UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Japanese manufacturing greenfields : the provincial location decision 1993

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_1993_spring_brown_iain_andrew.pdf [ 4.31MB ]
ubc_1993_spring_brown_iain_andrew.pdf
Metadata
JSON: 1.0086082.json
JSON-LD: 1.0086082+ld.json
RDF/XML (Pretty): 1.0086082.xml
RDF/JSON: 1.0086082+rdf.json
Turtle: 1.0086082+rdf-turtle.txt
N-Triples: 1.0086082+rdf-ntriples.txt
Citation
1.0086082.ris

Full Text

JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision by IAIN ANDREW BROWN B.Comm., The University of British Columbia, 1982 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN BUSINESS ADMINISTRATION in THE FACULTY OF COMMERCE AND BUSINESS ADMINISTRATION (Department of International Business) We accept this thesis as confirming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1993 ©Iain Andrew Brown In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Signature)  Department of International Business The University of British Columbia Vancouver, Canada Date April 21, 1993 DE-6 (2/88) ABSTRACT This paper examines why Japanese manufacturing greenfields locate in a particular Canadian province. We find that the location preference is based primarily on the present distribution of Japanese and Canadian firms. A secondary factor is market access, which includes the transportation costs of exporting the Canadian manufactured product to Japan. Other important factors are energy and labour costs. Having utilized quantitative methods to determine that the presence of Japanese and Canadian firms are the main reasons why new greenfields select the province they will locate in, we question the value of using tax dollars to attract investments to locations lacking substantial industry activity. TABLE OF CONTENTS ABSTRACT ^  ii TABLE OF CONTENTS ^  iii LIST OF TABLES LIST OF FIGURES^  vi ACKNOWLEDGEMENT  vii 1. INTRODUCTION ^  1 2. JAPANESE MANUFACTURING INVESTMENT IN CANADA . . . 3 2.1. The Greenfield Component of Foreign Investment ^ 3 2.2. Japanese FDI ^  8 3. RELATED RESEARCH  10 4. COMPILING OUR DATA^  15 4.1. Dependent Variable - Japanese Manufacturing Greenfields^15 4.2. Independent Variables - Provincial Characteristics ^ 16 4.3. Ship Rank, Energy and MetGDP ^  19 4.4. Industry Agglomeration as a measure of Manufacturing Activity ^  22 4.5. Data Collection Summary  24 5. OUR CONDITION LOGIT ECONOMETRIC MODEL ^ 26 6. RESULTS ^  30 7. CONCLUSION  36 7.1. Factors Influencing Japanese Manufacturing Greenfield's Location Decision ^  36 7.2. Limitations of Our Results  38 7.3. Policy Ramifications for Provinces Seeking FDI ^ 40 7.4. Further Study and Extensions ^  41 8. BIBLIOGRAPHY ^  43 APPENDIX 1: FIFTY-FIVE JAPANESE MANUFACTURING GREENFIELDS IN CANADA ^  46 APPENDIX 2: PROVINCIAL CHARACTERISTICS ^ 49 APPENDIX 3: NOTES TO JAPANESE MANUFACTURING GREENFIELDS AND PROVINCIAL DATA ^ 52 APPENDIX 4: JAPANESE MANUFACTURING GREENFIELDS - PROVINCIAL INDUSTRIAL LEVEL FOR 1987 (WAGES, FUEL, and REVENUE)   53 APPENDIX 5: JAPANESE MANUFACTURING GREENFIELDS - NATIONAL INDUSTRIAL LEVEL FOR 1987 (WAGES, FUEL, and SHIPMENTS)   60 APPENDIX 6: CONDITIONAL LOGIT REGRESSION RESULTS - VARIABLES & THEIR SOURCE ^  62 LIST OF TABLES TABLE 4.1: Number of Japanese Manufacturing Greenfields in Canada ^ 16 TABLE 4.2: 1989 (1980) Provincial Ranking ^  25 TABLE 5.1: Energy Intensity ^  29 TABLE 6.1: Final Results  30 TABLE 6.2: Substituting Low Energy Price ^  31 TABLE 6.3: Testing Unionization, Tax & Crime Rates ^ 32 TABLE 6.4: Testing Factor Intensities ^  33 TABLE 6.5: Replacing MetGDP with Quebec Dummy ^ 34 TABLE 6.6: Adding a Quebec Dummy Variable  35 LIST OF FIGURES FIGURE 2.1: Canada's Liabilities to Non-Residents ^ 5 FIGURE 2.2: Acquisition and Greenfield FDI Inflows  7 FIGURE 2.3: 1989 Worldwide Gross Outflow of FDI ^ 8 FIGURE 4.1: Provincial Lowest Energy Costs  21 FIGURE 4.2: Provincial Average Energy Costs ^  21 FIGURE 6.1: MetGDP Variable ^  34 ACKNOWLEDGEMENT I thank all those who have been inspirational during my work on this thesis. I am particularly indebted to John Ries for his assistance throughout the preparation of this paper. My gratitude for John's contribution is only dwarfed by my respect for his unselfish dedication to the project. Keith Head's assistance in the development of our model was most appreciated. I thank Jim Brander for helping to polish and structure this paper for final presentation. Anna Kwan deserves recognition for compiling this manuscript. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 1 1. INTRODUCTION In recent years Canada has sought new Japanese manufacturing investments. (New investments are often referred to as greenfield investments). Public policy interest in greenfield investment arises in part because of the political benefits of job creation that are associated with them. Another reason that provinces are interested in attracting Japanese manufacturing greenfields is technology transfer. However, job creation tends to be the host province's main motivation in soliciting Japanese manufacturing greenfields. Regardless of the reason(s) why provinces may wish to attract Japanese manufacturing greenfields, in order to have better success in attracting such investments, it would be useful for the provinces to know what attracts Japanese manufacturing greenfields to a particular province. This thesis focuses on this issue by asking the question what provincial characteristics influenced the greenfields to locate in the particular province they selected? It turns out that between 1980 and 1991, 84% of Japanese manufacturing greenfield investment locating in Canada established their new facilities in either Ontario (64 %) or British Columbia (20 %). Both these provinces attracted more greenfield investment than their national share of manufacturing. By using condition logit regression we discover that Japanese manufacturing greenfields prefer to locate in provinces with a concentration of Canadian firms in their industries, and in provinces with other Japanese manufacturing greenfields. Other important provincial characteristics in determining location preference include wage and energy lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 2 costs as well as geographic proximity to Japan. Past research on American state location decisions, which used similar statistical methodology as we employ, supports our findings in that state characteristics, similar to our provincial characteristics, are identified as reasons for manufacturing firms establishing new facilities in a particular location. However, our results may suggest evidence of pure agglomeration effects. That is, if the combination of our variable designed to measure geographic proximity to Japan and our variable that counts the number of already existing Canadian firms in the establishing greenfield's industry, capture the endowment effect, then the significance of our variable measuring the size of existing Japanese investment may be evidence of pure agglomeration effects. Section 2 provides some background on Japanese investment. Section 3 reviews related research. Section 4 explains how the data for our variables was compiled. Section 5 describes our econometric model. Section 6 summarizes our results. Section 7 is our conclusion that includes a discussion on the limitation of our results, policy implications and research questions that follow naturally from this study. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 3 2. JAPANESE MANUFACTURING INVESTMENT IN CANADA The purpose of this section is to familiarize the reader with foreign investment so that an appreciation is gained for what we mean by a Japanese manufacturing greenfield. We point out some of the possible benefits that are associated with foreign direct investment. However, we are not saying that other forms of investment are any less interesting or beneficial, nor that our methodology would only work for Japanese investments. Our methodology could have been used to study, for example, the provincial location decision of German greenfields, but we elected to study Japanese greenfields and hence this why we provide an overview of Japanese manufacturing investment in Canada. 2.1. The Greenfield Component of Foreign Investment Before addressing the issue of why we chose to measure the greenfield component of Japanese foreign direct investment (as opposed to, say, American direct investment) let us first distinguish between foreign portfolio investment (FPI) and foreign direct investment (FDI). FDI is ownership (with control) of real domestic assets by a foreigner. Statistics Canada considers foreign ownership to exist when a foreigner owns more that 10% of the equity of an investment'. Usually, FDI is undertaken by corporations to take advantage of the 1^As Appendix 1 shows the lowest level of Japanese ownership was 17%, with over half the greenfield investments being wholly owned subsidiaries. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 4 comparative advantage that foreign production offers or to preserve access to markets in an environment where true global free trade does not exist (i.e., Japan exporting to Canada or the U.S., and for that matter, vice versa). FDI typically involves the transfer of capital as well as technology, marketing and organizational skills. Management practises such as just-in-time (kanban) inventory control procedures lessen inventory carrying costs, quality circles (kaizen) reduce the number of defects, and consensus decision making (nemawashi) increase productivity, all originated in Japan. However, through technology transfer these management practises are no longer solely the property of Japanese transplants operating in Canada. They have also spread to Canadian owned manufacturing facilities, which now benefit from them. These examples of technology transfer, along with job creation, may motivate provinces to attract FDI, even though an increase in foreign ownership of this domestic industry will occur as ownership of these assets remaining with the foreign firm. This ownership can take the form of acquisitions or the construction of new facilities (greenfields). Other forms of FDI involve the establishment of a wholly-owned subsidiary or the formation of joint ventures and strategic alliances between firms. (Although strategic alliances may not involve the transfer of capital they do tend to have the other benefits that are associated with FDI). The alternative type of business structure to be used in the host country is important as the choice of organizational form often has significant implications for the transfer of knowledge and other firm-specific skills. Canada's Liabilities to Non-res (Cdn BI) Corp. Bonds 46.5Cdn. Gov . Bonds 63.2 STOCKS 20.8 Other. 108.7 Money Market Sec. 25.4 FDI 126.6 Prov. Gov . Bonds 65 Figure 2.1 (Source: Statistics Canada, 1990). lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 5 Finally, as FDI involves the transfer relatively intangible resources, the corporations involved typically operate with a fairly long time horizon, making flows of FDI stable in comparison to flows of FPI. However, as we see from Figure 2.1, FDI is about one quarter of the stock of foreign investment in Canada as of 1990. FDI is represented by the black wedge in Figure 2.1, and the remaining pieces of the pie are collectively known as foreign portfolio investment. FPI occurs when foreigners own financial assets from another country. Figure 2.1 2 illustrates that in 1990 foreigners owned $329.6 billion worth of Canadian financial assets. FPI is undertaken mainly by individuals, institutional investors and governments (i.e. central banks). FPI involves transactions in the capital markets (e.g., stocks, bonds, loans or money market securities). Those involved in FPI have different objectives, risk preferences, and investment time horizons than those involved in FDI. The risk-adjusted real rates of return 2 The 108.7 of Other Liabilities to Non-residents includes SDRs (Special Drawing Rights) and other official government flows. Also theses stocks are at historical costs and if the FDI was made some time ago its market value may exceed its book value by a considerable margin. Thus, its is possible that if market figures were available that FDI would represent a bigger slice of the pie. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 6 between assets in different countries is a major determinant of FPI flows. Thus, FPI flows are sensitive to changes in inflation, interest and exchange rates as well as portfolio diversification requirements. Accordingly, FPI tends to be in financial assets that are relatively liquid and mobile with fixed maturity dates. Just as FDI and FPI are less than perfect substitutes, acquisitions and greenfields are different types of FDI, if only for the reason that the former does not involve the same location decision. The location of a new plant is clearly a choice; whereas, it could be argued that an acquisition did not involve a location analysis because the site already exists. When a company is acquired, it may be acquired despite of the location of its manufacturing facilities. That is, in an ideal situation the purchaser would prefer the manufacturing facilities to be elsewhere (perhaps in another province). Nevertheless, if the purchase takes place, because the overall investment makes sense, then this non-ideal location was not actually chosen. However, in a greenfield investment, the ideal location would be selected. Thus, we have included only Japanese greenfields in our study. A secondary reason for only considering greenfields is unique to Canada: what we call "the FIRA (Foreign Investment Review Agency) effect." Prior to 1985 FIRA (a government organization designed to discourage foreign investment) may have inhibited acquisitions. Greenfields apparently did not offset the reduced level of acquisitions during the FIRA years (1974-1985). Presumably a Canadian greenfield was not the foreign firm's second choice after a Canadian acquisition and the direct investment simply went elsewhere. Acquisition and Greenfield FDI Inflows Cdn. Billions 14 12 10 8 6 4 2 0 1980 1981 1982 1983 1984 1985 1986 1987 198819891990• Greenfield* Acquisitions Investment Canada Figure 2.2 lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 7 In 1985 FIRA was replaced by Investment Canada, which is a government body with a mandate to promote foreign investment. The mid 80's were also a period of heightened merger and acquisition activity. We suspect that a combination of these factors explains the increase of acquisitions in 1985 (refer to Figure 2.2). Regardless of the reason, acquisitions dramatically increased in 1985. This imbalance of acquisition activity from the beginning of the decade to the end of the decade is another reason we have elected to study the greenfield component of FDI. In summary, FDI and FPI cannot be treated as perfect or even close substitutes because they are motivated by different factors. We have elected to measure FDI rather than FPI. We selected the greenfield component of FDI to avoid the FIRA effect as well as any controversy over an acquisition being a "non-location choice." • France A Germany ■ Japan United Kingdom United States ^ Canada • Other Developed E Developing Countries •::: 1989 Worldwide Gross Outflow of FDI (%) France Developing Countries 9.9^4.5 United Kingdom 16.3 United Nation, World Investment Report United States 13.5 lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 8 2.2. Japanese FDI Another reason that we are interested in Japanese greenfield FDI is Japan's recent dominance as an exporter of FDI. For example, Figure 2.3 shows Japan was the leading exporter of FDI in 1989. With Japan being a prominent world exporter of FDI this is one reason to study Japanese FDI. Secondly, although American FDI into Canada is also significant, in future work we wish to objectively measure the effect of the Canada/U.S. Free Trade Agreement on FDI into this trade area. To do this we will need to measure the location decision of an outside party and by definition this excludes the U.S. Figure 2.3 For these reasons we have chosen to analyze Japanese greenfield FDI. We have elected to look at manufacturing greenfields (as opposed to say distribution warehouses) for reasons such as the technology transfer and job creation benefits associated with these investments. The time frame of our study is 1980 to 1991. This time frame was selected because it has really only been the last decade that Japanese FDI has become such a noticeable factor in industrial host countries. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 9 However, host countries are generally made up of regional political territories and in Canada these are called provinces. Often these provinces try to out bid each other (i.e., with tax holidays) for greenfield investments, possibly because of the job creation or technology transfer associated with these new manufacturing facilities. Regardless of the reason(s) why provinces seek to host FDI, to gain insight into how host provincial governments may be more successful in attracting Japanese manufacturing greenfields we will study the provincial location decision of these greenfields. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 10 3. RELATED RESEARCH Before we describe how the data was collected for this study, we will first discuss related research to this paper. These studies have aided us in selecting the variables that we will present in the next section and in designing the conditional logit econometric model that we will describe in Section 5. In this regard we note that logit regression was used by Bartik (1985), Carlton (1983), Coughlin, et. al. (1991), Head, Ries, and Swenson (1993), Luger and Shetty (1985), Schmenner, et. al. (1987), and Woodward (1992). Bartik (1985) looked at the influence of state characteristics on new manufacturing plant location decisions, emphasizing unionization and taxation. The location data included the Fortune 500 companies' manufacturing plants from 1972-1978. He found a significantly large negative effect of a state's unionized rate, while wages had only a marginal negative significance. The corporate tax rate also had a negative effect on plant location. Bartik estimated a powerful effect of existing manufacturing activity on new business location, i.e., "a state with 10% greater existing manufacturing activity will have an 8% or 9% greater number of new plants" (Bartik, 1985). Schmenner, et. al. (1987) also examined new plant openings by Fortune 500 firms between 1970-1980. Their innovations were to include plant-specific characteristics (data was derived from surveying managers of the firms) in magnifying or tempering the state-specific lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 11 effects. They argued that state characteristics such as wage rates would be more important to particular plants (i.e., wages will matter more to labour intensive plants). By interacting state and plant characteristics, Schmenner found that "Labor unionism, climate, population density, and building costs are prominent influences on location. . . Tax variables, on the other hand, make a weak showing" (Schmenner, et. al., 1987). Also the interaction terms were usually significant. Carlton (1983) concentrated on a few industries that were selected to ensure equivalence (i.e., location choice is not hindered by local supply and demand factors). His results included: [the] "wage effect cannot be measured very precisely; energy [prices] have a large effect; taxes and state incentive programs do not seem to have major effects; [and] existing concentrations of employment matter a great deal with the effect being stronger for industries with smaller average plant size" (Carlton, 1983). Insignificance of taxes was also evident in Luger & Shetty (1985), who measured the manufacturing activities of three specific industries. Their objective was to measure the elasticity of FDI in relation to a state's promotional programs, as well as the effect of agglomeration and urbanization economies, and labour market conditions. They concluded "that agglomeration economies and wage rates are the most important determinants of new plant location. . . [and] public policies do not have a uniform effect on industries" (Luger & Shetty, 1985). To represent agglomeration attraction a measure of total manufacturing activity was used; thus they did not provide evidence on industry-level agglomeration. Moreover, unmeasured state lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 12 endowments (forests, research universities, or other omitted state characteristics) might be reflected in the coefficient on manufacturing activity. Like Luger & Shetty before them, Smith & Florida (1992) chose to explore a select industry (automobile-related) and the industry-specific relationships associated with it. They concluded that Japanese auto-related parts suppliers tend to locate near their assembly plant and have a preference for areas with greater aggregate manufacturing activity. Taxes had a marginal showing and unionization was not a factor, while wages had a positive relationship. This positive relationship could be due to the need for technical skills that might be reflected in wages. For similar reasons, education level also had a significant positive effect. Coughlin's (1991) research data included all FDI transactions (acquisition, equity increase, joint venture, merger, new plant or plant extension). His results indicated that FDI located in states with higher per capita incomes and higher levels of manufacturing activity. Coughlin also found that FDI is attracted to areas with lower wages and higher unemployment rates, while higher taxes deterred FDI. Unionization had a marginal positive significance, which could be "because of the increased productive efficiency in manufacturing stemming from unionization" (Coughlin, et. al., 1991). Woodward's (1992) study involved a measure of total manufacturing activity, rather than selecting certain industries. He analyzed Japanese-affiliated manufacturing investments in the U.S. between 1980-1989. The model worked on the assumption that lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 13 "Japanese firms, like all firms, seek branch locations with the highest expected profits" (Woodward, 1992). This relationship revealed a preference for strong markets and low unionization rates. In addition, "Japanese manufacturing plants are most likely to select counties characterized by manufacturing agglomeration, low unemployment and poverty rates, and concentrations of educated, productive workers" (Woodward, 1992). Head, Ries, and Swenson (1993) found that Japanese firms tend to locate near both other Japanese firms and U.S. firms that were in the same specific industry. They were able to use industry-specific variables for both Japanese and American companies and found both to be significant. They also found that the attractiveness of a state is increased by the level of industrial activity on bordering states. They showed that the positive relationship between industry activity and location is partly due to agglomeration externalities, not simply an endowment effect. As not all the research measured the same variables, it is difficult to generalize the findings of these studies. However, we are able to draw some conclusions from these papers. Intuitively, unionization, wages, and taxes should pay significant roles in the location decision. Bartik (1985), Schmenner (1987), and Woodward (1992) suggested unionization is negatively correlated with FDI, while Coughlin (1991) showed a marginal positive relationship. Luger & Shetty (1985) and Coughlin (1991) concluded a negative significance for wages, while Smith & Florida (1992) took the opposite view. The differences in these results may stem from industry- specific differences, in that capital intensive industry may be willing to "put up" with higher lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 14 wages and unionization in exchange for better trained and more productive workers. While labour intensive industries would place more importance on wages. Hence factor intensities could be the reason for the different findings. Taxes are generally found to be insignificant or inconclusive, with the exception of Bartik (1985) who found taxes had a negative effect. While never quite defined the same, when a variable is used to measure the level of manufacturing activity the results are usually that "existing manufacturing" attracts new manufacturing. In our next section we build on this research to explain why Japanese manufacturing greenfields locate in the particular province that they do, once they have elected to enter the Canadian market. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 15 4. COMPILING OUR DATA 4.1. Dependent Variable - Japanese Manufacturing Greenfields Data collection for this study was itself a major task as this data is not readily available. Furthermore, data available on Japanese-owned manufacturing facilities in Canada do not always clearly distinguish greenfield operations from acquisitions. To construct our detailed list we obtained much of our information from Japanese sources (Toyo, Jetro, Dodwell). At times these directories of Japanese companies in Canada provided conflicting information. Not one of them included all the companies that appear in Appendix 1. Furthermore, some information had to be verified by contacting the firms directly. This was often the case when Jetro claimed that a firm was a distribution manufacturing firm, but this firm was not listed in other manufacturing directories. It usually turned out that these firms were solely distribution companies. We found Toyo to be most reliable. We supplement the data from these three directories with information on Japanese companies in Ontario compiled by the Government of Ontario. We compiled a comprehensive list of fifty-five Japanese manufacturing greenfields that had located in Canada since 1980 (see Appendix 1). These investments were spread over six provinces, but were mainly concentrated in British Columbia (BC) and Ontario. Table 4.1 lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 16 portrays the distribution and growth of these firms in Canada during the period from 1980 to 1990. For example, we note that British Columbia received one investment in 1980 and two more Japanese greenfields located in the province in 1983; thus, from 1980 to 1983 three greenfields had located in this province. TABLE 4.1 NUMBER OF JAPANESE MANUFACTURING GREENFIELDS IN CANADA (Accumulated Totals Since 1980) Province 1980 1981 1982 1983 1984 I^1985 1986 1987 1988 1989 1990 Alberta 0 0 0 0 0 0 0 0 0 0 1 BC 1 1 1 3 3 3 3 3 5 8 11 NB 1 1 1 1 1 1 1 1 1 2 2 Ontario 4 5 6 7 10 13 20 27 30 33 35 Quebec 0 0 0 0 0 0 2 2 2 4 4 Sask. 0 0 0 0 0 0 0 1 2 2 2 Canada 6 7 8 11^I 14^I 17 26 34 40 49 55 BC = British Columbia^NB = New Brunswick^Sask. = Saskatchewan 4.2. Independent Variables - Provincial Characteristics With our dependent variable data collected we turned our attention to establishing and gathering factors (our independent variables) that may influence the selection of the province to locate in. To help determine what these factors may be we relied on recent work [Bartik (1985), Carlton (1983), Coughlin (1991), Head, Ries, and Swenson (1993), Luger and Shetty lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 17 (1985), Schmenner (1987), Smith & Florida (1992), and Woodward (1992)] that has looked at similar location decisions in the United States. These papers focused on state choices and thus were very applicable for our study on provincial location choices. The following list represents independent variables for which we collected Canadian provincial data. A brief description of each is contained in this list. (For more detail and actual time series data see Appendix 2, with notes to this data in Appendix 3). With the exception of our Ship Rank, Average Energy and MetGDP variables (which are described in further detail in the next section) and our Canadian Industries and Japanese Greenfields variables (described in Section 4.4) the following variables are relatively self explanatory. 1. Provincial Population (Cansim'). This is a measure designed to capture the relative size of the provincial economy. 2. Percentage of population living in Metropolitan areas (Cansim). Although the population of the province may be larger than another province, if it is very spread out this tends to hamper economic activity. Thus, a measure of concentration of the province's population is useful. In Canada metropolitan areas include cities with over 100,000 people. 3. Crime rate, a probabilistic variable is a measure of reported offenses per 100,000 people (Cansim). This variable is included as a measure of quality of life, which is increasingly receiving more attention. Thus, we thought that it was appropriate to see if the quality of life decision effected the provincial location process of Japanese manufacturing greenfields. 4. Unemployment rate, which has both positive and negative attributes associated with it. For instances, high unemployment is positive in terms of available work force, but is a negative in terms purchasing power (Cansim). Nevertheless, Computerized on-line database of Statistics Canada. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 18 despite the mixed effects of this variable, it is a general economic indicator; thus, its inclusion in our study. 5. Unionization rate is for total provincial labour force; however, for the years we checked, the manufacturing unionization rate did not significantly differ (Statistics Canada, Advisory Services). Unionization is include in our model as manufactures often are said to avoid it. However, unionization may also be a measure of quality of the work force and as such capital intensive industries could see it as a desirable quality. 6. Area in square kilometres not including fresh water (Canada Year Book). We include this variable as a measure of room to grow and natural resources. 7. All production managers tend to be concerned with the cost of average manufacturing Wages (Manufacturing Industries of Canada). However, high Wages may also be associated with skilled productive workers and as such high Wages in themselves may not be a bad thing. 8. Provincial GDP (Cansim). This variable is included as a measure of market size. 9. Highway Kilometres per square kilometre (Roads and Transportation Association of Canada) is designed to measure the level of existing infrastructure, which is thought to be a factor in attracting new industry. 10. Average Energy is a weighted average cost of fuel in the province (Energy Statistics Hand Book). The weights are based on existing industrial use for 1992, of each fuel, in each province. The cost of energy is generally of concern to manufacturers and as such is included in our study. 11. Low Energy is a measure of the lowest priced form of energy in the provinces at the time the Japanese manufacturing greenfield was established (Energy Statistics Hand Book). The cost of energy is important to manufacturers and as such is included in our study. 12. Airports with control towers (Transport Canada) are thought to represent a measure of transportation infrastructure as well as the preferred choice of executive transportation. Hence its inclusion. 13. Provincial corporate Tax rate (Canadian Tax Reporter). As businesses are usually established to make profits, the tax rate that they will face is an important factor and possibly may determine the manufacturing facility's provincial location. However, high taxes in themselves are not necessarily bad. It depends on what the tax dollars are spent on, and how efficiently this process is carried out. For lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 19 example, if tax dollars are spent effectively on improving needed infrastructure, then this may be seen as a positive. 14. Percentage of Provincial Labour Force in Manufacturing is a measure of provincial manufacturing activity (Cansim). It has been included in our study as a possible reason for plant location because Japanese manufacturing greenfields may only be attract to the relatively more industrial provinces. 15. The percentage of Canadian Manufacturing Labour Force that resides in the province is a measure of manufacturing activity (Cansim). It has been included in our study as a possible reason for plant location because Japanese manufacturing greenfields may only be attracted to the relatively more industrial provinces. 16. Agglomeration of specific Canadian Industries in each province is a measure of industry-specific manufacturing activity (Manufacturing Industries of Canada) in each province. This variable is included as firms in the same industry may be attracted to each other. 17. Agglomeration of Japanese Greenfields in each province is a measure of Japanese manufacturing activity (Toyo, Jetro, Dodwell). We have included the data in Table 4.1, as we wish to test if Japanese manufacturing greenfields are attracted to provinces that already have a base of such firms. 18. Ship Rank is a measure of ease of water access to Japan and is important when the Japanese-owned firm is shipping the manufactured product back to Japan. This variable could also be of some important when parts are being shipped to Canada to be assembled (although closeness to the market may be more important). 4.3. Ship Rank, Energy and MetGDP Our Ship Rank variable is a combination of geographic location and port facilities and is a measures of ease of water access to Japan. This is important not only for shipping over parts to be assembled in Canada, but also to ship back finished goods to Japan, such as lumber from Japanese-owned sawmills in British Columbia. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 20 British Columbia is ranked highest because it has deep water ports and the closest access to Japan. New Brunswick, with a deep water port, but requiring ships bound for Japan to utilize the Panama Canal, is second. Quebec, which requires a trip up the St. Lawrence Seaway, is third. Ontario fairs worse than Quebec, due to the canal that must be travelled to enter the Great Lakes from Montreal. The remaining two provinces hosting Japanese manufacturing, are the only two land locked provinces in the country; thus, Alberta and Saskatchewan score worse in this category (Appendix 2 presents a Ship Rank for all provinces). Apart from water transportation, manufacturing goods are commonly moved by both rail and truck. However, comparing the length of rail tracks in each province is less favourable than comparing the total paved highway kilometres, because of the history of Canadian rail development. Rail development was more prominent in the east during the earlier years of Canadian industrialization; whereas, road transportation developed later and as such lacks the eastern bias associated with rail development. Thus, we have elected to concentrate on highway, rather than rail kilometres. Along with our shipping variable we also devised our own method for measuring energy costs. Our Low Energy variable is a measure of the lowest priced fuel (among crude oil, natural gas or electricity) that prevailed at the time the Japanese decided to establish the greenfield in Canada. The second variable we constructed (which was also calculated by converting fuel prices to dollars per million Btu's) was a weighted average cost variable. Our Average Energy variable was calculated utilizing weights based on existing industrial Provincial Lowest Energy Costs 7 6— 5- B4— 0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 —Alberta^+British Columbia *New Brunswick + Ontario Quebec^+ Saskatchewan Statistics Canada, Energy Statistics Hand Book Provincial Average Energy Costs 10 2— 0 ^ 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 --Alberta^+ British Columbia *New Brunswick + Ontario * Quebec^+ Saskatchewan Statistics Canada, Energy Statistics Hand Book Figure 4.1 Figure 4.2 lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 21 consumption of fuel in each province. These two variables are displayed graphically in Figure 4.1 and Figure 4.2. Finally, let us explain our MetGDP variable, which is designed to capture the characteristics of the provincial market, which is also important to foreign investors. In general, a strong market is found where there is a concentration of people with purchasing power. Thus, we would expect the combination of urban concentrations and gross domestic provincial product to be an influential factor. Urban concentration is represented by the percentage of the provincial population living in a city with a population greater than 100,000 people. To arrive at our single measure of market size (MetGDP) we multiplied GDP by Metropolitan. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 22 4.4. Industry Agglomeration as a measure of Manufacturing Activity Industry agglomeration in itself may attract additional industry. That is industry attracts industry. Krugman (1991) suggests that users and suppliers of intermediate inputs will locate around one another in an effort to minimize transportation costs and encourage economies of scale. Also, technological spillovers may cause firms in the same industry to locate in proximity to each other. Smith & Florida (1992) point out that Japanese auto part suppliers tend to locate near Japanese auto manufacturers. Bartik (1985) attributes "existing manufacturing activity" as the reason for other manufactures locating there. Our study addresses this issue of whether the existence of manufacturing attracts new different kind of manufacturing to the area. If so, we would expect our variables measuring manufacturing intensity (i.e., variables 14 and 15 in Section 4.2) to be reasons for Japanese greenfields locating in the province that they do. If however, the level of provincial overall manufacturing activity is not as good a measure as industry-specific clusters are of provincial location choice, then we would expect our Canadian Industries and Japanese Greenfields variables to be more significant. Building on the work of Head, Ries, and Swenson (1993), our Canadian Industries and Japanese Greenfields variables are counts of establishments in specific industries, accumulated at the provincial level. To determine Canadian Industries we utilised SIC classifications (based on 4, or in some cases 3, digits). That is, on a provincial basis we counted all Canadian industries that had the same SIC classification as the entering Japanese greenfield. Appendix 4 contains the industry-specific counts on a provincial basis for the greenfields, with Appendix 5 providing the national counts. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 23 We also used the notion that Japanese firms are attracted to locate in proximity to other Japanese manufacturing facilities. Unfortunately, the industry-specific data (based on SIC classifications) that was collected for Japanese manufacturing facilities already in Canada, was of little use, because of the small number of industry-specific Japanese firms present in each province at time of entry. Therefore, rather than use a Japanese count of existing firms in the same industry we utilized our data presented in Table 4. I', which provided a count of Japanese manufacturing greenfields that existed in the province prior to the new Japanese entrant. While there is a possibility that the general level of manufacturing will in itself attract additional manufacturing facilities to locate there, we suspect that location choice is driven by industry-specific establishment counts. However, as Head, Ries, and Swenson (1993) point out, the location of like industry, an agglomeration effect, should not be confused with endowment effects. For example, if a new pulp mill locates in a less industrial province (say, British Columbia rather than, say, Ontario) then it may have done so for either industry-specific reasons (most Canadian pulp mills are in B.C.) or endowment effects (most trees close to the Pacific Ocean are in B.C.). In general, it is difficult to distinguish between agglomeration and endowment effects. 4 For a true measure of Japanese greenfields we also included greenfields established prior to 1980. Accordingly, this increased the accumulated totals in Table 1. Alberta had one; British Columbia had two; Ontario had two; Quebec had four; and Saskatchewan and New Brunswick had no Japanese manufacturing greenfields locate in the province prior to 1980. Lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 24 4.5. Data Collection Summary We were able to collect Canadian data for all the variables listed in Section 4.2. We summarize our variables in Table 4.2. In this table each provincial variable is ranked for the years 1989 and 1980. The 1980 ranking only appears in brackets if the ranking actually changed. We observe that approximately one third of the rankings do change. Generally, these changes are minor rearrangements. Only 3% of the rankings actually changed more than three places and apart from the Ontario's unemployment rate becoming the lowest in Canada in 1989, and New Brunswick becoming more unionized in 1989, all other major changes in provincial rankings occurred in the area of provincial corporate tax rates. TABLE 4.2 1989 (1980) PROVINCIAL RANKING Prov Rank Pop Met Crime UI Union Area Wage GDP Hw Engy Air Tax P-m C-m Ship ON 1 1 1 6 1(4) 3(5) 3 9(8) 1 6 5(4) 2 7(6) 1 1 6 PQ 2 2 2 1(2) 6(7) 9(8) 1 7(6) 2 8 6 3 1(4) 2 2 5 BC 3 3 5 10(8) 5 7(9) 2 10 3(4) 9 3 1 2(10) 5(3) 3 1 AB 4 4 3 8(10) 2(1) 1(2) 4 8(9) 4(3) 5 1 4 4(2) 8 4 9 MN 5 5 4 7 4(3) 8(7) 6 3(2) 5(6) 7 4(5) 7 10(8) 4 5 7 NS 6 7 7 5 7(6) 2(4) 9 4 7 2 9(8) 9 6(5) 6 6 3 SK 7 6 6 9 3(2) 5(1) 5 6(7) 6(5) 3 2 6 5(7) 10 8 10 NB 8 8 9 3 8(9) 6(3) 8 5 8 4 8(7) 8 8(3) 3(5) 7 2 PE 9 10 10 4 9(8) 4(6) 10 1 10 1 7(9) 10 3(1) 9 10 8 NF 10 9 8 2(1) 10 10 7 2(3) 9 10 10 5 9 7 9 4 Population 1 = Highest PROVINCE^TOTAL SCORES % of Population living in Metropolitan Area 1 = Highest 1. Ontario 53 Crime Rate 1 = Lowest 2. Quebec^57 Unemployment Rate 1 = Lowest (+ economic indicator) 3. British Columbia^69 Unionization Rate 1 = Lowest 4. Alberta^69 Area in Square Km 1 = Biggest 5. Manitoba 86 Average Manufacturing Wages 1 = Lowest 6. Nova Scotia^89 Provincial GDP 1 = Highest 7. Saskatchewan 90 Highway Km per Square Km 1 = Highest 8. New Brunswick^95 Average Energy 1 = Lowest 9. Prince Edward Island^106 Airports with Control Towers 1 = Most 10. Newfoundland^111 Provincial Corporate Tax Rate 1 = Lowest % of Provincial Labour Force in Manufacturing 1 = Highest NOTES: % of Cdn Manufact. Labour Force in Province 1 = Highest 31% of Rankings changed. Shipping - ease of water access to Japan 1 = Easiest 3% of Rankings changed 3 or more places. - 25 - lain Brown JAPANESE MANUFACTURING GREENFIELD& The Provincial Location Decision^page 26 5. OUR CONDITION LOGIT ECONOMETRIC MODEL Building on related research (Section 3) our model also seeks to explain the location selection process, by determining which factors (from our list in Section 4.2) were significant. That is, why did a Japanese manufacturing greenfield select one province over another, after having made the decision to locate in Canada? The structure of our model enables us to analyze the fifty-five Japanese manufacturing greenfields that have located in Canada during 1980 to 1991. While in theory these firms could have selected any one of ten provinces and two territories, as we know from Table 4.1 and Appendix 1, these investments only involved a total of six provinces. Recall that almost two thirds of the new investment went to Ontario (35 investments), with British Columbia (11 investments) accounting for a fifth. This leaves Alberta with one investment, Saskatchewan with two investments, Quebec with four investments and New Brunswick with two investments. Based on this it is not realistic to say that the whole of Canada represented a choice set, accordingly, we limited our choice set to only the provinces that had been picked in the last decade. (However, for comparative purposes, Table 4.2 and Appendix 2 contain provincial data on all ten provinces). Thus, there are 330 (6 X 55) dichotomous choices that stem from our 55 dependent variables. While there is no hard and fast rule on sample size, our lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 27 small sample size may be problematic; however, we have relied on McFadden (1974), who suggests that logit regression is suitable for sample sizes over 50. We assume that a Japanese manufacturing greenfield will choose to invest in a particular province if and only if it will maximize profit. Formally, the jth province is chosen by the i th firm, if and only if lIy = max {IL; m =^ (1) where Hy denotes the profit of the i th firm given that it locates in the jth province (for j = 1,...,6). Following Carlton (1983) and Coughlin (1991), we assume that Ilw = c + Xift + E y^(2) where c is the constant term; X, is a vector of observable characteristics for the jth province; ft is a vector of unknown coefficients to be estimated; and E y is the random term. If the error term is independent and has a Weibull distribution, McFadden (1974) shows that P = exp {X,J6} / E exp {X,fi}^(for k = 1,...,6)^(3) whereft is obtained by maximum likelihood estimation; and Pi denotes the probability a Japanese manufacturing greenfield locating in province j. This decision depends on the level of the lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 28 province's characteristics that affect profits relative to the other provinces. In general, the explanatory variables are provincial characteristics which are independent of the investment. Canadian Industries, Japanese Greenfields, Labor and Energy, however, contain provincial characteristics that are specific to each investment, and the variation is due to industry differences.The first two variables were introduced in Section 4. Labor and Energy are two additional variables we use to measure factor intensities. We personalize the data for each Japanese establishment (based on work by Schmenner, 1987) by interacting industry labour (energy) intensity and wages (energy) costs in each province. Table 5.1 illustrate the calculation use to arrive at our Energy variable. The Energy Intensity column for firm 10 is pulp industry fuel costs divided by pulp industry sales. To determine Energy, Energy Intensity is multiplied by the Average Energy price in each province prevailing at the time the greenfield is established. The calculation to arrive at our variable to measure labour intensity (Labor) uses the same methodology. To calculate Labor, Wage Intensity is multiplied by Wages prevailing at time of entry in each province. Table 5.1 also shows the variation in our data. For example, we see that energy prices vary among provinces (i.e. Alberta's Average Energy cost in 1990 was half of Quebec's 1990 Average Energy cost). Variation also occurs over time (i.e. from 1986 to 1990 Quebec's Average Energy cost increased by $.61 per 1,000,000 Btu's, whereas the Average Energy cost in Alberta decreased by $.26). In addition to these time and provincial variations that are occurring throughout our model, we also experience industry variation amongst Canadian lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 29 Industries, Japanese Greenfields, Labor and Energy. In this regard Table 5.1 points out the number of pulp mills in British Columbia is greater than the number in Ontario. This is an example of industry-specific variation. TABLE 5.1 ENERGY INTENSITY (In Thousands) Firm # SIC # Energy Intensity Province Avg. Energy $/MBtu Cdn Energy Indust. Count Alberta 3.87 0.2362 2 10 2711 355/5817 B.C. 5.74 0.3503 16 N.B. 8.61 0.5255 6 Mitsubishi Pulp =0.06103 Ontario 6.81 0.4156 5 1990 Quebec 7.51 0.4583 8 Sask. 5.05 0.3082 1 Alberta 4.13 0.0106 0 28 3231 100/39093 B.C. 5.89 0.0151 3 N.B. 8.04 0.0206 0 Honda Auto =0.00256 Ontario 6.32 0.0162 15 1986 Quebec 6.90 0.0177 5 Sask. 5.28 0.0135 1 Finally Table 5.1 shows how our model works. Mitsubishi can chose from six locations, and while Honda will also chose from the same six provinces, between the two investments there are a total of twelve alternatives. Honda and Mitsubishi face different location choices because of different investment times and the investments being in different industries. In our next section we reveal which provincial characteristics influenced the provincial location decision of the Japanese manufacturing greenfields. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 30 6. RESULTS The following final results were obtained by estimating equation (3). TABLE 6.1 FINAL RESULTS LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES Parameter^Estimate -55.0821 55 330 Standard Error t-statistic Manufacturing Wages -8.35985 5.35172 -1.56209 Average Energy -7.11482 2. 83188 -2.51241 * 5 Ship Rank -.350917 .204142 -1.71898 Canadian Industries .657384 .281352 2.33652 * Japanese Greenfields 2.16394 .612262 3.53434 * Provincial MetGDP -.880592 .340797 -2.58392 * The log of each variable was taken, with exception of the Ship Rank variable. Testing the theoretical relationships presented in the previous section was severely constrained by the limited sample size. The variables presented in Table 6.1 are those that had consistent effects across various specifications. To narrow down our list of twenty independent variables to the model above, the variable's significance was measured in terms of its coefficient and t- statistic (10 % significance level). For example, recall that as a measure of transportation services we had collected provincial data on Highway Kilometres, Airports, and Ship Rank variable. We selected Ship Rank because throughout our testing it remained generally significant, while the other two were not significant. 5^Asterisked variables are statistically significant. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 31 To a manager both the lowest available energy cost and the average energy costs are important. The choice between the two measures of energy cost was not intuitively obvious because both could be a guiding force for new plant location. To select the most appropriate measure of energy costs it was necessary to independently test both of our energy cost measures. As Table 6.2 illustrates, when Low Energy replaced Average Energy, the model was not improved. Thus, average energy prices are more influential in a location decision than the price of the lowest available energy source. TABLE 6.2 SUBSTITUTING LOW ENERGY PRICE LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES Parameter^Estimate -55.9915 55 330 Standard Error t-statistic Manufacturing Wages -3.58545 4.25787 -.842076 Low Energy -3.14202 1.48912 -2.10999 * Ship Rank -.068488 .135675 -.504793 Provincial MetGDP -.750461 .324385 -2.31349 * Canadian Industries .710716 .272957 2.60377 * Japanese Greenfields 1.92872 .592186 3.25695 * Recall Table 4.2 where we found that the variables that changed the most in provincial ranking from 1980 to 1989 were Tax, Unionization, and Unemployment. We did not examine the influence of unemployment, given the previously discussed ambiguity of high unemployment as a location factor (i.e. , being a drawing card for some, yet a deterrent to others). However, we decided to test the provincial crime rate, as safety is considered to be important to the Japanese. Our Crime variable was offenses per 100,000 divided by provincial population. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 32 TABLE 6.3 TESTING UNIONIZATION, TAX & CRIME RATES LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES -54.1813 :^55 330 Parameter^Estimate Standard Error t-statistic Manufacturing Wages -11.2525 6.69886 -1.67977 Average Energy -8.59998 3.25604 -2.64124 * Ship Rank -.360553 .397184 -.907774 Canadian Industries .651787 .289458 2.25175 * Japanese Greenfields 2.18708 .806161 2.71296 * Provincial MetGDP -1.27346 1.39549 -.912557 Unionization .033940 .089459 .379387 Tax .179514 .165387 1.08542 Crime -14.3812 31.7353 -.453161 Unlike Bartik (1985), who found tax and unionization to be significant in the United States, as Table 6.3 illustrates, we did not find this result. Nor did Crime have any influence in the location decision of Japanese manufacturing greenfields. Thus, we could not support the stereotype that high rates of unionization, crime and taxes drive business away. However, our results show that Wages and Average Energy are negative and fairly significant throughout our testing. At first this may appear to contradict our findings in Table 4.2. In Table 4.2 we see that Ontario and British Columbia have the highest wages and neither has the lowest energy prices, yet they account for 84% of new Japanese manufacturing investment in Canada in the last decade. This is not a contradiction, because while Japanese greenfields prefer to avoid high wages and energy costs, they have an even stronger preference for like industries and presence of other Japanese greenfields. Thus, they are willing to pay the higher price for inputs in order to be in areas of industry specific concentration. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 33 Next we tested to see whether labour intensive industries were more concerned about the level of wages, and energy intensive industries were more concerned about energy costs. As portrayed in Table 6.4 inconsistent results were found in that neither the Labor or Energy variable were significant. In our sample, it appears that firms in labour or energy intensive industries tended to choose Ontario or British Columbia in spite of high wages and mid-priced energy costs. TABLE 6.4 TESTING FACTOR INTENSITIES LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES Parameter^Estimate -59.0074 :^55 330 Standard Error t-statistic Labor 12.47260 12.5914 .990568 Energy -3.207280 30.5082 -.105128 Ship Rank .025374 .129654 .195703 Provincial MetGDP -.500709 .288151 -1.73766 Canadian Industries .690563 .270134 2.55637 * Japanese Greenfields 1.15395 .401528 2.87389 * In our section on independent variables we explained how we calculated MetGDP. We theorized that Japanese investors prefer large, concentrated markets, so we created this variable by multiplying the percentage of provincial population living in a city with provincial GDP. As Table 6.1 illustrated the coefficient for MetGDP is negative and this counter-intuitive relationship also occurs for GDP (see Appendix 6). Re-examining the data suggested that Quebec might be unduly influencing the model since it has the second highest GDP and Metropolitan population (see Table 4.2 and -A=*=/t=# 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 34 Figure 6.1) yet has only had four Japanese greenfield investments in the last decade. To explore this, we created a Quebec dummy variable. The results in Table 6.5 were inconclusive in that when Quebec replaced the MetGDP variable it too Figure 6.1 was negative and significant; however, when both the Quebec dummy variable and the MetGDP variable were included in the model, as illustrated by Table 6.6, neither was significant. TABLE 6.5 REPLACING MetGDP WITH QUEBEC DUMMY LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES Parameter^Estimate -56.1173 55 330 Standard Error t-statistic Average Energy -5.12543 2.6080 -1.96527 Manufacturing Wages -10.5204 5.71940 -1.83942 Canadian Industries .523389 .271664 1.92660 Japanese Greenfields 1.10012 .424306 2.59274 * Ship Rank -.485113 .217339 -2.23206 * Quebec Dummy -1.42687 .677821 -2.10509 * Lain Brown JAPANESE MANUFACTURING GREEIVFIELDS• The Provincial Location Decision^page 35 TABLE 6.6 ADDING A QUEBEC DUMMY VARIABLE LOG OF LIKELIHOOD FUNCTION NUMBER OF CASES NUMBER OF CHOICES Parameter^Estimate -55.0720 55 330 Standard Error t-statistic Manufacturing Wages -8.71039 5.89270 -1.47817 Average Energy -7.00878 2.92989 -2.39216 * Ship Rank -.366620 .232202 -1.57889 Canadian Industries .648374 .28773 2.25341 * Japanese Greenfields 2.08530 .822973 2.53387 * Quebec Dummy -.157868 1.10961 -.142273 Provincial MetGDP -.816484 .564654 -1.445990 Intuitively the explanation is that the Japanese are merely following a Canadian trend of avoiding Quebec due to government intervention into areas such as language legislation. Furthermore, one would suspect that a Japanese greenfield may be even more concerned about the separation issue than a Canadian firm. This would be particularly true if the Japanese FDI is made in order to tariff jump. Locating in the "Country of Quebec" may not automatically mean that the industry can benefit from existing free trade agreements within North America. While we can not confirm this, we can confirm that Japanese manufacturing greenfields have located in Ontario, which is close to Quebec's wealthy urban population, yet out of its jurisdiction. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 36 7. CONCLUSION In this section we summarize our results and identify the main weaknesses of our analysis. Then we discuss policy ramifications and research questions that stem naturally from our results. 7.1. Factors Influencing Japanese Manufacturing Greenfield's Location Decision We found that new Japanese manufacturing facilities locate next to similar Canadian industry-specific firms and other Japanese firms. Ontario and British Columbia attract a greater proportion of Japanese manufacturing than their national share of manufacturing would suggest. In this regard our data shows that Ontario has approximately half of Canadian manufacturing, yet they have attracted 64% (35/55) of Japanese greenfields. The case in British Columbia is even more profound. The number of new Japanese manufacturing greenfields (11/55 or 20%) is double British Columbia's share of national manufacturing. Thus, we conclude that industry-specific establishment counts are a more important determinant of a Japanese manufacturing greenfield's location than the general level of provincial manufacturing. Industrial distribution of Canadian firms is not the only factor underlying the location choice. Our results show that Japanese firms prefer provinces where there are relatively more Japanese manufacturing greenfields. That is to attract Japanese industry the presence of lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 37 similar Canadian industry (i.e., with the same SIC classification) is necessary and once Japanese industry is attracted then this in itself is a factor in attracting other industries from Japan. Furthermore, if two provinces share similar levels of industry-specific Canadian firms then geographic proximity to Japan may be important. To summarize, Canadian Industries and Japanese Greenfields were always significant, with Ship Rank generally being significant. Apart from these three variables, Average Energy, Wages, and MetGDP were also part of the equation; however, their importance is minor in determining where Japanese manufacturing greenfields will locate in Canada. This brings us to the question of whether the positive coefficient on the Canadian Industries variable reflects an agglomeration effect or an endowment effect. Greenfields in Ontario tend to be in the automotive sector, which follows a traditional Canadian pattern. However, while both British Columbia and Ontario have high counts of forestry specific industries the Japanese have elected to locate mostly in British Columbia presumably due to its easier access to the Japanese market. In short, it is the only province with trees across the water from Japan. This ease of transportation is captured by the significance of our Ship Rank variable, which measures the province's water access to Japan. Referring back to Table 4.2 we see that British Columbia's geographic position gives it the comparative advantage over Ontario's geographic position. Head, Ries, and Swenson (1993) suggested that the concentration of Japanese wood product firms in Washington State 6 is the result of the state's forests and 6^Washington State is British Columbia's southern neighbour, and both are bounded by the Pacific Ocean on the west. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 38 Washington's geographic proximity to Japan. Thus, endowments--British Columbia's forests and proximity to Japan--may explain the concentration of Japanese forestry investment in British Columbia. It is difficult to distinguish if clustering is due to agglomeration or endowment. That is, are Japanese pulp mills in British Columbia because it is the only province with trees across the water from Japan (an endowment effect) or because the province of British Columbia has more pulp mills than any other Canadian province (which could be due to agglomeration or an endowment effect)? While the answer is not intuitively obvious, our results may suggest evidence of pure agglomeration effects. That is, if all endowment effects are captured by the combination of the Canadian Industries and Ship Rank variables, then the positive coefficient on the significant Japanese Greenfields variable may be evidence of pure agglomeration effects. 7.2. Limitations of Our Results The small number of Japanese manufacturing greenfields that have entered the Canadian economy in the last decade limits our ability to assess the determinants of investment. The many possible explanatory variables and little information to distinguish among them (especially in the presence of multicollinearity) adds further to the difficulties in determining actual location factors that attract Japanese manufacturing greenfields. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 39 The possibility of omitted explanatory variables exists. Utilizing past research we identified and then tested the vast majority of factors that could influence the location decision. However, it is possible that our error term includes an independent variable that we were not able to identify and test for its significance. If we have omitted variables that are correlated with other independent variables, the estimated coefficients for those variables will be biased. Another possible problem with our model may be lack of variation amongst non industry-specific provincial characteristics. Table 4.2 illustrates that there is variation among these characteristics. However, the range of variation between these characteristics is not always, from a statistical point of view, as great as we would like it to be. This limited variation of non industry-specific provincial characteristics constrains our ability to distinguish the individual effects of each variable. With the combination of Ontario (64 %) and British Columbia (20 %) receiving 84 % of Japanese greenfield investment our distribution is skewed. This skewed distribution compounds the problem of our limited variation. Furthermore, this distribution also gives rise to the possibility that a unique omitted feature of either Ontario or British Columbia could be driving the results. Nevertheless, by measuring variation in both non industry-specific provincial characteristics and industry-specific characteristics, despite our small skewed sample size, we feel that our study still provides results that provincial policy makers may find useful. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 40 7.3. Policy Ramifications for Provinces Seeking FDI Regardless of whether we think foreign investment is good or bad for Canada it has become somewhat of a necessity given our desire to spend more than we earn. Accordingly, we find ourselves in the position of needing to attract foreign investment. Because of possible job creation and potential spillovers of management expertise, FDI has been highly sought after in North America. In the U.S. large subsidies have been paid by the individual states trying to attract Japanese greenfield investments. However, our study suggests that unless Japanese manufacturing greenfields and industry-specific Canadian firms are already in the province, and that the province provides market accessability, the province is unlikely to attract much Japanese investment. Hence, increasing schemes directed at attracting FDI will not have a significant effect if the industry- specific base of firms does not already exist. If the investment is to export products to Japan, then market access to Japan is determined by geography (a factor not easily modified). For example, although both Ontario and British Columbia have a forest industry, because the products being produced are for Japanese consumption, Japanese-owned plants tend to locate in British Columbia. If the market is Canada (or North America) then location is in Ontario with its central location and substantial industrial activity. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 41 Our results indicate that it is difficult for a province to try to attract Japanese greenfield FDI without a market for the product, or without an existing industrial base. For a province to try to attract this type of FDI with neither, it is near impossible, unless governments intervene. However, even government intervention may do little to encourage location preferences. Our study suggests that corporate tax rates and unionization were not significant. Therefore, rather than offer subsidies or direct government resources to reducing unionization, perhaps tax dollars would be better spent on developing an industrial strategy to encourage clustering of industry. If tax dollars are efficiently spent on needed infrastructure that creates a competitive advantage, then this may encourage clustering of firms that benefit from this infrastructure. Should Canadian industries cluster, then Japanese manufacturing greenfields in the same industry are more likely to be attracted to this area. With a base of Japanese manufacturing greenfields, more Japanese manufacturing greenfields are likely to be attracted. This possible Japanese agglomeration effect could then create industry diversification. 7.4. Further Study and Extensions This study questions government efforts to attract Japanese greenfield investments to areas that do not already have a base of firms in similar industries or Japanese manufacturing greenfields. Our study also drew our attention to a possible free trade effect, in that we noticed that 75% of greenfields locating in B.C. during the last decade did so after free trade. Contrary lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 42 to British Columbia, only 25% of Ontario's Japanese manufacturing greenfields located in Ontario after free trade. While this may be a free trade effect, Ontario's relative fewer investments could reflect the change in business climate due to the more socialist government being elected in 1990. We feel it is also worth studying why the majority of Japanese manufacturing greenfields locate in Ontario, which is close to Quebec's wealthy urban population, yet out of its jurisdiction. This too may be a free trade effect in that the Japanese manufacturing greenfields wish to have access to North American free trade, which is something that is not guaranteed if Quebec should separate from Canada. Alternatively, the possible avoidance of Quebec could be a language preference. Future work will take into account non-free trade issues such as language or changing political climates so that we may study whether the Canada/U.S. Free Trade Agreement has shifted FDI amongst the provinces and/or away from Canada to the U.S. We plan to study this by comparing Japanese manufacturing greenfields locating in Canada to those locating in the U.S. We will seek to not only determine whether one country has attracted relatively more Japanese manufacturing greenfields than the other, but also to determine if industry location is due to comparative advantage as FDI for tariff jumping is no longer required within the new North American trade zone. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 43 8. BIBLIOGRAPHY T. J. Bartik, "Business Location Decisions in the United States: Estimates of the Effects of Unionization, Taxes and Other Characteristics of States," Journal of Business & Economic Statistics, vol. 3, no. 1, pp. 14-22. 1985. C.C.H. Canadian Limited, Canadian Tax Reporter, vol. 1, pp. 1602-1603. 1992. D. W. Carlton, "The Location and Employment Choices of New Firms: An Econometric Model With Discrete and Continuous Endogenous Variables," Review of Economics and Statistics, vol. 65 pp. 440-449. 1983. C. Coughlin, et. al., "State Characteristics and the Location of Foreign Direct Investment with the United States," Review of Economics and Statistics, pp. 675-683. 1991. Dodwell Marketing Consultants, Auto Parts Industry of Japan (Tokyo: 1989). External Affairs Canada, Canada - U.S. Free Trade Agreement. 1988. N. Glickman & D. P. Woodward, "The Location of Foreign Direct Investment in the United States: Patterns and Determinants," International Regional Science Review, vol. 11, no. 2, pp. 137-154. 1988. E. M. Graham & P. Krugman, Foreign Direct Investment in the United States (Washington D.C.: Institute for International Economics 1990). L. C. Hamilton, Regression with Graphics: A Second Course in Applied Statistics, (Pacific Grove, California: Brooks/Cole Publishing 1992). K. Head, J. Ries, and D. Swenson, "Agglomeration Benefits and Location Choice: Evidence from Japanese Manufacturing Investments in the United States," University of British Columbia Working Paper. March 1993. Investment Canada, Annual Report. 1991. Investment Canada, International Investment's Canadian Developments in a Global Context. 1991. Japan, Japanese Direct Investment Aboard in Fiscal 1990. Ministry of Finance. 1991. JETRO, Directory: Japanese-Affiliated Companies in USA & Canada. 1991-1992. lain Brown JAPANESE MANUFACTURING GREENFIELDS: The Provincial Location Decision^page 44 M. I. Luger & S. Shetty, "Determinants of Foreign Plant Start-ups in the United States: Lessons for Policymakers in the Southeast," Vanderbilt School of Transnational Law, vol. 18, pp. 223-245. 1985. P. Krugman "Geography and Trade," Cambridge, MA: The MIT Press. D. McFadden, "Conditional Logit Analysis of Qualitative Choice Behaviour," in P. Zarembka, ed., Frontiers in Econometrics (New York: Academic Press 1974). D. McFadden, "Cost, Revenue, and Profit Functions," in D. McFadden and M. Fuss, eds., Production Economics: A Dual Approach to Theory and Applications, vol. 1 (Amsterdam: North-Holland 1978). R. J. Newman & D. H. Sullivan, "Econometric Analysis of Business Tax Impacts on Industrial Location: What Do We Know, and How Do We Know It?" Journal of Urban Economics, vol. 23, pp. 215-234. 1988. Ontario, Canada, Japanese Investment Profile. 1992. Roads and Transportation Association of Canada, Canada's Roadway Infrastructure Selected Facts and Figures, (Ottawa, Canada 1990). A. M. Rugman, International Business in Canada Strategies for Management (Scarborough: Prentice-Hall 1989). R. W. Schmenner, J. C. Huber, and R. L. Cook, "Geographic Differences and the Location of New Manufacturing Facilities," Journal of Urban Economics, vol. 21, pp. 83-104. 1987. D. Smith & R. Florida, "Agglomeration and Industry Location: An Econometric Analysis of Japanese-Affiliated Manufacturing Establishments in Automotive-Related Industries," H. John Heinz III School of Public Policy and Management (Carnegie Mellon University) Working Paper. 1992. Statistics Canada, "CALURA," Catalogue 71-201 Annual. 1991. Statistics Canada, "Canada Year Book," Annual. 1991. Statistics Canada, "Canada's International Investment Position." 1991. Statistics Canada, "Cansim Data Retrieval Program," (On-Line Database System, 1993). Statistics Canada, "Consumption of purchased fuel and electricity," Catalogue 57-208 Annual. 1991. lain Brown JAPANESE MANUFACTURING GREE1VFIELDS: The Provincial Location Decision^page 45 Statistics Canada, "Energy Statistics Hand Book," Catalogue 57-601 Quarterly. September 1992. Statistics Canada, "Manufacturing industries of Canada: national and provincial areas," Catalogue 31-203 Annual. 1991. Statistics Canada, Pacific Region Advisory Services, "Inquiry for Union Statistics," February 9, 1993. Statistics Canada, "Quarterly report on energy supply-demand in Canada," Catalogue 57-003 Quarterly. June 1992. Toyo Keizai Company, Japanese Overseas Investment (Tokyo: 1990). Transport Canada, "Airports Group - Airport Site Directory." 1993. United Nations, Foreign Direct Investment and the Service Sector and International Banking, the United Nations Centre for Transnational Corporations. 1991. United Nations, World Investment Report: the Triad in Foreign Direct Investment, the United Nations Centre for Transnational Corporations. 1991. D. P. Woodward, "Locational Determinants of Japanese Manufacturing Start-Ups in the United States," Southern Economic Journal. January, 1992. APPENDIX 1 FIFTY-FIVE JAPANESE MANUFACTURING GREENFIELDS IN CANADA (Dependent Variable) PROV YEAR FIRM # CANADIAN COMPANY JAPANESE INVESTOR OWNER- SHIP SIC CODE #CDN FIRM IN SIC PRODUCT DESCRIPTION AB 1990 1 Tomen Alberta Timber Industries Toyo Menka Kaisha 100.0 2512 1093 Lumber BC 1989 2 Advanced Energy Technology NTT 45.0 3391 27 R&D rechargeable batteries BC 1989 3 Atsugi Nylon Canada Inc Atsugi Nylon 24.0 1811 31 Nylon BC 1991 4 Campbell River Fibre Ltd C Itoh and Co Ltd 90.0 2512 1093 Woodchips BC 1983 5 Canadian Autoparts Toyota Inc Toyota Motor Corp 100.0 3255 47 Aluminum wheels BC 1988 6 Canadian Chopstick Mfg Co Ltd Mitsubishi Corp 100.0 2599 272 Chopsticks BC 1980 7 Daiwa (Canada) Ltd Daiwa Seiko Inc 3931 209 Golf clubs BC 1983 8 Dominion Malting Ltd Sumitomo Corp 35.0 1131 48 Liquor malt BC 1991 9 I.S. Forest Products Inland Kogyo 17.0 2512 1093 Forest product BC 1990 10 M.C. Forest Investment Mitsubishi Corp 100.0 2711 39 Pulp BC 1988 11 Primex Fibre Ltd Sanyang Pulp 50.0 2711 39 Pulp, chips BC 1989 12 S.M. Cyclo of Canada Sumitomo Heavy Ind. 3199 772 Speed reducers & variators, motors NB 1989 13 Ampal Pallets Inc Mitsui & Co Ltd 57.9 3099 463 Steel pallets NB 1980 14 NBIP Forest Products Inc Oji Paper Co/Mitsui & Co 33.0 2712 Newsprint ON 1987 15 ABC Nishikawa Industries Nishikawa Kasei Co Ltd 49.0 3256 96 Plastic autoparts & armrests ON-OP 1988 16 Bellemar Parts Ind. Canada Honda Motor Co Ltd 100.0 3259 190 Seats for vehicles & tire assembly ON 1986 17 CAMI Automotive Inc Suzuki Motor Co Ltd 50.0 3231 27 Vehicles - 46 - ON-OP 1989 18 Canada Mold Technology Inc Nagase Ltd 100.0 3062 539 Prototype molds ON-OP 1990 19 Cangel Inc Nitta Gelatin Inc 100.0 1011 526 Gelatin & lard ON 1987 20 Copar International Toyo Radiator Co Ltd 46.0 3251 50 Radiator & oil coolers ON 1988 21 DDM Plastics Inc Daikyo/Suzuki/Mitsui 100.0 3256 96 Automotive plastics ON-OP 1985 22 Denon Canada Inc Nippon Columbia Co Ltd 95.0 3341 25 Car stereo, cassette & CD software ON 1990 23 DNN Galvanizing Corp Nippon Kokan (NICK) 40.0 2912 26 Hot dip galvanizing steel sheets ON 1981 24 Epson Canada Ltd Seiko Epson Corp 19.0 3361 147 Computers, printers, software products ON 1980 25 Epson Manufacturing Ltd Seiko Epson Corp 3361 147 Printers, ribbons & technical products ON 1980 26 F&P Mfg Inc F. Tech Inc 55.0 3257 20 Auto part, pedal bracket ON 1987 27 General Seating of Canada Ltd NHK Spring Co 65.0 3259 190 Seats for automobiles ON 1986 28 Honda of Canada Mfg Inc Honda Motor Co Ltd 100.0 3231 27 Automobiles ON 1988 29 Inoac Canada Ltd Inoue MTP Co 50.0 3257 20 Automotive interior panels & armrests ON 1990 30 IDS Fitel Inc Furukawa Electric Co Ltd 50.0 3562 153 Passive fibre optic components ON 1986 31 Kao-Didak Ltd Kao Corporation 93.0 3399 72 Floppy disks ON 1984 32 Kuriyama Canada Ltd Kuriyama Corp 100.0 1621 77 Industrial plastic hose & plumbing ON 1983 33 Mitsubishi Electronic Ind Mitsubishi Electric Corp 100.0 3341 25 Colour cathode ray tubes ON 1987 34 Miura Boiler Co Miura Boiler Co 99.3 3011 43 High pressure steam boiler ON 1980 35 Murata Erie North America Ltd Murata Mfg Co Ltd 100.0 351 90 Ceramic capacitors ON-OP 1987 36 Nichirin Inc Nichirin Co Ltd 100.0 3259 190 Hydraulic hoses for autos & motorcycles ON 1985 37 NKC of Canada Inc Nakanishi Metal Works 100.0 3192 533 Conveyor systems ON 1986 38 Quality Safety Systems Co Tokai Rika Co Ltd 40.0 3259 190 Seat belts & auto components ON 1986 39 Rockwell Int'l Suspension Syst Mitsubishi Corp 40.0 3254 35 Coil springs & torsion bars ON 1982 40 Sanyo Cdn Machine Works Inc Sanyo Machine Works Inc 100.0 3081 1464 Automatic assembly & welding machine - 47 - ON-OP 1985 41 SM Yttrium Canada Ltd Shin-Etsu Chemical Co 100.0 3731 94 Silicon ON 1987 42 SMC Pneumatics Canada Inc SMC Corporations 3092 44 Cylinders & valves ON 1986 43 Toyota Motor Mfg Canada Inc Toyota Motor Corp 100.0 3231 27 Automobiles ON 1985 44 Trutech Canada Inc Nihon Parkerizing Co 3041 288 Paint finishing system, rolling oil conc ON 1989 45 UCAR Carbon Canada Inc Mitsubishi Corp 50.0 3399 72 Artificial graphic electrodes ON 1980 46 UNIC International Corp 319 1601 Sandblasting equipment ON 1987 47 Vdo-Yazaki Ltd Yazaki Corporation 50.0 391 603 Meters ON 1985 48 Woodbridge Inoac Inc Inoue MTP Co 50.0 3257 20 Automotive interior trims ON 1989 49 Yachiyo of Ontario Mfg Inc Yachiyo Industries 100.0 3259 190 Fuel tanks PQ 1986 50 Cree Yamaha Enterprise Ltd Yamaha Motors 40.0 3281 327 FRP boats PQ 1986 51 H Aida Enterprise Inc Tokia 100.0 1021 414 Processing seafood PQ 1989 52 Kobe Aluminium Canada Kobe Steel Ltd 100.0 2961 71 Aluminium PQ 1989 53 Miura Boiler Company Miura Company 3011 43 Boilers SK 1988 54 Hitachi Canadian Ind Ltd Hitachi Ltd 100.0 337 281 Electric power equipment SK 1987 55 SK Turbine Ltd Marubeni Corporation 100.0 3194 116 Turbines NOTE: "-OP" signifies that this is the year of operation, rather than year of establishment APPENDIX 2 PROVINCIAL CHARACTERISTICS (Independent Variables) COLUMN INDEPENDENT VARIABLE POP Population in 1000's % Metro Percentage of Population Living in Metropolitan Areas Crime Reported Offenses per 100,000 People UI Unemployment Rate Union % Unionization Rates Area Land Area (excluding fresh water) Square KM (constant 1980-90) MAN$/CAP Average Annual Manufacturing Pay per Worker GDP/CAP Provincial GDP per Capita Highway KM/SQ.ICM Highway Kilometres per Square Kilometre of Area (Constant 1980-90) AVG NRG $/MBtu Weighted Average Cost of Crude Oil, Natural Gas, and Electricity Prices per Million Btu's Low NRG $/MBtu Lowest Cost of Either Crude Oil, Natural Gas, or Electricity Prices per Million Btu's AIRPORT Number of Airports with Control Towers TAX Provincial Corporate Tax Rate % PROV LAB in MFG Percentage of Work Force in Manufacturing (Manufacturing Labour Force/Total Labour Force) % CDN MFG LAB/PROV Percentage of Canadian Manufacturing Work Force in the Province (Provincial Manufacturing Labour Force/Total Canadian Manufacturing Labour Force) Ship Rank Shipping 1=BC (yearly deep water; no canal) 2=NS, NB, NFL (yearly deep water) 3 = Que (mostly open deep water) 4= Ont (mostly open shallow water) 5 =Man (partly open deep water) 6 =PEI (deep water but no port facilities) 7 = Ab, Sask (no water access) PROV YEAR POP %METRO CRIME UI UNION AREA APPENDIX 2 CONT'D MAWCAP GDP/CAP HIGHWAY AVG NRG LOW NRG AIRPORT TAX %PROVLAB%CDN MFG SHIP 000's RATE % % SQ.KM KM/SQ.KM $/MBtu $/MBtu^i % IN MFG LAB/PROV RANK BC 1980 2666.0 54.4 141.77 6.8 39.2 929730 22231 14343.21^0.071^3.17 1.64 26 15.0 12.30 8.65 1 BC 1981 2744.2 54.0 150.63 6.7 40.0 929730 24388 16285.62^0.071^4.20 2.39^26 16.0 11.69 8.58 1 BC 1982 2787.7 54.5 161.03 12.1 40.0 929730 27059 16542.31^0.071^5.32 3.29 26 16.0 10.25 8.24 1 BC 1983 2813.8 54.8 156.34 13.8 40.4 929730 29242 17112.45^0.071^5.73 3.31^26 16.0 9.65 7.97 1 BC 1984 2847.7 55.2 155.39 14.7 38.3 929730 29905 17950.98^0.071^6.14 3.49 26 16.0 9.45 8.00 1 BC 1985 2870.1 55.5 154.99 14.1 36.9 929730 31389 18988.54^0.071^6.33 3.52^26 16.0 9.49 7.63 1 BC 1986 2889.0 56.7 161.88 12.5 41.2 929730 31895 19848.39^0.071^5.89 3.31 26 16.0 9.19 7.38 1 BC 1987 2925.0 57.2 165.63 11.9 37.5 929730 32375 21496.07^0.071^5.56 2.33^26 15.0 9.59 7.62 1 BC 1988 2980.2 57.8 160.57 10.3 38.1 929730 33542 23325.28^0.071^5.44 2.36 26 14.0 10.06 7.82 1 BC 1989 3048.3 58.3 166.60 9.1 36.4 929730 34841 25234.06^0.071^5.37 1.97^26 14.0 10.07 8.07 1 BC 1990 3132.5 58.3 177.51 8.3 38.0 929730 35000 25764.09^0.071^5.74 2.28 26 14.0 10.08 8.10 1 AB 1980 2140.6 60.7 153.14 3.7 22.0 644390 19003 20156.97^0.265^2.67 1.36^7 11.0 7.25 4.39 7 AB 1981 2237.3 58.1 157.16 3.8 23.3 644390 21466 22318.87^0.265^3.48 1.92 7 11.0 7.21 4.66 7 AB 1982 2314.5 56.4 149.12 7.7 22.7 644390 24248 22854.18^0.265^3.84 1.96^7 11.0 6.47 4.64 7 AB 1983 2338.7 56.1 145.16 10.6 23.9 644390 25989 23682.39^0.265^4.31 2.16 7 11.0 5.86 4.33 7 AB 1984 2338.5 55.9 129.87 11.1 23.6 644390 27241 25204.62^0.265^4.56 2.07^7 11.0 5.76 4.33 7 AB 1985 2348.5 55.7 128.19 10.0 22.7 644390 27864 27826.70^0.265^4.61 2.13 7 11.0 5.98 4.23 7 AB 1986 2375.1 61.4 134.52 9.8 26.4 644390 29027 24132.46^0.265^4.13 2.12^7 11.0 6.03 4.22 7 AB 1987 2377.7 61.7 145.05 9.6 24.8 644390 29132 25051.52^0.265^4.24 2.14 7 15.0 6.17 4.20 7 AB 1988 2388.7 62.1 148.41 8.0 26.1 644390 29453 26059.36^0.265^3.87 1.86^7 15.0 6.72 4.46 7 AB 1989 2425.9 62.3 143.24 7.2 25.8 644390 30638 27059.24^0.265^3.79 1.40 7 15.0 7.02 4.66 7 AB 1990 2473.1 62.6 145.49 7.0 26.6 644390 31000 28533.02^0.265^3.87 1.30^7 15.0 7.10 4.69 7 SK 1980 959.4 32.5 142.96 4.4 18.8 570700 17526 12924.74^0.341^3.05 1.70^5 14.0 4.85 1.15 7 SK 1981 968.3 32.8 152.39 4.6 28.9 570700 19690 14808.43^0.341^3.76 2.28 5 14.0 4.82 1.16 7 SK 1982 977.0 33.3 139.13 6.1 28.3 570700 21926 15107.47^0.341^4.54 2.90^5 14.0 4.37 1.17 7 SK 1983 989.3 33.7 135.83 7.3 28.2 570700 23495 15399.78^0.341^4.91 3.12 5 14.0 4.03 1.13 7 SK 1984 1000.5 33.8 139.08 8.0 28.4 570700 25385 16381.81^0.341^5.26 3.17^5 16.0 3.98 1.14 7 SK 1985 1008.4 34.2 144.83 8.1 27.1 570700 25272 17290.76^0.341^5.40 3.15 5 16.0 3.97 1.09 7 SK 1986 1010.2 38.3 153.95 7.7 32.7 570700 25356 16971.89^0.341^5.28 3.15^5 17.0 3.95 1.07 7 SK 1987 1015.8 38.6 155.71 7.4 30.5 570700 26141 16954.12^0.341^5.23 2.72 5 17.0 4.04 1.06 7 SK 1988 1013.5 39.0 152.73 7.5 31.1 570700 27557 17945.73^0.341^4.54 2.52^5 17.0 4.19 1.05 7 SK 1989 1006.7 39.2 148.37 7.4 32.2 570700 28379 19326.51^0.341^4.66 2.49 5 15.0 4.43 1.08 7 SK 1990 997.1 39.7 145.92 7.0 32.3 570700 29000 20332.97^0.341^5.05 2.90^5 15.0 4.50 1.07 7 MN 1980 1024.9 56.8 116.36 5.5 30.7 548360 15637 10916.19^0.154^4.22 2.25^4 15.0 12.00 3.14 5 MN 1981 1026.2 57.2 124.75 5.9 30.2 548360 17546 12824.01^0.154^4.62 2.81 4 15.0 11.67 3.08 5 MN 1982 1033.3 57.3 131.48 8.5 30.0 548360 19460 13562.37^0.154^5.03 3.48^4 15.0 10.78 3.13 5 MN 1983 1045.6 57.4 133.19 9.4 26.6 548360 20393 14260.71^0.154^5.37 3.67 4 16.0 10.13 3.07 5 MN 1984 1055.1 57.2 130.40 8.4 28.2 548360 21325 15657.28^0.154^5.65 3.69^4 16.0 10.00 3.11 5 MN 1985 1064.0 57.5 132.60 8.2 28.7 548360 22157 16598.68^0.154^5.87 3.63 4 16.0 9.93 2.91 5 MN 1986 1071.2 58.4 139.29 7.7 35.5 548360 23009 17196.60^0.154^5.49 3.40^4 17.0 10.05 2.92 5 MN 1987 1079.0 58.6 143.43 7.4 35.4 548360 23775 18064.87^0.154^5.86 3.26 4 17.0 10.18 2.90 5 MN 1988 1084.1 58.9 124.06 7.8 35.3 548360 24653 19895.77^0.154^5.88 2.94^4 17.0 10.70 2.94 5 MN 1989 1086.3 59.0 115.93 7.5 36.7 548360 25926 21135.05^0.154^6.22 2.84 4 17.0 10.55 2.88 5 MN 1990 1089.0 59.4 121.84 7.2 36.8 548360 27000 21769.51^0.154^6.85 2.76^4 17.0 10.40 2.91 5 ON 1980 8569.7 64.8 116.24 6.8 29.7 891190 18083 13418.67^0.187^3.83 2.45^16 14.0 20.92 49.19 4 ON 1981 8624.7 65.2 120.10 6.6 29.5 891190 20251 15285.29^0.187^4.61 3.13 16 14.0 20.44 49.20 4 ON 1982 8702.5 65.4 116.39 9.7 30.2 891190 22375 15778.22^0.187^5.42 3.87^16 14.0 18.81 49.75 4 ON 1983 8798.0 65.6 111.40 10.3 32.5 891190 24179 17270.40^0.187^5.92 4.33 16 15.0 18.39 50.20 4 ON 1984 8901.7 65.8 109.71 9.0 32.1 891190 27996 19265.87^0.187^6.18 4.33^16 15.0 17.38 49.03 4 ON 1985 9006.4 66.1 106.13 8.0 31.8 891190 27009 20381.17^0.187^6.46 4.31 16 15.0 19.16 51.60 4 ON 1986 9113.0 69.4 111.70 7.0 31.6 891190 27870 22244.05^0.187^6.32 3.56^16 15.5 19.23 51.70 4 ON 1987 9265.0 69.6 113.75 6.1 30.9 891190 28743 24162.12^0.187^6.15 3.68 16 15.5 19.18 51.36 4 ON 1988 9431.1 69.7 113.42 5.0 31.1 891190 30303 26758.07^0.187^6.23 3.24^16 15.5 19.41 51.04 4 ON 1989 9589.6 69.9 112.68 5.1 31.1 891190 31559 28380.85^0.187^6.54 3.51 16 15.5 19.53 51.69 4 ON 1990 9749.6 70.2 116.12 6.3 31.7 891190 32500 28420.14^0.187^6.81 3.46^16 15.5 19.70 51.34 4 - 50 - 17.67 28.54 3 17.31 28.37 3 16.09 28.22 3 15.63 28.42 3 11.76 (3' 29.39 27.78 3 3 15.86 27.84 3 16.00 27.92 3 16.28 27.69 3 15.68 26.60 3 16.00 26.83 3 11.51 1.74 2 11.02 1.70 2 10.11 1.67 2 9.87 1.70 2 9.84 1.73 2 9.82 1.67 2 10.39 1.76 2 10.36 1.74 2 10.85 1.77 2 11.12 1.83 2 11.30 1.86 2 10.94 2.11 2 10.81 2.09 2 9.57 2.02 2 9.30 2.05 2 9.10 2.12 2 9.13 1.99 2 9.10 1.97 2 9.43 2.02 2 9.75 2.04 2 9.73 2.04 2 9.70 2.07 2 5.62 0.16 6 5.74 0.16 6 5.47 0.17 6 5.57 0.18 6 5.29 0.18 6 5.84 0.19 6 5.79 0.19 6 5.88 0.19 6 6.06 0.19 6 5.94 0.19 6 5.90 0.19 6 8.54 0.95 2 8.75 0.98 2 8.10 0.99 2 7.47 0.95 2 7.30 0.96 2 7.36 0.92 2 7.82 0.96 2 8.35 1.00 2 8.30 0.99 2 7.83 0.95. 2 8.00 0.99 2 13.0 13.0 8.0 5.5Z . 13.0 13.0 13.0 13.0 13.0 12.0 13.0 14.0 14.0 15.0 15.0 15.0 15.0 16.0 16.0 16.0 13.0 13.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 16.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 15.0 15.0 15.0 15.0 15.0 15.0 16.0 16.0 16.0 16.0 16.0 16.0 16.0 16.5 17.0 PQ 1980 6386.1 61.5 77.80 9.8 35.9 1356790 16720 11308.94 0.080 4.91 2.69 15 PQ 1981 6438.2 61.3 81.52 10.3 37.9 1356790 18720 12660.84 0.080 5.81 3.38 15 PQ 1982 6462.2 61.4 78.66 13.8 36.5 1356790 20586 13343.44 0.080 6.73 4.41 15 PQ 1983 6474.9 61.7 73.88 13.9 38.2^1356790 21859 14251.03 0.080 7.23 4.73 15 PQ 1984 6492.0 61.8 74.29 12.8 38.7 1356790 23054 15556.22 0.080 7.23 5.06 15 PQ 1985 6514.2 61.9 77.42 11.8 38.7 1356790 24040 16570.57 0.080 6.95 4.73 15 PQ 1986 6540.2 63.4 78.23 11.0 38.5 1356790 24833 17964.74 0.080 6.90 3.60 15 PQ 1987 6592.6 63.4 79.77 10.3 37.9 1356790 25830 19683.13 0.080 7.05 4.26 15 PQ 1988 6640.8 63.6 78.93 9.4 37.8^1356790 27133 21372.73 0.080 7.09 3.28 15 PQ 1989 6698.2 63.7 78.50 9.3 40.2^1356790 28810 22342.12 0.080 7.21 3.86 15 PQ 1990 6768.2 63.9 85.18 10.1 40.0 1356790 30000 22763.22 0.080 7.51 4.11 15 NB 1980 695.4 16.4 88.41 11.0 26.8^72090 16353 7217.43 0.284 6.18 3.21 4 NB 1981 696.4 16.4 89.37 11.5 32.3^72090 18081 8548.25 0.284 7.39 4.82 4 NB 1982 696.6 16.5 91.22 14.1 34.8^72090 19690 9369.8C 0.284 8.02 5.77 4 NB 1983 703.2 16.4 89.14 14.8 37.5^72090 20878 10655.57 0.284 8.57 6.22 4 NB 1984 707.9 16.3 85.79 14.8 31.8^72090 22128 11830.77 0.284 8.74 6.29 4 NB 1985 709.9 16.5 87.10 15.1 29.3^72090 23494 12687.70 0.284 9.42 6.54 4 NB 1986 710.4 17.1 90.53 14.3 30.0^72090 24026 14187.78 0.284 8.04 3.60 4 NB 1987 712.3 17.1 90.59 13.1 34.6^72090 25856 15274.46 0.284 8.37 4.26 4 NB 1988 714.3 17.2 89.99 12.0 34.1^72090 26997 16470.67 0.284 7.88 3.28 4 NB 1989 717.8 17.1 89.00 12.5 35.3^72090 27517 17617.72 0.284 8.17 3.86 4 NB 1990 722.4 17.2 96.51 12.1 36.9^72090 28000 18403.9 1 0.284 8.61 4.76 4 NS 1980 845.1 32.4 115.07 9.7 27.0^52840 16069 7445.27 0.486 6.62 3.21 3 NS 1981 847.4 32.7 119.89 10.1 29.5^52840 17829 8667.69 0.486 7.40 4.82 3 NS 1982 849.5 32.9 123.32 13.1 30.4^52840 19437 9961.15 0.486 7.86 5.77 3 NS 1983 857.0 32.9 105.88 13.2 29.5^52840 20747 11235.71 0.486 9.29 6.22 3 NS 1984 864.4 33.1 99.58 13.0 28.0^52840 22170 12379.69 0.486 9.58 6.29 3 NS 1985 871.0 33.4 100.24 13.6 28.3^52840 22039 13701.49 0.486 9.70 6.54 3 NS 1986 873.2 33.9 103.41 13.1 27.8^52840 23206 14923.27 0.486 8.79 3.60 3 NS 1987 878.0 34.2 103.97 12.3 30.9^52840 23981 15897.49 0.486 9.28 4.26 3 NS 1988 881.9 34.2 103.86 10.2 30.5^52840 24688 17084.70 0.486 8.81 3.28 3 NS 1989 888.3 34.5 106.62 10.3 30.8^52840 26413 18089.61 0.486 9.37 3.86 3 NS 1990 895.1 34.9 116.69 10.5 30.4^52840 27100 19011.28 0.486 10.11 4.76 3 PE 1980 122.8 0.0 92.39 10.6 30.0^5660 13003 6889.25 0.919 6.63 3.21 1 PE 1981 122.5 0.0 93.51 11.2 30.1^5660 14074 8236.73 0.919 9.43 4.82 1 PE 1982 122.4 0.0 97.04 12.9 26.4^5660 14964 8586.60 0.919 11.35 5.77 1 PE 1983 123.7 0.0 98.42 12.2 31.2^5660 16226 9417.95 0.919 11.66 6.22 1 PE 1984 125.1 0.0 101.78 12.8 22.9^5660 16766 10367.71 0.919 12.30 6.29 1 PE 1985 126.0 0.0 100.87 13.3 21.6^5660 16966 10476.19 0.919 12.89 6.54 1 PE 1986 126.6 0.0 96.47 13.4 23.0^5660 16975 11832.54 0.919 8.33 3.60 1 PE 1987 127.3 0.0 88.66 13.2 29.0^5660 17195 12482.33 0.919 7.44 4.26 1 PE 1988 128.5 0.0 100.71 13.0 29.5^5660 18855 13859.92 0.919 7.25 3.28 1 PE 1989 129.9 0.0 104.26 14.1 31.3^5660 19375 14603.54 0.919 7.51 3.86 1 PE 1990 130.7 0.0 107.74 14.9 33.7^5660 20000 15233.36 0.919 8.69 4.76 1 NF 1980 565.6 27.1 71.96 13.2 43.7^371690 15645 7240.10 0.032 6.81 3.21 7 NF 1981 567.7 27.1 79.74 13.8 49.2^371690 17631 8178.62 0.032 7.99 4.82 7 NF 1982 566.2 27.6 80.13 17.3 49.4^371690 19707 8935.01 0.032 9.60 5.77 7 NF 1983 571.4 27.7 78.19 19.2 51.9^371690 20605 9599.23 0.032 10.19 6.22 7 NF 1984 572.4 28.0 77.98 20.6 47.5^371690 20308 10389.59 0.032 10.55 6.29 7 NF 1985 571.5 28.1 74.60 20.9 40.3^371690 20125 11142.61 0.032 13.10 6.54 7 NF 1986 568.3 28.5 76.99 18.7 45.6^371690 20723 11933.84 0.032 12.20 3.60 7 NF 1987 568.1 28.6 77.37 16.8 48.8^371690 21336 12997.71 0.032 11.26 4.26 7 NF 1988 568.8 28.5 75.74 15.2 51.8^371690 22835 13990.86 0.032 11.94 3.28 7 NF 1989 571.1 28.6 82.30 15.4 52.0^371690 24252 14841.53 0.032 11.13 3.86 7 NI' 1990 572.7 28.6 82.93 17.1 55.1^371690 25200 15343.11 0.032 11.02 4.76 7 - 51 - APPENDIX 3 NOTES TO JAPANESE MANUFACTURING GREENFIELDS AND PROVINCIAL DATA 1. If the year of establishment was 1991 then for our purposes we classified this establishment as being in 1990, because data on provincial variables was incomplete for 1991. 2. If the year of establishment was actually the year of operation, then to approximate the year of establishment we subtracted one year. 3. Highway miles were assumed to remain constant over the eleven year period. Provincial figures for 1988 were used. 4. Crude oil prices were not available from Statistics Canada for New Brunswick and were incomplete for Quebec. However, given the national energy policy and data that was available these missing figures were estimated. The Alberta amount plus $10.00 was substituted for Quebec and New Brunswick was assumed to face the same oil prices as Quebec. APPENDIX 4 JAPANESE MANUFACTURING GREENFIELDS PROVINCIAL INDUSTRIAL LEVEL FOR 1987 WAGES, FUEL, and REVENUE (In Thousands) SIC Firm # Prov Estabs Wages Fuel Revenue 2512 1 1 45 86281 12998 357328 2 336 1158091 137610 5356553 3 66 58145 8953 270520 4 185 202220 31180 867672 5 338 365629 53416 1794828 6 10 15995 2303 66928 3391 2 1 0 2 4 3 0 4 16 45225 4347 345477 5 5 7308 986 38924 6 0 1811 3 1 1 2 1 3 0 4 16 163922 33971 1004544 5 12 6 0 2512 4 1 45 86281 12998 357328 2 336 1158091 137610 5356553 3 66 58145 8953 270520 4 185 202220 31180 867672 5 338 365629 53416 1794828 6 10 15995 2303 66928 3255 5 1 0 2 1 3 1 4 33 191871 14738 1006041 5 10 6 0 2599 6 1 14 3712 267 12342 2 9 22 3462 3 8 4 99 5 80 27802 2840 123278 6 4 549 12 955 2512 7 1 45 86281 12998 357328 2 336 1158091 137610 5356553 3 66 58145 8953 270520 4 185 202220 31180 867672 5 338 365629 53416 1794828 6 10 15995 2303 66928 - 53 - 1131 8 1 2 3 4 5 6 7 8 3 15 4 3 24954 45047 170444 194485 11664 2680 3486 14605 14208 1199 132941 219923 1180198 679285 61786 2512 9 1 45 86281 12998 357328 2 336 1158091 137610 5356553 3 66 58145 8953 270520 4 185 202220 31180 867672 5 338 365629 53416 1794828 6 10 15995 2303 66928 2711 10 1 2 2 16 322950 149825 2629804 3 6 125641 68704 892708 4 5 125540 63380 944065 5 8 94275 44371 716661 6 1 2711 11 1 2 2 16 322950 149825 2629804 3 6 125641 68704 892708 4 5 125540 63380 944065 5 8 94275 44371 716661 6 1 3199 12 1 26 7903 314 31486 2 62 30489 1057 112847 3 7 3300 111 9812 4 450 526267 25434 2681006 5 191 207077 8114 771907 6 7 3099 13 1 39 13182 837 55700 2 43 12532 832 45750 3 2 4 249 205147 19611 902069 5 103 52915 3613 217599 6 10 1254 133 6365 2712 14 1 0 2 4 262219 171838 1719818 3 3 4 9 348143 169559 1808450 5 21 727865 401159 3679968 6 0 3256 15 1 2 2 6 3 0 4 75 244844 19772 1119061 5 10 21581 2040 118368 6 0 3259 16 1 2 3 4 5 6 3 12 4 137 23 1 644112 20550 37826 925 3300384 90632 3231 17 1 0 2 3 3 0 4 15 1849987 89923 36834766 5 5 6 1 3062 18 1 4 1284 44 2422 2 14 2368 99 6907 3 0 4 427 271055 9612 679773 5 86 27339 982 62324 6 0 1011 19 1 64 119395 9747 2340450 2 45 74661 5597 625625 3 6 4 198 336150 28638 3237259 5 138 183860 25182 2198977 6 27 34637 3314 477239 3251 20 1 0 2 4 3 1 4 37 594076 51832 3407927 5 5 544 53 2142 6 0 3256 21 1 2 2 6 3 0 4 75 244844 19772 1119061 5 10 21581 2040 118368 6 0 3341 22 1 0 2 1 3 0 4 16 5 8 6 0 2912 23 1 3 2 4 3 0 4 12 5 5 6 2 3361 24 1 2 3 4 5 6 8 17 0 86 29 1 7716 33102 238360 60226 77 339 6017 1370 18670 79239 1380276 261013 3361 25 1 8 7716 77 18670 2 17 33102 339 79239 3 0 4 86 238360 6017 1380276 5 29 60226 1370 261013 6 1 3257 26 1 3 499 17 1866 2 0 3 0 4 10 195106 3818 873087 5 4 167 14 858 6 2 3259 27 1 3 2 12 3 4 4 137 644112 37826 3300384 5 23 20550 925 90632 6 1 3231 28 1 0 2 3 3 0 4 15 1849987 89923 36834766 5 5 6 1 3257 29 1 3 499 17 1866 2 0 3 0 4 10 195106 3818 873087 5 4 167 14 858 6 2 3562 30 1 13 4372 325 24050 2 17 7451 316 27461 3 2 4 65 116417 12157 464558 5 41 28854 1835 150520 6 4 3399 31 1 2 2 3 3 0 4 43 83529 9900 366685 5 21 37673 6340 178511 6 1 1621 32 1 2 3 4 5 6 8 9 3 36 15 3 7149 7794 44759 23691 1063 616 6299 2564 51757 39474 366865 312982 3341 33 1 0 2 1 3 0 4 16 5 8 6 0 3011 34 1 4 6615 230 21898 2 3 3 2 4 24 111223 3336 464761 5 6 27561 1029 158376 6 2 351 35 1 10 2 5 3 5 4 45 64697 20306 289807 5 14 21030 7955 69494 6 3 3259 36 1 3 2 12 3 4 4 137 644112 37826 3300384 5 23 20550 925 90632 6 1 3192 37 1 117 127396 6455 539744 2 68 57044 2224 247175 3 7 7331 316 36916 4 222 384532 14476 2115972 5 88 73083 3165 363639 6 13 5407 301 28464 3259 38 1 3 2 12 3 4 4 137 644112 37826 3300384 5 23 20550 925 90632 6 1 3254 39 1 1 2 5 1265 139 5164 3 0 4 23 5 6 4637 825 28465 6 0 3081 40 1 2 3 4 5 6 113 180 22 652 369 45 33221 61610 7846 212299 84173 12519 1491 2556 418 9074 4073 588 87487 171660 22358 556718 241621 45003 3731 41 1 9 33247 11876 524571 2 7 6225 685 71206 3 0 4 52 154734 44597 1723356 5 26 52056 18250 642122 6 0 3092 42 1 11 2 4 2185 46 8212 3 1 4 22 26161 1062 165016 5 5 6 0 3231 43 1 0 2 3 3 0 4 15 1849987 89923 36834766 5 5 6 1 3041 44 1 20 11069 1527 43682 2 24 7894 865 19699 3 1 4 178 172338 24917 705846 5 50 22309 2964 69130 6 6 2190 287 5997 3399 45 1 2 2 3 3 0 4 43 83529 9900 366685 5 21 37673 6340 178511 6 1 319 46 1 163 2 177 3 20 4 820 1134868 49033 5697737 5 345 345967 14774 1383431 6 22 11326 701 57585 391 47 1 51 26408 516 94609 2 69 3 8 4 276 459214 15372 1985639 5 139 6 19 3257 48 1 2 3 4 5 6 3 0 0 10 4 2 499 195106 167 17 3818 14 1866 873087 858 3259 49 1 3 2 12 3 4 4 137 644112 37826 3300384 5 23 20550 925 90632 6 1 3281 50 1 3 717 51 2079 2 79 20594 841 90377 3 18 4 101 31106 1823 149543 5 65 26920 1122 141114 6 0 1021 51 1 1 2 48 100725 6435 1067023 3 73 54517 5384 452911 4 19 5 40 38372 4005 207321 6 1 2961 52 1 4 2 9 3 0 4 32 127082 19981 878524 5 23 71529 11928 707052 6 1 3011 53 1 4 6615 230 21898 2 3 3 2 4 24 111223 3336 464761 5 6 27561 1029 158376 6 2 337 54 1 17 2 22 14877 481 56592 3 2 4 154 479350 17245 1752783 5 71 88879 3601 399546 6 4 5267 934 23034 3194 55 1 3 2 10 8303 166 29236 3 4 1175 84 3698 4 63 139225 5218 493724 5 33 36346 2311 167758 6 1 PROVINCE CODES 1 = Alberta^2 = British Columbia^3 = New Brunswick 4 = Ontario 5 = Quebec^ 6 = Saskatchewan - 59 - APPENDIX 5 JAPANESE MANUFACTURING GREENFIELDS NATIONAL INDUSTRIAL LEVEL FOR 1987 WAGES, FUEL and SHIPMENTS (In Millions) SIC Establishments Wages Fuel Shipments 319 1601 1851 79 8638 337 281 624 23 2390 351 90 107 33 413 391 603 591 18 2451 1011 526 851 82 9811 1021 414 530 43 4111 1131 48 502 42 2557 1621 77 89 11 801 1811 31 211 43 1199 2512 1093 1919 252 8862 2599 272 63 5 275 2711 39 757 355 5817 2712 43 1572 927 8537 2912 26 81 13 204 2961 71 215 33 1701 3011 43 148 4 652 3041 288 217 30 850 3062 539 303 10 753 3081 1464 436 19 1194 3092 44 71 2 274 3099 463 290 25 1257 3192 533 666 27 3387 3194 116 187 7 703 3199 772 802 36 3724 3231 27 2116 100 39093 3251 50 598 52 3430 3254 35 168 16 720 - 60 - 3255 47 209 16 1086 3256 96 270 22 1252 3257 20 196 3 877 3259 190 672 39 3429 3281 327 93 5 442 3341 25 83 1 822 3361 147 367 8 2006 3391 27 66 6 420 3399 72 122 16 552 3562 153 161 14 688 3731 94 246 75 2961 3931 209 144 8 788 APPENDIX 6 CONDITIONAL LOGIT REGRESSION RESULTS VARIABLES & THEIR SOURCE ^LW =^log of manufacturing wages as listed in Appendix 2. ^ LE =^log of weighted average energy prices as listed in Appendix 2. LE2 =^log of low energy prices as listed in Appendix 2. SHIPRANK =^as listed in Appendix 2. LMETGDP =^log of (metropolitan times GDP per Capita times population, from Appendix 2). LCAN =^log of the number of Provincial firms in like industry as listed in Appendix 4. LJPN =^log of the number of Japanese manufacturing greenfields in the province as per Table 1. QUEBEC =^A dummy variable to measure the effect of this large province with few Japanese greenfield investments. LABOR = Using data from Appendix 5 to estimate the labour. factor input coefficient for each specific industry for the Cobb - Douglas Production Function. ENERGY =^Using data from Appendix 5 to estimate the energy factor input coefficient for each specific industry for the Cobb - Douglas Production Function. UNION_P =^Unionization rate in the province as listed in Appendix 2. TAX =^Provincial Corporate Tax Rate as listed in Appendix 2. CRIME =^Provincial crime rate per 1000 divided by the provincial population. Both of these numbers are in Appendix 2. 55 Japanese Manufacturing Greenfields CHOICE FREQUENCY PERCENT 1 (AB) 1 1.8182 2 (BC) 11 20.0000 3 (NB) 2 3.6364 4 (ON) 35 63.6364 5 (PQ) 4 7.2727 6 (SK) 2 3.6364 Without a Measure of Market Size (i.e. GDP) LOG OF LIKELIHOOD FUNCTION^: -58.3579 NUMBER OF CASES^: 55 NUMBER OF CHOICES : 330 Standard Parameter^Estimate^Error^t-statistic LCAN .489813^.28027^1.74765 LJPN^1.21409^.435358^2.78872 SHIPRANK^-.324323^.196032^-1.65444 LW^-5.31907^4.93843^-1.07708 LV -4.91472^2.68017^-1.83373 lw = log(avgmfgy), lv = log(avgengy) A better fit is achieved with GDP LOG OF LIKELIHOOD FUNCTION^: -55.1337 NUMBER OF CASES^: 55 NUMBER OF CHOICES : 330 Standard Parameter^Estimate^Error^t-statistic LCAN .677792^.285464^2.37435 LJPN^2.26426^.641879^3.52755 SHIPRANK^-.316939^.203115^-1.56039 LW^-8.23081^5.30678^-1.5510 LV -6.42729^2.72486^-2.35876 LGDP^-1.29039^.501009^-2.57558

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
Japan 6 0
United States 5 5
China 4 4
Germany 2 7
United Kingdom 1 0
City Views Downloads
Tokyo 6 0
Beijing 4 2
Unknown 3 7
Wilmington 2 0
Dolton 1 0
Redmond 1 2
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items