INFORMATION TO USERS This m anuscript has been reproduced from the microfilm master. U M I films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality o f the copy subm itted. B roken or indistinct print, colored or p o o r quality illustrations and photographs, print bleedthrough, substandard margins, and im proper alignment can adversely affect reproduction. In the unlikely event that the author did not send U M I a com plete m anuscript and there are missing pages, these will be noted. Also, if unauthorized copyright m aterial had to be removed, a note will indicate the deletion. O versize m aterials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. E ach o rig in al is also p h o to g ra p h e d in one exposure an d is in clu d ed in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. H igher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact U M I directly to order. University M icrofilm s International A Bell & Howell Information C om pany 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600 O rder N u m b e r 9 4 3 1 3 0 0 Im p acts o f th e S u b d iv isio n C on trol A c t o f 1967 on land fragm en tation in M ich ig a n ’s to w n sh ip s Norgaard, Kurt Jay, Ph.D. Michigan State University, 1994 C op y rig h t © 1 9 9 4 b y N o r g a a r d , K u r t Jay. A ll r ig h ts reserved . UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 IMPACTS OF THE SUBDIVISION CONTROL ACT OF 1967 ON LAND FRAGMENTATION IN MICHIGAN’S TOWNSHIPS By Kurt Jay Norgaard A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1994 ABSTRACT IMPACTS OF THE SUBDIVISION CONTROL ACT OF 1967 ON LAND FRAGMENTATION IN MICHIGAN’S TOWNSHIPS By Kurt Jay Norgaard Michigan has experienced an increase in the amount of land that is in unplatted lots less than 20 acres. A common perception is that the Subdivision Control Act of 1967 (SCA) has contributed in the growth in the number of 10 acre lots. By defining "subdivision" as the creation of more than 4 lots, 10 acres or less, within a ten year period, the SCA created an incentive to create unplatted 10 plus acre lots. Michigan law requires platting when a subdivision is created. Landowners desiring to create more than 4 lots and avoid platting would be required to create lots greater than 10 acres. This research examined the land fragmentation process in Michigan and the impact of the SCA on land fragmentation. The level and pattern of land fragmentation was determined by counting the number of lots, 1-19 acres in size in 72 sample townships, using plat maps for the years 1960, 1970, 1980, and 1990. Townships were grouped into three population density groups, low, medium, and high. Three hypotheses were tested to determine the rate and pattern of land fragmentation in Michigan. The results of these tests indicate: (1) that the number of unplatted lots 1-19 acres in size have increased over time; (2) the percentage increase in number of unplatted lots 1-19 acres in size are not equal across township groups; (3) the mean percentage of new unplatted lots 10-11 acres in size are not equal over time or township groups. A fourth hypothesis, stated that land division patterns are impacted by ecological, economic, social, and institutional factors but that the SCA has had no affect on the number of 10-11 acre lots. An econometric model was used to estimate four equations testing this hypothesis. In three of the four equations, the null hypothesis that the SCA had no affect was rejected. These results indicate that the SCA caused the creation of an additional 15 to 51 lots between 10-11 acres in size per township during the time period from 1960 to 1990. Given the results of this research, policy makers may want to examine the SCA to determine whether formation of a large number of 10-11 acre lots is consistent with the objectives stated in the Act. Copyright by KURT JAY NORGAARD 1994 Dedicated to my wife, Sandy, and children, Caleb and Stephanie whose encouragement, sacrifice, and love made this possible. v ACKNOWLEDGMENTS First, I would like to acknowledge my major professor, Dr. Lynn Harvey, who has contributed to this endeavor in so many ways. Secondly, my committee consisting of Dr. Lester Manderscheid, Dr. David Schweikhardt, and Mark Wyckoff were excellent. And finally, Drs. Jim Bonnen, Larry Connor, and Allen Schmid for their interesting dialogue and confidence in me. TABLE OF CONTENTS Page LIST OF TABLES ...............................................................................................................xi LIST OF F IG U R E S ..............................................................................................................xii Chapter 1 ................. 1 1.1 IN TRODUCTION..................................................................................... 1 1.2 PROBLEM STA TEM EN T....................................................................... 3 1.3 PURPOSE OF THE S T U D Y ................................................................. 5 1.4 RESEARCH O B JE C T IV E S .................................................................... 6 1.5 FORMAT OF THE STUDY ................................................................. 7 THE SUBDIVISION CONTROL ACT OF 1967 ..................... 8 2.1 INTRODUCTION..................................................................................... 8 2.2 SUBDIVISION CONTROL R E G U L A T IO N S.................................... 9 2.3 LAND DIVISION: A NEW SITUATION .......................................... 11 2.4 UNINTENDED CO N SEQ U EN CES..................................................... 17 2.5 ATTEMPTED AMENDMENTS TO THE SUBDIVISION CONTROL ACT ................................................................................... 19 INTRODUCTION AND PROBLEM DEFINITION Chapter 2 2.6 SUMMARY . ........................................................................................... 21 A MODEL OF THE FR A G M E N TA TIO N .......................... 23 3.1 INTRO DU CTIO N ..................................................................................... 23 3.2 THE LAND M A R K E T ............................................................................ 23 3.3 IMPERFECT M A R K E T S ....................................................................... 24 3.4 ECOLOGICAL DETERMINANTS ..................................................... 25 3.5 ECONOMIC D ETER M IN A N TS........................................................... 29 3.6 SOCIAL DETERMINANTS ................................................................. 33 3.7 INSTITUTIONAL DETERMINANTS ................................................ 37 3.8 CONCLUSION ........................................................................................ 38 Chapter 3 Chapter 4 HYPOTHESES AND ECONOMETRIC M O D E L .................... 39 4.1 INTRO DU CTIO N..................................................................................... 39 4.2 LAND FRAGMENTATION H Y PO TH E SIS....................................... 40 4.3 LAND FRAGMENTATION CONSISTENCY HYPOTHESIS . . . 41 4.4 CONSISTENCY OF LAND DIVISION PATTERNS H Y PO T H E SIS......................................................................................... 41 4.5 LAND DIVISION PATTERNS H Y P O T H E S IS ................................. 42 4.6 MULTIPLE REGRESSION A N A L Y S IS ............................................ 43 4.7 SUMMARY 50 ............................................................................................. viii Chapter 5 EMPIRICAL R E SU L T S........................................... 51 5.1 INTRODUCTION.................................................................................... 51 5.2 SELECTION OF S A M P L E ................................................................... 51 5.3 ASSUMPTIONS UNDERLYING STATISTICAL TESTS ............. 58 5.4 LAND DIVISION PATTERNS H Y PO T H E SIS................................. 70 5.5 SUMMARY ............................................................................................ 86 Chapter 6 CONCLUSIONS AND APPLICATIONS .......................... 87 6.1 IN TRODUCTION.................................................................................... 87 6.2 RESEARCH O B JE C T IV E S................................................................... 87 6.3 RESEARCH RESULTS ........................................................................ 88 6.4 POLICY ALTERNATIVES FOR REDUCING LAND FR A G M EN TA TIO N ............................................................................. 92 SUGGESTIONS FOR FUTURE R ESEA R C H ................................... 98 6.5 APPENDIX A Sample Townships Included in S tu d y ...................................................................99 APPENDIX B Table 1 Number of Unplatted Lots Counted in Sample Townships in 1960 .......... 100 APPENDIX B Table 2 Number of Unplatted Lots Counted in Sample Townships in 1970 .......... 101 APPENDIX B Table 3 Number of Unplatted Lots Counted in Sample Townships in 1980 .......... 102 APPENDIX B Table 4 Number of Unplatted Lots Counted in Sample Townships in 1990 ............ 103 APPENDIX C Results From F-tests Required For P o o lin g ...................................................... 104 APPENDIX D Table 1 Statistics of Variables Used in Regression Models; MED AGE and INCOME . ................................................................................105 APPENDIX D Table 2 Statistics of Variables Used in Regression Models; TOTLAB and EMP .............................................................................................. 106 APPENDIX D Table 3 Statistics of Variables Used in Regression Models; TOTPOP and TOTHOU ..................................................................................... 107 APPENDIX D Table 4 Statistics of Variables Used in Regression Models; AGP, DFREE, DCITY, REC, and LAND ...................................................... 108 APPENDIX D Table 5 Statistics of Variables Used in Regression Models; MORTNOM, MORTREAL, PGAS, CPI_NG, CPI_ALL, and FRMINC . 109 BIB LIO G R A PH Y ...............................................................................................................110 x LIST OF TABLES Page Table 5.1 Number of Townships Selected and Used in Each Population C ategory .................................................................................................... 54 Table 5.2 Number of Lots Under 20 Acres for 72 Sample T ow nships 55 Table 5.3 Number of Acres in Lots Under 20 Acres for 72 Sample Tow nships.................................................................................................. 57 Table 5.4 Percentage of Township Land Area in Unplatted Lots Under 20 A c re s .......................................................................................................... 58 Descriptive Statistics and Results of Statistical Tests for Land Fragmentation H y p o th e sis..................................................................... 61 Number of Lots under 20 Acres for 72 Sample Townships, 1960 to 1990, by Population Density Group .............................................. 62 Descriptive Statistics and Results of Statistical Tests for Land Fragmentation Consistency Hypothesis .............................................. 64 Distribution of Created Lots Under 20 Acres by Population Density Group, 1970 to 1990 ............................................................... 66 Descriptive Statistics and Results of Statistical Tests for Consistency of Land Division Patterns Hypothesis .......................... 69 Table 5.10 Variable Specification for the Land Fragmentation M o d e l............... 72 Table 5.11 Estimates of the Land Fragmentation Regression M o d e l.................. 77 Table 5.12 Estimated Range of Impact on the Number of 10-11 Acre Lots From the Land Fragmentation M o d e l ................................................. 82 Table 5.5 Table 5.6 Table 5.7 Table 5.8 Table 5.9 xi LIST OF FIGURES Page Figure 3.1 Acres of Forest by Counties in M ich ig an ............................................ 27 Figure 3.2 Acres of Water by Counties in M ic h ig a n ............................................ 29 Figure 3.3 Nominal Mortgage Rates, 1963-1990 .................................................. 31 Figure 3.4 Percentage of Michigan’s Population Over 6 5 .................................... 34 Figure 3.5 Change in Population By County, 1980-90 ......................................... 36 Figure 3.6 Possible Land Divisions Not Qualifying as "S ubdivision"............... 38 Figure 5.1 Seventy Two Townships Selected in Study S a m p le ........................... 53 Figure 5.2 Estimated Impact of the SCA on Number of 10-11Acre Lots 79 xii ... Chapter 1 INTRODUCTION AND PROBLEM DEFINITION 1.1 INTRODUCTION The State of Michigan consists of 36.1 million acres of forest land, agricultural land, wetlands, and urbanized areas. Of this total, approximately 30 million acres in 1987 were considered rural land held by others than the federal government (U.S. Bureau o f the Census, 1992). Healy and Short specify three long term trends that are observable in rural land use. Rising prices, different demands by new owners and the changing size of parcels were identified as important trends that have existed since World War II but have accelerated greatly since the late 1960’s (Healy and Short, 1981). Many factors are considered important in influencing these trends but a central issue appears to be the property rights governing land ownership. The term "property rights" is defined in this study as a person’s rights with respect to a resource. Furthermore, "in our society the existence of property rights presupposes the presence of: (1) an owner together with other persons who can be excluded from the exercise of ownership rights; (2) property objects that can be held as private or public possessions; and (3) a sovereign power that will sanction, and if 1 2 necessary protect, the property rights vested in individuals or groups" (Barlowe, 1978, pp. 395-6). Property rights associated with land ownership are continually evolving and must be considered dynamic. Land ownership typically has been defined as a bundle of rights. The individual who holds title or ownership to a parcel of land can sell, lease, mortgage. subdivide, devise and grant easements to that property. Other separable rights include air, water, mineral, and development rights. The public’s rights include the power to tax, control use (police power), escheat, and take for public use. Both private and public property rights have "rules" that characterize and define these rights. For example, the public has the right to confiscate land for public use but compensation must be given for such takings. The rules regarding land ownership have received increased attention during the past 25 years. Since 1970, the population of rural areas in Michigan has grown more rapidly than that of urban areas. This migration creates increasing pressure on traditional land uses, such as agriculture and forest management, as the demand for land for low density residential use has increased. As an indication of this demand, the total amount of farm land in Michigan has declined 30% from 1960 to 1990 (Michigan Agricultural Statistics Service, 1961, 1991) while the percent of Michigan’s population classified as rural has increased from 27% to 30% (U.S. Bureau o f the Census, 1963, 1993). This growth in population has been uneven, with some areas experiencing net losses whereas others have experienced growth. 3 1.2 PROBLEM STATEMENT Subdividing, or the division of land into multiple lots, is the result of the real estate market functioning within the existing set of property rights. Originally surveyed as large parcels suitable for agriculture or forestry, rural land is being divided into smaller lots, most commonly used for residences. Land use planners often define fragmentation as the division of large parcels into smaller unplatted parcels. Platted parcels are generally referred to as a subdivision or plat. The effects of land fragmentation have not gone unnoticed. These and related issues have been at the forefront of land use literature since the 1970’s. Although academic interest from the 1970’s regarding the significant loss of agricultural land has diminished, public attention is refocusing on this issue, along with special attention to the preservation of prime agricultural land (Misseldine, 1992). Other concerns related specifically to land divisions in rural areas are loss of open space, loss of rural character, increased cost of public services and facilities, and an increase in the number of conflicts between agricultural and nonagricultural residents (Wyckoff, 1986). As evidence of the growing concern over land use and land fragmentation, a 1992 report ranked the "absence of land use planning that considers resources and the integrity of ecosystems" among the most pressing environmental issues facing the State of Michigan (Michigan Relative Risk Analysis Project, 1992). Several of the problems mentioned in this report include: farmland loss, wildlife habitat modification, loss of open space, timber management, and urban sprawl/urban flight. While the report attributed these problems to inadequate local regulations and a lack of appropriate land use planning, some planning experts have attributed these issues in part to the state statute defining the rules regarding land division. According to an article in Planning & Zoning News, the applicable State of Michigan statute, the Subdivision Control Act of 1967 (SCA), has allowed an unknown number of parcels to be created without platting and has created an incentive to develop 10 + acre lots. Landowners desiring to avoid the cost of subdividing would create 10 + acre lots after the fourth split in 10 years. The SCA is described in detail in Chapter 2. Relatively lax platting requirements, combined with these incentives, have created opportunity for large amounts of land to be fragmented. The division of land into lots larger than 10 acres converts more acres than if the same number of lots had been created of smaller size, thus causing a greater amount of land to be fragmented (Wyckoff, 1986). Although there is some suspicion that the SCA has contributed to the fragmentation of land, the validity of this belief has not been tested. The significance of this allegation is amplified because of the nature of land fragmentation. Land developed principally for use as a residence is typically not also used for commercial forestry or agricultural production. In addition, residential use, for all practical purposes, is an irreversible choice for the life of the residence. Perhaps more importantly, an indiscriminate pattern of large lot land divisions in an area historically suited for agricultural and/or forest use reduces the viability of the entire area for resource production. This is because a few 10 + divisions: 1) take the land out of production; 2) introduce incompatible land uses into an area; 3) create an 5 uneconomic land resource management unit; and 4) tend to result in higher tax assessments on the larger parcels used for resource production (Dunford, 1979). In Michigan, the data needed to determine the scope of land fragmentation are not easily available. Records of land division, transactions, and ownership are maintained by local jurisdictions. Any attempt to aggregate and analyze the data for any significant number of jurisdictions is extremely difficult. Additional questions related to where fragmentation is occurring and at what rates are not easily answered because of the lack of congruity in the data. These questions must be answered in order to examine the impact of the Subdivision Control Act of 1967 on land fragmentation patterns. 1.3 PURPOSE OF THE STUDY The objective of this study is to analyze the impact of the Subdivision Control Act of 1967 on land fragmentation in Michigan. Public Program Evaluation (PPE) methodology is used as a framework to conduct this study. The purpose of PPE is "to measure the effects of a program against the goal it set out to accomplish as a means of contributing to subsequent decision making about the program and improving future programming" (Hatry, et al, 1981). In the first step of this methodology, the program’s or in this case, the SCA’s history, goals, and potential unanticipated consequences are described. By contrasting the SCA with the Plat Act of 1929 (Plat Act), which the SCA replaced, important elements in the SCA affecting land fragmentation are identified. The unique characteristics of the real estate market is described next. Differences between the real estate market and the perfectly competitive ideal are considered. Additionally, factors influencing the division of land in Michigan are identified. Data for these variables are given on a state-wide basis to provide a background for this study. Using this information, four hypotheses regarding fragmentation are formed. To test these hypotheses the level of fragmentation is established. The amount of fragmentation is measured by determining the total acreage of land in lots less than 20 acres. Because of the difficulty in acquiring the data, the analysis is limited to a sample of 90 townships in Michigan. An econometric model of the land division process in Michigan is developed. Specific variables are identified that represent the influential factors identified in this study. This model is then used to examine the contribution of the Subdivision Control Act of 1967 to land fragmentation in Michigan. 1.4 RESEARCH OBJECTIVES The objectives of this study are: (1) To describe the Subdivision Control Act of 1967, identify the statute’s objectives and potential unanticipated consequences; (2) To illustrate the land market, the land division process, and the factors affecting this process; (3) To form testable hypotheses regarding land divisions; (4) To develop an econometric model of the land division process in Michigan; 7 (5) To use the econometric model developed to estimate the impact of the Subdivision Control Act of 1967 on land fragmentation patterns in Michigan. 1.5 FORMAT OF THE STUDY Chapter 2 examines the Subdivision Control Act of 1967. By contrasting the SCA with the Plat Act, a shift in property rights regarding land division is highlighted. In Chapter 3, the land market and factors affecting the land division process in Michigan are described. Testable hypotheses and the methodology for testing these hypotheses are developed in Chapter 4. The levels of fragmentation discovered and the results of the analysis are presented in Chapter 5. Chapter 6 concludes the study, giving attention to future areas of research. Chapter 2 THE SUBDIVISION CONTROL ACT OF 1967 2.1 INTRODUCTION Subdivision regulation is an exercise of the police power directed to the division of land, in order to insure proper access, size and shape, safe water supply and waste disposal, and a buildable area under other related state or local laws. In so doing, subdivision regulation seeks to ensure a proper survey and recording of legal lots thereby protecting future buyers and public taxable interests. In all of these ways, subdivision regulation promotes the general health, safety and welfare of the community. Subdivision regulation differs from zoning in that zoning regulates the use of land, often maintaining the status quo until development occurs, while subdivision control regulates the way in which the land is divided and prepared for building development or other uses (Wyckoff, 1986). Development of land resources is often beneficial to communities. The division of land can promote economic growth and provide needed housing, commercial or industrial lands for the community. However, the fragmentation of land may promote urban sprawl and its associated costs {Environmental Protection Agency, 1974). In addition, the fragmentation of renewable land resources means that such land is converted 8 9 from another use such as agriculture or forestry. Loss of these renewable resources needs to be considered when determining the benefits gained by development. 2.2 SUBDIVISION CONTROL REGULATIONS Material for the next sections, which focus on the SCA, is taken generously from a law review article written by Professor Roger A. Cunningham (Cunningham, 1968) and conversations with Richard Lomax, Manager of the Subdivision Control Unit of the Michigan Department of Commerce (Lomax, Personal Communication, Nov. 1993). Legislation to assure uniform methods of recording subdivision plats was adopted in the Michigan territory in 1821. Since this original legislation, several acts were added in piecemeal fashion until 1929, when the Plat Act was enacted to unify and reorganize all laws relating to plats. In 1931, Michigan enacted the Municipal Planning Act, which was patterned after the U.S. Standard City Planning Enabling Act. The Municipal Planning Act gave municipalities the right to regulate subdivisions after forming planning commissions. However, some question remained as to whether the definitions of municipalities empowered to adopt such regulations included counties and townships. In 1945, subdivision control powers were eliminated from county planning commissions by the enactment of the County Planning Commission Act. This Act provided that the powers exercised by all county planning commissions should "be those specified for . . . county commissions in the terms of the Act," and since the County Planning Commission Act conferred no power on county planning commissions to regulate subdivision, subdivision control powers were therefore eliminated (Cunningham, 10 1967). Questions about the authority of townships were resolved in 1952 when an amendment was enacted to redefine the term "municipality" to include townships, charter townships, cities, villages, and other incorporated political subdivisions. To clarify the role of townships, the legislature adopted the Township Planning Act in 1959. This Act authorized townships to "make and adopt a basic plan as a guide for the development of unincorporated portions of the township" and to perform certain advisory functions concerning land subdivision regulations. In 1967, the Michigan legislature passed the Subdivision Control Act (SCA), which repealed and replaced the Plat Act of 1929 (Plat Act). The SCA, like the Plat Act it replaced, did not make any references to the Municipal Planning Act or the Township Planning Act. In summary, Michigan has three separate subdivision control statutes, none of which makes reference to the others. The Municipal Planning Act and the Township Planning Act are primarily enabling Acts while the SCA is largely mandatory (Cunningham, 1967). Cities, villages and townships in Michigan have the legal authority to govern all land divisions under these statutes. However, according to a survey completed in 1978 only 233 townships and 196 cities and villages had any type of subdivision regulation (Michigan Planning & Zoning Survey, Office of Intergovernmental Relations, State Department of Management & Budget, 1979 quoted in Planning & Zoning News, January 1986, p. 14). 11 2.3 LAND DIVISION: A NEW SITUATION Introduction The Subdivision Control Act of 1967 (SCA) replaced the Plat Act to provide the minimal mandatory requirements of subdividing. The passing of the SCA by the state legislature inaugurated a new era for dividing land in Michigan. From a historical perspective, one can see that the new law in unison with other changes around that time, created a fundamentally different set of property rights for landowners regarding land division. Subdivision Control Act versus the Plat Act The SCA and the Plat Act, were similar in that both ordinances regulated the land division process in Michigan. In the next section, these laws are contrasted to highlight differences regarding subdivision definitions and other aspects of platting. The division of land into smaller lots is commonly called the subdivision process. For purposes of this study, "subdividing" or "subdivision" is defined as the process of land division meeting certain criteria as described in the state statute or local subdivision control ordinance. When a land division qualifies as a subdivision, the legal requirements are quite different from a division of land that is not a subdivision. Several sections taken from the SCA and the Plat Act are cited below to display how they differ in defining subdivision. Articles written about the SCA have consistently emphasized that this definition is important (Cunningham, 1967; Wyckoff, 1986). 12 Michigan Compiled Laws Annotated (MCLA) Section 560.103 of the SCA, titled "Subdivision of land; surveys and plats, when required" addresses the question of when platting is required. Part (1) of Section 560.103 states: "Any division of land which results in a subdivision as defined in section 102 shall be surveyed and a plat thereof submitted, approved and recorded as required by the provisions of this A ct." Therefore, to know when platting is required, the definition of subdivision as stated in Section 102 must be known. Subdivision from Section 560.102 is defined as follows: "Subdivide" or "subdivision" means the partitioning or dividing of a parcel or tract of land by the proprietor thereof or by his heirs, executors, administrators, legal representatives, successors, or assigns for the purpose of sale, or lease of more than one year, or of building development, where the act of division creates 5 or more parcels of land each of which is 10 acres or less in area; or 5 or more parcels of land each of which is 10 acres or less in area are created by successive divisions within a period of 10 years." This language of the 1967 Subdivision Control Act should be contrasted with that of the 1929 Plat Act. Section 3 of the Plat Act states: "Any proprietor who shall hereafter subdivide any lands shall make and record a plat thereof in accordance with the provisions of this Act and said plat shall be made, approved, filed, recorded, altered and vacated in the manner hereinafter provided." Therefore, to know when platting is required under the Plat Act, the definition of "subdivision" as stated in Section 2 must be known. Section 2 of the Plat Act defines subdivide: "The word "subdivide," when used in this Act, shall mean the partitioning or dividing of a lot, tract or parcel of land into 5 or more lots tracts or parcels of land: provided, however, that this limitation shall not apply to the partitioning or 13 dividing of agricultural lands into tracts or parcels of land 10 acres or more in area for continuing agricultural use." In summary, once landowners have divided their land in such a manner as defined as a subdivision, the platting process mandated by the statute must be followed. Landowners wanting to avoid this process must divide their land in a way that avoids the definition of subdivision. In comparing the two Acts, both define subdividing as the dividing of a lot into five or more lots. In the Plat Act, lots of 10 acres or more in size are not to be included if they are for continuing agricultural use. The definition of subdivision in the SCA excludes parcels created greater than 10 acres. The result would be that 10 acre parcels created would not be considered as a lot under the Plat Act but are under the SCA. The difference between "greater than 10 acres" or "10 acres or greater" is minor. The major issue is the specific size of the lot required to be excluded. For example, lots that are 10.1 acres would not be considered as a lot when counting for platting requirements under the SCA. Another difference between the definitions of subdivision was the addition of the 10 year period clause in the SCA’s definition. This 10 year period applies to the new lots created and the original piece of land. Under the SCA, 10 years after a lot is created, a landowner can split the lot into a maximum of 4 lots under 10 acres without platting. In effect, the counting of the number of lots created 10 acres or less is reset on each parcel created 10 years after the split. The number of lots on the original or parent parcel is counted as a running total. Ten years after a split of a lot under 10 14 acres, that split is not counted as a division according to the SCA. Therefore, another lot under 10 acres could be created and platting would not be required. No time period was given in the Plat Act. A fifth creation of a lot smaller than 10 acres of the original parent parcel would require platting. The 10 year time period in the SCA lawfully allows land to be split into more parcels without platting than under the Plat Act. The SCA also had additional language describing the landowner(s). "Proprietor" in the Plat Act was expanded to "proprietor thereof or by his heirs, executors, administrators, legal representatives, successors." This additional description is necessary because the vague wording of "proprietor" in the Plat Act "The opportunity for avoidance of the Plat Act requirements is obvious when one considers the possibility that the original proprietor may convey each of his first four lots to relatives, or to corporations controlled by him especially formed for the purpose" (Cunningham, 1967, p. 54). A 1966 study concluded that the imprecise definition of subdivision in Section 2 of the Plat Act make it "virtually unenforceable and that consequently there have been thousands of subdivision of land without recording of plats, with resulting serious problems to the community, road commissions and fire departments because of roads too narrow or too poorly designed to permit entry of snow plows and fire vehicles and inadequate storm water and sanitary drainage" {Bureau o f Local Government Services, 1967?). Research on legal records did not find any convictions regarding violations of the Plat Act (Cunningham, 1967). 15 The SCA was different from the Plat Act in several areas that ultimately established a new situation for land divisions in Michigan. In some cases these differences raised the cost of platting and in other instances may have prevented platting. In the SCA, the county road commission was given greater authority over final plat approval. Section 183 in the SCA states: " . . . for all highways streets and alleys in its jurisdiction or to come under its jurisdiction and also for all private roads in unincorporated areas to require . . ." and a list of specifications for roads followed. This is important because the county road commissions in the state were updating their road specifications at approximately the same time as the SCA was enacted. This may have significantly increased the cost of platting depending on the new specifications. With the increased cost, landowners may choose to avoid platting and these higher costs, and create lots in a manner that do not qualify as a subdivision. Before the SCA was enacted, landowners would go through the platting process. With the new rules they would no longer plat, but rather create 10 + acre lots after the creation of four lots in at 10 year period. Another difference was the SCA included definitions of "land suitability." In the Plat Act the interpretation of land suitability was at the discretion of the individual municipality. Section 105 of the SCA states: "By conditioning approval of both preliminary and final plats upon compliance with rules of the water resources commission of the state department of conservation adopted for the determination and establishment of floodplain areas of rivers, streams, creeks or lakes, . . . as published in the state administrative code" and "with rules of the department of public health as published in the state administrative code relating to suitability of soils for subdivisions not served by public water and public sewers." 16 This defining of land suitability by the SCA gave precise guidelines for acceptable land for subdivisions. This standard definition could decrease the amount of land available to be subdivided. Thus, owners of land, unsuitable for subdivisions, would divide their land in a manner that does not qualify as a subdivision. The enactment of the SCA provided a new start for local municipalities regarding subdivision control. By exempting all land divisions up to January 1, 1968, enforcement could begin with a clean slate. Moreover, the Survey Recording Act of 1970 required that all new lots created be surveyed. By surveying all new lots, records would become more accurate and small lots disguised as large lots would be eliminated. Also, in this period of time land use regulation at the township level experienced great growth. Some estimates have the percent of townships with land use regulations increasing from 9% in 1965 to 63% ten years later (Lomax, Personal Communication, Nov. 1993). This large expansion in number of townships with land use regulations is evidence of increased public concerns regarding private use of land. 17 2.4 UNINTENDED CONSEQUENCES OF SUBDIVISION CONTROL ACT The objectives of the SCA were described as follows: "AN ACT to regulate the subdivision of land to promote the public health, safety and general welfare, to further the orderly layout and use of land; to require that the land be suitable for building sites and public improvements and that there be adequate drainage thereof; to provide for proper ingress and egress to lots; to promote proper surveying and monumenting of land subdivided and conveyed by accurate legal descriptions; to provide for the approvals to be obtained by subdividers prior to the recording and filing of plats; to establish the procedure for vacating, correcting and revising plats; to control residential building development within floodplain areas; to provide for reserving easements for utilities in vacated streets and alleys; to provide for the filing of amended plats; to provide for the making of assessors plats; to provide penalties for the violation of the provision of this Act and to repeal certain Acts and parts of Acts." The objectives of this Act can be divided into two parts. First, the reason for having a subdivision control act was "To promote the public health, safety and general welfare, to further the orderly layout and use of land." And, the second part of the objective states in general terms the means on how the SCA will facilitate the reasons given above. The bulk of the 193 sections of the Act are the specific provisions designed to meet these objectives. The common perception is that in many areas the cost of complying with the SCA is great enough to cause a number of landowners to avoid platting by creating lots of 10 + acres after the fourth split under 10 acres in a 10 year period. In the language of Public Program Evaluation, this would be the unanticipated consequence of the Subdivision Control Act. As stated earlier, a landowner has the right to divide and sell property, conditional to the rules applicable to these rights. Many reasons exist why individuals might choose 18 to sell, but it is assumed that, when selling all or a portion of land, landowners would desire to maximize their returns on the sale. The returns would be positively impacted by the number of lots sold and by the price of the lots while the costs associated with land division would negatively affect total returns. Another assumption is that the lots are primarily sold for use as residential building sites. Usually local regulations specify minimum lot size and minimum road frontage required. For landowners to maximize their number of lots, they would divide their land at the minimum road frontage and size until all the land was divided. With the definition of subdivision as specified in the SCA, landowners would incur the costs of platting once more than four lots 10 acres or under were created in a 10 year period. If lots could be sold at a high enough price to justify incurring the costs of platting, 10 + acre lots possibly would not be created. However, in many areas the demand and price of the lots is not of a magnitude and landowners would not incur the costs to plat and would only create lots greater than 10 acres after the fourth split under 10 acres. To maximize the number of lots with the required road frontage, the lots could be long narrow lots with the minimum amount of road frontage required. The result can be seen on plat maps as long narrow lots, sometimes called "bowling alley lots." The Michigan Department of Natural Resources, Office of Land Use, published a working report in 1974, seven years after the enactment of the SCA. A major conclusion in this draft was that a large number of 10 + acre lots had been created in order to circumvent the Act. This report noted that this creation of large lots had led to "increased service costs to local units, accelerated and often poor development of 19 northern resources, and loss o f agricultural and forestry land in manageable size tracts (40 acres or more)." Included in this draft were recommendations to enlarge the scope of the definition of subdivision from 10 to 40 acres (Michigan Department o f Natural Resources, 1974?). 2.5 ATTEMPTED AMENDMENTS TO THE SUBDIVISION CONTROL ACT Since the SCA became law in 1968, several attempts have been made to revise this Act. In a few cases, the proposed bills were extensive, covering many areas of the Act. In the following paragraphs, portions of these amendments that would have had an impact on land fragmentation are highlighted. House Bill No. 4151 was introduced in February 1977. The first amendment of this bill would have changed the section on definitions. The most notable change was the definition of "subdivide" or "subdivision." The 10 acre minimum was changed to a 2 acre minimum and the 10 year period clause was dropped. An indication of the significance of this definition change can be gained from an analysis of the bill by Department of Treasury. This report contrasted the positive and negative aspects of the bill and also included some suggested amendments. One positive aspect of the bill identified was that by establishing a minimum width of 165 feet for lots, this would require that many 10 + acre parcels of width of 360 feet could only be divided into two lots without platting. A negative aspect of this bill identified that failing to require the 10-year limit would lead to successive divisions that could not be controlled. This uncontrollable 20 division could result in lengthy descriptions and cumbersome tax rolls. To correct this the Department of Treasury recommended that all divisions less than 10 acres require that a certified survey be presented to the Office of Register of Deeds (Michigan Department o f Treasury, 1977). This would correct the lengthy descriptions and cumbersome tax rolls but would still fail to control the number of successive divisions. In 1985, another attempt was made by the legislature to revise the SCA. House Bill 5152, or the "Land Division Control Act," was a major proposed revision covering 105 of the 118 sections. The revisions pertinent to this study are discussed below. As in the prior revision attempt, this bill would have changed the definition of subdivision. Wyckoff (1986) felt that while the new definition of subdivision would remove the incentives to create 10 + acre lots, the nature of land divisions would not be significantly altered. The shape of lots might be changed by the addition of maximum length to width ratio included in the provision but probably the general pattern of development would not have varied much, although a local review process would have been required. A 40 acre parcel could be spilt into either 4 or 9 lots depending on whether the community had zoning and/or subdivision regulations. The fact that property owners desire to maximize their returns when dividing land could have had a substantial impact on this process. This proposal created much discussion about the purposes of the SCA. Some insight can be gained from an article in Planning & Zoning News of January 1986. Several individuals involved with this amendment were interviewed in this article. A number of questions focused on the incentives to create the 10 + acre lots. While some 21 individuals thought this issue should be resolved by changing of the SCA, others said this was a subdivision control law rather than a land preservation law and that land fragmentation issues should be addressed by a separate land preservation law (Wyckoff, 1986). House Bill 5152, discussed and reviewed for several years, did not pass. Richard Lomax, Manager of the Subdivision Control Unit of the Michigan Department of Commerce, suggested that the impetus behind the bill was the building industry. By changing the SCA, the cost of doing business would decrease thus stimulating the industry. However, after economic growth accelerated and the construction business recovered, pressures for changing the SCA diminished (Lomax, personal communication, Nov. 1993). 2.6 SUMMARY The Subdivision Control Act of 1967 replaced the Plat Act of 1929 as the minimum state mandate for regulating the division of land in Michigan. When contrasted to the Plat Act, the SCA’s definition of subdivision had changed. The common perception is that the SCA is responsible for defining subdivision in part as more than four lots that are 10 acres and under which hypothetically sets up the incentive to create 10 + acre lots. This research has shown that the change in definition regarding the cutoff point at 10 acres was minor. Subdivision as defined in the Plat Act was vague and, as a result, was not being enforced. The SCA created a more precise definition and also included a 10 year redivision provision. Several additions in the SCA raised the cost 22 of platting. These changes in the SCA initiated a greater level of enforcement. More stringent enforcement and greater platting costs could impact the pattern of land divisions. In particular, the number of 10 + acre lots could increase as a result of landowners avoiding qualifying as a subdivision. Several attempts to revise the Act have been unsuccessful. One of the major points of contention is the definition of subdivision. Incentive to avoid the classification of a subdivision and thus avert the platting process exist because of the current definition of subdivision in the SCA. In doing so, many of the lots being created are larger than 10 acres and are causing more land to be fragmented than if the lots were smaller. The exact amount of land included in these 10 + acre lots is unknown because available land use records are difficult to aggregate for these types of questions. However, perusing plat maps at different points in time provide some indication of the number of 10 + acre lots. The SCA is only one of many possible factors that could affect the pattern of lot divisions and land fragmentation. demand for lots are investigated. In the next chapter, factors affecting supply and Chapter 3 A MODEL OF THE FRAGMENTATION PROCESS 3.1 INTRODUCTION Land economists generally describe the land market as having four dimensions: quantity, price, spatial location, and time. At any point in time the quantity, price, and location will both independently and interdependently be affected by certain factors. Barlowe groups these factors into three categories: ecological, economical and socialinstitutional (Barlowe, 1978). With the emphasis of this study on the Subdivision Control Act of 1967, the social-institutional group will be separated into two sections. These factors (ecological, economical, social, and institutional) can influence both the supply and demand for lots in rural areas. In this chapter, these factors are examined as to their impacts on land fragmentation. Chapter 2 stated that some believe that the SCA has caused an increase in land fragmentation. The SCA and other factors will be discussed in this chapter and their appropriateness in the model to be considered. Characteristics of the land market will be addressed before these factors are examined. Some knowledge o f the land market provides a background for the model. 3.2 THE LAND MARKET 23 24 Healy describes four principal ways in which land is transferred from one owner to another: (1) non-market transfers such as a gift or inheritance; (2) sale by private agreement; (3) sale through a real estate broker; and (4) sale at an auction by verbal or sealed bid. The volume of each type of transfer is difficult to determine, but a 1978 survey found that 62% of the land had been acquired by purchase from a non-relative (USDA Landownership Survey, 1978). Therefore, a large percentage of transfers are executed through the market. Questions about the financing of land purchases were also asked on the 1978 survey. Only 17% of the acreage transferred from 1975 to 1977 was an all cash transaction. The balance of the land was purchased with borrowed capital. 3.3 IMPERFECT MARKETS The land market differs from the perfectly competitive ideal for a number of reasons. First, land is a fixed location. The immobility of land could restrict the number of potential buyers causing the price to be less than in a perfectly competitive market. Potential buyers outside the geographical area would not bid on the land. Also because land is in a fixed place the location of the property could have a profound effect on its value. Second, land is not homogeneous. Location, land uses permitted, soil type, and quality of structures are variables that affect the value of land. Potentially great differences between individual lots explain why information is important for the real 25 estate market. In those cases where the buyer or seller does not have adequate information, real estate agencies can provide that service for a fee. Another reason that land differs from other goods in a perfectly competitive ideal is that land is usually sold in relatively expensive units. This high cost of land could also limit the number of potential buyers. The availability of financing would also effect the number of potential buyers. Land is subject to certain transfer costs that may not pertain to other commodities. A longer time frame for settlement, fees for title searches, legal fees and brokerage costs all combine to create relatively large transaction costs. These costs could reduce the number of land transfers. These differences from a perfectly competitive ideal make the land market complicated to study. Given the dissimilarities between lots it is difficult to measure price behavior by inspecting sales records alone. Also sales and ownership records are kept at the local level. Thus, information is difficult to aggregate since the data may be maintained differently at each locale (Healy and Short, 1981). In this study, the factors affecting the land market were grouped into four categories. In the next several sections these factors are discussed as to how they affect land fragmentation. 3.4 ECOLOGICAL DETERMINANTS The physical factors of land can effect both the supply and demand of land for housing. Some natural amenities are appealing for residential use, whereas others are 26 limiting either because of physical constraints or laws regarding the use of those amenities. For instance, in Chapter 2, definitions of land suitability in the SCA were discussed. Soil Types Soil types can limit the supply of land for lots. In rural areas, septic systems are the predominant means for waste disposal. Not all soils are conducive to a septic system, and therefore, standards specify which types of soils can support septic fields. Land with inadequate soils for septic fields are not suitable for residence lots without the additional cost of an engineered field. Land suitable for agriculture is often acceptable for septic fields. Consequently, farm land is often in demand for use as residential lots. Additional knowledge about ecosystems can have an impact on land use. In the past, wetlands might have been filled or drained for other uses but are now being protected from development by various statutes. Thus, wetlands are eliminated as potential building sites, which reduces the potential amount of land available to be developed. Forests Approximately 51 percent of Michigan’s land area is covered with forests. Of the total 18.4 million acres, 95 percent of this total is commercial forest (Source: USDA, Resources Bulletins NC-60,62,64,66, 1982). The percent of land in forest is shown on a county basis in Figure 3.1. 27 Figure 3.1 Acres of Forest by Counties in Michigan Acres of Forest 1 2 0 to 77 Source: USDA, Resources Bulletins NC-60,62,64,66 The physical characteristics of forest land can play an important role in the desirability of a specific location for residence use. Generally the cost of the developing of forest land is higher because clearing of trees increases the cost of the lots. However, the mature trees of a forest may be more appealing than land without such amenities. 28 Also, a significant portion of Michigan’s wetlands are forested. These lands are generally unsuited for building construction. Lakes Access to lakes and rivers might also affect the demand for land for lots. Michigan has 11,037 inland lakes and 36,350 miles of rivers. Figure 3.2 displays the number of acres of water per county. The close proximity to water conceivably makes some areas more desirable than others. 29 Figure 3.2 Acres of Water by Counties in Michigan Source: USDA, Resources Bulletins NC-60,62,64,66 3.5 ECONOMIC DETERMINANTS Income The demand for lots is affected by the economic ability of the prospective buyers. Over the last 30 years the average per capita income in real terms in Michigan has 30 increased from $7,902 to $14,061 (U.S. Bureau o f the Census, 1993). This increase in per capita income could be a contributing factor of having larger residential lots. Results from the USDA 1978 Land Ownership Survey indicate that a large percentage of the lots are purchased with borrowed capital. Proof of employment and sufficient income is often required to qualify for financing. The increasing per capita income could be an argument for a greater demand for lots. Interest Costs The cost of borrowed capital, or interest cost, could affect the demand for lots. In years of high interest rates, demand would be impacted negatively and in years with low rates, positively affected. Nominal mortgage rates from 1963 to 1990 are shown in Figure 3.3. 31 Figure 3.3 Nominal Mortgage Rates, 1963-1990 C/3 & CO cr -»(/) — < CD k_ 0 ■4—> c 4 i i i-1 f- - i—i— i t i i i i i + - h - i i i i i i i i i i 1963 1967 1971 1975 1979 1983 1987 Years Source: U.S. Department of Commerce, Business Statistics, 1963-91 Mortgage rates rose dramatically in the late 1970’s early 1980’s and then dropped nearly as dramatically in the late 1980’s. While the relatively high interest rates might negatively affect demand, supply of land for lots might increase because of the high cost of holding land. These two separate effects make the impacts of changing interest rates difficult to determine. 32 Agricultural Income Another factor affecting the supply of land for lots would be income from the agricultural use of land. Present and future income from owning the land would be contrasted with profits from selling immediately. Ability to capture greater economic rents by dividing their land may induce property owners to divide land. Though highly variable, total net farm income in Michigan has been generally decreasing in real terms the last 30 years. This may encourage farmers to divide their land and sell lots for residential use. This selling of lots could have negative spill-over effects on agriculture. New construction in the area could raise the state equalized values on all property, thus raising property taxes. Increasing property taxes are blamed as part of the decreasing farm income (Dunford, 1979). The possibility of future land division can influence agriculture. Both the uses and prices of land could be affected. Land with development potential would be in high demand by speculators which would raise prices. Farm land that is being held awaiting development could be left to lay idle. Costs of Travel This section identifies various factors related to costs of travel. The topic "costs of travel" is interpreted broadly to include time, money and convenience of travel. Demand for land for residence use would be inversely related to distance from an urban area, taking into account items related to travel time. Such items as distance from a major city, accessibility to a freeway and type of roads could substantially change 33 travel time. Travel time to work, shopping, and other activities would be weighed by potential buyers when choosing where to live. Those areas further away and not having access to a freeway would be less desirable than those closer to a freeway (Ronald Briggs, 1983). The cost of gasoline is also an important factor related to costs of travel. With little public transportation in rural areas, the main form of transportation is the automobile and the cost of gasoline is a major component in the cost of driving. 3.6 SOCIAL DETERMINANTS Non-economic factors can also effect the supply and demand for lots. A survey of farmland owners in Vermont who sold land listed the top three reasons in this order: health, age and "received a good offer for the land." Divorce and death are other common reasons for selling o f land (Bancroft et al, 1977). As the survey indicated, an aging landowner population could increase the supply of land. Desires to capitalize on their investment in land would prompt landowners to sell some or all of their land. Another impact of the aging population would be the increase in the number of retirees. Figure 3.4 displays the percent of the population over 65 years of age. As the number of retirees increases, residences in rural areas might become more appealing because of lower housing costs, slower life style and not having to commute to work. In addition, an aging population may mean that a greater percent of the population are prospective home buyers. 34 Figure 3.4 Percentage of Michigan’s Population Over 65 Years Source: U.S. Bureau o f the Census, 1961, 1991 If the demand for lots is affected by the number of prospective buyers, changes in population and age of the population could have an impact on demand. Areas of substantial population growth would have relatively greater demands for lots. Some have attributed the increase in rural populations to the expanded employment opportunities. New technology and economic development strategies have enticed employers in new areas, thus creating jobs. 35 Michigan’s average annual rate of population growth since 1960 has been approximately 3% (U.S. Bureau o f the Census, 1990), but this low growth can be misleading, Some areas have experienced tremendous growth in contrast to other areas that experienced net loss in population. Figure 3.5 displays population growth per county from 1980 to 1990. The areas with the largest percentage increases in population are not the major urban areas. 36 Figure 3.5 Change in Population By County, 1980-90 P e rc e n t C h a n g e □ f""j -9.7% to —13.5% 1% to -7.8% 0.0% to 9.7% H I 10.1% to 19.3% ■ 20.8% to 29.5% Source: U.S. Bureau o f the Census, 1961, 1991 Another phenomena of Michigan’s population is the decrease in the average size of households. From 1960 to 1990 the average size of household has dropped from 3.42 to 2.66 (U.S. Bureau o f the Census, 1961, 1991). Noting that households are the principal buyers of houses, this decrease in average size of households could reflect a greater demand for lots because of a greater number of households, populations being equal. 37 3.7 INSTITUTIONAL DETERMINANTS Institutional determinants can be defined as any application of the powers of government, (i.e., police, taxation, spending, proprietary, and eminent domain powers to direct land use). These are designed to enhance the welfare of the public in those instances where market forces are insufficient to bring about acceptable land use patterns. Some form of institutional restrictions over land use have been employed in many municipalities for nearly a century. However, in many rural areas of Michigan, the land use restrictions are spotty and weak (Michigan Planning & Zoning Survey, Office of Intergovernmental Relations, State Department of Management & Budget, 1979 quoted in Planning & Zoning News, January 1986 p. 14). Local zoning powers can affect the resultant lot size of land divisions. Requirements of minimum road frontage, a lot’s length to width ratios, and minimum lot size, could be included in a local zoning ordinance. These requirements would affect the maximum number of lots available for creation from a given parcel. Land divisions in Michigan are regulated by the Subdivision Control Act of 1967 (SCA). Under this Act, a landowner would incur the costs of platting after five lots of 10 acres or less are created in a 10 year period. It is hypothesized that with these rules in place many landowners would not incur the costs associated with platting and would create lots greater than 10 acres after the fourth division. This behavior, multiplied many times, could have a significant impact on the sizes and patterns of lots in Michigan. An example of how a 40 acre parcel could be divided and not platted is given in Figure 3.6. 38 Figure 3.6 Possible Land Divisions Not Qualifying as "Subdivision" Source: Author 3.8 CONCLUSION This chapter described the land market and its unique characteristics. A number of factors affecting the land market were presented and discussed. These factors and their impact on the demand and/or supply of lots were presented. In the next chapter, variables representing these factors will be identified. These variables will then be used to create an econometric model to examine the impact of SCA on land fragmentation in Michigan. Chapter 4 HYPOTHESES AND ECONOMETRIC MODEL 4.1 INTRODUCTION Chapter 3 identified factors affecting the land division process in Michigan. These factors were arranged into ecological, economical, social, and institutional categories. Supported by this background analysis, several hypotheses and appropriate general statistical models to test these hypotheses are formulated in this chapter. This study examines four major hypotheses. These hypotheses build on one another, culminating with the fourth hypothesis. The first two hypotheses focus on the subject of land fragmentation. The third hypothesis concentrates on patterns of land division over time and the fourth examines the impact of the Subdivision Control Act of 1967 on land division patterns. In Michigan, all land area is either in a city or a township. The 83 counties are divided into 1,241 townships and 273 cities. The township is chosen as the unit of study in this analysis for the following reasons: (1) most of the land area in Michigan is in townships; (2) a majority of unplatted land divisions are in townships; (3) data on unplatted land division are in plat maps on a township basis; (4) concerns about land fragmentation focus on land in townships; (5) townships, as compared to cities, are less likely to have local subdivision control laws; (6) townships are the smallest units for 39 40 many data sources; and (7) the use of townships allows for definitive locational and descriptive factors. The summary in Chapter 3 stated that the factors affecting the process of land fragmentation was suspected as being inconsistent throughout Michigan. An attempt was made to create homogeneous groups by utilizing population density. Townships were sorted by use of population density and then arranged into three groups. Chapter 5 further describes the method used. These groups are used in the following hypotheses. 4.2 LAND FRAGMENTATION HYPOTHESIS The first question to be answered is whether land fragmentation is occurring. After its resolution, the next question must explore whether the rate of land fragmentation is consistent over time. For this study, the levels of land fragmentation is measured by the number of lots less than 20 acres. The first hypothesis is in two parts. Part A of the Land Fragmentation Hypothesis states: "The mean number o f unplatted lots under 20 acres in size is equal over time. " and if A is rejected then Part B states: "The mean rate o f change in number o f unplatted lots under 20 acres in size is equal over tim e." This hypothesis implies that the mean number of lot divisions is not statistically different over time. The rejection of this hypothesis leads to questions of where land fragmentation is occurring. Questions in regard to spatial considerations are addressed by the next hypothesis. 41 4.3 LAND FRAGMENTATION CONSISTENCY HYPOTHESIS If land fragmentation is occurring, the next question would ask where fragmentation is taking place. The Land Fragmentation Consistency Hypothesis is: "For any decade, the mean change in the number o f unplatted lots less than 20 acres in size is equal across township groups." In order to test this hypothesis the townships must be grouped in some manner. The method of township grouping is addressed in a later section titled "Selection of Sample". In discussing land fragmentation it was hypothesized that certain patterns were the result of the Subdivision Control Act. Lot division pattern questions are addressed in the third hypothesis. 4.4 CONSISTENCY OF LAND DIVISION PATTERNS HYPOTHESIS In Chapter 2, differences between the SC A and the Plat Act were examined and, as a result, several conclusions were reached. First, upon enactment of the SCA there was a fundamental change in the rules regarding land divisions. Second, wording in the SCA in combination with economic impetus created incentives to encourage landowners to create 10 + acre lots. Because the SCA was implemented January 1, 1968, land division patterns after this time were contrasted to land division patterns before 1968 to demonstrate whether any changes in land division patterns existed. The Consistency of Land Division Patterns Hypothesis is in two parts. Part A of the hypothesis states: "The mean percentage o f all new unplatted lots that are 10-11 acres in size is equal over time. * and, if Part A is rejected, Part B would be: 42 "The mean percentage o f all new unplatted lots that are 10-11 acres is equal across township groups." By testing the first three hypotheses, the foundation is built for the hypothesis on factors affecting land division patterns. 4.5 LAND DIVISION PATTERNS HYPOTHESIS In Chapter 3, factors were identified as potentially affecting land division patterns. The SCA was highlighted as a possible contributor to the number of lots 10 + acres. The Land Division Patterns Hypothesis states: "Land division patterns are impacted by ecological, economical, social, and institutional factors. However, the Subdivision Control Act o f 1967 has had no effect on the number o f 10 + acre lots. " This hypothesis is the nucleus of this study. The first three hypotheses are used as building blocks for the fourth hypothesis. By using time trend analysis to test these hypotheses, land fragmentation trends and patterns can be identified. However, factors affecting land fragmentation can not be established. Multiple regression analysis will be used to test the fourth hypothesis, which will account for the factors identified as affecting land division patterns. 43 4.6 MULTIPLE REGRESSION ANALYSIS The Land Division Pattern Hypothesis focused on the impact of one of multiple factors affecting the pattern of land division patterns. The least squares multiple regression analysis will be used to account for these other factors. The General Model The General Statistical Model is a functional relationship between the number of 10 + acre lots and the factors identified in Chapter 3. This model can be specified in the general form: Qit f (Ejit) Sm, ECjit, Inut* eit) Qit The number of lots 10 + acres in township i in year t; Eju A set of n ecological factors (j = 1 ...n) affecting the number of where: lots in township i in year t; A set of p social factors (k = 1 .. .p) affecting the number of lots 3kit in township i in year t; E ca = A set of q economic factors (1 = 1 ...q) affecting the number of lots in township i in year t; Imit A set of r institutional factors (m = 1 ...r) affecting the number of lots in township i in year t; An error term. 44 Variable Specification The factors affecting land division, examined in Chapter 3, will be used to specify the variables included in the model. A description of the data used to construct each variable is given. With the SCA enacted January 1, 1968, the period of study is 1960 to 1990 with four distinct time periods. The years 1960, 1970, 1980, and 1990 are used because of the availability of data from the U.S. Bureau of the Census. For a few variables, the value at that one point in time (i.e., 1960), would not adequately demonstrate the effects of the factor over time. For those variables, the values over the four years previous and that year are averaged to create a proxy for that factor. For example, the interest rate of home mortgages in 1960 would be an average of the home mortgages from 1956 to 1960. Number of Lots The dependent variable in this model is a description of the pattern of development believed to be caused by the SCA. The fourth hypothesis implied that the SCA has affected the number of lots 10 + acres. Data for this variable is obtained by counting lots using plat maps (Rockford Map Publishers, Inc, Rockford, Illinois). These plat maps are updated approximately every three years. For each time period desired, the plat map that most closely met the date needed was used. For example, plats for 1970 would come from a 1971 plat map rather than a 1968 plat map. In this model, the number of lots 10-11 acres is used as a representative of this pattern of 10 + acres. Fractions on most plat maps are commonly reported as whole 45 numbers. Therefore, to capture the number of lots 10.1 to 11.0 acres, 10 acre lots must be counted. Lots that are truly 10 acres would not be the pattern of development sought, however, this number is assumed to be small. Distinguishing Between Groups The townships were divided into three groups to account for differing impacts from the factors affecting land division. To reflect this in the model, two dummy variables were created to distinguish among the different groups. These dummy variables were of value "1" when it was the group identified and "0" otherwise. For example, variable D1 for a township equaled "1" when it was a low density townships, and equaled "0" when the township was a high density township. Ecological Determinants All variables in this section were township specific but not time specific. To construct variables specifying the ecological determinants, data from the Michigan Resource Information System (MIRIS) published by the Land and Water Management Division, Department of Natural Resources was used. These data were compiled from 1978 aerial photography, therefore information on a township would be how it appeared in 1978. Ideally, information on land types in 1960 would be used. However, since 1960 data is not available and the 1978 data provides fne detail needed, this is the best choice. 46 The ecological factors stated as affecting the demand for lots because of recreational amenities were the presence of lakes, rivers and forest in a particular area. From MIRIS, the acreage of all types of forest and all water for each township was obtained. Summing these values and dividing by the number of total acres in the township created a variable that represented the percent of the township that was considered recreational. The supply of lots was affected by the acreage of agricultural land and soil types. The number of acres of agricultural land from MIRIS was divided by the number of total acres in the township to create a variable representing the percent of the township in agriculture. The acreage of land with suitable soil types for residence use was not available, therefore, a variable for soil type was not included. Economic Determinants Several factors were identified as economic determinants. Economic ability of prospective buyers was measured by two different items, (1) per capita income and (2) employment figures. Data from the U.S. Bureau of the Census for years 1960, 1970, 1980, and 1990 were used. Some data was not available on a township basis for 1960 but were available on a county basis. As a proxy for township data in 1960, the ratio of the township to county in 1970 was multiplied by the county data in 1960 to extrapolate a township figure for 1960. To calculate per capita income, the total income per township was divided by the total population in that township. This variable was 47 deflated using a consumer price index published in Statistical Abstract o f the United States, 1992. Total income data for the year 1960 was not available on a township basis, however, it was on a county basis. For a proxy, the ratio of township to county for the year 1970 was used to extrapolate a figure for per capita income by township in the year 1960. An employment rate was created by dividing the number of employed persons by the number of people in the work force in each township. The number of employed persons was not available for 1960 on a township basis. Therefore, the township to county ratio for 1970 was multiplied by the county data for 1960 to determine a proxy. Again, the data needed for these calculations came from the U.S. Census Bureau. The cost of financing was identified as a factor that could affect land division. This is specified in the model by using home mortgage interest rates for new homes, Federal Housing Authority (FHA) insured (Statistical Abstract o f the United State, 1970, 1984, 1992). Ideally the average rate in Michigan would be used, but attempts to find specific data for Michigan were unsuccessful. Chris Peters, an economist employed by the U.S. Federal Reserve in Chicago, suggested that there would be no significant difference between the Michigan average and the U.S. average (Peters, personal communication, November 1993). Average interest rates on home mortgages in 1960 would not identify changes in the years leading up to 1960. Therefore, interest rates for that year and the previous 48 four years were averaged for a proxy of effects of interest rate over time. This method was also used for 1970, 1980, and 1990. Costs of travel were also identified in Chapter 3 as an economic factor affecting the demand for lots. Distance to a freeway interchange and from a major city were two factors considered as indicators of the cost of travel. To obtain distances to a freeway interchange, road maps from the Michigan Department of Transportation (MDOT) for 1960, 1970, 1980, and 1990 were used in combination with a detailed township map. Using the detailed township map, the distances to freeway interchanges were established for townships in 1960. The MDOT maps exhibited additions of freeways over time. Once changes in freeways were identified, new distances were calculated for those townships affected. The center of the township was used for measuring all distances. Distances to a major city were measured slightly differently. A major city was defined as having a population of 50,000 or more in 1990. Once the major cites were identified, a mapping software program, Atlas Pro, was used to measure the distance from the center of the township to the center of the nearest major city. Another factor identified as a cost of travel was the price of gasoline. Ideally, the average price of gasoline for Michigan would be used. However, the closest descriptive historical price was the average price of gasoline for Detroit, Michigan (American Petroleum Institute, 1993). To capture changes in prices over time the price of gas used in the model was an average for the five years ending in the specified year (the gasoline price used for 1960 for example, was the average price for 1956 through 49 1960). This variable was deflated using a consumer price index which excluded energy {Statistical Abstract o f the United States, 1992). The amount of paved roads in a township was also considered as a factor affecting the demand for lots. Such data were not available at a township level and this variable was not included in the model. One of the factors identified as possibly increasing the supply of lots was that periods of economic hardships would force farmers to sell their land as lots. As a proxy, average net farm income per farm for farms in Michigan was utilized. To determine this figure, total net farm income (U.S. Department o f Agriculture, 1986) was divided by the number of farms in Michigan (Michigan Agricultural Statistics Service, 1992). For each year, the average income of the five years ending on that specified year was used. Social Determinants In Chapter 3, total population, number of households and the age of the population were identified as factors affecting the land market. Township data for total population and total number of households were used to measure this impact. Median age of the population was used as a measure of the changes or differences in ages of the population. The U.S. Bureau of the Census published this data on a township basis except for median age in 1960. The ratio of median age in the township to median age in the county in 1970, in conjunction with the median age in the county in 1960 was used to estimate a median age in the township in 1960. 50 Institutional Determinants It was noted earlier that several institutional variables would affect the rate and pattern of land divisions. Township zoning ordinances would have an impact on land divisions. However, townships have different types of ordinances and these ordinances change at different times, therefore, it would be infeasible to represent these in the equation. An institutional determinant represented in the model is the Subdivision Control Act. This Act is treated as a shock variable and is specified in the model as a dummy variable. Because the real estate market has a significant time lag, it was concluded that the SCA should be included as a dummy variable with "0" for 1960 and 1970, and "1" for 1980 and 1990. 4.7 SUMMARY Four hypotheses were developed in this chapter. Questions regarding levels and patterns of land fragmentation were presented in such a manner to allow for testing. The first three hypotheses will be tested using trend analysis. The fourth hypothesis, which is the central question of this study, requires multiple regression analysis. An econometric model was developed using the factors identified in Chapter 3. The manner in which the factors identified in Chapter 3 are incorporated in the model and the sources of the data are included in this chapter. The empirical results from testing the hypotheses are reported in Chapter 5. Chapter 5 EMPIRICAL RESULTS 5.1 INTRODUCTION This chapter presents the empirical results of this research. The first section focuses on the manner in which land fragmentation is measured. The amount and distribution of land fragmentation and changes of lot patterns over time are then presented. These data are used to test the first three hypotheses proposed in Chapter 4. Estimates of the multiple regression model, that was constructed to test the fourth hypothesis, are presented in the next section, The final section summarizes the findings of the study. 5.2 SELECTION OF SAMPLE The State of Michigan consists of 1,241 townships. Due to the difficulty in acquiring local land use information, a random sample of 90 townships was used in this study. Because of possible inconsistent effects of the factors affecting the land fragmentation process, population density was used to stratify the sample townships into homogeneous segments. A measure of population density was created for each township by dividing the 1990 population by the number of square miles in the township. 51 After ranking all 52 townships according to this factor, the patterns in the data suggested natural discontinuities at the cutoff points of 100 and 600 persons per square mile, creating three distinct strata. The low density stratum included 940 townships, the medium density stratum included 246 townships, and the high density stratum included 54 townships. A random number generator was used to select a random sample of thirty townships in each stratum. Townships that were selected more than once were replaced with different townships so that 30 unique townships were selected per stratum. Figure 5.1 depicts the townships selected. The townships included in the sample are listed in Appendix A. 53 Figure 5.1 Seventy Two Townships Selected in Study Sample Eighteen townships were eventually deleted from the sample because of missing plat maps or because of radical change^ in the size of the township during the study period. Table 5.1 shows the number of townships in each group in the study. 54 Table 5.1 Number of Townships Selected and Used in Each Population Category Title Selected Used Low Density (0-100 persons per square mile) 30 22 Medium Density (101-600 persons per square mile) 30 28 High Density (Over 600 persons per square mile) 30 22 5.3 DETERMINING LAND FRAGMENTATION The level of land fragmentation in each township was determined by counting the number of unplatted lots of less than 20 acres for the years 1960, 1970, 1980, and 1990. The limit of 20 acres was chosen because a lot greater than 20 acres could be further divided into lots of 10 4- acres and still be greater than the 10 acre maximum established by the Subdivision Control Act (SCA). Also, lots greater than 20 acres are generally considered large enough to still be used for agricultural purposes. All data are collected using plat maps published by Rockford Publishing Inc. Number of Lots A summary of the number of total lots counted in the sample townships is given in Table 5.2. The complete listing is included in Appendix B. In presenting the results, the number of lots were divided into five groups: 1-9 acres, 10-11 acres, 12-19 acres, 1-19 acres, and 10-19 acres. These distinctions are made because of the definition of 55 subdivision in the SCA. With incentives in the SCA to create 10 + acre lots, the 10-11 acres group was generated to demonstrate any changes in this size category. The 1-9 acres and the 12-19 acres groups included the number of lots on either side of the 10-11 acres group. The 1-19 acres group provides a summary of all lots counted. The 10-11 acres group is combined with the 12-19 acres group to create the 11-19 acres group, which represents the number of large lots identified in the plat count. These large lots are important because of their potential contribution to land fragmentation. Table 5.2 Number of Lots Under 20 Acres for 72 Sample Townships, 1960-1990 1-9 Acres 10-11 Acres 12-19 Acres 10-19 Acres 1-19 Acres 1960 1970 1980 1990 4,625 1,538 1,340 2,878 7,503 7,294 2,563 1,955 4,518 11,812 12,933 5,321 3,040 8,361 21,294 16,271 7,197 3,753 10,950 27,221 % Change 1960-90 251.8% 367.9% 180.1% 280.5% 262.8% For the 72 townships used in this study, the number of total lots under 20 acres increased from 7,503 in 1960 to 27,221 in 1990. Thus, in a span of 30 years 19,718 new parcels were created, a 262.8% increase in the number of lots of this size. This increase was not consistent throughout the period studied. The number of lots created 56 from 1970 to 1980 more than doubled the number of lots created from 1960 to 1970. The number of lots created in 1980 to 1990 was notably less then in 1970 to 1980. Comparing lots of 1-9 acres, 10-11 acres and 12-19 acres demonstrates that the increase was not equal across lot groups. The number of lots 1-9 acres increased by 251.8%, 10-11 acres lots increased by 367.9%, while the 12-19 acres lots increased by 180.1%. This evidence suggest that the number of 10 + acre lots increased relatively faster than other sized lots. This increase in the number of large lots is a concern because of their potential contribution to land fragmentation. Number of acres Another method of examining the data is to calculate the number of acres in the lots under 20 acres. The total acreage in each size category was computed by multiplying the number of lots by their respective size and summing across all lots. The total number of acres each year are shown in Table 5.3. 57 Table 5.3 Number of Acres in Lots Under 20 Acres for 72 Sample Townships, 1960-1990 1-9 Acres 10-11 Acres 12-19 Acres 10-19 Acres 1-19 Acres 1960 1970 1980 1990 20,758 15,563 20,373 35,936 56,694 30,598 25,882 29,806 55,688 86,286 52,450 53,675 45,837 99,512 151,962 64,404 72,581 56,499 129,080 193,484 % Change 1960-90 210.3% 366.4% 177.3% 259.2% 241.3% The total acreage in lots under 20 acres increased from 56,694 acres in 1960 to 193,484 acres in 1990, a 241.3% increase. For all groups, the percent change from 1960 to 1990 decreased slightly when measured as acreage rather than number of lots. Once again, the largest total increase was in the 10-11 acres group, which increased by 366.4% Using the total acreage in lots under 20 acres, the percent of land in a township in unplatted lots can be calculated. These results are presented in Table 5.4. 58 Table 5.4 Percentage of Township Land Area in Unplatted Lots Under 20 Acres Percent Maximum 41.1% Minimum 1.2% Mean 11.7% Median 10.7% The percentage of township land in unplatted lots under 20 acres ranged from 1.2% to 41.1%, with the mean equaling 11.7% and the median 10.7%. One cannot assume that those townships with relatively low percentage of lots under 20 acres do not have much land division activity. Land division activity could have been platted and therefore, would not be included in this figure. In addition, total acreage in a township includes villages and the land divisions inside village limits were not included in this study. 5.3 ASSUMPTIONS UNDERLYING STATISTICAL TESTS The choice of valid statistical models requires assumptions regarding the sample data. The data from counting the number of lots were assumed to be normally distributed and the mean was assumed to be the appropriate statistic of central tendency. Two different procedures were used in testing the hypotheses, depending on whether the 59 means were independent or dependent. When the means were dependent, or coming from the same source, the data were compared using the difference in their numerical value, (a paired difference test). The two dependent means were compared using the observed mean of the resulting paired differences. When the means are independent, pair-wise tests of equality among means are based on the differences between the means. If the differences between the means are statistically significant using t-test analysis, then the null hypothesis is rejected (Snedecor and Cochran, 1980). Land Fragmentation Hypothesis The Land Fragmentation Hypothesis was stated in Chapter 4 as: (A) "The mean number o f unplatted lots under 20 acres in size is equal over tim e." and (B) "The mean rate o f change in the number o f unplatted lots under 20 acres in size is equal over tim e." Because the means were dependent, the paired differences procedure was used to test Part A of this hypothesis. The null hypothesis for Part A was that the mean difference in the number of lots between each decade was equal to zero. This hypothesis was rejected at the .01 significance level for each of the decade comparisons and for the 1960 to 1990 comparison. This implies that there were statistically significant changes in the number of lots during this period. The results indicate that, on average, there were 60 more 1-19 acres lots in 1970 than in 1960. There were 132 more lots, on average, in each township in 1980 than in 1970 and 81 more lots in 1990 than in 1980. There were on average, 273 more lots in each township in 1990 than in 1960. 60 For Part B of this hypothesis, the mean change in the number of lots from 1960 to 1970 was compared to the change in the number of lots between 1970 to 1980 and 1980 to 1990. Similarly, the mean change from 1970 to 1980 was compared to 1980 to 1990. The null hypothesis was rejected because the mean change in the number of lots for all comparisons were significantly different at the .01 level. The results indicate that, on average, there were 72 more 1-19 acres lots created in the time period 1970 to 1980 than in 1960 to 1970. There were 49 more lots, on average, created in each township during 1980 to 1990 than in 1970 to 1980 and 23 more lots created during 1980 to 1990 than in 1960 to 1970. Table 5.5 presents the results from the tests. 61 Table 5.5 Descriptive Statistics and Results of Statistical Tests Fragmentation Hypothesis Paired Differences Mean Difference Std. Calculated Dev. * t-value ** for Land df 2-tail Prob. Comcarison of number of total lots in the vears: 1960 1970 1980 1960 and and and and 1970*" 1980 1990 1990 59.8 131.7 80.9 272.5 52.1 115.5 59.5 195.2 9.75 9.68 11.55 11.81 71 71 71 71 0.000 0.000 0.000 0.000 85.2 97.3 62.2 7.15 4.31 3.06 71 71 71 0.000 0.000 Comparison of the increase in number of total lots between the vears: 1960-70 and 1970-80 1970-80 and 1980-90 1960-70 and 1980-90 71.8 49.4 22.5 0.003 * Std. Dev. of the Mean Difference “ The t-valuo was calculated as t= (mean difference)/((Std DevV(sq. root of N» For example, to test this hypothesis, the number of lots in each township was counted in 1960 and in 1970. The number of lots in 1960 was subtracted from the number of lots in 1970 to calculate the paired difference from 1960 to 1970. These differences for each township were used to calculate the mean and standard Land Fragmentation Consistency Hypothesis The study townships were sorted by population density and using cutoff points of 100 and 600 persons per square mile, created three distinct strata, classified as: low, medium, and high density. The number of lots per density group are given in Table 5.6. 62 Table 5.6 Number of Lots under 20 Acres for 72 Sample Townships, 1960 to 1990, by Population Density Group Total Number of Lots 1960 1970 1980 1990 % Change 1960-90 Low Densitv GrouD (less than 100 nersons Der sauare mile) 1-9 Acres 10-11 Acres 12-19 Acres 10-19 Acres 1-19 Acres 396 133 177 310 706 901 257 246 503 1,404 1,894 679 413 1,092 2,986 2,804 1,029 628 1,657 4,461 608.1% 673.7% 254.8% 434.5% 531.9% Medium Densitv Groun 1100 < Dersons Der sauare mile < 600) 1-9 Acres 10-11 Acres 12-19 Acres 10-19 Acres 1-19 Acres 2,381 767 604 1,371 3,752 3,816 1,393 961 2,354 6,170 7,208 3,200 1,567 4,767 11,975 8,807 4,356 1,917 6,273 15,080 269.9% 467.9% 217.4% 357.5% 301.9% Hieh Densitv Grouo (Greater than 600 Dersons Der sauare mile) 1-9 Acres 10-11 Acres 12-19 Acres 10-19 Acres 1-19 Acres 1,848 638 559 1,197 3,045 2,577 913 748 1,661 4,238 3,831 1,442 1,060 2,502 6,333 4,660 1,812 1,208 3,020 7,680 152.2% 184.0% 116.1% 152.3% 152.2% Comparing across densities, the lowest percentage increase in the number of total lots was in the high density townships (152.2%). This result might be expected because landowners in high population density areas, would most likely develop subdivisions to 63 maximize lots per acre because of heightened development pressure created by limited land and greater population. The highest percentage increase was in the low density townships (531.9%). Though this was the largest percentage increase, the medium density group had the largest increase in total number of lots under 20 acres (11,328). There is typically a substantial demand for lots in townships in the medium density group. However, the demand is not so great to encourage landowners to develop the more costly subdivisions. The Land Fragmentation Consistency Hypothesis was used to test whether these differences are significant. The results and discussion from testing this hypothesis are given in the next section. Land Fragmentation Consistency Hypothesis The Land Fragmentation Consistency Hypothesis was stated in Chapter 4 as: "For any decade, the mean change in the number o f unplatted lots less than 20 acres in size is equal across township groups." This hypothesis was tested by comparing the mean change in the number of lots of each density group. The data and results of this test are presented in Table 5.7. 64 Table 5.7 Descriptive Statistics and Results of Statistical Tests Fragmentation Consistency Hypothesis Mean Std. Calculated Dev. * t-value ** df for Land 2-tail Prob. A veraae increase in the number of total lots 1960-70 Low Density Medium Density High Density 30.1 84.1 58.6 22.5 62.4 44.2 T est of Low and Medium Density T est of Medium and High Density T est of Low and High Density 1970-80 Low Density Medium Density High Density 68.7 209.9 95.1 T est of Low and Medium Density Test of Medium and High Density Test of Low and High Density 62.5 118.0 56.8 48 48 42 0.000 -4.75 3.50 -1.42 48 48 42 0.000 -4.03 3.82 0.41 48 48 42 0.000 0.000 0.111 0.010 32.8 136.1 80.6 T est of Low and Medium Density Test of Medium and High Density Test of Low and High Density 1980-90 Low Density Medium Density High Density -3.86 1.62 -2.70 0.001 0.162 35.1 56.5 56.1 * Std. Dev. of the Mean Difference ** The t-value was calculated as t= (mean difference)/((Std Dev)/(sq. root of N)) 0.687 65 The null hypothesis that the rate of change of land fragmentation was equal across all density groups was rejected for all years and all density groups except for 1980 and 1990, when the mean number of lots created were not statistically different between the high and low density townships at the .05 significance level. This rejection of the null hypothesis implies that differences in the rate of land fragmentation exist among townships of different population densities. In the two cases when the null hypothesis was not rejected, one would expect the number of new unplatted lots to be similar for a variety of reasons. In low density townships, one would expect a low number of unplatted lots because the average demand for new lots would be low. In high density townships there would be a greater demand for lots, therefore, landowners would develop subdivisions to maximize lots per acre. Having tested the level and rates of land fragmentation, the remainder of this research will focus on land fragmentation patterns. In the next section, the number of 10-11 acres lots are examined. Consistency of Land Division Patterns The number of 10-11 acre lots is examined in detail because of the incentive in the SCA to create 10 + acre lots. The data are presented as the percent of lots created in each period in order to observe patterns of land fragmentation. The data are also sorted by density group to present any differences between density groups. These results are given in Table 5.8. 66 Table 5.8 Distribution of Created Lots Under 20 Acres by Population Density Group, 1970 to 1990 1-9 acre 10-11 acre 12-19 acre All Townships 1960-70 1970-80 1980-90 61.9% 59.5% 56.3% 23.8% 29.1% 31.7% 14.3% 11.4% 12.0% Low Density Group 1960-70 1970-80 1980-90 72.3% 62.8% 61.7% 17.8% 26.7% 23.7% 9.9% 10.6% 14.6% Medium Densitv Group 1960-70 1970-80 1980-90 59.3% 58.4% 51.5% 25.9% 31.1% 37.2% 14.8% 10.4% 11.3% High Densitv Group 1960-70 1970-80 1980-90 61.1% 59.9% 61.5% 23.1% 25.3% 27.5% 15.8% 14.9% 11.0% For all 72 sample townships, 24 percent of the lots created between 1960 to 1970 were 10-11 acres in size. During 1970 to 1980, 29 percent of the lots created were 1011 acres in size and during 1980 to 1990, 32 percent of the lots created were 10-11 acres in size. This increase in the percentage of 10-11 acre lots was not a downsizing from 67 12-19 acre lots because as a percentage the number of lots 1-9 acres decrease from 61.9% to 56.3%. The pattern of having a greater percentage of lots at 10-11 acres was consistent for 1970 to 1980 and 1980 to 1990. The medium density group had the largest percentage of lots, 10-11 acres, of the three groups in all three time periods with the maximum in 1990 at 37%. For both the medium and high density groups, there was a trend of a continual increase in the percentage of lots being created being 10-11 acre lots. However, for the low density group the percentage in the final period decreased as the percentage of 12-19 acre lots increased. The next section addresses the hypothesis on the pattern of land fragmentation. The hypothesis and test results are presented. Consistency of Land Division Patterns Hypothesis The Consistency of Land Division Patterns Hypothesis focused on the pattern of new lots being created. It was hypothesized that the SCA introduced an incentive to create 10 + acre lots. This hypothesis investigates this pattern in two parts: {PC) "The mean percentage o f all new unplatted lots that are 10-11 acres in size is equal over time. " and (B) "The mean percentage o f all new unplatted lots that are 10-11 acres in size is equal across township groups." Several tests were conducted to test this hypothesis. For Part A of the hypothesis, the mean percentage of new lots in each township that were 10-11 acres in size, are tested across the different time periods. Because the means were dependent, the paired 68 difference procedure described earlier was used. For Part B of the hypothesis, each density group is compared to the other groups in each time period. The results of these tests are reported in Table 5.9. 69 Table 5.9 Descriptive Statistics and Results of Statistical Tests for Consistency of Land Division Patterns Hypothesis Mean Std. Dev. * Calculated t-value ** 2-tail Prob. df Paired Differences ComDarison of Dercent of 10-11 acre lots in vears: 1960-70 and 1970-80*** 1970-80 and 1980-90 1960-70 and 1980-90 8.5 1.5 8.5 14.3 16 20.2 5.5 21.0 14.4 5.4 24.7 22.9 5.02 0.78 3.58 70 70 71 0.000 0.435 0.001 2.90 0.98 1.79 48 46 42 0.006 6.700 0.081 3.89 3.43 0.83 48 48 42 0.001 0.409 4.26 3.98 0.05 48 48 42 Average increase in percent of 10-11 acre lots laSQrZQ Low Density Medium Density High Density Test of Low and Medium Density Test of Medium and High Density Test of Low and High Density 1970-80 Low Density Medium Density High Density 18.9 65.5 23.1 12.4 55.1 20.3 Test of Low and Medium Density Test of Medium and High Density Test of Low and High Density 1980-90 Low Density Medium Density High Density Test of Low and Medium Density Test of Medium and High Density Test of Low and High Density 15.6 42.6 15.4 0.000 12.4 27.5 18.5 0.000 0.000 0.962 * Std. Dev. of the Mean Difference ** The t-value w as calculated as t= (mean difference)/((Std Dev)/(sq. root of N)) *"* For example, to test this hypothesis, the number of lots in each township w as counted in 1960 and in 191 The number of lots in 1960 w as subtracted from the number of lots in 1970 to calculate the paired differe from 1960 to 1970. T hese differences for each township w ere used to calculate the m ean and standard deviations of the differences listed in columns one and two. 70 For the entire sample of townships, the mean percentage of 10-11 acre lots was significantly different between 1960 to 1970 and 1970 to 1980 or 1980 to 1990. However, the mean percentage of 10-11 acres lots during 1970 to 1980 were not statistically different than the mean percentage of lots during 1980 to 1990. These results suggests that the SC A might have some effect on the percentage of 10-11 acre lots. With the SCA enacted in 1968, the 1960 to 1970 period would be the land fragmentation pattern before the Act, while the land patterns in the other two periods would be after the Act. When examining the differences between the values for the township groups in the 1960 to 1970 time period, the only significant difference was between the low and medium density groups. In 1970 to 1980, there was no significant difference between the low and high density groups. This pattern was repeated in 1980 to 1990. Based on this evidence both Part A and B of this hypothesis are rejected, suggesting that the pattern of land fragmentation has changed over time. The next step is to identify the impact of factors affecting the land fragmentation process in the sample townships. The Land Division Patterns Hypothesis in the next section focuses on these issues. 5.4 LAND DIVISION PATTERNS HYPOTHESIS The Land Division Patterns Hypothesis states: "Land division patterns are impacted by ecological, economical, social, and institutional factors. However, the Subdivision Control Act o f 1967 has had no effect on the number o f 10 + acre lots. ” 71 This hypothesis is tested using the econometric model presented in Chapter 4. The regression model and results of this hypothesis test are presented in the following sections. Econometric Model In Chapter 4, an econometric equation modeling land division was specified using the factors identified in Chapter 3. The results of the estimated equation are reported here. Table 5.10 lists the variables used and their descriptions. The methods of deriving the variables were described in Chapter 4. 72 Table 5.10 Variable Specification for the Land Fragmentation Model VARIABLE DESCRIPTION AGP The percent of the township’s land area devoted to agricultural uses. D1 Dummy for low density townships, 1 for low density, 0 otherwise. D3 Dummy for high density townships, 1 for high density, 0 otherwise. DSCA Dummy for the SCA, "0" for 1960 and 1970, "1" otherwise. DCITY* Distance to a city of 50,000 in 1990. DFREE* Distance to a freeway interchange. EMP Total employment in the township. FRMINC Real average farm income in Michigan (thousands of 1983 dollars). INCOME Per capita income by township (thousands of 1983 dollars). LAND Total number of acres in the township (thousands). MEDAGE Median age of residents in the township. MORTNOM Nominal rate of interest on home mortgages. MORTREAL Real rate of interest on new mortgages. PGAS Average price of gas in Michigan (1983 dollars). REC Percent of the township that is either forest or water. TOTHOU Total number of households in a township (thousands). TOTPOP Total population in a township (thousands). There were four townships that had values from 150 to 300 miles while the rest of the townships had values under 68 miles. For those townships greater than 68 miles from either a large city or freeway interchange, the values were changed to 68 thus truncating the data. 73 Dependent Variable The dependent variable in the model could be specified in one of three ways. The alternative specifications include: (1) the number of 10-11 acre lots; (2) the acreage in 10-11 acre lots; or (3) the percent of the township’s land area in 10-11 acre lots. Regression models were estimated using each of these specifications and the results were compared to determine which dependent variable would be used. Two equations were estimated for each dependent variable, with interest rates for new mortgages specified as either a real or nominal interest rate. The results from these equations were compared and new equations were estimated using different combinations of variables (i.e., variables that were consistently insignificant and unstable in sign were omitted). The adjusted R2s, signs of the regression coefficients, and standard errors for each equation, were compared to determine the most appropriate specification of the dependent variable. Based on these results, the number of 10-11 acre lots was selected as the dependent variable. Deleted Variables Three variables, DFREE, PGAS, and EMP, were deleted from the model. Two variables, DFREE and DCITY, were specified to measure the influence of the distance from cities or freeways on land fragmentation. It was hypothesized that access to a freeway would reduce traveling time thus increasing demand for lots in an area. DFREE was consistently insignificant and was deleted from the model. The variable, DCITY was retained in the model to measure the effect of distance on land fragmentation. 74 Another variable designed to measure the cost o f travel, PGAS, was also deleted from the model because it was insignificant in all versions of the model. PGAS was defined as a five year moving average of the price of gasoline in the city of Detroit. Possibly it was not descriptive of the price of gasoline in Michigan or that the price of gas was not a factor in demand of 10-11 acre lots. EMP, the employment rate variable, was the third variable deleted from the model. The EMP variable and the per capita income variable, INCOME, were included as measures of economic ability. EMP was insignificant in most cases and therefore, was not included in the final equations. Population Testing and Pooling The test of the Consistency of Land Divisions Hypothesis concluded that the number of 10-11 acre lots being created were sometimes significantly different across the three groups of townships. In light of this result, some question remained about whether these differences should be accounted for when pooling to estimate the regression for the entire sample. To test whether the townships were different, the same equation was estimated using the entire sample, and the three different population density groups. The null hypothesis that all estimated regression coefficients were equal among the four equations can be tested using an F-test as the relevant test statistic (Kementa, 1971). The test results are given in Appendix C. Since the null hypothesis that all regression coefficients were equal was rejected, this result indicates that some regression parameters were different for some density 75 groups. Therefore, in order to pool the three groups, those variables that had unequal regression coefficients must have a separate variable for each density group. To create these new variables, the variable that was identified as having a different coefficient for each density group, was multiplied by the dummy variables D1 and D3, thus creating two additional variables that would measure the different impact of population density. To identify which variables had unequal regression coefficients, a test was created using the regression coefficient and the standard error for each variable. A range for each variable was created by adding and subtracting the standard error from the estimated coefficient. If one of the estimated coefficients fell outside the range of the others, then two new variables were created by multiplying this variable by D1 and D 3.1 Once all coefficients were examined and new variables created, the three density groups were pooled and the equation was re-estimated. The estimated coefficients of the newly created variables were then examined to determine if they were different from one another using the same process as before. If the coefficients on the newly created variables were overlapping, these variables were deleted and the equation was estimated again. 1 For example, the estimated coefficient on TOTPOP was 7.839 and the standard error was 2.265 for the low density township equation. The estimated coefficient on TOTPOP was 10.430 and the standard error was 3.174 for the medium density township equation. The estimated coefficient on TOTPOP was .903 and the standard error was .719 for the high density township equation. Since the estimated coefficient on TOTPOP for high density townships was outside the range of the other two, it was concluded that the TOTPOP variable had different coefficients for each density group. Thus, a new variable was created by multiplying TOTPOP by D1 and D2. 76 Regression Results Most of the variables specified were time and townships specific. However, the DSCA, MORTNOM, MORTREAL, and FRMINC variables were time specific, with only one value used for these variable in each cross sections (i.e., the MORTNOM variable was 5.492 for all townships in the 1960 sample). It was discovered the model yielded inconclusive results when more than two of these time specific variables were included in the model. Therefore, in reporting the findings of this study, results from four equations were given, with no more than two of these time specific variables included. In the first equation only DSCA included. DSCA and FRMINC. The second equation includes The third equation includes DSCA and MORTNOM and the fourth equation includes DSCA and MORTREAL. When the regression coefficients were tested for equality using the procedure described in the preceding section, the variables, LAND and TOTPOP were found to be different in all four of the final equations. The variable DSCA was different for the equations, 1 and 4. The INCOME variable was different in all but equation 4. The DCITY and TOTHOU variables were determined to be different in equation 4. However, after pooling and estimating a new equation, the newly created variables in this equation for DCITY and TOTHOU were determined to be similar. Therefore, they were deleted and the equation was estimated again. The results of the four final equations estimated are given in Table 5.11. 77 Table 5.11 Estimates of the Land Fragmentation Regression Model VaiisM* AGP 01 D3 DSCA DSCAD1 DSCAD3 DCITY FRMINC INCOME INCOMED1 INCOMED3 LAND LANDD1 LANDD3 MEDAGE MORTNOM 1 2 3 •0.10 -0.10 -0.25 (-0.464) (-0.489) (-1.169) 256.55 280.60 280.25 200.56 (3.768)” * (4.198)“ * (4.200)” * (3.254)*” 266.54 280.09 279.38 17C.39 (3.911)*” (4.163)” * (4.160)” * (2.728)” * 51.14 23.63 15.45 6812 (4.806)” * (1.818)* (0.S38) (6.802)“ -26.40 -37 64 (-1.762)* (-2799)— -0.92 -16.39 (-0.063) (-1-192) -0.10 -0.12 -0.12 6.20 (-0.475) (-0.568) (-0.574) (-1.005) 5.22 (1.660)* 10.02 11.80 11.80 (3.959)” * (4.964)” * (4.993)” * -9.55 -10.74 -10.74 (-3.425)” * (-4.080)” * (-4.097)*” 8.91 9.01 8.99 8.65 (3.456)*” (3.498)*” (3.494)” * (3.357)— -8.98 -9.08 -9.06 6.72 (-3.485)” * (-3.529)” * (-3.528)” * (-3.385)*” -6.36 •6.33 -6.35 6.68 (-2381)” (-2383)” (-2379)” (-2135)” -1.14 -1.16 -1.17 -1.82 (-1.597) (-1.604) (-1.650* (-2375)“ 6.87 (1.909)* 5.34 (3.351)— REC 0.28 (0.906) 0.34 (1.119) 0.34 (1.108) TOTHOU -1.85 (-1.053) -0.88 (-0.599) 6.98 (6.597) TOTPOP 9.96 10.02 (5.650)” * (6.210)*” TOTPOPD1 1.38 -0.03 (0.231) (-0.004) -9.77 -10.06 (-6.187)*” (-6.670)*“ (Constant) -25249 (-3.785) -313.63 (-4.363) Adjusted R 1 0.59596 0.59557 2266 23.83 F- Statistic 1.68 (1.191) -5.80 -9.71 -9.73 (-1 335) (-2484)” (-2495)” MORTREAL TOTPOPD3 4 -0.17 (-0.776) 0.30 (0.959) -238 (-1.350) 9.96 10.61 (6.179)” * (6.297)— 6.07 0.19 (6.012) (0.034) -10.01 -10.43 (6.645)” * (6.727)— -315.61 -170.16 (-4.461) (-2610) 0.59693 0.59659 2395 Numbers in parentheses are t-statistics Dep. variable = "Number of 10-11 acre Lots" 2392 *** ** * = 1 % Significance Level = 5 % Significance Level = 1 0 % Significance Level 78 Land Division Patterns Hypothesis The Land Division Patterns Hypothesis was tested using multiple regression analysis. Results from four different equations were reported in Table 5.11. The dummy variable, DSCA represented the impact of the SCA on the creation of 10-11 acre lots. The SCA was enacted in 1967 with enforcement beginning in 1968. Given the brief period between the start of enforcement in 1968 and the sample data in 1970, the DSCA variable was specified in the model as "0" for years 1960 and 1970 and as "1" for years 1980 and 1990.2 The estimated coefficient on the DSCA was positive in all of the final equations. The DSCA coefficient was significant at the one percent level in two of the equations and at the ten percent level in another equation. Thus, the null hypothesis is rejected and it can be concluded that the SCA did have a positive effect on the number of 10-11 acre lots in the sample townships. DSCA was dummied in two equations, implying that the SCA had differing impacts on the different density groups. To determine the impact of the SCA on low density townships, the estimated coefficients for DSC ADI is added to the coefficient estimated for DSCA. For high density townships, DSCAD3 is added to the coefficient estimated for DSCA. Figure 5.2 shows the impact of the SCA on the different density groups estimated by the four equations. 2 It was assumed that there was no Subdivision Control Act in force during the time period between 1960 and 1970, but the SCA was being enforced after 1970. 79 Figure 5.2 Estimated Impact of the SCA on Number of 10-11 Acre Lots Low Density m m Med. Density □ High Density Ail Densities Equations 1 M u m AOP. D1. 01. DSCA. DSCAD1. 0SCA3. OCtTY. otCORE. KCOHEOI. IH00ME0J. IAMD. LAMD01. U N 003. UEDAOE. REC, TOTHOU. TOTPOP. TOTPOPOI. TOTPOPOl 3 M u M AGP. 01. 03. DSCA. OCtTY. FHHMC. MCOME. NCOMEDt. MCOBED1. lANO, IAHD01. LAN001. UEQAGE, REC. TOTHOU TOTPOP. TOTP0P01. TOTPOPOl 1 M u m AGP. 01. 01. OSCA. OCTTY. WCOME. WCOWED1. 1RCCWE03. LAJffli LAN001. LAWOD3. UEQAGE. UORTNCMA REC TOTHOU. TOTPOP. TOTPOPOl. TOTPOPOl 4 M um aop. oi. ta, o k a . oacAOi. OSCAV oetTY. w c o a t land, lawsoi. u m a s . ueoage. m o rtreal. r e c . to th o u . to tp o p . to tp o p o i. to tp o p o i The regression results indicate that the passage of the SCA caused an increase in the number of 10-11 acre lots. For equations 2 and 3, the results do not differ among density groups and the estimated impacts on the number of 10-11 acre lots are somewhat smaller than for equations 1 and 4. In equations 2 and 3, which includes all townships, 80 the SCA is estimated to have caused the creation of an additional 15 or 23 lots of 10 to 11 acres in size (equations 3 and 2). In equations 1 and 4, the DSCA variable was dummied when the density groups were pooled. Consequently, there is a separate estimate for each density group. From equations 1 and 4, the range o f impact on the number of 10-11 acre lots is 25 or 30 lots for the low density township group. For the medium density group, the impact on number of 10-11 acre lots range from 51 to 68 lots. For the high density group, the estimates range from 50 to 51 lots. Equations 1 and 4 were similar in that they both estimate the impact on the middle density townships to be the greatest. The results of equations 1 and 4 suggest that more 10-11 acre lots were created in the medium density townships because of the SCA than the other density groups. A possible explanation is the likely differing levels of demand for lots in the different density groups. It is suggested that the demand for lots in the medium density townships is greater than in the low density townships but not as great as in the high density townships. In the high density townships, with the greater level of demand for lots, this would allow the increased costs due to platting be recouped because the higher level of demand allows for higher prices to be charged on the lots. Impact of Other Factors on Land Fragmentation The Land Division Patterns Hypothesis focused on the impact of the SCA on land fragmentation patterns, but the impact of other factors on land fragmentation patterns can 81 be calculated from the estimated equations. The impact o f the other variables on the number of 10-11 acre lots of the other variables are presented in Table 5.12, . 82 Table 5.12 Estimated Range of Impact on the Number of 10-11 Acre Lots From the Land Fragmentation Model LowDensity Variable Low High Low AGP -0.25 -0.10 DCITY -0.20 -0.10 FRMINC 5.22 INCOME 1.68 LAND MEDAGE -1.82 6.87 MORTREAL 5.34 REC 0.28 0.34 -2.38 -0.98 TOTPOP Med. Density Low High High Densitv Low High 2.07 4.23 10.02 11.8 0.47 1.06 -0.06 -0.08 8.65 9.01 2.55 2.98 9.89 11.36 9.96 10.61 -0.05 0.21 -1.14 MORTNOM TOTHOU High 83 AGP, the variable describing the percent of the township in agricultural land was not dummied in order to be pooled and had negative regression coefficients. As the percent of the township’s land area that is agricultural increased by one percent, the number of 10-11 acre lots decreased by .10 to .25. These results would suggest that agriculture is in competition for land for development. DCITY, the variable measuring mileage to a city of 50,000 people was not dummied and had negative regression coefficients in all versions of the model. Thus, as the distance to a large city increased by one mile, the number of 10-11 acre lots decreased by . 10 to .20 lots. This result was as expected, since it was expected that the fragmentation of land would decrease as the distance from an urban area increased. FRMINC, the variable for average farm income was included only in equation 2 and had a positive regression coefficient. As average farm income per farm increased by 1,000 dollars, the number of 10-11 acre lots increased by 5.22. FRMINC was included in the model because some claim that farmers sell land because of financial pressure. This result would suggest that decreases in income are not a cause of land fragmentation. INCOME, which is the variable for per capita income, had positive regression coefficients for all three density groups. Thus as the per capita income of the township residents increased by 1,000 dollars, the number of 10-11 acre lots increased by 2.07 to 4.23 in a low density township, by 10.02 to 11.8 in a medium density township, or by .47 to 1.06 in a high density township. These results suggest that the creation of large 84 lots is stimulated by growth in income but the magnitude of that change is dependant on where the land is located. LAND, the variable for the amount of land in a township had positive regression coefficients for both medium and high density groups and a negative coefficient for the low density group. Thus as the amount of land in a township increased by 1,000 acres, the number of 10-11 acre lots decreased by .06 to .08 in a low density township. Similarly, a 1,000 acre increase in the size of a township increases the number of 10-11 acre lots by 8.65 to 9.01 in a medium density township and by 2.55 to 2.98 in high density townships. As the amount of land in a township increases, the potential for the number of lots of this size increase. The negative coefficient for the low density group may be explained by presence in the sample of the very large townships in the Upper Peninsula that do not face development pressure. MEDAGE, the variable describing median age of the citizens in the townships, had negative regression coefficients. Thus, as the median age of the residents in a township increased by one year, the number of 10-11 acre lots decreased by 1.14 to 1.82. These results would suggest that as people age, they tend to prefer smaller lots. REC, the variable describing the percentage of the land in a township that was classified as forest or water had positive coefficients. As the percentage of the township’s land in forest or water increased by one percent, the number of 10-11 acre lots increased by .28 to .34. These results suggest that forest and water are not recreational amenities attracting development of large lots. 85 TOTHOU, the variable describing the number of households, had negative coefficients in all equations. As the number of households in a township increased by 1,000 households, the number of 10-11 acre lots decreased by .98 to 2.38. This is contrary to what was expected since it was expected that as population increased the number of 10-11 acre lots would increase. Though the sign on this variable was the opposite of what was expected, these coefficients were not statistically significant. TOTPOP, the population variable had positive regression coefficients for all three township groups in most of the results. This variable had a relatively large impact on low and medium density townships but very little impact on high density townships. Thus, as the total population in a township increases by 1,000 people, the number of 1011 acre lots increased by 9.89 to 11.36 in a low density township, increased by 9.96 to 10.61 in a medium density township, or decreased by .05 to increased by .21 in a high density township. These results indicating a low impact of population on 10-11 acre lots in high density townships suggest that landowners have the ability to recapture the costs of platting in high density townships. The range of estimated coefficients for variables other than Subdivision Control Act have been reported in this section. The signs of these coefficients were usually consistent with prior expectations and were often significant at the 10 percent level or higher. These results are further proof that the regression equations are a reasonable model of the land division process. 86 5.5 SUMMARY This chapter reports the results of the tests conducted on the hypotheses presented in Chapter 4. The amount of land fragmentation was determined by counting the number of unplatted lots under 20 acres on plat maps of each township in the sample. Using these data, the first three hypotheses were tested using trend analysis. Land fragmentation was determined to have occurred in the 72 sample townships during the time period from 1960 to 1990. Furthermore, the fragmentation of land was not consistent across time nor across township groups. Another conclusion was that the percentage of unplatted lots that were 10-11 acres in size being created during the time periods analyzed had increased. Townships with different population densities were found to have dissimilar patterns of land fragmentation. An econometric model was estimated to determine what factors affected this pattern of land fragmentation. The results from this model were presented as four different equations. In all four equations, the Subdivision Control Act had a positive impact on the number of 10-11 acre lots. For three of the equations, the SCA variable was significant at the ten percent level or higher. These results led to a rejection of the null hypothesis that the SCA had no effect on the number of 10-11 acre lots. The results of the regression model indicate that the passage of the SCA increased the number of 10-11 acre lots in the sample townships by 15 to 68 lots per township. In Chapter 6, the conclusions and applications from these results are given. Chapter 6 CONCLUSIONS AND APPLICATIONS 6.1 INTRODUCTION This research has examined the impact of the Subdivision Control Act of 1967 (SCA) on land fragmentation in the State of Michigan. This chapter reviews the objectives of this research and summarizes the results of the study. alternatives for reducing land fragmentation are then discussed. Some policy In the final section, suggestions for future research are made. 6.2 RESEARCH OBJECTIVES Landowners in Michigan are permitted to divide and sell land subject to local zoning ordinances and the conditions stated in the SCA. Rural land, originally surveyed as large parcels suitable for agriculture or forestry, is being divided into smaller lots most commonly used for residences. This fragmentation of land has been viewed as a concern because of its impact on open space, natural resources, and the cost of public service.A common perception is that the SCA has contributed to land fragmentation by creating an incentive to develop 10 + acre lots. The incentive is created by a provision 87 88 that allows landowners to avoid platting and the costs of platting, if no more than four lots 10 acres or less are created within any 10 year period. The public program evaluation technique was used as the framework to analyze the SCA. Using this framework, the impacts of the SCA were measured against the goals the Act set out to accomplish. The goals of the SCA are: "to regulate the subdivision of land; to promote the public health; safety and general welfare; to further the orderly layout and use of land." The manner in which these goals are to be accomplished is by the establishment of plats. In this study, an attempt was made to test whether the SCA had positively affected the number of 10-11 acre lots. No attempt was made to measure the effectiveness of the SCA on platted land divisions. The first objective of this study was to describe the SCA and identify the statute’s objective and potential unanticipated consequences. The second objectives were to describe the land market, identify factors affecting the land market, and form testable hypotheses regarding land division activity. The final objective was to develop an econometric model of the land division process in Michigan and use the model to estimate the impact of the SCA on land fragmentation patterns. 6.3 RESEARCH RESULTS The Subdivision Control Act of 1967 repealed and replaced the Plat Act of 1929. The common perception is that the SCA created an incentive to create 10 4- acre lots because "subdivision" was defined in part as more than four lots that are 10 acres or less. When a land division qualifies as a subdivision the landowner is then required to plat the 89 land division. To avoid qualifying as a subdivision, landowners can create lots of greater than 10 acres and thus avoid the cost of platting. By comparing the SCA to the Plat Act, this research has shown that the SCA was different from the Plat Act in at least two areas. One, the SCA defined subdivision more precisely than did the Plat Act. The change in definition allowed for stricter enforcement. Secondly, additional language in the SCA increased the cost of platting by changing the required standards. Thus, both the increased enforcement and the higher cost of platting provided an incentive to create 10 + acre lots to avoid qualifying as a subdivision and the necessity to plat. After examining the history of the SCA, the land market and the land fragmentation process were examined. The factors affecting land fragmentation were classified as ecological (forests, lakes, and rivers), economic (per capita income, employment, interest rates, farm income, the price of gas, and the distance to a freeway interchange or city), social (total population, total number of households, and median age), and institutional (SCA). These factors were used to build an econometric model capable of estimating the impact of the SCA on land fragmentation in Michigan. To study land fragmentation in Michigan a sample of 90 townships was selected from the 1,241 townships in the State. Population density was used to classify the townships into low, medium, and high density strata. stratum were randomly selected. maps from which to collect data. Thirty townships from each Of the 90 townships selected, 72 had adequate plat 90 To determine the level and pattern of land fragmentation in the sample townships, the number of lots, 1-19 acres in size in a township were counted using plat maps for the years 1960, 1970, 1980, and 1990. Land fragmentation rates were determined by counting the number of lots in the township in the next period. Using the data obtained from plat maps on the 72 townships, four hypotheses were tested to determine the rate and pattern of land fragmentation in Michigan. The first hypothesis focused on the change in the number of unplatted lots in the townships over time. It was found that the number of unplatted lots increased from 7,503 in 1960 to 27,221 in 1990. The null hypothesis that the mean number and rate of change of unplatted lots under 20 acres in size was equal during each decade was rejected. The second hypothesis compared the rate of fragmentation between the density groups. The percentage increase in the number of unplatted lots from 1960 to 1990 was 532 percent in a low density township, 302 percent in a medium density township, and 152 percent in a high density township. The null hypothesis that the mean change in the number of unplatted lots less than 20 acres in size was equal across townships groups was rejected. This is evidence that land fragmentation is not consistent across the state, with more rapid fragmentation occurring in medium density townships. The third hypothesis investigates the impact of the SCA on the number of 10 + acre lots by testing whether the mean percentage of all new unplatted lots that are 10-11 in size was equal over time and across township groups. The mean number of 10-11 acres unplatted lots created in the time period 1960 to 1970, was 5.5 in low density townships, 21.0 in medium density townships, and 14.4 in high density townships. 91 During the time period 1970 to 1980, the mean number of 10-11 acre lots created was 18.9 in low density townships, 65.5 in medium density townships, and 23.1 in high density townships. During the time period 1980 to 1990, the mean number of 10-11 acre lots created was 15.6 in low density townships, 42.6 in medium density townships, and 15.4 in high density townships. This hypothesis was rejected, implying that the mean number of 10-11 acre lots created had changed and that these changes were not equal across township groups. The impact of the SCA on the number of 10-11 acre lots was the largest in the medium townships. Once a pattern of an increased number of 10-11 acre lots had been established a fourth hypothesis was tested to determine the impact of the SCA on land fragmentation. The fourth hypothesis, building on the research on the land market, stated that land division patterns are impacted by ecological, economical, social and institutional factors but that the SCA has had no affect on the number of 10 + acre lots. Using the econometric model, four equations including different sets of variables were estimated. In three of the four equations estimated, the null hypothesis that the SCA had no affect was rejected. The results of these equations indicate that the SCA caused the creation of an additional 15 to 51 lots between 10-11 acres in size per township during the time period from 1960 to 1990. 6.4 POLICY ALTERNATIVES FOR REDUCING LAND FRAGMENTATION Given the results of this research, policy makers may want to examine the SCA to determine whether formation of a large number of 10-11 acre lots is consistent with 92 the objectives stated in the Act. Amending the SCA to change this pattern of development may be one policy alternative for reducing land fragmentation. Implementing land preservation techniques, such as open space zoning, are another approach to reduce land fragmentation. Possible changes to the SCA and other policy alternatives are considered in the next sections. One policy option to reduce land fragmentation is to amend the SCA to lessen the incentive to create the 10 + acre lots embedded in the Act. The SCA defines a land division as a subdivision when more than 4 lots, 10 acres or less, are created in a ten year period from a parent parcel. When a land division qualifies as a subdivision, platting is required. Therefore, if individuals want to avoid platting after the fourth split in a ten year period, they must create lots greater than 10 acres. Three approaches could be used to reduce the incentive to develop 10 + acre lots: (1) change the 10 acre provision in the definition of subdivision; (2) reduce the cost of platting; or (3) decrease the ten year period. These approaches may reduce the incentive to develop 10 + acre lots, but they also may create new incentives that could affect land fragmentation. Changing the 10 acre provision in the definition of subdivision may create a pattern of lots at the margin of the new standard. For example, reducing the 10 acre standard to 5 acres may lead to a new pattern of development of 5 + acre lots, while raising the standard to 20 acres may cause a pattern of 20 + acre lots. The resultant impact on land fragmentation of these new patterns would depend on the number of lots of this size created and the acreage limit included in the legislation. 93 Another alternative to amending the SCA would be to reduce the ten year period in which only four lots under 10 acres or less may be created without platting. This may reduce the incentive to create 10 + acre lots, because more lots of less than 10 acres could be created in the same amount of time. The affect of this option on land fragmentation would depend on the number of unplatted lots created. The third approach of reducing the number of 10 + acre lots would be to decrease the costs of platting. Costs would include the expense, time, and effort required for the platting process plus the required standards for the plat, i.e., road specifications. By lowering the costs to plat, this would reduce the incentive to avoid platting, thus possibly reducing the number of 10 + acre parcels. However, shortening the time frame or changing the required standards may have a negative impact on the intended objectives of the SCA as stated above. For example, the platting process could be shortened by streamlining the process, but this could result in an inadequate review of the plats, possibly leading to poorly designed subdivisions. Amending the SCA is only one approach to reducing the amount of land fragmented. Other alternatives to reduce land fragmentation are grouped into two categories: limiting or changing the nature of land divisions and enhancing the economic vitality of agriculture. The collective impact of each alternative could vary by location. 94 Limit or change the nature of land divisions. There are land preservation techniques that are available under the existing zoning regulations. Zoning has the ability to describe and control land use. However, zoning can be undermined when either zoning is not enforced or when use variances are given. A common zoning technique is large lot zoning where the minimum lot size is, for example, 10 acres. However, this could compound the problem of land fragmentation because for each lot created, the number of acres fragmented would be relatively large. For example, if the minimum lots were five acres in size, only half as much land would be fragmented as under a 10 acre rule. Another zoning technique is sliding scale zoning, which allows the landowners to create a certain number of lots based on the size of the original parcel. Large parcels are permitted fewer splits proportionate to total acreage than are small parcels (e.g. A 20 acre parcel might be permitted 4 splits while a 100 acre parcel might be only permitted 10 splits) This would reduce the number of allowable splits and possibly reduce the size of the lots depending on the minimum lot size as specified in the local ordinance. This could result in less land fragmentation. The objective of cluster zoning is to "cluster" residential development on one portion of the parent parcel while restricting development on the remainder of the parent parcel. This form of zoning is an attempt to maintain open space and the rural character of an area while still allowing a landowner to capture profits from development of the land. The allowable lots would be small, thus reducing the amount of land fragmented. 95 The restricted development would also eliminate land fragmentation on the remaining original parcel. Zoning techniques with the present structure are limited by the ability of the township board to pass and enforce the ordinance. Many times such an ordinance is not coordinated with other townships’ ordinances. This situation could be modified by changing the enabling laws to decrease the level of autonomy of the townships and increase the standardization of zoning laws among townships. This action could reduce the amount of land fragmented. Open space preservation techniques, such as purchase of development rights, could also be implemented. Public or private finances could be expended to buy the rights to develop the land. Public financing could be justified if the present use is preserved for the future. Land on which the development rights have been sold could not be fragmented in the future. Land will have different productivity levels for such inherent features as fertility but often the locational features are the driving force for the value of the property. The question is raised: who will capture the increased value of land located near centers of population or economic activity. The prospect of capturing large profits creates incentives to obtain zoning for maximum development. Schmid (1981) suggests that a tax which captures most of the appreciation gained would reduce the incentive to fragment land. However, public acceptance of such a tax is questionable. Enhance the economic vitality of agriculture 96 Enhancing the economic vitality of agriculture has been one approach recommended to limit the amount of agricultural land fragmented. It is argued that agricultural and forest land should be preserved because they are the principle economic base in rural areas, they are renewable resources, they retain natural environmental systems, and they provide open space and rural character. For this approach to be effective, either the number of land divisions must be reduced or, if there is a division of land, the land use must remain the same. It is often argued that enhancing the economic vitality of agriculture would reduce the number of forced sales due to low profitability, thereby reducing the incentive to sell off road frontage for residence lots. Similarly, when agricultural land is sold, higher profits in agriculture would permit farmers to better compete in bidding for land. Three methods of enhancing the economic vitality of agriculture are highlighted here. The adoption of a "differential property assessment" can reduce the amount of property tax paid by a farmer, thus improving the profitability of the agricultural firm. Under a differential property assessment, land is assessed according to the value at current use rather than according to market value. One of the criticisms of differential property assessment is the manner in which use value is determined. The very fact that use value is lower than market value is evidence that the property is worth more for another use than it is for agricultural use. restrictions on changes in land use. Sometimes, differential assessments have When there are land use restrictions, it is more likely that land with low development potential will be enrolled. Land with high development potential would probably not be enrolled being that the benefits of lower 97 taxes do not outweigh the potential of capturing increased profits from selling land for development. When the differential property assessment has no corresponding restriction on changes of land use, land that has been assessed at use value would be sold to the highest bidder regardless of uses intended by the purchaser. Tax credit techniques such as the Farmland and Open Space Act, Public Act 116, 1974, are another method of enhancing the economic vitality of agriculture. Under PA 116, farmers contract with the state to keep their land in agriculture for a minimum of 10 years, in exchange for a limit on the amount of property taxes paid. A lien in the amount of the last seven years of rebated property taxes is placed on the land regardless of length of the contract. The criticism of this approach is similar to that of the differential assessment approach. Those farmers owning agricultural land with high development potential are less likely to enroll in the program. Moreover, all land in PA 116 could be developed at some point in time by paying the lien once the contract has expired. The Right To Farm Law (RTF) is another attempt to assist agriculture. This bill defines acceptable farm practices and seeks to protect agriculture from nuisance suits and/or reduce the cost of litigation to agriculture. The fact that there is this legislation gives evidence that there are land use incompatibilities. The question remains whether enhancing the profitability of agriculture is a long term solution to preserving agricultural land. In the land division model developed in this research, farm income was found to have insignificant impact on land fragmentation. 98 This result suggests that enhancing the profitability of agriculture may have little impact on land fragmentation. This was an overview of some the options available for reducing land fragmentation. In the next section, suggestions for future research are identified. 6.5 SUGGESTIONS FOR FUTURE RESEARCH One of the limitations of this study was the inability to easily locate data on land use and some factors that may affect land use. This study was limited to 72 townships because of the time required to count the number of 1-19 acre lots in the township on plat maps (the number of lots in each township had to be counted in 1960, 1970, 1980, and 1990). A larger sample would probably improve the results of the econometric model. A number of factors were identified as possibly affecting land fragmentation. In two cases (mileage of paved roads in the township and types of local zoning ordinances) data could not be obtained within the time constraints. As a result these variables were not included in the model. Land use legislation is not unique to Michigan. Examining the impact of laws similar to the SCA in other states may yield additional insights into the land fragmentation process. An examination of other states’ laws may indicate how landowners would respond to amendments in the SCA and how land fragmentation might be affected if the SCA is changed. APPENDICES 99 APPENDIX A Sample Townships Included in Study J s s s m ............ Alam o OMOrOrai K alam azoo Low Aloha Austin C heboygan Lew M eoocta Low B angor Bay High B anian Berrien Medium Blackm an Jac k so n High B randon O akland Medium Brighton Livingston Medium B uchanan Berrien Medium C annon Kent Medium Cedarville M enom inee Low L eelanau Low Centerville Clay S aint Clair C layton Arenac Low Clinton M acomb High Medium C olum bia Van Boren Lew D a Witt Clinton Medium Detoi Dexter Ingham High W a sh te n aw Medium B aton R apids B bridga Eaton High Low F kw erfield O ceana S aint Jo sep h Freedom W ashtenaw Low G enesee G enesee High G enoa Livingston M edium Low G reen O ak Livingston Medium G unpiain Allegan Medium H am pton Bay Medium Handy Uvingston Medium H arnson M acomb Livingston High O sceola Lew Hartiand H eraey Medium H ighland O akland Medium Howard C ass Medium Independence Kimball Lansing O akland High S aint Clair M edium Ingham High Lincoln Berrien High Lowell Kent Medium M aple Valley S anilac Low M arcaSua C ass Low M entor C heboygan Low M eridian Ingham High MMord Oakland M onroe Monroe Medium High M ount Morris O gden G enesee High Lenaw ee Low Orion O akland High P entw ater O c ean a Low Q uincy B ranch Medium R aisin Lenaw ee Medium Richland R obinson Kalam azoo Medium O ttaw a Medium R ose O gem aw High S sg o la Dickinson Low S anborn Alpena Low Schoolcraft Kalam azoo M edium Hillsdale Low S d p io Solon Kent High Southfield Spring Lake Oakland High Ottaw a S t Clair S aint Clair Medium S t Ignace S urrey M ackinac Low Clare Low Talim adge Ottawa Medium T a w ts Iosco High T ecum seh Thetford L enaw ee Medium G e n esee Medium W e s t Bloomfield Oakland High W heatland W hite Lake Sanilac Low Oakland High Ypeilanti W ashtenaw High High 100 APPENDIX B Table 1 Number of Unplatted Lots Counted in Sample Townships in 1960 TcmraiuD Alano Aiofa. A utfn B ngor Bmxmx ntwJgwn Ekwdaa Gkihtfln Buchanan Omm GedwiUo Cmterrille C ky Gaytan Q ntaa C okm ba DeSu Dewitt Doxfeer Eaton Rapids Elhndge Flowar Field Freedom Qenwe* OON O ra n Oak GHnpkai Ih n y tu n ta J y TT. J IM S C Q Ife tk o d Hereey H o llan d Howard Kmbell Lm sng Lionofai Lowell Maple Vtilcy Mvoattue M ericr M aidian M ifcrd Monroe Mount Moms O tfim Qaoa Fulw aier Q m cy R am Rkfclmd RobtBoa Row Sagola 9mbcsn SdiDolcn& S c fio S okn Southfield SpnngLake S t Clair Stlgnaoe Suney Tattmadgc Tawas Teannaeh Thatford W a t Bloomfield W tenlfiid W hs. Lake Ypaiknti Yew1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 I960 1960 1960 1960 I960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 I960 1960 1960 I960 1960 1960 1960 I960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 I960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 1960 StzeofLoto snAaee 1 2 3 4 10 15 5 5 45 4 5 2 25 16 42 11 7 6 1 2 2 4 9 6 14 7 2 3 17 8 8 25 27 31 5 17 2 6 7 2 8 2 9 3 90 14 2 17 10 42 37 9 5 2 7 3 2 11 8 6 3 4 13 8 8 10 2 7 7 4 3 10 14 1 3 14 16 60 8 1 2 1 IS 1 50 16 4 I 12 1 21 17 59 11 8 9 19 1 1 5 2 3 2 7 2 1 2 4 3 6 1 3 3 7 7 12 51 2 5 2 16 4 5 5 10 6 8 6 14 5 14 9 33 12 4 5 10 1 I 1 5 2 2 2 3 3 17 2 3 1 2 6 1 1 1 13 11 33 5 I 5 2 15 13 8 7 75 5 10 6 17 107 9 15 15 13 7 5 23 8 18 20 17 51 11 15 23 23 22 7 12 12 16 2 20 20 3 1 30 25 10 15 3 4 6 1 3 6 2 5 2 4 8 6 9 5 5 5 38 21 21 7 5 2 2 I 4 5 1 15 20 20 25 4 20 21 18 8 5 5 2 7 2 3 5 4 9 5 5 4 12 4 1 3 8 6 4 15 11 15 2 7 8 1 1 10 17 9 4 3 2 2 13 3 4 8 9 3 5 4 3 3 1 8 7 5 5 3 5 5 3 7 10 12 20 15 15 9 4 4 -> 15 4 5 14 5 7 9 27 10 9 7 -> 10 5 20 22 20 10 8 6 9 8 9 8 9 6 5 11 5 10 10 8 3 1 2 2 6 3 2 1 10 18 3 12 3 4 2 1 2 1 30 22 4 34 8 10 9 2 1 44 1 1 3 1 33 11 4 8 4 6 1 17 3 1 20 3 2 7 23 12 2 6 3 4 2 3 1 5 2 1 20 3 3 1 22 10 9 10 7 9 1 3 57 5 5 7 2 6 72 5 3 22 8 5 14 31 7 9 4 3 18 6 1 U 10 5 4 15 6 5 6 3 2 *» 3 1 2 11 7 1 9 2 2 8 3 4 6 5 3 11 5 2 3 1 2 6 3 9 14 6 3 12 19 12 3 2 3 10 11 19 8 5 4 5 1 8 6 4 3 4 4 4 5 I I 1 2 4 2 1 8 3 11 6 7 11 1 2 13 22 -> 8 4 6 1 11 4 3 10 11 13 1 1 2 4 35 17 2 4 11 34 6 25 2 31 5 4 39 2 2 28 5 5 1 29 15 2 25 28 4 37 6 7 2 9 4 2 1 21 34 2 23 2 22 3 18 3 54 6 8 1 26 1 2 1 13 3 15 60 11 61 1 19 1 69 9 16 4 4 4 15 32 5 40 2 24 6 5 10 19 41 13 6 3 4 4 22 9 45 33 4 5 21 4 4 47 19 29 29 7 2 10 3 3 12 2 8 2 13 1 4 2 5 8 7 3 2 3 5 3 1 6 10 7 3 4 2 1 2 1 4 8 1 5 3 4 1 2 2 14 15 13 16 4 5 1 2 3 5 4 5 3 11 7 3 2 5 2 5 4 3 1 2 1 1 1 2 3 4 1 2 2 6 9 3 6 10 10 1 3 5 2 2 1 2 1 5 3 3 I 3 2 3 1 3 3 6 4 9 5 1 t 4 2 1 5 3 15 4 1 1 1 7 2 3 3 1 1 3 5 2 5 5 5 4 15 4 10 14 3 19 4 5 2 1 10 7 3 4 3 6 2 2 10 6 8 4 3 3 1 4 3 3 5 1 6 10 3 2 3 5 9 2 5 4 3 3 1 3 1 3 7 3 3 3 4 1 3 1 1 4 4 3 1 -» 2 6 1 1 1 5 1 2 4 4 ■y 3 1 1 5 1 I 4 5 4 3 3 1 2 7 1 9 1 3 1 1 3 *> 2 4 6 7 4 6 2 •y 5 3 2 2 2 4 19 2 5 4 4 2 5 5 3 4 4 2 14 I 3 19 4 5 6 4 3 ■y 5 2 3 4 1 2 3 2 4 4 3 2 5 17 18 3 5 1 3 1 3 2 1 5 2 3 3 15 3 5 1 2 3 5 14 3 3 9 4 3 5 3 4 t 3 2 3 3 1 2 I 2 4 3 8 2 5 3 4 2 2 3 1 1 3 1 3 2 2 4 2 2 4 5 5 6 2 3 4 2 1 7 2 1 2 3 2 2 15 8 6 2 6 •y 4 1 2 1 2 2 2 4 101 APPENDIX B Table 2 Number of Unplatted Lots Counted in Sample Townships in 1970 S ia ofLoto n Acres ToweviuD A tan o AJoba Au s tn Bmgar Bemen S ts d a a m tte m lrn Brighton Buchanan C anon Codavifle CeAavills 3 6 4 2 11 8 6 6 6 10 1 3 2 i 5 4 3 1 1 1 3 4 1 2 4 1 9 5 1 3 5 2 3 3 7 9 4 11 11 1 5 1 3 6 1 2 1 1 1 10 24 9 6 7 5 3 104 54 1 5 4 5 4 7 5 8 8 2 8 18 5 5 I 41 21 11 7 13 2 45 52 4 50 11 2 2 14 11 1 12 6 3 2 1 1 3 6 2 3 1 1 13 4 1 14 15 3 1 2 1 2 3 3 3 1 1 4 3 1 4 4 I 2 1 1 1 3 1 1 5 3 1 1 5 1 1 3 1 5 2 2 4 1 1 1 2 3 I 2 1 2 2 1 3 1 1 2 5 1 2 4 1 2 1 3 2 1 -> 2 4 2 4 3 2 4 2 1 5 3 1 1 4 3 1 9 2 3 1 3 1 2 1 3 4 1 2 2 I 1 1 1 2 2 1 2 2 3 1 3 1 2 2 2 4 1 4 1 1 2 3 1 1 1 3 1 8 1 7 3 1 5 1 1 2 3 1 3 3 2 1 1 5 3 4 2 2 2 2 4 2 2 1 1 2 4 1 3 1 6 1 1 I 3 1 5 1 4 5 1 5 2 1 3 6 2 4 I 3 1 3 1 5 2 3 2 1 1 2 I 1 1 2 5 6 6 2 2 *> 2 1 1 2 1 2 2 2 1 2 3 1 1 2 4 4 3 4 4 3 3 9 2 1 6 2 1 5 1 1 5 6 1 2 1 1 2 1 I 1 2 4 1 1 I 1 2 3 2 2 3 I 6 2' 2 1 1 1 3 1 I 1 1 1 1 1 3 1 1 1 3 1 4 1 1 1 1 1 2 1 1 3 1 1 2 1 1 1 2 1 1 2 2 1 1 1 1 2 2 1 6 1 9 3 1 1 1 2 1 2 6 4 -> 28 16 2 1 3 2 1 1 I 1 1 3 3 6 1 5 5 1 I 1 1 3 1 2 1 12 10 1 2 1 1 1 3 2 1 1 1 I 3 3 1 1 4 6 12 15 95 6 5 1 9 13 6 2 11 3 6 20 3 2 ! 2 4 3 60 3 7 1 15 1 •) 3 3 2 2 1 3 1 1 1 1 2 1 19 2 1 1 2 3 1 4 1 18 1 2 5 2 1 1 1 17 1 I 1 1 2 16 2 1 3 1 5 3 8 8 6 4 2 7 19 2 4 9 4 2 11 1 1 12 9 1 2 I 3 15 30 8 4 1 6 7 3 7 4 1 9 10 15 5 6 23 5 1 3 1 12 1 3 4 21 10 59 6 5 1 8 5 3 1 2 3 6 13 2 11 2 11 1 11 2 10 4 3 3 1 6 1 3 1 20 3 12 5 1 5 1 7 1 1 I 3 4 2 4 1 102 APPENDIX B Table 3 Number of Unplatted Lots Counted in Sample Townships in 1980 S isc o fL o to m A erw 1 4 2 3 14 11 15 t o 1980 S 6 7 25 8 7 8 6 1980 5 18 1 1980 3 3 1980 11 4 Y e* 3 8 6 8 5 5 1980 25 1 28 2 15 13 15 2 6 1 I 3 5 1980 1980 20 24 20 2 12 4 11 21 5 20 12 5 4 6 4 5 15 20 21 310 47 3 10 17 35 10 32 14 6 4 15 6 16 26 7 15 5 13 4 2 I 12 4 2 2 2 2 5 5 1 S 2 2 2 6 6 3 2 8 3 1 3 3 6 1980 1980 1980 1980 1980 2 2 1980 1980 1960 2 1 4 3 20 9 4 4 4 5 3 1 9 10 12 4 13 4 15 16 17 50 11 4 14 8 1 5 3 1 1 2 4 12 28 1 1 1 4 22 25 247 1 3 1 4 1 y 6 4 3 4 7 3 88 17 60 4 14 2 7 1 5 1 6 •> 1 3 13 1 2 10 10 11 9 1980 1980 22 17 10 20 7 12 20 11 6 9 3 24 6 3 7 2 3 7 1 4 5 3 5 61 1980 46 45 31 38 17 8 8 12 6 1 1 15 2 2 10 3 13 11 3 50 19 1980 11 1 2 1 18 1980 43 12 3 23 7 4 2 4 1 4 2 2 1 3 2 2 1 1 I 1 3 25 2 4 3 1 2 3 4 2 4 2 2 2 1 1 20 1 2 12 6 5 88 4 4 64 3 7 1980 1980 38 64 11 5 6 2 4 1 1 1 5 38 7 4 10 27 1 10 11 10 7 8 4 1980 3 6 39 2 1 1 4 5 1980 1980 24 24 131 7 12 3 6 6 3 3 12 26 14 5 6 20 4 14 I9 6 0 21 10 1980 I9 6 0 15 14 15 18 10 4 29 13 5 19 23 4 1960 28 25 28 1980 15 12 28 17 13 41 2 5 5 14 6 10 7 20 11 38 10 17 2 1 4 1 1 4 30 2 4 4 y 1 4 3 2 1 1 2 5 1 2 1 1 4 2 2 1 3 3 3 2 2 3 7 7 12 4 6 8 4 73 5 12 4 1 1 1 7 3 2 1 2 2 7 1 ■> 4 3 T 3 2 1 1 1 3 1 4 9 2 1 5 5 y 2 2 4 3 4 3 3 5 3 3 2 3 8 11 1 5 5 68 6 2 22 2 2 19 1 5 4 3 13 9 3 1 2 1 1 4 2 2 3 1 2 1 112 9 10 3 1 I 1 9 4 2 8 1 20 2 1 1 2 19 4 3 2 3 1 1 1 2 1 3 1 6 8 10 14 3 1 5 I9 6 0 2 8 1 3 1980 11 10 8 1980 8 5 1 2 4 13 2 6 1 I 1980 1980 105 55 57 55 60 6 6 9 3 7 2 2 3 2 2 3 5 13 1 3 3 1 10 10 10 10 5 5 3 29 3 4 10 10 3 5 2 7 2 9 4 5 I 3 2 2 1 1 1 1 3 3 20 78 77 1 1 2 9 3 4 1 4 4 6 2 41 30 20 2 2 1 4 2 4 1 2 2 4 1 2 2 18 6 2 1 4 2 4 1 1 2 I 5 1 4 1 2 2 1 1 1980 1980 I9 6 0 10 5 3 3 1980 27 2 4 4 9 5 1980 12 10 15 15 10 14 7 11 37 15 33 7 1 11 21 51 17 15 5 4 1 6 1 5 1980 1980 1980 1980 1980 1980 1980 1980 I9 6 0 1980 1980 29 7 4 9 1 13 2 5 10 7 10 11 9 9 3 6 5 3 3 13 5 3 8 20 11 23 4 3 11 13 16 2 7 2 10 1 7 5 2 4 51 4 1 1 5 1 1 2 3 21 8 2 6 17 1 14 1 11 3 1 1 1 20 131 10 28 5 7 2 26 40 1 2 3 9 3 18 1 3 1 5 1 5 6 39 1 1 2 1 y 1 3 3 y 2 5 T 1980 42 12 40 18 33 7 1980 15 15 12 1980 23 30 25 10 20 8 10 20 10 4 2 9 13 8 43 2 1 3 3 14 36 15 3 13 1980 1980 1 2 8 10 1980 9 3 4 3 1980 5 1 5 24 3 y 10 3 3 15 6 13 4 13 3 15 10 33 5 3 3 1 2 2 7 1 1 1 5 y 4 1 1 3 2 5 1 3 1 1 1 1 1 2 2 1 1 1 1 2 1 3 3 1 6 3 1 1 2 1 1 2 6 2 11 1 y 1 1 4 2 3 U 6 1 1 2 2 1 y 2 1 2 3 1 t 2 1 2 1 1 1 1 2 4 6 3 2 12 5 1980 1980 1 6 1 2 26 5 6 7 1980 4 2 1 1 3 2 2 1980 1980 1 1 5 6 1 9 4 I 3 1 1 111 28 50 4 2 4 20 1 5 5 20 1 1 4 4 26 2 3 4 8 4 1960 1 1 t -> y y 58 10 1980 2 2 2 4 42 11 3 1 2 4 2 1 1 2 2 2 1 4 11 4 2 1 9 I 3 1 2 2 10 1 2 1 y 2 15 9 5 1 7 1 10 4 4 2 4 13 1 4 1 6 7 2 2 5 1980 5 6 1 2 3 8 1980 1 4 3 1 2 7 3 2 y 3 19 1980 3 2 3 3 2 1980 3 8 2 8 y 3 13 1 4 1 3 1 1 1 12 1 5 8 1980 1980 8 4 4 7 1980 4 15 8 7 6 16 29 19 y 1 3 6 18 1 1 5 2 2 2 4 2 2 2 1 1 1 1 48 5 4 12 3 2 2 1 2 4 4 2 y 1 103 APPENDIX B Table 4 Number of Unplatted Lots Counted in Sample Townships in 1990 Ym 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 1990 3iza ofLoto a iA a w 1 2 3 4 15 15 15 15 2 4 2 7 2 I 1 9 2 4 17 17 13 8 8 10 3 3 5 4 3 1 2 4 5 5 2 1 I 7 10 10 2 5 3 16 9 1 10 7 10 10 25 10 7 10 15 10 9 29 23 20 10 15 5 15 15 10 10 31 25 5 6 5 5 5 15 15 10 10 7 5 23 6 5 7 20 10 11 15 10 15 11 10 10 14 12 4 4 7 10 10 15 2 5 14 6 10 5 5 10 2 2 3 2 10 10 13 20 25 9 3 5 5 25 1 2 1 2 2 1 5 5 3 6 5 2 20 20 20 10 U 10 6 5 18 1 1 4 4 6 23 23 2 7 6 3 8 5 9 5 3 4 6 2 5 2 3 3 2 2 3 3 8' 7 7 5 4 10 2 5 6 5 6 5 16 10 5 2 2 11 3 9 17 4 26 7 2 13 2 4 16 20 16 8 10 3 25 1 30 11 6 1 10 6 11 1 2 3 2 2 2 3 2 5 2 4 4 3 6 4 5 4 2 5 2 3 3 5 2 6 4 2 3 2 3 4 6 2 2 2 14 2 2 2 5 6 4 3 7 2 3 3 1 27 36 7 4 20 15 15 10 20 20 21 13 25 30 25 25 1 6 4 3 3 5 3 2 17 14 8 4 10 11 13 9 10 12 11 9 18 15 15 10 3 10 15 17 19 13 7 8 3 10 28 10 10 3 5 19 16 5 21 3 41 15 8 1 4 4 14 10 1 1 1 15 2 3 8 6 2 4 9 3 4 8 2 4 5 3 2 3 2 2 3 3 3 5 2 6 8 2 2 3 5 3 5 5 2 2 2 10 25 3 7 3 8 14 43 83 4 76 10 13 14 9 3 25 23 8 47 24 4 21 22 10 87 69 44 16 41 11 24 74 37 22 32 9 48 2 4 78 13 17 25 1 28 5 35 2 4 3 3 30 39 50 54 11 9 9 5 65 4 2 3 2 6 3 2 12 5 13 5 1 1 1 3 5 2 2 5 2 5 1 14 3 1 7 4 1 2 3 9 1 2 1 1 3 2 1 1 3 2 1 2 2 3 3 1 2 1 1 2 1 1 4 4 5 1 4 6 2 3 2 4 2 I 2 1 1 1 3 2 3 2 3 4 1 4 1 1 1 2 1 1 1 2 3 2 I I 1 2 2 1 I 1 1 3 1 1 3 4 2 3 4 4 6 3 3 3 3 2 1 1 1 1 1 2 1 2 1 1 1 2 1 I 2 2 3 1 1 1 3 1 1 2 2 1 2 2 1 1 2 3 4 2 3 1 2 6 3 1 3 3 3 1 3 2 I 1 I 1 2 1 2 3 2 1 2 1 2 4 1 2 1 9 3 1 3 1 1 1 3 1 1 1 2 1 1 1 1 1 1 3 2 1 1 2 2 1 1 I 4 1 2 2 n 3 3 12 2 2 3 2 2 1 5 1 1 4 3 4 4 19 3 1 5 7 1 3 5 1 9 5 1 4 4 1 6 4 7 1 1 5 16 17 18 1 2 1 1 5 3 10 8 15 4 2 I 2 1 2 1 1 1 1 1 3 3 1 1 1 5 1 1 1 2 1 4 1 2 5 19 1 2 1 3 2 4 2 1 12 8 10 10 10 6 10 10 10 6 9 1 7 6 8 19 8 17 16 5 5 20 10 6 8 5 10 10 2 8 10 2 4 3 6 2 3 3 3 4 80 15 30 37 20 15 22 4 7 11 3 2 2 1 1 1 1 3 1 4 10 27 1 2 3 6 2 1 2 1 2 1 5 2 7 1 1 2 1 1 1 1 7 1 3 1 I 1 2 1 3 2 1 1 3 1 *> 1 1 1 *) T 1 4 I 3 1 1 1 5 3 104 APPENDIX C Results From F-tests Required For Pooling F= (SSEC-SSE1-SSE2-SSE31/K (SSE1 +SSE2+SSE3)/(N+M+P-3K) F 10,40 = 2.132 F= EQUATION..! SSE1 = SSE2= SSE3= SSEC= 15,033 301,245 99,528 807,846 K= N= M= P= SSE1 = SSE2= SSE3= SSEC= 15,013 295,684 98,832 803,677 K= N= M= P= 14,961 294,412 98,504 800,969 K= N= M= P= SSE1 = SSE2= SSE3= SSEC= 13,022 291,966 95,605 767,953 K= N= M= P= 3.24 11 21 27 22 F= EQUATION 4 3.24 11 21 27 22 F= EQUMIQN 3 SSE1 = SSE2= SSE3= SSEC= 10 21 27 22 F= EQ.UM.IQN 2 3.77 3.08 11 21 27 22 105 APPENDIX D Table 1 Statistics of Variables Used in Regression Models; MEDAGE and INCOME Townshio Density Alamo Aloha Austin Bangor Berrien Blackman Brandon Brighton Buchanan Cannon Cedarville Centerville Clay Clayton Clinton Columbia DeWItt Delhi Dexter Eaton Rapids ElbrkJge Floweffieki Freedom Genesee Genoa Green Oak Gunpiain Hampton Handy Harrison Hartland Mersey Highland Howard Independence Kimball Lansing Lincoln Lowell Maple Valley Marcetlus Mentor Meridian Milford Monroe Mount Morris Ogden Orion Pentwater Quincy Raisin Richland Robinson Rose Sagota Sanborn Schoolcraft Scipio Solon Southfield Spring Lake St. Clair St. (gnace Surrey Tallmadge Tawas Tecumaeh Thetford West Bloomfield Wheatland White Lake Ypsilanti Low Low Low High Medium High Medium Medium Medium Medium Low Low Medium Low High Low Medium High Medium High Low Low Low High Medium Medium Medium Medium Medium High Medium Low Medium Medium High Medium High High Medium Low Low Low High Medium High High Low High Low Medium Medium Medium Medium High Low Low Medium Low High High High Medium Low Low Medium High Medium Medium High Low High High MEOAGE INCOME 1960 1970 1980 1990 1960 1970 28.3 27.0 27.3 26.2 28.9 31.0 24.4 26.8 27.3 27.0 42.1 32.3 29.6 27.4 22.9 34.5 24.7 28.6 26.1 28.5 25.8 312 24.6 30.2 30.2 23.5 28.3 22.2 27.3 25.0 26.7 26.3 25.8 25.6 23.8 24.9 25.0 33.9 29.3 27.7 25.2 24.7 31.1 37.6 23.7 26.5 21.5 25.0 27.5 39.1 29.9 3Q.7 24.7 24.5 23.9 33.4 38.8 26.6 26.6 25.3 23.1 30.2 24.3 24.1 26.3 29.3 24.1 31.8 28.6 26.2 25.1 26.3 26.0 26.4 26.0 25.2 27.8 29.1 24.1 25.3 26.2 25.1 41.4 32.0 32.7 26.5 22.7 31.6 23.9 26.5 24.5 26.7 24.9 25.8 23.1 28.2 28.5 22.2 27.6 21.4 25.8 24.B 25.2 25.1 25.5 25.2 23.5 27.5 23.1 32.6 27.2 25.2 24.8 24.2 28.8 37.2 23.1 24.8 21.3 24.7 26.5 38.1 29.7 28.2 23.9 26.5 24.3 33.3 35.6 25.7 26.3 25.0 22.4 33.4 25.9 26.1 25.5 27.4 23.9 29.7 28.3 23.8 24.8 24.7 23.2 33.3 30.2 28.6 30.9 30.8 26.3 29.1 31.9 29.4 46.8 29.6 32.5 27.8 27.2 32.6 29.4 27.6 30.1 29.3 27.5 29.2 31.1 27.0 29.6 25.8 29.7 28.4 27.0 28.4 27.9 28.6 26.3 31.1 29.3 26.8 29.8 29.7 26.8 29.6 29.8 33.6 28.0 28.9 27.8 25.5 29.0 27.5 41.7 29.2 27.2 30.6 27.2 33.1 31.4 28.0 29.3 28.2 28.1 39.5 30.6 29.6 31.0 32.5 26.9 33.2 32.8 25.9 32.1 29.9 27.5 26.3 36.2 41.2 34.4 34.9 34.7 34.5 30.6 32.9 32.1 32.6 44.8 35.1 37.1 33.0 32.8 34,9 34.8 32.1 34.5 34.6 30.5 31.4 36.6 32.2 35.2 32.8 33.1 35.5 29.5 32.8 33.7 32.3 30.9 35.8 34.4 31.0 34.3 34.7 30.6 31.7 33.2 39.9 32.3 33.2 31.9 30.9 32.3 32.3 45.1 32.5 31.9 34.9 30.0 36.4 36.4 34.4 33.8 31.8 32.4 42.0 34.1 33.6 31.6 36.0 31.3 39.3 36.0 30.8 36.7 30.6 32.4 29.8 2.513 1.578 1.672 2,787 2,243 2,200 2,468 2.834 2,775 3.620 1,784 1,645 3,601 1.527 3.079 1,875 2,678 3,138 2,882 2,321 1.552 2,149 2,568 2,573 2,710 2,424 2,593 2.680 2,277 3,376 2,451 2,157 2,659 2,321 2,989 2,751 3,499 3,051 2,182 1,652 2,019 2.135 3,502 2,912 2,750 2,518 2,222 2.865 2,044 2,762 2,357 3,111 2,489 1,867 2,072 2,383 2.714 1,809 1,790 6,430 3,298 2.755 2,200 1,630 2,413 2,231 2.461 2,633 4,392 1,296 3,031 2.763 2,927 2,045 2,380 3,297 2,456 2,514 2,958 3,682 3,038 4,006 2,213 2,293 3,408 1,988 3.569 2,219 3,383 3.706 3,678 2,987 1,826 2,562 3,278 2,942 3,522 3,150 2,903 3,170 2,958 3,913 3.184 2,752 3,187 3,024 3,582 2,604 4.132 3,340 2,415 2,538 2,631 2,787 4.138 3,490 3,450 2,880 2,695 3,434 2,405 2,219 2.858 3,622 2,807 2,441 2,526 2,587 3,160 2,126 2,496 7,707 3.719 2,608 2,450 2.080 2,721 2,321 2,984 3.012 5,263 1,989 3.632 3,527 1980 7,248 5,154 5,439 7,557 6,047 5,863 8,068 9,935 6,800 8,824 6,068 6,208 7.759 4,611 8.862 5.226 8,135 8,038 9,092 7,546 5,913 8,640 8,085 7,417 9,077 7,934 8,074 7,355 6,700 9,428 9,069 5,101 7,713 6.837 9,874 5.750 8,838 8,192 7,019 5,567 6,044 5.128 9.997 9,057 7.340 6,985 7,218 9,022 7,275 6,363 7.151 9,676 6,208 4,480 5,339 5.845 7,818 8,130 6,329 18,063 8,173 8,742 5,162 5,402 7,249 5,491 9.098 7,621 15,124 5,963 8,706 8,527 1990 14,324 10,056 10.462 14,776 12,352 10,291 15,874 20,360 14,044 20,013 11,165 11,852 16,848 7,578 16,864 15,854 15,275 14,505 19,080 14,194 8.360 12,793 17,611 12,531 21,274 17,272 14,183 12,796 13,194 18,183 17,690 8,850 15,716 11,885 21,271 11,776 15,105 16,231 14,439 11,595 12,605 7,851 20.728 17,745 13,277 12,092 11,336 17,773 13,144 11,794 12,790 19,401 13,102 8,024 9,862 10,479 14,315 9,950 11,494 38,418 17,606 15,604 8,503 8,985 14,203 9,965 17,959 13,976 31,845 9,992 16,750 14,977 106 APPENDIX D Table 2 Statistics of Variables Used in Regression Models; TOTLAB and EMP Township TOTLAB 1960 1970 Alamo Aloha Austin Bangor Berrien Blackman Brandon Brighton Buchanan Cannon Cedarville Centerville Clay Clayton Clinton Columbia DeWHt Delhi Dexter Eaton Rapids Elbridge Flowerfield Freedom Genesee Genoa Green Oak Gunpiain Hampton Handy Harrison Hartland Hersey Highland Howard Independence Kimball Lansing Lincoln Lowell Maple Valley MarceMus Mentor Meridian Milford Monroe Mount Morris Ogden Orion Pentwater Quincy Raisin Richland Robinson Rose Sagola Sanborn Schoolcraft Scipio Solon Southfiold Spring Lake St. Clair St. Ignace Surrey Tallmadge Tawas Tecumseh Thetford West Bloomfi Wheatland White Lake Ypsilanti 748 168 130 5,722 1,263 4,681 1,259 1,454 1,272 1,190 76 206 34,564 190 10,899 560 3,101 4,344 580 510 269 183 385 7,735 1,116 1,764 1,093 2,194 830 3,867 596 282 2,117 1,809 4,636 33,283 3,931 4,170 628 80 651 40 7,865 1,876 3,021 8,838 454 4,606 303 1,320 1,454 1,146 609 213 297 583 1,731 417 206 4,484 2,174 14,917 236 540 1.313 408 359 1,853 7,241 47 3,912 9,889 949 206 183 6,382 1,413 5,295 1,802 2,302 1,423 1,415 78 252 2,375 213 18,190 650 4,166 5,845 866 736 330 257 574 9,245 1,767 2,794 1,331 2,447 1,315 6,453 944 320 3,029 2,313 6,635 2,287 5,289 4.666 747 307 832 49 10,582 2,684 3,781 10,564 529 6,591 346 90 1,693 1,454 850 243 289 598 2,197 477 252 6,417 3,036 1,025 199 778 1,833 508 418 2,215 10,363 180 5,599 14,759 1930 1,385 294 339 8,154 1,816 6,777 4,226 5,294 1,646 2,617 89 307 3,530 319 35.321 823 5,016 8,723 1,923 1,285 367 581 723 10,807 4,243 4,872 2,127 4,655 1,978 11,517 2,696 464 7,787 3,008 10,387 2,872 5,539 6,904 1,786 407 1,085 196 15,675 4,869 5,038 11,511 506 11,151 562 1,953 2,471 2,507 1,408 411 418 941 3,200 589 465 7,253 4,902 1,675 324 1,167 2,977 603 750 3,767 20,461 261 10,593 23,532 1930 1,600 320 455 8,191 2,333 6,880 6,547 7.905 1,805 4,260 87 402 4,484 358 46,602 3,197 5,884 10,815 2,443 1,620 381 757 790 10,798 5,823 6,228 2,517 4,706 2,669 13,611 3,763 584 9,397 3,323 13,936 3,632 5,003 7,399 2,579 492 1,285 195 20,084 6,510 5,676 2,773 515 13,653 608 1,994 2,824 2,906 2,059 418 493 1,049 3,565 706 635 7,090 5,897 2,088 435 1,218 3,426 679 839 4,299 30,055 241 12,297 25,565 TOTEM3 1960 1970 729 152 119 5,230 1,216 4,444 1,125 1,366 1,186 1,108 58 193 32,939 176 10,171 533 2,918 4,190 560 491 241 183 371 7,243 1,022 1,645 1,058 2,074 764 3,583 569 279 2,000 1,732 4,278 32,380 3,842 3,983 595 78 60S 35 7,584 1,776 2,829 8.255 449 4,271 284 1,289 1,342 1,119 592 180 257 539 1,678 399 180 4,383 2,115 13,840 193 512 1,222 368 337 1,714 6,964 40 3,569 9,173 919 186 164 5,943 1,368 5,066 1,608 2,186 1,332 1,310 60 237 2,160 193 17,350 618 3,960 5,575 834 714 283 252 553 8,724 1,636 2,633 1,271 2,330 1,223 6,111 911 303 2,859 2,235 6,115 2,143 5,112 4,474 703 293 782 43 10,091 2,539 3,658 9,943 523 6,106 334 63 1,563 1,411 813 213 259 557 2,117 446 221 6,266 2,906 516 170 715 1,679 468 392 2,064 9,956 149 5,102 13,661 1980 1,318 222 299 7,128 1,627 6,095 3,782 4,845 1,470 2,422 79 244 2,916 269 31,735 677 4,709 8,097 1,837 1,187 313 541 674 9,190 3,763 4,407 1,963 4,106 1,680 10,327 2,510 422 6,827 2,745 9,694 2,473 5,199 6,444 1,659 351 960 148 14,935 4.442 4,423 9,696 465 10,225 486 1,685 2,243 2,428 1,262 352 372 759 3,003 498 408 7,030 4.548 1,479 247 995 2,821 517 695 3,260 19,498 240 9,630 20,621 1990 1,492 275 415 7,557 2,214 6,381 6,091 7,624 1,695 4,150 79 363 4,193 301 43,783 2,946 5,617 10,209 2,408 1,563 328 688 773 9,569 5,628 5,959 2,393 4,247 2,440 12,761 3,537 517 8,859 3,115 13.283 3,280 4.702 7,133 2,465 449 1,200 176 19.428 6,169 5,195 2,555 491 12,929 560 1,850 2,615 2,796 1,978 378 447 950 3,369 647 614 6,881 5,651 1,952 372 1,094 3,300 641 816 3,901 29,007 224 11,533 23,948 107 APPENDIX D Table 3 Statistics of Variables Used in Regression Models; TOTPOP and TOTHOU Townshio Density Alamo Aloha Austin Bangor Borneo Blackman Brandon Brighton Buchanan Cannon Cedarville Centerville Clay Clayton Clinton Columbia De Witt Delhi Dexter Eaton Rapids Elbridge Flowerfield Freedom Genesee Genoa Green Oak Gunplain Hampton Handy Harrison Hartland Hersey Highland Howard Independence Kimball Lansing Lincoln Lowell Maple Valley Marcellus Mentor Meridian Milford Monroe Mount Morris Ogden Orion Pentwater Quincy Raisin Richland Robinson Rosa Sagola Sanborn Schoolcraft Scipio Solon Southfield Spring Lake St. Clair St. Ignace Surrey Tallmadge Tawas Tecumseh Thetford West Bloomfield Wheatland White Lake Ypsilanti Low Low Low High Medium High Medium Medium Medium Medium Low Low Medium Low High Low Medium High Medium High Low Low Low High Medium Medium Medium Medium Medium High Medium Low Medium Medium High Medium High High Medium Low Low Low High Medium High High Low High Low Medium Medium Medium Medium High Low Low Medium Low High High High Medium Low Low Medium High Medium Medium High Low High High TOTPOP 1960 1970 1.905 274 434 11.666 3,183 16,060 3,187 2,875 2.410 2,525 218 577 6,948 572 25,688 1,619 7,649 16,590 1,698 1,597 767 722 1.065 21,011 2.402 4,631 2,796 5,387 2,890 12,910 1,436 645 4,855 4,622 10,890 6,266 14,387 4,462 1,567 765 1,814 202 13,884 5,871 8,343 20,633 1,305 11,844 1,146 3,129 3,061 2,574 1,618 566 952 1.413 4,418 1,069 2,422 11,319 8,016 2,416 686 1,653 3,243 1,104 775 3.843 14,994 544 8,381 25,950 2,492 643 492 15,990 3,905 17,060 4,813 5,882 3,252 3,690 158 629 6,337 766 48,910 1,951 9,945 13,730 2,225 1,985 918 729 1,227 25,576 4,623 7,598 3,231 6,774 3,630 18,793 2,574 718 8,372 5,497 17,377 5,983 11,145 10,900 2,109 737 2,033 141 23,818 7,256 9,351 29,349 1,465 17,096 1,175 233 4,248 3,728 2,025 816 946 1,625 5,289 1,283 735 17,495 8,013 3,303 547 2,291 4,883 1,517 1,165 5,970 28,574 554 14,292 33,278 1980 2.909 726 898 17,494 4,302 19,741 9,526 11,222 3,571 4,983 212 709 8.518 987 72,400 2.248 10.038 17,144 3.872 2.823 899 1,290 1.436 25,065 9,261 10,802 4,298 10,418 4,681 23,649 6.034 1,229 16.958 6.524 21,537 7,180 10.097 13,520 3,972 1,009 2,463 462 28,754 10,187 11,654 27,928 1,224 22.473 1,424 3,929 5,499 4.677 3,018 1,085 1,146 2,297 6,435 1,352 987 15,031 9,588 3,965 706 3,101 5,927 1,463 1,480 8,499 41,982 582 21,870 44,511 1990 3,278 707 1,102 16,028 4,697 20,492 12,051 14,815 3,402 7,928 185 838 8,882 908 85,686 6,308 10,448 19.190 4.407 3,003 820 1,418 1,486 24,093 10,620 11,604 4,754 9,520 5,488 24,685 6,860 1,455 17,941 6,376 24,722 7,247 8.919 13,604 4,774 1,022 2,450 518 35,644 12,121 11,909 6,236 1,146 24,076 1,422 4,003 5,848 5,099 3,925 1,260 1,166 2,196 6,705 1,479 1,268 14,255 10,751 4,614 932 3,221 6,293 1,465 1,539 8,333 54,518 513 22,608 45,307 TOTHOU 1960 1970 493 88 111 3,275 914 2,988 822 787 669 676 65 134 2,193 168 6,468 504 2,053 4,479 476 435 215 215 315 5,343 719 1,271 733 1,357 854 3,325 399 194 1,323 1,274 2.711 1,594 4.071 1,331 449 218 582 65 3,735 1,568 2,285 5,177 385 3,170 394 981 788 704 424 179 295 368 1,355 290 658 3,167 2,294 665 199 452 629 299 211 963 3,929 141 2,226 6,534 641 163 148 4,656 1,074 3,697 1,277 1,691 948 986 76 137 2,095 205 13,026 572 4.108 787 2,900 646 578 334 219 257 384 6,761 1,457 2,028 880 1,808 1,064 5,519 724 227 2,304 1,697 1,671 178 3,705 3,260 630 247 618 80 6,428 1,970 7,585 5,301 348 415 1,077 94 1,201 1,061 562 267 306 468 1,623 333 583 4,833 2,337 910 159 691 1,309 392 300 1,613 7,333 4,025 1980 9,397 252 298 5,903 1,284 5,325 2,898 3,349 1,204 1,578 94 236 3,000 300 23,908 780 3,440 6,053 1,256 849 277 432 486 8,305 2,996 3,193 1,412 3,765 1,580 8,677 1,744 402 5,264 2,242 6,652 2,209 4,316 4,777 1,182 329 842 157 10,952 3,260 4.121 8,706 408 7,487 594 1,442 1,668 1,596 905 402 402 762 2,247 442 327 5,110 3,380 1,283 226 1,091 1,770 463 474 2,715 12,877 185 7,037 16,162 1990 1,131 283 391 6,039 1,544 5.988 3,980 4,711 1,195 2,574 87 300 3,354 326 32,594 2,412 4,043 7,014 1,584 995 284 473 540 8,830 3,684 3,859 1,594 3,799 1,873 9,951 2,196 505 5,898 2,324 8,410 2,430 4,029 5,262 1,481 327 883 202 13,989 4,159 4,469 1,876 385 8,548 612 1.480 1,826 1.870 1,165 459 439 805 2,532 509 431 5,530 4.121 1,575 344 1,222 1,972 532 533 2,825 19,215 177 7,834 17,743 108 APPENDIX D Table 4 Statistics of Variables Used in Regression Models; AGP, DFREE, DCITY, REC, and LAND Townshio Density Alamo Aloha Austin Bangor Berrien Blackman Brandon Brighton Buchanan Cannon Cedarville Centerville Clay Clayton Clinton Columbia De WHt Delhi Dexter Eaton Rapids Elbridge FlowerileJd Freedom Genesee Genoa Green Oak Gunplain Hampton Handy Harrison Hartland Hersey Highland Howard Independence Kimball Lansing Lincoln Lowell Maple Valley Marcelius Mentor Meridian Milford Monroe Mount Morris Ogden Orion Pentwater Quincy Raisin Richland Robinson Rose Segda Sanborn Schoolcraft Scipto Solon Southfield Spring Lake St. Clair St. tgnace Surrey Tallmadge Tawas Tecumseh Thetford West Bloomfield Wheatland White Lake Ypsilanti Low Low Low High Medium High Medium Medium Medium Medium Low Low Medium Low High Low Medium High Medium High Low Low Low High Medium Medium Medium Medium Medium High Medium Low Medium Medium High Medium High High Medium Low Low Low High Medium High High Low High Low Medium Medium Medium Medium High Low Low Medium Low High High High Medium Low Low Medium High Medium Medium High Low High High DFREE 1d60 1970 DCITY 1980 AGP REC LANO 1990 3.0 10.0 3.0 3.0 160.0 12.0 12.0 12.0 60.0 6.0 6.0 6.0 30.0 3.0 3.0 3.0 15.0 15.0 15.0 15.0 3.0 0.5 0.5 0.5 30.0 7.0 7.0 7.0 5.0 3.0 3.0 3.0 18.0 12.0 12.0 12.0 13.0 7.0 7.0 7.0 260.0 160.0 160.0 160.0 160.0 60.0 60.0 60.0 40.0 9.0 9.0 9.0 70.0 8.0 8.0 8.0 15.0 3.0 3.0 3.0 14.0 12.0 12.0 12.0 6.0 7.0 1.0 1.0 15.0 1.5 1.5 1.5 16.0 6.0 6.0 6.0 30.0 15.0 9.0 9.0 15.0 15.0 6.0 0.0 15.0 4.0 4.0 4.0 6.0 11.0 6.0 6.0 4.0 4.0 4.0 4.0 6.0 2.0 2.0 2.0 3.0 2.0 2.0 2.0 6.0 2.5 2.5 2.5 32.0 6.0 6.0 6.0 20.0 1.0 1.0 1.0 16.0 2.0 2.0 2.0 10.0 1.0 1.0 1.0 75.0 38.0 8.0 8.0 6.0 6.0 6.0 6.0 30.0 23.0 23.0 23.0 20.0 1.0 1.0 1.0 45.0 3.0 3.0 3.0 4.0 1.0 1.0 1.0 4.0 1.5 1.5 1.5 1.5 1.5 1.5 1.5 40.0 30.0 20.0 15.0 16.0 16.0 16.0 16.0 150.0 5.0 5.0 5.0 20.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 2.0 2.0 2.0 2.0 0.5 0.5 0.5 0.5 10.0 10.0 10.0 10.0 26.0 4.0 4.0 4.0 18.0 18.0 3.0 3.0 22.0 5.0 5.0 5.0 15.0 15.0 15.0 15.0 9.0 9.0 9.0 9.0 8.0 3.0 3.0 3.0 20.0 20.0 14.0 14.0 300.0 150.0 150.0 150.0 SO.O 60.0 60.0 60.0 8.0 3.0 3.0 3.0 16.0 12.0 12.0 12.0 24.0 3.5 3.5 3.5 12.0 5.0 5.0 2.0 15.0 4.0 4.0 4.0 40.0 0.5 0.5 0.5 105.0 5.0 5.0 5.0 75.0 1.0 1.0 1.0 4.0 4.0 4.0 4.0 90.0 30.0 30.0 30.0 14.0 14.0 14.0 14.0 7.0 7.0 7.0 7.0 10.0 5.0 5.0 5.0 65.0 55.0 45.0 37.0 11.0 11.0 11.0 11.0 1.0 1.0 1.0 1.0 9.3 160.0 42.9 15.7 42.4 29.1 14.0 17.9 50.5 7.0 260.0 160.0 21.3 48.7 5.6 24.3 7.1 5.4 13.6 14.5 57.8 18.1 12.2 5.9 24.2 13.5 13.1 15.5 23.8 10.2 20.8 60.7 16.8 41.2 8.4 57.5 3.3 49.5 12.2 40.0 20.5 150.0 6.9 15.0 24.1 5.9 36.6 7.3 68.9 40.0 25.1 9.3 20.5 69.2 300.0 104.4 11.1 33.3 17.8 3.9 27.1 30.7 105.0 56.8 9.6 61.7 21.9 11.1 5.8 57.2 10.8 16.8 47.51% 9.87% 21.75% 10.00% 32.60% 70.34% 34.65% 22.43% 28.61% 27.80% 8.60% 45.14% 29.06% 2.81% 32.45% 10.58% 46.05% 10.91% 27.76% 30.07% 48.35% 30.01% 61.63% 48.24% 55.60% 62.04% 18.55% 19.28% 16.44% 50.17% 67.32% 61.43% 0.66% 27.43% 26.65% 21.11% 54.93% 9.87% 26.97% 3.71% 38.29% 42.37% 88.55% 57.08% 4.08% 20.65% 11.55% 57.81% 29.69% 91.28% 12.99% 0.91% 73.87% 73.23% 54.15% 36.75% 14.42% 2.15% 30.10% 59.60% 67.79% 45.71% 7.32% 41.87% 19.62% 23,332 21.71% 22,383 23.82% 22,117 23.76% 22,863 38.38% 23,872 81.01% 50,693 39.34% 23,063 22.08% 18,215 40.03% 20.563 5.46% 18,114 46.02% 22,657 18.44% 23,116 9.67% 21,118 31.54% 21,175 16.31% 23,017 32.09% 23,203 28.32% 23,014 17.86% 22,936 14.02% 22,896 26.66% 23,322 30.20% 23,714 24.60% 23,329 2.13% 16,808 13.83% 21,083 7.26% 9,475 29.29% 23,850 45.55% 22,628 21.21% 23,141 20.46% 22,709 22.87% 23,228 31.62% 23,992 9.19% 23,220 14.08% 16,077 34.25% 21,333 5.76% 22,803 21.31% 21,926 80.34% 22,951 15.15% 23,453 25.00% 20,868 7.10% 11,377 8.99% 23,434 6.81% 50,440 22.67% 22,155 78.00% 9,090 15.38% 23,201 14.04% 23,451 20.69% 23,179 47.63% 25,646 66.69% 33,522 84.99% 104,359 51.86% 28,267 17.44% 23,047 17.61% 22,680 31.22% 23.350 0.00% 45.02% 21,050 21.79% 25,051 5.32% 52.36% 29.24% 69.01% 49.01% 3.59% 73.41% 16.20% 28.38% 62.87% 27.09% 48.26% 11.20% 21.98% 24.43% 14.48% 29.84% 17.60% 23,283 20,984 22,892 9,634 22,699 21,079 22,233 23,444 22,179 20,192 23,327 23,760 20,391 109 APPENDIX D Table 5 Statistics of Variables Used in Regression Models; MORTNOM, MORTREAL, PGAS, CPI_NG, CPI_ALL, and FRMINC Variable FRMINC MORTNOM MORTREAL PGAS CPI NG CPI ALL 1960 6,687 5.492 3.332 30.9 30.4 29.6 1970 8,976 7.486 2.886 33.8 40.3 38.8 1980 11,343 10.302 1.022 81.16 81.9 82.4 1990 10,997 10.194 6.074 98.74 134.7 130.7 BIBLIOGRAPHY 110 BIBLIOGRAPHY American Farmland Trust, Michigan’s Land and Soil Resources: A Status Report. Michigan Farmland Preservation Project, 1985. Americana Foundation, New Directions Toward Integrated Land Use Planning. Americana Foundation, 1993. "Analysis of Bill No. H.B. 4151," Michigan Department of Treasury, 1977. Andrew, Neil J. "Use Value Assessment and Land Use Policy," Options in Land Use Policy for Michigan. Department of Resource Development, Michigan State University, 1974, pp. 29-41. American Petroleum Institute, "Average United States Retail Gasoline Prices for Selected Cities," Section VI, Table 12b, 1993. Bancroft, Robert L. et al., "Attitudes Toward Preserving Agricultural Land in Vermont," Miscellaneous Publications 93, Burlington: University of Vermont, Agricultural Experiment Station, 1977. Barlowe, Raleigh "Changing Rights in Land: The Dynamics of Property," Protection of Essential Lands. Department of Resource Development, Michigan State University, 1973, pp. 1-10. ________ . Land Resource Economics: The Economics of Real Estate. Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1978. Board of Governors of the Federal Reserve System, Construction and Mortgage Credit Balance Sheet of Agriculture. 1960. August 1960. Briggs, Ronald, "The Impact of the Interstate Highway System on Nonmetropolitan Development, 1950-75," Beyond the Urban Fringe: Land Use Issues of Nonmetropolitan America. University of Minnesota Press, 1983, pp. 83-105. Bureau of Local Government Services, Michigan Department of Treasury, 1967?. Brown, David L. & Calvin L. Beale, "Socioeconomic Context of Nonmetropolitan Land Use Change," Beyond the Urban Fringe: Land Use Issues of Nonmetropolitan America. University of Minnesota Press, 1983, pp. 61-80. Cherry, Jason, "Subdivision Regulation," Planning & Zoning News. January 1986, pp. 2-3. Ill Clawson, Marion, Suburban Land Conversion in the United States: An Economic and Governmental Process. The Johns Hopkins Press, Baltimore, 1971. Coughlin, Robert E., David Berry & Thomas Plaut, "Differential Assessment of Real Property as an Incentive to Open Space Preservation and Farmland Retention," National Tax Journal. Vol. 40 No. 2, pp. 165-179. Cunningham, Roger A. "Public Control of Land Subdivision in Michigan: Description and Critique," Michigan Law Review. Vol. 66, No. 1, November 1967. Daniels, Thomas L., Mark B. Lapping and John W. Keller, "Rural planning in the United States: fragmentation, conflict and slow progress," Rural Land-Use Planning in Developed Nations. Unwin Hyman, pp. 152-176. Dunford, R. W ., "Farmland Tax Relief Alternatives: Use Value Assessment vs. CircuitBreaker Rebates," College of Agriculture Research Center, Washington State University, 1979. Environmental Protection Agency, "The Cost of Sprawl: Detailed Cost Analysis." the Council on Environmental Quality; the Office of Policy Development and Research, Department of Housing and Urban Development; the Office of Planning and Management, April 1974. Espel, Thomas Kent, A Conversation Between Buyers and Sellers of Land or the Theoretical Basis for Estimating Land Values: A Market Equilibrium Approach. Master of Science thesis, Michigan State University, 1981. Gujarati, Damodar, Basic Econometrics. McGraw-Hill, Inc., New York, 1978. Hagman, Donald G. and Dean J. Misczynski, Editors, Windfalls for Wipeouts: Land Value Capture and Compensation. American Society of Planning Officials, 1978. Harvey, Lynn R. "Balancing Benefits in Rural Life," Michigan Planner. Winter 1992, pp. 3-5. Hatry, Harry P ., Richard E. Wayne and Donald M. Fisk, Practical Program Evaluation For State and Local Governments. The Urban Institute, 1981. Healy, Robert G. and James L. Short, The Market for Rural Land: Trends. Issues. Policies. The Conservation Foundation, 1981. ________ . "Changing Markets for Rural Lands: Patterns and Issues," Beyond the Urban Fringe: Land Use Issues of Nonmetropolitan America. University of Minnesota Press, 1983, pp. 109-133. Heimlich, Ralph E. Marlow Vesterby, & Kenneth S. Krupa, Urbanizing Farmland: Dynamics of Land Use Change in Fast-Growth Counties. Economic Research Service, U.S. Department of Agriculture, Agriculture Information, Bulletin Number 629, August 1991. Kamenta, J., Elements of Econometrics. Macmillan Co., N.Y. 1971 Kasowski, Kevin, "The Cost o f Sprawl, Revisited," Planning & Zoning News. January 1993, p. 8. Libby, Lawrence W ., "Essential Lands and The Problem of Choice," Protection of Essential Lands. Department of Resource Development, Michigan State University, 1973, pp. 57-66. Lucier, Gary, Agnes Chesley and Mary Aheam, Farm Income Data: A Historical Perspective. U.S. Department of Agriculture, Economic Research Service, Statistical Bulletin Number 740, May 1986. Marans, Robert W., Don A Dillman and Janet Keller, Perceptions of Life Quality in Rural America: An Analysis of Survey Data from Four Studies. Survey Research Center, Institute for Social Research, The University of Michigan, 1980. Michigan Agricultural Statistics Service, "Michigan Agricultural Statistics, 1960," 1961. ________ . "Michigan Agricultural Statistics, 1970," 1971. ________ . "Michigan Agricultural Statistics, 1980," 1981. ________ . "Michigan Agricultural Statistics, 1990," 1991. Michigan Department of Natural Resources, "Working Paper Aimed at Four Specific Aspects of the Subdivision Control Act," Office of Land Use, 1974. Michigan Department of Transportation, "Official Transportation Map, Michigan," 1960. ________ . "Official Transportation Map, Michigan," 1970 ________ . "Official Transportation Map, Michigan," 1980. ________ . "Official Transportation Map, Michigan," 1990. "Michigan Planning & Zoning Survey," Office of Intergovernmental Relations, State Department of Management and Budget, 1979. 113 Michigan Relative Risk Analysis Project, Michigan’s Environment and Relative Risk. Michigan Department of Natural Resources, 1992. Michigan Resource Information System(MIRIS), Land and Water Management Division, Department of Natural Resources, 1978 Data. Michigan Statutes, Act. No. 288, P. A. of 1967, Subdivision Control Act of 1967. Misseldine, Carol K. "Planning for Prime and Unique Farmlands," Michigan Planner. Winter 1992, pp. 8-9, 12. Misseldine, Carol K. and Mark Wyckoff, Planning and Zoning for Farmland Protection: A Community Based Approach. American Farmland Trust, January 1987. Moyer, D.David and Gene Wunderlich, Editors, Transfer of Land Rights. The Farm Foundation, 1985. Nino, Ronald F., A Comparative Analysis of Michigan’s Subdivision Control Act and Critique. Master of Science thesis, Michigan State University, 1970. Raup, Philip M ., "Policies for the Protection of Prime Agricultural Lands," Options in Land Use Policy for Michigan. Department of Resource Development, Michigan State University, 1974, pp. 13-22. Rockford Map Publishers Inc., Land Atlas and Plat Book Map. Rockford, Illinois, 196090. Snedecor, George W. and William G. Cochran, Statistical Methods. Iowa State University Press, Ames, Iowa, 1980. Thomas, Holly L., "The Economic Benefits of Land Conservation," Planning & Zoning News. January 1993, pp. 12-15. Tinberg, Cynthia Penner, An Economic Analysis of Agricultural Land Use Policies for Southern Michigan. Doctor of Philosophy dissertation, Michigan State University, 1978. U.S. Bureau of the Census, U.S. Census of Population: 1960. Vol. I, Characteristics of the Population, Part 24, Michigan, U.S. Government Printing Office, Washington, D .C ., 1963. ________ . Statistical Abstract of the United States: 1970. Washington, DC, 1992. 114 ________ . U.S. Census of Population: 1970. Vol. I, Characteristics of the Population, Part 24, Michigan, U.S. Government Printing Office, Washington, D .C ., 1973. ________ . County and Citv Data Book. 1983. U.S. Government Printing Office: 1983, Michigan, U.S. Government Printing Office, Washington, D .C ., 1983. ________ . Statistical Abstract of the United States: 1984. Washington, DC, 1984. ________ . Statistical Abstract of the United States: 1992 (112th edition), Washington, DC, 1992. ________ . U.S. Census of Population: 1990. Vol. I, Characteristics of the Population, Part 24, Michigan, U.S. Government Printing Office, Washington, D .C ., 1993. U.S. Department of Agriculture, Who Owns the Land?. Agriculture and Rural Economy Division, Economic Research Service, 1979. ________ . Timber Resource of Michigan’s Western Upper Peninsula. 1980. North Central Forest Experiment Station, Forest Service, NC-60. ________ . Timber Resource of Michigan’s Northern Lower Peninsula.1980.North Central Forest Experiment Station, Forest Service, NC-62. ________ . Timber Resource of Michigan’s Eastern Upper Peninsula.1980.North Central Forest Experiment Station, Forest Service, NC-64. ________ . Timber Resource o f Michigan’s Southern Lower Peninsula.1980.North Central Forest Experiment Station, Forest Service, NC-66. ________ . Economic Indicators of the Farm Sector: State Financial Summary. 1990, Agriculture and Rural Economy Division, Economic Research Service, December 1991. U.S. Department of Commerce, Business Statistics 1963-1991. 27th Edition, June 1992. Vlasin, Raymond D ., "Protection and Preservation Problems and Challenges," Protection o f Essential Lands. Department of Resource Development, Michigan State University, 1973, pp. 135-145. Warbach, John D ., "Lucas Remanded: May Be A Taking," Planning & Zoning News. July 1992, pp. 6-10. Wyckoff, Mark A., "Summary of H.B. 5152," Planning & Zoning News. January 1986, pp. 4-7. 115 "The Fiscal Impacts of Growth," Planning & Zoning News. January 1993,