treats the prize as exogenous (Taylor, 1995), (Fullerton & McAfee, 1999). However, there are models where the prize was internalized in order to capture the profit-maximizing nature of R&D activity of the firms (Evenson & Kislev, 1976), (Baye & Hoppe, 2003). Given the nature of research question in this thesis, particularly when the innovation contest prize is predetermined by the public agency, this set of literature is not directly related to the topic of this research.  3.2  Time Inconsistency in Economic Literature Besides the literature on research tournaments, this thesis draws from another no less outstanding line of research on inconsistency of optimal plans. It was originated by 2004 Nobel Prize winners Finn E. Kydland and Edward C. Prescott with their famous work on time inconsistency of economic policy. The idea offered by these two outstanding scientists is that policy makers are tempted to renege on originally announced policies in order to  43 capture short-run benefits. Consequently, the crucial role of credibility in making the policy successful is emphasized. The pioneering work of Kydland and Prescott  (Kydland & Prescott, 1977) showed that the policy meant to be optimal at the beginning might not be optimal anymore in later periods. In response to how many economists perceived inconsistency to be a result of changing preferences of the agents or inefficiency of information distribution, Kydland and Prescott demonstrated that it can arise simply because of inability of the policymakers to commit to announced policy. Kydland and Prescott conclude that time inconsistency, which often results when policy is a discretionary one, always results in suboptimal social welfare. Blackburn and Christensen (Blackburn & Christensen, 1989) produced an extensive review of the literature on policy credibility, stressing that credibility is a much broader notion than time inconsistency. Policy-making normally lends in game-theoretic interpretation, as was pointed out by the authors, and this approach is adopted in this research. The problem of time inconsistency was most often studied with application to public finance, fiscal and environmental policies, and the private business sector. There exists an extensive set of papers looking at time inconsistency of optimal monetary policy. This strand of literature builds on a widely cited paper by Calvo (Calvo, 1978). He shows that in the absence of lump-sum taxation, the government that optimizes welfare may have incentives to deviate from announced monetary policy towards a higher rate of money supply resulting from a “disharmony between governments’ and individuals’ objectives”. Furthermore, “Any economy where individuals are sensitive to the announcement of future policies has, in principle the seeds of time inconsistency” (Calvo, 1978). This generalization can easily be  44 adopted for a commercialization competition environment where the ultimate goal of hosting agency and participating agents differ. Several authors studied the issue of time inconsistency with application to a wide range of problems from the area of environmental economics. Golombek et al. (Golombek, Greaker, & Hoel, 2010) notice that innovative firms make investments into climate-friendly technologies based on announced climate mitigation policy. But after investments are made, government has an incentive to renege on the policy causing a time inconsistency problem. As a solution, the author suggest implementation of a first-best subsidy. Gulati and Vercammen (Gulati & Vercammen, 2006) point out time inconsistency inherited in some resource conservation contracts. They point out that agents who are required to reserve resources according to the temporal contract with a principal, in fact have different incentives before and after the contract is signed providing him with incentive to exit the contract earlier. The principal recognizes the potential of this time inconsistent outcome and responds by offering a contract that is shorter than optimal. The authors suggested two alternative ways of dealing with this distortion, such as the use of a penalty for early termination of the contract or an upward slopping conservation payment schedule. Helm, Hepburn and Mash (Mash, Helm, & Hepburn, 2003) analyze  the problem of carbon policy and, similarly to the case studied in this thesis, demonstrate that optimal government policy is different ex ante and ex post. They notice that government has a dual objective and consequently may set environmental taxes that encourage firms to make irreversible investment decisions and later renege on initially announced tax policy.  The authors also show that with discretionary government policy output, subsidy does not resolve the issue.  45 In competitions such as New Ventures and CAT, there is a very high chance that one or several projects are obviously more socially promising than the others. The superiority of the projects may stem from either the amount of investments made by a participant or area of project application, etc. Prendergast (Prendergast, 1999) pointed out that in tournaments where one contestant has a higher chance of winning (biased tournaments) the resulting productive effort is lower than the optimal effort in non-biased tournaments. In extreme cases, this causes some contestants to withdraw from the tournament completely. Summarizing the review of literature above, we conclude that up to data there was not a single study focusing on time inconsistency of research or commercialization technology competitions. Although an enormous amount of academic effort has been devoted to studies of different technical aspects of labor and research competitions, distortions arising under some circumstances are lacking researches’ attention.  3.3 Model of Commercialization Competition The model presented in this section is a two-stage game with imperfect information. One party of this game is a risk-neutral public agency that holds a commercialization competition. By doing so, the agency avoids high monitoring costs usually attached to project assessment, and creates   incentives for entrepreneurs to invest into innovation projects and carry all the associated risks. The game also includes two risk-neutral entrepreneurs who exert innovation effort in anticipation of a pre-announced award. They make their investment decisions based on their assessment of the competition credibility, i.e. their realization of the outcome of the competition. Sub-section 3.3.1 begins with derivation of optimal innovation effort from an expected social welfare maximizer’s point of view. Sub-  46 sections 3.3.2 and 3.3.3 show the optimal efforts of entrepreneurs when innovating both with and without the commercialization competition. Finally, sub-section 3.3.4 demonstrates how the anticipation of a time inconsistent award allocation decision affects the level of entrepreneurs’ investments into innovation.  3.3.1. A Public Agency’s Problem At stage zero of the game, consider a competition held by a public agency in order to assist the commercialization of the most valuable business innovations. The public agency announces that it will pay an award of size P to the entrepreneur whose project, after successful commercialization, would have the highest social value attached to it. Since entrepreneur’s innovative efforts are unobservable and the outcome is generally not verifiable, the public agency can use a prize to induce entrepreneurs to invest into innovation accounting not only for the private value of the project but also for its social importance. To capture this aspect, a parameter    (     ) representing marginal external social benefit from the project, and a parameter    (     ) reflecting its marginal private market value were introduced. Let K (     ) be commercialization cost of the innovation project. The investment choice of the entrepreneur is presented by a quadratic function C=  where β >0 indicates the degree of its convexity and   is entrepreneur effort. The cost function of the entrepreneur is strictly increasing, strictly convex and twice differentiable for all       Next we introduce a random shock component    that is uniformly distributed and is intended to reflect the uncertain quality of the innovation idea.  47 A public agency considers a project socially desirable when its value W is positive. Social value is the sum of private and external value,(   )  plus random quality,  minus commercialization cost K. Thus,    requires:   (   )                         (3.1) Equation (1) can be rewritten as:  eK )(                           (3.2) The expected social surplus is expressed by:  ( )  ∫ [(   )     ]       (   )                               (3.3)  The integrated terms is a measure of the expected value of the project conditional on     (i.e. the project is launched). Equation (3.3) evaluates to:  ( )  [(   )   ][    (   ) ]   [   [  (   ) ] ]    =   [  (  (   ) ]                            (3.4) The level of innovation effort that maximizes expected social surplus solves:   ( )   Thus, [  (  (   ) ][   ]                                         (3.5)  Solving (5) for e results in the level of the firm’s innovation effort which is considered optimal from the public agency’s standpoint:     (   ) (   )                  (3.6)  48  Notice that     is an increasing function of private and external marginal value    , and is a decreasing function of the marginal cost of innovation effort,  , and commercialization cost K.  3.3.2. Entrepreneur’s Optimal Effort when Commercializing Innovation without Government Assistance Now assume that it is the firm rather than the agency which chooses commercialization effort e. At stage one of the game, a firm has developed an innovation investing        into the project. In order to commercialize it, the firm has to pay commercialization cost K. The firm makes a decision whether to proceed with commercialization at stage two of the game after observing the random quality component which is uniformly distributed,  ~[0,1] .  The problem that this firm faces is essentially the same as that facing the public agency except the firm ignores the external value of the innovation project  . The stage 2 private value of the firm’s project can be expressed as: V= Ke                       (3.7) where    is a private market value parameter capturing the marginal profitability of the project. The firm will commercialize its innovation at the stage 2 if V≥0, or when ε eK                       (3.8) Thus, the solution of optimization problem is the same as that of the agency with )1(* 2 Ke                                    (3.9) The optimal amount of investments in innovation from the firm’s standpoint and the public agency’s standpoint are different. Comparing (3.6) and (3.9) gives:  49   (   (   )) (  (   ) )(    ) (   )                                                                                (3.10) Intuitively, when the project has a positive social benefit  , the socially desired level of innovation effort     is always higher than the privately optimal level   . In other words equation (3.10) is always positive. A formal explanation of this conclusion is contained in the Appendix C.  3.3.3. Entrepreneur's Optimal Effort if Participating in Commercialization Competition In this sub-section a two-firm commercialization competition is considered. Firms are competing for a fixed prize P, which will be allocated to the most socially valuable project according to the competition rules announced at stage one of the game. If the award is sufficiently large, at stage two of the game each firm decides to participate and makes decision about the level of private R&D investments denoted by   . As before, at this stage the level of capital allocated by each firm is combined with a stochastic random shock variable    , which influences both the level of private R&D capital invested into the project development and the expected post commercialization value of the project. The assumption is made that the environment in which the two firms compete is not oligopolistic, but a competitive one since the players do not know who their opponents are.  For the derivation of the optimal level of innovation investments for the firms participating in commercialization competition, we use Cournot’s model (1838) and rely on the Nash equilibrium concept. In the original Cournot model, firms choose the level of output and the equilibrium establishes the price.  In the model offered here, each firm independently chooses its investment strategy to maximize profits from innovation and takes the optimum  50 strategy of its opponent as given. Since in current setting the firms have no influence over the competition flow, the Nash-Cournot approach can be applied. The competition represents a highly stylized market for innovation. For the purposes of proposed model, innovation is viewed as a homogeneous good. The profit function of each firm is a difference between revenues and costs. Each firm chooses its level of R&D investments to satisfy the usual first order conditions. The equilibrium level of effort is decreasing in the investment of the other firm, so if a firm chooses to increase its effort, the other firm is motivated to decrease its effort. In the Nash equilibrium, both firms have an identical level of optimal effort because the reaction functions are symmetrical. Firm 1 wins the prize of size P if the following two conditions of “feasibility” and “sociability” hold: 1) Feasibility condition V= 011  PKe    or                                (3.11) For convenience, superscripts 0 and 1 are used to distinguish between situations “no prize” and “winning the prize”. Subscripts are used to distinguish between firms 1 and 2. If the random quality term for firm 1 is denoted by 1 1 , then feasibility holds if 0 1 1 1   , where 1 0 1 ePK   . The feasibility restriction holds if the economic conditions are sufficiently favorable (i.e. the value of stochastic term must be large enough). Particularly,    should exceed a threshold level determined by the commercialization cost, the value of the prize and the value of the project. 2) Sociability condition 2211 )()(   ee                            (3.12)  51 This condition specifies that for the firm 1 to win the prize the social value of its project must be higher than the social value of the firm 2’s project. The public agency’s decision about which firm receives the prize is based on this condition. From equation (3.12) it follows that firm 1 wins the prize if )( 2 1 1 1 1   , where: 2212 1 1 ))(()(   ee                            (3.13) Notice that  (  )   if  (   )(     )  , or    (   )(     )  . In other words,  (  )   provided that     where   (   )(     )  . Thus, the firm 1 wins when:    and  (  ) This part of expression can be combined and re-written in integral notation. Thus, the probability of the firm 1 winning the competition prize is:     (     )  ∫ ∫    (  )     ∫ ∫                    (3.14) The first term corresponds to the case where meeting sociability automatically satisfies feasibility and the second term corresponds to the opposite case. Let  (  ) denotes total expected surplus for firm 1 which is comprised of the expected profits and the expected prize:  (  )   (  )      (     )                 (3.15) The expression for the expected profit  (  ) is the same as in (3.3) from the public agency optimization problem with    . Thus, making the appropriate substitution gives:  (  )  ∫ [(        )     ]   [∫ ∫    (  )     ∫ ∫       ]   (3.16) where: 1 0 1 ePK   ,    (   )(     )  , and  52   (  )   (   )(     )  . Firm 1 chooses the level of innovation effort 1e to maximize social surplus given the opportunity to win the prize. Then, first order condition is:   (  )     (  )    (    (     ) )                 (3.17) Let’s consider the first and second terms in (3.17) separately:   (  )    (       )                     (3.18)  (    (     ) )    [(   )(     )    (       )     (     )   (   )]                                  (3.19) Thus, combining (3.20) and (3.21), the first order condition becomes:   (  )   (    )   (   )   [(   )(     )    (       )    (     )   (   )] Solving for firm 1’s best response function, it follows that:    (     )       (        )  (   )   [(   )(     ) ( (   )]       (        )             (3.20) When P=0, the result in (3.22) is exactly the same as obtained in (3.9) when there is no innovation contest and the innovation does not carry additional social benefit value but only private market value. As was discussed in sub-section 3.3.2, 0 2  , thus      . Intuitively, the reaction function of firm 1 must be upward slopping.  If firm 2 raises its effort, then probability of firm 1 winning the competition goes down. If firm 1 decreases rather than increases its innovation effort, then the probability of winning would go down  53 even further. Hence, it must be the case that firm 1 increases its effort in response to an increase in effort by firm 2.  Let  (     )       (        )   and (   )   [(   )(     ) ( (   )]       (        )  , then the Best Response Functions can be written in familiar linear form: 21 BeAe   and 12 BeAe  Inverse firm 2’s reaction function to get: 12 1 e BB A e   The Nash equilibrium level of effort can be derived by jointly solving the pair of reaction functions, as shown in the following set of equations: )( 11 BeABAe  )1()1( 21 BABe  A B B e 21 )1( 1     ) 1 1 1( 22 B B BAe    = A B B 21 1    So, eventually:        for   {   }                  (3.21)  The graph presented in Figure 8 shows best response functions for each firm. It also shows how increasing the level of prize shifts out the reaction curves and raises the equilibrium level of innovation effort.  54 Figure 8: Relationship between Prize Size and Firms’ Best Response Functions             3.3.4. Time Inconsistency This section discusses how the agency’s deviation from the ex ante announced rules may distort optimal innovation efforts of the firms. This distortion is a direct result of the time inconsistency problem. At stage three of the game, the agency makes the award decision. If the random outcomes are such that neither firm has a project that can be successfully commercialized after the award is paid, then the award is not paid. If both firms require the award to commercialize their innovations then the firm holding the project with the highest social value earns the award, and the outcome of the competition is ex post consistent. The interesting dimension can be observed in a situation when the public agency is aware of the Rx Firm 2 Rx Firm 1 𝑒  𝑒  Response for higher P  55 fact that the random outcomes are such that one firm with the best project has generated sufficient external capital to finance its commercialization without the award, and the other firm with the second best project is in position to commercialize it only if the award is received. According to ex ante criteria of the competition, the firm with the best project should be awarded a prize. However, in this particular situation the public agency can raise social welfare by allocating the award to the firm with the second best project because this will result in both projects rather than one project being commercialized. In stage one, consequently, the firms will recognize that the competition may be time inconsistent and will allocate a lower than optimal level of private investments towards their innovation projects. It is also possible that some firms may choose to not participate in the competition, thus compromising the idea of government support being available to those who need it. The agency will be required to raise the award amount to compensate for time inconsistency, and this will eventually raise the overall cost of delivering the program.  As was discussed in Chapter 2, most real world business innovation competitions, including CAT and New Ventures, require participants to reveal their current and past financial situation as well as the amount of financial resources committed to the project. This increases the agency’s buyer power and creates a potential for distortions. The economic literature views research competitions as a tool to be used when the innovation input and output values cannot be verified by the agency. In the real world, this does not hold. For the purpose of modeling, suppose firm 1 has the most socially valuable project and further suppose that firm 1 can commercialize it without the prize. At the same time, firm 2 has a socially valuable but privately unprofitable project due to lack of own finances to commercialize it. The framework for deriving optimal effort of the firm 1 under the presence  56 of time inconsistency in winner selection process is generally the same as in the previous sub-section, where the problem of competition between the two firms was considered. What makes modeling in this sub-section different is the probability of receiving the prize. Under time inconsistency, the expression for probability of receiving the prize is different, and captures the whole spectrum of situations including those when the firm 1 deserves to be awarded the prize, but the prize is allocated to the firm 2 in order to assure that both innovations, firm 1’s and firm 2’s, are eventually commercialized. With time consistent competition outcome, the probability of winning the prize is:     (     )  ∫ ∫    (  )     ∫ ∫                        (3.22) When time inconsistency is present, some probability mass must be subtracted from this function to account for the fact that in scenarios discussed at the beginning of this sub- section, firm 1 satisfies both feasibility without the prize condition and sociability condition, while at the same time firm 2 satisfies feasibility with the prize but not without the prize. By allocating the prize to firm 2 instead of firm 1, the public agency achieves its goal of raising social welfare by making both firms able to commercialize their innovations. This effect can be shown using the Reaction Curve graph (Figure 9). The solid lines on the graph represent the firms’ best responses under time consistent commercialization competition. The dotted lines show that the optimal innovation effort of the firms is lower if prize allocation is time inconsistent. Time inconsistency shift reaction curves in, and lowers equilibrium innovation effort of the firms.   57 Figure 9: Effects of Time Inconsistency on Firm’s Innovation Efforts    Rx Firm 2 Rx Firm 1     Impact of time inconsistency 𝑒 (𝑇𝐼)  𝑒 (𝑇𝐼)  𝑒  𝑒  58 Chapter  4: Experimental Study of Fairness versus the Greater Good in Winner Selection Process of Commercialization Competitions This chapter examines theoretical aspects of experimental research in economics and describes the construction of an experiment that allows testing for the fairness versus the greater good in judges’ decisions at a hypothetical commercialization competition designed after the BCIC CAT competition and the BCIC New Ventures competition. It offers a formulation of the hypothesis and discusses an empirical way to test for it. It describes scope conditions of the proposed experiment and translates the hypothesis into concrete operations. It also provides a description of experimental treatments and manipulations.  4.1 Motivation behind the Experiment   It was noticed that the possibility for competition judges to deviate from the original set of decision-making rules, which constitutes the essence of time inconsistency problem, involves a competition fairness component. Particularly, the judges face an ethical trade-off between the competition individual fairness and the maximization of the greater good. One of the research objectives is to establish the outcomes of this trade-off. The most natural way to demonstrate this possibility is to use real world data about all participated and awarded projects over the years and estimate the consistency of awards allocation. Examples of data required for this type of analysis would be the actual data used by the agency for selecting the winners at the final stage of the competition: - expected market value of the project; - the initial investment into innovation project done by entrepreneur on the moment of application;  59 - amount of money that has been raised by entrepreneur towards the project from sources others that the BCIC program; - the amount of money needed for the successful commercialization of the project; -  the list of the winners. Unfortunately, this data is not available and this makes studying research tournaments harder. As was noted by Cozzarin (2008) and other authors that have attempted to study different aspects of R&D programs in Canada, the main stumbling block for conducting any type of analysis is data availability. It is especially true when a researcher is interested in the decision-making process. Statistics Canada does not collect this type of data. Another source could potentially be audit reports. Although it is possible to find audit reports for larger federal innovation/commercialization programs in the public domain, none of them are publicly available for BCIC programs. When contacted in 2010, BCIC informed us that it keeps record of all the participating projects and the assessment process in the CAT competition, but would not supply this data due to high confidentiality. Conducting interviews with members of expert panels could potentially contribute to the understanding of the decision-making process in BCIC CAT and BCIC New Ventures. However, this probably would not supply sufficient amount of information to conduct a statistical study. Also, identity of the panel members are held secret, which makes the publication of these interviews impossible. Additional concern about the necessity to interview panelists was raised because of the fact that the final decision about prize allocation lies with BCIC. Eventually, absence of data brought up the idea of conducting an economic experiment in order to simulate the decision-making process of a jury and collect enough  60 information to test for the outcome of fairness versus the greater good trade-off during a prize allocation process.  The results of this study will inform public agencies by emphasizing the aspects of the competitions that are likely to cause inconsistency and, if recognized by participants, will likely result in lower innovation effort.  4.2 Social Welfare Function and Individual Fairness Concept with Relation to Commercialization Competition In this section, the notion of individual fairness and social welfare functions are reviewed based on the analysis of existing economic literature. Next, the model of Social Welfare versus Fairness is presented with application to the commercialization competition. We start by reviewing how the problem of individual fairness was addressed in economic literature. The first significant step towards theoretical modeling of individual and social preference relations was made by Harsanyi (1955) who proposed the utilitarian social welfare function   ∑      as a weighted sum of individual utilities    . He assumed that the individuals and the social planner use von Neumann-Morgenstern expected utilities to evaluate risky outcomes and that the Pareto principle holds (Harsanyi, 1955). The model proposed by Harsanyi was criticized by Diamond (1967) for not incorporating fairness considerations on the part of the social planner (Diamond, 1967). In essence, the Diamond’s argument indicates that in situations when the social planner may be indifferent between two alternative allocations, but the individual agents have strictly conflicting preferences as for these allocations, the Harsanyi’s model is not able to capture the aspect of fairness, and thus it is strictly preferable to use a fair lottery to determine the fair allocation (Karni, 1996).  61 To address the issue of individual fairness being incorporated into Harsanyi’s social welfare function, S. Trautmann developed a notion of all-inclusive utility. He assumed that the description of social allocations already include individual preferences for fairness and possible inter-agents comparisons. Thus, all-inclusive utility function allows accounting for the individual fairness preferences while the expected utility nature of the function is retained and the Pareto-efficiency assumption is satisfied (Trautmann, 2010). In particular, Trautmann proposed a two-stage model which is based on Sugden’s (2000) approach. At the first stage, agents evaluate individual outcomes of risky options based on self-interested von Neumann- Morgenstern (vNM) utilities. These can be utilities of different health states or different possible outcomes of an innovation activity. At the second stage, these self-interested vNM utilities are incorporated into all-inclusive vNM utilities and become a basis for social welfare evaluation. The author applies Fehr and Schmidt (1999) model of individual fairness and its modifications at this stage. The type of the model of individual fairness chosen at this stage determines how individual fairness is accounted for and is crucial for the preciseness of the model. It is important to realize that for the purpose of the modelling, Trautmann uses the equality principle as the definition of fairness. Obviously, in many real world situations the “fair allocation” principle will go beyond this definition and may become very complex. For example, Gachter and Riedl (2006) address the issue of fair distribution of an estate funds and applies three different fairness norms at the agent level. These norms differ in the size or proportion of the estate share allocated to the agents. One of the fairness norms assumes allocation of funds in proportion to the agents’ claims (Gachter & Riedl, 2006).  This thesis presents a model of trade-off between Social Welfare maximization and Fairness of commercialization competition. Let   denote an ex ante prize allocation rules of  62 the competition, and  ̂ represents the ex post rules. Based on the ex ante rules, entrepreneur i makes innovation investments    to maximize its expected utility. The ex ante optimal innovation effort of the entrepreneur i,  ( ), is obtained from the following optimization problem:       ( (    )                   (4.1) The set of optimal innovation efforts of all participating entrepreneurs results in an ex ante distribution of prizes    , where total amount of prize money allocated in the competition is:  ̅  ∑                         (4.2) The prize allocation based on ex ante rules does not maximize ex post social welfare since prizes may be awarded to the entrepreneurs that can commercialize their projects even without the prize money, whereas other projects if not awarded a prize, may be potentially lost for the society. Thus, ex post social welfare is maximized if fairness of the competition is ignored and the new prize allocation rule  ̂ is applied, which results in ex post allocation of prize money     , where  ̅  ∑     . The fairness of prize allocation to a particular entrepreneur is the difference between ex ante and ex post prize allocated to him/her, and is captured by a parameter    . Then the maximum unfairness of the prize money allocation in the competition is:      ∑                        (4.3) This ex post prize re-allocation results in social welfare gain:       (     )   (    )               (4.4)   63 Figure 10 depicts the relationship between the level of unfairness of individual judge, whose ultimate goal is maximization of social welfare, and the resulting social welfare gain. Figure 10: Trade-off between Competition Fairness and Efficiency Gain         At the origin, the level of unfairness is minimal, and social gain is zero. In other words, at origin, an individual judge’s prize allocation decision completely complies with the ex ante prize allocation rules. An individual judge can choose an allocation of prizes anywhere between the perfectly fair and maximum unfair outcome of the competition. This deviation from the ex ante prize allocation rules results in positive incremental social welfare changes. When making prize allocation decisions, an individual judge trades-off competition fairness to the “greater good” based on individual “social welfare gain versus competition fairness” indifference curves (Figure 11). The tangency of this curve to the feasible choice determines how much he/she is ready to deviate from the original decision-making rules in order to achieve social welfare gain. From the Figure 11, this particular competition judge is Level of Unfairness of the ex post prize allocation, F Soc  l w l  r 𝐹𝑚𝑎𝑥  𝑚𝑎𝑥 Trade-off between fairness and efficiency gain  64 willing to incorporate    of unfairness into his final decision which results in    social welfare gain. Figure 11: An Individual Judge Trade-off between Social Welfare Gain and Fairness of the Competition    ∆^0     As was shown, the issue of individual fairness is related to the problem of time inconsistency. The experiment, the description of which follows in section 4.4, studied the issue of fairness of judges’ prize allocation decisions when they strive to achieve the “greater good” by maximizing social welfare. But first, a brief retrospective look at experimentation in economics is taken in section 4.3.  4.3 Economic Experiments in the Literature Although economists started to fragmentarily use experimentation as a research tool a long time ago, formal records of their successes on these grounds started to be available as of the middle of the last century. With the works of Allais, Friedman, Stone, Smith and others the whole new era in economic research was started (Roth, 1988). At that time, Level of Unfairness of the ex post prize allocation, F Soc  l w l  r 𝐹𝑚𝑎𝑥  𝑚𝑎𝑥 Trade-off between fairness and efficiency gain Indifference curve of individual judge  𝐹  65 experimentation was mostly concerned with market mechanisms and public goods. Particularly, according to Ross, most of the literature dealing with economic experiments is concentrated on several general areas: experiments modeling two-person bargaining, experiments studying free-rider problem (including prisoner’s dilemma), individual choice behavior experiments and experiments investigating different types of auctions (Roth, 1988). During recent decades, experimentalists continued questioning rationality of market agents through their work. Vernon Smith and his followers set a new trend and extended economic experiments into the area of economic institutions initiated by Vernon Smith and his followers. Further development of experimental practice in economics in the 1990s shifted research focus to studying situations “when rationality and information processing are less than perfect, and when social considerations such as fairness… play an important role” (Eckel, 2007). The experiment presented in this thesis follows this fascinating trend. As will be explained further in more detail, the one way to see the main focus of the proposed experiment is actually to look at it as the matter of fairness and economic rationality under condition of a non-transparent decision-making environment. This type of environment creates imperfect flow of information and results in distortions that were modeled in the previous chapter. Commercialization competitions in the format of the BCIC-New Ventures Competition and BCIC-CAT that are subject of this study, by their nature, are very close in several aspects to all-pay auction. The bidders, which in this case are participating firms, submit their “bids” in the form of market value of their innovations to the “auctioneer”, BCIC. After the “bids” are compared, the participants with the highest “bids” are announced to be the winners. As in the all-pay auction, the participants submit their valuations based on  66 presumed market value of their projects. Of course, in the case of BCIC programs, the participants do not directly pay their bids to the “auctioneer” but they sunk innovation cost of their projects before entering the competition. The “auctioneer” possesses a lot of power over the outcome of the auction, which makes it even more intricate. Much of the existing literature on economic experiments is dedicated to format and level of efforts in auctions. Researchers recreate the auction in the lab and investigate predictions of the theory. For example, Dufwenberg and Gneezy experimentally tested a first-price sealed-bid auction under three levels of information disclosure. They compared the results with Nash equilibrium outcome (Dufwenberg & Gneezy, 2002). Karla and Shi (2001) confirmed that in competitions with a fixed budget, higher effort of a participant can be elicited in multiple prizes format rather than in winner-take-all format. At the same time, interesting experimental results were obtained by Schotter and Weigelt (1992). They found that in two-contestant tournaments, which are asymmetric in terms of contestants’ endowments, their effort is lower than in symmetric tournaments. They also showed that contestants who have initially lower endowments or are disadvantaged in any other way, tend to overinvest. When applied to the BCIC New-Ventures and BCIC-CAT competitions, these findings indicate, that it is reasonable to expect increased innovative effort of the participants due to multiple prize format used by BCIC. At the same time, due to asymmetric nature of these competitions, over exertive effort of the participants with a priory fewer resources will probably occur.  Allocation of nonrefundable grants (R&D subsidies) through some sort of competition is a common practice in the US, Germany, and the UK (Giebe, Grebe, & Wolfstetter, 2006). Giebe et al. focus on two deficiencies of this process. The first is  67 allocation of funds based on ranking of the projects instead of ranking of the complete allocation of funds. The second concern of the authors is excessive funding when projects are funded based on predetermined percentage of refundable project cost. They propose a particular auction mechanism that would eliminate these deficiencies and broaden the pool of potential subsidy recipients. Although the focus of this study is rather different, experimental study of public programs providing support to socially valuable projects that otherwise would not be feasible, closely relates this paper to the research interest of this thesis.  Next, a brief review of the literature used in determining the format of the experiment is presented. The respondent’s preferences are often elicited by choice modeling, which includes several techniques. The respondent may be asked to simply indicate the most preferred option, rank the alternatives given for his/her consideration, or score the options based on the provided scale. All these approaches were originally used in marketing for eliciting consumer preferences (Morkbak, Christensen, Gyrd-Hansen, 2010). During the last decade, they became well appreciated in health-care research estimating patients’ preferences towards treatment, service or medication options (Park et al., 2011, Van Houtven et al., 2011, Howard, Jan, Rose et al., 2011). Evidence of their application was also found in psychology, as well as in natural resource management (Fieberg et al. 2010). The experiment proposed in this thesis exploits one of the forms of choice modeling called contingent ranking. In contingent ranking, experiment respondents are typically asked to rank a set of presented alternatives on a scale from most preferred to the least preferred. Usually, each alternative is described by several attributes varying in levels across alternatives (Speelman, Farolfi, Frija, & Van Huylenbroeck, 2010). This allows determining not only the most preferred choices, but also establish role of each attribute in decision-making. In BCIC-New Ventures and  68 BCIC-CAT, judges are asked to rank projects participating in the competition considering a number of attributes of different levels (e.g. financial need, probability of commercialization and environmental friendliness). As such, the feature of contingent ranking was used in order to separately value the impact of judges’ awareness about the financial situation of the entrepreneurs on their final ranking decisions. As a result, a choice-contingent technique was proposed in the experiment in this thesis.  4.4 Description of the Experiment This section explains how the experiment is constructed, and how the outcome of ethical trade-off between the competition fairness and the greater good maximization can be established based on information obtained in the experiment. It also states the hypothesis tested in the proposed experiment, and explains how the decision on size of the sample was made.  4.4.1. Questionnaire Instruments The proposed experiment tests an assumption of possibility for the agency to deviate from the announced decision rules, which results in efficiency distortions described in Chapter 3. In particular, it tests judging rationale when judging is subjective and influenced by a variety of factors. The main question that the experiment was called to answer was how judges would allocate prizes in a hypothetical competition when presented with choice between maximizing social welfare and maintaining ethical integrity. There were conducted four experimental sessions. The experiment was conducted in the form of a survey with two survey questionnaires being used. Four experimental sessions were held with four groups of  69 respondents that were randomly distributed one of the two survey questionnaires. This resulted in two groups of 36 University of British Columbia undergraduate students each filling out two questionnaires (72 participants in total). Each experimental session lasted approximately 20 minutes. Participants were paid a $10 participation fee in the form of a gift card. In order to keep the measurement error at minimum, subjects of the experiment were randomly recruited and randomly assigned to the groups. In addition, random assignment of experimental treatments to groups of subjects was assured. Based on the format of BCIC-CAT and BCIC-New Ventures competitions, two survey instruments (Questionnaires A and B) presented in Appendices D and E were developed for this study. To develop a survey, the format and decision making process of CAT and New Ventures competitions were scrupulously studied. At this early stage, an interview with BCIC officials was conducted to clarify the details of the winner selection process. Then, a list of decision-making criteria at every stage of the winner selection process was established. Based on this list, to reconstruct a real-world competition, hypothetical quality descriptions of the projects, probability of their commercialization and personal contribution of entrepreneurs were developed. In addition, a list of questions on socio- demographic characteristics of participants was created. Both questionnaires were scrupulously pre-tested in a group of 14 volunteer UBC students. Finally, required changes were fully incorporated. Survey instruments consisted of two parts. In the first part, participants were asked to provide general information about them. The questionnaires asked their gender, languages spoken in their childhood homes, which was expected to reflect their ethnic background, average grades received during their last year in high school, mother’s and father’s levels of  70 education and participants’ attitude towards a government being supporting business innovation by public money. This information was expected to help in understanding decisions made by respondents in the second part of the survey. In the next, major section of the questionnaires, each respondent was asked to act as a competition judge. First, respondents in both groups had to provide a complete ranking of eight hypothetical projects described by a set of attributes based on presented criteria. In particular, the projects were characterized by their patentability potential, novelty level and environmental sustainability. The description of the projects presented to respondents in both groups was absolutely the same. Table 3 provides an example of how two projects were described. Table 3: Examples of Projects Described by Three Quality Characteristics Project Summary A This project faces 60 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 3/5 with respect to environmental sustainability F This project faces 100 percent chance of receiving a patent, ranks “excellent” in terms of novelty, and scores 5/5 with respect to environmental sustainability  Next, respondents in both groups were presented with additional information on the probability of projects being successfully commercialized with and without prize money. In addition to this, respondents in Group A were also provided with information on personal contributions of entrepreneurs into their projects. This piece of information was supposed to provide respondents with an idea of the amount of investments lost for society in case the project does not win the prize and cannot proceed to the commercialization stage on its own. It is important that Group B respondents were not presented with this second piece of  71 information which, according to the hypothesis, should had resulted in different rate of re- rankings in Group B. Table 4 shows the actual financial information that the respondents were given at this second stage of the survey, and the last column with personal contribution data was included only into Group A questionnaires. Table 4: Probability of Successful Commercialization and Personal Contribution by Entrepreneur  Project Amount Awarded Personal Contribution  $0 $100,000 $175,000 $250,000 Probability of Successful Commercialization A 0.1 0.3 0.4 0.5 $9,500 B 0.9 0.95 0.99 1 $50,000 C 0.8 0.85 0.9 0.95 $40,000 D 0 0 0.1 0.9 $10,000 E 0.1 0.1 0.2 0.3 $8,000 F 1 1 1 1 $60,000 G 0.6 0.65 0.7 0.8 $25,000 H 0.7 0.8 0.8 0.9 $9,600  It was expected that after learning what the chances are for each project to be commercialized with and without the prize, respondents will consider to change initial rankings, taking into account potential loss of social welfare from some projects being unable to proceed without prize money. It was also expected that the rate of re-rankings will be higher in Group B where respondents had less incentives to worry about social justice. Thus, respondents trade competition fairness to the greater good if, when compared between groups, re-ranking decisions of respondents are significantly different, whereas their initial ranking decisions do not exhibit a significant difference.  72 In summary, in the first part of the survey, the original ranking is performed based on project quality exclusively. This procedure involves aggregation of various quality indexes in order to come up with single rank for each project. In the next part, participants are informed that the government managing this competition is interested in maximizing social welfare. It implied that this goal can be achieved through not allocating prizes to the “top three” projects for which the probability of commercialization is high without the prize, and instead allocating the prizes to a “non-top three” projects for which the probability of commercialization is low without the prize and high with the prize. Participants realize that entrepreneurs were informed that the prize allocation will only be based on the quality of their business plans. Nevertheless, they may realize that ranking decisions are confidential and highly subjective, so the projects can be re-ranked without running a risk of being accused of following inappropriate decision-making procedure and being inconsistent with initial rules. This ethical trade-off of maximizing social welfare at the expense of individual fairness is particularly strong for Group A participants, who were exposed to a personal financial contribution from the entrepreneurs. This trade-off is not nearly as strong for Group B participants who were not presented with this information. The study was particularly preoccupied with estimating the difference in re-ranking decisions across the two groups of respondents as an indicator of judges being willing to maximize the greater good at an expense of fair competition outcome in real-world commercialization competitions. Actual experimental data showed that out of 72 respondents only six (8.3%) decided to leave their original rankings unchanged, and four of them were from Group A. The rest of respondents reconsidered their original ranking based on quality. Table 5 provides typical ranking and re-ranking taken from the actual data set for Group A and Group B.  73 Table 5: Example of Ranking and Re-ranking in Each of Two Groups Projects Group A Group B Ranking Re-ranking Ranking Re-ranking A 6 7 6 6 B 2 1 2 5 C 3 2 3 4 D 4 4 5 1 E 8 8 8 7 F 1 3 1 8 G 5 5 4 3 H 7 6 7 2  4.4.2. Hypothesis As was illustrated in the previous sub-section, the majority of individual respondents tended to re-rank the projects after being presented with additional information. Does this change constitute a particular pattern in each group of respondents? Is the change in respondents’ rankings statistically significant? How do the re-ranking patterns compare between the two groups? These are the major questions posed in this research. If the tendency to deviate from original ranking is consistent for both groups, and if the respondents in one of the two groups deviate significantly more than the other, then the potential for wiliness to trade competition fairness for the greater good is established. By analyzing socio-demographic information provided by the respondents, the determinants of respondents’ re-rankings behaviour can be established. Thus, the following two sets of hypothesis were formulated: 1) Fairness versus the greater good hypothesis:  74     there is no statistical difference in original rankings of Group A and Group B participants, whereas there is a statistical difference between re-rankings across two surveys, as opposed to     there is no statistical difference both between rankings and re-rankings across the two surveys. 2) Socio-demographic hypothesis:   : factors such as gender, ethnic background, parents’ education, attitudes toward importance of innovation support by the government and average percentage grade received in final year at high school have no impact on Survey A respondents’ statistically significant re-rankings, as opposed to     some socio-demographic factors had statistically significant impact on re-ranking decisions in that group of respondents.  4.4.3. Size of the Sample One of the crucial questions any experimenter faces is how big the number of participants should be. There is no one precise rule to answer this question. Studies eliciting consumers’ preferences by the means of contingent ranking usually use a large number of participants ranging from several hundred to several thousand (Howard, Jan, Chadban et al., 2011; Van Houtven et al., 2011). Despite the fact that a method of contingent ranking was used in this experiment, the sample size usually used in contingent ranking experiments cannot serve as guidance, since the research questions, hypothesis and analysis are different. Based on the reviewed literature, the only determinant of sample size in the current study was statistical analysis procedures required to analyze the obtained data. Based on the literature review, it was established that for the purpose of assessing intra- and intergroup concordance,  75 the size of the sample does not have to be large. For example, when studying multi-group concordance, Vidmar and Cernogoj used total of 47 respondents divided into groups of 4-6 in each of 9 groups (Vidmar & Cernogij, 2005). To establish the level of concordance between labour productivity and wages, Verbic and Kuzmin analyzed two rank structures of 10 observations each (Verbic & Kuzmin, 2009).  When applying their method of testing for homogeneity of populations of rankings, Bu, Cabilio and Zhang used two examples: first, with 43 and 48 respondents, and second with 14 and 13 respondents in each group (Bu, Cabilio, & Zhang, 2009). Thus, it was decided to keep the size of each of the two groups under 40 participants. Eventually, 72 volunteer participants took part in the experiment, and were split into two groups of 36 individuals.   76 Chapter  5: Data Analysis This chapter presents analysis of the data obtained in the experiment with the purpose of establishing the possibility of competition judges to deviate from the decision based on project quality only when presented with additional information. It starts with a relevant literature review indicating that there is no one perfect way to analyze the relationships between groups of rankings. It was possible to discover three approaches that are usually applied when estimating the agreement of rankings provided by several groups of judges. Next, these three methods are applied to experimental data and the results are provided. Further, socio-demographic factors influencing respondents’ decision to re-rank the projects are established and discussed.  5.1 Relevant Literature First attempts to develop non-parametric methods based on ranks started with significant contributions of Maurice  G. Kendall and Babington Smith (1939), Frank Wilcoxon (1945) and H. Mann and D. Whitney a couple of years later. Their works focused on measuring the concordance of ranked sequences within an individual rank structure, i.e. inter-group concordance.  In 1955, Kendall proposed a rank correlation coefficient for measuring the agreement between two sets of rankings. This coefficient is based on the order of sets of rankings that are used to calculate the number of inversions required to transform one set of rankings into another. Finally, Pearson correlation coefficient is applied to the binary measures obtained. Whereas an agreement exists on testing the unanimity of judges’ decision within a group that most often involves Kendall’s rank correlation coefficient, analysis of the  77 literature shows that there is no one widely accepted non-parametric test to estimate agreement of judges’ opinion between the groups. The analysis of the existing literature indicates that the major cause is the absence of common agreement on the set of hypothesis. The first statistic test to establish the level of agreement between the two groups of judges was offered by Schucany and Frawley (1973).  Although useful and definitely easily applicable, this method is extensively criticized in the literature for several drawbacks. The most debatable is the hypothesis statement. The null hypothesis after Schucany and Frawley is formulated in such a way that, if rejected, leads to any number of alternative hypotheses being accepted (SNELL, 1983). In 1978, Hollander and Sethuraman first pointed this out and made an attempt to overcome this problem. They offered a test in which the null hypothesis tested was the identity of mean ranks across groups. Snell (1983) conducted a comparative analysis of both approaches and testified that the Hollander-Sethuraman test does not improve the Schucany- Frawley statistics much. (SNELL, 1983).  Their hypothesis requires not only agreement of judges within the groups, but also calls for the same intensity of concordance in both groups. The major problem of degree of concordance required within each group for the test to be valid still was not completely resolved. Another measure of concordance between two groups of rankings, U-Statistics, is based on the average rank correlation between rankers from each group. This approach was described by Palachek and Kerin (1982) and is an extension of Quade’s (1972) analysis of internal rank correlation between rankings within one group.  It is noted that the hypothesis stated in Schucany-Frawley and Hollander–Sethuraman tests are more restricting than in U- Statistics test. The latter is more general and its application in practice is wider. The  78 versatility of the U-Statistic is explained by opportunity to construct a confidence interval for the estimate of the strength of agreement between two groups. Palachek and Kirin gave special attention to U-Statistic as most versatile in their opinion. (Palachek & Kerin, 1982). Coefficient of Structural Concordance as a measure of concordance for two analogous rank structures was proposed by Verbic and Kuzmin (2009). The nature of this coefficient is in using the differences in absolute terms between the ranks of the two groups. It was derived as a measure applicable for diverse range of analysis with no restriction on what the rank structures represent. The authors position it as a “pure intergroup measure of concordance designed as a complement to Kendall’s intragroup coefficient of concordance” (Verbic & Kuzmin, 2009). This important feature of the coefficient makes it an appealing tool to be used in two-group concordance analysis. Unfortunately, the distribution of this measure is unknown. Authors suggested use of boot-strap procedure, as in Bradley Efron (1979), for creating an empirical distribution and determining a confidence interval. This chapter presents application of two of before mentioned methods of estimating intergroup concordance. Schucany-Frawley statistics was often criticized in the literature, as was mentioned. An empirical study conducted by Martin C. Snell showed that application of the Hollander-Sethuraman approach does not provide much improvement over the Schucany- Frawley method (SNELL, 1983). Thus, these tests were eliminated from analysis.  5.2 Description of the Data This section provides a description of two types of data collected in the experiment. Sub-section 5.2.1 focuses on rankings and re-rankings obtained in both groups, whereas subsection 5.2.2 summarizes socio-demographic information about respondents.  79 5.2.1. Ranking Data Respondents in each of two groups were asked to rank and re-rank eight projects participating in the hypothetical commercialization competition. The collected data represents four sets of rankings. There were 36 respondents in each of two groups. Thus, each set of rankings is in the form of matrix of size 36x8. Based on mean rank, projects F, C, and B were consistently ranked as first, second and third in original rankings by both groups of respondents (Table 6). All three projects on average were re-ranked lower after the respondents were presented with additional information, but this difference was bigger in Group B which did not know about entrepreneurs’ personal contribution to their projects. Table 6: Descriptive Statistics of Ranks Assigned to Projects in Both Groups Project Group A Group B Ranking Re-ranking Ranking Re-ranking Mean H L Mean H L Mean H L Mean H L A 6 5 7 6.2 2 7 5.9 4 6 6 1 8 B 2.1 1 5 2.5 1 8 2 2 2 2.6 1 7 C 2.9 1 3 3.1 1 6 2.9 1 3 3.3 2 6 D 4.9 3 7 5.6 1 8 4.8 4 6 4.8 1 8 E 7.6 2 8 7.4 1 8 7.9 7 8 7.2 2 8 F 1.1 1 4 2.1 1 8 1.1 1 3 2.4 1 8 G 4.4 4 8 4.4 1 8 4.3 4 6 4.8 3 6 H 7 5 8 4.9 2 7 7.1 7 8 4.9 2 7 Note: 1. Lower value of mean indicates higher average rating of the project. Highest possible rank is    1, the lowest possible is 8. 2. H=the highest rank received, L=the lowest rank received  80 Figure 10 provides graphical representation of mean ranks assigned to each project and visualizes the tendency in re-rankings in both groups of respondents. Figure 10: Comparison of Mean Ranks Assigned to Projects at Original Ranking and Re-ranking by Two Groups of Respondents   Note: Mean rank =1is the highest, mean rank=8 is the lowest Figure 10 shows that there is discrepancy in mean ranks assignment to projects by judges in both group. For example, project F received on average the highest rank of 1 in original ranking by both groups of respondents. After respondents were presented with additional information specific to each group, they re-ranked project F. It ended up being second in the first group when received a higher average rank of 2.4 in the second group of respondents. Similar tendency remains for projects B and C that initially were ranked on average as second and third. Thus, respondents in group two which were not exposed to the personal contribution of entrepreneurs, during the re-ranking exercise assigned lower ranks the projects F, B, and C (the top three based on quality, in both groups of respondents), comparing to respondents in group one, in which the participants new entrepreneurs’ contributions. Respondents in both groups were almost unanimous about keeping the ranking 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 A B C D E F G H Mean Rank in Group 1 Original Ranking Mean Rank in Group 2 Original Ranking Mean Rank in Group 1 Re- ranking Mean Rank in Group 2 Re- ranking  81 of project A unchanged and re-ranking project H from 7 th  place to an average of 5 th  place. Project E, which both groups initially considered of the lowest quality, was re-ranked much higher by respondents who knew about the personal contribution of entrepreneurs than those who were unaware of it. Interesting insight comes from observation of highest and lowest ranks assigned to the projects. Both groups agreed that project F is the highest quality project. Its mean rank was very close to 1, the highest rank earned was obviously 1 in both groups and its rank was never lower than 4 in the first group and 3 in the second. At the same time, when re-ranked, the lowest rank assigned to this project was 8 in both groups. This signifies that despite general agreement about the exceptional quality of this project, when becoming aware of its high potential to be commercialized on its own and the financial commitment of the entrepreneur, some respondents re-ranked this particular project significantly lower. The same tendency is true for project B, which was ranked second and also had a comparatively high commercialization potential. Interestingly enough, whereas the highest quality project, project F, did receive rank 8 when re-ranked by both groups, project C which was ranked as third based on quality exclusively, was never re-ranked lower that sixth in either group. Thus, based on this visual analysis of descriptive statistics, the top two quality projects that were also the strongest two in terms of commercialization potential experienced higher spread of ranking when re-ranked and were negatively affected by re-ranking more than the third best in terms of quality.     82 5.2.2. Summary of Socio-Demographic Information Provided by the Respondents Both Group A and Group B included 36 respondents who were assigned to the groups in random order. The socio-demographic structure of both groups was quite similar. This can be observed in Appendix F. Almost 30% of respondents in each group were males. Most of the respondents were domestic students whereas the international students’ share constituted around 20% on average. The question about the language spoken in childhood at home was included into the questionnaire as one that might help to establish the impact of ethnicity on the ranking and re-rankings decisions. It was thought that in a multicultural environment such as Vancouver, it might not be easy for some students to report their ethnicity. Thus, information about the languages spoken at home was expected to reflect the ethnicity of the respondent. Almost half of the students reported Chinese language, either alone or in combination with English, to be the language spoken in their childhood home. English alone was spoken in childhood by 25-30% of participants. Due to the fact that innovation is produced by the most progressive and educated members of society, it was expected that questions about education level of the participants and their parents might supply information on the impact these factors have on the way students allocate prizes to competing projects. When asked about the highest degree earned by their mothers, approximately 45% of all respondents indicated “university graduate”. The same answer with respect to their fathers was given more than 60% of the students. At the same time, above 65% of all respondents earned average grades above 85% in high school. Finally, the respondents were asked to rank (on a scale from 1 to 10) a statement about the importance for government to use tax payers’ money to support business innovation in society. Their responses were supposed to directly indicate the awareness about  83 the role of innovation in society wealth growth, as well as the role of the government in this process. It appeared that 20% of the participants strongly agreed with the provided statement assigning the highest rank of 10, while more than half ranked the statement higher than 8.  5.3 Testing the Agreement of Respondents’ Rankings and Re-rankings within Each Group There are several coefficients available to test for intragroup concordance of respondents’ rankings. They all stem from bivariate ordinal correlation. The most widely used in the literature is Kendall’s W proposed by Maurice G. Kendall and B. Babington Smith in 1939 (Kendall & B. Babington Smith, 1939). When n judges are asked to rank k objects, the authors called this the problem of n-rankings.  They introduced a Coefficient of Concordance:     ∑ (  (   )  )      (    ) where     is the sum of ranks for the    object. Usually, the null hypothesis is no concordance among the judges, which essentially means the judges assign ranks at random. This hypothesis is rejected, at chosen level of  , if  (  )   where   is the chi-square critical value with (k-1) degrees of freedom. When judges perfectly agree on the rankings, W=1. Total disagreement exists when W=0. Kendall’s measure is strictly positive and belongs to the interval 0    . The presence of ranking preference pattern with application to rankings and re- rankings of respondents in Group A and Group B was tested by calculating Kendall’s W’s and establishing their significance. Results are reported in Table 7 for both groups. Each  84 group was comprised of 36 subjects. The number of ranked objects is 8, thus number degrees of freedom is 7. Table 7:  Kendall’s W Values of Intragroup Concordance  Ranking Re-ranking W Calculated critical value (7 df) W Calculated critical value (7 df) Group A 0.905 0.025 0.581 0.028 Group B 0.977 0.027 0.470 0.028  As can be observed from Table 7, the test-statistics are far below the critical values at 7 degrees of freedom, which signifies that high level of agreement among the respondents and the difference between the respondents’ rankings inside each of the four sets of rankings is caused by chance. Since the test-statistic is in the critical region, the null hypothesis of no concordance between judges’ opinions is rejected. Thus, it is possible to conclude that respondents’ rankings and re-rankings in both groups were not random and there was a high level of agreement of respondents on ranks within each set of rankings.  A closer look at the absolute value of the coefficients suggests that there were less agreement among respondents during the re-ranking exercises comparing to ranking activity.  Drawing parallels to real world commercialization competitions, this suggests that when judges are presented with financial information about participants, this would cause some judges to change their opinion. Thus the judges will be less unanimous as for final award allocation comparing to degree of their agreement on original ranking based solely on project quality. In addition, if judges are aware of personal contribution of the entrepreneur the level of agreement on prize  85 allocation is higher than when they are not informed about the firms’ financial commitments to the project.  5.4 Estimation of Intergroup Concordance As was briefly mentioned in the section of literature review, there is no unanimous opinion among researchers on one perfectly suitable measure of concordance for two or several sets of rankings. For the purposes of the analysis in this thesis, the U-Statistic estimator of intergroup concordance (sub-section 5.4.1) and Coefficient of Concordance (sub-section 5.4.2) were chosen. They are easy to apply and formulation of the null hypothesis in these tests matches the purpose of the analysis.  5.4.1. U-Statistic Estimator of Intergroup Concordance U-Statistics analysis allows estimating concordance of two groups of rankings based on the level of concordance between rankings of individual judges from group one and from group two. Then, the expected rank correlation between rankers from each group is calculated, which serves as a test of agreement level between the two groups of judges. In the spirit of Palachek and Kerin (1982), let R(     ) be the rank correlation coefficient between rankings of judge i from group 1 and judge j from group 2. Correlation between their rankings is determined using either Spearman or Kendall rank correlation coefficients. In current analysis Spearman coefficient was used:   (     )     (    )∑ (   )(   )  86 A parametric measure of concordance for two groups is estimated by:     ( (     ) Palachek and Kirin suggested using unbiased U-Statistic estimator of     :    ∑ ∑  (    )     to test the null hypothesis of no agreement within one or both of the groups, or between the groups, versus the alternative hypothesis that there is agreement within the groups and between the groups:          versus        . Palachek and Kirin showed that standard normal distribution can be used for approximate tests and confidence intervals. The test rejects the null hypothesis in favor of the alternative if:   ̂   where asymptotic variance of √    can be estimated using method of Puri and Sen (1971):  ̂             ∑ [   ∑  (    )    ]         ∑ [   ∑  (    )    ]     Results of the test and hypothesis testing across two groups are provided in Table 8.     87 Table 8: U-Statistics as a Measure of Concordance between Rankings and Re-rankings Across Two Groups of Respondents  Measure of strength of agreement Test- statistic  Critical value   ̂   Hypothesis test Group A rankings and Group B rankings 0.937 0.026 0.059 Accept Ho Group A re-rankings and Group B re-rankings 0.5476 0.015 0.035 Accept Ho  Test-statistics are greater than critical values. Thus, the composite null hypothesis of no agreement within one or both of the groups or between the groups is accepted. In sub- section 5.3, with high degree of significance it was established that respondents inside both groups agreed on original rankings and re-rankings. When combining that result with the Kendall’s Coefficient of Concordance results, it is reasonable to credit most of the disagreement effect to inter-group interaction. The closer the absolute value of    to zero, the less agreement is established. Thus, although the respondents in both groups did not perfectly agree on the original rankings, 94% of concordance shows that generally they ranked the projects with high degree of unanimity. The re-rankings in both groups agreed only on 55%, indicating that projects were re-ranked differently due to information presented to respondents in the second part of the questionnaire.  To get a sense of which of the two groups re-ranked the projects more, the difference between original rankings and re-rankings in each of the two groups was analyzed. The  88 obtained results (Table 9) show the significant difference in both groups, with slightly less deviation of re-rankings in Group A (0.698>0.626). Table 9: U-Statistics as a Measure of Concordance of Rankings and Re-rankings within Each of the Two Groups of Respondents  Measure of strength of agreement Test- statistic  Critical value   ̂   Hypothesis test Group A rankings and re-rankings 0.698 0.019 0.046 Accept Ho Group B rankings and re-rankings 0.626 0.017 0.041 Accept Ho Since Group B was not presented with information on the personal contribution of entrepreneurs to their projects, it was expected that respondents will tend to re-rank projects more than in group A, due to less intense ethical dilemma. Confirming this expectation, the value of   is lower in Group B.  5.4.2. Coefficient of Structural Concordance The Coefficient of Structural Concordance was proposed as a universal measure of measuring an agreement across two groups of ranking without any restrictions put on the groups besides those determined by the logic of such an analysis. The following methodology is presented after Verbic and Kuzmin (2009). If   is a rank in the rank structure A of the value    of the variable X assigned by the value of    of the ranking variable Y and   is a rank in the rank structure B of the value of the variable W assigned by the value    of the ranking variable Z, and variables X and Y  89 consist of values            with N being the number of ranks in each set of ranks, and          with M being the number of sets of ranks, then Coefficient of Structural concordance   may be calculated by:    (    ) ∑ ∑ |   |   for odd number of ranks in each set of ranks, and    ∑ ∑ |   |   for even number of ranks in each set of ranks. This coefficient a priori cannot be negative and belongs to the interval      . Its absolute value estimates level of agreement between the two groups of rankings, with values close to 0 indicating absence of concordance in judges’ opinions, and values close to 1 signifying strong association between the two groups of responses. In order to avoid misleading inference based on the obtained estimates, accurate estimation of the uncertainty associated with them is required. Usually, this can be done by construction of confidence interval which is claimed to include the true coefficient value with a pre-specified probability. Distribution of Coefficient of Structural Concordance, which is a non-parametric measure, is not obvious. In this thesis it was approximated with a bootstrap procedure in the spirit of Bradely Efron (Efron B., 1986). Bootstrapping is an approach to assign measures of accuracy to sample estimates. Usually, in order to make conclusions about the value of population parameter , a random sample   from that population is drawn, and an estimate  ̂( ) of the value of   from the sample is constructed. The bootstrap procedure allows obtaining information about the relationship between   and the random variable  ̂( ) by looking at the relationships between  ̂(    ) and  ̂(  )   where    is a resample characterized by the sample      (Carpenter J., 2000). For the purpose of current  90 estimation,    was constructed by sampling with replacement from the data vector      using Bootstrapping add-in in Excel. It was achieved by following the general algorithm below: 1. Sampling 36 observation  randomly with replacement from      to construct a bootstrap data set    (the number of observations corresponds to the number of observations in the sample); 2. Calculation of the bootstrap version of the Coefficient of Structural Concordance,  ̂   ̂(  ); 3. Repetition of (1) and (2) 1000 times to approximate the bootstrap distribution. Results of bootstrapping are presented in Appendix G. Using obtained standard errors  ̂ for the estimator  ̂(  ), approximate confidence intervals for unknown parameter   were constructed using the following well-known formula:    ̂   ̂ ( ) where  ̂ is a sample Coefficient of Structural Concordance,  ( ) is the 100   percentile point of standard normal variate ( (    )       ) (Efron B., 1986). Table 10 provides estimated values of Coefficient of Structural Concordance as a measure of agreement between original rankings across the two groups of respondents, and agreement of re-rankings across the two groups. In brackets, a standard 95 percent approximation confidence interval for   is reported.     91 Table 10: Coefficient of Structural Concordance as a Measure of Agreement between the Two Groups of Respondents  Group B Ranking Re-ranking  Group A Ranking 0.920188 (0.918, 0.928) x Re-ranking x 0.66667 (0.654, 0.679)  The obtained coefficients are good estimators of population coefficient of concordance in both cases, and their absolute value confirms the research hypothesis of this study. In particular, after being presented with different additional financial information about entrepreneurs and their projects which called for different amount of social justice, the respondents in two groups re-ranked projects differently (66.7% of concordance and  33.3% of disagreement).  5.5  Socio-Demographic Determinants of Respondents’ Re-ranking Decisions In this section, the linear regression model estimating which socio-demographic characteristics of respondents had significant impact on their re-ranking decisions was constructed. In the survey, each respondent provided information about their gender, language spoken in childhood home, mother’s and father’s highest education level,  respondents’ highest average mark in last year at school, their attitude towards innovation being supported  92 by tax payers money and whether they are international students or not. It was expected that these characteristics might have the strongest explanatory power of respondents’ re-ranking decisions. Thus, these characteristics were included into the model as explanatory variables. Kendall’s measure of concordance between two sets of rankings (also referred as Kendall’s rank correlation coefficient or tau) was chosen as a response variable. For each respondent in both groups, Kendall’s rank correlation coefficient ( )was computed to measure the level of discrepancy between original ranking and re-ranking. These computations as well as application of the Ordinary Least Squares (OLS), description of which follows, were carried using R software environment for statistical computing and graphics. The subsequent regression analysis was carried for Group A and Group B (36 observations each) separately. This decision was predetermined by the structure of the experiment. As was explained in Section 4.3, the only difference in questionnaires presented to participants of Group A and Group B was the inclusion or omission of information on personal contribution of entrepreneur to the project. This was supposed to reveal the respondents’ ethical choice, either to maximize the greater good or ensure the individual fairness. Thus, the focus is to establish which factors influence this choice in each case. In the questionnaires, the question about language was open-ended, and mother’s and father’s highest education had several categories. Thus, various categorical answers were received. For example, the answers regarding language spoken in the childhood home were segregated into four major categories: Chinese, English, Chinese and English, Spanish. In order to determine the most significant categories of answers for each of the above mentioned questions, ANOVA analysis was carried out. This allowed reducing the number of  93 categories by choosing the significant categories only. Eventually, “Chinese and English” was included as the explanatory variable for Language. For mother’s education and father’s education “University Graduate” was chosen for inclusion into model. The following model was estimated with regression by using the Ordinary Least Square:    = where:    = Kendall’s rank coefficients calculated for each respondent, 1    ;     = gender, a dummy variable;     = language routinely spoken in respondent’s childhood home, dummy variable;      = the highest education completed by respondent’s mother, dummy variable;      = the highest education completed by respondent’s father, dummy variable       = average grade received in respondent’s final year at school;     = ranking of statement about innovation on a scale from one to eight;     = random errors. Stepwise algorithm based on Akaike’s Information Criterion (AIC) was applied to sequentially reduce the variables in the model, keeping the variables that minimized the criterion. The regression results of estimated coefficients are presented in Table 11.      94 Table 11: Regression Results for Factors Determining Re-ranking Decisions Explanatory Variables Model for Group A Model for Group B Intercept [      ]  [      ] Highest average Grade (HAS) [     ]  [     ] Language (L) – Chinese-English [      ]  [      ] Observations Adjusted F-statistic (           )   (         df)  Note: t-statistic is presented in square brackets; * P<0.5.  As the table 11 shows, the adjusted    is quite low in both models. The models explain only 20.3% of variation in rankings in Group A and only 7% of variation in group B. The only significant characteristics that were determined, and that influence decisions of respondents in Group A were respondents’ grades earned in last year at school (HSA) and the fact that the respondents spoke English and Chinese in their childhood. Based on analysis conducted in sections 5.3, in Group A, when judging includes the ethical component, respondents’ re-rankings were significantly different from their original rankings. The F- statistic, although not very high but significant, confirms that the two significant variables, the grades received in high school and possibility to speak Chinese in combination with English in childhood, definitely played an important role in this. Interesting conclusions arise from the analysis of obtained coefficients in Group A regression results. The value of the HAS coefficient estimate is positive, meaning that when  95 ethical trade-off is involved and respondents know exactly which allocation is “fair” and “unfair”, the higher the grades of the respondents in school, the less likely they tend to change their original ranking which contributes to increased fairness of their decision. Here, it is important to remember that the dependent variable is Kendall’s rank correlation coefficient ranging from (-1) to (+1). Thus, the higher HSA, the higher Kendall’s measure, and less disagreement between the original ranking and re-ranking of the projects provided by the respondent is expected. Another significant variable that was established in the analysis of Group A data was Language. This dummy variable distinguishes respondents who spoke both Chinese and English in their childhood home from respondents speaking other combinations of languages. In a situation when intense ethical trade-off is involved, if respondent speaks both Chinese and English, he/she tends to deviate from his/her original ranking since the estimated coefficient is negative (-0.269). Thus, if assuming that the languages spoken in the childhood reflect ethnicity, than in a situation when the best projects might be commercialized without prize money and personal financial commitment of entrepreneurs is known, due to their ethnic background, some judges might still re-rank these projects lower in order to let other lower quality projects to go forward.  The results obtained after running the regression model for Group B where no information about personal contribution of entrepreneurs was provided to the respondents show quite similar pattern. The same two variables entered the model and the coefficient estimates are very close. But the fit of this model is not very high. Coefficient of Determination (F-statistic) is low and its significance is 89%. Also, only 7% of variability in  96 the response variable is explained by the included explanatory variables. The high school average (HSA) and language (L) were not significant even at      .  In summary, the regression analysis showed that respondents’ level of education obtained in high school and Chinese background were significant factors influencing re- ranking decisions when strong ethical trade-off was involved, as in Group A. It was established that the higher grades respondents had in school, the less likely they tend to change their original rankings in order to accommodate commercialization of lower quality projects at the expense of competition fairness. When the ethical component is not as strong, there are other factors responsible for the large portion of variation in dependent variable that has not been captured by the model.               97 Chapter  6: Conclusions and Discussion  6.1  Summary and Conclusions The main objective of this thesis was to research publicly funded commercialization competitions in the format of New Ventures Competition and Commercialization of Agricultural Technology Competition administered by BCIC and show that the potential for time inconsistent award allocation incorporated into the competition structure causes efficiency distortions. Business is a powerful source of innovation. Analysis of business innovation development in Canada has shown that by productivity growth, which economists consider a major result of innovation, Canada is behind such OECD countries as Korea, Finland, USA and Japan. Investments by Canadian business into R&D are constantly increasing. They reached the $15 billion mark in 2011, but if expressed in percentage of GDP, they are still much lower than in countries that are innovation leaders of the world. Canadian government addresses this issue with continuous efforts in support of business innovations, and their commercialization in particular. Government uses commercialization competitions as a cost effective tool to determine the most socially valuable business innovations that require public funding to successfully penetrate the market. Commercialization programs are one component of a broad and diverse mix of business innovation assistance initiatives that have been implemented by Canadian government for years. These programs are a direct support measure and currently account for 26% of all direct federal funding in support of business R&D, and 2.5% total federal business R&D support. A review of commercialization  98 programs structure indicated that most of them are organized as a multi-stage assessment process very often involving evaluation by an independent industry expert panel. At the same time, final awards allocation decisions are usually made by the agency and are subjective. From discussion with BCIC officials, and from revision of several other agencies audit reports, it was established that this final decision stage is very confidential and either poorly documented or not documented at all. Besides project description, most of the reviewed programs require participants to submit detailed information about their financial position and the level of funding that has been committed to the projects. Thus, it was established that the expert panels and the agency officials are usually aware of the capacity of each participant to commercialize the innovation independently without the competition prize. Stylized commercialization competition model involving one risk-neutral principal and two risk-neutral homogeneous participants was developed by using a game-theoretic approach. It was shown that without commercialization competition socially optimal level of innovation investments differs from the level that is optimal for a firm. The size of this difference depends on the relationship between the social value and private value parameters and the slope of the firm’s marginal innovation cost curve. When commercialization competition is introduced into the model by using a Cournot-Nash equilibrium concept and incorporating feasibility and sociability conditions, the firms’ best response functions were derived. A situation when public agency is aware that the firm that owns the most socially valuable innovation accumulated sufficient amount of resources to commercialize it without the competition prize, while the other firm with less quality but still socially important innovation can commercialize it with the prize only, was considered. Since the public agency maximizes social welfare through commercialization of  99 as many innovations as possible, it allocated the award to the firm with the “second-best” project which contradicts the originally announced rules. The firm’s optimal level of innovation effort accounting for time inconsistent award allocation was derived. Computer simulation results confirmed that recognition of potential time inconsistent outcome of commercialization competition lowers the firm’s innovation efforts. Economic experiment to establish the possibility for the judges to trade competition fairness for the greater good was conducted. Seventy-two volunteer students from the University of British Columbia were asked to act as individual judges of a hypothetical commercialization competition. They were randomly divided into control and experimental groups. In both groups, participants were asked to rank and re-rank eight innovation projects based on presented information. Group B participants were provided with project quality indicators and probability of successful commercialization with prize money. Besides this, Group A participants also knew the amount of money each firm has invested into the project. The hypothesis at question was that there is no statistical difference in first rankings between Group A and Group B, whereas there is statistical difference in re-rankings between Group A and Group B. If this is confirmed, the conclusion about the possibility of judges to deviate from originally announced competition rules would be established, and consequently the possibility for the judges to allocate competition prizes by increasing the greater good at cost of competition fairness is shown. The obtained experimental data were analyzed by estimating the agreement inside and between groups of rankings. The intragroup concordance was estimated with Kendall’s rank correlation coefficient. With high degree of significance presence of a preference pattern in how respondents in both groups ranked and re-ranked the projects was established,  100 which signifies a high degree of intragroup concordance. As for intergroup concordance estimation, the relevant literature review showed that there is no unanimity among the researchers as for one statistical procedure that should be used for testing hypothesis about groups of rankings. Thus, two non-parametric methods that are extensions of Kendall’s rank correlation coefficient with application to comparison of several groups of rankings were used. The U-Statistic estimator of intergroup concordance and Coefficient of Structural Concordance were estimated. When comparing original rankings between Group A and Group B, no significant difference between them was found. Therefore, a high level of agreement was established. When considering re-rankings in Group A and Group B, statistical difference was established with high level of significance. This confirmed that respondents in Group A and Group B re-ranked the eight projects differently and do not agree on re-ranking. Thus, the presentation of information about personal contribution in Group A resulted in different re-ranking patterns. Thus, allocation of competition prizes in Group A significantly differed from allocation in Group B. Since it was shown that the original rankings were similar in both groups, this resulted in top prizes being allocated to different projects in both groups which compromised the competition fairness. An empirical analysis to estimate an impact that some socio-demographic characteristics of respondents had on their prize allocation decisions was conducted. By applying OLS method of estimation, it was found that the respondents’ level of education obtained in high school and Chinese background were the only significant factors influencing re-ranking decisions when strong ethical trade-off was involved.    101 6.2 Restrictions and Recommendations The major difficulty of the researched topic is the subjectivity of decision-making process in commercialization competitions. We focused on finding an opportunity to confirm that the ethical trade-off between competition fairness and the greater good, which is closely related to the problem of time inconsistency of competition outcome, is possible in real life competitions. The process of final awards allocation in BCIC-CAT competition is not documented, and thus a lack of sufficient amount of relevant information was one of the major obstacles. However, it was overcome with implementation of the experimental study. Another obstacle that was experienced in working under this thesis was related to the nature of the data obtained in the experiment. The data came in the form of two samples, two groups of rankings each, and scarce relevant publications were not unanimous on appropriate statistical procedures. Also, based on the literature review, there is no readily available statistical software to estimate the level of agreement between groups of rankings of several objects. The issue was resolved by applying several statistical tests which provided similar results. Through implementation of the experiment, this thesis showed the possibility of competition outcome to be unfair from the point of view of individual participating firms due to deviation of judges from original decision-making rules.  Future research can be conducted in the direction of experimentally testing how level of the firms’ innovation efforts is distorted under time inconsistency. Despite the restrictions mentioned above, this thesis has fully investigated the theory of time inconsistency problem arising in commercialization competitions and empirically tested the aspect of fairness versus the greater good. By addressing the distortion that has not  102 been addressed in economic publications before, this research will contribute to the literature on research tournaments. It can be beneficial to innovation policy makers at federal and provincial levels. Extensive review of business innovation support in Canada can be of interest to a wide range of researchers and the general public. And finally, the raised issues and obtained results can be enlightening and useful for the British Columbia Innovation Council and other public agencies administering innovation competitions.      103 References  Agriculture and Agri-Food Canada. (2011). Audit of the agri-opportunities program Retrieved February, 23, 2012, from http://www4.agr.gc.ca/AAFC-AAC/display- afficher.do?id=1315589750421&lang=eng#two Agriculture and Agri-Food Canada. (2012). Agricultural innovation program guide Retrieved February, 23, 2012, from http://www4.agr.gc.ca/resources/prod/pdf/SPT_aip- pia/aip-pia_guide_eng.pdf Atlantic Canada Opportunities Agency. (2011). Quality assuarance audit of business development program - final report (2006). Retrieved 02/19, 2012, from http://www.acoa- apeca.gc.ca/eng/Accountability/AuditsAndEvaluations/Documents/BDP_QAA_EN.pdf Atlantic Canada Opportunities Agency. (2012). Repayable contributions portfolio of business development program - 2009. Retrieved 02/18, 2012, from http://www.acoa- apeca.gc.ca/eng/publications/ParliamentaryReports/Pages/RepayableContributionsPortfo liooftheBusinessDevelopmentProgram.aspx Baye, M. R., & Hoppe, H. C. (2003). The strategic equivalence of rent-seeking, innovation, and patent-race games. Games and Economic Behavior, 44(2), 217-226. doi:10.1016/S0899-8256(03)00027-7  104 BCIC New Ventures Competition. (2011). BCIC-new ventures competition: Agritech 2011. Retrieved 09/30, 2011, from http://www.newventuresbc.com/the-competition/agritech- competition/ BCIC New Ventures Competition. (2012). Rules 2012. Retrieved March/02, 2012, from http://www.newventuresbc.com/the-competition/rules/#2 Blackburn, K., & Christensen, M. (1989). Monetary-policy and policy credibility - theories and evidence. Journal of Economic Literature, 27(1), 1-45. British Columbia Innovation Council. (2011). 11th annual BCIC-new ventures competition now open. Retrieved 09/30, 2011, from http://www.bcic.ca/media-room/media- room/media-releases/1533-11th-annual-bcic-new-ventures-competition-now-open British Columbia Innovation Council, Life Sciences. (2009). BC innovation council commercialization of agricultural technology (CAT) competition guidelines. Retrieved Bu, J., Cabilio, P., & Zhang, Y. (2009). Tests of concordance between groups of incomplete rankings International Journal of Statistical Sciences, 9(Special Issue), 97-112. Calvo, G. A. (1978). On the time consistency of optimal policy in a monetary economy. Econometrica, 46(6), 1411-1428. doi:http://www.econometricsociety.org Canada Economic Development for Quebec Regions. (2012). Formative evaluation of the community diversification and business and regional growth programs - evaluation report. Retrieved 02/18, 2012, from http://www.dec- ced.gc.ca/eng/publications/agency/evaluation/205/index.html  105 Carpenter J., B. J. (2000). Bootstrap confidence intervals: When, which, what? A practical guide for medical statisticians. Statistics in Medicine, 19, 1141-1164. Che, Y. K., & Gale, I. (2003). Optimal design of research contests. American Economic Review, 93(3), 646-671. Coughlan, A. T., & Schmidt, R. M. (1985). Executive compensation, management turnover, and firm performance: An empirical investigation. Journal of Accounting and Economics, 7(1-3), 43-66. doi:http://www.elsevier.com/wps/product/cws_home/505556 Cozzarin, B. P. (2008). Data and the measurement of R&D program impacts. Evaluation and Program Planning, 31(3), 284-298. doi:10.1016/j.evalprogplan.2008.03.004 Diamond, P. A. (1967). Cardinal welfare, individualistic ethics, and interpersonal comparison of utility: Comment. Journal of Political Economy, 75(5), pp. 765-766. Dufwenberg, M., & Gneezy, U. (2002). Information disclosure in auctions: An experiment. Journal of Economic Behavior & Organization, 48(4), 431-444. doi:DOI: 10.1016/S0167-2681(01)00235-9 Eckel, C. (2007). Economic games for social scientists. In M. J. Webster, & J. Sell (Eds.), Laboratory experiments in the social sciences (First Ed. ed., pp. 497-515). Burlington, MA: Elsevier Inc. Efron B., T. R. (1986). Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Statistical Science, 1(1), 54-75.  106 Ehrenberg, R. G., & Bognanno, M. L. (1990). Do tournaments have incentive effects? Journal of Political Economy, 98(6), 1307-1324. doi:http://www.journals.uchicago.edu/JPE/ Evenson, R. E., & Kislev, Y. (1976). A stochastic model of applied research. Journal of Political Economy, 84(2), 265-281. doi:http://www.journals.uchicago.edu/JPE/ Final Report of the Expert Panel on Commercialization. (2006). People and excellence: The heart of successful commercialization Retrieved March/08, 2012, from http://publications.gc.ca/collections/Collection/Iu4-78-2006E-I.pdf Fu, Q., & Lu, J. (2009). Contest with pre-contest investment. Economics Letters, 103(3), 142-145. doi:http://www.elsevier.com/wps/find/journaldescription.cws_home/505574/description #description Fullerton, R. L., Linster, B. G., McKee, M., & Slate, S. (2002). Using auctions to reward tournament winners: Theory and experimental investigations. Rand Journal of Economics, 33(1), 62-84. Fullerton, R. L., & McAfee, R. P. (1999). Auctioning entry into tournaments. Journal of Political Economy, 107(3), 573-605. Gachter, S., & Riedl, A. (2006). Dividing justly in bargaining problems with claims - normative judgments and actual negotiations RID C-9787-2009. Social Choice and Welfare, 27(3), 571-594. doi:10.1007/s00355-006-0141-z  107 Giebe, T., Grebe, T., & Wolfstetter, E. (2006). How to allocate R&D (and other) subsidies: An experimentally tested policy recommendation. Research Policy, 35(9), 1261-1272. doi:10.1016/j.respol.2006.01.008 Glazer, A., & Hassin, R. (1988). Optimal contests. Economic Inquiry, 26(1), 133-143. Golombek, R., Greaker, M., & Hoel, M. (2010). Carbon taxes and innovation without commitment. B.E.Journal of Economic Analysis and Policy: Topics in Economic Analysis and Policy, 10(1) doi:http://www.bepress.com/bejeap/topics/ Green, J. R., & Stokey, N. L. (1983). A comparison of tournaments and contracts. Journal of Political Economy, 91(3), 349-364. doi:http://www.journals.uchicago.edu/JPE/ Gulati, S., & Vercammen, J. (2006). Time inconsistent resource conservation contracts. Journal of Environmental Economics and Management, 52(1), 454-468. doi:http://www.elsevier.com/wps/find/journaldescription.cws_home/622870/description #description Harsanyi, J. C. (1955). Cardinal welfare, individualistic ethics, and interpersonal comparisons of utility. Journal of Political Economy, 63(4), pp. 309-321. Industry Canada. (2011). Canadian innovation commercialization program: Addressing the innovation pre-commercialization gap. Retrieved 02/09, 2012, from http://www.ic.gc.ca/eic/site/ich-epi.nsf/eng/02158.html Karni, E. (1996). Social welfare functions and fairness. Social Choice and Welfare, 13(4), pp. 487-496.  108 Kendall, M. G., & B. Babington Smith. (1939). The problem of m rankings. The Annals of Mathematical Statistics, 10(3), pp. 275-287. Kydland, F. E., & Prescott, E. C. (1977). Rules rather than discretion - inconsistency of optimal plans. Journal of Political Economy, 85(3), 473-491. Lazear, E. P., & Rosen, S. (1981). Rank-order tournaments as optimum labor contracts. The Journal of Political Economy, 89(5), pp. 841-864. Mash, R., Helm, D., & Hepburn, C. (2003). Time inconsistent environmental policy and optimal delegation. Unpublished manuscript. Ministry of Finance Canada. (2011). Canada's economic action plan, budget 2011 . Retrieved 02/08, 2012, from http://www.budget.gc.ca/2011/home-accueil-eng.html Nalebuff, B. J., & Stiglitz, J. E. (1983). Prizes and incentives: Towards a general theory of compensation and competition. Bell Journal of Economics, 14(1), 21-43. National Research Council Canada. (2009). IRAP - benefits to canadians. Retrieved 02/09, 2012, from http://www.nrc-cnrc.gc.ca/eng/ibp/irap/about/benefits.html Natural Sciences and Engineering Research Council of Canada. (2010). Evaluation of the centres of excellence for commercialization and research - 2009 report Retrieved 02/16, 2012, from http://www.nce-rce.gc.ca/_docs/reports/CECREvaluation2010-EN.pdf Nova Scotia Department of Agriculture. (2011). Innovation and commercialization of new opportunities for agri-based products programs guidelines 2011-2012. Retrieved  109 February, 23, 2012, from http://www.gov.ns.ca/agri/prm/programs/inn_comm_new_opps.pd OECD. (2010a). Collaboration in innovation - innovation today. Retrieved 02/15, 2012, from http://www.oecd.org/dataoecd/49/55/45185075.pdf OECD. (2010b). Measuring the innovation: A new perspective . Retrieved 02/15, 2012, from http://www.oecd.org/dataoecd/29/33/45188105.pdf Office of the Auditor General Newfoundland and Labrador. (2011). Annnual report january 2011. Retrieved February, 23, 2012, from http://www.google.ca/url?sa=t&rct=j&q=commercialization%20program%20newfoundl and%20and%20labrador%20assessment&source=web&cd=1&ved=0CCEQFjAA&url= http%3A%2F%2Fwww.ag.gov.nl.ca%2Fag%2FannualReports%2F2010AnnualReport% 2F2.12%2520- %2520Investments.pdf&ei=JtxGT8eMKIGxiQKYl_TaDQ&usg=AFQjCNHXMvQOh4 E-b2oTyG2RXsD3BntZ_Q Palachek, A. D., & Kerin, R. A. (1982). Alternative approaches to the two-group concordance problem in brand preference rankings. Journal of Marketing Research, 19(3), pp. 386-389. Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 37(1), pp. 7-63.  110 PRO INNO Europe. (2009). Innovation and innovation policy in canada, report 2009 . Retrieved 02/10, 2012, from http://www.google.ca/url?sa=t&rct=j&q=innovation%20policy%20canada&source=web &cd=1&ved=0CDEQFjAA&url=http%3A%2F%2Fwww.proinno- europe.eu%2Fpage%2Finnovation-and-innovation-policy- canada&ei=LXg1T_b4JOrjiAKhzvnACg&usg=AFQjCNFhdyuUWPdilvseb1LoQBBhu f8HjQ Review of Federal Support to Research and Development. (2011). Federal R&D panel reports six major recommendations Retrieved 02/10, 2012, from http://rd- review.ca/eic/site/033.nsf/eng/home Review of Federal Support to Research and Development - Expert Panel Report, Industry Canada, (2012). Innovation canada: A call to action. Retrieved 02/10, 2012, from http://rd-review.ca/eic/site/033.nsf/vwapj/R-D_InnovationCanada_Final- eng.pdf/$FILE/R-D_InnovationCanada_Final-eng.pdf Rosa, J., & Rose, A. (2007). Report on interviews on the commerciazation of innovation, statistic canada, science, innovation and electronic information division. No. Catalogue no. 88F0006XIE, no. 004). Ottawa: Ministry of Industry. Retrieved from http://www.statcan.gc.ca/pub/88f0006x/88f0006x2007004-eng.pdf Roth, A. E. (1988). Laboratory experimentation in economics: A methodological overview. Economic Journal, 98(393), 974-1031.  111 Schmookler, J. (1959). Bigness, fewness, and research. Journal of Political Economy, 67(6), pp. 628-632. Schoettner, A. (2008). Fixed-prize tournaments versus first-price auctions in innovation contests. Economic Theory, 35(1), 57-71. doi:10.1007/s00199-007-0208-9 SNELL, M. (1983). Recent literature on testing for intergroup concordance. Applied Statistics-Journal of the Royal Statistical Society Series C, 32(2), 134-140. doi:10.2307/2347292 Speelman, S., Farolfi, S., Frija, A., & Van Huylenbroeck, G. (2010). Valuing improvements in the water rights system in south africa: A contingent ranking approach RID A-7342- 2012. Journal of the American Water Resources Association, 46(6), 1133-1144. doi:10.1111/j.1752-1688.2010.00480.x Standing Committee on Government Operations and Estimates. (2012). Effectiveness of the office of small and medium enterprises and the canadian innovation commercialization program - report 2011. Retrieved February 20, 2012, from http://www.google.ca/url?sa=t&rct=j&q=cicp%20report&source=web&cd=2&ved=0CC kQFjAB&url=http%3A%2F%2Fpublications.gc.ca%2Fcollections%2Fcollection_2011 %2Fparl%2FXC70-1-411-01-eng.pdf&ei=-AhDT7- GEOmZiAK4uaykAQ&usg=AFQjCNHRksVR138bBSFfuM4FzjFZYdr4-w Stein, J. C. (1997). Internal capital markets and the competition for corporate resources. Journal of Finance, 52(1), 111-133. doi:http://www.blackwellpublishing.com/journal.asp?ref=0022-1082  112 Taylor, C. R. (1995). Digging for golden carrots - an analysis of research tournaments. American Economic Review, 85(4), 872-890. Terwiesch, C., & Xu, Y. (2008). Innovation contests, open innovation, and multiagent problem solving. Management Science, 54(9), 1529-1543. doi:10.1287/mnsc.1080.0884 Trautmann, S. T. (2010). Individual fairness in harsanyi's utilitarianism: Operationalizing all- inclusive utility. Theory and Decision, 68(4), 405-415. doi:10.1007/s11238-008-9104-4 Treasury Board of Canada Secretariat. (2011). Performance information by strategic outcome of national research council of canada. Retrieved 02/09, 2012, from http://www.tbs-sct.gc.ca/ppg-cpr/so-rs- eng.aspx?Rt=1049&Pa=0&Gc=0&So=10275&Dt=55 Verbic, M., & Kuzmin, F. (2009). Coefficient of structural concordance and an example of its application: Labour productivity and wages in slovenia RID B-5144-2009. Panoeconomicus, 56(2), 227-240. doi:10.2298/PAN0902227V Verbic, M., Majcen, B., Ivanova, O., & Cok, M. (2011). R&D and economic growth in slovenia: A dynamic general equilibrium approach with endogenous growth RID B- 5144-2009. Panoeconomicus, 58(1), 67-89. doi:10.2298/PAN1101067V Vidmar, G., & Cernogij, M. (2005). Studying norms in small groups by means of multi-group concordance analysis. Retrieved March/03, 2011, from http://ibmi.mf.uni- lj.si/ibmi/biostat-center/predtiski/PO_Vidmar_Cernigoj_ANACONDA.pdf  113 Western Economic Diversification Canada. (2012). Evaluation of western diversification program (2008). Retrieved 02/16, 2012, from http://www.wd.gc.ca/eng/10937.asp     114  Appendix A    Federal Programs Designed to Support Commercialization of Business Innovation Western Development Program (WDP) is administered by Western Economic Development Canada (WD) since 1987. It is currently the largest commercialization program in Canada with funding of $368 million in 2010-2011. Reflecting the significance of supporting Canadian innovating firms at the commercialization stage of innovation development, the funding of WDP has steadily increased over the years. For comparison, the accumulated funding of WDP during 2003-2007 amounted to $390 million. Geographically, the program covers British Columbia, Alberta, Saskatchewan and Manitoba. This funding is normally available to not-for-profit organizations, such as industry associations, economic development organizations, provincial governments and crown corporations, etc. The program focuses on areas of commercialization of technology, productivity growth and increasing rural diversification. As most of the programs delivered directly by WD, WDP is administered in the form of grants and contributions to cover specific costs. The application process includes submission of the Proposal and Funding Request. Through submission of the second document, participants reveal major financial information to the approval committee. Particularly, in the Funding Request Summary applicants must provide total project costs, WD funding requested and other sources of proposed funding (including committed and accepted).  If the proposal is in the priority areas, the applicant must submit a recent financial statement of the organization and a detailed budget. Another significant federal program aimed to support research and commercialization is the Centers of Excellence for Commercialization and Research (CECR) Program launched  115 in 2008. Being administered by the same three agencies as Network of Centres of Excellence initiative, CECR was created to stimulate the private sector’s investments in innovation with clear emphasis on commercialization. Consistent with the key areas of Canada’s competitive advantage, CECR supports business innovation in environmental science and technology, areas of natural resources and energy, information and communication and health and life sciences. During the five-year fiscal period of 2007-2008 to 2011-2012, CECR distributed $277.65 million. This money is intended to fund operating expenses and commercialization costs of funded centres. The program provides grants to centres on a five-year basis with expectation of the outcomes to be received by the end of this time span. There are three major types of centres with regard to their specialization: commercialization, research and research and commercialization. At the moment, 14 out of 17 existing centres that are funded through CECR deal with both research and commercialization projects, whereas three of them are solely commercialization focused (Natural Sciences and Engineering Research Council of Canada, 2010). The centres are playing a role in a hub that connects expertise of research and development institutions and facilities, as well as industry and business networks. They may reach for international expertise if needed and if possible. The assessment process is two-stage. At the first stage, letters of intent are reviewed by the Private Sector Advisory Board (PSAB) based on Program selection criteria and short- lists of the applicants. The second stage assumes submissions of full application, which is thoroughly considered by relevant federal agencies and an Expert Panel comprised of national and international specialists. After reviewing each application, the Expert Panel meets with applicants to discuss the main aspects of the application and provides a thorough assessment of the proposal. A final decision is made by the NCE Steering Committee based  116 on PSAB and the Expert Panel reports and funding recommendations. It is important to realize that the applicants in the case of CECR are centres themselves rather than firms or entrepreneurs. They receive CECR grants on a five-year basis. This money is called to cover operating, salary and knowledge dissemination costs of the centres, as well as commercialization expenses. The applicants are usually non-for-profit entities comprising of education and research institutions, as well as firms and non-for-profit organizations. Evaluation of CECR was conducted for the first time in 2009 one year after the program was launched. The Report indicated success of the program. All stakeholders, including funded centres management, expert panelists and provincial government representatives expressed satisfaction with the implementation of the program. Due to highly diligent and detailed and well-documented review and assessment, the selection process was recognized in general as “fair” and “transparent”. Unfunded applicants provided less optimistic insight, but none of them questioned the transparency of the selection process. Still, according to the recommendation in the Evaluation Report 2009, the competition was modified in terms of increased standardization of the Expert Panel (Natural Sciences and Engineering Research Council of Canada, 2010). Network of Centres of Excellence (NCE) was originally created in 1989 with the purpose to connect research and development and promote final commercialization of innovations. It is administered by three agencies: Natural Sciences and Engineering Research Council of Canada (NSERC), the Social Sciences and Humanities Research Council of Canada (SSHRC) and the Canadian Institutes of Health Research (CIHR) in a partnership with Industry Canada. In 2008, it was expanded through implementation of three new initiatives, one of which was CECR.  117 The applications are assessed by a peer-review system which assumes review by experts in a specific field.  A Selection Committee comprising of international experts uses selection criteria to review the letters of intent. Participants chosen at this stage will be invited to submit a proposal. An expert panel provides reports about the participating projects. Based on projects rating provided by the Selection Committee, expert reports and selection criteria, NCE Steering Committee makes funding allocation decisions. Business and Regional Growth Program (BRG) was launched in April, 2007 by Canada Economic Development for Quebec Regions and is expected to last until March 2012. The major goal of this program is to promote competitiveness and sustainable growth of Qu ́bec small and medium enterprises and regions. One type of activities for which grants and contributions are available to individuals and organizations of any type (including not- for-profits) is projects which through innovation development increase competitiveness of Qu ́bec and its regions. Annually, the program distributes slightly above $90 million in contributions and $720,000 in grants. The total funding for the period from 2007-2008 to 2011-2012 fiscal years was almost $463 million. The application process starts with applicants contacting a CED advisor serving a particular region who verifies the eligibility for the program. The next stage is a submission of application for financial assistance. Besides a general business and project profile, the application must enclose financial history of the firm for the previous two years and forecast financial statements for the next three years. Applicants must also submit the description of the project funding, including the information on confirmed and unconfirmed sources of financing. Thus, the Treasury Board that makes the final allocation decision is totally aware of the financial situation and needs of the participants at the moment of decision making.  118 For the purpose of funding allocation to a particular project, the budget allocation model for grant and contribution amounts is used. Seventeen business offices receive budgets based on an average unit cost of processing application for all offices. The Agency adopted a six-stage project approval structure. Unfortunately, the publicly available information about this process was limited to “this structure involves several steps, from forwarding by the business office to the appropriate branch, to the Deputy Minister’s signature”. Within the most recent Formative Evaluation of BRG program undertaken in 2008, nine focus groups were conducted with participants who received financial assistance. Discussions indicated that participants were generally satisfied with the CED’s services. Their major recommendation was related to long application processing times. In addition to this, some funding recipients identified other factors that negatively affected the way files were processed by CED. Particularly, “the rules for establishing project funding levels were unclear”. Neither interviews nor focus groups with unfunded applicants were conducted to establish their insight on assignment of funds (Canada Economic Development for Quebec Regions, 2012). Business Development Program’s (BDP) goal is to assist in starting-up businesses and expand or modernize the existing ones in Atlantic Canada. Its innovation element provides funding for R&D projects and adaptation of technology. The program focuses on small and medium size enterprises, but non-for-profits benefit from this initiative as well. Funding is available in the form of interest-free unsecured repayable contributions. Since the program was introduced in 1995, $1.0763 billion of loans were provided under BDP. The repayment term is usually 5-7 years. The annual default rate varies around 4% in recent years (Atlantic Canada Opportunities Agency, 2012).  119 The important component that makes this program relevant to this review is that the program provides help with bidding and obtaining public and private procurement contracts to develop and commercialize innovative ideas. Expenses related to innovation are eligible for 75% of financing through BDP. BDP is administered by Atlantic Canada Opportunities Agency (ACOA). The final Report of Quality Assurance Audit of BDP (2006) was mostly concerned with assurance that the agency exercised due diligence in the delivery of BDP. The careful analysis of the document reveals how in fact subjective the funding approval process is. Agency uses Personal Summary Forms (PSF) to document the approval process. An audit found that most of the PSF contained the information substantiating funding decision; however, some of the files were missing such key pieces of information as funding sources and details of a project’s costs (Atlantic Canada Opportunities Agency, 2011). The commercial applications are assessed by ACOA in accordance with the agency’s guidelines against three criteria: incrementality, commercial viability and repayability, and net economic benefit. Incrementality criteria establish that “assistance from ACOA is a significant factor in a decision to proceed with the project and that the alternative sources of funds cannot be used to cover the need in financing”. According to the document, the application review process starts with an evaluation by an account manager who uses a set of these three program criteria. At the moment of the funding decision, decision makers are aware of the financial situation of an applicant. In the Application for Assistance form, along with the project’s profile, an applicant must provide summary of costs, proposed financing, and financial obligations towards the project, if any. The attached must be a financial statement for the last  120 fiscal year and 1-3 years of projected statement of income and expenses, as well as personal net worth statement.             Under Budget 2010, the Canadian Government announced a $40 million two-year’ initiative, the Canadian Innovation Commercialization Program (CICP), which had to be implemented through four calls for proposals. The CICP was launched on September 24, 2010 and the last call for proposals is expected to take place in spring 2012. This nationwide program is administered by the Department of Public Works and Government Services Canada (PWGSC), the Office of Small and Medium Enterprises (OSME). It addresses commercialization challenges by connecting innovating firms with potential government users of innovation. Recognizing that actual testing of prototypes in labs may be in fact costly process for many entrepreneurs, Canadian government awards contracts to eligible applicants to test and showcase pre-commercialization stage innovations before launching them to the market. Priority areas are environment, safety and security and health and enabling technologies. The program targets small and medium size enterprises that constitute almost 98% of all businesses in Canada (Industry Canada, 2011). A financial proposal cannot exceed CAN $500,000. The procurement evaluation process consists of two stages. First, in accordance with the Call for Proposals solicitation document, evaluators selected from NRC–IRAP and various government departments rank submissions in descending order in accordance with technical evaluation criteria. The criteria used at this stage are never mentioned at PWGSC’s publicly available resources. Next, the Innovation Selection Committee (ISC) comprising of industry experts mostly from the private sector who have extensive experience in investment, entrepreneurship and innovation and commercialization review and validate these rankings.  121 The Committee members have signed a non-disclosure agreement and a conflict of interest agreement in order to protect applicant’s interests. After the evaluation is complete, the pre-qualified projects are added to an on-line pool of proposals. PWGSC uses the pool to match the top-ranked proposals with a technical department for commercialization. It is done “based on the funding available for the request for proposals”. After the match is found, PWGSC negotiates a contract with the business. A recent Report on Effectiveness of CICP revealed a weakness in the program selection process. The majority (up to 70%) of ISC is from the private sector and might potentially have a strong conflict of interests despite them signing a conflict of interest agreement. Although there was no evidence of conflict of interest presence established, it was recommended to change the composition of the Committee towards higher representation of university members instead (Standing Committee on Government Operations and Estimates, 2012). Applicants must submit their proposal electronically. Part of the submission is a financial proposal detailing costs of the project and financial capacity of the applicant. The latter may include financial statements for the last three years and a confirmation of all other sources of financing has to be used by the applicant, including cash flow, etc. This makes the final decision-maker aware of the financial resources available to the applicant, as well as costs incurred. Despite this very well defined submission process, PWGSC’s online resources do not provide any mentioning or explanation of the criteria used by evaluators and the Innovation Selection Committee, as well as matching processes employed by PWGSC. Agricultural Innovation Program (AIP) – Commercialization Stream is administered by Agriculture and Agri-Food Canada (AAFC). The program is comparatively new and was  122 launched in November 2011. It is expected to continue until March 2013. This federal initiative is a $50 million program aimed to support development and commercialization of new products and new processes in the Canadian agriculture (Ministry of Finance Canada, 2011). AIP consists of two streams that logically complement each other: “knowledge creation and transfer” and “commercialization” streams. The funding within the commercialization stream comes in the form of no-interest re-payable loans to innovating entrepreneurs in agricultural sector. It must be repaid during the 10 year period after the completion of the project. The program is discretionary and non-entitlement. The open application submission process when applications can be filed at any time until the budget was allocated completely makes the competition nature of this program less noticeable. In fact, if several projects are submitted simultaneously they become a competition. The applications are reviewed by AAFC and by external experts if necessary. A final allocation decision is made by AAFC based on a list of criteria including the need for AIP funding, financial capacity to finance the project and repay the loan, technical feasibility, etc. (Agriculture and Agri-Food Canada, 2012). With regard to the assessment process, the program guide lists the whole range of financial documents that have to be submitted for review, including financial statements for three previous years along with project business plan and financing plan.    123 Appendix B  Overview of Several Business Commercialization Support Programs Implemented at the Provincial Level The review of government efforts in the innovation area would be incomplete without regional dimension. All provincial governments in Canada offer a variety of programs and investments to promote business innovation which are reviewed in this sub-section. Figure 13: Gross Domestic Expenditures on R&D Funded by Provincial Government Sector in Canada, total, million $  Source of data: Statistics Canada, CANSIM, 2012, Table 358-0001 Commercialization Support for Business Program (CSBP), Manitoba. In June 2011, a new $30-million Commercialization Support for Business Program has been established by Manitoba Entrepreneurship, Training and Trade. With the major goal of supporting small and medium size Manitoban businesses in seeking to develop and commercialize innovative products, this five-year initiative contributes to Manitoba’s advancement by supporting product and process commercialization and business development in all sectors of the province. The program was introduced with feedback from 0 200 400 600 800 1000 1200 1400 1600 1800 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 Total R&D expenditure Business R&D Expenditure  124 the Manitoba Innovation Council and other provincial authorities that suggested a broader support for commercialization activities in the province. CSBP helps entrepreneurs at every stage of the business lifecycle by providing financial resources and support services. Particularly at the stage of commercialization, this program offers a support of up to $200,000 for moving from prototype to market-ready product. The program offers 50/50 sharing of eligible costs. Applications are evaluated against a set of business and technical criteria to establish eligibility of the project for the program. At the time of application, applicants must demonstrate their ability to fund the project. Applicants will need to provide information to support their claims, such as: company year to date financial statements (income statement and balance sheet), company or personal bank records, letters from financial institution confirming loans, lines of credit, or other forms of proof of financial position. From the information available in public domain, it is not clear how the funding decision is made. Development and Commercialization Fund (DCF), Prince Edward Island. This new initiative was launched in the province in November 2011. It is built as supporting researchers and businesses that already have pre-tested innovation and need help in moving it into the market. Focus areas include bioscience (including agriculture and fisheries), information technology, and renewable energy, aerospace and advanced manufacturing. The program welcomes proponents from both for-profit and not-for-profit sectors. Approved participants are eligible for up to $100,000 of direct funding. The project assessment process is slightly different from the majority of this type of programs. A clearly stated set of criteria and point system used for evaluation of proponent  125 projects makes this program special. Assessment starts with each proposal being evaluated by the Program Officer. Applications preselected at this stage are then evaluated by a Peer Review Committee, which consists of researchers and professionals from across Prince Edward Island. The Committee determines successful proponents evaluating projects against a set of program criteria. In fact, the application is built based on these clearly stated set of criteria. Before filling out the application, innovators know the maximum points they can get for each part of the application that sum up to 100. For example, a project may receive up to 10 points for identification of a commercialization strategy for the developed product. As a result, applicants enter the competition with clear understanding of the ranking process and relevant importance of major areas in application. In addition, applicants have an opportunity to suggest up to three peer reviewers to minimize conflict of interests. The Peer Review Committee develops calculations and ranking of the participating projects, which are a basis for the final funding decision.  Although the peer review process is very detailed, it is not clear how the final funds allocation decision is made by Innovation Prince Edward Island. The program guidelines also state that applicants may be required to submit financial statements to Innovation Prince Edward Island prior to contract negotiations. This requirement makes the decision makers aware of financial position of applicants. Commercialization Program (Newfoundland and Labrador) is implemented by the Department of Innovation, Trade and Rural Development (INTRD). In term of goals, the targeted segment and implementation mechanism of this program is similar to CSBP in Manitoba. It was established in 2007 and has provided $6.95 million up to the 2010 fiscal year (Office of the Auditor General Newfoundland and Labrador, 2011). It targets incorporated businesses that need help with moving their innovative products or services  126 from research stage to market introduction. It provides direct investments and repayable contributions up to 75% of cost to a maximum value of $500,000 per project. During the period of 2007-2010 fiscal years, the aggregated budget of the program constituted $14.76 million and actual expenditures reached only 50% ($6.95 million). Applicants have to attach a business plan of the project and an annual financial statement and balance feet for the current year and two previous years of operation; or if it is a proprietorship, one recent and two previous years of income tax returns. Applicant’s investments into the project must also be shown in application form. An integral part of the application is certification confirming that assistance from INTRD is a significant factor to proceed with the project. Development Officers, Program Directors and Management Committee play key roles in the process of project assessment. The Development Officer establishes initial eligibility of the project. The Officer also completes a comprehensive technical and financial assessment of the project, and submits a Presentation for Funding which includes a recommendation of the appropriate type of investment (repayable loan or equity) to the Program Manager. The Director of the region from which the project originates individually makes the funding decision for projects under $100,000 of requested funding. Funding in amounts over $100,000 is allocated by the Management Committee comprising of INTRD highly ranked stuff. The most recent audit of the program was conducted in 2010 and revealed multiple issues. Since the focus of this thesis is funding distribution by commercialization programs, the following violations contributed to diminished transparency of the funding decisions. In two instances, INTRD had not completed all required the checks of applicant financial standing. Presentations for Funding were not signed by the  127 Development Officer prior to submission to the Management Committee for the final allocation decision (Office of the Auditor General Newfoundland and Labrador, 2011).  The Oil and Gas Manufacturing and Services Export Development Fund (OGMSEDF) is managed by the Department of Innovation, Business, and Rural Development of Newfoundland and Labrador government. The province has committed $2.0 million in 2011/12 for the OGEDF. The Fund supports the manufacturing and fabrication of equipment and provision of services that is used primarily in the upstream or downstream supply activities of the oil and gas industry. One of the Program’s objectives is to support commercialization and introduction of new technologies and processes. Applications for funding under the program are subjects to assessment by the Department. Successful applicants under the OGEDF will normally be eligible to receive up to 50% of total eligible project costs in a non-repayable contribution. As one of the information requirements, applicants must submit a detailed Business Plan including Financial Plan (historical and proforma). Project proposals are submitted to the Assistant Deputy Minister, Department of Business, for review and evaluation. Proposals and supporting documentation supplied by the applicant will be assessed against the program objectives and eligibility criteria that are not described in public domain. The funding allocation decision is made by the Department, but the exact mechanism is not available to the public. Private Sector Emerging Technology Program (Northern Ontario Heritage Fund Corporation). Established in June 1, 1988, The Northern Ontario Heritage Fund Corporation (NOHFC) is a crown corporation and development agency of the Ontario government that invests in northern businesses and municipalities through conditional contributions, forgivable performance loans, incentive term loans and loan guarantees. Eligible initiatives  128 for this program include biotechnology and life sciences projects, etc. Among other forms, funding is provided in the form of non-repayable contributions of up to $100,000 on a cost- shared basis. Financing type and funding sources must be specified. Financing is subject to availability of funds.              The Innovation & Commercialization of New Opportunities for Agri-based Products Program is a federal-territorial-provincial initiative delivered in Nova Scotia by Agriculture and Agri-Food Canada and Nova Scotia Department of Agriculture. It is commercialization focused and specifically addresses the needs of agri-food and agri-based product sector. Applicants are eligible for up to $200,000 of direct contributions during two years. In the first stage, applicants submit a Letter of Intent which is evaluated by the Pre-Screening Committee based on a LOI evaluation form. Applicants approved at this stage submit full proposal which is evaluated and ranked based on Full Proposal Selection Criteria and Weightings Form publicly available. Projects with the highest ranks will be approved budget permitting. No public information is available on the selection of funding recipients which is made by the Selection Committee (Nova Scotia Department of Agriculture, 2011). I-3 Technology Start-Up Competition by Innovacorp (Nova Scotia). Innovacorp is a provincial Crown corporation, falling under the jurisdiction of Nova Scotia Economic and Rural Development and supporting the provincial government’s corporate path for creating a prosperous Nova Scotia. The Innovacorp team moves at the speed of business to help high potential Nova Scotia-based knowledge companies overcome traditional hurdles to business growth. The description of the program can be found at http://innovacorp.ca/about-us/i-3- technology-start-competition Submissions will be evaluated based on: business plan credibility, management experience, a high barrier to competitive entry, a large addressable  129 market and the probability of obtaining a fully-funded business plan. Ultimately, the winning entrants are those determined by the judges to have the highest probability of entering the market in a competitively sustainable fashion and the most commercialization potential.   130 Appendix C  Comparison of Socially Desired and Privately Optimal Levels of Innovation Effort Let us first consider the case when only the private benefits of the innovation project are accounted for, as examined in this sub-section. Since the random shock variable   is uniformly distributed on [0, 1], it follows that         . The left-hand side of this inequality implies   . The right-hand side of it implies     when        It follows that    is the upper bound of the optimal innovation effort, whereas the lower bound is 0. Then, setting )1(* 2 Ke       greater than 0, and remembering that K has to be less than 1 it follows that      . So, when considering only the private benefits of the project, the following conditions have to be satisfied:      and Similar reasoning can be applied to the case when both private and social benefits are considered in calculation of the optimal innovation effort. The random shock variable is between 0 and 1, and     (   )   .  Thus,   , which represents the upper bound of the optimal innovation effort in this case. Then setting    (   ) (   ) greater than 0 requires   (   )  for K<1. So, when accounting for both private and social benefits of the innovation project, the following conditions have to be satisfied:   (   )   and    (   ) Assuming that innovation projects always have positive external value (   ), both the numerator and denominator of  131   (   (   )) (  (   ) )(    ) (   )  are always positive, which implies that        always has to be positive.                       132 Appendix D  Survey Questionnaire    (Group A) Please, imagine yourself being the only judge in the final stage of an innovation competition. After several preliminary rounds, eight innovation projects, each with an estimated value of about $1 million if commercialized, have made it to the final stage. Examples of the short listed projects include a novel food product, a smart phone app and a new livestock gene marker technique. Part A.  Ranking of Projects  In Table 1 the projects have been evaluated by independent industry experts according to three quality criteria: (1) Patentability; (2) Novelty; and (3) Environmental Friendliness. These three quality criteria are key desirable attributes of a successful innovation project. When entrepreneurs first learned about the innovation competition (in the early stages of project development) they were told that patentability, novelty and environmental friendliness carry approximate equal weight for the purpose of selecting the winning projects. Part A.1: Use the data in Table 1 to rank all eight projects according to their overall quality, assigning 1 to the highest quality project and 8 to the lowest quality project. Mark your decision using Arabic numbers 1 through 8 in Table 2. Construct a ranking in a way that makes sense to you. There is no single “correct” approach because in the real world judges such as yourself use many different ranking techniques. The managers of the innovation competition will use your list of ranked projects to award a first place prize equal to $250,000 to the project you ranked as #1, a second place prize equal to $175,000 to the project you ranked as #2 and a third place prize equal to $100,000 to the project you ranked  133 as #3. This experiment lasts approximately 20-25 min., and we suggest you to spend no more than 1/3 of your time on this task.  1. Experts’ Evaluation of Participating Projects Project Summary provided by experts A This project faces 60 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 3/5 with respect to environmental sustainability B This project faces 80 percent chance of receiving a patent, ranks “excellent” in terms of novelty, and scores 5/5 with respect to environmental sustainability C This project faces 80 percent chance of receiving a patent, ranks “very good” in terms of novelty, and scores 4/5 with respect to environmental sustainability D This project faces 60 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 4/5 with respect to environmental sustainability E This project faces 40 percent chance of receiving a patent, ranks “moderate” in terms of novelty, and scores 2/5 with respect to environmental sustainability F This project faces 100 percent chance of receiving a patent, ranks “excellent” in terms of novelty, and scores 5/5 with respect to environmental sustainability G This project faces 80 percent chance of receiving a patent, ranks “very good” in terms of novelty, and scores 3/5 with respect to environmental sustainability H This project faces 40 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 2/5 with respect to environmental sustainability   134 2. Rank Projects based on Quality  Project Assign “1” to your favorite, “2” to your second favorite, …, “8” to your least favorite. A B C D E F G H   Part A.2: Briefly describe the procedure used for coming up with the ranking in Table 2.  135 Part B. Re-Ranking of Projects  Part B.1. Now suppose you are given additional information about the projects, and you have the option to change your project ranking. Specifically, the managers of the innovation competition indicate that their ultimate goal is to assist in the commercialization of the greatest number of high quality projects. In other words, the prize money that you allocated above may be better spent on other projects that can make better use of the financial assistance. The competition managers recognize some individual entrepreneurs may be disadvantaged by your re-ranking, but nevertheless feel that higher wealth for many is in society’s best interest. The managers stress, however, that the re-ranking decisions are yours alone and will likely have an ethical component. The method you use for re-ranking must be described in the space below. Moreover, your re-ranking should be defensible noting that entrepreneurs were previously told that approximate equal weight will be given to the three quality variables when the winning entries are being selected. To assist you with the re-ranking competition managers have provided you with two types of data (see Table 3 on the next page). The first set of data is the probability that each project will be successfully commercialized with: (i) no prize; (ii) with a first place prize; (iii) with a second place prize; and (iv) with a third place prize. The second set of data is the Personal Contribution of each entrepreneur. It reflects the amount of personal money that was allocated to the project by each entrepreneur. This funding is lost to the society if the project fails to be commercialized. Given the information provided to you by competition managers and given the data in Table 3 please re-rank all 8 projects in Table 4.  136  3. Probability of successful commercialization and personal contribution by entrepreneur  Project Amount Awarded Personal Contribution* $0 $100,000 $175,000 $250,000 Probability of Successful Commercialization A 0.1 0.3 0.4 0.5 $9,500 B 0.9 0.95 0.99 1 $50,000 C 0.8 0.85 0.9 0.95 $40,000 D 0 0 0.1 0.9 $10,000 E 0.1 0.1 0.2 0.3 $8,000 F 1 1 1 1 $60,000 G 0.6 0.65 0.7 0.8 $25,000 H 0.7 0.8 0.8 0.9 $9,600 * Personal Contribution refers to the amount of personal money that was allocated to the project by the entrepreneur. This funding is lost for society if the project fails to be commercialized.  4. Re-ranking of projects after reading background information Project Assign “1” to your favorite, “2” to your second favorite, etc. A B C D E F G H Part B.2: Briefly describe the procedure used for coming up with your re-ranking.    General Information about the participant   137 1. What is your gender? Male       _______ Female   _______ 2. What language or languages were routinely spoken in your childhood home?  ______________,  _______________,  ________________ 3. Are you international student? Yes______ No ______  4. What is the highest education completed by your mother and father? Mother Father Grades 0-6  Grades 0-6 Grades 7-9  Grades 7-9 Grades 10-12  Grades 10-12 Some post-secondary  Some post-secondary University graduate  University graduate  5. Approximately, what was your average in your final year of high school? _____ _____ % 6. Please rate your attitude to the statement below. It is important for governments to use tax payer money (e.g., grants, R&D tax credits, innovation competitions), to encourage innovation.  Strongly disagree   1    2    3    4    5    6    7    8    9    10     Strongly agree    THANK YOU VERY MUCH FOR YOUR PARTICIPATION!    138 Appendix E  Survey Questionnaire    (Group B) Please, imagine yourself being the only judge in the final stage of an innovation competition. After several preliminary rounds, eight innovation projects, each with an estimated value of about $1 million if commercialized, have made it to the final stage. Examples of the short listed projects include a novel food product, a smart phone app and a new livestock gene marker technique. Part A.  Ranking of Projects  In Table 1 the projects have been evaluated by independent industry experts according to three quality criteria: (1) Patentability; (2) Novelty; and (3) Environmental Friendliness. These three quality criteria are key desirable attributes of a successful innovation project. When entrepreneurs first learned about the innovation competition (in the early stages of project development) they were told that patentability, novelty and environmental friendliness carry approximate equal weight for the purpose of selecting the winning projects. Part A.1: Use the data in Table 1 to rank all eight projects according to their overall quality, assigning 1 to the highest quality project and 8 to the lowest quality project. Mark your decision using Arabic numbers 1 through 8 in Table 2. Construct a ranking in a way that makes sense to you. There is no single “correct” approach because in the real world judges such as yourself use many different ranking techniques. The managers of the innovation competition will use your list of ranked projects to award a first place prize equal to $250,000 to the project you ranked as #1, a second place prize equal to $175,000 to the project you ranked as #2 and a third place prize equal to $100,000 to the project you ranked  139 as #3. This experiment lasts approximately 20-25 min., and we suggest you to spend no more than 1/3 of your time on this first task.  1. Experts’ Evaluation of Participating Projects Project Summary provided by experts A This project faces 60 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 3/5 with respect to environmental sustainability B This project faces 80 percent chance of receiving a patent, ranks “excellent” in terms of novelty, and scores 5/5 with respect to environmental sustainability C This project faces 80 percent chance of receiving a patent, ranks “very good” in terms of novelty, and scores 4/5 with respect to environmental sustainability D This project faces 60 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 4/5 with respect to environmental sustainability E This project faces 40 percent chance of receiving a patent, ranks “moderate” in terms of novelty, and scores 2/5 with respect to environmental sustainability F This project faces 100 percent chance of receiving a patent, ranks “excellent” in terms of novelty, and scores 5/5 with respect to environmental sustainability G This project faces 80 percent chance of receiving a patent, ranks “very good” in terms of novelty, and scores 3/5 with respect to environmental sustainability H This project faces 40 percent chance of receiving a patent, ranks “good” in terms of novelty, and scores 2/5 with respect to environmental sustainability   140 2. Rank Projects based on Quality  Project Assign “1” to your favorite, “2” to your second favorite, …, “8” to your least favorite. A B C D E F G H   Part A.2: Briefly describe the procedure used for coming up with the ranking in table 2 2.   141 Part B. Re-Ranking of Projects  Part B.1. Now suppose you are given additional information about the projects, and you have the option to change your project ranking. Specifically, the managers of the innovation competition indicate that their ultimate goal is to assist in the commercialization of the greatest number of high quality projects. In other words, the prize money that you allocated above may be better spent on other projects that can make better use of the financial assistance. The competition managers recognize some individual entrepreneurs may be disadvantaged by your re-ranking, but nevertheless feel that higher wealth for many is in society’s best interest. The managers stress, however, that the re-ranking decision is yours alone and will likely have an ethical component. The method you use for re-ranking must be described in the space below. Moreover, your re-ranking should be defensible noting that entrepreneurs were previously told that approximate equal weight will be given to the three quality variables when the winning entries are being selected. To assist you with the re-ranking competition managers have provided you with additional information (see Table 3 on the next page). The data is the probability that each project will be successfully commercialized with: (i) no prize; (ii) with a first place prize; (iii) with a second place prize; and (iv) with a third place prize.  Given the information provided to you by competition managers and given the data in Table 3, please re-rank all 8 projects in Table 4.   142 3. Probability of successful commercialization  Project Amount Awarded $0 $100,000 $175,000 $250,000 Probability of Successful Commercialization A 0.1 0.3 0.4 0.5 B 0.9 0.95 0.99 1 C 0.8 0.85 0.9 0.95 D 0 0 0.1 0.9 E 0.1 0.1 0.2 0.3 F 1 1 1 1 G 0.6 0.65 0.7 0.8 H 0.7 0.8 0.8 0.9  4. Re-ranking of projects after reading background information Project Assign “1” to your favorite, “2” to your second favorite, etc. A B C D E F G H  Part B.2: Briefly describe the procedure used for coming up with your ranking in Table 4.    General Information about the participant 1. What is your gender?   143 1. What is your gender? Male       _______ Female   _______ 2. What language or languages were routinely spoken in your childhood home?  ______________,  _______________,  ________________ 3. Are you international student? Yes______ No ______  4. What is the highest education completed by your mother and father? Mother Father Grades 0-6  Grades 0-6 Grades 7-9  Grades 7-9 Grades 10-12  Grades 10-12 Some post-secondary  Some post-secondary University graduate  University graduate  5. Approximately, what was your average in your final year of high school? _____ _____ % 6. Please rate your attitude to the statement below (by circling the appropriate number).  It is important for governments to use tax payer money (e.g., grants, R&D tax credits, innovation competitions), to encourage innovation.  Strongly disagree   1    2    3    4    5    6    7    8    9    10     Strongly agree   THANK YOU VERY MUCH FOR YOUR PARTICIPATION!    144 Appendix F  Socio-Demographic Characteristics of the Respondents  Table 12: Socio-Demographic Characteristics of the Respondents Socio-demographic characteristics Group A Group B Male  11 10 Female 25 26 International  students 6 Domestic students 30 Languages routinely spoken in childhood home      Chinese-English 15 12      English 11 9      Chinese 1 7      Korean 2 -      Korean-English 2 -      Other 5 8 Highest education completed by mother      University graduate 17 15      Some post-secondary 15 13      Grades 10-12 4 5 Highest education completed by father      University graduate 23 22      Some post-secondary 11 8      Grades 10-12 2 3      Other  - 3 High school average      90+ 11 12      85-99 12 12      80-84 8 7      75-79 3 4      below 75 2 1 Ranking of a statement about innovation      8-10 19 22      5-7 15 12  145 Appendix G Bootstrapping Results 1. Coefficient of Structural Concordance of re-rankings in two groups  Simulation Stats        1000 repetitions        153 seconds  Summary Statistics Notes  Average 0.923    SD 0.0190    Max 0.969    Min 0.854                      0.8525 0.8725 0.8925 0.9125 0.9325 0.9525 Histogram of data rankings  146 2. Coefficient of Structural Concordance of Original Rankings in two groups Simulation Stats        1000 repetitions        157 seconds  Summary Statistics Notes  Average 0.667    SD 0.0459    Max 0.793    Min 0.523                         0.52 0.57 0.62 0.67 0.72 0.77 Histogram

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0072776/manifest

Comment

Related Items