UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Capitalization effects of creative mortgage financing Graham, Barbara Elizabeth Anne 1985

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

Item Metadata

Download

Media
831-UBC_1985_A4_6 G73.pdf [ 6.4MB ]
Metadata
JSON: 831-1.0095930.json
JSON-LD: 831-1.0095930-ld.json
RDF/XML (Pretty): 831-1.0095930-rdf.xml
RDF/JSON: 831-1.0095930-rdf.json
Turtle: 831-1.0095930-turtle.txt
N-Triples: 831-1.0095930-rdf-ntriples.txt
Original Record: 831-1.0095930-source.json
Full Text
831-1.0095930-fulltext.txt
Citation
831-1.0095930.ris

Full Text

CAPITALIZATION E F F E C T S O F C R E A T I V E M O R T G A G E FINANCING By B.E. A N N E G R A H A M B.Comm., The University of British Columbia, 1980 A THESIS SUBMITTED IN PART IAL F U L F I L L M E N T O F T H E REQUIREMENTS FOR T H E D E G R E E O F M A S T E R O F SCIENCE in Business Administration in T H E F A C U L T Y O F G R A D U A T E STUDIES Commerce and Business Administration We accept this thesis as conforming to the required starud^f d T H E UNIVERSITY O F BRITISH COLUMBIA November 1985 Q B.E. Anne Graham, 1985 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of URBAN LAND ECONOMICS The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date OCTOBER 1, 1985 DF-fi r^/R-n ii A B S T R A C T This thesis studies the impact of creative financing on single family house prices, focussing on instruments such as Agreements-for-Sale, assumed mortgages and vendor-financed loans. Benefits from financing at below-market interest rates are expected to be capitalized into higher sale prices of houses, however, the literature is not clear on the appropriate adjustment. If the purchase price of houses increased by an amount equal to the present value of the payment savings, there would be no direct benefit for purchasers and they would effectively be using a financing scheme similar to a graduated payment mortgage. A Canadian sample of housing transactions has been used, specifically in the Lower Mainland area of British Columbia between 1980 and 1982. The sample was selected to include home sales at different price levels and during different periods of market activity. In this way, the capitalization effect of creative financing instruments could be tested at different house price levels and market conditions, as well as for alternative types of creative funds. Creative financing arrangements can be classified as institutional or non-institutional depending on the loan origination source, and it was hypothesized that assumption loans (institutional origination) would be properly capitalized in the price while vendor-financed loans (non-institutional origination) would be overcapitalized in house prices. The reason for the dual pricing response hinges on the discount rate used to calculate the benefit of below-market financing; it should be higher for vendor-financed loans due to the higher costs in loan origination and servicing for the seller as compared to a financial intermediary. The results indicated that assumption loans were properly capitalized in the price of single family homes as the coefficient of the creative financing variable assumed a value of approximately 1. The vendor-financed sample revealed that the benefit for below-market financing was overcapitalized in the price as the coefficient i i i was generally in excess of 3. The research suggests that the market interest rate is not the appropriate discount rate to use in deriving the benefit from vendor-financed loans and possibly a higher rate should be used. This is an area which requires further study. A sample of house sales which includes information on the secondary yields of vendor-financing from sellers to other mortgage market participants should help to identify the true market rate for these non-institutional loans and provide for a more precise calculation of the present value of the payment savings in the case of vendor-financed loans. iv T ABLE OF CONTENTS Abstract i i Table of Contents iv List of Tables v i List of Figures v i i Acknowledgements v i i i Chapter 1.0 - Introduction 1.1 Comments on the Existing Valuation Literature 1 1.2 Significance of Thesis 3 Chapter 2.0 - Review of Creative Financing Instruments and Valuation 2.1 Development of Alternative Mortgage Instruments 7 2.2 Features and Implications of Creative Financing Instruments 12 2.3 Valuation Issues and Methodology 18 2.4 Summary 23 Chapter 3.0 - Review of Capitalization Effects 3.1 Introduction 27 3.2 Empirical Studies 3.2.1 Cash-Equivalent Method 27 3.2.2 Capitalization of Creative Financing 31 3.3 Factors Affecting the Capitalization of Creative Financing 42 3.4 Summary 44 Chapter 4.0 - Empirical Models of House Prices 4.1 Literature Review 46 4.2 Specified Housing Price Equation 51 Chapter 5.0 - Data Description 5.1 Sampling Issues 56 5.2 Sample Statistics 68 Chapter 6.0 - Housing Price Equation Results 82 Chapter 7.0 - Summary and Pricing Implications 96 Bibliography 100 Appendix A: Vancouver (East Side) Neighbourhood Location Map 105 Appendix B: Richmond Neighbourhood Location Map 106 Appendix C: Vancouver (West Side) Neighbourhood Location Map 107 Appendix D: North Vancouver Neighbourhood Location Map 108 Appendix E: West Vancouver Neighbourhood Location Map 109 Appendix F: Raw Data Input Format 110 Appendix G : Conventional Residential First Mortgage Rates, 1980 112 Appendix H: Conventional Residential First Mortgage Rates, 1981 113 Appendix I: Conventional Residential First Mortgage Rates, 1982 114 Appendix 3: Secondary Financing, Conventional Residential Second Mortgage Rates, 1980 - 1982 •. 115 Appendix K: Neighbourhoods by Enumeration Areas 116 vi LIST O F T A B L E S Table Page 2.1 Forms of Creative Financing 9 2.2 Relationship of Mortgage Values 22 3.1 Calculation of Vendor Financing Benefit to Borrower 33 3.2 Summary of Capitalization Studies 35 5.1 Trends in House Prices by Area, 1980 - 1983 62 5.2 Multiple Listing Sales Volume by Area, 1980 - 1982 64 5.3 Quarterly House Sales by Neighbourhood, 1980 - 1982 65 5.k Preliminary Sample Statistics 67 5.5 Final Sample by Year and Neighbourhood 68 5.6 Final Sample by Year and Financing Arrangement 69 5.7 Conventional First Mortgage Rates, 1970 - 1982 70 5.8 Incidence of Assumed and Vendor-Financed Loans by Year 72 5 .9 List of Regression Variables 73 5.10 Annual Sample Financing Statistics by Neighbourhood 75 5.11 Means and Standard Deviations of Variables 77 5.12 Correlation Matrix of Variables 79 6.1 Equation for the Full Sample 85 6.2 Results by House Price Level and Year of Sale 87 6.3 Results for the Vendor-Financed Sample 89 6A Results for the Loan Assumption Sample 91 vii LIST O F FIGURES Figure Page 5.1 Regional Location Map 57 5.2 Conventional First Mortgage Rates, 1980 - 1982 59 5.3 Average Residential Selling Price in Greater Vancouver, 1965 - 1981 . . . . 61 viii A C K N O W L E D G E M E N T S I would like to thank the members of my thesis committee, Dr. George Gau, Dr. Stanley Hamilton and Dr. John Claxton, for their help in completing this research. Dr. George Gau, thesis chairman, provided tremendous guidance throughout the project. I appreciate the assistance given by Dr. Larry Jones, Dr. Michael Goldberg, and Mr. George Oikawa and Cumberland Realty Group Limited to begin the masters program. I am particularly grateful to Shirley Koelling and Tina Hansen who made it possible to complete this thesis. Canada Mortgage and Housing Corporation provided financial assistance. 1 1.0 INTRODUCTION 1.1 Comments on the Existing Valuation Literature The real estate valuation literature stresses the need to eliminate, in appraising a property, differences in the sale prices of comparables which result from financing terms: "Financing can and often does make a difference in price ... thus, to the extent that financing arrangements for a comparable differ from the "normal market" financing pattern for the type of property under appraisal, an adjustment would be required."* This relationship between financing and value (price) indicates that it is important to establish which contractual arrangements are typical of normal financing in the market as well as the appraisal procedures to account for differences in financing terms. However, little is known on the actual incidence of creative financing arrangements and a consensus has not been reached on the appropriate adjustment process; these two topics form the basis of this thesis. A number of authors have indicated that transactions which involve other than a cash payment or cash equity plus a conventional first mortgage should be adjusted to a "cash-equivalent" basis; 2 this is done by adding the market value of any mortgages to the down payment amount. This approach has been criticized by others as not accurately adjusting for the effects of nonconventional financing for several reasons:^ 1. The use of the conventional first mortgage rate as a standard may cause an incorrect adjustment since this rate is not consistent with the concept of 'normal' financing in the market. 2. The use of the conventional first mortgage rate may lead to an overadjustment when applied to loans with shorter terms (e.g. loan assumptions) and greater default risk (e.g. second mortgages). 3. An overadjustment will result if the entire amount of the premium for favourable financing is not actually paid by the purchaser. In the context of this thesis, the term "creative financing" is meant to include those mortgages which have non-market contract interest rates, typically below existing rates; however, in some cases, contract rates on creative instruments are 2 above market levels for reasons presented later. The instruments generally studied in this research include Agreements-for-Sale,^ assumed mortgages and vendor-financed loans, and a key issue to be addressed is why these types of arrangements have become increasingly popular. The arrangements are described as follows: (a) Agreement-for-Sale: A contract whereby the buyer agrees to buy real property through installment payments, while the seller retains legal ownership to the property until the obligation is fully repaid with interest. (b) Assumed Mortgage: The buyer assumes all obligations under the terms of an existing mortgage. However, the original borrower (and seller of the property) is responsible should the buyer default on the assumed loan, without retaining legal title to the property. (c) Vendor-Financed Loan: Although it may take a variety of forms, the vendor provides the buyer with financing which bypasses traditional sources for a period of time, typically involving lower cash downpayments and below-market interest rates. Creative financing instruments commonly have below-market interest rates in the initial years of the loans, thereby providing borrowers with savings on their monthly mortgage payments during the life of the creative financing. However, if the acquisition price of houses increased by an amount equal to the present value of these payment savings, there would be no direct gains for purchasers. In fact , borrowers would have smaller monthly payments while the creative financing was outstanding but larger mortgage payments upon maturity of this financing to compensate for the higher acquisition prices paid. Consequently, they would end up with a financing arrangement similar to a graduated payment mortgage. Graduated payment mortgages have payments which increase over some, or al l , of the life of the loan in an attempt to match a growing payment with a borrower's increasing income. The initial payments on this loan are usually not sufficient to pay the interest that is due, and any interest not paid is added to the outstanding balance of the loan, accruing interest at the contract rate. The risk elements on the graduated payment mortgage stem from both the increasing indebtedness over the initial years of the loan (capital risk to the lender) and the increasing size of payments over time (default risk to the borrower).^ 3 If the savings benefits from creative financing are not fully capitalized in house prices, it suggests that borrowers are reacting to a market imperfection which allows them to reduce their overall financing costs. On the other hand, if the benefits are fully capitalized, creative financing must be considered as a graduated payment mortgage, an alternative mortgage design which is not common in the Canadian residential mortgage system at this time. Creative financing instruments allow variations in the terms of the loans including the interest rate, principal amount, loan-to-value ratio and payment stream, thereby providing flexibility to the parties of the transaction. While it has been reported that the incidence of creative financing increases during periods of high interest (thus, high cost of borrowing from traditional sources),** there is no current data on the incidence of non-conventional instruments in a Canadian real estate market. Furthermore, evidence from past studies do not clearly establish whether the benefits from creative financing instruments are fully capitalized in the price of single family houses, primarily due to ambiguities in the sampling information. In the U.S., residential loans are typically for longer terms and the interest portion of the mortgage payment is tax deductible. As a result, the benefits of a below-rate loan accrue over a longer period, making the present value of the savings much larger. 1.2 Significance of Thesis This study attempts to improve on the existing research which utilizes data on U.S. home sales, by collecting a random sample of Canadian housing transactions to use in the capitalization tests. The benefit of this research rests on the use of a Canadian data base where the transactions are not affected by tax and interest rate risk considerations. In Canada the maximum term of residential loans is generally five years based on a partially amortized, constant payment loan, and interest on owner-occupied homes is not tax deductible. Creative financing arrangements can be classified as "institutional" or "non-institutional", with the former category including It primarily loan assumptions and the latter category comprising vendor-financed mortgages. The distinguishing feature of these two groups is that "institutional" creative funds are loans originated by a financial intermediary. The loan assumption is based on a conventional mortgage provided by a standard mortgage lender, and the calculation of the present value of the payment savings should be based on the existing market rate for similar conventional mortgages (i.e., similar loan-to-value ratio, amortization period, term, etc.). On the other hand, the vendor-financed loan is originated outside of the normal intermediation channels since the seller acts as the mortgage lender. The appropriate discount rate to determine the payment savings should be higher for a vendor-financed loan than an assumed loan since the vendor does not have the economies of scale in loan origination, servicing or analysis of default risk that are afforded by large financial institutions. Assumption loans and vendor-financed loans also differ in that the assumption loan is fixed in amount at the sale date while the vendor-financed loan is a function of the price of the housed A sample which allows independent analysis of both "institutional" and "non-institutional" creative instruments should provide greater insight of the capitalization affect . A sample of single family home transactions in the Lower Mainland area of British Columbia between 1980 and 1982$ has been used. Neighbourhoods within the Lower Mainland area were specifically selected to obtain a representative sample of low, middle and high-priced house transactions, thereby permitting housing submarkets to be studied. The 1980 - 1982 time frame was chosen to represent a cyclical period, during which time residential property values and conventional mortgage rates peaked and declined dramatically. The thesis is divided into seven chapters. Chapter Two outlines basic characteristics of creative financing, the valuation issues arising from these arrangements and the current appraisal methodology employed in pricing these instruments. Chapter Three reviews the literature on the extent to which creative 5 financing is capitalized in the sales price of residential property and the possible reasons for the lack of perfect capitalization. Chapter Four discusses the capitalization models and specifies the housing price equation to be used in this analysis. Chapter Five describes the data base which was selected and the market conditions during the study period. Chapter Six presents the methodology used and the housing price equation results. Chapter Seven summarizes the pricing implications for creative financing arrangements and suggests areas of further research. 6 F O O T N O T E S 1 Corgel and Goebel, "Financing Adjustments via Cash Equivalency," The Real  Estate Appraiser and Analyst, 49 (Spring 1983), p. 55. 2 See Garcia, p. 9; Schwartz, p. 35; Lipscomb, p. 23; Friedman and Lindeman, p. 43; Maes, p. 88. 3 See Corgel and Goebel, p. 55; Findlay and Fischer, p. 65. 4 In the final sample of properties used in this thesis, only 7 of the 350 transactions involved Agreements-for-Sale. Consequently, the focus of this research is on assumed and seller-financed loans. 5 S.W. Hamilton, et al , Real Estate Finance in a Canadian Context, September 1984, p. 10-14. 6 See Maes, p. 87; Schwartz, p. 35; Case, p. 45; Koch et al , p. 4. 7 Claurentie notes that research on seller-supplied financing have used hedonic pricing models which may produce biased estimates of the capitalization process. Unlike the assumption loan where the loan amount is fixed, the owner-financed loan is a function of the price of the house which causes a problem of simultaneity bias. This bias is discussed in Chapter 4.0. 8 The study extends from the first quarter of 1980 to the beginning of the fourth quarter of 1982 as this was the period of rising and falling interest rates and house prices in the Lower Mainland of B.C. 7 2.0 REVIEW O F C R E A T I V E FINANCING INSTRUMENTS A N D V A L U A T I O N , 2.1 Development of Alternative Mortgage Instruments In Canada, the principal mortgage instrument in the NHA-guaranteed and conventional mortgage markets since 1969 has been the rollover, renewable term mortgage.* Basically, this type of loan has a maximum term of five years, amortized over a period of 20 - 30 years in the case of conventional mortgages. The mortgage is subject to renewal at the end of the term at the prevailing market interest rate with no penalty for repayment of any portion of the outstanding loan balance by the borrower. The rollover design arose from provisions in the Federal Interest Act of 1931 which permitted the borrower to repay the mortgage at any time after five years with a maximum penalty of three months interest. 2 Thus, in the attempt to prevent the potential problem of repayment at the borrower's discretion, Gau suggests that lending institutions responded with the rollover design. It was only in 1969 that the National Housing Act was amended to alter the NHA mortgages from their minimum term of 25 years to allow the renewable term, rollover design. Discussions of housing affordability in recent years have focussed on the problems of rising interest rates and house prices. Together, these elements have "caused the required payment obligations of new residential mortgages to grow at a faster rate than Canadian nominal income".^ The gap created between the real after -tax ownership costs of housing and down payment and the mortgage payment requirement has been described as follows: ". . ."real after-tax ownership cost" is a long term concept that reflects the wealth that will accrue to the owner in the form of capital appreciation and tax savings. Loan qualification standards consider only the buyer's current ability to handle the cash flow demands that the mortgage imposes upon him."^ 8 The use of the standard, level-payment mortgage during periods of high and variable rates of inflation has an adverse affect on the demand and supply of housing. Two major shortcomings are cited with the traditional standard mortgage:^ 1. The standard mortgage involves equal nominal payments throughout the term which suggests that during periods of high inflation, the real pattern of payments (actual payments deflated by the inflation index) declines. In an effort to offset the decline, lenders demand higher debt service payments in the early years of the instrument, creating a 'tilt' problem for the borrower. These higher initial payments occur at a time when households can least afford it , and they do not allow households to acquire housing services commensurate with their preferences and future income expectations. 2. Although financial institutions attempt to match the rates given on their deposit-liabilities and mortgage-assets, interest rates on the traditional mortgage instrument cover a long interval of time and liabilities are generally issued with much shorter maturities. With increasing and unstable rates of inflation and interest, many lenders have watched their short term cost of funds increase rapidly as they continue to hold long term fixed rate mortgages. This interest rate risk problem is commonly referred to as the "asset-liability mismatch". Alternative mortgage instruments have been proposed as one method of correcting the aforementioned shortcomings in the housing finance system. While the rollover mortgage is the predominant instrument in the Canadian residential mortgage market, there are several alternative designs which can be created by varying the conditions of the loan contract. Table 2.1 outlines several forms of creative financing, and Gau identifies three main changes to the conventional design: 1. Manipulating the length of the term as well as the loan specifications which govern interest rate adjustments (fixed payment mortgage, variable rate mortgage); 2. Altering the pattern of either the nominal (graduated payment mortgage) or the real (price level adjusted mortgage) monthly loan payments; and 3. Incorporating an equity participation by the lender in any capital gains of the borrower resulting from a rising value of the collateral property (shared appreciation mortgage). The main objectives of alternative mortgage instruments include reducing the fluctuations in housing demand and permitting households to better match housing preferences with permanent income projections.^ Several of the instruments are directed largely at solving the 'tilt' problem (e.g. graduated payment mortgage and 9 TABLE 2.1 FORMS OF CREATIVE FINANCING 1. A D J U S T A B L E INTEREST M O R G A G E (AIM): The interest charged is based on a selected index. Payments may be changed as the index changes or, alternatively, the term of the loan is extended by the interest not covered in the fined payment. 2. A D J U S T A B L E M O R T G A G E L O A N (AML): The payment term and/or interest may be changed according to a selected indicator(s). 3. A S S U M E D M O R T G A G E : The buyer agrees to pay all obligations due under a loan existing on a property being purchased. 4. D E F E R R E D INTEREST M O R T G A G E (DIM): A l l payments are used to reduce the amount owed (principal) with interest deferred and paid at a later date. 5. D U A L INTEREST M O R T G A G E (DIM): Initial payments are based on a long-term repayment plan at a lower interest rate and payment, with the amounts owned calculated on a current higher money market rate. The difference between the amounts paid and owed are added to the outstanding principal balance. 6. S H A R E D APPRECIAT ION M O R T G A G E (SAM): The lender gives favorable lending terms (usually in the form of a lower interest rate) in return for a share in future property increases. 7. F IXED P A Y M E N T : The traditional mortgage where a level payment is made regularly and a portion is first credited to the interest due and the remainder to reduce the principal due. 8. SINKING F U N D ASSISTED M O R T G A G E S (SFAM): A portion of loan proceeds is placed in an interest bearing savings account with amounts periodically withdrawn and credited to principal reduction. 9. G R A D U A T E D P A Y M E N T M O R T G A G E S (GPM): Total payments are related to changes in a selected indicator with early payments lower than later payments. The term, interest and principal due may be flexible. 10. G R O U N D L E A S E : Land is rented on a long term basis while the improvements may be sold separately. 11. INTEREST O N L Y L O A N S : Al l loan payments are credited to the interest due and the principal is paid as a lump sum at a later date. 12. INTEREST B U Y D O W N : The borrower (or seller) makes a cash payment to the lender which is applied to the interest due in order to reduce loan payments and borrwer qualification. 13. JUNIOR INSTRUMENTS: A lending instrument which has less priority in recovering amounts due than instruments recorded ealier (i.e., 2nds, 3rds, 4ths, etc.). 10 TABLE 2.1 (cont'd) 14. L A N D C O N T R A C T : The property is sold but the seller retains the title and receives periodic payments which are credited towards the price. Similar to A G R E E M E N T S FOR S A L E in Canada. 15. PRICE L E V E L A D J U S T E D M O R T G A G E (PLAM): Fixed interest with the value of the property adjusted to reflect changing prices; the interest rate is applied to the adjusted price. 16. P U R C H A S E M O N E Y M O R T G A G E : The seller takes back a mortgage as partial payment for the property. Also referred to as VENDOR F I N A N C E D M O R T G A G E S . 17. R E V E R S E ANNUITY M O R T G A G E (RAM): The property area receives periodic payments which must be repaid at a later date. Commonly used by property owners with substantial equities who wish to supplement their monthly incomes. 18. R O L L O V E R (CANADIAN) : The term of the mortgage is renegotiated at the end of the loan term (usually five years) with the borrower having the right to find a new lending source at no penalty; the loan is amortized usually over 25 years. Also referred to as the P A R T I A L L Y A M O R T I Z E D , C O N S T A N T P A Y M E N T M O R T G A G E . 19. VARIABLE R A T E M O R T G A G E (VRM): Changes in the interest rate are reflected by changes in the payment, the term of the loan, or combination of both. S O U R C E : F .E . Case, "Creative Financing Instruments," The Real Estate Appraiser  and Analyst (Spring 1982), pp. 49, 50. 11 price level adjusted mortgage), while other designs have been promoted as a way of lessening the asset-liability mismatch (eg. variable rate mortgage).^ The provision of alternative mortgage instruments requires consideration of the supply of mortgage funds and development of secondary markets. Changes in the housing finance industry require solutions to the interest rate risk and affordability problem of the traditional, long term, fixed rate mortgage under inflationary times as well as a supply of alternative financing sources to the deposit institutions.^ While conventional mortgage sources may continue to offer new design to address the risks in lending, the inability of conventional sources to supply adequate mortgage funds will result in a large number of transactions involving cash or creatively financed sales.^ Funds from non-conventional sources do not have institutionally imposed criteria and may be based on the desire to sell the property rather than assessing loan yields and risks. Treadway suggests that the secondary market will need to rely more heavily on private sources of funds such as pension funds and a private mortgage-backed security. However, he acknowledges that pension fund portfolios have not generally sought the outright purchase of mortgages. Development of a successful conventional passthrough security by the private market will require the creation of a more homogeneous instrument with possibly a central or limited number of issuers to prevent fragmentation, modification of regulatory and accounting practices, offerings at competitive yields and terms, as well as the establishment of a forward commitment market. Another major controversy regarding the use of alternative mortgage instruments rests on the "determination of a 'fair' or 'appropriate' price for these instruments, either in absolute terms or relative to the standard fixed payment mortgage".10 According to Cassidy and Field, the problem of establishing the probable range of market prices for these new designs is complicated by the following items: - it is difficult to compare different non-interest rate terms, even when the contract rates are similar; 12 - the associated uncertainty concerning the interest rate and default risks; and - the short time that alternative mortgage instruments have been used and the fragmentary data of their prices and performance. Despite the concerns, much of the literature indicates that the conventional level payment mortgage instrument will remain as the principal design in the U.S. and Canadian residential mortgage markets (see Gau, Cohn & Fischer, Swan). Alternative mortgage instruments are viewed as a way of improving a household's financing choice since the standard fixed rate mortgage is no longer considered appropriate for all borrowers. Because "different mortgage instruments appeal to different segments of the market, the availability of a range of instruments is necessary to meet the needs of both borrowers and lenders".* 1 Colton (et al) advises that while it is desirable to offer consumer choice, it is important that borrower education and safeguards also be provided. In order that such choices are available, financial institutions will have to find other means of matching their deposit liabilities to assets which will emphasize the use of the secondary market .* 2 2.2 Features and Implications of Creative Financing Instruments Creative financing refers to lending arrangements which allow variations in the terms of the agreement including the interest rate, principal amount, loan-to-value ratio and payment stream. *^ Although it has been defined as "any financial arrangement for a real estate transaction involving a non-institutional supplier of funds", *^ this definition is not appropriate since loan assumptions and builder buydowns are originated by institutional lenders. The Case article cites a number of creative instruments incorporating variations of the terms and lists five basic concepts of non-conventional financing: 1. assumed mortgages - the buyer assumes all obligations under the terms of an existing mortgage; 2. deferred interest - while interest due is calculated, it is paid at some later date; 13 3. flexible interest payments - the borrower and lender agree to changes in the interest and payments; 4. equity participation - the lender participates in any price increases in the property in exchange for providing favourable loan terms; and 5. indexing - interest rates and payments are based upon changes in an index such as the Consumer Price Index. Creative financing instruments compiled for this research involve the first and third items. Rosen (1982) presents three major creative financing techniques including assumed mortgages, interest 'buydowns' and seller-financed loans. In the case of a buydown, the seller (usually a new home builder) makes a lump sum payment to a lender to have a lower rate loan (in the early years) provided to buyers. This buydown essentially amounts to a price discount, arising from a lower mortgage interest rate (hence lower monthly payments) rather than a lower sales price, and it enables a larger proportion of prospective purchasers to qualify for new loans. Seller financing may take on a variety of forms such as first or secondary loans, wraparound mortgages, leasing arrangements and Agreements-for-Sale. Although the terms of the instruments differ, they are similar in that the vendor provides the buyer with an extended financing package which bypasses traditional sources for a period of t ime . l ^ Seller-financed arrangements frequently have lower cash downpayments, below-market interest rates and several years of interest only payments followed by a lump sum payment or by level principal payments, as compared to the standard self-amortizing mortgages generally provided by institutional lenders. Typical provisions in seller-financed transactions include release clauses and prepayment privileges: ^ Release Clause: This clause is largely used with seller-financing on land, whereby some portion of the property is released as collateral upon a principal payment. Friedman and Lindeman indicate that institutional mortgages rarely contain release privileges as the entire property is kept as collateral regardless of the size of the outstanding loan balance. Prepayment Clause: While the terms of this clause usually differ, it is intended to permit repayment of principal in excess of the regular payment (usually composed of interest and principal). 14 The growth of creative financing instruments has been attributed to several factors including: - the increase in housing prices and interest rates in the late 1970's (Case, p. 45). As house prices and interest rates increased, it became more difficult for purchasers to meet conventional lending criteria and qualify for new loans; - the absence or high cost of third party financing (Maes, p. 87). The inability of conventional sources to meet demand has encouraged cash or seller-financed sales; - the large number of new instruments created with flexibility in the terms of the lending agreement (Case, p. 46). Despite the growth in non-conventional financing arrangements in the residential market, there is l imited evidence of the incidence of these instruments as the sole or secondary source of funds. Two studies in California estimated the incidence of creative financing in 1981 to be as high as 75% - 80% (and 50% of the value of home sales).*7 A study of residential real estate financing in the southeastern United States between January and June 1982 by Koch (et al) found that 53% of the housing transactions were creatively financed. There is no recent evidence regarding the use of creative financing in Canadian mortgage markets, and it may differ considerably from the U.S. case due to differences in the structure of conventional loans (i.e., loan term, amortization period, tax treatment of interest payments). Earlier work by Eger on the housing market of the Greater Vancouver Regional District in B.C. between 1954 and 1963 indicated that the non-conventional sectors of the mortgage market (such as the Agreements-for-Sale and assigned mortgage sectors) reacted in a cyclical manner;*^ in other words, as interest rates increased, increasing proportions of these sectors were used to finance housing sales. Eger estimated that assigned mortgages on existing housing represented 22% - 32% of the total extended mortgage market, with private vendor financing accounting for 13% - 21% during the 1954 - 1963 period. The size of the submarkets was considered significant, each being approximately 20% of the extended mortgage market which allows a submarket to influence the housing and mortgage markets. One 15 of the objectives of this thesis is to provide a contemporary estimate of the incidence of creative financing from a random sample of home sales during a cyclical period of the real estate market. Borrowers and lenders face a number of risks (eg. interest rate risk, capital (principal) risk, default risk) which depend upon the terms of the lending agreement. Regardless of the instrument design, borrowers are concerned with the size of the periodic payment, the payment required and equity position at the end of the loan term, and their ability to pay over the life of the loan. Risks for the lender generally include repayment by the borrower according to the loan contract, the relationship between the yield received and cost of funds loaned, the relationship between the remaining loan balance and value of the property, and the expense of administering the loan and potential foreclosure costs. Case proposes that the uncertainty associated with selecting the appropriate form of creative financing is compounded by the flexibility of such lending contracts: "In the face of uncertainty and unknown risk, lenders will seek ways of transferring more of the financial risk to borrowers and this has emerged as a major characteristic of creative f inanc ing . "^ The risks of changing interest rates are shared by the lender and borrower with the Canadian rollover mortgage design, since the lender absorbs the interest rate risk during the life of the mortgage and the borrower bears the risk at the end of the loan term.20 Default risk associated with residential mortgages includes the risk of delinquency as well as foreclosure. Delinquency risk relates to the borrower failing to make a payment(s) and is largely influenced by the variability of the incomes and expenditures by borrowers. Foreclosure risk involves a borrower totally ceasing to make mortgage payments and forfeiting the house (collateral) which secures the loan; it is a function of homeowners equity as borrowers with positive net equity positions (current market value of the property is greater than the outstanding mortgage balance) could convert a portion of their equity into debt and use the funds to meet the mortgage payments. During periods of rising interest rates, the greatest risk of 16 default is experienced with the standard rollover instrument since borrowers renew at higher interest rates causing higher mortgage payments. However, an advantage of this mortgage design is that borrowers may select among different loan periods (term of loan) to best fit their expected future income flow. The literature indicates a number of advantages to the seller in financing the transaction. The property could sell faster since it is exposed to a larger group of potential buyers with favourable financing t e r m s . 2 1 Seller financing may be necessary to complete the transaction due to: - characteristics of the property (i.e. collateral and liquidity concerns); - characteristics of the borrower (i.e. default risk); and - characteristics of the money market (i.e. high cost or shortage of funds) . 2 2 Seller-financed loans may also offer higher interest rates (and yields) to the vendor than alternative opportunities (e.g. savings certificates) with the property serving as collateral for a mortgage given to the buyer. Realization of the sale proceeds may be spread over a longer time frame by periodic payments, thereby providing vendors with an income stream rather than a lump sum payment. A higher sales price is likely to be achieved with favourable financing which effectly converts ordinary income into capital gains. A lower interest rate loan, in return for a higher contractual price, increases the seller's capital gain and lowers his ordinary interest income. However, in the case of ownership for investment purposes (i.e. not for occupancy as the principal residence), the entire gain is taxed, and it would not encourage the use of below-market financing and a higher contractual purchase price. It has even been suggested that realtors encourage seller financing to raise the sales price since their commission is paid on the nominal sales price. Some of the disadvantages of seller-financed loans to the vendor include the problem of collection, the opportunity costs of receiving the sale proceeds at some time in the future, and the lack of an organized secondary market for trading these instruments. Since the seller finances a large portion of the property's equity, he may 17 be forced to forego other investment opportunities over the term of the instrument in return for a fixed rate of interest. Often, these contracts are not marketable in the secondary market just for the reasons the financing was provided by the seller and not by a third party; in other words, the property may not be adequate security for the funds loaned, the borrower may not meet standard lending criteria, or the yield on the loan may be lower than conventional mortgages if written at a below-market interest rate. Consequently, a considerable discount from the face value of the loan may be required to resell the instrument. There are advantages to the purchaser in using seller financing in the transaction. Third party funds are not always available and contract rates can be lower than prevailing market interest rates. Higher leverage is possible by unrestricted loan-to-value ratios, and there can be greater flexibility in prepayment terms, release privileges and collateral substitutions. In addition, the buyer may avoid institutionally imposed qualification criteria, loan fees and other closing costs. On the other hand, the buyer is likely to pay a higher price for the property to receive the financing concessions, thereby facing a larger debt repayment at the loan expiry and potential difficulty of refinancing by conventional or non-conventional sources. Because many of the mortgages arranged by non-traditional sources are considered to involve higher risks, other options have been cited to finance homes. Koch suggests two alternative approaches: 2^ 1. negotiating the transaction on the basis of the effective sales price rather than financing terms. 2. negotiating with financial institutions to obtain a financing arrangement similar to that offered by sellers. Their survey reported examples where the purchaser's cost and the seller's return was not affected by reducing the sales price of the house. On that basis, the buyer would finance a smaller amount through a conventional source at a higher interest rate and at a similar mortgage payment. Since interest payments in the U.S. are tax deductible, the buyer would benefit with this relatively higher proportion of the 18 payment being an interest component; this does not apply to owner-occupied Canadian homes. While the seller would receive a lower sales price, the cash could be invested at a higher interest rate with less risk and greater liquidity than providing a low-rate loan to the buyer. It was noted that buyers can receive the same rates sellers are willing to accept by using alternative mortgage instruments, and sellers avoid risk by not holding a mortgage. Although creative financing may help to sell a home, they stress that the risks of the loan and the security provided by the real estate need to be carefully assessed. 2.3 Valuation Issues and Methodology Although a number of the differences between conventional and non-conventional financing instruments have been discussed, it is also necessary to consider how financing terms affect comparable sales data. Lipscomb notes that "favourable financing terms should affect the sales price of real estate only when financing arrangements become a factor in the price negotiations between buyer and seller".2^ A 'cash sale' involves the seller receiving all cash at the closing regardless of whether the purchaser obtained financing, as long as the seller did not participate in providing funds. According to Lipscomb, seller involvement in the transaction requires examining the terms of financing with several examples listed below: 2 ^ - the buyer assumes an existing mortgage held on the property by the seller; the seller agrees to carry a first or second mortgage; - the seller pays points to acquire third-party financing for the buyer (U.S. case); and - the seller acquires an advance loan commitment for the benefit of the buyer. It is important to identify whether the financing terms are more favourable than those typically available in the market in order to separate the price paid for the real estate and the price paid for favourable financing. As DeLacy concedes, if a seller makes a financing concession, the property with the concession will sell sooner and at 19 a greater price than the property without such an alternative.26 The methodology to estimate the premium warranted for favorable financing, or conversely, the discount for unfavorable financing must be defined.27 The Market Approach to Value is an appraisal method commonly used to value single family homes. This approach involves the collection and analysis of data for similar properties which have recently sold in arm's length transactions. It assumes that the value of the subject property will be equal to the price paid for similar properties and is based on the Principle of Substitution. It is the most common, simple, and accurate method used, as it directly reflects market behavior and requires minimum subjective judgement of the appraiser.28 While identical properties (or interests in properties) do not exist in practice, it is possible to find properties which are considered similar by participants in the market. Boyce states that the reliability of the Market Approach depends on four f a c t o r s : ^ 1. the availability of comparable sales data; 2. the verification of the sales data; 3. the degree of comparability or extent of adjustment for time differences; and 4. the absence of nontypical conditions affecting the sale price. The growth of available financing packages in the residential sector has made it increasingly difficult to satisfy the fourth condition. Financing often causes a difference in price, as buyers are willing to pay a premium for houses having below-market financing. An adjustment should be made to the extent that financing on a comparable differs from the "normal market" financing arrangements on a particular type of property.^ u The appraisal literature reveals that the definitions of Market Value have been changed by appraisers to address contemporary issues as well as the possibilities of alternative applications to reflect various market phenomena.^* Garcia suggests that 20 the appraisal profession has not established how financing is to be included in the concept of Market Value, since some members advocate total disregard and others feel only cash sales should be used.32 He notes the brief reference made to financing by the industry in analyzing sales data: " . . . it is desirable to know whether the price paid was cash or a partial cash payment with balance secured by a mortgage. Unless otherwise defined, the value stated in an appraisal report usually means a cash payment or cash equity plus ordinary first mortgage financing. It should be determined whether or not the sale was financed by easy payment arrangements, a purchase money first mortgage, or a second mortgage. Such arrangements usually indicate a price greater than the cash price would have been p a i d . " 3 3 Friedman notes in an active market with the majority of sales involving conventional first mortgages, the appraised property value could be based on the cash market value of the first mortgage plus cash to the seller at closing. Kinnard feels that the typical pattern of financing available to the most likely type of investor should be considered in estimating the market value of the fee simple estate in the property. The issue then becomes whether creative financing, such as seller financing, can be considered as the typical current pattern of financing for a type of property in a local market. According to accepted appraisal terminology, the Market Value of a property is defined as: "the most probable price in terms of money that a property should bring in a competitve and open market under all conditions requisite to a fair sale, the buyer and seller each acting prudently, knowledgeably and assuming the price is not affected by undue s t i m u l u s . " ™ This definition presumes the completion of a sale on a specific date and the transfer of title from seller to buyer under the following conditions: 1. buyer and seller are typically motivated; 2. both parties are well informed or well advised and acting in what they consider their own best interests; 3. a reasonable time is allowed for exposure in the open market; k. payment is made in cash or its equivalent; 21 5. financing, if any, is on terms generally available in the community at the specified date and typical for the property type in its locale; and 6. the price represents a normal consideration for the property sold, unaffected by special financing amounts and/or terms, services, fees, costs, or credits incurred in the transaction. Explicit in the definition and concept of Market Value is the assumption that the price is a normal consideration for the property sold, unaffected by special financing amounts, terms and costs in the t ransact ion .^ The criteria listed in the preceding definition makes three references to the price paid in the transaction: namely, the payment is made in cash or its equivalent (No. 4); the available financing is typical for the property in its location (No. 5); and the price is unaffected by special financing amounts (No. 6). Accordingly, sales of comparable properties with different financing packages must be adjusted to a common unit of comparison and cash-equivalency has been cited as the appropriate m e t h o d . ^ The cash-equivalent price represents the sum of two components: - the cash consideration paid (down payment); and - the market value of all non-cash portions of the nominal sales price at the effective date of sale. The cash-equivalent price of a property may differ from its nominal price, and it should reflect the present worth at the sale date of all cash and other considerations paid for the real property as opposed to other portions of stated consideration paid for fees, services and/or other non-realty items. The Market Value of a mortgage is an estimate of the amount that might be received if the loan was sold in an arm's length transaction in the market under current conditions. A potential purchaser of the mortgage would pay to receive the remaining payments over the life of mortgage and 22 the outstanding balance at the end of the term, discounted at the prevailing market interest rate for that particular loan: Market Value of Mortgage: PV = Payments a ^ i + Outstanding Balance (l + i ) - n Where n = number of remaining periods in term i = market interest rate Payments = calculated at contract interest rate Outstanding Balance = calculated at contract interest rate The relationship between the Market Value and Book Value 3 ? of a mortgage may be expressed as follows: TABLE 2.2 RELATIONSHIP OF MORTGAGE VALUES If C O N T R A C T INTEREST R A T E = M A R K E T R A T E then M A R K E T V A L U E = BOOK V A L U E (The mortgage is said to sell at par) If C O N T R A C T R A T E < M A R K E T R A T E then M A R K E T V A L U E < BOOK V A L U E (The mortgage is said to sell at a discount) If C O N T R A C T R A T E > M A R K E T R A T E then M A R K E T V A L U E > BOOK V A L U E (The mortgage is said to sell at a premium) There is a direct relationship between contract and market interest rates and the market and book values of a mortgage. For instance, an investor buying a mortgage written at a lower than market rate will pay less for the loan than its book value to increase the return. The f ifth condition of the definition stipulates that the specifics of the market must be considered in each transaction. If the prevailing market reflects cash transfers of property, then the Market Value concept means that estimates of value should reflect a cash transaction. If the standard transaction involves creative financing with particular terms, then those terms should be reflected in the estimate of market value. Albritton points out that appraisers should interpret the market and use terms of the typical transaction to estimate Market Value. While cash-23 equivalency is not consistent or reflective of all markets and conditions, several applications have been noted: 3 ^ - to 'equalize* sales transactions involving diverse financing terms for data analysis purposes; to estimate cash market value; - to advise vendors/purchasers in structuring a range for the justifiable price depending on the terms; and - to estimate cash acquisition prices when sales of typical properties in the market are highly levered with seller financing. Despite the distinction, the uses appear similar with the goal of converting sales to a common standard for comparison purposes. The problem not yet addressed is whether the basis of cash-equivalency, or the methodology it uses, can be universally applied to adjust any financing package to a cash-equivalent price. Furthermore, it is very difficult to reach a consensus of the type of terms available or typical in a real estate market in order to analyze the non-cash components in terms of typical market transactions. 2A Summary The purpose of this chapter was to describe the main characteristics of creative financing arrangements and introduce the valuation problem created by these designs. With the growth of non-conventional financing in the sales of single family homes, comparable sales data must be adjusted to a common unit of comparison to eliminate financing differences. Cash-equivalency has been cited as the appropriate technique for adjustment and it is based purely on a present value calculation of the financing; the cash-equivalent price is the sum of the cash paid plus the market value of any mortgage(s). The difference between the nominal (unadjusted) sales price of a property and its cash-equivalent price is the premium paid for favorable financing (benefit) or the discount paid for unfavorable financing (cost). It is not clear, however, if these benefits or costs of non-market financing are entirely capitalized in the price 24 or passed on to buyers. While the actual amount of the adjustment can be calculated, the price paid for the property may not include the total amount of the premium or discount for reasons addressed in Chapter 3.0. As noted by Corgel & Goebel, most of the cash-equivalency literature centres on how it should be applied without questioning the relevancy of the method. 25 F O O T N O T E S 1 G.W. Gau, An Examination of Alternatives to the Rollover Mortgage, Paper prepared for Canada Mortgage and Housing Corporation, (June 1981), p. 1. 2 Ibid, p. 2. 3 Ibid, p. 5 4 R.J . Curcio and J .R. Webb, "Creative Finance and Affordable Housing," Real  Estate Review, 11 (September 1981), p. 79. 5 K.W. Colton, et al , "Borrower Attitudes Toward Alternative Mortgage Instruments," A R E U E A Journal, 7 (1979), p. 582. 6 Ibid, p. 583. 7 Ibid. 8 P. Treadway, "Housing Finance in the '80»s," Real Estate Today, 14 (July 1981), p. 15. 9 M.A. Maes, "The Emergence of Cash Equivalency in Valuation," Real Estate  Review, 12 (Fall 1982), p. 87. 10 H. Cassidy and A . Field, "The Pricing of A Mi's," Alternative Mortgage  Instruments Research Study, 111 (November 1977), p. 1. 11 Colton, p. 603. 12 W. Baldwin, "Where Will the Money Come From?" Forbes, (September 14, 1981), p. 154. 13 F .E . Case, "Creative Financing Instruments," The Real Estate Appaiser and  Analyst, (Spring 1982), p. 46. 14 D.L. Koch, et a l , "The Risks of Creative Financing," Economic Review, Federal  Reserve Bank of Atlanta, LXVII (December 1982), p. 11. 15 A . L . Schwartz, "Influences of Seller Financing Upon Residential Property Sales Prices," The Real Estate Appraiser and Analyst, 48 (Winter 1982), p. 36. 16 J .P. Friedman and J .B. Lindeman, "Seller Financing and Cash Equivalence," Real  Estate Appraiser and Analyst, (May - June 1979), p. 46. 17 See Rosen, pp. 1, 5; Salkin and Durning, p. 13. 18 A . F . Eger, Financing the Market for Existing Housing; An Alternative Source of  Funds, U B C Thesis, (September 1976), p. 31. 19 Case, p. 51. 26 20 Gau, p. 9. 21 Schwartz, p. 36. 22 Friedman and Lindeman, p. 46. 23 Koch, et a l , p. 12. 24 J.B. Lipscomb, "Discount Rates for Cash Equivalent Analysis," The Appraisal  Journal (January 1981), p. 24. 25 Ibid, p. 25. 26 P.B. DeLacy, "Cash Equivalency in Residential Appraising," The Appraisal  Journal, (January 1983), p. 82. 27 It is possible that a discount could be paid in a transaction if a mortgage was given or assumed at an interest rate above market levels. 28 S.W. Hamilton, et al , Real Estate Finance in a Canadian Context, September 1984, p. 17-8. 29 B.N. Boyce, Real Estate Appraisal Terminology, American Institute of Real Estate Appraisers, p. 136. 30 See "Below Market Financing," p. 4; Friedman, p. 416. 31 H.D- Albritton, "A Critique of the Prevailing Definition of Market Value," The  Appraisal Journal, (April 1980), p. 199. 32 K. Garcia, "Sales Prices and Cash Equivalents," The Appraisal Journal, (January 1972), p. 9. 33 Ibid. 34 DeLacy, p. 83. 35 Maes, p. 88. 36 Maes describes the cash-equivalent price as "a price in terms of cash as distinguished from a price which is expressed all or partially in terms of the face amount of notes or other securities which cannot be sold at their face amount". 37 Book Value is the original face value of the loan less the amount of principal repaid, or the amount outstanding at a particular point in t ime. At origination, the book value and face value are the same. 38 Albritton, p. 204. 27 3.0 REVIEW O F CAPITALIZATION E F F E C T S 3.1 Introduction The need to adjust sales data for the effects of financing has been discussed throughout the appraisal literature, however, there does not appear to be a consensus on the appropriate adjustment process. Recent studies on the influence of financing on house prices generally focus on two topics: 1. the ability of the Cash-Equivalent Method to accurately adjust creatively financed sales to a fair market price (see Schwartz (1982), Corgel & Goebel (1983), Findlay & Fischer (1983)); and 2. the extent to which creative financing is capitalized in the nominal sales price (see Zerbst & Bruggeman (1977, 1979), Colwell (et al) (1979), Gunterman (1979), Rosen (1982), Sirmans (et al) (1983), Bible (1980)). In the former case, it is hypothesized that the prices of conventionally financed homes should be the same as the cash-equivalent prices on 'identical' non-conventionally financed homes. These studies implicitly assume that any creative financing benefits are fully capitalized in the price and the adjustment for financing is strictly a present value calculation. Research in the latter group is largely centered on loan assumptions and FHA/VA transactions in the U.S. (involving discount points charged by the seller) and suggests that creative financing is 32.2% to 104.6% capitalized in the price of single family homes. Consequently, the Cash-Equivalent Method is not considered to be the appropriate adjustment process in converting all creatively financed sales to a cash-equivalent value. A brief summary of the empirical literature in these two areas follows. 3.2 Empirical Studies 3.2.1 Cash-Equivalent Method Schwartz (1982) examined the influence of seller financing on property sales prices in nominal and cash-equivalent terms, finding that seller-financed properties have higher sales prices on a nominal (unadjusted) basis and sell for less than their 28 nonseller financed counterparts when adjusted to a cash-equivalent value. The analysis involved comparing pairs of MLS condominium sales in Hawaii during the first nine months of 1981, with each pair consisting of an Agreement-of-Sale* financed unit and a cash sale or conventionally financed unit. To qualify as a matched pair, the units had to be similar in size, view, and number of bedrooms and baths, and any sales involving loan assumption or FHA/VA terms were excluded for comparability reasons. A cash-equivalent value was calculated for the seller-financed sales by discounting each Agreement-of-Sale to the present value by the conventional mortgage rate at the time of sale, and this discounted value was then added to the downpayment. While the creatively financed units sold for higher nominal prices, on average, than the conventionally financed units, their cash-equivalent value was lower. Schwartz concluded that seller-financed properties do not achieve higher values in real terms using the cash-equivalent method. Although he notes that the discounting process may overadjust for the real adjustment made in the market, his methodolgy assumes that a fixed, first mortgage loan rate can be applied to the Agreements-of-Sale. In fact , these contracts may differ significantly in the terms when compared to conventional first mortgage loans due to the greater flexibility and negotiation between the borrower and lender (seller). The use of the cash-equivalent method assumes that the benefits from creative financing arrangements are capitalized 100% in the price, which has not been shown in this study. Accordingly, it is difficult to draw any specific conclusions from the Schwartz research. Corgel & Goebel (1983) studied the ability of cash-equivalency to adjust for the effects of favorable financing by regressing the prices of conventionally financed homes on the cash-equivalent prices of comparable homes which sold with creative financing arrangements. Their data consisted of 38 pairs of MLS sales in Florida between January and June 1981 and 30 pairs in Texas selling between March 1978 and January 1981. A simple linear regression of the conventionally financed price on the non-conventionally financed comparable implied that cash-equivalent analysis 29 overstated the adjustment for favorable financing. If the properties could be matched to produce a similar pair, the prices of conventionally financed homes and the cash-equivalent prices on non-conventionally financed homes were expected to be identical. However, this presumes that the entire premium calculated for favorable financing is paid by the buyer. The Florida sample revealed that a $1 increase in sales price resulted in a $.95 to $.99 increase in the cash-equivalent price, depending on how long the property was assumed to be held before it resold; the Texas sample showed $.98 -$1.00 increase for each $1 increase in sale price, again, based upon the duration of the holding period. The results indicated that buyers (borrowers) did not pay the full amount of the premium for favorable financing, as the cash-equivalent price was less than its conventionally financed counterpart. The bias of cash-equivalency to overstate the adjustment required for favorable financing appears smallest when the shortest holding periods were used. This sensitivity to the holding period used should have less impact on a Canadian sample due to the considerably shorter loan maturities. They concluded that cash-equivalency may not accurately capture the effects of creative financing since it overstated the adjustment in virtually all of the holding scenarios. These two studies are significant as they suggest that the Cash-Equivalent Method may not yield accurate results in adjusting for creative financing. Corgel & Goebel point out that, while cash-equivalency may be an appropriate technique in adjusting for financing effects, it is crit ical to consider what assumptions are valid in applying cash-equivalency to get an accurate reflection of market conditions. 2 A holding period limited to 3 - 5 years for the creative funds is recommended since it reflects the typical length of short term mortgage contracts and probable holding period.^ The Corgel & Goebel study supports this case as their empirical evidence indicates that cash-equivalency significantly overestimates the adjustment for favorable financing with the exception of the short (three year) holding period scenario. The studies emphasize the sensitivity of the results to the underlying 30 assumptions used in adjusting for creative financing which has been considered in the sampling and analytical work of this thesis. Furthermore, the term of the Canadian mortgage loans in this thesis sample was known so there was no need to make assumptions regarding the anticipated holding period of the creative funds. Findlay & Fischer (1983) presented a theoretical assessment of the effects of creative financing on residential sale prices, finding cash-equivalency to be an invalid approach to valuation. They argue that cash-equivalency provides an excessive adjustment for financing due to the use of a fixed, first mortgage rate to discount mortgage cash flows that generally have shorter maturities and lower loan-to-value ratios than fixed rate, first mortgage loans. In other words, the selected rate is too low to calculate the present value of the payment savings and the estimated savings amount is, therefore, also too low. They developed a Financial Fee Valuation Adjustment model which estimates the present value of alternative financing based upon a given level of monthly payments afforded by the buyer. The F F V A adjustments are lower than cash-equivalency, and no adjustment is made for small first mortgages as they felt these loans would not be assumed. However, in this thesis, a number of small first loans had been assumed and these properties were not excluded from the sample. While is has been suggested that discounting long term financing should be restricted to 3 - 5 years to reflect likely holding periods or the length of standard short term mortgages, Findlay & Fischer use a 15 year period in their research. They did not indicate how they selected the holding periods or future financing rates used in their analysis. Unrealistic holding period assumptions in cash-equivalency analysis could be an explanation for overstating the adjustment indicated by the market and it reduces the validity of the Findlay <5c Fischer study. Based upon the foregoing discussion, it appears that cash-equivalency may not be the appropriate technique to use in adjusting comparable sales data. The main issue raised is the extent to which the benefits from favorable or unfavorable financing are capitalized or realized in the sales price. If it can be shown that creative financing is 31 fully capitalized in the price paid for homes, then the Cash-Equivalent Method is considered useful. If, on the other hand, the effects of creative financing are not fully capitalized, cash-equivalency will inaccurately estimate the adjustment. 3.2.2 Capitalization of Creative Financing In examining the capitalization impact of creative financing on house prices, there have been a number of studies on the shifting of FHA/VA mortgage discount points from sellers to buyers as well as the capitalization of favorable financing using samples of vendor-financed and/or assumed loans. While the studies use samples of either mixed creative financing schemes (i.e. assumed and seller-financed loans) or are restricted to a particular type of instrument (i.e. loan assumptions), none of them compare the results from a mixed sample against a sample using only one instrument. In addition, none of the research is based exclusively on vendor-financed loans. Three items are hypothesized in this thesis regarding the impact of the type(s) of creative financing schemes used in empirical analysis: - a sample of mixed creative financing instruments (i.e. loan assumptions and vendor-financed loans) will distort the capitalization results; - a sample of loan assumptions will show capitalization effects of approximately unity; and - a sample of vendor-financed loans will not be consistent with full capitalization. It might be argued that the different forms of creative financing can be classified as either 'institutional' or 'non-institutional', with the former group including mortgage assumptions and builder buydowns and the latter group involving vendor-financed loans. The differentiating characteristic of these groups is that the underlying financing in the institutional category involves a mortgage originated by a financial intermediary. To calculate the present value of the payment savings, institutional creative loans should be discounted by the conventional mortgage rate at the time of sale on loans having similar terms to the creative instrument. The loan was originally granted on the basis of institutional lending criteria which carefully evaluates the 32 borrower's ability and willingness to repay the funds borrowed and the value of the property as security for the loan. Institutional mortgage lenders are noted for providing the most complete and formal borrower qualification compared to private lenders.^ The lending institution is an independent third party which has economies of scale in loan initiation and administration. The average cost of originating and administering a loan decreases as the volume of applications and loans increases, so the actual dollar expenditure per loan is usually considerably lower for large mortgage lenders than it is for private lenders (vendors). Institutions with vast mortgage holdings are also able to reinvest the cumulative monthly payments received at a considerably higher rate than a vendor holding a small portfolio. A vendor may be forced to reinvest the mortgage receipts in a savings account, for instance, which bears a much lower rate of interest than that available on new mortgage loans; this situation is commonly referred to as the re-investment rate problem. The non-institutional category of creative financing comprises loans which are originated outside of normal lending channels, and the economies of scale in loan origination, servicing and risk analysis are not available to the vendor (lender). The lender in this case is involved in two legal relationships with the purchaser, namely, negotiating the transaction price and providing the purchaser with funds to help finance the sale. The desire to sell the property may result in the vendor giving a loan based on inadequate security analysis or poor underwriting criteria without consideration of the enduring relationship created. Generally, credit analysis for proposed private loans should involve greater care and scrutiny than for institutional lenders due to the higher level of risk often associated with private lending. Private lenders may be approached because institutional financing cannot be obtained, and they do not have the same ability to spread their risk of loss over a large number of mortgage loans and other investments. Institutional and private lenders also differ as the maximum uninsured loan-to-value ratio is set by statute at 75% for federally 33 TABLE 3.1 CALCULATION OF VENDOR FINANCING BENEFIT TO BORROWER Example: Vendor provides a loan of $80,000 amortized over 15 years at a rate of 10% per annum compounded semi-annually. Market interest rate is 14% per annum compounded semi-annually for comparable loan. Assume monthly payments. Loan = $80,000 isa = 10%, i m o = .8164845 Payments = $849.82 PV Calculation ® 14%: isa = 1^%, imo = 1-134026 PV = $849.82 a 18(fll. 134026 = $65,093.89 Benefit = $80,000 - $65,093.89 = $14,906.11 PV Calculation Q 16.5% (to account for vendor's higher loan origination/servicing costs): i s a = 16.5%, i m o = 1.3299859 PV = $849.92 a Hoi 1.3299859 = $57,972.54 Benefit = $80,000 - $57,972.54 = $22,027.46 34 chartered (Canadian) financial institutions and there is no maximum specified for private lenders. Accordingly, it appears that the discount rate should be higher for non-institutional creative loans than those in the institutional group. The use of too low a discount rate will cause an understatement of the financing premium which could result in an overcapitalization in house prices of the incorrectly calculated vendor financing premium. A numerical example is presented in Table 3.1. It should be noted that the creative financing benefit could also be calculated by discounting the difference between the monthly payments for equivalent standard financing at the market rate and for creative financing at a below-market rate. In the U.S., lenders provide Federal Housing Administration (FHA) insured and Veteran Adminstration (VA) guaranteed loans which have regulated contract rates and restricted loan origination fees. Due to the ceiling on interest rates and the difference between loan origination fees and costs, lenders charge discount points to achieve the same yields on FHA/VA mortgages as on conventional mortgages with similar loan terms, risk, and servicing costs. F H A regulations prohibit lenders from charging borrowers the points, which means sellers pay the discount. Zerbst & Bruggeman (1977) point out that sellers respond by shifting the discount points to FHA/VA purchasers with higher property prices since the points are a part of the financing costs incurred by borrowers. Table 3.2 summarizes the capitalization studies discussed in the remainder of this chapter. Zerbst & Bruggeman (1977) studied the extent to which sellers pass on discount points in FHA/VA transactions using a sample of 276 MLS sales in June 1973 from Columbus, Ohio. They proposed that if sellers successfully shifted discount points over a large number of transactions, FHA/VA buyers should systematically pay a higher percentage of asking prices when compared to the percentage of asking prices paid by conventional buyers. Their sample comprised 14.7% F H A , 22.1% VA , and 63.2% conventional sales. The ratio of sales price to asking price was regressed on dummy 35 TABLE 3.2 SUMMARY OF CAPITALIZATION STUDIES Research Data Base (Loan Type) Study Period Dependent Variable Significant Independent Variables Equation R2 Estimated Capitalization Comments Zerbst & MSL Sales 3une 1973 Sale Price t FHA/VA dummies; Bruggeman FHA/VA Loans N = 276 Asking time to sell. (1977) Price 13% (FHA) 56% (VA) Colwell et al (1979) FHA Loans Jan. 1970-Dec. 1975 N = 2,1(08 Sale Price (Log) Dwelling age & size; property frontage; no. bathrooms; air-conditioning & fireplace dummies. .892 77% (100% Not Rejected) FHA dummy not signficicant and wrong sign; highly correlated with points variable. Gunterman (1979) FHA Loans Jan. 1970-Dec. 1975 N = 2,108 Equation by Price; $5,000 - $11,999 $15,000 - $21,999 $25,000 + Sale Price FHA dummy; dwelling age 4; size; time to sell. .83 N = 811 (Similar Equation to Main Model) .51 N = 1,098 N = 169 .19 .11 73% 12% 70% FHA dummy not significant. FHA dummy significant. FHA dummy significant. Rosen ( 1982) Assumed and Vendor Financed Loans June 1981 Sale Price N = unknown Savings dummy; no. of bathrooms; dwelling age & size; location. .683 101.6% Mixed sample of vendor financed and assumed loans. Gunterman (1982) Assumed Loans (FHA) Nov. 1975-June 1978 N = 367 Sale Price (Log) Discount/predicted price; dwelling age & size; time to sell; air-conditioning; no. of bathrooms; property frontage. .818 67% Savings variable (discount/predicted price) highly related to sales price. Sirmans et al (1983) Listing Service Sales Primarily Assumed Loans July 1980-Dec. 1980 N = 108 Sale Price Discount variable; no. of bathrooms; air-conditioning and fireplace dummies; basement; age. .60 32.2% Mixed sample. Vendor loans as junior financing. Bible (1980) MLS Sales Nov. 1978-Assumed Loans Feb. 1979 N = 280 Sale Price Dwelling size & age; neighbourhood quality; dwelling construction quality. .85 Equation by Price; $15,000 - $11,910 N=130 (Similar Equation to Main Model) .73 $11,911 - $105,000 N = 150 .80 Assumption variable not significant. Assumption variable significant. Assumption variable not significant. * Dummy variable for assumption loans indicated it was worth $969.10 to seller. 36 variables representing whether a transaction was F H A , VA , or conventionally financed, the loan-to-value ratio and the number of days required to sell the property. The results indicated that discount points were passed on to FHA/VA buyers, as they systematically paid a higher proportion of asking prices than the sample of conventional buyers. In fact, it was estimated that 43% of the discount points were shifted to F H A buyers and 56% were passed on to VA buyers. This was the first empirical work to establish that sellers shift discount points to buyers, although complete shifting did not occur. Colwell et al (1979) crit icized the Zerbst & Bruggeman model for not treating the asking price of the house as a function of the probability of a non-conventional loan. They argued that this probability would vary between neighbourhoods and that sellers might refuse to sell at the asking price with F H A financing if they anticipate a higher price (net of points) with a conventionally financed sale. "Recognition that some financing outcome can be reasonably anticipated by the seller gives rise to the possibility that sellers adjust their asking prices to reflect anticipated f inancing." 5 The failure of Zerbst & Bruggeman to incorporate the expected F H A or VA points in the asking price was cited as the reason for the low estimate of shifting. Colwell (et al) specified a model with sales price as the dependent variable and a financing dummy variable (FHA or conventional financing) and other housing characteristics as independent variables. The sample consisted of 2,408 house sales in Lubbock, Texas which occurred from January 1970 to December 1975, divided equally between F H A and conventionally financed transactions. The model was transformed by taking the natural logarithms of both sides of the equation and then estimating the coefficients by ordinary least squares. Approximately 77% of the points were shifted to buyers in the form of higher selling prices, suggesting a substantially stronger relationship than the Zerbst & Bruggeman model. In addition, a statistical test of whether the shifting, indicated by an estimate of the model, differed significantly from unity (i.e., 100% of the points shifted) revealed that the hypothesis of all points 37 being shifted could not be rejected. However, the F H A financing variable was not significant and did not have the expected sign, probably due to the high correlation with the points variable. Zerbst & Bruggeman (1979) replied to the criticism of their model, stating that the shifting estimate could be inaccurate if sellers systematically included the expected effects of FHA/VA financing in the asking price. Asking prices should be materially higher on FHA/VA financed units than comparables conventionally financed. In their sample, the sale prices equalled or exceeded asking prices in roughly 60% of the FHA/VA transactions, while only 24% of conventional transactions had equal sales and asking prices (sales price never exceeded the asking price). This suggested in a large number of FHA/VA transactions, sellers had not incorporated the effects of expected financing in asking prices or had underestimated the impact on asking prices. Zerbst & Bruggeman also stressed that their results were not intended as a precise measure of shifting which would remain constant over time or within different urban housing markets. The Colwell research is critized for using two highly correlated variables in their model (the points and F H A dummy variables) as well as misspecifying it by omitting variables (i.e. neighbourhood influences). The allocation of F H A discount points between buyers and sellers was studied in greater detail by Gunterman (1979) who addressed two new issues: 1. the relationship between the proportion of points shifted and the level of house prices; and 2 the relationship between the proportion of points shifted and the number of discount points charged. Using the same data as Colwell (et al), Gunterman regressed the sales price against house size, a time variable, and a dummy variable for F H A financing. The results showed that purchasers using F H A financing paid approximately $475 more, on average, for a house than if they had used conventional financing. This increase reflected a 73% shift of the F H A discount points from sellers to buyers, given the average discount of $654 paid by the seller over the study period. The results 38 generally confirmed the Zerbst & Bruggeman findings, but indicated that a larger share of the discount is paid by house purchasers. A refinement of the model was done by dividing the data into three levels of house prices, estimating a regression equation for each group. While the F H A variable was not significant for the lowest priced housing, it was for the two higher priced categories. The middle priced group showed 42% of the discount points being pased on to buyers, and over 70% of the points were shifted in the highest priced group. They concluded that sellers are increasingly aware of the cost of mortgage discount points as house prices increase and they respond by passing on an increasing porportion of the costs to borrowers. The discount in the lowest priced category would be small, particularly relative to transaction costs such as real estate commissions and other closing expenses, and it is possible the sellers would be less sensitive to this discount. Gunterman also estimated regression equations for each price category when two, three and five points were charged. The F H A variable was not statistically significant when two points were charged or in any of the lowest priced scenarios. Purchasers generally paid a larger proportion of the points as the number of points increased, which was attributed to uncertainty about the discount points at closing and market conditions at the time of sale. Gunterman hypothesized that discount points would not be fully shifted when they had increased from the previous quarters and would be overshifted in the case of decline. Therefore, the model was re-estimated for quarters when discount points increased and decreased from the previous quarter, and discovered that sellers would fully shift F H A points if they could be perfectly estimated. Gunterman's work supported the earlier findings of Zerbst & Bruggeman that F H A mortgage discount points are shifted from sellers to buyers through increased sale prices, establishing that the relationship is much more complex. The extent to which below-market rate financing is capitalized in higher housing prices was studied by Rosen (1982) using sales from Northern California during Dune 1981. The transactions comprised 66.6% with assumable first mortgages, 61% with 39 seller second mortgages and 52.6% with both instruments. Sales price was the dependent variable which was regressed on the downpayment proportion, a variable representing the present value of the cash payment savings to the buyer from favorable financing, as well as a number of housing characteristics. The coefficient on the savings variable would reflect the extent of capitalization, with 1.0 reflecting full capitalization. Since his equation produced a significant coefficient of approximately 1.0, he concluded that the cash-equivalent method was a reasonably good estimation of the adjustment required for creative financing. Details were not available on the size of the sample used, and his final model produced an R of .6828 which meant that the equation cannot explain about 32% of the change in price. Rosen's mixture of institutional and non-institutional creative financing is considered to cloud the analysis and, with the incorrect derivation of the vendor-financing benefits, it is not clear whether the results support the full capitalization hypothesis. While it is generally accepted in the literature that favorable financing results in higher sale prices, Gunterman (1982) suggested that houses with assumption financing might have lower selling prices than if they had new loans; in other words, the benefit from creative financing would be viewed as a 'discount' rather than a 'premium'. Since assumption houses tend to have lower loan-to-value ratios, he argued that sellers with large equities relative to the selling price might be able to negotiate better purchase terms. Gunterman stated: "The reason for the discount, instead of a premium, is because of the increase in interest rates which makes the assumption attractive. A t the higher current rate of interest, the present value of the remaining payments is less than the actual loan balance assumed by the purchaser. Finance theory would predict that the purchaser would attempt to discount the mortgage, and realize this out of the sellers equity." This reasoning is incorrect because the mortgage is a liability, not an asset, once incurred by the homebuyer. A mortgage with a below-market interest rate would be traded as an asset in the secondary market at a discount from the face value depending upon the relationship between market and contract rates.^ 40 The purpose of Gunterman's article was twofold: 1. to determine the extent of capitalization of the presumed 'discount' into a lower selling price with assumption financing; and 2. to evaluate the effect of the purchaser's equity on the sales price. The sale prices of the properties were regressed against the ratio of the discount to the predicted sales price of the property/ the loan-to-value ratio and a number of housing variables. The model was transformed by taking the natural logarithms of both sides and estimating the coefficients with ordinary least squares. A l l of the independent variables, excluding the loan-to-value ratio, were statistically significant. When the equation was re-estimated without this variable, the coefficient of the discount variable was - .672, suggesting that roughly 67% of the discount was capitalized into a lower selling price for assumption houses. Sirmans noted that the significant negative coefficient of the discount variable resulted from misspecification of this model, as there was a negative correlation between the sales price (the dependent variable) and reciprocal of the predicted price; the discount variable was not considered to offer any positive explanatory power. Sirmans et al (1983) also looked at the impact of assumption financing on sale prices of single family homes, finding that properties financed with assumption loans have higher selling prices than those conventionally financed. Using a random sample of 108 sales in the Atlanta area during the second half of 1980, they developed a model where the sale price was regressed on a number of standard explanatory housing price variables, the loan-to-value ratio and a discount variable. Four different equations were specified by manipulating the loan-to-value ratio and discount variable due to multi-collinearity, and the results indicated that 28.2% to 35.9% of the discount was passed on to the buyers through higher selling prices. However, similar to the Rosen study, vendor-financed loans were also included in their sample which were used only in a junior financing role. The use of a mixed sample is considered to cloud the capitalization results. 41 Bible (1980) took a sample of 280 homes which sold in Tallahassee, Florida and regressed the sales price on several housing characteristics as well as a dummy variable representing financing terms. The loan assumption variable was not significant in the estimated equation and it did not support the hypothesis that loan assumptions had a considerable impact on sale prices. He then refined the data on loan assumptions by dividing this sample into two price ranges, finding that the loan assumption variable was significant in the lower price category. The smaller equity (downpayment) required for lower priced homes may have made these loan assumptions more desirable. Bible's findings are important for two reasons: 1. they support Gunterman's research that there are different housing submarkets requiring different adjustments for the effects of financing; and 2. the presence of a loan assumption requires a dollar adjustment to estimate an adjusted sales price within a specified housing submarket. With the exception of Gunterman's (1982) research, the literature reveals that creatively financed homes sell at a premium to comparable properties financed with conventional mortgages. According to the previous discussion, creative financing is capitalized anywhere from 32.2% to 104.6% depending on the data set used. It is significant that several authors have sought refinements of their models to study different submarkets within the samples (see Gunterman (1979), Bible (1980)) or have restricted their analysis to a particular type of creative financing arrangement (Colwell et al (1979), Bible (1980)). With the exception of Rosen and Sirmans (et al) who used vendor-financed and assumed loans in their samples, all of the studies have utilized loan assumption data and have not dealt specifically with vendor-financed loans. This thesis will consider the capitalization effect of each loan type as well as in a combined sample. These further specifications of the price equation help to provide a clearer interpretation of the impact of creative financing instruments on house prices. 42 3.3 Factors Affect ing the Capitalization of Creative Financing While the literature indicates that creative financing is capitalized to some extent in the price of housing, the capitalization effect does not appear to be unity. Rosen noted that it would be inappropriate to assume that a direct adjustment, based upon the discounted value of the savings from the financing package, to the sales price would be made due to factors influencing the creative finance price premium: - a larger downpayment may be required from the buyer as a significant portion of the old low-rate loan may have been paid off, thereby lowering the creative finance price premium. - a creative package with a balloon second due in 1 - 3 years might be regarded as risky to the buyer and therefore command a smaller price premium, despite a lower interest rate. - since mortgage interest payments in the U.S. are tax deductible, taxpayers in high marginal tax brackets would not value the savings as highly as in the case of no interest tax deduction. Rosen suggested that generally the lower the interest rate and longer the term of the creative financing package, the more valuable it becomes. This is less significant in the Canadian context, since mortgage interest is not tax deductible and the maximum loan term is typically l imited to five years. Zerbst & Bruggeman offered several explanations for the FHA/VA discount points not being fully shifted from sellers to buyers: there is a time interval between the dates when a) the contract sale price is established, b) the F H A insurance or VA guarantee is issued, c) the loan commitment is obtained by the buyer, and d) the loan is actually closed. the discount points have not been incorporated into the FHA/VA appraised value. - consumers purchasing durable goods do not accurately estimate financing charges associated with purchases. In the first case, unanticipated interest rate changes may occur during the interval, causing higher unanticipated discount points to be paid by sellers who inaccurately estimated them at the time the contract sale price was established. Although Zerbst & Bruggeman found a lower proportion of points passed on than 43 Colwell (et al) or Gunterman, their sample was taken during a period of rapidly rising interest rates. It was expected that buyers and sellers would have more difficulty in estimating points at future closing dates than in a more stable mortgage market.** The exclusion of discount points in appraised values implies that they would not be financed as part of the FHA/VA mortgage loans. Accordingly, borrowers would pay the points through a higher down payment as sale prices would exceed appraised values. In order to close a sale, sellers might not fully pass on the discount points to buyers. In the third example, reference is made to a study of automobile financing transactions in which 50% of the borrowers could not state the actual rate of interest paid within a 50% range of accuracy. Gunterman (1979) suggests it is also possible that discount points are a matter of negotiation between the purchaser and seller, similar to other terms of the transaction that are negotiated. Therefore, the full discount might not be passed on to the purchaser despite precise estimates. Sirmans (et al) list three factors which affect the price paid for the discount: 1. The Buyer's Marginal Tax Rate: Because interest payments are tax deductible in the U.S., the difference between the present value and face value of an assumable loan decreases as the purchaser's tax rate increases, and a lower price is paid for the discount as the marginal tax rate increases. 2. The Length of Time the Purchaser Intends to Hold the Mortgage: A lower price will be paid for the discount as the holding period declines and the variability in conventional interest rates increase; this is caused by increased variability in the expected value of the discount. 3. The Negotiating Skills of the Buyer and Seller: Creative financing benefits may be treated as any other term of the transaction to be negotiated and, therefore, not fully capitalized despite accurate calculations of the amount. The first two factors have less importance in this research due to the use of Canadian data with no interest deductibility for income taxation or long loan terms. The third item is important since many vendors and purchasers in the single-family market are not sophisticated in their understanding of finance details and would not be aware of the exact amount of the present value of the benefit from favorable financing. 44 3.4 Summary Although the literature indicates that creative financing is capitalized in the sales price of housing between 32.2% and 104.6%, the general concensus supports a less than unity ef fect . Despite the number of recent studies, there are few explanations of the capitalization impact, particularly when different submarkets have been analyzed. Consequently, the appropriate pricing mechanism of creative financing remains unclear. The foregoing research is significant to this thesis by establishing that the relationship between creative financing and the sale price of single family homes is much more complex than was indicated by early studies. Several questions will have to be addressed in this research to determine the pricing mechanisms of non-conventional financing techniques: 1. Do refinements in specifying the house characteristics, neighbourhood and financing variables improve the model of capitalization effects? 2. Is creative financing capitalized to the same degree at different levels of house prices? 3. Do market conditions such as rising and falling mortgage interest rates and house prices affect the extent to which creative financing is capitalized in the price of housing? 4. Are alternative creative financing techniques capitalized differently in the price of single family homes? 45 F O O T N O T E S 1 An Agreement-of-Sale contract involves a buyer who signs an installment contract to buy real property and a seller retains legal title to the property until the obligation is fully repaid with interest. Generally, there is little or no repayment of principal prior to maturity. (Schwartz, p. 36.) 2 Corgel & Goebel, p. 61. 3 DeLacy, p. 85. 4 Hamilton, et al , p. 7 - 20. 5 Colwell , et a l , p. 1049. 6 Sirmans, et a l , p. 2. 7 The predicted selling price variable for the assumption properties was estimated by a regression equation based on conventionally financed houses in the sample. 8 Their study was replicated during a period of more stable interest rates using data from Chicago and more shifting did occur. (See Bruggeman & Zerbst, "Discount Points and Housing Prices: A Reply," The Journal of Finance, XXX IV (September 1979), p. 1057). 46 4.0 EMPIRICAL M O D E L S O F HOUSE PRICES 4.1 Literature Review The hedonic price model is generally used to study house price determination, whereby the market price of a dwelling is regressed on an array of characteristics. 1 Hedonic residential property valuation models assume that buyer preferences for specific housing characteristics are directly related to the prices paid for houses. The principal conclusions of the hedonic relationship are outlined below: 2 the hedonic function is a joint envelope (in characteristic space) of demanders' bid curves and sellers' offer curves; - supply or demand for characteristics cannot be identified from the hedonic function since it is defined over the characteristic space and its arguments consist exclusively of characteristics; and - the partial derivatives of the function should be interpreted as the implicit marginal characteristic prices prevailing at a particular market equilibrium. Observed product prices and the specific amounts of characteristics associated with each good define a set of 'hedonic' or implicit prices which means we can measure the impact of each variable (or characteristic) on the price of a house. The predominant statistical method used in appraising property is least squares regression estimates of the dependent variable as a linear function of independent variables, where the parameters are estimated from recent sales information. 3 The dependent variable is the estimated selling price, and the independent variables include characteristics of the property (land and improvements) such as dwelling, neighbourhood, location, and financing features. Properties possess value to buyers and sellers based on observable and unobservable characteristics which are unique to each location and set of improvements. For linear least squares, the specified regression equation must have linear coefficients (no coefficient appears as a power in a product or quotient of another coefficient) but may be nonlinear in the property 47 characteristics. Thus, the reduced form function is linearized for m observable characteristics as:^ P[ = AQ + A I Z H + ... + A m Z i m + A m + i N i i = 1, n for properties; j = 1, m for characteristics; An is a constant; where Aj are the linear coefficients to be estimated from the data on property sales; Zjj are the specific characteristics or combination of characteristics for properties; and Nj is the random term (which is assumed to be zero on average). The resulting value prediction for the selling price of a property may be generalized if it is assumed that all properties' supply and demand functions possess an identical relationship to their particular set of characteristics (i.e. the specific form and parameters of the structural equation remain stationary for all property through time). Therefore, the relative importance of the value determinants is assumed to remain constant over the period of analysis. Sales data can be used to establish the expected or probable sales price for all properties classified by a specific group of characteristics.^ Church notes that the underlying supply and demand relationship of this technique may be likened to the conventional cost and market valuation techniques.^ The characteristics in the supply function reflect the materials and services and their prices which are required for the construction of a new residence or replacement (less depreciation) of an existing one. The supply function is similar to the Cost Approach to valuation and it summarizes the prices for which suppliers (builders and owners) are willing to sell. The Market (Sales Comparison) Approach estimates sale prices by matching, as closely as possible, the attributes of sold and unsold properties. The actual sales price of a 'nearly' identical parcel which has sold is used to estimate the selling price of an unsold property. 48 There are a number of problems cited with the hedonic function and, particularly, the linear form least squares regression equation. Straszheim comments that the central problem in estimating hedonic equations rests with the delineation of homogeneous submarkets.? He argues that the urban housing market is a set of compartmentalized and unique submarkets with demand and supply influences likely to result in different price structures within each; consumers have distinct preferences for housing characteristics which are directly associated to the capital stock, and certain neighbourhood characteristics have major effects on household bids at any given location. Straszheim uses the example of estimating the effects of race on housing prices, indicating that independent variables within the price equation may differ for different racial groups. Pooled data and a single dummy variable to represent racial background is not considered to account for potential differences between prices of comparable housing between different racial groups or in different markets. In estimating the effects of financing, the thesis will segregate the data as follows: - homogeneous submarkets based on house price levels; - submarkets of house sales based on particular periods of market activity (i.e. rising or falling house prices and interest rates); and - submarkets of house transactions based on the particular creative financing techniques used (i.e. loan assumptions or vendor-financed loans). Buyers at the lower price level may value certain characteristics related to the selling price in a different manner than those at higher levels. For instance, the literature has indicated that the benefits of creative financing may be of more value at the higher price levels due to the larger amount of financing involved.** During a period of rising prices and interest rates, mortgage funds may be rationed by institutions as other investments (such as bonds) compete for depositor's funds. Financing, which is a determinant of price, may be valued differently by the market than during a period of available funds. 49 The distinction between different types of creative financing was discussed in the previous chapter. Clauretie notes a problem generated in samples of vendor-financed homes since the amount of this loan is a function of the price of a house. The use of hedonic price models causes the price of a house to appear on both sides of the equation which causes a correlation of the disturbance term and explanatory variables. He suggests using the amount of equity as an explanatory variable with a reduced form equation as a potential solution to this problem. It should be noted that the biased estimate of the financing coefficient occurs if owner-financed houses represent a significant portion of properties in the sample; the vendor-financed sales in this thesis represent a small portion of the sample. Butler and Linneman point out that misspecification bias is introduced into the price equation when some of the relevant independent variables are omitted. In principle, all of the characteristics which are relevant to the determination of a market price (i.e., those that both yield utility to residents and are costly to produce) should be incorporated.9 in practice, all of the characteristics cannot be included due to the large number of traits involved and the unavailability or poor quality of data. Butler's empirical work on a fully specified and a severely restricted model to estimate market values suggests that the practical impact of biases from omitted variables is minimal. Linneman looked at the relative importance of classes of characteristics in the determination of property prices (namely neighbourhood and structural traits), finding that neighbourhood traits are major determinants of a site's value since they explain between 17% and 48% of the standardized variation of a site's value. 10 As a result, characteristics reflecting neighbourhood traits have been incorporated in the model. Multicollinearity among the independent variables has been identified as a problem in many housing studies due to the basic nature of the d a t a . H Intercorrelations among the explanatory variables may become large enough to make it difficult to separate the effect of each independent variable on the dependent 50 variable (sales price) and obtain an accurate estimate of their relative impacts. Mark indicates that structural characteristics such as the number of rooms, bedrooms and square feet of living space are inherently related as are neighbourhood traits such as median income, racial composition and school years of education. The condition is signalled when the simple correlation coefficients of the variables are large (i.e., larger than the R of the equation) or if the estimated standard errors of the regression coefficients are large in relation to the size of the coefficient itself. This problem may be reduced or avoided by limiting the number of variables in the equation, particularly those which are interrelated, or by using statistical methods such as component analysis to reduce the set of variables in the final price equat ion . 1 2 Another limitation imposed by linear regression is that the coefficients estimated from the regression equation are constant. Constant multipliers or coefficients are unlikely in many cases, since the value of the improvements does not increase at a fixed dollar amount per additional square foot of living area; in fact , the value per additional unit of space would likely decrease after a certain s i z e . 1 3 However, transformation of the variables (using logarithic, inverse and squared relationships) and the use of noncontinuous variables can improve the explanatory power of the estimated equation. Noncontinuous characteristics include items such as a swimming pool, air-conditioning and double glazed windows, and they are represented by dummy variables in the equation which take on a value of 1 if the factor is present and a value of 0 if not present. Also to be discussed here is the assumption that the probability distribution of the random error term (nj) remain constant for all properties. As Church comments, the variance of r\{ will likely increase with the market value of a property (since there is more latitude for bargaining), the type of property (residential or commercial) and possibly other factors. He suggests classifying properties into categories possessing constant variance for nj and then estimating separate equations for each category. The more homogeneous the properties are within a class, the more accurate the model 51 will be in determining the probable selling price. This is another reason for segregating the thesis sample into various submarkets. In conclusion, there are a number of issues to be considered in estimating the housing price equation to be used in this research. The variables to be included in the price equation will have to be carefully selected to reduce misspecification of the model and multicollinearity of the independent variables; also, the structural form of the equation will have to be considered. 4.2 Specified Housing Price Equation In selecting the specific variables of housing characteristics to be used in this analysis, a brief review of the hedonic residential property valuation models was completed. Lang & Jones note that much of the previous research in the area has focussed on two distinct types of variables: ^ (1) measures of neighbourhood amenities; and (2) measures of physical home characteristics (eg. size of house, number of rooms, baths, etc.). They cite a set of nine physical property characteristics which are typically used in housing price studies: whether or not the property had a basement; whether it had a fireplace; whether it had a carport or garage; whether it had central air conditioning; the size of the finished living area in the house; the size of the lot; - the number of bathrooms; the total number of rooms; and - the age of the structure. 52 These features were used in this model as a minimum, with the exception of central air-conditioning which is not a common dwelling characteristic to the study neighbourhoods selected. Measures of neighbourhood amenities vary considerably between the valuation models, however, most research in this area has relied almost exclusively on objective proxies rather than direct subjective measurement (i.e. visual attractiveness, relative preference and prest ige) . 1 5 Lang & Jones studied the use of proxy measures for neighbourhood amenities, using objective and subjective models, concluding that direct subjective measures yield only modest improvements in price predictions. Proxy measures simplify the data collection and are cost efficient for large scale valuation work; they have been used in this research to account for neighbourhood effects. While Grether & Mieszkowski provide an extensive list of dwelling characteristics in their m o d e l , 1 6 it appears that many of the studies limit the number of these variables used due to data collection and validation problems. It is postulated that the sales price of a house is an additive function of its structural characteristics, property characteristics, neighbourhood amenities and the financing arrangements negotiated in the transfer. Since several different geographical locations were used in this research, a common data source for each class of characteristics was desired. Information was compiled from records at branches of the Land Title Off ice (for financing details), the British Columbia Assessment Authority (for property characteristics) and Statistics Canada (for neighbourhood traits). These government sources survey the Lower Mainland area of the Province using similar data collection and recording procedures for offices in different municipalities. While the selection of variables was constrained to the information available from these sources, they provided an extensive data base which was not considered to limit this study. The entire set of variables used in the empirical analysis described in Chapter 6.0 is presented in the following chapter as Table 5.9. Several of them represent transformations of other variables to improve the accuracy of the equation. 53 For instance, P V S A L E and P V L A G are two similar variables, reflecting the present value of the benefits or costs of nonmarket financing; they only differ in the discount rate used. A brief description of the variables follows: D E P E N D E N T V A R I A B L E The sales price (SPRICE) for each house was taken from the transfer documents registered at the Land Title Off ices. INDEPENDENT VARIABLES Financing Characteristics: The amount of the sale price which was financed is reflected by the loan-to-value ratio (LNTOV) as well as the actual dollar value of all outstanding financing (LOAN). These two variables are expected to have positive signs in the estimated equations since buyers with larger equities are more able to negotiate better purchase prices. The present value of the benefit or cost for nonmarket financing was calculated for each property using either the market interest rate at the date of sale (PVSALE) or using an interest rate six weeks prior to the sale date (PVLAG). Discussions with local realtors and appraisers suggested that the average period of time between signing the Interim Agreement and closing the deal was roughly six weeks, so the appropriate discount rate for present value calculations would be at the initial point of negotiation. These variables are expected to have a positive sign and previous literature has suggested coefficients of 1.0. Several time dummy variables are used (YEAR1, Y E A R 2 , TIM EM) to sort the data for estimating annual price equations and for runs capturing different periods of rising and falling markets. They are also used to illustrate changes in market conditions since the study took place over a relatively volatile period. Dwelling Characteristics: The structural and lot characteristics for each house are expected to have positive signs in the price equation, with the exception of A G E which should show a negative sign. While most of the variables are self-explanatory, R E V E N U and O W N O C C are used to capture investment opportunities of the dwelling, as R E V E N U involves rental of the basement of the dwelling and O W N O C C shows rental of the entire dwelling. Neighbourhood Characteristics: The first four traits were included to capture the differences in neighbourhoods between the study areas. The ethnicity (ETHNIC) and density (FAMSZ) are expected to have a negative effect on price. The proportion of single family development (5FDU) and amount of homeownership as opposed to rental housing (HOMOWN) are expected to have positive effects on price. Alternatively, dummy variables were used for location to pick up differences in geographic situation such as desirability and prestige of an area. The expected signs for the North Shore neighbourhoods (LOCNS) and Vancouver (West Side) neighbourhoods (LOCVW) are positive. The Richmond neighbourhood (LOCRI) is expected to have either a positive or negative sign, and the Vancouver (East Side) area (LOCVE) should show a negative impact. 54 The following chapter details the data used and the characteristics of the sample. An explanation of the data base input is in Appendix F, and several additional variables to those previously noted are used to categorize the transactions for discussion purposes. The additional variables which were not used for empirical analysis have not been discussed in this chapter. 55 F O O T N O T E S 1 M.S. Johnson and M.J. Lea, "Differential Capitalization of Local Public Service Characteristics," Land Economics, 58 (May 1982), p. 190. 2 R.V. Butler, "The Specification of Hedonic Indexes for Urban Housing," Land  Economics, 58 (February 1982), p. 96. 3 A . M . Church, "An Econometric Model for Appraising," A R E U E A Journal, 3 (Spring 1975), p. 17. 4 Ibid, p. 23. 5 Ibid, p. 21. 6 Ibid. 7 M. Straszheim, "Hedonic Estimation of Housing Market Prices," The Review of  Economics and Statistics, LVI (1974), p. 404. 8 Refers to the research done by Gunterman (1979) where the price equation was estimated at different levels of housing prices. 9 Butler, p. 97. 10 P. Linneman, "Some Empirical Results on the Nature of the Hedonic Price Function for the Urban Housing Market," Journal of Urban Economics, 8 (1980), p. 56. 11 See Church, p. 28; Mark, p. 109; Curcio et al , p. 46; Butler, p. 97. 12 Church, p. 29. 13 Ibid, p. 24. 14 J .R. Lang and W.H. Jones, "Hedonic Property Valuation Models: Are Subjective Measures of Neighbourhood Amenities Needed?" A R E U E A Journal, 7 (Winter 1979), p. 452. 15 Ibid. 16 D .M. Grether and P. Mieszkowski, "Determinants of Real Estate Values," Journal of Urban Economics, 1 (1974), pp. 132, 133. 56 5.0 D A T A DESCRIPTION 5.1 Sampling Issues The data set used in this analysis was obtained from sampling single family home transactions in the Lower Mainland area of British Columbia during the 1980 - 1982 period. Figure 5.1 outlines the general areas sampled and maps showing each neighbourhood are presented in the Addenda as follows: Appendix A : Vancouver (East Side) Neighbourhood Location Map; Appendix B: Richmond Neighbourhood Location Map; Appendix C : Vancouver (West Side) Neighbourhood Location Map; Appendix D: North Vancouver Neighbourhood Location Map; and Appendix E: West Vancouver Neighbourhood Location Map. The areas are considered to be located within the inner or central core of the Region, and there were several reasons for choosing these particular neighbourhoods: 1. While neighbourhoods selected involved low, middle and high-priced house transactions, together they were considered to be a representative sample of the overall real estate market in the Lower Mainland area. 2. Significant differences in property values are found within this relatively concentrated geographic area. Consequently, a sample containing transactions at different price levels could be collected without requiring major adjustments for area differences. 1 3. It was possible to obtain detailed information on the sale of properties and their structural characteristics from two principal government sources. The Land Title Off ice (LTO), with branches in Vancouver and New Westminster, records the registered sales price ("Declared Market Value") and the outstanding charges (i.e. liens, mortgages, easements, etc.). The British Columbia Assessment Authority (BCAA) maintains records on the physical characteristics of dwellings which are regularly updated to establish assessed property values. These two sources were considered to offer a vast amount of data measured consistently between municipalities. The selection of specific neighbourhoods from which housing transactions were randomly sampled was based upon relatively comparable transportation (commuting) t ime, rather than distance, to the Central Business District of Vancouver. Discussions with B C A A appraisers helped to identify neighbourhoods which were representative of 57 FIGURE 5.1 REGIONAL LOCATION MAP DELTA j V SURREY 58 a specific level of house prices (e.g., low, middle, high) within a particular municipality. Five neighbourhoods were selected to reflect three levels of house prices, and they correspond to the subdistrict areas below which were established by the B.C. Assessment Authority: House B C A A Price Subdistrict Level Neighbourhood Number 1. Low Vancouver (East side) 22 2. Middle Richmond 1 3. High Vancouver (West Side) 13, 15 North Vancouver 20 West Vancouver 16, 17, 18, 19 The B C A A produces an annual summary of all property transactions within each subdistrict making it possible to analyze the volume of sales activity in relatively small defined neighbourhoods. The areas can then be systematically sampled, as the total number of sales can be tallied and properties sequentially selected (i.e. select every 3rd transaction). Due to the limited number of sales in North Vancouver and West Vancouver, two areas on the west side of Vancouver were also included to supplement the high priced category. The lowest priced category of house prices is Subdistrict #22 of Vancouver (Appendix A) situated on the east side of the City and generally formed by the boundaries of King Edward Avenue to the north, Victoria Street to the east, East 41st Avenue to the south and Main Street to the west. The middle category comprises Subdistrict #1 of Richmond (Appendix B) located in the northwestern area of the Municipality and bounded by River Road to the north, Railway Avenue to east, Williams Road to the south and the Gulf of Georgia to the west. The highest priced group consists of an area in the southwestern portion of Vancouver (Appendix C) roughly centered on Cypress Street and West 49th Avenue as well as two North Shore 22.0 21.5 21.0 I 20.5 N 20.0 T 19.5 E 19.0 R 18.5 E 18.0 S 17.5 T 17.0 16.5 R 16.0 A 15.5 T 15.0 E 14.5 S 14.0 13.5 13.0 12.5 12.0 FIGURE 5.2 CONVENTIONAL FIRST MORTGAGE RATES, 1980-1982 3 F M A M 3 3 A S O N D 1980 S O U R C E : G.W. Gau, Faculty of Commerce, U .B .C . J F M A M J J A S O N D 1981 T I M E ( M O N T H S ) J F M A M 3 3 A S O N D 1982 60 areas (Appendices D and E) north of the Trans Canada Highway and west and east of the Capilano River. The time frame of this analysis was specifically chosen to include a cycle of rapidly increasing and decreasing residential real estate values and mortgage interest rates, contrary to most studies which have used a static study period. A dramatic increase and decrease in interest rates and house prices occurred in the Lower Mainland area between 1980 and 1982 as interest rates exceeded 20% by the third quarter of 1981 and house prices peaked by the second quarter of 1981. Figure 5.2 illustrates the rise and fal l of conventional first mortgage rates throughout this period, and Figure 5.3 reveals the tremendous increase in average selling prices in Greater Vancouver after 1980. While average sale prices remained relatively constant at approximately $70,000 between 1975 and 1980, prices more than doubled by mid 1981 reaching levels of roughly $160,000. Analysis of quarterly house prices between 1980 and 1983 within the general study a r e a , 2 as presented on Table 5.1, indicates that the largest increase occurred between October 1980 and February 1981, after which time values stabilized and then declined. By October 1982, house prices were comparable to values in mid 1980. Much of the activity in the real estate and mortgage markets has been attributed to the following combination of events : 3 (a) High Immigration: In 1980 it was estimated that Vancouver gained 50,000 immigrants. (b) Limited Prospects for New Housing: A shortage of land exists due to government restrictions on the development of farmland near Vancouver and the geographic location of the Region (with water and mountains restricting outward growth). (c) A Shortage of Rental Accommodation and Pent-Up Demand for Owner- Occupancy: A decline in interest rates during the early part of 1980 encouraged households to obtain housing ownership. (d) Inflation Psychology and Speculation: As house prices increased, investors and speculators bought properties for immediate resale at inflated prices (and for anticipated profits). In addition, many homeowners, retired or nearing retirement, decided to sell as price increases pushed forward plans to "cash in" their property. FIGURE 5.3 AVERAGE RESIDENTIAL SELLING PRICE IN GREATER VANCOUVER. 1965 - 1981 S O U R C E : The Vancouver Sun, January 2, 1982. TABLE 5.1 TRENDS IN HOUSE PRICES BY A R E A . 1980 - 1983 Vancouver (East) Richmond Kerrisdale North Vancouver West Vancouver Quarterly Quarterly Quarterly Quarterly Quarterly Change Change Change Change Change Date Price (%) Price (%) Price (%) Price (%) Price (%) F e b . 80 N/A N/A $ 84,500 N/A $220,000 N/A $139,000 N/A $172,500 N/A Jun. 80 $168,500 N/A $145,000 + 71.60 $250,000 + 13.64 $147,000 + 5.76 $192,000 + 11.30 O c t . 80 $173,000 + 2.67 $185,000 + 27.59 $275,000 + 10.00 $189,000 + 28.57 $230,000 + 19.79 F e b . 81 $205,000 + 18.50 $250,000 + 35.14 $365,000 + 32.73 $230,000 + 21.17 $320,000 + 39.13 Apr . 81 $187,000 - 8.78 $250,000 N/A $380,000 + 4.11 $230,000 N/A $320,000 N/A Ju l . 81 $185,000 - 1.07 $235,000 - 6.00 $340,000 - 10.53 $230,000 N/A $298,000 - 6.88 O c t . 81 $176,000 - 4.86 $230,000 - 2.13 $305,000 - 10.29 $199,000 - 13.49 $298,000 N/A Jan . 82 $170,000 - 3.41 $195,000 - 15.22 $300,000 - 1.64 $185,000 - 7.04 $260,000 - 12.75 Apr . 82 $140,000 - 17.65 $180,000 - 7.69 $275,000 - 8.33 $172,500 - 6.76 $260,000 N/A Ju l . 82 $145,000 + 3.57 $180,000 N/A $265,000 - 3.64 $162,500 - 5.80 $230,000 - 11.54 O c t . 82 $133,000 - 8.28 $140,000 - 22.22 $240,000 - 9.43 $140,000 - 13.85 $200,000 - 13.04 Jan . 83 $138,000 + 3.76 $155,000 + 10.71 $250,000 + 4.17 $148,000 + 5.71 $210,000 + 5.00 S O U R C E : Royal Trust Survey of Canadian House Prices (Detached, 2 Storey). r-J 63 Periods of rising and falling real estate and mortgage markets are studied in this thesis since previous research by Zerbst & Bruggeman and Gunterman suggested that creative financing would not be capitalized at a constant proportion between different periods of market activity. One of the main sampling issues addressed in this thesis was whether Multiple Listing Service (MLS) data from the Greater Vancouver and Fraser Valley Real Estate Boards could be used. Research has generally relied upon MLS data since it is readily accessible and much faster to compile than randomly sampling all residential transactions. This source provides the sale and asking prices of a property, the length of time it had been listed on the market, as well as physical characteristics of the dwelling. However, it is difficult to establish whether MLS data is representative of all housing transactions including homes which sell by exclusive or private listings. While local realtors suggested that MLS transactions probably account for as much as 75% of all sales (on average), they indicated that this proportion may vary considerably depending upon market conditions. During periods of high activity, it was expected that a smaller proportion of vendors would list their property with the multiple service due to the higher commission than if the property was listed privately or by exclusive contract. In slower periods, a larger percentage of sellers were expected to use the MLS because of the better market exposure it offers. Figures obtained from the Greater Vancouver Real Estate Board tend to confirm this notion, since the number of MLS sales in 1981 was almost 76% lower than 1980 despite the substantial increase in total sales activity. As indicated on Table 5.2, the MLS sales volume for the general areas which include the study neighbourhoods reached a low in the first half of 1981, while sales almost doubled between the first and second halves of 1982. 64 TABLE 5.2 MULTIPLE LISTING SALES VOLUME BY AREA, 1980 - 1982 Months 1980 Jan. - June July - Dec. 1981 Jan. - June July - Dec. 1982 Jan. - June July - Dec. Vancouver (East) 670 714 415 450 598 1,011 Richmond 339 385 223 269 385 755 Vancouver (West) 105 229 156 231 309 648 North Vancouver 536 809 419 311 470 848 West Vancouver 160 261 159 135 159 296 S O U R C E : Greater Vancouver Real Estate Board. Although the number of MLS listings increased significantly between 1980 and 1981, it is reported that the proportion of MLS listings to all property listings fell during this period. It is possible that MLS sales may be systematically different than other sales due to the realtor/vendor/purchaser relationships. In a private transaction, the purchaser and vendor would likely negotiate independent of realtors (and their corresponding desire to close the sale to receive the commission fee). Since it could not be established that MLS sales are representative of all sales, MLS transactions were not used as the data base from which the sample was chosen. Instead, all arm's length, single family sales within the designated study areas were used as the population, which should be an improvement of the previous research. The B C A A compiles a list of all sales recorded at the Land Title Offices and they classify them according to the type of property (i.e. use) and the type of transaction (i.e. arm's length or nonarm's length). This information is based on physical inspections of the dwellings as well as analysis of the documentation involved in the sale. The properties TABLE 5.3 QUARTERLY HOUSE SALES BY NEIGHBOURHOOD, 1980 - 1982 Time of Sale (Quarter) 1980 - 1 2 3 4 Totals 1981 - 1 2 3 4 Totals 1982 - 1 2 3 4 Totals Vancouver Vancouver North and West Total Sales (East Side) Richmond (West Side) Vancouver by Year 136 98 69 65 157 107 60 93 234 153 137 130 184 122 84 70 711 480 350 358 1,899 (54.4%) 100 70 55 66 291 72 62 47 45 226 58 58 50 27 193 59 41 22 7 129 839 (24.9%) 76 76 102 N/A 254 43 63 77 N/A 183 31 24 42 N/A 97 27 31 37 N/A 95 629 (18.7%) Total Sales Over Study Period by Area (%) 1,256 (37.3%) 889 (26.4%) 640 (19.0%) 582 (17.3%) 3,367 (100.0%) 66 which were considered to represent a non-arm's length transaction, were omitted from the population base. A standard formula can be used to determine the necessary size of a sample which is independent of the population size.'* A representative sample of housing transactions utilizing creative financing is desired in this analysis, however, the incidence of creative financing in all residential property sales is unknown for the Region and each neighbourhood. Although U.S. studies imply the incidence is in the range of 50 - 8 0 % , 5 the rate is likely to be lower in Canada due to the short Canadian loan terms and thus shorter term benefits accruing from favourable financing. Due to the time and costs involved in identifying the actual incidence of creative financing in the Region, the formula was used as a guide for a reasonable estimate of sample size. If the worst case scenario is assumed here, where the population variance is at its maximum, the proportion of creative financing in all property transfers would be given as 50%. The formula to estimate sample size indicates a maximum size of 400. 6 Table 5.3 illustrates the total number of non-arm's length house sales tabulated for each of the five study neighbourhoods on a quarterly basis throughout the study period. The areas were then systematically sampled as a group to provide a preliminary sample of approximately 400 transactions, with the distribution of sales summarized in Table 5.4. 67 TABLE 5A PRELIMINARY SAMPLE STATISTICS Average Sample Sales Sales Neighbourhood Size Sold on MLS (%) Price Vancouver (East Side) : 1980 94 22 (23.4%) $ 87,100 Subdistrict #22 1981 39 15 (38.5%) $123,400 1982 31 20 (64.5%) $ 94,300 Richmond : 1980 69 20 (29.0%) $105,600 Subdistrict #1 1981 33 10 (30.3%) $152,200 1982 27 14 (51.9%) $120,100 Vancouver (West Side) : 1980 28 6 (21.4%) $157,900 Subdistrict #13 1981 15 5 (33.3%) $224,400 1982 14 7 (50.0%) $174,300 Vancouver (West Side) : 1980 22 6 (27.3%) $206,600 Subdistrict #15 1981 12 3 (25.0%) $359,900 1982 10 3 (30.0%) $266,500 North Vancouver : 1980 31 18 (58.1%) $135,000 Subdistrict #20 1981 10 8 (80.0%) $193,900 1982 9 7 (77.8%) $123,600 West Vancouver : 1980 20 13 (65.0%) $230,600 Subdistrict #16-19 1981 8 3 (37.5%) $379,900 1982 5 3 (60.0%) $223,400 Total Sample 477 183 (38.4%) The preliminary sample comprised 264 sales in 1980 (55.4%), 117 sales in 1981 (24.5%) and 96 sales in 1982 (20.1%). While the data indicates that only 38.4% of the transactions randomly selected sold on MLS, it is significant that MLS sales represent 32.2% of the transactions in 1980, 37.6% in 1981 and 56.3% in 1982. These figures suggest that MLS data is not necessarily representative of all sales during periods of different market activity and, as noted by a number of realtors, the incidence increased considerably with the downturn in the real estate market. Physical characteristics of each dwelling and financing details regarding each transaction were obtained from the B C A A and L T O offices, respectively. However, 127 properties had to be excluded from this preliminary sample due to unavailable or ambiguous information, leaving a net sample of 350 housing transactions. The sample size used in 68 this analysis is comparable to the majority of capitalization studies presented in Chapter 3.0. 5.2 Sample Statistics The final sample of housing transactions has been itemized by area and the year of sale in Table 5.5: TABLE 5.5 FINAL SAMPLE BY Y E A R AND NEIGHBOURHOOD Total Neighbourhood 1980 1981 1982 (By Area) Vancouver (East Side) 74 32 23 129 ( 37%) Richmond 52 24 16 92 ( 26%) Vancouver (West Side) 29 13 12 54 ( 16%) North and West Vancouver 47 16 12 75 ( 21%) Total (By Year) 202 (58%) 85 (24%) 63 (18%) 350 (100%) Since the lowest priced category of house sales represents 37% of the sample, 26% is from the middle range and 37% is from the upper end of values, the data set is considered to have reasonably even distribution of housing transactions at each price level. The breakdown of properties by year of sale reveals that 58% of the transactions occurred in 1980, 24% in 1981, and 18% in 1982. This result was expected since values and sales increased throughout 1980, peaking in early 1981. The sample has also been classified in terms of the incidence and type of financing arrangements used, since the documentation of debt financing for each sale was available at branches of the Land Title Of f ice . The certif icate of title for each property was analyzed to identify the registered financial charges, and the transfer document was inspected to obtain the registered sales price and the parties to the sale. Table 5.6 provides a breakdown of the sample by the year of sale and type of financing used. 69 TABLE 5.6 FINAL SAMPLE BY YEAR AND FINANCING ARRANGEMENT Free & Conventional Creative Conventional Year Clear (Only) (Only) & Creative Total 1980 36 82 59 25 202 ( 58%) 1981 18 26 30 11 85 ( 24%) 1982 13 22 22 6 63 ( 18%) Total 67 (19%) 130 (37%) 111 (32%) 42 (12%) 350 (100%) Properties which sold on an all-cash basis or by conventional financing accounted for 56% of the sample, with creative financing used 44% of the t ime. The incidence of creative financing is similar to the proportion selected earlier to estimate a reasonable size of sample. Creative instruments were the only source of debt funds in 32% of the transactions, and it was used in conjunction with conventional loans 12% of the time. As the sole source of funds, the most prevalent form of creative financing was loan assumptions (48%), followed by seller-financed mortgages (28%), loan assumptions plus seller mortgages (18%) and Agreements-for-Sale (6%). When used in conjunction with conventional loans, creative financing took the form of loan assumptions in 60% of these cases, seller-financed loans in 38% and a combination of both in only 2%. The incidence of creative financing in the sample was 42% in 1980, increasing to 48% in 1981 and declining to 44% in 1982. As predicted, the incidence is lower than that noted in the U.S. literature. Despite the fluctuations in the real estate and residential mortgage markets during the 1980 - 1982 period, the sample showed a relatively constant use of non-conventional financing. This is particularly interesting given the tremendous increase in interest rates during 1980 and 1981, peaking in September - October 1981, which should suggest that creative financing instruments would have been more prevalent in this period. It is possible that non-traditional sources were used for reasons other than interest rate relief such as reduced loan TABLE 5.7 CONVENTIONAL FIRST MORTGAGE RATES. 1970 - 1982 I n t e r e s t Y e a r M o nth R a t e 1970 J a n u a r y 10.58 F e b r u a r y 10.54 M a r c h 10.58 A p r i l 10.60 May 10.58 J u n e 10.53 J u l y 10.38 A u g u s t 10.40 S e p t e m b e r 10.36 O c t o b e r 10.35 November 10.28 December 10.16 1971 J a n u a r y 9.94 F e b r u a r y 9.72 M a r c h 9.28 A p r i l 9.20 May 9.25 J u n e 9.34 J u l y 9.46 A u g u s t 9.53 S e p t e m b e r 9.55 O c t o b e r 9.55 November 9.26 December 9.10 1972 J a n u a r y 9.04 F e b r u a r y 8.93 M a r c h 8.97 A p r i l 9.03 May 9.16 J u n e 9.37 J u l y 9.41 A u g u s t 9.41 S e p t e m b e r 9.38 O c t o b e r 9.35 November 9.30 December 9.22 1973 J a n u a r y 9.09 F e b r u a r y 9.02 M a r c h 9.07 A p r i l 9.15 May 9.30 J u n e 9.52 J u l y 9.71 A u g u s t 9.91 S e p t e m b e r 10.13 O c t o b e r 10.13 November 10.08 December 10.02 1974 J a n u a r y 10.02 F e b r u a r y 10.01 M a r c h 10.04 A p r i l 10.70 May 11.26 J u n e 11.37 J u l y 11.60 A u g u s t 11.85 S e p t e m b e r 12.05 O c t o b e r 12.05 November , 12.00 December 11.88 I n t e r e s t Y e a r M o nth R a t e 1975 J a n u a r y 11.81 F e b r u a r y 10.95 M a r c h 10.65 A p r i l 10.67 May 10.99 J u n e 11.23 J u l y 11.35 A u g u s t 11.52 S e p t e m b e r 11.94 O c t o b e r 12.15 November 11.97 December 11.89 1976 J a n u a r y 11.84 F e b r u a r y 11.80 M a r c h 11.90 A p r i l 12.03 May 11.99 J u n e 11.93 J u l y 11.86 A u g u s t 11.83 S e p t e m b e r 11.76 O c t o b e r 11.60 November 11.56 December 11.27 1977 J a n u a r y 10.75 F e b r u a r y 10.25 M a r c h 10.25 A p r i l 10.25 May 10.38 J u n e 10.35 J u l y 10.40 A u g u s t 10.33 S e p t e m b e r 10.32 O c t o b e r 10.34 November 10.34 December 10.33 1978 J a n u a r y 10.32 F e b r u a r y 10.31 M a r c h 10.33 A p r i l 10.42 May 10.43 J u n e 10.32 J u l y 10.31 A u g u s t 10.31 S e p t e m b e r 10.67 O c t o b e r 10.93 November 11.26 December 11.53 I n t e r e s t Y e a r M o nth R a t e 1979 J a n u a r y 11.28 F e b r u a r y 11.26 M a r c h 11.11 A p r i l 11.05 May 11.06 J u n e 11.16 J u l y 11.20 A u g u s t 11.80 S e p t e m b e r 12.25 O c t o b e r 13.50 November 14.46 December 13.58 1980 J a n u a r y 13.26 F e b r u a r y 13.50 M a r c h 14.69 A p r i l 16.94 May 13.99 J u n e 12.92 J u l y 13.09 A u g u s t 13.44 S e p t e m b e r 14.50 O c t o b e r 14.87 November 15.00 December 15.60 1981 J a n u a r y 15.17 F e b r u a r y 15.27 M a r c h 15.75 A p r i l 16.45 May 17.82 J u n e 18.55 J u l y 18.90 A u g u s t 21.30 S e p t e m b e r 21.46 O c t o b e r 20.54 November 18.80 December 17.79 1982 J a n u a r y 18.21 F e b r u a r y 18.97 M a r c h 19.41 A p r i l 19.28 May 19.11 J u n e 19.10 J u l y 19.22 A u g u s t 18.72 S e p t e m b e r 17.49 O c t o b e r 16.02 November 14.79 December 14.34 SOURCE: Bank o f Canada R e v i e w - J a n u a r y 1 974, 1978 and 1 9 8 3 . 71 qualification criteria or higher loan-to-value ratios. If fact , a discount was calculated for non-conventional financing in 16 of the transactions since the contract rate was higher than market rate of interest. The loan-to-value ratio exceeded the maximum limit of 75% for uninsured mortgages in only two of these cases; however, vendors were involved in providing funds in 11 of the sales which implies that buyers bypassed traditional sources for either loan qualification reasons or reduced loan origination fees. The sample revealed that there were three cases where the vendor provided funds at roughly market interest rates. Table 5.7 illustrates the level of conventional first mortgages from 1970 to 1983,7 w i t n monthly rates generally in the range of 9 - 11% during the 1970's. It. was only by the fourth quarter of 1979 that rates exceeded 13% to reach a high of 21% in the third quarter of 1981. Between 1980 and 1982, rates were from 13.26% to 21.46%, which was substantially higher than rates of the previous decade. Despite the rise and fall of interest rates during the study period, rates may still have been perceived as high in relation to previous levels, thereby resulting a relatively constant use of creative funds. Furthermore, there was substitution between the types of creative financing used during the study period. Table 5.8 illustrates the number of transactions involving either loan assumptions or vendor-financed loans as the only source of funds (sole source) or in conjunction with conventional funds (secondary source). The annual figures below are less than those presented in Table 5.6 because cases of mixed creative funds (i.e. a transfer using both assumed and seller financing) and Agreements-for-Sale are excluded. 72 TABLE 5.8 INCIDENCE OF ASSUMED AND VENDOR-FINANCED LOANS BY Y E A R Type of Creative Funds 1980 1981 1982 Total Loan Assumptions - Sole Source 39 10 6 - Secondary Source 16 7 2 55 17 8 80 (63%) Vendor-Financed Loans - Sole Source 12 10 8 - Secondary Source 8 4 4 20 14 12 46 (37%) Total 75 (59%) 31 (25%) 20 (16%) 126 (100%) Loan assumptions represented 73% of the creatively financed sales in 1980, 55% in 1981 and 40% in 1982. Evidence suggests that while the incidence of creative financing remained relatively constant during the 1980 - 1982 period, there was a decline in the use of flow-through funds such as loan assumptions and an increased use of vendor financing within the creative financing market. Since the accuracy relating to the incidence of creative financing declines as separate submarkets are analyzed, it is not relevant to detail the incidence and types of creative funds used in the individual neighbourhoods. The complete set of variables used in estimating and refining the price equation is displayed in Table 5.9. Variables related specifically to financing the transaction include PVSALE , P V L A G , L O A N and LNTOV . PVSALE represents the present value of the benefit or cost of non-market financing (i.e. funds provided at below or above market interest rates). As previously noted, there were 16 cases of non-market financing at above market rates. The premium or discount from creative financing is calculated by subtracting the market value of the mortgages from the outstanding balance of the loan at the date of sale. A similar calculation was done for P V L A G except, in this case, a lagged market discount rate was used to account for the time delay between signing the Interim 73 TABLE 5.9 LIST OF REGRESSION VARIABLES Variable Name Description D E P E N D E N T V A R I A B L E SPRICE Registered "Declared Market Value" from the Land Tit le Of f ice . INDEPENDENT VARIABLES  Financing Characteristics: L N T O V Loan-to-value ratio of all outstanding financing. L O A N Actual dollar amount of outstanding financing. P V S A L E Present value of benefit/discount from creative financing using market interest rate at sale date. P V L A G Present value of benefit/discount from creative financing using market interest rate six weeks prior to sale date. C F Dummy variable for sale transaction involving only creative financing. SF Dummy variable for sale transaction involving creative financing but only as a secondary source of funds. Y E A R 1 Dummy variable for 1980 sale. Y E A R 2 Dummy variable for 1981 sale. Y E A R 3 Dummy variable for 1982 sale. T IMEM Month of sale (1-36) during study period. Dwelling Characteristics: HSIZE Livable area of house (square feet). LSIZE Area of lot (square feet). A G E Age of dwelling (years). P O O L Dummy variable for swimming pool. F IREPL Number of fireplaces. C A R P T Dummy variable for carport. TABLE 5.9 (cont'd) Variable Name Description Dwelling Characteristics (cont'd): G A R A G E Dummy variable for garage. NROOMS Number of rooms (including bedrooms) in dwelling. NBEDS Number of bedrooms. NBATHS Number of bathrooms. B A S E M T Dummy variable for basement. D E C K Dummy variable for deck. D G L A Z E Dummy variable for double-glazed windows. R E V E N U Dummy variable for potential basement rental. O W N O C C Dummy variable for occupancy by registered owner. Neighbourhood Characteristics: ETHNIC Percentage of non-english speaking residents by enumeration area. F A M S Z Average number of people per family by enumeration areas. S F D U Percentage of single family dwellings by enumeration areas. HOMOWN Percentage of homeownership by enumeration areas. L O C N S Dummy variable for North Shore location. LOCVW Dummy variable for Vancouver (West Side) location. L O C R I Dummy variable for Richmond location. L O C V E Dummy variable for Vancouver (East Side) location. 75 Agreement and closing the sale. Consultation with a number of local realtors suggested that this lag would amount to approximately six weeks, on average. The mortgage rates used to calculate the present value of the savings benefits/costs are displayed in Appendices G , H, I and 3 which provide the average weekly interest rates for conventional first mortgages and the monthly rate for second mortgages. Since all financing and transfer documents were obtained for each sale, the appropriate discount rate could be identified based on the term, commencement date of the financing and the type of loan (i.e. first or junior loan). L O A N represents the dollar amount of all outstanding financing, and L N T O V is the proportion of the sales price that has been financed. Annual financing statistics were calculated for each neighbourhood and are presented in Table 5.10. TABLE 5.10 ANNUAL SAMPLE FINANCING STATISTICS BY NEIGHBOURHOOD Neighbourhood Vancouver (East Side) 1980 1981 1982 Richmond 1980 1981 1982 Vancouver (West Side) 1980 1981 1982 North Shore 1980 1981 1982 Average Sales Price $ 86,399 125,088 95,591 105,751 150,335 116,725 178,900 260,115 183,792 171,157 260,531 167,333 Average L/V Ratio .52 .52 .52 ,56 ,38 ,52 .37 .38 .41 .52 .41 .43 Average Benefit From Creative Financing^ $ 790 1,795 986 1,069 2,409 2,985 1,206 2,187 3,653 2,245 5,069 5,926 % o f Average Sales Price .914 1.435 1.028 1.011 1.602 2.557 .674 .841 1.988 1.312 1.946 3.541 76 The summary of annual financing details by neighbourhood confirms that house prices rose between 1980 and 1981 and generally dropped in 1982 to a level slightly above 1980 prices. Average loan-to-value ratios ranged from 37% to 56% over the study period which implies that creative funds were not being used to supplement conventional sources for lower downpayments. The average benefit from creative financing represents only a marginal proportion of the sales prices ranging from 0.674% - 3.541% depending on the year and area selected. The average discount was 5.34% of the sales prices in the Zerbst & Bruggement (1977) study, 3.51% in the Gunterman (1979) paper and 9.85% overall in the Rosen research. Although the proportions are generally lower in this thesis, it was expected since a Canadian sample of housing transactions was used. Characteristics of the neighbourhoods were compiled from 1981 Census data of Statistics Canada, matching the geographic boundaries of each subdistrict with the corresponding enumeration areas (refer to Appendix K) . Several traits were identified as being important determinants of a "neighbourhood" including percentage of owner occupied dwellings, multi - family development in the area, ethnicity, population density and education levels. Selection of preliminary neighbourhood variables was based on available census data with the following intended to reflect a particular measure: Desired Trait Census Data (by Enumeration Area) Owner-Occupancy: Percentage of housing stock owned or rented. Population Density: Average number of persons per family. Density of Development: Percentage of private occupied dwellings which are single family . Ethnicity: Mother tongue (percentage of population with a language other than english as their mother tongue). Since the focus of this research is on the capitalization effect of creative financing, neighbourhood traits were used only to help define the pricing equation. As noted in TABLE 5.11 MEANS AND STANDARD DEVIATIONS OF VARIABLES Standard Variable Mean Deviation SPRICE 138,758.914 71,084.122 HSIZE 1,312.191 527.569 LSIZE 5,229.035 5,174.859 A G E 28.223 21.466 R E M O D 0.160 0.367 P O O L 0.043 0.242 F IREPL 1.169 0.789 C A R P T 0.340 0.474 G A R A G E 0.454 0.499 D E C K 0.509 0.501 NROOMS 7.377 1.906 B A S E M T 0.740 0.439 P V L A G 1,669.916 4,043.632 R E V E N U 0.071 0.258 NBATHS 1.606 0.710 L N T O V 0.486 0.298 P V S A L E 1,888.689 4,464.155 O W N O C C 0.726 0.447 ETHNIC 0.280 0.156 F A M S Z 3.175 0.189 S F D U 0.854 0.176 H O M E O W N 0.823 0.150 78 Chapter 6.0, these traits were not significant in the analysis and were subsequently replaced with location dummy variables. Analysis of Table 5.11, which contains the means and standard deviations for the sample, indicates that the "average" dwelling in the sample sold for $138,759 with 1,312 square feet of livable area on a 5,229 square foot lot. The average dwelling age was 28.2 years, and 16% of the sample had undergone renovations. On average, there were 7.4 rooms, 3.4 bedrooms, 1.6 bathrooms and 1.2 fireplaces per dwelling. While only 4% of the sample had a swimming pool, over 50% had a deck, 34% had a carport, and 45% had a garage. The incidence of basements was 74%, with 7% being potential rental situations since there was a kitchen on both the basement and main floors. The homes were predominantly owner occupied (72.6%), and the average family size was 3.2 persons per dwelling. The dwellings were located largely in areas with a low proportion of multiple family dwellings (14.6%) as well as a low proportion of rental units (17.7%). The ethnicity of the area, measured by the percentage of residents who do not list English as their mother tongue, was 28%. The sample revealed an average benefit from creative financing of $1,889 using the market interest rate at the sale date (PVSALE) and an average benefit of $1,670 with a lagged market interest rate (PVLAG). The average loan-to-value ratio was 48.6% which is reasonable given the relatively high proportion of cash sales in the sample. The correlation matrix, as displayed on Table 5.13, showed high positive correlations between sales price and house size (.652), house size and number of bathrooms (.632), number of rooms and bedrooms (.801), number of bedrooms and bathrooms (.610), and proportion of single family dwellings and tenure (.682)). Since bedrooms can be used for other uses (such as a study), the number of rooms is considered a better measure of impact on sales price. The relationship between the type of dwelling unit and tenure is expected, since all of the neighbourhoods are zoned for single family development. There does not appear to be a linear relationship C O R R E L A T I O N S P R I C E H S I Z E AGE REMOD POOL F I R E P L CARPT GARAGE DECK NROOMS BASEMT P V L A G REVENU S P R I C E 1 000 0 652 -0 251 0 259 0 383 0 450 0 192 0 071 0 129 0 448 0 045 0 277 -0 0 4 7 H S I Z E 0 652 1 000 -0 41 1 0 215 0 399 0 417 0 163 0 151 0 058 0 558 -0 175 0 143 0 01 1 AGE -0 251 -0 411 1 000 -0 076 -0 083 -0 504 -0 275 -0 1 18 -0 000 -0 361 0 202 -0 076 0 009 REMOO 0 259 0 215 -0 076 1 OOO -0 013 0 085 0 032 -0 007 0 180 0 131 0 1 17 -0 035 -0 06 1 POOL 0 383 0 399 -0 083 -0 013 1 000 0 007 0 098 0 052 0 056 0 214 -0 003 0 100 0 089 F I R E P L 0 450 0 417 -0 504 0 085 0 007 1 000 0 191 0 147 0 094 0 430 0 135 0 130 -0 03 1 C A R P T 0 192 0 163 -0 275 0 032 0 098 0 191 1 000 -0 449 -0 127 0 162 -0 097 -0 04 1 0 035 GARAGE 0 07 1 0 151 -0 118 -0 007 0 052 0 147 -0 449 1 000 0 047 0 145 0 044 0 069 -0 053 DECK 0 129 0 058 -0 000 0 180 0 056 0 094 -0 127 0 047 1 OOO 0 213 0 316 0 040 0 162 NROOMS 0 448 0 558 -0 361 0 131 0 214 0 430 0 162 0 145 0 213 1 000 0 182 0 125 0 283 BASEMT 0 045 -0 175 0 202 0 1 17 -0 003 0 135 -0 097 0 044 0 316 0 182 1 OOO 0 018 0 089 P V L A G 0 277 0 143 -0 076 -0 035 0 100 0 130 -0 041 0 069 0 040 0 125 0 018 1 OOO -0 028 R E V E N U -0 047 0 01 1 0 009 -0 061 0 089 -0 031 0 035 -0 053 0 162 0 283 0 089 -0 028 1 000 NBATHS 0 504 0 632 -0 437 0 122 0 308 0 426 0 089 0 204 0 227 0 610 0 107 0 068 0 154 LNTOV -0 228 -0 222 0 142 - O 075 -0 127 -0 182 -0 059 -0 064 -0 049 -0 175 -0 054 O 193 0 005 P V S A L E 0 286 0 147 -0 070 -0 029 0 101 0 125 -0 032 0 076 0 006 0 106 0 006 0 975 - 0 031 OWNOCC 0 056 0 098 -0 062 -0 01 1 0 083 0 067 0 157 -0 056 -0 092 0 007 -0 029 -0 042 -0 07 8 E T H N I C -0 447 -0 384 0 309 -0 169 -0 142 -0 328 -0 280 0 002 0 129 -0 237 0 179 -0 131 0 258 F A M S Z 0 012 0 067 -0 106 -0 063 0 079 0 047 0 014 -0 033 0 096 0 131 0 061 -0 018 0 148 S F D U 0 094 0 066 0 071 0 063 0 122 0 1 10 0 086 0 066 0 015 0 034 0 081 0 045 -0 015 HOMOWN 0 244 0 311 -0 266 0 088 0 136 0 269 0 197 0 044 -0 044 0 167 -0 1 15 0 081 -0 133 N B A T H S LNTOV P V S A L E OWNOCC ETHNIC FAMSZ SFDU HOMOWN S P R I C E 0 504 -0 228 0 286 0 056 -0 447 0 012 0 094 0 244 H S I Z E 0 632 -0 222 0 147 O 098 -0 384 0 067 0 066 0 31 1 AGE -0 437 0 142 -0 070 -0 062 0 309 -0 106 0 071 -0 266 REMOD 0 122 -0 075 -0 029 -0 01 1 -0 169 -0 063 0 063 0 088 POOL 0 308 -0 127 0 101 0 083 -0 142 0 079 0 122 0 136 F I R E P L 0 426 -0 182 0 125 0 067 -0 328 0 047 0 1 10 0 269 C A R P T 0 089 -0 059 -0 032 0 157 -0 280 0 014 0 086 0 197 G A R A G E 0 204 -0 064 0 076 -0 056 0 002 -0 033 0 066 0 044 DECK 0 227 -0 049 0 006 -0 092 0 129 0 096 0 015 -0 044 T A RI F 5 11 NROOMS 0 610 -0 175 0 106 0 007 -0 237 0 131 0 034 0 167 1 l\DLL. JmLj B A S E M T 0 107 -0 054 0 006 -0 029 0 179 0 061 0 081 -0 1 15 P V L A G 0 068 0 193 0 975 -0 042 -0 131 -0 018 0 045 0 081 R E V E N U 0 154 0 005 -0 031 -0 078 0 258 0 148 -0 015 -0 133 NBATHS 1 000 -o 226 0 059 0 046 -0 090 0 186 -0 022 0 126 LNTOV -0 226 1 000 0 194 0 021 0 070 -0 029 -0 063 -0 128 P V S A L E 0 059 0 194 1 000 -0 032 -0 133 0 002 0 065 0 095 OWNOCC 0 046 0 021 -0 032 1 OOO -0 105 -0 005 0 084 0 149 E T H N I C -0 090 0 070 -0 133 -0 105 1 OOO 0 183 -0 03 1 -0 413 FAMSZ 0 186 -0 029 0 002 -0 005 0 183 1 OOO 0 339 0 313 S F D U -0 022 -0 063 0 065 0 084 -0 031 0 339 1 0 0 0 0 682 HOMOWN 0 126 -0 128 0 095 0 149 -0 4 13 0 313 0 682 1 OOO 80 between the loan-to-value ratio and either creative financing variable unlike some of the published research (See Sirmans (et al)). In conclusion, the sample is considered to be well distributed amongst areas and price level categories, with a relatively stable proportion of creative financing used throughout the study period despite the cycl ical pattern of house prices and interest rates. However, substitution between the types of creative financing occurred as vendor financing became an increasing proportion of the creatively financed sample. The average annual sales price for each area appears to conform to the pattern of prices in the market based upon statistics from the Greater Vancouver Real Estate Board, and evidence suggests that MLS sales may not be representative of all transactions particularly during different periods of market activity. 81 F O O T N O T E S 1 The Bible study suggested that the level of house prices had a significant impact on the capitalization effects of creative financing. 2 The tables in this chapter are often divided into 'neighbourhoods' or 'areas'. The distinction between the two groups is that 'neighbourhoods' comprise the entire sample at each price level, and they represent a subset of the 'area' classification. 3 Refer to the articles noted in The Vancouver Sun, The Wall Street Journal, The  Province. 4 D.A. Aaker and G.S. Day, Marketing Research. John Wiley & Sons, Inc., 1980, p. 255. 5 See Rosen, Salkin & Durning. 6 Ibid, p. 256. N = C 2 P(1-P) Where N = sample size (Error) 2 C = 2 for a 95% confidence level P = population proportion = .50 Error = the allowed interval estimate error 2 2 ( .50) ( l - .50) = 4(0.25) = 400 (0.05) 2 (.0025) This case assumes the population proportion is to be estimated within an error of .05 (or five percent points) at a 95 percent confidence level. Although reducing the confidence level and increasing the allowed error would reduce the sample size, the value of more accurate information through a larger sample is considered to outweigh the additional costs. 7 The interest rates are based on five year loan terms. 8 Based upon the market interest rate at the date of sale including any cases of below-market financing. 82 6.0 HOUSING PRICE EQUATION RESULTS To study the capitalization effects of creative financing on house prices, many formulations of the estimated price equation were tried using different explanatory variables, house price levels, periods of market activity and types of creative financing instruments. Crit icism of previous research is largely centred on the use of samples exclusively from MLS sources, with mixed financing arrangements or house price levels and restricted time frames during which the data was compiled. Because of the specific nature of the data, it has been difficult to generalize the results to other real estate markets and conditions. It should be noted that while a number of the studies report highly predictive equations for house prices, l imited statistics are provided regarding the relationship between explanatory variables or the fit of the models. The first step in developing this model is to identify the relevant explanatory variables to predict house prices based upon the full sample and list of financing, dwelling and neighbourhood traits. Using ordinary least squares regression, the objective is to statistically derive models with the most accurate estimate of value for the dependent variable while maintaining the lowest amount of estimation error. Regression analysis has two uses: (1) to predict the estimated sales price from a set of independent variables; and (2) to indicate the relative contribution of each trait. While the former application will define price equations for various submarkets of the sample, the latter application is more important here in assessing the capitalization impact of non-market financing and the validity of the cash-equivalent method. 83 Once the preliminary model is estimated, the next step is to divide the sample into several homogeneous submarkets in order to address the questions raised in Chapter 3.0. The following submarkets are identified: (i) House Price Levels - A plot of sale prices reveals that the sample should be segregated into low and high price groups based on a price of $126,000. The low price category is a sample of 183 transactions and the sample size of the high price group is 166. Bible and Gunterman had contrary conclusions based on low and high house price samples. (ii) Market Conditions - Variations in the level of interest rates and house prices occur each year of the study period. Market activity gradually increases during 1980, involving 202 transactions. There is rapid growth in the first half of 1981 followed by a decline, with a sample of 85 house sales. Prices and interest rates decline and stabilize in 1981, with 63 transactions during this period. It is not clear whether the annual distinctions will accurately capture changes in market activity and the corresponding capitalization impact. (iii) Type of Creative Financing - The source of the loan proceeds can be used to differentiate creative financing instruments into institutional (loan assumptions) or non-institutional (vendor-financed loans) categories. An institutional sample of 185 sales and the non-institutional sample of 160 transactions do not represent the full sample since each group comprises the specific type of financial instrument plus all sales conventionally financed; consequently, the two groups are not mutually exclusive. Since the focus of this research is on creative financing effects on house prices, the final step is to specifically analyze the two financing submarkets. At each step, statistics are generated to check that the models conform to the assumptions of regression analysis and to evaluate the validity of the results. Chapter 4.0 deals with a number of the inherent weaknesses in this type of analysis, outlining ways of eliminating or reducing their impacts. The final equations are selected using four major statistics which comprise the coefficient of determination (R^), the standard error of the estimate (SEE), the F value and the individual t-statistics for the explanatory variables and constant, also comparing them with previous research. The coefficient of determination measures the extent to which the sample information conforms to the linear relationship indicated by the model, and it is the proportion of variation of SPRICE explained by the regression model.* As the data more closely fits the estimated regression, R^ approaches a value of 1 (100%). SEE is 84 the standard deviation of the errors of residuals between the actual and predicted values of the dependent variable. The smaller this statistic becomes, the better the fit of the model. The F value is used to statistically test whether the regression model is significant, representing the relationship between the explained and unexplained variation in the model. Accordingly, a high F value suggests that the model is significant since the explained variation is large relative to the unexplained variation. The t-statistics are the ratios of each coefficient to its standard error and they indicate how well each explanatory variable contributes to the value of the dependent variable. Another measure of the influence of each explanatory variable on sale price involves the beta coefficients which are the regression coefficients adjusted to be free of the original units of measurement; the larger the beta coefficient, the stronger the impact of the variable on the dependent variable.2 While these coefficients were considered in analyzing the results, they are not presented in any of the tables. Plots of the sample data and correlation analysis reveal outliers and variables which potentially create estimation problems (i.e., high correlations with other variables; insignificance), thereby causing initial deletion or transformation of these characteristics. For instance, the four neighbourhood variables were often insignificant or had the incorrect sign, and they did not appear to capture locational differences between neighbourhoods possibly due to misspecification (i.e., incomplete list of traits). They were subsequently replaced with location dummy variables. Similar models using either PVSALE or P V L A G reveal that PVSALE is a more appropriate measure of the present value of the savings benefit based on R^, SEE and the signs and significance of the coefficients. PVSALE is calculated using the discount rate at the time of sale while P V L A G is calculated using a six week lagged rate. This lagged discount rate may not be reasonable for transactions in very active markets (i.e., short closings) or transactions of high price homes (i.e., longer closings to arrange financing). The average discount was also lower for P V L A G than PVSALE which reduces its impact on sales price. 85 Table 6.1 presents the best preliminary house price equation for the full sample. While the removal of outliers (where n = 341) marginally improved the estimations in terms of R 2 and SEE, numerous variables had the wrong signs. TABLE 6.1 EQUATION FOR THE FULL SAMPLE 3 SPRICE = 5,312.20 + 1.819 PVSALE - 9,936.77 Y E A R 1 (.674) (4.145)* (-1.935) + 47,222.43 Y E A R 2 + 40.30 HSIZE + 14,106.09 NBATHS (8.05)* (7.209)* (3.786)* + 11,309.39 C A R P T + 10,861.85 F IREPL + 13,362.65 BASEMT (2.661)** (3.747)* (2.771)** + 48,987.48 POOL + 45,113.51 L O C N S + 68,389.76 LOCVW (5.4867)* (8.199)* (12.105)* R 2 = .76638 SEE = 34,912.75 F = 100.8 n = 350 The equation explains about 77% of the variation in sale price which is comparable to results by Rosen and Sirmans (et al) and superior to the Zerbst & Bruggeman model. The variables have the expected signs and are significant with the exception of the constant and Y E A R 1 . The time and location dummies reflect their relative impacts on price so that Y E A R 1 indicates lower prices in 1980 than 1981 and not falling house prices during that year. Although not shown in Table 6.1, the beta coefficients suggest that the location and time dummy variables and house size have the strongest impact on house prices. The significant PVSALE variable shows that the present value of the payment savings from creative financing is being overcapitalized in sale prices. It was previously noted that vendors would have additional costs of analyzing, originating and servicing loans outside of the conventional mortgage system, thereby requiring a premium to be compensated. The results suggest that this premium is approximately 82% of the savings benefit. The SEE is about 25% of the mean sales price of homes in the sample (Y = $138,758.91), indicating considerable room for valuation error in this model. Nonetheless, it may compare to previous research, and the F value of 100.8 is high enough to conclude that the overall regression results are statistically significant. 86 A division of the sample into two levels of house price reveals that the financing variable was not significant in the low price category and significant in the high price category, contrary to Bible's research.* Gunterman's study showed that as house prices rise, sellers are more aware of the cost of mortgage discount points and shift an increasingly larger proportion of the cost to purchasers. Annual price equations suggest that the financing variable is only significant in 1981 which was a volatile period of market activity. In general, the equations illustrated in Table 6.2 have weaker predictive power than the full model, particularly the models based upon house price level. Although the PVSALE variable is not significant in the low price equation, it is in the high price model with a coefficient suggesting considerable overcapitalization of the savings benefit from creative financing. The mean values of PVSALE are $1,248 and $2,582, respectively, for the two samples which represent only about 1.4% of mean house prices. The small size of the discount may be the reason why the financing premium is not capitalized in the low sample group. Given the small relative amount of the savings benefit in the high price group, it is not clear why this amount is being overcapitalized aside from the possibility that buyers in this group are more sophisticated or aware of financing impacts due to the high purchase prices. The work by Bible uses comparable sample sizes but relies exclusively on one type of financing instrument (loan assumptions). It is presumed that the mixture of financing types distorts the capitalization impacts noted here. Annual price equations have significantly better predictive ability than the previous two price equations with similar error estimates to the full model. Creative financing is not capitalized in the 1980 or 1982 models which represent periods of increasing and decreasing market activity. In all three models, HSIZE and LOCVW had the strongest impact on sales price as indicated by the beta coefficients. In 1980, PVSALE is a small amount on average ($1,260) which likely results in it being insignificant; a monthly time dummy variable accurately indicates the growth in property values. In 1981, PVSALE (mean value of $2,633) is not related to the monthly 87 TABLE 6.2 RESULTS BY HOUSE PRICE LEVEL AND YEAR OF SALE SEE_ PVSALE Coefficient Model n Y R 2 (% of Y) F (t-statistics) Low Price 183 91,174 .40 17,150 16.49 .945 Sample (18.8%) (1.756) High Price 166 191,294 .59 46,488 26.96 2.65 Sample (24.3%) (3.99)* 1980 Sample 202 124,382 .75 30,590 65.57 1.27 (24.6%) (1.714) 1981 Sample 85 178,363 .79 41,890 32.26 3.46 (23.5%) (3.94)* 1982 Sample 63 131,424 .72 33,213 15.46 1.13 (25.3%) (1.178) 88 time dummy which may account for its significant coefficient. House prices and interest rates increased dramatically only to drop off towards the end of the year, making both the impact of financing difficult to ascertain in this erratic period and the time dummy variable insignificant. A more useful distinction of market activity is the division of the sample into periods of rising and falling house prices and interest rates since the real estate and mortgage markets peaked at different times. The insignificant PVSALE variable in 1982 has a mean value of $2,833, and the time dummy, although not significant, had the correct sign. PVSALE is positively correlated with L N T O V in this model which accounts for the insignificance of both variables. These three yearly models are not considered particularly useful since several independent variables are either not significant or exhibit the incorrect signs. It is proposed that the division of the sample by year does not capture changes in market conditions due to the variations of rates and prices, particularly in 1981. Preliminary regression analysis of the full sample indicates that the following list of characteristics are important in predicting house prices: PVSALE , LNTOV, annual time dummies (YEAR1, Y E A R 2 , YEAR3) , HSIZE, NBATHS, A G E , BASEMT, POOL and location dummies (LOCNS, LOCVW, LOCRI , LOCVE) . Numerous price equations are estimated using variations of these variables to check if refinements do improve the overall model. However, a mixed sample of creatively financed transactions is not expected to produce clear results of the appropriate adjustment process for the reasons Rosen's model was crit icized. Nonetheless, the R 2 increases from .67 to .76 with a corresponding drop in SEE from 41,620 to 35,360 based on the poorest and strongest models derived.^ It is clear that modifications to the set of variables, either through transformation or omission, affect the house price equation which is certainly illustrated by other research and is not new to this study. It is important here that in all of the estimated equations, the PVSALE variable is signficant at the .01 level and the coefficient has a value of approximately 2, suggesting that the financing premium is overcapitalized in the price of single family TABLE 6.3 RESULTS FOR THE VENDOR-FINANCED SAMPLE Model Description 1. Drop L O A N 2. Drop A G E 3. Drop HSIZE; Add BASEMT, NBATHS 4. Add BASEMT, NBATHS 5. Monthly TIME 6. Rising Prices 7. Falling Prices 8. Rising Rates 9. Falling Rates Sample Size (No. of Creative Sales) 160 (30) 160 (30) 160 (30) 160 (30) 160 (30) 109 (14) 51 (16) 127 (19) 33 (11) R2 (SEE) .68 (40,339) .72 (37,933) .72 (38,532) .74 (37,130) .72 (37,868) .89 (19,628) . 83 (42,541) .84 (31,495) .81 (18,270) F Value 36.1 43.0 33.8 34.4 39.0 76.2 19.9 60.2 9.3 PVSALE Coefficient (t-statistics) 3.79 (4 .90)* 3.19 (4 .41)* 3.48 (4 .67)* 3.23 (4 .48)* 3.34 (4 .56)* 2.11 ( 2 . 6 0 * 3.00 (2 .47)* 3.86 (5 .53)* .31 (.25) No. 9 Explanatory Variables  Wrong Sign Not Significant 1 1 11 12 10 10 10 10 10 1 1 oo 90 homes. The next step in evaluating the explanatory power of the creative financing variable is to segregate the data according to the type of financing arrangement used in the transaction. The division of the full sample by type of creative financing arrangement produces a vendor-financed group of 160 sales and a loan assumption group of 185 sales. It should be noted that each submarket comprises transactions exclusively financed by coventional sources or by the particular creative instrument. Since the PVSALE variable assumes a value of 0 for conventionally financed transactions, the coefficient will capture the impact from the creative transactions. Incorporating the conventionally financed sales in either submarket increases the sample sizes substantially, as there are 30 transactions which are exclusively financed by the vendor and 55 loan assumption cases (refer to Table 5.8). As previously discussed, two pricing responses are anticipated: (1) the non-institutional group, or vendor-financed submarket, should overcapitalize the savings benefit and indicate a PVSALE coefficient in excess of 1. The vendor faces higher loan origination and servicing costs than a financial intermediary and will require compensation through a higher than market yield. Calculation of the savings benefit at the market rate will result in an understatement of the financing benefit. (2) the instutitional group, or loan assumption submarket, should correctly capitalize the savings benefit since the funds were actually originated by the financial intermediary. The P V S A L E coefficient may not be exactly 1 for reasons outlined in Chapter 3.0, but it is expected to approximate unity. Based upon the earlier analysis, house price equations are derived for the two submarkets using different explanatory variables and assessing the use of annual models to reflect market conditions. A summary of the results are provided in Tables 6.3 and 6A. Since market conditions cannot be isolated by annual distinctions, it is necessary to determine the periods of rising and falling prices and interest rates. Property values increased between January 1980 and April 1981, dropping after that time. Mortgage rates rose from January 1980 to peak in October 1981, six months after the peak in house prices. The first five equations on each table represent TABLE 6.4 RESULTS FOR THE LOAN ASSUMPTION SAMPLE Model Description 1. Drop L O A N 2. Drop A G E 3. Drop HSIZE; Add BASEMT, NBATHS 4. Add BASEMT, NBATHS 5. Monthly TIME 6. Rising Prices 7. Falling Prices 8. Rising Rates 9. Falling Rates Sample Size (No. of Creative-Sales) 185 (55) 185 (55) 185 (55) 185 (55) 185 (55) 140 (45) 45 (10) 157 (49) 28 (6) R2 (SEE) .63 (33,760) .73 (29,099) .73 (29,048) .75 (28,280) .74 (28,517) .87 (20,168) .79 (29,698) .84 (23,329) .68 (24,300) F Value 33.4 51.7 42.7 42.1 49.2 82.3 14.9 77.8 4.3 PVSALE Coefficient (t-statistics) - 1 . 1 6 ( - .977) .802 (.762) .912 (.865) .766 (.746) .969 (.939) .96 (.913) - . 7 8 ( - .477) .704 (.745) .58 (.261) No. 9 Explanatory Variables  Wrong Sign Not Significant 2 3 1 11 12 10 10 9 10 92 changes to the list of explanatory variables, while the last four divide the samples according to market activity. It is not particularly disturbing that several variables are not significant or have the incorrect sign in consideration of the prior discussion on multicollinearity; there are no correlations between explanatory variables which exceed the R 2 of the equations. Logarithmic transformations are done to the house size and age variables. The number of observations per equation should be greater than the degrees of freedom which is always the case; however, the small number of creatively financed transactions during the period of falling rates is considered to limit the use of the last equation. The nine vendor-financed equations on Table 6.3 have an R 2 of .68 to .89 and, with the exception of no. 9, the F values are high enough to conclude that the overall results are statistically significant. In general, the PVSALE coefficient is significant ranging from a value of 2.11 to 3.86. Variations of the explanatory variables show that the most complete equation (no. 4) is the best predictor of SPRICE which is the anticipated result. In fact, the inclusion of HSIZE in no. 4 does significantly add to the model when compared to equation no. 3.6 While the results are statistically comparable to some published research, the capitalization effect is unlike previous studies. The distinction here, however, lies on the use of an exclusively vendor-financed sample, rather than a mixed sample or a loan assumption group. The increase in house prices and rates during each growth period is much greater than the subsequent periods of decline. Increased variability in interest rates would create increased variability in calculating the discount and would lead to an even greater miscalculation or underestimate of the vendor financing benefit. The average present value of the savings benefit is only $490 when prices are rising (Y = $127,205), compared with $2,339 when prices are falling = $143,463), which likely accounts for the lower relative beta coefficient and role in predicting sales price in the former case. Also, in the vendor-supplied loan, the loan proceeds are a function of the price of house which causes greater estimation problems with variability in house prices. 93 Loan proceeds were similar, on average, for the rising and falling price equations at approximately $75,000, despite having a greater impact on price in the rising price sample. This, may be a result of the corresponding increase in interest rates and potential rationing of conventional mortgage funds which would make vendors an important loan source. The capitalization impact of creative financing based on market activity does differ depending on the criteria used (i.e., house prices versus interest rates). The loan assumption sample achieves R 2 ranging from .63 to .87 for the nine equations shown on Table 6.4. Again, the best predictive model using variations of the explanatory variables is no. 4, and the addition of another variable does add to the model. The PVSALE variable has the expected coefficient which is roughly 1. The insignificance of the coefficients is not surprising in light of the smaller loan amount and short term of loan assumptions relative to vendor loans, with the average savings benefit generally being less than $900. The predictive ability of the models is comparable to other research, although the standard errors of the estimates indicate there is notable valuation error if used to directly predict Lower Mainland house prices. The small amount of the discount is unlike U.S. studies where creative financing benefits generally accrue over a much longer time frame and represent a larger proportion of the purchase price. In summary, two distinct price responses are derived for creatively financed single family homes. Transactions using vendor-financed loans suggest that the discount or savings benefit to the purchaser is overcapitalized in house prices using a market discount rate to calculate the creative financing premium. Since the vendor incurs higher costs in analyzing, originating and servicing a loan compared with the conventional financial lender, his discount rate should be higher. The use of an inappropriately low discount rate results in an understatement of the present value of vendor financing benefits and suggests the compensation would then be through overcapitalization of the benefit in house prices. Despite being a form of creative 94 financing, assumption loans differ from vendor loans as they are originated by conventional mortgage channels. In this case, the appropriate discount rate in estimating the discount is the market rate used by financial intermediaries. The savings benefit from vendor loans was overcapitalized in house prices, and the benefit approached unity (although not significant) in the assumption loan case. The small amount of the loan assumptions is attributed to the short terms of Canadian mortgages which results in a considerably shorter assumption period than U.S. samples. 95 F O O T N O T E S 1 The computerized regression lists three items involving R, with the coefficient of determination being R S Q U A R E . 2 Beta coefficients are calculated as = bj(sXi/sy). They are the change in terms of the standard deviation of Y that is expected if X.[ were increased by one standard deviation of Xj holding the other independent variables constant (See Aaker & Day, p. 449). 3 The t-statistics are shown in parenthesis below the coefficients: * = coefficients signficant at .01 level; * * = coefficients significant at .05 level. 4 The division of this sample by house price level also represents a geographic distinction, and it is possible that lenders discriminate between areas. Therefore, the results may be somewhat biased in using the entire metropolitan sample instead of analyzing house price levels within a particular neighbourhood. 5 The poorest model replaced the annual time dummies with a monthly indicator over the entire study period and BASEMT and NBATHS were not included; of the ten explanatory variables, one had the wrong sign and four variables were not significant. In the strongest model, all explanatory variables were included using annual time dummies. 6 Reject the hypothesis that the additional variable adds no explanatory power if F > F (k 2 - k j , n - k 2 - 1); F = 13.32 > F . 0 5 ( l , 158) = 3.84. The calculated F Value is based on the precise R 2 values rather than the rounded figures shown in Table 6.3. 96 7.0 S U M M A R Y A N D PRICING IMPLICATIONS The objective of this thesis was twofold: (i) to determine the incidence of creative financing techniques in a Canadian real estate market during a cyclical period of market activity; and (ii) to determine the adjustment process or pricing mechanism for creative financing instruments. Previous research on the capitalization effects of creative mortgage financing differs from this research in a number of ways: - the use of U.S. housing transactions where loan terms are considerably longer, causing financing benefits to accrue over a longer period and to amount to a larger proportion of the purchase price; - the use of U.S. housing transactions where mortgage interest is tax deductible which results in higher tax bracket purchasers discounting the benefit from low-rate financing; - the use of data enclusively from MLS sources when it has not been shown that this source is representative of all housing transactions; the use of either mixed financing samples or a sample involving a particular type of financing where vendor-financing has not been specifically studied or analysis has not been completed comparing a mixed and single financing sample; the use of data compiled during a static period of market activity where the results are difficult to generalize to other markets and market conditions; and - the use of assumptions regarding the holding periods of creative funds. This research attempts to improve on previous studies by using a sample of Canadian housing transactions, unaffected by interest deductibility or long-term loans, which were taken during a cyclical period of real estate and mortgage market activity. The sample was derived from all arm's length house sales in a defined area, and the financing details were obtained for each transfer. The sample was divided into 97 submarkets to study the capitalization effects of different types of creative financing and three items were hypothesized: (1) a sample of mixed creative financing instruments will distort the capitalization results; (2) a sample of loan assumptions will accurately capture the financing benefits; and (3) a sample of vendor-financed loans will overcapitalize the financing benefits. The sample of 350 house sales in the Lower Mainland of British Columbia between 1980 and 1982 revealed that the incidence of creative financing was relatively constant despite considerable fluctuations in real estate prices and mortgage rates. The incidence of transactions using non-conventional funds in a first or secondary role was approximately 45% throughout the study period which is comparable to other U.S. reports. However, the proportion of creative financing did not react in cycl ical manner with changes in interest rates which is contrary to earlier research in the Greater Vancouver area. It is possible that market participants perceived the level of mortgage rates in the 1980's to be substantially greater than rates of the previous decade. Accordingly, the tremendous increase in rates seen in 1981 was not accompanied by a large increase in the relative use of creative financing. A more reasonable explanation is found in the substitution between types of creative financing with vendor-supplied loans becoming a more significant source of creative funds while fewer loans were being assumed. Although the overall incidence of creative financing was stable during the 1980 - 1982 period, the proportion of seller-financing did increase. Creative financing instruments can be distinguished on the basis of where the loan was originated, with loan assumptions being originated initially by a financial intermediary and vendor loans originating from the seller. The distinction between these two groups is important as the vendor does not have the economies of sale in loan analysis, origination and servicing afforded by financial institutions. The use of 98 market rates to calculate the present value of the savings benefit will cause an understatement of the benefit in the case of vendor-financed sales. Evidence indicates that the vendor is compensated for the higher costs of providing funds through a higher discount rate or overcapitalization of the benefit in sale prices of single family homes if a market discount rate is used. The results confirm the three pricing responses anticipated. The capitalization results using a mixed financing sample was not clear since a coefficient of approximately 2 was estimated. The loan assumptions suggested that creative financing was accurately capitalized into higher prices. The vendor loan sample indicated that creative financing was overcapitalized in the sales price due to the use of an incorrect discount rate. An area of further study would be researching the secondary yields on vendor financing. The yields on these loans from sellers to other mortgage market participants should indicate the true rate required by lendors of non-institutional loans and provide for a more accurate calculation of the present value of the payment savings from vendor-financing loans. Four questions were raised in this research regarding the capitalization impact of creative financing. It is clear that refinements to the house price equation improve its predictive ability but not necessarily the contribution of the financing variable if the sample set is misspecified. Creative financing does not appear to be capitalized in house prices to the same extent given different levels of house prices or periods of market activity. Alternative creative financing instruments do not produce similar pricing adjustments as noted earlier. While this sample of Canadian house transactions is considered to provide a clearer interpretation of the capitalization tests with the absence of tax and interest rate risk considerations common to U.S. studies, the short term of Canadian rollover mortgages means that the savings benefit is substantially smaller and of less significance in influencing sale prices. In conclusion, it appears that current appraisal practice using the cash-equivalent method is appropriate for sales involving loan assumptions, but is not 99 correct for property transfers financed through seller loans. The evidence suggests that appraisers should use a higher than market discount rate to calculate the benefits from a vendor-supplied loan. Also, it appears that borrowers do not derive any direct benefit from low-rate loan assumptions, as the purchase (sales) price of houses would have increased by an amount roughly equal to the present value of the payment savings. Borrowers seem to have used this type of funds to produce a financing scheme similar to a graduated payment mortgage. 100 BIBLIOGRAPHY Aaker, D.A. and Day, G.S. Marketing Research. John Wiley & Sons, Inc., 1980, 254-257. Albritton, H.D. "A Critique of the Prevailing Definition of Market Value," The  Appraisal Journal, (April 1980), 199-205. Baldwin, W. "Where Will the Money Come From?" Forbes, (September 14, 1981). Ball, M.J . "Recent Empirical Work on the Determinants of Relative House Prices," Urban Studies, 10(1973). "Below-Market Financing Clouds Home Price Statistics," Real Estate Quarterly, 2 (Winter 1983), 3-4. Benage, W.F. , Jr. "Which Offer is Worth More?," The Real Estate Appraiser and  Analyst, (Spring 1981), 28-30. Bible, D.S. "Financing Conditions as Related to Residential Value," The Real Estate  Appraiser and Analyst, 46 (May-June 1980), 39-41. Bible, D.S. and Crunkleton, J.R. "The Effects of Financing on the Sale of Multi -family Properties," The Real Estate Appraiser and Analyst, 49 (Summer 1983), 33-37. Blumberg, D. "Demystifying Creative Finance," Real Estate Review, 12 (Summer 1982), 18-20. Boyce, B.N. Real Estate Appraisal Terminology. American Institute of Real Estate Appraisers, 1975, 136-137. Brigham, E.F. "The Determinants of Residential Land Values," Land Economics, (February 1965), 325-334. Bruggeman, W.B. and Zerbst, R .H. "Discount Points and Housing Prices: A Reply," The Journal of Finance, XXXIV (September 1979), 1055-1060. Butler, R.V. "The Specification of Hedonic Indexes for Urban Housing," Land  Economics, 58 (February 1982), 96-108. Case, F .E . "Creative Financing Instruments," The Real Estate Appraiser and Analyst, (Spring 1982), 45-58. Cassidy, H. and Field, A . "The Pricing of Alternative Mortgage Instruments," Alternative Mortgage Instruments Research Study, 111 (November 1977), 1-30. Church, A . M . "An Econometric Model for Appraising," A R E U E A Journal, 3 (Spring 1975), 17-31. Clauretie, T . M . "A Note on the Bias in House Price Capitalization Models," A R E U E A  Journal, 11 (Winter 1983), 521-524. 101 Cohn, R .A . and Fischer, S. "Alternative Mortgage Designs." New Mortgage Designs  for Stable Housing in an Inflationary Environment. Conference Series No. 14 (1975), 47-69. Colton, K.W., et a l . "Borrower Attitudes Toward Alternative Mortgage Instruments," A R E U E A Journal, (Winter 1979), 561-609. Colwell , P.F. "A Statistically Oriented Definition of Market Value," The Appraisal  Journal, (January 1979), 53-58. Colwell , P.F., et al . "Discount Points and Housing Prices: Comment," The Journal of  Finance, XXXIV (September 1979), 1049-1054. Corgel, J .B. and Goebel, P.R. "Financing Adjustments via Cash Equivalency," The  Real Appraiser and Analyst, 49 (Spring 1983), 55-61. Curcio, R.J. , et a l . "Bayesian Regression Procedures Applied to the Valuation of Residential Real Estate," The Real Estate Appraiser and Analyst, (Winter 1981), 46-48. Curcio, R.J . and Webb, J .R. "Creative Finance and Affordable Housing," Real Estate  Review, 11 (Summer 1981), 77-80. D'Ardenne, R.H. "Valuation Subject to Existing Financing," The Appraisal Journal, (October 1980), pp. 512-520. Danek, C E . "Creative or Favorable Financing and Its Effect on Value," The Real  Estate Appraiser and Analyst, 49 (Summer 1983), 31-32. DeLacy, P.B. "Cash Equivalency in Residential Appraising," The Appraisal Journal, (January 1983), 81-88. DeMagistris, R. "Impact of 'Buy Downs' on Affordability and Home Prices," The  Federal Reserve Bank of New York, Quarterly Review, (Summer 1982), 41-45. Dickerson, F .G . "A Technique for the Analysis of Owner-Financed Sales," The  Appraisal Journal, XLIV (January 1976), 55-59. Doherty, J .L . "Favorable Financing Effects of Values of Single-Family Residences," The Appraisal Journal, XXX IX (July 1970), 398-405. Dorchester, J .D. , Jr. "A Response to 'Valuation Subject to Existing Financing'," The  Appraisal Journal, (January 1981), 102-103. Eger, A . F . Financing the Market for Existing Housing: An Alternate Source of Funds. Ph.D. Thesis, University of British Columbia, September 1976. Findlay, M .C . and Fischer, M.C . "On Adjusting the Price of 'Creatively Financed' Residential Sales: Cash Equivalence vs. F F V A , " Housing Finance Review, 2 (January 1983), pp. 63-79. Fischer, J . and Sander, A . B . "Capitalization Rates and Market Information," The  Appraisal Journal, (April 1981), 186-198. 102 Friedman, J .P. "Valuation of Encumbered Acreage," The Appraisal Journal, (July 1975), 410-420. Friedman, J .P. and Lindeman, J .B. "Cash Equivalent Analysis," The Appraisal Journal, (January 1979), 35-43. Friedman, J .P. and Lindeman, J.B. "Seller Financing and Cash Equivalence," The Real  Estate Appraiser and Analyst, 45 (May-June 1979), 46-50. Garcia, K. "Sales Prices and Cash Equivalents," The Appraisal Journal, (January 1972), 7-16. Garrett, J . "Mortgage Warehousing: Creative Financing for Residential Tract Builders," Real Estate Review, 12 (Fall 1982), 47-50. Gau, G.W. "Capitalization of Creative Financing into Single Family House Prices." Unpublished paper, 1984. Gau, G.W. An Examination of Alternatives to the Rollover Mortgage. Paper Prepared for Canada Mortgage and Housing Corporation, June 1981. Grether, D .M. and Mieszkowski, P. "Determinants of Real Estate Values," Journal of  Urban Economics, 1 (1974), 127-146. Gullickson, A . E . and Hewitt R. "Adjusting to Reality," The Real Estate Appraiser and  Analyst, 47 (Summer 1981), 22-27. Gunterman, K .L . "Adjusting Comparable Sales Data for the Effects of F H A Financing," The Real Estate Appraiser, 44 (July-August 1978), 12-18. Gunterman, K .L . " F H A Mortgage Discount Points, House Prices and Consumer Behavior," A R E U E A Journal, 7 (Summer 1979), 163-176. Gunterman, K .L . "Financing and Selling Price of Single-Family Houses," Research in  Real Estate, 1 (1982), JAI Press, Sir mans, C F . (ed), 255-274. Hale, C . G . "Buying-Down Interest Rates: New Form of Typical Financing?" The Real  Estate Appraiser and Analyst, 49 (Summer 1983), 21-26. Hamilton, S.W. et al . Real Estate Finance in a Canadian Context. University of British Columbia, September 1984. Heckman, J .J . "Sample Selection Bias as a Specification Error," Econometrica, 47 (January 1979), 153-161. Jackson, J .R. "Intraurban Variation in the Price of Housing," Journal of Urban  Economics, 6 (1979), 464-479. Johnson, M.S. and Lea, M.J . "Differential Capitalization of Local Public Service Characteristics," Land Economics, 58 (May 1982), 189-203. Koch, D.L . et a l . "The Risks of Creative Financing," Economic Review, Federal  Reserve Bank of Atlanta, LXVII (December 1982), 4-13. 103 Lang, 3.R. and Jones, W.H. "Hedonic Property Valuation Models: Are Subjective Measures of Neighborhood Amenities Needed?" A R E U E A Journal, 7 (Winter 1979), 451-463. Linneman, P. "Some Empirical Results on the Nature of the Hedonic Price Function for the Urban Housing Market," Journal of Urban Economics, 8 (1980), 47-68. Lipscomb, J.B. "Discount Rates for Cash Equivalent Analysis," The Appraisal Journal, (January 1981), 23-33. Lucas, R .E .B . "Hedonic Price Functions," Economic Inquiry, 13 (June 1975), 157-177. Maes, M.A. "The Emergence of Cash Equivalency in Valuation," Real Estate Review, 12 (Fall 1982), 87-90. Mark, J .H . "A Preference Approach to Measuring the Impact of Environmental Externalities," Land Economics, 56 (February 1980), 103-116. Martin, R.S. "Estimating the Effect of Existing Financing on Market Value Using an After Tax Approach," The Real Estate Appraiser and Analyst, 48 (Summer 1982), The M L A Style Sheet. Second Edition. New York: The Modern Language Association of America, May 1970. Polinsky, A . M . "The Demand for Housing: A Study in Specification and Grouping," Econometrica, 45 (March 1977), 447-461. The Province. "House Resale Market Slumps in Vancouver," (August 19, 1981). Reenstierna, E.T. "Value Spread: The Effects of Occupancy, Financing, and Buyer/Seller Motivations on Most Probable Selling Price," The Real Estate  Appraiser and Analyst, 49 (Summer 1983), 39-43. Rhodes, F .G . "Inflation Discounting: How Appraisers Can Deal With Inflation," The  Appraisal Journal, (April 1981), 234-247. Roberts, R .M. "The Value of an Assumption," The Appraisal Journal, (July 1982), 428-433. Rosen, K.T. Creative Financing and House Prices: A Study of Capitalization Effects, (Working Paper 82-52). Center for Real Estate and Urban Economics, University of California, Berkeley. (August 1982). Rosen, S. "Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition," Journal of Political Economy, 82 (January-February 1974), 34-55. Salkin, M.S. and Durning, D. "What is a House Really Worth?" Mortgage Banking, 43 (October 1982), 10-20. Schafer, R. "The Effect of BMIR Housing on Property Values," Land Economics, 8 (1982), 282-286. Schwartz, A . L . , Jr. "Influences of Seller Financing Upon Residential Property Sales Prices," The Real Estate Appraiser and Analyst, 48 (Winter 1982), 35-39. Sirmans, G.S. et a l . "Assumption Financing and Selling Price of Single-Family Homes," Journal of Finance and Quantitative Analysis. Smith, D.V. "An Appraiser Looks at Multiple Regression," Appraisal Institute  Magazine, 23 (November 1979), 37-39. Straszheim, M. "Hedonic Estimation of Housing Market Prices," The Review of  Economics and Statistics, LVI (1974), 404-406. Swan, C . "Alternative Mortgage Instruments and Mortgage Defaults," Alternative  Mortgage Instruments Study, (November 1977), ix(l)-ix(35). Treadway, P. "Housing Finance in the '80's," Real Estate Today, 14 (July 1981), 13-15. The Vancouver Sun. "Peak Seen in House Prices for Lower Mainland Area," (January 20, 1981). "They're Cashing In and Moving Out," (March 28, 1981). "Housing Ride Derailed Homeowners, Builders, Realtors," (December 26, 1981). The Wall Street Journal. "Vancouver's Housing Collapse Shows Dangers of Speculation," (August 26, 1981). Wiens, E .G. and Wales, T . J . "Capitalization of Residential Property Taxes: An Empirical Study," The Review of Economics and Statistics, LVI (1974), 329-333. Wright, A . L . and Gilli land, C E . "Seasonal Variations in Home Sale Prices and Time Adjustments in the Market Data Approach," The Real Estate Appraiser and  Analyst, 47 (Winter 1981), 29-33. Young, F. "Calculation of the Adjustments for Beneficial Mortgage Financing," Appraisal Institute Magazine, 27 (August 1983), 38-39. Zerbst, R .H . and Brueggeman, W.B. " F H A and VA Mortgage Discount Points and Housing Prices," The Journal of Finance, XXXII (December 1977), 1766-1773. Zumpano, L.V. and Marsh, G .A . "California Mandates Disclosure in Creative Financing Arrangements: Implications for Real Estate Professionals," Real Estate Issues, 8 (Spring/Summer 1983), 15-16. 105 APPENDIX A VANCOUVER (EAST SIDE) NEIGHBOURHOOD LOCATION MAP (B.C.A.A. SUBDISTRICT //22) APPENDIX B RICHMOND NEIGHBOURHOOD LOCATION MAP (B.C.A.A. SUBDISTRICT ffi) **** Sea Island fRASEft McConachie Bridgeport Road LVancouver International Airport Westminster / Hwy. C3 Cambie Road Granville Ave Blundell Road Francis Road C3 Williams Road Steveston Hwy STEVESTON 107 APPENDIX C VANCOUVER (WEST SIDE) NEIGHBOURHOOD LOCATION MAP (B.C.A.A. SUBDISTRICT #13, 15) 108 APPENDIX D NORTH VANCOUVER NEIGHBOURHOOD LOCATION MAP (B.C.A.A. SUBDISTRICT //20) VANCOUVER 109 APPENDIX E WEST VANCOUVER NEIGHBOURHOOD LOCATION MAP (B.C.A.A. SUBDISTRICT #16-19) 110 APPENDIX F RAW DATA INPUT FORMAT Column Width Explanation of Data  1-6 Registered Title Number (L.T.O. //) 7-8 Dummy Variable for Year of Sale: 1980 = 10 1981 = 01 1982 = 00 9-14 Sale Price (Registered "Declared Market Value") 17-18 Month of Sale within the Sample Period (1-36) 19-20 Month of Sale within the Year (1-12) 22-25 Dwelling Size (Square Feet) 27-28 Age of Dwelling (Years) 30 Dummy Variable for Renovation of Dwelling 32 Dummy Variable for Swimming Pool 34 Number of Fireplaces 36 Dummy Variable for Carport 38 Dummy Variable for Garage 40 Dummy Variable for Deck 42 Number of Bedrooms 45 Number of Rooms Including Bedrooms 47 Dummy Variable for Basement 49 Dummy Variable for Double Glazed Windows 51 Dummy Variable for Basement Rental (Two Kitchens Per Dwelling) 53-55 Number of Bathrooms 57-59 Loan-to-value Ratio of all Outstanding Financing 61-68 Present Value ($) of Benefit/Discount from Creative Financing Discounted at Market Interest Rate 69 Dummy Variable for Owner-occupancy of Dwelling 71-73 Percentage of Non-english Speaking Residents by Corresponding Enumeration Areas 75-77 Average Family Size by Corresponding Enumeration Areas 79-82 Percentage of Single Family Dwellings by Corresponding Enumeration Areas I l l Column Width Explanation of Data  84-86 Percentage of Homeowner ship by Corresponding Enumeration Areas 88-95 Present Value ($) of Benefit/Discount from Creative Financing Discounted at Lagged Market Interest Rate 97 Number of Mortgages Related to the Sale 99 Types of Financing Used in the Sale: A = Conventional First Mortgage B = Conventional First and Second Mortgage C = Conventional First and B.C. Second Mortgage D = Conventional Mortgage and Assumed Mortgage E = Conventional and Vendor-Financed Mortgage F = Conventional and Assumed and Vendor-Financed Mortgage G = Conventional and Assumed and B.C. Second Mortgage H = Conventional and Vendor Financed and B.C. Second Mortgage I = Assumed Mortgage J = Vendor-Financed Mortgage K = Assumed and Vendor-Financed Mortgages L = Assumed and B.C. Second Mortgages M = Agreement-for-Sale N = Free and Clear of Financing (Cash Sale) 0 = Conventional First and Second and Assumed Mortgages 100 Classification of Financing Arrangements: 1 = Free and Clear 2 = Creative Financing Only 3 = Creative Financing as Secondary Financing 4 = Conventional Financing Only 101-103 Dummy Variable for Property Location: North Shore = 100 Vancouver (West) = 010 Richmond = 001 Vancouver (East) = 000 104 Dummy Variable for the Sale Being Creatively Financed 105 Dummy Variable for a Creatively Financed Sale Where Creative Funds are the Secondary Source of Financing APPENDIX G CONVENTIONAL RESIDENTIAL FIRST MORTGAGE RATES, 1980 Average Interest Month/Week Rate (%) January 8 13.54 15 13.52 22 13.52 29 13.52 February 5 13.58 12 13.79 19 13.88 26 13.98 March 4 14.46 11 14.52 18 14.94 25 15.54 April 1 16.00 8 16.63 15 16.85 21 17.00 28 16.50 May 5 15.88 12 14.73 16 14.02 26 13.50 June 2 13.06 9 13.08 16 13.00 23 13.04 30 13.02 Average Interest Month/Week Rate(%) July 7 13.10 14 13.10 21 13.13 28 13.15 August 5 13.25 11 13.35 18 13.40 25 13.67 September 2 13.88 8 14.04 15 14.21 22 14.29 29 14.44 October 6 14.65 14 14.75 20 14.69 27 14.69 November 3 14.73 10 14.85 17 14.90 24 14.90 December 1 15.00 8 15.27 15 15.52 22 15.77 29 15.75 S O U R C E : G . W . Gau, Faculty of Commerce, U .B .C . (NOTE: Three year term.) APPENDIX H CONVENTIONAL RESIDENTIAL FIRST MORTGAGE RATES, 1981 Average Interest Month/Week Rate (%) January 5 15.73 12 15.65 19 15.48 26 15.44 February 2 15.42 9 15.44 16 15.47 23 15.49 March 2 15.49 9 15.65 16 15.83 23 15.88 30 15.90 April 6 16.00 13 16.42 20 16.60 27 16.79 May 16.98 11 17.77 19 18.11 25 18.20 June 1 18.17 8 18.47 15 18.58 22 18.58 29 18.67 Average Interest Month/Week Rate(%) July 6 18.69 13 18.67 20 18.73 27 19.43 August 4 20.42 10 21.28 17 21.42 24 21.64 31 21.64 September 8 21.64 14 21.50 21 21.50 28 21.44 October 5 21.44 13 20.69 19 20.29 26 20.09 November 2 20.09 9 19.75 16 18.66 23 18.00 30 17.66 December 7 17.66 12 17.66 21 17.75 28 17.81 S O U R C E : G . W . Gau, Faculty of Commerce, U . B . C (NOTE: Three year term.) APPENDIX I CONVENTIONAL RESIDENTIAL FIRST MORTGAGE RATES, 1982 Average Interest Month/Week Rate (%) January ti- 17.91 l l 18.00 18 18.00 25 18.31 February 1 18.42 8 18.66 15 18.69 22 19.25 March 1 19.31 8 19.31 15 19.19 22 19.14 29 19.01 April 5 19.04 13 19.00 19 19.00 26 19.00 May 3 19.02 10 19.00 17 18.98 25 18.96 31 18.94 June 7 19.00 14 19.04 21 19.21 28 19.42 Average Interest Month/Week Rate(%) July 5 19.42 12 19.52 19 19.46 26 19.29 August 3 19.08 9 19.02 16 19.02 23 18.88 September 1 18.13 7 18.02 13 17.84 20 17.79 27 17.32 October 4 17.16 12 16.32 18 15.46 25 15.29 November 1 15.21 8 15.13 15 14.62 S O U R C E : G . W . Gau, Faculty of Commerce, U . B . C (NOTE: Three year term.) 115 APPENDIX 3 CONVENTIONAL RESIDENTIAL SECOND MORTGAGE RATES* , 1980 - 1982 1980 1981 1982 Month (%) (%) (%) January 15 - 17 17 - 18 22 - 24 February 15 - 17 16.5 - 17.25 22 - 24 March 15.25 - 17 17.25 - 18 22 - 24 Apri l 18.5 - 20 17.25 - 18 21 - 23 May 15.25 - 17 1 8 - 1 9 • 22 - 24 June 15 - 17 20 - 21 19 - 22 July 15 - 17 20 - 21 22 - 23 August 14.5 - 16.5 24 21 - 23 September 16.5 - 17.25 22 - 23 N/A October 16.5 - 17.25 23 N/A November 17 - 18 19.5 - 19.75 N/A December 16.5 - 17.25 N/A N/A The ranges indicated represent the term of the loan with the lower limit generally being a one year mortgage and the upper limit being a three to five year mortgage. S O U R C E : Prepared for the IC & I Division of the Real Estate Board of Greater Vancouver by Cumberland Realty Group Ltd . 116 APPENDIX K NEIGHBOURHOODS BY ENUMERATION AREAS Census Classification Neighbourhood Vancouver (East Side) Richmond Vancouver (West Side) Federal Electoral District 025 026 020 026 027 North Vancouver West Vancouver 002 002 Census Tract 018.01 018.02 033 034 019 030 142.02 149.02 020 022 023 027 006 007 008 009 010 119 120 122 134 135 Enumeration Area 066, 067 (17, 22, 27, 33), 068 060 - 064 151 - 155 070, 071 001 ( -33 , -34 ) , 002 (-47) 011 - 013 213 - 222 272 - 277 009 ( - 1 1 , - 1 2 ) , 010 (-29) 101 - 105 106 ( -35 , - 3 6 ) , 108 (-34), 109 (-33) 111 ( -17 , - 1 9 , - 2 0 , -21) 171 ( -75 , -76) 201 (30), 202 (29), 203, 204, 205 ( -29 , - 9 2 , - 9 3 , - 9 4 , -97) 217 ( -23 , - 2 4 , - 4 0 , -41) 121, 208 - 215 114, 115, 120 (-75) 059, 118 114 ( -37) , 115 - 117 118 - 120, 121 (101) 175 (114, 115, 121) 122 ( -1) , 169 (-51, -52, -53 , -54), 170, 171 * Unless otherwise noted, all of the blocks within each enumeration area have been included. The numbers in parenthesis indicate the blocks only to be included from a specific enumeration area or, if noted by a " - " sign, the particular blocks to exclude from an enumeration area. S O U R C E : Statistics Canada, Enumeration Area Maps. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0095930/manifest

Comment

Related Items