IN FO R M A TIO N TO USERS This was produced from a copy of a document sent to us for microfilming. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the material submitted. The following explanation o f techniques is provided to help you understand markings or notations which may appear on this reproduction. 1.Th e sign or ‘‘target’’ for pages apparently lacking from the document photographed is "Missing Page{s)’\ If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure you of complete continuity. 2. When an image on the film is obliterated with a round black mark it is an indication that the film inspector noticed either blurred copy because of movement during exposure, or duplicate copy. Unless we meant to delete copyrighted materials that should not have been filmed, you will find a good image of the page in the adjacent frame. If copyrighted materials were deleted you will find a target note listing the pages in the adjacent frame. 3. When a map, drawing or chart, etc., is part of the material being photo­ graphed the photographer has followed a definite method in "sectioning” the material. It is customary to begin filming at the upper left hand corner of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete. 4. For any illustrations that cannot be reproduced satisfactorily by xerography, photographic prints can be purchased at additional cost and tipped into your xerographic copy. Requests can be made to our Dissertations Customer Services Department. 5. Some pages in any document may have indistinct print. In all cases we have filmed the best available copy. University. Microfilms International 300 N ZE E B RD , A N N A R B O R , Ml 48106 8212415 Kufiior, Osei Kwaku (a/lt/a Francis Oliver Arthur) A N ANALYSIS O F PRIVATE L A N D FRAGMENTATION BY L A N D HOLDINGS O F LESS T H A N 11 ACRES IN MICHIGAN PH.D. 1981 Michigan State University University Microfilms International 300 N. Zeeb Road, Ann Arbor, M I 48106 AN ANALYSIS OF PRIVATE LAND FRAGMENTATION BY LAND HOLDINGS OF LESS THAN II ACRES IN MICHIGAN By Ose1 Kwaku Kufuor [Also known as Francis Oliver Arthur] A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Resource Development 1981 ABSTRACT AN ANALYSIS OF PRIVATE LAND FRAGMENTATION BY LAND HOLDINGS OF LESS THAN 11 ACRES IN MICHIGAN By Francis Oliver Arthur Subdivision of land into parcels of less than 11 acres (parcellation) is on the increase in Michigan. Counts of the actual acreages included in non-platted subdivisions in 30 counties of the state between 1963 and 1977 indicate that the total of these subdivisions increased from around 673,000 to 1,200,000 acres during the period. Differences do exist in the amount of acreages and the rate of increase in these subdivisions among and within counties, districts and regions. High levels of par c el ­ lation are significantly related to socio-economic activities but the high rate of increase is also associated with environmental conditions and institutional factors such as the Subdivision Control Act (SCA) of 1967. This study examined the spatial distribution and the time trends of non-platted and approved subdivisions in Michigan in an attempt to provide empirical evidence about small-tract land parcellation. The study also projected parcellation trends to the year 2000. A method was developed Francis Oliver Arthur to classify counties and regions into highly, moderately and least affected parcellation areas. Two regions were primarily affected by the process of small-lot parcellation; these are rural lands, fringing urban centers and countrysides with clean environmental conditions. Two types of demand were, therefore, noted — job oriented demand and recreational demand for homesites. Correlation and multiple regression analyses revealed that the most important factors which contributed to the spatial and trend variations in the parcellation process were: (a) personal incomes, (b) population concentrations and movements and associated demand for homesites, (c) de­ mand for recreational and environmental amenities and (d) certain public land use policies. Serial Correlation analysis, based on quasi-experimental design, indicated that the Subdivision Control Act of 1967 contributed sig­ nificantly to the increased parcellation of land into 10 and 10+ acres, especially after 1970. Its impact on 10- acre parcels and on approved subdivisions was not signifi­ cant. A simple linear extrapolation projection to the year 2 0 0 0 would result in a total and acreages of about 200,000 holdings and 1 ,1 0 0 , 0 0 0 acres of these non-platted parcels in another 20 years. Parcellation data were obtained by counting all small tracts of non-platted subdivisions less than 11 acres for Francis Oliver Arthur 465 townships of 30 counties, all regions of the state. systematically selected from County Atlas and Plat Books provided the sample frame and figures covering a period of 15 years with 1963, 1970 and 1977 as the study periods. The linear extrapolation projection technique was based on ceteris paribus assumption. The importance of clearly identifying the parcella­ tion process is emphasized by the study. A policy such as the Subdivision Control Minimum acreage provision would not necessarily discourage relatively large-lot parcella­ tion and the creation of idle lots. It is, therefore, recommended that the 1 0 acre m in imum lot size provision in the Subdivision Control Act, either be removed or the minim um be increased to a higher acreage level such as 40-60 acres. ACKNOWLEDGEMENTS The study was funded by the Michigan State University Agricultural Experimental Station through the Research Directorship of Dr. Raleigh Barlowe of the Department of Resource Development. I w ou ld like to express my deepest gratitude to Dr. Barlowe for obtaining support for the study and for his untiring effort to direct the whole dissertation even after he retired from active professorial duties. Sincere thanks to all my other Committee members, Dr. M.H. S t e i n m u e l l e r , Dr. D.E. Chappelle, Dr. W.J. Kimball, and Dr. Anthony Koo who in one way or the other — economic, psychological and social, academic, contributed to my study at MSU and to this dissertation. I would also like to make a special mention of Mrs. Joann Silsby who prepared this document and Mr. Paul Schneider for his excellent graphics and advice on d i s ­ sertation formats. Special thanks are also extended to my friends and colleagues, Mr. Essama Nssah and Mr. Kwabena Abeka Whyte of University of Michigan and Michigan State University respectively, for their contribution to the statistical and computer aspects of this thesis. Mr. Robert Burke also offered invaluable constructive criticisms and heated discussions on various subjects — thanks Bobby. Acknowledgement is also due to Mr. Richard Lomax of Michigan Treasurer's Department, Plat Section, for allowing me access to Plat files, data, and documents. It is not possible to mention all the names of County Exten­ sion Directors who responded to the feedback survey let­ ters; my sincere gratitude to all of them for their in­ valuable information. Acknowledgement is due to Rockford Map Publishers, I n c . , whose Land Atlas and Plat Books were used as a sample frame. Hardly can I fully thank individuals in the Resource Development Department and the Department as a body for all the roles they played in making my academic career at MSU a success. My three year study in the Department will always conjure happy memories. Finally, last but not least, special gratitude to my patient and devoted wife, Gyamfua-Amonu, who has spent almost 12 years out of our 17 years of married life waiting on an almost professional student and for taking upon her­ self the responsibilities of a father towards his children in addition to her own maternal duties. moral and psychological support Her patience, in the face of extreme financial hardships, sometimes baffled me, her wisdom confounding. Thanks, darling. This acknowledgement would be incomplete if I do not mention my little boy who was born during my graduate ii study days and hence had to bear the blunt of paternal neglect for a while. Ghana, To Kwame and his brothers in I say, cheer up kids, dad will now make up all losses to you! iii TABLE OF CONTENTS CHAPTER PAGE LIST OF T A B L E S .................................... viii LIST OF F I G U R E S................................... xi ONE........... INTRODUCTION............................. The P r o b l e m .................................... Purpose of S t u d y . .............................. Definition of Terms ........................... Limitations of S t u d y ........................... Organization of D i ss er ta ti on .................. TWO RESEARCH METHODS AND ANALYTICAL FRAMEWORK . . . Analytical Approach ........................... Sources and Nature of D a t a .................... Primary Data: Non-Approved Parcels . . . . Secondary Data: Approved Subdivisions. . . Study V a r i a b l e s ................................ Data Collection P r o c ed ur es .................... Sample Districts and Regions .............. Broad Comparative R e g i o n s ............ 24 Levels of A n a l y s i s .............................. THREE GENERAL BACKGROUND TO S T U D Y .................... Introduction.................................... Descriptive Framework ......................... Social Determinants ........................... Michigan Population and Demand for Ho me s i t e s ............................ 28 Trends in Michigan Population (1940-1980). Housing U n i t s .......................... 41 Conclusions on Social Determinants . . . . Economic Determinants ......................... Economic Acti vi ti es ................... 45 Michigan Agriculture .................... Land V a l u e s .......................... 52 54 I ncomes.............................. E m p l o y m e n t .......................... 55 iv 1 3 5 6 7 8 10 10 14 14 15 17 19 20 24 26 26 27 28 30 45 45 47 CHAPTER PAGE Biophysical Determinants ....................... Ecological F a c t o r s ........................... Forests .................................... Recreational Re so u r c e s .................... Institutional Determinants .................... Government R egulations....................... The Subdivision Control Act .............. Conclusions .................................. 57 57 58 59 61 61 63 72 FOUR HYPOTHESES AND STATISTICAL M O D E L S ................ 74 Working Hypotheses . . . . . .................. Assumptions Underlying Hypo th es es .............. Assumptions Underlying Statistical Tests . . . Hypotheses Testing ............................. Partial Correlation Matrix .................... Multiple Regression Analysis .................. The General Model ........................... Quasi-Experimental Design Model ........... Method of Estimating State Acreages from S a m p l e ......................................... Weighted Parcellation Density (WPD)........... Derivation of WPD S c o r e s .................... FIVE RESEARCH F I N D I N G S .................................. Non-Platted P a r c e l s ............................. Holding of Non-Platted Land Parcels . . . . Acreages of Non-Platted Land Parcels. . . . Trends in Amount of Pa rc ellation ........... Trends in Parcellation Categories ......... Summary on Time Trend Analysis of Parcel­ lation based on 30 Counties in Michigan . Spatial Distribution of Land Parcellation in Michigan .................................. Statistical Analysis of Variations in P ar cellation.................................. Correlation Results ......................... Multiple Regression Results ................ Analysis of AAPAS ......................... Analysis of ASPAS and L A G P A S .............. Analysis of AMPAS and ALPAS .............. Spatial Variations in Parcellation Trends . Variations Among D i s t r i c t s ................ Variations Among Study R e g i o n s ........... Variations Among the Two Broad Regions (Northern and Southern Michigan). . . . Approved Subdivisions. . . . . ................ Trends in Approved Subdivisions ........... Acreages Under Subdivisions .............. Ecological F a c t o r s ........................... v 74 77 79 84 86 86 87 92 93 97 98 103 103 103 105 106 111 115 119 137 139 144 144 148 154 157 157 159 161 169 169 173 173 CHAPTER PAGE Trends in Approved Subdivisions by District and R e g i o n ....................... C o n c l u s i o n ....................................... SIX SUMMARY, CONCLUSION AND RECOMMENDATIONS 175 177 . . . . 180 S u m m a r y ......................................... Spatial Distribution ....................... Time Trend . . . . . . .................... C o n c l u s i o n s .................................... Subdivision Regulation: Subdivision Control Act of 1 9 6 7 ................................ . . . . . Recommendations .................... Recommendations Stemming from the Study F i n d i n g s .................................. Recommendations for Future Research. . . . Final C o m m e n t s .................................. Planning for the F u t u r e .................... 180 182 183 184 186 190 190 192 194 194 APPENDICES Appendix 2-A Appendix 3-B 3-B-l 3-B-2 Appendix 4-C Appendix 4-D Appendix 5-E Appendix 5-F Sample of Land Atlas and Plat Book Map of Portage Township, Houghton County, Upper P e n i n s u l a ............ 196 Sample of Correspondences with County Extension Directors for Feedback In­ formation on Land Parcellation . . . . 197 Letter Written to County Extension Directors (CEDs) ....................... 197 Selected Responses of C E D s ....... 193 Frequency Histograms with Fitted Smooth Curves for Categories of Parcellation D a t a .................. Descriptions of Variables used in Regression and Correlation Models. 206 . . 207 Statistics on Non-Platted Parcels Table E-l to Table E - 8 ............ 209 Quinquannial Projection Estimates of Amount of Parcellation in Michigan by Size Unit C1 9 8 0 - 2 0 0 0 ) ......... 217 vi APPENDICES Appendix 5-G Appendix 5-H Appendix 5-1 Appendix 5-J PAGE Summary Data for - Parcellation Density Scores for 1963, 1970 and 1977; Pro­ jected Estimates for Year 2 0 0 0 218 Trends in District and Regional Parcel­ lation Acreages; Table H-l to Table H - 4 ................................ 222 Subdivision Trend Equations by D i s t r i c t .................................. 225 Subdivision Trend Graphs by Region. SELECTED BIBLIOGRAPHY . . .................................. vii 226 227 LIST OF TABLES TABLE 2-1 PAGE Location Regions and Districts for Selected Study Counties in M i c h i g a n .................... 22 3-1 Trends 31 3-2 Number of Occupied Housing Units in Michigan ( 1 9 4 0 - 1 9 7 9 ) .................... 42 Regional Distribution of Housing Units in Michigan (Occupied and Vacant) 1970 and 1 98 0.............................................. 44 Average Values of Farm Real Estate per Acre, Michigan and Selected States, 1973-1979 . . . 53 Pearsonian Coefficient of Skewness for Sample Acreage Distribution by Year and by Size U n i t ................................ 80 3-3 3-4 4-1 in Michigan Population ... 4-2 Values of ip 4-3 Estimated Mean Parcellation Amounts for the State of Michigan and the Sample Means for Periods 1963, 1970 and 1977 .............. 96 The $ Values for the Estimation of Paj*cellation Density Index for 1963, 1970 and 1977 . . . . 101 4-4 5-1 5-2 5-3 5-4 for 1963, (1940-1980) 1970 and 1 9 7 7 ........... Number of Holdings of Land Parcels by Size Unit and by Period of Study for the 30 Selected Counties ................................ 95 104 Amount of Parcellation by Size Unit and by Period of Study for the 30 Selected C ou nt i e s ......................................... 106 t-Test for Difference Between Means of Parcel­ lation Acreages for Time Periods 1963, 1970 and 1977 .................................. 107 Increases in Parcellation Acreages by Size Unit for 30 Selected C o u n t i e s .................. viii 110 TABLE 5-5 5-6 5-7 5-8 5-9 PAGE Percentage Shares of Parcellation Categories of Total Amount of Parcels per P e r i o d .............. H3 Percent Share Distribution of Total Increments in Parcellation Acreages Between "Small" and "Large" Parcels... ................................. 114 Estim at ed Total Nonapproved Parcels in Michigan by Period of Study, 1963, 1970, 1 9 7 7 .............. 117 Quinquannial Projection Estimates of Amount of P arcellation Under 11 Acres in Michigan ( . 19 80 -2 00 0) ..................................... W eighted Parcellation Density (WPD) Index for 1963, 1970 and 1977 and Projected for 2000. 118 . 121 5-10 Description of Ea ch of the Groups of Counties Categorized by W P D ............................. 123 5-11 Acres of Land Under Parcels Less than 11 Acres and Large Parcels (10-10.9) with Corr es po nd ­ ing Parcellation Density for Ea ch Group of Counties Based on 197 7 D a t a .................. 124 5-12 Density and Acreages by Group and County for Group X ( 1 9 7 7 ) ................................... 126 5-13 Density and Acreages for Group..2 ............... l2^ 5-14 Density and Acreages for Group 3 ............... 132 5-15 Correlation Coefficients Between Dependent and Independent Variables ......................... 138 5-16 Correlation Coefficients Between Pairs of Independent Variables ......................... 143 5-17 Multiple Regression Statistics for AAPAS for the T h ree Study Periods ....................... 145 5-18 Multiple Regression Statistics for ASPAS. . . . 149 5-19 Multiple Regression Statistics for LAGPAS . . . 150 5-20 Multiple Regression Statistics for AMPAS. . . . 155 5-21 Multiple Regression Statistics for ALPAS. . . . 156 5-22 Trends in Average Annual Rates of Change in LAGPAS Actual and Projected by District and R e g i o n ....................................... ix 158 PAGE TABLE 5-23 Multiple Regression Statistics for Parcellation Size Units for Southern M i c h i g a n .............. 164 5-24 Multiple Regression Statistics for Parcellation Size Units for Northern M i c h i g a n .............. 165 5-25 Approved Subdivisions for Sample Counties and for the State ( 1 9 6 9 - 1 9 7 9 ) .................... 170 x LIST OF FIGURES FIGURE PAGE 2-1 Analytical F ra me w o r k............................. 11 2-2 Subsample Structure of the Dependent Variable, Counted Parcels and Approved Subdivisions . . 18 Michigan: Regional and District Distribution of Selected C o u n t i e s ........................... 21 Analysis of the Effects of Minimum Lot Ceiling on P a r c e l l a t i o n ................................ 69 Trends in Nonapproved Parcels, 1963, 1970, 1977 and Projected to 2000 by Samp le ......... iq8 Trends in Nonapproved Parcels 1963-1977 and Projected to 2000 by Study R e g i o n ........... 160 Map Showing Broad Study Regions for Statistical A n a ly si s ......................................... 162 Trends in Approved Subdivision Plats with Fitted Curve, Sample and State, (1969-1979)......... 171 Trends in Subdivision Acreages and Lots, Sample, State, (1969— 1 9 7 9 ) .............................. 174 2-3 3-1 5-1 5-2 5-3 5-4 5-5 xi CHAPTER ONE INTRODUCTION Parcelling of private land holdings into smaller tracts has been a common practice with privately owned lands in Michigan and the nation for many years. Major contributing factors include urbanization and suburban­ ization of rural and semi-rural lands, the recent phen­ omenon of reversed migration which involves urban dwellers reaching out for homesites in a rural environment, seeking of recreational amenities, and other socio-economic fac­ tors such as escape from the social problems of most American cities. These determinants (and others) explain the demand side of the fragmentation process. On the supply side, land owners have offered more and more land as demand bras increased and real estate prices have gone up. The conditions in the real estate market have served as a "pull force" that has attracted land owners and encouraged them to parcel out their lands to individual developers or home builders. Other factors have served as a "push force" to rein­ force the pull factors. Some land owners have been b u r ­ dened by increasing property taxes associated with 1 2 increasing expected market values of their properties which in turn have resulted from increasing urban and suburban pressures on rural lands. Ripening costs on such urban fringed lands have soared and owners have adjusted to this added burden by selling all or parts of their lands pre­ maturely to developers and others. In sum, the process of land fragmentation and parcel­ lation must be examined in the light of the "push" and "pull" forces operating on both the supply and demand sides of the real estate market. Forces that exist on the demand side may also be categorized as "push" and "pull". Increasing general affluence has brought an increase of emphasis on the "quality" as compared with the "quantity" components of individual levels of living. Urban con­ ditions (socio-economic) associated with oversized cities or diseconomies of size and negative externalities of p o p ­ ulation, economic and social localization are the major "push" forces. "Pull" factors are the perceived environ­ mental or ecological amenities of the rural areas, lower land values, and in general, "better quality" of life associated with rural areas. A recent Michigan Public Opinion Survey, conducted at Michigan State University, indicated that about 41% felt that the government should spend more tax revenues on protecting "prime" or "important" lands from urban 3 developments.^- The report of Governor William Milliken's Community Development Cabinet indicates that agriculture in the state is threatened by the increasing amount of farmland lost to urban developments. The report calls for a number of measures to bolster agriculture in the state, including special tax incentives, zoning regulations, "right-to-farm" legislation and economic assistance to the 2 food and fiber industries. The Problem Urbanization and suburbanization have created a tre­ mendous demand for new residential building sites. The "urban exodus" of the last decade has added greatly to the these demands. Developers have been providing the lots, but in many cases the development process has not been orderly and has involved the creation of larger lots or parcels than appear either necessary or desirable. Typical lots in suburban communities call for onethird to one-half acre per house. General observations, ^Kimball, W. J. et. al., Report on Results of the Michigan Public Opinion Survey, Michigan Citizens Speak Out on Community Problems, Preferences and Government S p e n d i n g . Michigan State University, Agricultural Experiment Station (East Lansing, Development and Public Affairs, No. 378, July 1979). 2 Report of the Community Development Cabinet to Governor William G. M l l l i k e n . on "Agricultural Preservation Strategy for Michigan" by Tim Noworyla, Policy Analyst (Lansing, Michigan. November 1980). 4 however, indicate that large numbers of lots are larger than one acre and that many are 10 acres or slightly larger. lots. Two explanations are often given for the 10 acre Local zoning ordinances often use a minimum 10 acre size presumably as a means of discouraging the development of certain areas. The Subdivision Control Act of 1967 also affects the situation with its requirement that sub­ dividers go through the formal platting process when they divide land bolding into five or more tracts of less than 10 acres. General observations indicate that many subdividers have circumvented the intent of both the zoning ordinances and the Subdivision Control Act by creating and selling 10+ acre lots. The main results are: (1) an ignoring of the benefits that platting process should bring to both the community and the land purchaser; (2 ) a wasteful use of land when buyers could or would be content with smaller holdings; and (3) possible higher costs to local govern­ ments in providing services. Increasing numbers of the Michigan citizenry have been expressing concerns about these trends in land use, specifically to the extent that the process is affecting important agricultural and forest lands. tical evidence, Yet, no statis­ so far, exists about the level and extent of land parcellation in the state. 5 An ongoing study of parcels in excess of 10 acres in Shiawassee county by Uentius has indicated that most lot purchasers would prefer smaller lots if they were availa b l e .1 Consumers are forced by circumstances to buy lots in excess of 10 acres. These lots, being too large for standard single family housing units, are often underuti­ lized and, in general, part of the property is usually left idle. Furthermore, it is argued that the 10 acre limitation is contributing to premature parcellation and leap-frogging of residential developments, particularly in counties and townships where zoning regulations are not strictly enforced. One danger of the process is that it is taking large areas of prime agricultural land that should be retained in their present use. Purpose of Study The lack of empirical evidence and the scantiness of statistical data base on less than ll«=*icre parcels of nonplatted lots in the state motivated this study. jectives of this study, therefore, 1. The ob ­ are: To indicate the extent and trends of small, nonplatted lot parcellation in the state; ^Mentius, F. S., "Venice Township: A Study of Land Fra g­ mentation by 10.1 Acre Parcels, 1968-1978." Prelim­ inary Research Report, 1978. 6 2. To examine the interrelationships between land parcellation and certain ecological, social, eco­ nomic and institutional factors as they affect the spatial distribution and time path of the process; 3. To examine trends in approved subdivisions; 4. To relate small lot parcellation to needed changes in state land use regulations. and An attempt also is made to project trends in parcellation to the year 2000 A.D. under the ceteris paribus assump­ tion. Definitions The words "fragmentation", "parcellation" and "land partitioning" are used interchangeably in this study. They all connote the dividing or subdividing of tracts of land into smaller parcels— rural holdings of less than 40, 80, or 160 acre survey units. The term "small lots" and "large lots" are used in this dissertation to refer to nonplatted parcels less than 10 acres and 10-10.9 acre parcels respectively. They are relative and not meant to compare with any other parcels outside the cut-off point (larger lots, 11 acres and above, are also created in the state). The term "level" of par­ cellation means amount of parcellation as measured by 7 number of holdings; "extent” of parcellation connotes amount of parcellation as measured by acreage and "degree" of parcellation refers to amount of parcellation weighted by population and land area— it is referred to as Weighted Parcellation Density (W.P.D.). The term "subdivision" in this dissertation refers exclusively to parcellation of land into tracts as provided in the Subdivision Control Act— also referred to as approved parcellation as opposed to non-approved parcellation. Non-platted subdivisions are not usually submitted to local, county or state authorities for official approval, and should be distinguished from subdivisions that are for­ mally platted. The parcellation process involves indi­ vidual plots or parcels sold or leased. Over periods of several years, the process can involve the "chipping" away of numerous smaller tracts from larger units of land. A "parcel" of land as treated here, therefore, means any holding less than 11 acres. Limitations of the Study The use of land atlas and plat books as sample frame confines most analysis and inferences to counted parcels. No attempt was made to interview or reach land owners and developers through questionnaires to determine their p e r ­ ceptions about the process of land parcellation. Thus, 8 motives underlying both the demand and supply sides of the real estate market could not be ascertained adequately to reflect on the data that were obtained. Information about the supply of and demand for small parcels was obtained from secondary sources. This limits the scope of the study to the primary data that we re collected. Organization of the Dissertation This dissertation deals with land parcellation in Michigan; its current level and spatial distribution; trends over time, past to present, and to future. its The di s ­ sertation also tries to provide explanations for the p a r ­ cellation levels and distributions over time and space and finally attempts at relating trends to needed policies in land use. Chapters are organized to follow this logical sequence. Research methods and analytical techniques are discussed in Chapter Two, where the dependent variable is explicitly described. Chapter Three provides a general background to the study and factors influencing the p r o ­ cess of parcellation are discussed. Chapter Four states the various hypotheses to be tested and formulates sta ti s­ tical models w hi ch establish relationships between the de­ pendent and independent variables. Research findings are presented in Chapter Five and results of statistical tests are used to explain findings. Chapter Six provides a 9 summary and recommendation for both future land use policies and land use research. CHAPTER TWO RESEARCH METHODS AND ANALYTICAL FRAMEWORK Analytical Approach The process of land parcellation is theoretically conceptualized as a four-dimensional problem, quantity involving (number of parcels or acres of land), value (price of parcels), time (dynamic). spatial distribution (counties) and The analytical approach, therefore, cludes static economic analysis, in­ (value and quantity), comparative static analysis (value, quantity and discrete time periods) and dynamic locational analysis (quantity distribution over space and over time). Figure 2-1 illus­ trates the analytical approaches to the problem of land par ce ll at i on . The problem can conceptually be likened to a cuboid moving through time and space. In Figure 2-1 the four important conceptual aspects of parcellation are shown. Q refers to the quantity of parcels existing at any point in time and place, V refers to the value of a parcel of land (unit price), S refers to the location of the parcels or the spatial distribution of parcels and t is the time period of analysis (in the dissertation, 1977 are the discrete periods). 10 1963, 1970 and Q b (1977) (1963) S SpotiolLocation distribution FIGURE 2-1 ANALYTICAL FRAMEWORK 12 Classical economists have treated land as an economic commodity or factor of production within the V-Q space or quantity-price dimension. Ricardo, Adam Smith, and even Marshall, all dealt with land from a two-dimensional, static perspective. The price of land indexed by land productivity was usually related to the quantity of land available at any point in time, given quality and all other factors. Most land economists start with this two-dimensional approach, but also recognize that land use decisions in­ volve spatial and time considerations .1 This dissertation follows this pattern by relating land values to the quan­ tity of parcels created at three given points in time at various locations in the state. To provide a clear perspective of the land parcel­ lation phenomenon, it is important to determine what has happened over an extended time period. This study, there­ fore, goes beyond the three-dimensional approach by adding the time variable. The discrete time periods provide a comparative static analysis of land parcellation in the state. These equilibrium points (1963, 1970 and 1977) are compared with the initial condition defined for 1963 as the zero base year. The equilibrium path is then projected to the year 2 0 0 0 . M a r l o w e , R. Land Resources Economics: The Economics of Real E s t a t e , 3rd Edition, Prentice Hall, N.J. (1978). 13 Projection of the equilibrium time path usually in­ volves assumptions about initial conditions. In this dis­ sertation the assumption was that conditions will behave as they have been doing during the 15-year period of study. The weakness of such an assumption is clearly recognized, but this study does not aim at exactly predicting or fore­ casting the future; the projection of the equilibrium time path of land parcellation is not a best judgment estimate of what will actually exist at the turn of the century. Rather, it is most useful in providing a boundary notion of where the present trends are likely to lead in the ab­ sence of significant changes in the underlying forces. It is recognized, however, that changes not yet anticipated will occur eventually. Despite its recognized limitations, time analysis may prove useful for long-term planning by land developers, agribusiness and governmental institu­ tions. At any specific point in time, value, quantity and space are the main framework of analysis and the question of how the equilibrium conditions in, say, 1963 got to 1970 or 1977 are not addressed. Nevertheless, attempt is made to explain the whys of change by examining the factors that are likely to influence the time path of quantity, location and value of land parcels independently and in their combinations as well as their interactions over time. 14 Sources and Nature of Data Determination of the exact number of parcels for any given moment on a statewide basis is, of course, sibility. an impos­ N e w parcels are constantly being created; former parcels are either resubdivided into smaller tracts or re­ consolidated into bigger lots or both; existing small lots are being combined w i t h existing bigger holdings; existing big land holdings continue to be "chipped" away, parcel by parcel and old approved subdivisions w h ic h never reached a development stage are being sold outright for new uses; these and other forms of combinations and recombinations, partitioning and repartitioning go on constantly and c o n­ tinuously . Primary Data: Non-Approved Parcels (1963. 1970, 1977) The counted parcels which were created through s u b ­ divisions exempted from the provisions of the Subdivision Control Act of the state reflect the number of such p a r ­ cels (parcels less than 11 acres) e x isting in the selected counties during the three specific time periods of study. Any inference drawn from the sample data to provide state estimates for the three periods reflect the actual number of parcels existing in the state assuming that the sample frame reports accurate observations and the estimation technique is valid. continuous; However, they are periodic. data are not serially 15 The difference between the amounts or numbers of parcellation for any two periods reflect the net additions to the total parcellation figure of the preceding period. The net additions or incremental figures do not show the process of partitioning that resulted in the totals. They provide no indication about the methods of partitioning. The incremental figures are, therefore, representative of all kinds of partitioning processes which result in the creation of parcels less than 11 acres, during some inter­ val periods. An increase or decrease in the total number of all categories may be the result of several processes operat­ ing individually or in their various combinations. Increase may be due to the creation of new parcels entirely from large holdings, excluded from the definition of parcella­ tion in this study or may be due to the repartitioning of the relatively larger parcels (10-10.9) (parcels less than 10 acres) or both. into smaller units A decrease in parcel ■ numbers may be due to a recombination of smaller parcels into bigger ones, which may or may not be Included in the d e f in it io n. Secondary Data: Approved Subdivisions (1969-1979) Approved subdivision data were obtained from a sec­ ondary source and data are continuous annually from 1969 to 1979 for 30 counties, and from 1970 to 1978 for the 16 whole state. Therefore, data were truncated at both ends and hence do not reflect the actual total number of sub­ divisions existing in the state at the terminal year. Annual figures represent solely the newly created subdivi­ sions; they do not include subdivisions awaiting approval, or those in the process of being created or those held at local levels. Each year's subdivision figure represents only the number of new plats approved that year, usually for the fiscal year ending June 30 (up to 1975) and Sep­ tember 30 (1976 onwards). Incremented figures for the period between 1975-1976, therefore, represent a total for 15 months fiscal period.^A cumulative total number of approved subdivisions from 1969 to 1979 reflects total new subdivisions created during the 11 years only. Subdivision data are, therefore, used mainly for time series analysis. Its use for spatial trend analysis is limited since the past subdivision dis­ tribution is unknown. Counted parcels are used for both spatial and trend analysis since the number counted at any period in time represent what actually existed at the time. The year 1963 is considered the base and all comparisons are made with reference to that year. *See State of Michigan, Annual Report of the State T r e a ­ surer . October 1, 1977 to September 30, 1978. 17 Study Variable The main dependent variable is land parcellation, measured by number of holdings of land less than 11 acres and approved subdivision parcels (Figure 2-2). Number of parcels (non-platted) were counted from Land Atlas and Plat Books 1 on township basis aggregated into county, 2 and the resulting figures were district and regional totals (2-2). Data are both cross-sectional (30 counties) and longitudi­ nal (1963, 1970, 1977).3 Figures obtained for the counted parcels in holdings were converted into acreages (Chapter Four, conversion procedure) and then grouped into three size units, viz: (1) parcels less than 10 acres (10- acre parcels), (2) pa r ­ cels of 10 acres (10 acre parcels) and (3) all parcels in excess of 10, but less than 11 acres (10+ acre parcels). Eleven acres is the cut-off point (all larger parcels Land Atlas and Plat Books are published by Rockford Map Publishers, Inc. , annually and distributed by various counties. Acknowledgement is due to R . M . P . , Inc., 4525 Forest View A v e . , P.O. Box 6126, Rockford, IL 61125. 2 A sample of Land Atlas and Plat Book map is provided in Appendix 2-A. 3 Other possible sources of data include county deed rec­ ords, tax records and files on building permits. These were not used because of the time and expense that would have been involved and also because it was felt that the county plat books provided a reliable source of information on land parcels. PARCELLATION DEPT. VARIABLE (P) APPROVED SUBOIV. PARCELS (Snail Lots) 1969-79 COUNTED PARCELS ALL 11- ACRE PARCELS (1963-77) 10 Acre Parcels HOLDINGS (Parcel 1. Level) 10 Acre Parcels 10 + Acre Parcels ACREAGE (Parcell. Extent) WEIGHTED DENSITY (Degree of Parc.) Av. ACREAGE PER HOLDING PLATS ACREAGES ACRES PER PLAT ACRES PER LOT FIGURE 2-2 SUUSAI1PLE STRUCTURE OF OEPENOENT VARIABLES COUNTED PARCELS AND APPROVED SUBDIVISION PARCELS LOTS PER 19 are excl ud ed ) .1 Total acreage parcelled is the principal unit measure for analysis. Acreage is a continuous va r i­ able measure w h ic h permits parametric statistical treato ment of the data. Another (dimension-less) measure based on acreage figures is parcellation density (WPD). It is parcellation acreage weighted by population and land area of counties. 3 The weighted parcellation density scores are meant to m e a ­ sure degree of parcellation and to provide the basis of comparing level and extent of parcellation among counties* districts and regions. Data Collection Procedure The unit of data collection is the township. counties The 30 (36 percent of total counties in the state) The cut-off point is arbitrary. The writer recognizes that parcellation involving 11-40 acre parcels is as important as those less than 11 acres. However, time and financial limitations required that the dependent variable be defined narrowly. This opens up a po s ­ sibility for further research into the much bigger parcels (11-40 acres). 2 For full discussion of statistical measurement, reader may consult the following authors: (a) Ya-Lun-Chou, Stattistical Analysis w i t h Business and Economic Ap p l i c a ­ tion . 1969, p. 477; (b) Borg, W . R . , and Gall, M . D . , Educational Research: An I n t r o d u c t i o n . 2nd Edition, 1971, p p . 312-315; (c) Hucks, S.W., Cormier, W . H . , and Bounds, W.G. Jr., Reading Statistics and R e s e a r c h . 1974, pp. 197-198. 3 Parcellation density is fully discussed in Chapter Four, pages 20 systematically selected (see list in Table 2-1, pp. 22-23) contain about 465 townships or 37 percent of the total townships in the state. Counties are grouped into dis­ tricts and districts into regions (Figures 2-3) for compara­ tive study. The three time periods of 1963, 1970 and 1977 provide base year data for trend analysis and projections. Projection takes into consideration current population den­ sities and growth rates in the state. Sample Districts and Regions The state was first divided into eight blocks or study districts, (Figure 2-3; Table 2-1); each district contained approximately 10 to 11 counties. About 3 to 4 counties were selected from each district and parcels were counted for all the townships that constituted the selected 30 counties. level. No further sampling was done at the township The eight districts provided the first stage stratification of the 83 counties as well as a spatial frame for comparative analysis. Figure 2-3 also shows the four study regions. were combined into regions. Districts The regions were designated as R-I (East Southern Lower Peninsula), R-II (Vest Southern Lower Peninsula), R-III (North Lower Peninsula) and R-IV (The Upper Peninsula). Table 2-1 shows the list of regions and their abbreviated form with the corresponding counties. 21 Region E C Region H I counties Reg ion H jWwum*! mtmi J Region I ;<{.a FIGURE 2-3 MICHIGAN: REGIONAL AND DISTRICT DISTRIBUTION OF SELECTED COUNTIES 22 TABLE 2-1 Location Regions and Districts For Selected Michigan Counties Selected County Name Study D i s tr ic t8, Study . Region Hillsdale Livingston Macomb Monroe Bay Clinton Huron Lapeer Allegan Berrien Calhoun S t . Joseph Montcalm Newaygo O ttawa Alpena Cheboygan Crawford Iosco Antrim Clare Grand Traverse Manistee Delta Mackinac Schoolcraft Gogebic Houghton Menominee Iron S.E.S.L.P. SaEiS>LiP> S.E.S.L.P. S.E.S.L.P. C.E.S.L.P. C.E.S.L.P. C.E.S.L.P. C.E.S.L.P. S.W.S.L.P. S.V.S.L.P. S.W.S.L.P. S.W.S.L.P. C.W.S.L.P. C.W.S.L.P. C.W.S.L.P. E.N.L.P. E.N.L.P. E.N.L.P. E.N.L.P. W.N.L.P. W.N.L.P. W.N.L.P. W.N.L.P. E.U.P. E.U.P. E.U.P. W.U.P. W.U.P. W.U.P. W.U.P. E.S.L.P. E.S.L.P. E.S.L.P. E.S.L.P. E.S.L.P. E.S.L.P. E.S.L.P. E.S.L.P. W.S.L.P. W.S.L.P. W.S.L.P. W.S.L.P. W.S.L.P. W.S.L.P. W.S.L.P. N.L.P. N.L.P. N.L.P. N.L.P. N.L.P. N.L.P. N.L.P. N.L.P. U.P. U.P. U.P. U.P. U.P. U.P. U.P. 23 TABLE 2-1 (Continued) d i s t r i c t s are designated as: District District District District District District District District 1: 2: 3: 4: 5: 6: 7: 8: South East Southern Lower Peninsula (SESLP) South Vest Southern Lower Peninsula (SWSLP) Central East Southern Lower Peninsula (CESLP) Central West Southern Lower Peninsula (CWSLP) East Northern Lower Peninsula (ENLP) West Northern Lower Peninsula (WNLP) East Upper Peninsula (EUP) West Upper Peninsula (WUP) (See Figure 2-3) u Regions are designated as follows with county distri­ butions: Region I: Region II: Region III: Region IV: East Southern Lower Peninsula (ESLP), 8 counties West Southern Lower Peninsula (WSLP), 7 counties Northern Lower Peninsula (NLP)t 8 counties Upper Peninsula (UP), 7 counties 24 Delineations followed county boundary lines closely. The four regions provide a second stage stratification and further spatial frame for broader comparative study than districts. Broad Comparative Region Apart from the eight districts and four regions, com­ parative analysis was occasionally based on three broad regions according to how homogeneous counties were. A line, stretching from the Northern County boundary of Oceana to Bay County usually divided the Lower Peninsula into Southern and Northern halves for broad comparison, with the U.P. remaining a unit most of the time. tiple regression purpose, two regions, For m u l ­ the whole state was divided into the Southern Michigan, southern study regions (E.S.L.P. embracing the two and W.S.L.P.), and the Northern Region which included the Northern Lower Peninsula (N.L.P.) and the Upper Peninsula (U.P.) (see Figure 5-3, pp. 162) . Levels of Analysis Analysis begins with the whole sample (30 counties) as representing the state (by Statistical Inference Tech­ nique); broad generalizations are made about the nature and scope of parcellation in the whole state based on sample evidence. Districts and regions are compared, 25 providing several stages and levels of analysis. Stages of analysis are vital to the detailed study of parcellation because factors underlying land parcellation (or demand for land parcels) vary considerably from region to region, dis­ trict to district and even county to county. For example, in the southern Lower Peninsula counties the demand for parcels may be described as "job oriented home site d e ­ mand", a spillover effect from the industrial urban cen­ ters. In the northern counties, the main factor is recre­ ation. Region-by-region and district-by-district analyses permits differential emphasis on relevant spatial factors. CHAPTER THREE GENERAL BACKGROUND TO STUDY Introduction The concept of land as a "commodity" is based on the fundamental social institution of "property rights". Fee simple ownership rights in land allow individuals to d i s ­ pose of their lands in any manner that enhances indivi­ dual's interests and satisfaction. Land parcellation is one mode of real estate transaction through w h i ch private citizens transfer their property rights in land to others at any point in time and for any value agreed upon. Several factors influence land parcellation. The amount of parcellation that exists at any specific time period, and the number of parcels w hi ch is created over time and space reflect the underlying d etermining factors of the parcellation process both on the demand and supply sides of the real estate market. This chapter briefly examines some of the factors which contribute to the land parcellation process in the State of Michigan. Major factors discussed are population, incomes, demand for various types of residential units, recreational and physical resource amenities, 26 and a few 27 public policies affecting land use. grouped under the broad headings of: economic, These factors can be (a) ecological, (c) social-institutional determinants. (b) The above classification of factors provide a threefold framework for the analysis of the determinants of land parcellation.^ Descriptive Framework Demand for subdivided property by small-lot parcella­ tion has been one segment of a larger increase in the de­ mand for and supply price of rural lands generally in the nation after the Second World War. Studies have identi­ fied, as the main driving force behind rural land demands, the following factors: (1) population dynamics, (2) socio­ economic developments and associated increases in per cap­ ita personal or family incomes, (3) improved communication systems which have opened up and made accessible remote areas, (4) agricultural technologies leading to capital labor substitution and the freeing of rural labor for ur­ ban industries, (5) increased demand for recreational re­ sources and for rural environmental amenities which 1Barlowe, Raleigh. Land Resource Economics (Third Edition), Prentice Hall, Inc., Englewood Cliffs, New Jersey. (Chapter One, p. 5-9, discusses "The threefold frame­ work affecting land use"), 1978. 28 reflects the changing consumption pattern with emphasis shifting from "quantity of life" to "quality of life" and finally, (6) general societal opulence.1 An understanding of the parcellation issue requires a brief examination of some of the major underlying deter­ minants and their interrelationships with land parcella­ tion process. In Chapter Five of this dissertation m u l ­ tiple regression and time series analyses are employed to determine the significance of the relationships. In this chapter, discussion is confined to the explanatory va ri ­ ables . Social Determinants Michigan Population and Demand for Homesltes Between 60 percent to 80 percent of all lands parcel­ led out are required for residential purposes. A study of 10+ acre parcels conducted in 1978 by Frank Mentius in the township of Venice, Shiawassee County, Michigan, reported that about 57 percent of the buyers of small lot parcels ^ e a d y , E. 0., and Whiting, L. R . , "Rural Development Problems and Potentials", in Rural Development in a Land Use P e r s p e c t i v e , Soil Conservation Society of America, 1974. 29 surveyed acquired their properties for rural residential purposes.1 Responses of about 13 County Extension D i re c­ tors in Michigan to letters sent by the writer of this dissertation, requesting a report on their perceptions about the 10+ acre parcels also revealed that most of the o "large" lots were acquired for rural housing estates. Parcellation is, therefore, closely related to real estate residential demand which, in turn, is usually associated with population levels, growth or movements. impact of population is, nevertheless, mixed. The direct 3 Increasing population requires new housing construc­ tions to accommodate the excess population. Where increa­ sing population is associated with rising incomes, the de ­ mand for homesites is accentuated. Housing construction implies demand for housing sites and subdivision lots developments. Even though the effects of population are Mentius, F.S. "Venice Township: A Study of Land Fra g­ mentation by 10.1 Acre Parcels, 1968-1978," Prelim­ inary Research R e p o r t . Unpublished Research Paper, 1978. o The writer, through Dr. Raleigh Barlowe, sent letters to about 26 County Extension Directors in selected counties for feedback information. Copies of some of the responses are in Appendix 3-B. 3 Demand for parcels is not always closely related to po p ­ ulation. This is especially the case for second and recreational resort homesites where purchasers may actually be living in other counties, regions or even in other states. The Northern Lower Peninsula is a case in point. 30 mixed, the following aspects of population are discussed to provide a background to the parcellation process: population growth, (a) (b) population distribution and m o v e ­ ments, and (c) population concentration or urbanization and its trend spatially and over time. Trends in Michigan Population (1940-1980) Michigan population grew at above national average rate during the 1940's and 1950's. Table 3-1 reports total population and decennial percentage change in the popula­ tion from 1940 to 1980.1 The rise of the automobile in­ dustries in the southeastern portion of Michigan, espe­ cially after 1900 and revived after the 1930 slump, pro­ vided an impetus for rural population drift towards the industrial region from other parts of the state, and from other states. Factory jobs attracted hundreds of workers from the rural areas, south. 2 from nearby states and from the rural Since 1970, Michigan's population growth has slackened considerably (Table 3-1). Increasing population and job opportunities during the 1950-1970 period contributed to the increased *The 1980 data are based on the recently published prelim­ inary report of the census. Figures are liable to change. 2 "Introduction to Michigan Population," Michigan Statisti­ cal A b s t r a c t , 1979, p. 3-5. 31 T AB LE 3-1 Trends In Michigan Population (1940-1980) Year N umber of Persons Decennial Percentage Change 1940 5,256,106 1950 6,371,766 21.0 1960 7,823,194 22.8 1970 8,881,826 13.5 1980 9,228,128 4.0 aPercent decennial gain between 1930 and 1940. population is not reported. Sources: 8 .5 a 1930 1. (1940-1960) Michigan Statistical Abstract, 1979. 2. (1970-1980) Preliminary Report, Census of Population and Housing, 1980. 32 urbanization and suburbanization, especially in the south­ ern half of the state and particularly the southeastern portion around Detroit, Flint, and Saginaw-Bay City areas where increased personal incomes enabled workers to acquire lands for homes or buy developed subdivision residential units. The 1980 population count suggests that urbaniza­ tion is still proceeding and spreading to rural counties. The focus is gradually shifting from the old urban centers to new areas, especially counties outlying the old me tro­ politan centers.*1 Table 3-1 shows that between 1970 and 1980 Michigan gained only four percent in total population, only one-third as much as the 1960-1970 rate of gain and one-sixth the 1950-1960 rate of gain. migration loss. This decline is mainly due to net A continued decline in population may cause a decline in future land parcellation since vacant homes may increase and absorb any future population in­ creases through natural growth. In this regard, the effects of population on future land parcellation would be negative. ^Michigan Preliminary Population Counts, Report on Popu­ lation and Housing, 1980. (Also see Ching-Li Wang and Lawrence S. Rosen, Preliminary Population Counts, Office of the Budget, Department of Management and Budget, 1980. 33 Prior to World War I, Michigan's population was b a s ­ ically rural and growth had been steady. About 60 percent of the total population was classified as rural by the 1900 United States population census. By 1940, 65.7 percent of the people of Michigan were urban; the rural population was split between rural nonfarm (18 percent) and rural farm population (23 percent). By 1970, 73.8 percent were urban and in 1980, 7,469,991 or 80.9 percent were classified as urban. With these changes, cities grew rapidly and sprawled over rural lands.1 Rural population declined and continues to decline despite the 2 reversed migration trend noted during the last decade. However, the decline in the rural population of Michigan varies from region to region. A partial reversal of the post-war urbanization trend was noted in 1970. Michigan population increased only 4.0 percent during the decade, the lowest rate of increase in ^"People and Society," Atlas of M i c h i g a n , Michigan State University, 1977, p. 62-89. (Also, Rathge, R . W . , and Beegle, J.A., "Urban and Rural Population Change in Michigan Counties, 1960-1975." Rural Sociology Studies No. 7, Michigan State University, July 1978. 2 Sofranko, A . J . , e d . , Rebirth of Rural America; Rural Migration in the M i d w e s t . NCRC for Rural D e v e l o p m e n t , Iowa State University Press, Ames, Iowa. 1980. 34 its history. Large percentage increases were, nevertheless, experienced by most of the northern lower peninsula coun­ ties and most particularly by those located in the center of this region. Larger numerical increases and above average percentage increases also occurred in many once rural southern counties and areas located adjacent to metropolitan population centers. The most notable loss of population (-345,671) was in Wayne County while significant percentage losses were also reported for Chippewa, Keweenaw, Ontonagon and Gogebic counties. Altogether, the 42 north­ ern counties showed an 18.5 percent increase in population. Total numbers increased in all but six northern counties and in 39 of the 41 southern counties. The population reversal trend proceeded between the last intercensal decade of 1970-80. The period saw most of the large cities losing population. Considerable m i ­ gration to outskirts and areas around major cities occurred. Secondary migration movement to central counties of the northern lower peninsula accelerated with increasing rec­ reational and environmental demands. These internal movements did not affect the general loss of Michigan's population to other states. When allowances are made for the surplus of birth over deaths, Michigan lost about four percent of its population to out-migration between 1970 35 and 1980.* Parcellation distribution closely reflects pop2 ulation distribution in most cases. However, population level and parcellation are not always closely related since second home ownership by absentees can be a major factor in land parcellation. Thus, the distribution of land par­ cellation in the state is dichotomous. lation in the E.S.L.P. oriented homesites Increasing parcel­ is associated with demand for job- where industrial workers seek large lots in urban fringe counties, far enough out beyond high residential land values, but still close within commuting distances of jobs. This demand for parcel is closely associated with population. Example of heavily parcelled dormitory counties are Livingston, Macomb, Washtenaw and Lapeer. The increase in population in the northern lower pe n ­ insula, associated with recreational and environmental d e ­ mands, explains some of the increased parcellation in the region. Presently, most of the parcels in the north are ^Preliminary Report on Michigan Population and Housing, 1980 (ibid.). 2 The relationship between parcellation and population is generally positive on the average but it is not always positive. People from the county or region may p u r ­ chase lands in another county or region. Absentee land and homesite ownerships are common in the northern lower peninsula. 36 owned by individuals who live in other counties and in other states, or by speculators and s ub di viders.1 Most of the developed parcels are for second and third homes, rec­ reational residences, marinas and tourist condominiums. According to Fletcher, between 1973 and 1977 a total of 47 recreational land developments were registered in Michigan by outside developers. This involved some 66,589 2 lots and 18 subdivisions. Fletcher points out that most of the subdivision lots in the nation are vacant and are purchased for speculation or investment purposes or for the purpose of building future leisure homes. Some of the lots are targeted for resort condominiums in anticipation of expected boom in the tourist industry. This observation is relevant to the parcellation situation in the northern lower peninsula of the state. County Extension Directors from the region pointed out that premature parcellation is 3 a common phenomenon ^his in the area. Given current trends in information was obtained through correspondence with County Extension Directors. 2 Fletcher, J.E. "A Systematic Approach to the Analysis of Land Sales Regulatory Programs: A Case Study of the Michigan Land Sales Act of 1972". (Ph.D. D i s s e r t a ­ tion, Michigan State University, 1978), p. 157. 3 County Extension Directors, Antrim and Emmet Counties. Ibid. (Appendix B-3). 37 Michigan's recreational industry, it is likely that second home ownership by absentees who like to spend vacations in the north will remain as the m aj or reason for parcellation in the region for many years. It is expected that the wide differences in the dis­ tribution of levels and rates of change of population in the state should reflect the variations in the amount and trends of land parcellation in the state. For example, lots already created in metro counties will be left idle, but new parcels would be created in the new center of p o p ­ ulation gravity. A declining trend in parcellation should be observed in the source regions while an increasing trend should be apparent in the receiving regions. The impact of migration on land use is closely related to the migrant character. An analysis of migrants by age indicates that the northern lower peninsula is attracting mostly adults and older individuals. These migrants have interest in real estate and tend to stimulate parcellation activities. The southern lower peninsula mostly attracts youth m i gr at in g towards job and education centers. direct Their impact on land is a derived one— they create a con­ centrated market for food and housing. Since they tend to be t r a n s i e n t , developers are wary to rely on them. Their purchasing power is also too low to induce brisk housing construction activities. Adult migrants tend to be more 38 permanent and constitute effective demand for land. Their impact on land has been observed in some of the responses by County Extension Directors in the N.L.P. counties. R e­ versed migration involving urban to rural movements have several negative consequences on land use pattern and de­ velopments. Urbanization and Urban Exodus in Michigan Urbanization had its first stage impact on land use when the movement of workers to towns and cities called for the development of land areas for compact urban com­ munities. A more complicated second stage emerged with the swelling of the urbanization and suburbanization move­ ments after World War II. New urban residents brought tremendous demands for additional housing and urban facil­ ities. But at the same time, increased individual mobility, demands for larger lots, the declining attractiveness of central city living, and public policy inducements for suburban development favored an outward movement of city residents to suburban and sometimes rural locations. This stage has been followed in many instances by one of active flight of residents, industries, businesses, and jobs from central cities and the consequent deterioration of once viable downtown areas. It remains to be seen whether this process can be reversed or if there would continue to be 39 incessant demands for the development of new lands for urban and suburban uses around the fringes of cities while their older sections suffer from underutilization, blight, and d e c a y . Suburbanization, scatteration of residential develop­ ments in the rural-urban fringe areas around cities, and the outward spread of urban-oriented uses has resulted in a luxurious and often wasteful use of lands for these purposes. Southern Michigan had 669,000 acres used for various urban purposes in 1940, and 1,722,000 acres in 1961. 1,058,000 acres in 1955, By 1977, the acreage used for urban and built-up areas both in and around cities had increased to 3,287,000 acres.1 Both suburbanization and urban exodus to countrysides have several implications for rural communities and rural industries and economy, besides their impacts on cities that serve as source regions. In the rural areas, developments may occur, but local farming communities are usually unprepared for the in­ creasing tax burden, the high assessed value of their farm­ lands and the conflicts in land use which eventually arise. Where the exodus is followed by industries job seekers of lower income class follow suit and some of the urban p r o ­ blems are transferred to the rural areas. Parcellation ^Barlowe, R . , e t . al., Preliminary Report on Michigan Land Resources by the Michigan Task Force; 1981. 40 proceeds at faster rates and home lots tend to be larger and secluded. Farmlands tend to shift to rural residential and industrial uses as a result of "push" and "pull" forces on the farmers. "Push" forces are associated with in­ creasing assessed values and tax burden on their land as well as court cases involving pollution, nuisance, etc.; "pull" forces are related to capital gains as land values rise. Eventually the agricultural sector quickly gives way to residential and industrial land use. From the standpoint of land conservation, the process is double-edged. Urban land use, abandoned in cities, have little or no potential for any rural industrial activities. Rural farmers cannot move to urban areas to pursue agricultural activities. However, rural lands are quickly transformed to urban uses. In both instances, good lands are locked up in urban uses irretrievably, especially since most Michigan cities grew and sprawled on good farmlands. Residential developments associated with this type of movement usually consumes large amounts of lands for rural industries. The urban demand for land will be affected by a number of factors such as individual choices concerning residential and building sites, employ­ ment patterns, energy developments, respecting land use. and public policies In this regard, the topic of housing demand requires particular attention. 41 Housing Units Population trends are only one indicator of growth and development in any community. Another important indicator is new residential developments which have considerable impact on land use pattern. The rapidly declining family size, as has occurred in the state and generally in the nation in recent years, can often mask significant new residential developments. Approximately 50-65 percent of all small-lot parcel­ lation are utilized for residential purposes. Increasing demands for new homes implies increasing parcellation. Table 3-2 reports the number of occupied housing units in Michigan from 1940 to 1979. Data are decennial. The total number of occupied housing units has been increasing numerically and in absolute terras over the last 40 years. Incremental rate has, however, been declining over the period. The total and annual percentage figures in Table 3-2 reflect a steady declining rate. Seasonal and m i ­ gratory bousing units have been increasing during the last 20 years. Seasonal, second and third homes have become important in the state's recreational and fringe counties. Most of the second and third homes are found in fringe and rural counties as well as in counties with tourist attractions and recreational parks. 42 The declining rates in housing developments are p a r ­ ticularly more evident in S.M.S.A. areas of the N.L.P. counties than rural and counties contiguous to metropolitan TABLE 3-2 Number of Occupied Housing Units in Michigan: 1940-1979 Number of Occupied Housing Units and Percent Change + Total % Change Annual % Change Year Number 1940 1,396,014 1950 1,790,702 28.3 2.8 1960 2,239,079 25.0 2.5 1970 2,653,059 18.5 1.9 1979 3,029,000 14.1 2.0 Source: centers. - - Michigan Statistical A b s t r a c t . 1979, 14th Ed., Research Division, Graduate School of Business Administration, Michigan State University, East Lansing, Michigan. For most of the rural counties, housing develop­ ments are increasing at increasing rates even though the number of housing units remain very low. Increasing rates of development of housing units should be associated with 43 increasing rates of land parcellation while high existing levels of housing units should be directly associated with high parcellation levels. It should, however, be noted that the demand for homesites is dichotomous— one for industrial worker residential units, the other for recre­ ational activities. Furthermore, in some cases, the re ­ lationship between number of developed housing units and demand for parcels for homesites lags. Information from County Extension Directors confirms that most recreational second home lots remain undeveloped and a developer pointed out that he expected only 20 per­ cent of his recreational second home lots to have houses on them by the year 2000.1 This points to the fact that lots around cities are usually built upon soon after pur­ chase, while a large number of lots in the recreational areas are bought as investments which owners will build on or sell later. This may account for regional variations in the effects of housing units on parcellation. Table 3-3 reports regional distribution of number of housing units in Michigan, based on most recent census data. The high percentage change figure for the N.L.P. p a r ­ ticularly reflects the impact of reversed migration and the increasing numbers of second and recreational homes in *This information was obtained second source from Dr. Raleigh Barlowe, May 19, 1981. 44 the region. The U.P. figure is generally attributed to in­ flux of students and workers into counties such as M ar ­ quette, Houghton and Delta. Since demand for housing units TABLE 3-3 Regional Distribution of Housing Units (Occupied and Vacant) 1970 and 1980 Number of Housing Units Region 1970 1980 Percent Change 1970-1980 I E.S.L.P. 2,042,884 2,319,213 13.5 II W.S.L.P. 600,599 809,245 34.7 III N.L.P. 205,676 299,460 45.6 IV U.P. 108.144 153.305 41.8 2,957,303 3,625,003 22.6 Source: Preliminary 1980 Population Census Report. is closely associated to population and its impact on land parcellation is direct (occupied or unoccupied housing unit locks up land) it is preferred over population in the specification of the regression model, though both were retained. 45 Conclusions The above population characteristics indicate that if population and parcellation are positively correlated, then: a) The S.L.P. should show very high level land frag­ mentation but the creation of new parcels should exhibit declining trend. b) The N.L.P. should show moderate to low level parcellation but the creation of new parcels should exhibit increasing trends, especially lots of large size categories (10 to 10.9 acre parcels). c) The U.P. should show low to very low level p a r ­ cellation with constant to slight increasing rates in the creation of new parcels. d) Parcels in the S.L.P. should be relatively small in sizes, while the lot sizes in the N.L.P. and the U.P. should be relatively large. Hence the problem of 10 and 10+ acre parcels should be more acute in the very rural areas of the S.L.P. and the N.L.P. e) Parcellation density should be high in the south, decreasing northwards. 46 Economic Determinants Economic Activities Population concentration and economic activities are almost synonymous. The two factors are extremely related. Economic opportunity in an area tends to attract population; at other times, population concentration creates a market which in turn attracts economic activities. The two fac­ tors are both the cause and the effect of their inter­ actions. The strong causal relationships between population and economic activities make it almost impossible to dis­ cuss them in isolation. The supply of land is affected by the market price of land which influences the land owner's willingness to sell or parcel out raw land, or to convert used land to alter­ native uses. Trends in land values generate speculative behavior among land owners and developers who usually aim at increasing their capital gains. Effective land supply depends not only on the real estate market condition but also on a complex interrelationship between ecological, institutional and personal factors. Several institutional arrangements may facilitate or impede the theoretical frictionless real estate transactions. These are dis­ cussed under the other determinants. Demand for land given biophysical considerations is influenced by land values, information about land and 47 prices, institutional arrangements and governmental poli- cies, and the preferences of the buyers and sellers. Effective demand is a function of the ability of the in­ dividual to pay for the land. This may be facilitated by certain financial arrangements in the system— either p ri ­ vate or public, e.g. bank credit facilities, mortgage arrangements, interest rates, and public credit facilities. To examine the relationships between land parcellation and economic factors, assessed valuation of farmlands and in­ comes were selected for brief discussion. These two fac­ tors are included in the multiple regression model speci­ fication. Employment is so closely related to population that it is excluded from the models and hence is only briefly discussed. A very brief discussion of Michigan agriculture is also provided as a background to trends in land values and incomes. Michigan Agriculture The importance of Michigan's agricultural industry does not rest on its added value or the employment it offers, but on the fact that it has made the state selfsufficient in many products. Several citizens, especially farmers and public officials are, therefore, concerned about losing this industry. Michigan State University, The Land Grant University, is one of the largest in the 48 world. In terms of employment and added value, the manu­ facturing sector, led by the automobile and Its allied industried and the services sector— commerce, transpor­ tation, finance, etc.— are by far the most important. H ow ­ ever, agriculture is usually considered the second most important sector of the economy, followed by the growing tourist industry. Despite its acclaimed importance, Michigan agriculture has and will continue to face many challenges. The question of development of the resource has been a very important issue. There are many competing industries all desiring a substantial quantity of the landscape. Development of housing, upgrading of highways, expansion of airports, dustrial and commercial improvements, in­ and land for recre­ ation all compete with agriculture for Michigan's land surface. The development of any one of these industries may have an adverse affect on surrounding farms. A housing development in a predominantly agricultural area may result in loss o’f more land because of the noncompatibility of the two uses. Highways likewise directly take substantial acreage for each mile of new construction. Although most of the highways in the interstate system have been completed in Michigan, there are still several significant areas where highway development is contemplated. This development will result in the direct loss of surface 49 area and also indirectly impact agriculture as development of other types increase. Airport expansion has also util­ ized substantial agricultural land. Recreational demands are increasing, with more leisure time and apparently more money to invest in recreational vehicles. Citizens of Michigan are demanding more use of the land, and frequently trespass on private land with resultant crop damage. The concept known as right-to-farm has recently been revived in Michigan. Typically,, agriculture has been re­ quested to move out when some other "higher priority" use is considered. The concept of right-to-farm is simply that agriculture is, in its own right, a legitimate use of the land and that other uses must coexist with agriculture. Michigan's tax policy also does and will continue to have a very significant impact on Michigan agriculture. Property tax, sales tax, inheritance tax and income tax, all directly affect agriculture. Tax policy is a signi­ ficant force in preserving farmland or promoting the sale of farmland. As the state's second largest and most stable indus­ try, the decisions made on development, taxes, and the right-to-farm will have a significant impact on the viability of Michigan agriculture. 50 Based on the various U.S. Censuses of Agriculture^the number of farms, the acreages in farms and the acre­ ages of cropland declined steadily throughout the 1940-1978 period. The state has less than a third as many farms in 1978 as in 1940 while the acreage in farms dropped to 61 percent and the acreage of cropland to 70 percent of the earlier levels. Acreage of cropland harvested dropped at a slower rate, reached a low of 5.5 million acres in 1969 at a time when federal programs were still being used to hold down production, and then responded to improved market incentives by increasing to 6.8 million acres in 1978. Average farm size, however, increased during the period. These trends were similar for all regions even though more than 80 percent of the state's farmland is found in the southern 41 counties. Prior to 1974, Agricultural Census Reports were based on supposed complete enumeration. The Reports for 1974 and 1979, in contrast, are based on statistical sam­ ples. Both approaches do not yield comparable results and, in fact, there has been suspicion of underenum­ eration. The Bureau of the Census has estimated un ­ derenumeration of about 17 percent in 1969, 13 percent in 1974 and 11 percent in 1978. Information based on agricultural statistics in the state must, therefore, be interpreted and used with caution. According to the Report of the State of Michigan Task Force for Natural Resources and Public Policy, the uncorrected totals of 1978, reported by the operators of the sam­ ple farms probably provides a much more comparable standard for indicating actual trends in farmland use throughout the state. 51 Several observations may be made concerning Michigan's farmland use trends. The number of farms has declined at a rapid rate and will probably continue to decline for some time. This trend will be favored by the national trend towards larger farms and the fact that approximately 30 percent of the farm units in the state now involve holdings of less than 50 acres. The 1978 census indicated that only 45.3 percent of the farm operators regarded farming as their principal occupation and that 53.3 percent worked 100 days or more off their farms. More than half of the operators in the southwestern, northeastern and southeastern regions of the state indicated that they were part-time farmers who worked 100 days or more off their farms. The highest rates of off-farm employment were reported for the most part in areas near urban and industrial developments while lower rates were reported in the more strictly agri­ cultural areas. Michigan has probably gone through its period of major reductions in farmland and cropland acreage. However, the area in farms is still declining and will probably continue to do so as some operators discontinue their farming oper­ ations and as farmlands are acquired for various urbanassociated uses. Some cropland will be lost to agriculture in this process; but in the case of operators discontinuing operations, much of the better farmland will be taken over by other operators. 52 Despite the decline in number of operators and farm­ land acreage, it is interesting to note that the acreage of cropland harvested increased 1.2 million acres between 1969 and 1978. This acreage represented 80.7 percent of the reported cropland area in 1978 as compared with 66.1 percent in 1940, 70.6 percent in 1950, 71.9 percent in 1959 and 64.1 percent in 1969. Ups and downs in this total can be expected with variations in weather conditions and the economic climate for farming. domestic and international, Favorable markets, both for farm products are largely responsible for the higher 1978 acreage. One negative as­ pect of the trend relates to the fact that a smaller area of cropland was probably planted to soil conservation crops in 1978 than in earlier years. It must be recognized that those prime lands that lie in the path of urban growth are, generally, in danger of being taken unless a more strict protection policy is adopted. Major hindrance to adopting stricter policies than currently existing is a constitutional and political issue. Land Values Average value of farm real estate per acre has been increasing (Table 3-4). Average farmland values for Michigan from 1973-1979 for selected years are compared with other states from selected regions. 53 All the selected states show an increasing trend in farm real estate value per acre. This is due in part to TABLE 3-4 Average Values of Farm Real Estate Per Acre Michigan and Selected States (1973-1979) Region and State Value Per’ Acre in $ North East Region Maine Vermont New Jersey 1973 1975 1977 1978 $ $ $ $ $ 253 346 1,337 341 462 1,807 400 541 2,004 441 597 2,057 485 657 2 ,222 444 328 269 553 434 429 767 583 652 860 690 730 955 807 854 505 294 567 706 396 846 1,121 526 1,431 1,263 602 1,581 1,516 674 1,786 * 108 94 199 195 142 296 258 194 376 273 227 380 306 257 437 1979a Lake States Michigan Wisconsin Minnesota Corn Belt Ohio Missouri Illinois Northern Plains N. D ak o t a S. Dakota Kansas Source: Michigan Agricultural S t a t i s t i c s , July 1979, MDA. P r e l i m i n a r y Data the rising costs of capital investments on the land which reflects inflation, and partly to the increasing expected 54 value of the farmland itself as urbanization encroaches on them. States such as New Jersey, Ohio, Illinois and Michigan which are highly urbanized tend to have higher land values than rural states such as Maine, North and South Dakota and Kansas. Associated with increasing land values is assessed valuation and tax burden on land owners, a possible push factor in agricultural land transference. Increasing land values should, therefore, reflect increasing parcellation process. Income One of the most important factors underlying land p a r ­ cellation process in Michigan is personal incomes. The per capita personal income in Michigan has often exceeded national average since and as far back as 1921 when the ratio of Michigan average per capita income to that of the nation was about 1.13 or 113 percent (index). percentage has been declining over the years. However, the It reached a peak of 119.8 percent in 1953, began to fluctuate, reached the lowest figure of 102.2 percent in 1975 and since then has began to pick up again. 1978 estimate places the per­ centage at 108.1. Vith declining automobile demand, and competition from Japan and other automobile manufacturing countries, the Michigan average per capita personal income may converge to national average. However, it should be observed that 55 per capita income in dollars has been increasing nationally; in Michigan it increased from 1950 level of $1,704 to 1977 level of $7,606. National average increased from $1,430 to $7,026 correspondingly. Wide variations exist, however, among counties and regions. In 1977, per capita personal income in Michigan ranged from a low of $3,984 in Oscoda county to a high of $9,776 in Oakland county. The average that year was $7,606 and about 66 counties fell below this average. This indicates a highly positively-skewed income distribution among counties. The persistently high level personal income in the state, and the increasing purchasing power of families as two persons (husband and wife) begin to earn incomes together, have supported an effective demand for high quality residential units. Michigan real estate construc­ tion industry has usually enjoyed large domestic market. Drops in income observed for 1960, 1970 and 1974 coincided with unemployment highs. This close relationship between incomes and employment calls for a brief examination of Michigan's employment situation. Employment Employment effect on land parcellation in Michigan is dual in nature. Job opportunities have both concentrated Michigan's population and further increased Michigan's 56 people's purchasing power through increased per capita income. The housing construction industry is closely linked to the economy of the state. Increasing unemploy­ ment tends to slow down the housing industry as c o u p l e s , uncertain about their future earnings, refrain from taking mortgage loans. Income is, however, used in the regression model in place of employment as a proxy of economic activity. In general, the unemployment rate in the state has been above national average. In 1975, the unemployment rate reached a peak of 12.5 percent. 1958 had been a high unemployment year (13.7 percent) in the state. average of unemployment National (1977) was 7.0 percent and Michigan reported 8.2 percent that year. 1978 rate showed a slight decline in unemployment nationally (6.0 percent) and for Michigan (6.9 percent). began to increase. Since then, unemployment has again In 1980, Michigan's unemployment level reached 12.5 percent of its labor force. The February 1981 figure for unemployment in the state was a high of 14.2 percent. If such a trend continues, the housing industry will slow down and land parcellation and subdivision devel­ opments may slacken. As already noted, ment causes out-of-state migration. increasing unemploy­ Most of the migrants have been young college graduates and low income unskilled workers with relatively minimal impact on land parcellation. 57 Increasing income and e m p l oy me nt , nevertheless, in­ crease the purchasing power of the low income earners and enable a proportion of them to afford single family resi­ dential units. It could, creased employment therefore, be argued that in­ increases incomes and affluence, which in turn increase demand for rural amenities. However, a stronger urban pull as a result of general economic well­ being of the cities is much more likely to weaken the push factor; with proper land use control policies, communities may be able to deal with the pull factors of migration and discourage urbanites from invading rural lands. Biophysical Determinants Ecological Factors Physical factors such as soils, climate, forest, to­ pography, etc., play an important role in human settlements and activities (economic, social and political). It must be recognized that heavily urbanized and industrial areas are often located on productive soils for agriculture and forestry, where the climatic conditions are more favorable for the two industries and other human activities. 58 Forests The State of Michigan still possesses considerable forestlands and rural environment, with unique historic re ­ sources. About 54 percent of the state is classified as forested. Of the 19.1 million acres of forest in 1965, about 9.0 million acres or 47.2 percent were found in the Upper Peninsula and 7.5 million acres or 39.3 percent in the Northern Lower Peninsula. million acres or 33.3 percent The public owns about 6.3 (state, local and federal). These public lands are in general found in the Upper P e n i n ­ sula and Northern Lower Peninsula and are mostly in the form of undeveloped forest, parks, wildlife refuges and open spaces, wetlands, and so on; part of the public hold­ ings are managed for multiple uses. Government, To date, the Federal through the Forest Service, 2.7 million acres of land in the state. USDA, owns about Within the bo un ­ daries of the federally-owned lands National Park Services and Fish and Wildlife Services have extensive holdings. The State of Michigan, of Natural Resources, through the Michigan Department administers over 4.3 million acres of forest, parks and wildlife; they are m an aged on multiple- use basis and provide recreational facilities. ernments' Local gov­ land holdings and recreational resources are scattered and not very significant. 59 Areas with public resources attract many tourists from all parts of Michigan and from other states. Demand for natural environmental resources is on the increase in the whole nation and Michigan's richness in forests, lakes, water, historic monuments, wildlife and other scenic attractions have made tourism and recreation a prominent industry in the state. Demand is particularly high for water frontage and counties such as Antrim, Charlevoix, Cheboygan and Newaygo, and almost all the U.P. counties have considerable recreational potential. This potential is associated with the demand for small parcels. It may, therefore, be concluded that parcellation in the North is not directly related with county population, but rather correlated with recreational homesites. Recreational Amenities A study of Michigan recreational activities indicate that the north of the L.P. and the U.P. are the focus of most summer recreational activities. From the metropolitan regions of the South and the Chicago-Gary areas, individuals converge on State and national forests for skiing, camping, deer hunting, fishing and boating. Major centers of at­ traction such as the Sleeping Bear Dunes in Leenanau county, Hartwick Pines in Crawford, Mackinac Island and Bridge, Fort Mic hi li ma ck in ac , Sault St. Marie Locks, 60 Tahquamenon Falls, Pictured Rocks of Alger and the P o r ­ cupine Mountains are all found up north of the state. Sub­ divisions line up all along the lake shores and counties such as Grand Traverse, Leelanau, Antrim, Keweenaw and Houghton are bustling with real estate activities in sec­ ond and recreational homes, marinas and resorts. Michigan has over 170,000 seasonal and second homes.1 Recreational activities in the south are over utilized. In 1975, about 600 million participant-days were recorded for 20 popular recreational activities— a tremendous pres­ sure on the state's recreational resources. Much of the activity was concentrated on relatively few heavily used areas of private, local government and state lands in the Southern Lower Peninsula. In general, the major factors attracting individuals to the north of the L.P. and the U.P. are recreation and environmental amenities and clearer air. Buyers interested in the out-of-doors are buying before prices rise further. Some owners may be wealthy, but most represent middle in­ comes and a surprising number are in low income brackets. Besides, construction and developments of real estates, multi-family housing units and condominiums are likely to ^Michigan Atlas, 1977, (ibid.), pp. 171-202. Also, see Santer, Richard A. (1977) M i c h i g a n , Heart of the Great L a k e s . Kendall/Hunt Publishing Co., pp. 254-264. 61 increase land withdrawals in this area. Speculation and premature land sales have already been noted in many of the West Northern Lower Peninsula counties (e.g. Antrim, Charlevoix, etc.) and many subdivided lands sold on a land contract basis are still not on records and hence do not appear on the Plat Atlas maps. An increasing land parcellation problem is expected in the recreational areas in the future as developers, sumers and subdividers converge on the region. con­ In the re­ gression model, public parks and air quality indexes are included as explanatory variables and proxies for ecologi­ cal determinants. Institutional Determinants Governmental Regulations Apart from the self-regulation implicit in the land development and sales industry, several governmental reg u­ lations in Michigan are designed to direct the pattern of land use. Regulations at local, state and federal levels affect the acquisition of landed property, registration of land titles, leasing arrangements, mortgages and land d e ­ velopments. However, the local governments (municipalities, county and township) have the major responsibility for con­ trolling the location and quality of land developments 62 through the exercise of the "police power" to protect the health, safety and general welfare of its citizens. Direct and indirect controls such as zoning and subdivision regu­ lations, development rights arrangements under Land Re ­ source Protection Act, Subdivision Control Act, building health and sanitary codes, e t c . , are a few of such public controls circumscribing the individual fee simple owner's property rights and therefore directly or indirectly in­ fringing on the sales and purchases of land parcels. It should be noted that these public instruments may be used to restrict or facilitate land transactions. They work in both ways, depending upon the objective of the government relative to land resources at any point in time. Land use policies are treated as shock variables in this study. The Subdivision Control Act of 1976 (PA 288) is selected as the main shock policy variable affecting trends in parcels less than 11 acres. Other Acts and land use controls such as zoning are not examined or included in the model. The impacts of zoning on land parcels for residential purposes have received ample study and exten­ sive discussion in the literature. For example, Nelson has excellent discussion of the impact of zoning on properties in his book, Zoning and Property Rights; An Analysis of the American System of Land Use Regulation 63 (1977).1 Jud has also studied the effects of zoning on single-family residential property v a l u e s , in Charlotte, North Carolina. 2 Jud cites several other empirical works on zoning and its impacts on land use. The Subdivision Control Act (PA 288 or 1967) The Subdivision Control Act was adopted as a vehicle to empower local units of government to pass ordinances to regulate planned unit and cluster developments as well as the conventional subdivision. A ccording to D'Amelio: "Properly applied, the provisions can combine zoning and subdivision control into a single administrative process by adopting a subdi vi ­ sion o r d i n a n c e ."3 The Subdivision Control Act was revised and enacted during 1967 and became effective in January 1968. has often be referred to as the Plat Act. The Act Although the Nelson, Robert H. Zoning and Property Rights: An Analysis of the American System of Land Use R e g u l a t i o n ; 1977, The MIT Press. 2 Jud, G. Donald. "The Effects of Zoning on Single-Family Residential Property Values. Charlotte, North C a r o ­ lina." Land E c o n o m i c s . Vol. 56, No. 2, May 1980, pp. 142-154. 3 D'Amelio, R.S., Director, Local Property Services D i v i ­ sion, Dept, of Treasury, "The Subdivision Control Act and PUD's" in The Michigan Survey N e w s l e t t e r , Vol. 10, no. 4, Fall i s s u e , 19757 64 majority of the Act is devoted to platting procedures, a key element of the Act is Section 102 (d) which allows only four land subdivision without platting, each of which may be less than 10 acres within a ten-year period. Should a landowner desire a fifth split less than 10 acres, that person must formally plat the land. be very financially rewarding, Even though this can it is also expensive and may require a year or more of time. The Subdivision Control Act has raised a number of questions relative to the limits to division and sale of land. Section 102 (d) Act 288 of 1967 defines "Subdivide” or "Subdivision” as: The partitioning or dividing of a parcel or tract of land by the proprietor thereof or his heirs, executors, adminstrators, legal representatives, successor, or assigned for the purpose of sale, or lease of more than one year, or of building developments where the act of division creates five or more parcels of land each of which is 10 acres or less in area; or five 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.1 The above definition, therefore, exempts any p a rt i­ tioning which creates parcels or lots slightly bigger than the 10 acre minimum. Act, Together with the Michigan Land Sales (PA 286, 1972), developers and subdividers could ^"Attorney General's Opinion 4804, April 25, 1974. 65 create as many as 24 individual parcels, each parcel slightly larger than 10 acres, at any single act of sub­ dividing. 1 The Act's impact on land parcellation in rural areas is a major concern for rural land use policy. Some policy makers contend that parcels of one or two acres in size are sufficient for residential development in rural areas. Among the reasons, they argue that less land will be re­ moved from agricultural use because the parcels are small. The opposite approach claims that acres minimum for a homesite, serve agricultural land. large parcels, e.g. 40 is a better method to p r e ­ This would limit the number of buyers who could afford to move into area while maintaining large blocks of land that could easily be leased to full­ time farmers. Both sides can present strong cases to sub­ stantiate their viewpoints. The Michigan Land Sales Act (PA 286 of 1972) is not a sub­ ject for discussion in this dissertation. However, it and the Subdivision Control Act (PA 288 of 1967) t o ­ gether constitute the most important land transaction regulations which determine the number and amount of parcels an individual in the state can create during any single act of land partitioning. Paragraph 565. 804, Section 4(b) of PA 286 of 1972 exempts land which is divided into fewer than 25 parcels from the p r ov i­ sions of the Act. The Act requires that a detailed statement of record with property report be filed with the Land Sales Div. of the Michigan Dept, of Licensing and Regulations. Thus, an individual who offers less than 25 lots, parcels, units or interests, including condominium and time share units, if offered as part of a common promotional plan, regardless of the size of the lot, is exempt from the provisions. 66 For example, an independent study conducted during 1979 examined land use on 10-acre parcels in four townships in Washtenaw county, viz., Freedom, Manchester, Pittsfield, and Salem. Each contains approximately 23,000 acres. Al­ though these were not selected according to the County in­ dex of "ruralness" or specifically to examine new housing, the findings are related to land use activity. The inves­ tigators determined that 134 10-acre parcels have been created in the four townships between 1964 and 1977. Eighty-three of these were in Pittsfield and Salem, two townships near the urban center, while the remaining 52 were located in the more distant townships of Freedom and Manchester. Only 13 of the 134 parcels remained in large scale agriculture. Twenty-eight of the 10-acre parcels are used for pleasure houses. Each of the 28 contained a relatively new home usually located on the road frontage of the property. Seventy-three parcels appeared to have nine acres of idle land with a new home and a large lawn, while the other parcels may be awaiting new housing in the near future. The remaining land use on the 10-acre parcels was primarily devoted to woods. In these instances, new expensive homes could be observed to occur frequently at the end of the private access roads. 67 Analysis of the Impacts of Subdivision Control M i ni mu m Lot Regulation on Rural Lands Empirical studies have shown that the setting of m i n ­ imum- lot size either by zoning or subdivision control re g ­ ulations tend to create supply of and to induce demand for lot sizes in excess of the regulated m i ni mu m lot size e s ­ pecially for simple family residential developments.^- The analyses of the effects of m i n im um lot regulation is very simple. Generally, m in im um lot regulations (by Zoning or S u b ­ division Control Acts) tend to distort the equilibrium real estate market. Given the supply of and demand for land for competing uses, market forces eventually create m i n i ­ mum lot sizes and parcels for each land use (ignoring the social and ethical implications). land use classifications Zoning regulations and (districting) w hich seeks to in­ troduce uniformity into community land use have strong positive effects on residential properties. Land purchasers seek mu ch uniformity in land use and w ou l d be wi lling to pay a p re mium for it. Where public controls are absent, a m e c h a n i s m w ou ld eventually arise to meet the expressed 2 needs of the residential consumer. XJud, G. D., 1980 (ibid.), p. 152. 2Jud, G. D., 1980 (ibid.), p. 151. 68 Minimum lot regulations which create larger lots than equilibrium lot size tend to lower the cost of single­ family residential housing units constructed on large lots.* If the 10-acre minimum lot size provided in the Michigan Subdivision Control Act is greater than what real estate market forces would have eventually determined (0.5-0.6 of an acre) for single-family residential units, it should be expected that more larger lots would be created and demanded for housing construction. Figure illustrates the analytics of the impacts of minimum lot regulation. The cost per unit lot construction is lowered when minimum lot size is set. This enables developers to re­ duce the unit price per residential unit to consumers. Setting minimum lot size above market equilibrium lot size, thus, increases the supply and reduces the price of large lot residential land. Figure 3-1 is based on the assump­ tion of a free real estate market and assumes various elasticities of demand and supply. In Figure 3-1, four panels are constructed to show the linkages between the real estate market sale of land for single-unit residences and minimum lot regulations. Panels should be read clockwise from a...b. In panel (a), *0hls, J. C . , e t . a l ., "The Effect of Zoning on Land Values." Journal of Urban E c o n o m i c s . October 1974, pp. 428-44. 69 C/L C/L Cost/Unit Lot Size So Co s c, s RL Lot Size No.of Residential Lots P/H P/H ^Bo Price/Unit House w3 O X Po "c v7 S T “~ r - - - - - Pi Ho 0 Ve S ' X - H, Po Pi C0 SLH No.of Single-lot Houses v6 Ho B, H, 0 SLH No.of Single-lot Houses FIGURE 3-1 ANALYSIS OF THE EFFECTS OF MINIMUM LOT CEILING ON PARCELLATION 70 real estate static equilibrium m i nimum lot size (lot size is in acres) at Cq . At these equilibrium lot size and cost, developers would purchase R q units of lot for single unit residential developments (panel b). traces the cost-lot demand relationships. The curve A A o o Declining lot cost implies increasing demand for more lots, given lot size. Assuming no speculative demand, all lots purchased are converted into housing units and sold to home consumers (panel c). The curve 13 13 o o in panel (c) indicates that as the price per unit of housing lot declines, more single lot houses would be demanded, supply at the point. given infinite elasticity of Thus at P Q , H Q d of single lot houses w ould be purchased and this is consistent wit h the lot size Lq , (via panel d). If the legislature adopts a minimum-lot regulation w hi ch sets the size at L, above market equilibrium lot size Lq , (panel a), market equilibrium is distorted, relevant supply curve for lot sizes is S ^ ^ ^ l ^ o * and the ^2 ^ e ” comes the equilibrium point of intersection between dd and and cQ falls to c^. The response between large lot sizes and unit cost has been ascertained empirically by several s t udies.1 With the fall in unit cost of lot size ^Grether, D.M. and Mieszkowski, Peter. 1974. "Determinant of Real Estate Values." Journal of Urban E c o n o m i c s , 1 (April) pp. 127-46. Also see: Maser, S.M., et. al. 71 for construction purposes, more large lots are demanded for single unit housing developments. Rq increases to R^ and more such housing units are provided for consumption increases to All things equal, the price per unit of large lot residential units fall correspondingly, PQ to from and this would be consistent with the regulated lot size £. In the case where penalty is incurred by creating lot sizes in the range of 0 — L (L inclusive) e.g. the 10-acre minimum lot size of Michigan, land owners and sellers (subdividers) would usually sell lot sizes slightly in excess of L in order to escape the platting expenditures. This further reduces the per unit lot costs to developers, who would then buy the larger lots for the development of single family residential units. Large lot houses tend to be relatively cheaper, by hypothesis, and usually provide extra land for wood lots and other ecological features. "The Effects of Zoning and Externalities on the Price of Land: An Empirical Analysis of Monroe Co., New York." Journal of Law and E c o n o m i c s , 20 (April) pp. 111-32. Moss, W . G . , 1977. "Large Lot Zoning, Prop­ erty Taxes and Metropolitan Area." Journal of Urban E co no m i c s . 4 (October) pp. 408-27. Sagalyn, L . B . , and Sternlict, G . , 1973. Zoning and Housing C o s t s . New Brunswick Rutgers University Center for Urban Policy Research. 72 The demand for large lot is usually Induced. The effects of mini mu m lot regulations on farm lands are there­ fore obvious. It leads to more farmlands being withdrawn into non-farm residential uses. It is expected that the Michigan Subdivision Control Act which provides for 10-acre m in im um lot size should reflect in the rate of change of lot sizes in the range of 1 0 . 1 to 10.9 acres in this study. Conclusion Chapter Three of this dissertation has revealed that there are several factors contributing, either independently or jointly through their interactions to land parcellation and subdivision developments in the state. Some of the factors such as population may be major determinants, but the impacts of population are mixed, depending on the m a g ­ nitude of population growth rates and densities at any given place and at any point in time as well as population movements and redistribution over space and over time. Its impacts on the rate of parcellation may be positive for frontier zones or negative for the more mature regions which are serving as source areas for the reversed m i g r a ­ tion trend. Agriculture and forestry tend to have negative impacts on parcellation in that areas with high agricultural a c t i ­ vities do not normally attract high density residential 73 concentrations, and concommitant manufacturing and com­ mercial activities. Recent trends, however, indicate that agricultural and forestland are perceived as recreational amenities and are attracting individuals who want to enjoy rural environment. Gradually, the forest and open spaces in Michigan as well as farm areas are being opened up, becoming more and more accessible and lands are gradually being transferred to non-primary uses. development. Farmlands are the easy targets for Recreational demand is, therefore, becoming a major factor in land use pattern. The trends in land parcellation in Michigan can be appreciated only if trends in these factors are kept in mind. CHAPTER FOUR HYPOTHESES AND STATISTICAL MODELS Chapters One through Three identified the study vari­ able (parcellation) and the possible determinants which may contribute to the problem of land parcellation in Michigan. Contributing factors were examined under the framework of Social, Economic, Biophysical and Institutional Deter­ minants. Based on the background analysis, hypotheses are stated and appropriate general statistical models to test these hypotheses are formulated in this chapter. Working Hypotheses This study explores three major hypotheses. The hy ­ potheses are tested statistically at the a * 0.05 (signifi­ cance level). Any amount of parcellation noted in 1963 suggests that land parcellation exists in the state. Based on this assumption (by study definition) three hypotheses are tested. The first hypothesis deals with trends in land parcellation; the second with time and spatial distribution and the third with the impacts of the Subdivision Control Act, (PA 288, 1967). 74 75 1. The Time Trend Hypothesis Based on the assumption that land parcellation does exist in the state (by definition) it is hypothesized that: "Land Fragmentation is proceeding at a constant rate in all areas of the State and for all cate­ gories ." This hypothesis implies that the difference between the mean parcellation amounts for the three periods of study are not statistically different. The rejec­ tion of this hypothesis leads to hypothesis 2 : 2. The Incidence and Spatial Trend Hypothesis If land fragmentation is significant and changing, then the logical questions are: (a) Where is the process occuring most significantly? (b) Why the specific spatial distribution or in­ cidence? (c) At what rate is the process occuring spatially and over time? (d) How can the rates be explained and what are the implications? Answers to these questions call for relating parcella­ tion to several explanatory factors. relationships, Based on possible it is hypothesized that: "Parcellation is associated with ecological, socio-economic and institutional factors; high level and rapid rate of land parcellation are associated with high level and rapid rates of socio-economic developments." 76 The test of this hypothesis is expected to demonstrate regional and district differences. The next question is if land parcellation is in­ creasing and is perceived as a p ro blem to the state, what ought to be done? This provides the basis for the third hypothesis which concerns public action al­ legedly contributing to the spread of the process: 3. The Impacts of the Subdivision Control Act 1967) Hypothesis (PA 2 8 8 , It is hypothesized that: "The Subdivision Control Act (PA 288, 1967) has had no impact on the creation of parcels and sub­ divisions in excess of the 10 acre mi n i m u m lot provision and no evidence exists to support the cl aim that parcels of land 1 0 + acres are on the increase in the state because of the 10 acre m i n i ­ m u m specification." This last hypothesis is the core of land parcellation problem in the state. It is argued that the minimum lot size of 10 acres specified in the Subdivision Control Act has stimulated the partitioning of land tracts into relatively larger lots, usually too large for residential purposes but too small for viable rural industry. By using trend regression with the Subdivision Control Act as a dummy variable, attempt is made to estimate the impacts of the Act on large lots. 77 Assumptions Underlying Hypotheses The three hypotheses are predicated on several assump­ tions, 1. important among which are: Biophysical (ecological) factors such as climate, topography, soils and living organisms w hich are recognized to influence the amount of parcellation that occurs in any area directly or indirectly are assumed constant over the period of study and p r o ­ jection. Human factors remain the major m ot i v a ­ tional force behind the process. 2. Psychological and cultural factors are also r e c og ­ nized to have considerable impacts on most land transactions. These are, however, assumed con­ stant and homogenous over time and over counties. 3. Some institutional factors, especially political and legal factors affecting land transactions are also considered exogenous to the model. However, certain policy measures are regarded as autonomous shock variables (parameters) which generate d i s ­ crete changes, e.g. revisions in subdivision acts, changes in property taxes, etc. These variables tend to be stable over longer time periods than the other endogeneous variables such as per capita incomes, number of housing units, population and so on. To project figures to the year 2000 A.D., the following assumptions are further made: 78 (a) There would be no major catastrophes such as war, earthquakes, global climatic changes, etc. (b) The land area of Michigan would remain fixed at current level of 56,817 square miles or 36,362,880 acres. Land area cannot increase significantly. (c) County boundaries remain fixed at current delineations. (d) Michigan population growth rates will continue at the current trend over the projection period; the population will continue to in­ crease at a very low and declining rate of about 1.0 percent or less. Zero population growth (ZPG) will not be attained by the state during the period of s t u d y .1 (e) Current state-of-arts remains constant. Michigan population increased by 13.5 percent during 19601970 intercensal period; this represented an annual compound growth rate of about 1.275 percent. Between 1970 and 1980 intercensal period, the population grew by only 4.02 percent or at a rate of 0.4 percent per annum. (Variations exist among counties and regions— see Preliminary Population Report - 1980, Chi-Li Wang and Lawrence S. Rosen, Office of the Budget, Dept, of Management and Budget). Projected figures between 1970-2000 indicate that the population of the state will grow at an average yearly growth rate of about 0.6 percent (Michigan Statistical Abstract, 1979). 79 Assumptions Underlying Statistical Tests Specification of valid statistical models require several assumptions about the sample data. To determine the type of statistical analysis and techniques, parametric or n o n - p a r a m e t r i c , that would be appropriate to analyze the collected data, runs, viz: the data were subjected to two test skewness and goodness-of-fit tests. Smoothed frequency polygons revealed a slight skewness to right (positive s k e w n e s s ).^ This observed skewness created a problem as to whether the mean or the median was the a pp ro ­ priate statistic of central tendency. The use of the median w ould rule out some of the most powerful classical st at is ­ tical tests, associated with parametric method of tests. A test of skewness was conducted on the various size unit categories and for three time periods. Results are re­ ported in Table 4-1. The Pearsonian coefficients of skewness ranged between 0.8 to 1.9, ( 0 .8 s study variables. s 1 .9) for all distributions of the * 0 implies perfect symmetrical fre- quency distribution and the mean and median would coincide. ^he 2 distribution graphs are provided in Appendix 4-C with a summary of the various discriptive statistics. Any introductory statistical analysis textbook discusses skewness and other descriptive statistics. For e x a m ­ ple, see: (a) Chou, 1969, ibid., p. 109; (b) Neter, Vasserman and Whitmore, 1966, ibid., p. 63. 2 80 Chou points out that theoretically, S^p varies within the limits of ±3, but in practice, the limits of ±1. values seldom exceed However, he further argues that for most social-behavioral non-laboratory research, Skp lying within the limits of ±1 for all practical purposes reflect mild assymetry and the mean can be considered as a good approxi­ mation of the population parameter.* It should, however, T ABLE 4-1 Pearsonian Coefficient of Skewness for Sample Acreage Distribution by Year and by Size Unit Size Unit 1963 1970 1977 11- Acre Parcels 1.5 1.2 1.0 10+ Acre Parcels 1.9 1.2 1.3 10 Acre Parcels • H 00 1.2 1.4 10- Acre Parcels 1.2 1.1 0.8 Large Parcels (10-10.9) 1.9 1.2 1.4 be recognized that positively skewed distributions are most common and reflects multiplicative forces operating on the variable. For most of the data collected on parcellation, both the mean and median values are reported. most of the statistical models, Also, for 1977 terminal data are used to reflect the most current relationships and situations of * Y a - L u n - C h o u , (ibid.), pp. 108-109. 81 land parcellation in the state. An examination of Table 4-1 and figures in Appendix 4-C reveal that most of the category distributions tend to be only moderately skewed, ranging between 0.8 to 1.4. The test for skewness provided only a partial and not entirely conclusive support for the use of parametric test in this study. An interview with Dr. D. E. Ch ap p e l l e 1 on the issue of normality assumption, and also with Mr. Essama Nssah, o revealed that the skewness test above is not a sufficient condition to establish normality or non-normality. distributions reveal skewed frequency distribution, may be normal; the converse is also true. Some but A test for good- ness-of-fit wa s suggested and run at cx * 0.05 and a = 0.01. Distribution was fairly normal at 0.05, but not at 0.01. Since all tests in the study are run at a = 0.05 the nor­ mality assumption is accepted. There are three sets of distribution for each p a r c e l ­ lation size unit— 1963, test 1970 and 1977. A relative variance [ C V (X)] revealed that the variances (or standard ^Dr. Chappelle is a professor in the Department of Resource Development, Michigan State University, and a member of the author's Guidance and D is se rtation Committee. The author acknowledges his invaluable assistance to this chapter. 2 Mr. Essama Nssah is a Ph.D. candidate in Theoretical E co no metrics at the University of Michigan, Ann Arbor, Michigan. 82 deviations) of the period distributions were not s i gn i fi ­ cantly different at a ■ 0.05. The coefficients of variation for 1963, 1970 and 1977 distributions of large-lot parcellation are 89.4 percent, 77.2 percent and 86 percent, respectively, and these were not statistically different. However, some serial co r r e l a ­ tion between period data sets is suspected since the amount of parcellation that occurs in a period is likely to impact the amount of parcellation that can occur in another period. For example, one of the respondents to the letters sent out for second stage survey, aptly pointed out: "A major portion of the land area (in Oakland County) has already been divided and sold in 10, 11, 12 or 15 acre parcels. It is no longerin process, for the most part is has happened." Mr. Nierman is suggesting that, categories saturation. (10, as far as those size 11, 12 or 15) are concerned, O a kl an d is near However, clearly, 15, 12 or 11 acre parcels can again be subdivided into 10 and still smaller acre lots, if they are not totally developed. In this regard, lot size units less than 10 acres are correlated wit h large size units, especially over time. Nevertheless, where parcels are subdivided for r es i d e n ­ tial purposes, partitioning occurs once and forever. No ^Nierman, Wayne. Oakland County Extension Director. sponse to letters. (Appendix 3-B). Re­ 83 consolidation or repartitioning would be possible unless the property developed on it is destroyed, or unless the whole parcel was not developed. In that case the portion of the parcel remaining idle can be resold or resubdivided. A simple chi-square test was run between parcellation density and large lot parcellation based on the hypothesis that large lot parcellation is independent of parcellation density (i.e. no ceiling effect). Counties were classified into two groups by density and amount of large lots as high and low and a 2-by-2 contingency table test at a * 0.05 r e ­ jected the hypothesis of no dependence. Critical region of chi-statistic of ld.f. at 5 percent confidence level was 3.84 S X 2 s 2 w and computed chi-squared (X ) was 11.230. How­ ever, dependent relation between large lot parcellation and parcellation density need not necessarily imply ceiling ef ­ fect. If high density is associated with low amounts of large parcels the ceiling effect exists; but if high density is associated with high level large lot parcellation, no ceiling effect can be assumed, though relationship still exists. Based on the skewness and goodness-of-fit tests, the assumption of normality is maintained. Variance homo- genity is also assumed based on relative variance test and chi-squared test of dependence. The sample mean is consid­ ered as the appropriate least square estimator of the p o p ­ ulation parameter. 84 Hypotheses Testing Based on the assumption that parcellation is the major mode of land transference in the state, but its nature, ex­ tent, scope and trends are posing problems to the state, by threatening certain "critical" land uses, the hypothesis about trends over time is te st e d .1 To test trends over the 15-year period of study, pa r ­ cellation means are paired and tested as: a) H 1963 = H 1970 and H 1963 < H 1970 (1* 1 b) H 1970 * H 1977 and H 1970 < H 1977 (1~ 2 H 1963 “ H 1977 and H 1963 < H 1977 ( 1~ 3 and C) All tests are done for one-tail at a = 0.05. Pairwise tests of equality among mean parcellation are based on Several independent studies have shown that parcellation exists in several localities of the state. Barlowe and Hostetler did a study covering six counties in the southwest corner of the SLP. Their result was positive. Mentius researched into land parcellation in Venice Township (1978) and the result was positive. The Mich­ igan Public Opinion Survey conducted by Kimball and others revealed that parcellation causing agricultural land withdrawals into nonfarm uses is a problem in the state. Research conducted in Washtenaw County on sub­ divisions turned out positive results— parcellation and subdivision were occuring and idle lots and inefficient use of land (for its highest and best use) were a grow­ ing problem. County Extension Directors of 13 counties reported increasing or mature process of parcellation (Appendix 3-B). Farmland Retention studies by Allen K. Montgomery Jr., Master's Technical Paper (1980), conducted in Canton Township, Wayne County, revealed advanced form of land parcellation in the area due to urban sprawl. 85 t-test analysis of the difference between me a n s . * If the difference between the period means are statistically sig­ nificant, and the null hypotheses are rejected, it is con­ cluded that parcellation process is increasing over time. The second hypothesis requires tests for relationships. The county, as the most common organizational subdivision, was chosen as the unit of interest. This approach permits a cross-sectional comparative analysis. Correlation and multiple regression techniques were used to study the relationships between parcellation (mea­ sured in acreage units) as the dependent variable and the selected socio-economic and biophysical determinants, dis­ cussed in Chapter Two. Correlation analysis measures the degree of association that exists between two independent variables; regression analysis quantifies the parameters of such an association, and provides estimates of the value of dependent variable from known values of one or more in­ dependent variables. These methods also permit statistical inference and testing hypotheses concerning population parameters. To test the significance of the relationships analysis of variance (AOV) subprogram is used. Hypothesis about the period means and variances are tested to establish any sig­ nificant variations among district and regional observations. ^ a - L u n Chou, Statistical Analysis with Business and Economic A p p l i c a t i o n . 1969, pp. 385-429. 86 Partial Correlation Metrix The partial correlation matrix (SUBPROGRAM PARCOR) re ­ veals the strengths and directions of the various relation­ ships among the explanatory variables and between the de­ pendent and explanatory variables. MATRIX are filled with Pearson The cells of the PARCOR 'r' coefficients w h i c h , b e ­ sides establishing relationships, strengths and directions, also help identify spurious correlations, or masked variables. and confounding Such variables are then discarded b e ­ fore the multiple regression model is specified, of variables showing strong, direct in terms and clear relationships. Multiple Regression Analysis The rationale for the use of multiple regression subprogram was to develop predictive models of land p a r ­ cellation. Such models will allow trend comparisons among counties and regions. Least square (LS) multiple regression was run using data for the 30 selected c o u n t i e s .1 This technique p e r ­ mit te d analysis of land parcellation for the whole state. ^The LS Multiple Regression was adopted as a final technique of analysis after several mu ltiple regression techniques have been tried, e.g. L STEP where the computer is asked to delete or include certain variables. The L STEP helped in rejecting and retaining certain variables. 87 Differences in demand for land parcels exist between north­ ern and southern Michigan. Separate equations were, there­ fore, developed for the two broad regions split along Oceana to Bay County line (see Figure 5-3, page 162) and designated as Southern Michigan (SM) and Northern Michigan (NM).1 The parcellation index consisted of five subcategories; multiple regression equations were specified for each sub­ category for the three time periods of 1963, 1970 and 1977. For each category two separate equations were developed based on the broad regional aggregative data for the 1977 period only. This method permitted the isolation and d i s ­ cussion of the different influences on parcellation re le ­ vant to the two broad regions. The model for any period of study tests only the re­ lationships between the variables at that particular point in time based on the hypothesis that: "Parcellation (extent or level) is associated with and is explained or caused by socio-economic, p h y s ­ ical determinants and certain other factors." • The General Model * The General Statistical Model ship between parcellation amount is a functional r elation­ (PAS) and those dete rm in ­ ants identified in the threefold framework of Chapter Three, 1 c . f . Figure 2-3, page 21, for subregion classification. 88 viz.: Biophysical (P), Economic Social (S). factors". (E), Institutional (I), The model is closed by adding Z for "other The General Parcellation Function (GPF) may be formulated in this form: Parcellation * F (Biophysical, Economic, Institutional, Social, Personal, Other f actors)................ (2-1) Symbolically, function (1) may be written as PAS - F (P, E, I, S, Q, Z ) ..........................(2-2) where Z, other factors, takes care of all those other forces which may not be accounted for by the specified elements of the model. PAS, P, E, I, S, Q and Z are all vectors which can be broken down into several components. The model can be sim­ plified by the following transformations. PAS ■ Y Let: ... (A vector of Parcellation) ... (A vector of Biophysical Factor) P ae X^ E = I = x 3 . . . (A vector of Institutional Factor) S = X, ... 4 Q - X- . . . (A vector of Personal Factor) o Z - xg ... (A vector of Other Things) X2 ’** (A vector of Economic Factor) (A vector of Social Factors) Then, Equation (2-2) becomes Y - G (X , X 2 , X 3 , X 4 , X 5 , X0 )..................... (2-3) and is the general "closed" parcellation model. The for­ mulation is "closed" because it includes any other possible 89 factors in Z or X g . To make the function operational, additivity is assumed along with all other classical assumptions underlying a multiple regression model.* The general regression equation based on the general functional equation in (2-3) may be set up as: where 'b' (0 ...6 ) are the multiple regression coefficients, (bQ - constant), and eg is the stochastic element; Y and X g are already explained. Each vector can then be broken into its component parts so that the expanded version of equation (2-4) may be presented as: + e 1 6 (2-5) Many basic statistical texts discuss the assumptions underlying the classical linear multiple regression; exten­ sive discussion on assumptions underlying multiple regression procedure can be found in: (a) Kmenta, J. Elements of E c o no me tr ic s, 1971, The MacMillan, N . Y . , Chapter 10; (b) Steel, R.G. and Torrie, J . H . , Princi­ ples and Procedures of Statistics: A Biometric A p p r o a c h , 2nd Ed., 1980, McGraw-Hill Inc., Chapter 14; (c) Ackoff, R.L. and others, Scientific Method: Optimizing Applied Research D e c i s i o n s , 1962, also discusses the topic on pages 329-341. The assumption of linear multiple re­ gression is made for expediency. Methods are avail­ able for non-linear regression analysis. 90 where the subscripts i..n refer to the individual component, the n's need not be equal and 0, 1, 2, 3 . . 6 are the major functional vectors in equation (2-4). Thus X 2i is a factor or variable i in vector X 2 or an economic factor. Each of the Yi's is isolated and run on all the d epen­ dent variables. The list below indicates that there are five different types of dependent variables, Y - y i, y 2 . . . Y 6 ...............................(2-6) where: Y1 *2 a AAPAS ...(Acreage All Counted Parcels, m ...(Acreage Small Parcels, ASPAS 11- Acres) 10- Acres) AMP AS Y3 - ...(Acreage Mean Parcels, Y 4 - ALPAS ...(Acreage Large-Lot Parcels, 10+ Acres) LAGPAS YS * ...(Acreage Large Parcels, 10 Acres) 10-10.9 Acres) Each of these is run on the explanatory variables and for each year. The final operative equation is: Y i " bo + b iiX ii + b 2iX 2i + b 3iX 3i + b 4iX 4i + b 5 U5 + e 6 *** (2-7) where X^ ■ Biophysical Factor Land Area per County in square miles (LACO). X^g * Biophysical Factor Public Recreational and Forest­ lands as percent of county land area (1974), (PUBREC). 91 X 13 * Biophysical Factor, Air Quality, (1974), mea­ sured by micrograms of particulates per cubic meter (AQUA). X 21 “ Institutional-Economic Factor, Property Value Assessment, for tax purposes, in current million dollars (PAV). X 22 * Economic Factor, Average per Capita Income, of a county in current dollars (PHPI). X 3^ «* Social Factor, total population, per county in number of persons (TOTPOP). X 32 “ Social Factor, Standardized Percent net migration for 1950-1960, 1960-1970 and 1970-1980 (SPNM*). X g 3 = Percent urban population (PURB). Xg^ * Number of housing units (NOHU). Equation (2-7) can now be put in the parcellation acronym form a s :^ ♦The percent Net Uigration (NU) figures were standardized to remove negative signs. The formula used is SPNM = 1 ± NM/100, depending upon the sign. For example, if the percent N M of a county is 32 percent, the stan­ dardized value, SPNM » 1 + 32/100 * 1 . 3 2 ; if the p e r ­ cent NM is -10 percent, then SPNM « 1 - 10/100 * 0.90, and so o n . 1 A11 subsequent equations use variable labels (acronyms) from the computer programs. See Appendix 4-D for complete list and definitions. 92 PAS - bo + .................................... (In te r c e p t ) b ^ L A C 0 ^ + b ^ 2 ]?tJBREC^2 + b^g AQ UA ^4+ .....(P-Factors) b24P A ^ 2 4 + ............................... < I-Factors) b ^ g P H P I g g * ............................ (E-Factor) b 4 ?T O T P O P + b 4 8 SPNM+. ............( S - F a c t o r ) b 4 9 P U R B 4 9 + b 4 1 0 N O H U 4l£ ......... (S-Factor) b 5 llU 5 ii+ ...................... (U-Factor) e g ............................ (Stochastic) . . . (2 - 8 ) All variables and subscripts are defined in Appendix 4-D. Selection of independent variables in the absence of previous state-wide research w a s based on a priori assump­ tions about the probable factors contributing to both the demand and supply of lots for urban uses. This meant a considerable background information on the real estate market as discussed in Chapter Three. Quasi-Experimental Design Model In order to isolate the impact of the Subdivision C o n ­ trol Act for detailed analysis, a trend m odel is also used. The time model is specified as: P A S 1963' S C A 1967* P A S 1970* P A S 1 9 7 7 ................(3-1) w here P A S ^ 9 g 3 1970 and 1977 is mean P a rcellation for the study periods and S C A ^ 9 g^ is the Subdivision Control Act of 93 1967 entered as the shock variable. The limited number of observations impose a constraint on the usefulness and pre­ cision of the method. Campbell and Stanley discuss the shortcomings of this formulation .1 This model 1 is mainly used to provide indications about trends in large lot parcsllation over time relative to the land use regulation. Method of Estimating State Acreages From Sample The previous section of this chapter provided a rational for using the sample mean as the least square estimator of the population mean. However, it was realized that a weighted mean would reflect closely the true popu­ lation given the distribution of some of the socio-economic determinants influencing the studied variable. Population and land area were selected as appropriate weights. amount of parcellation gory of parcellation) lation and land area. The (total acres of land under any cate­ is directly related to total popu­ The ratios of the mean sample popu­ lation and land area to mean state population and land area are used as weights on the sample mean amount of parcella­ tion. The mean amount of parcellation for the State is obtained by the formula: ^For detailed discussion of time series experimental design, see Campbell, D.T. and Stanley, J.C. Experimental and <^uasi-Experimental Designs for R e s e a r c h , 1966, pp. 37- 94 ) PAS (4-1) st where the subscripts m - Michigan s * Sample of 30 Counties t “ Time Period, and PAS * mean amount of parcellation in acres POP * mean population A land * mean area in acres, and the (bar) over these variables imply "mean" or "average". Since land area remains constant and is a known parameter mean population for any period ofstudy, the term in parenthesis in equation (4-1) may be considered as a para­ meter symbolized as . Equation (4-1) is, therefore, re­ duced to: (■■ 1A- ^ i ) 100 (5-6) Implicit in the derivation process are certain ass um p­ tions and hypotheses. Assumptions are made about the u nd er ­ lying distribution of the sample or population. Earlier sections of this chapter examined the statistical assumptions "^Weighted Parcellation Densities are standardized by 100 to remove decimal fractions. 100 in connection with the derivation of state parcellation estimates and the underlying distribution. The main hypotheses implicit in the density index estimator are: a) Parcellation pressure on a county is directly related to the population density of the area; b) It is directly related to the total acreage under parcellation; c) It is inversely related to the land area of the county. The higher the population density and the amount of parcellation, the higher the pressure or degree of p a r c e l ­ lation; however, the larger the land area, the less the pressure. as pointed out in the early part of this section, Thus, 1,000 acres of parcellation in a county such as Iron need not exert the same pressure on the land as 1,000 acres in Wayne or Macomb county even though the extent of parcellation in both counties may be the same. F r o m equation (2) it is clear that if and P ^ i * P A , then W P D A « 1. = D, A * That is, a county with p o p u ­ lation density equals to the mean state density, the land area equals the mean per county land area of the state and the amount of parcellation equals the sample mean amount of parcellation, will have wei gh te d density exactly equals a unity or 100. WPD « 1 or SWPD * 100 is generally assumed to be the delineation between high pressure and low pressure 101 densities. A unity or 100 may also be obtained by other combinations of the factors involved in the computation of WPD. Parcellation density index is a non-negative score where WPD £ 0; it cannot be less than zero. It can theo­ retically take any value between zero and infinity. TABLE 4-4 The $ Value for 1963, 1970 and 1977 v For Calculating Parcellation Density ^ Values * A/D*P^ Parameters 1963 Mean Sample Area (Acres) Mean Sample P o p u ­ lation Density Average Sample _ P a r c e l l . , (PA ) ♦t 704 (x 64 0) 82 1970 7 04(x640) 94 5,065 6,586 1.0848 0.7278 1977 704(x640) 107 11,495 0.3663 Population density data were obtained from 1960, 1970 and 1980 censuses. Table 4-4 reports values for the co­ efficient parameter for parcellation density ( ) 1970 and 1977. for 1963, The value of P^ are based on the mean amount of parcellation for each study time period. 102 The land area is given in acres instead of square miles to increase the 4> values for computational purposes. (The use of square miles gave a $ 1963.) value of 0.00165 for The effect of the 640 acre multiplier is cancelled out eventually since county land areas (as denominator) are also multiplied by 640 . 1 *The estimate for each county is: D. x 640 x P...+ WPD, — t B t X 640 A it x P At and the 640 acre factor is eliminated. CHAPTER FIVE RESEARCH FINDINGS Non-Platted Parcels Holdings of Non-Platted Land Parcels Number of holdings and acres of non-platted land parcels for each of the 30 selected counties and for the three time periods of 1963, 1970 and 1977 are provided in Appendix 5-E for the size units, 11-, 10-, 10, 10+ and 10-11 acre parcels. A total of about 28,316 parcels, ranging from as small as 0.1 of an acre to 10.9 acres were counted for the 30 counties in 1963. Table 5-1 summarizes the county data into totals by categories. 103 104 TABLE 5-1 Number of Holdings by Size Unit and By Period of Study for 30 Selected Counties in Michigan Holdings by Period 1963 1970 1977 29,183 49,871 5,275 6,645 13,317 855 1,190 2,204 6,130 7,835 15,521 28,316 37,067 65,392 10- Ac. Parcels (ASPAP)1 22,186 10 Ac. Parcels (AMPAS) 10+ Ac. Parcels (ALPAS) 10 and 10+ Ac. Parcels (L A G P A S ) 11- Ac. Parcels (AAPAS) Number of holdings increased to 37,067 parcels in 1970 and then to 65,392 parcels in 1977. Thus, between 1963 and 1970, about 8,751 new non-platted parcels were created in the study counties. The total new parcels represented an increase of 30.9 percent over the 1963 base year figure. An additional 28,325 new non-platted parcels less than 11 acres were created during the second 7-year p eriod between 1970 and 1977. This increment represented about 76.4 percent of the 1970 figure or slightly over 100 percent of the 1963 base year figure. Thus, more than twice the number of new parcels created between 1963-1970 was created between 1970-1977 period. ^Meanings of Acronyms were provided in Appendix 4-D. 105 Acreages of Non-Platted Land Parcels The amount of land that is being parcelled into small lots is best measured in acreages. therefore, converted into acres. Individual holdings are, Citizens are more inter­ ested in knowing the total acres of land that are affected by land parcellation rather than the number of holdings under the parcellation. The rest of this study analyzes parcellation by using an acreage index. holdings and acreages, County by county data on both by size unit and period are all provided in Appendix 5-E, Tables E-l to E-8.* There were about 151,952 acres of land under 11- acre p arcellation in the 30 counties in 1963. However, amount of parcellation varied greatly among the counties studied, ranging from a m ax im um of 19,344 acres in Berrien county to a m i n i m u m of 1,178 acres in Huron County that year (Appendix 5-E). Table 5-2 reports the total amount of parcels counted for each period of study by parcellation categories. ^Conversion technique was pp. 17-19. explained in Chapter Two, 106 TABLE 5-2 Amount of Parcellation by Size Units and By Periods of Study for 30 Counties in Michigan Parcels in Acres by Period Size Unit 1963 1970 1977 ASPAS (10-) 90,237 118,622 189,143 AMPAS (10) 53,110 66,450 133,170 8,605 12,496 23,146 61,715 78,946 155,716 151,952 197,568 344,859 ALPAS (10+) LAGPAS (10-11) AAPAS (11-) The amount of parcellation increased from 151,952 level of 1963 to 197,568 acres in 1970 and then to 344,859 acres in 1977 for the 30 sample counties. Trends in Amount of Parcellation An examination of Table 5-2 reveals that signifi­ cant amounts of new non-platted parcels were created during the 1970-1977 period of study. The hypothesis that there was no significant differ­ ence between the mean parcellation amounts (for all parcellation categories) of 1963 and 1970 could not be rejected on the basis of available evidence at the 5 per­ cent level of significance. However, the hypothesis was rejected for the means of 1970 and 1977. The computed 107 t-values of all categories for the difference between 1970 and 1977 were all greater than the critical value of 2.00 (Table 5-3). Table 5-3 reports the mean amounts of parcellation by size unit for each study period, the standard deviations (SD) and the t-statistics of the two interval periods ^(1970-1965) and t2 (1977-1970). TABLE 5-3 t-Test, for Difference Between Means of Parcellation Amounts for Time Period 1963-1970 and 1970-1977 Parcellation Category 1963 Mean 1970 SD Mean 1977 SD Mean SD t l *2 AAPAS (11-) 5,065 4,324 6,585 4,696 11,495 7,729 1.28 2.92* ASPAS (10-) 3,008 2,641 3,954 2,848 6,305 4,372 1.31 2.43* AMPAS (10) 1,770 1,472 2,215 1,581 4,439 3,490 1.10 3.13* ALPAS (10+) LAGPAS (10-11) 772 770 1.03 2.04* 2,057 1,841 2,632 2,034 5,211 4,009 1.10 3.09* .287 422 417 530 The increase in parcellation between 1970 and 1977 for 10 acre parcels (AMPAS) and 10-10.9) acre parcels (LAGPAS) had larger t-values, (t*3.13 and 3.09 respectively, Table 5-3) than the other size categories. This indicates that the impacts of the Subdivision Control Act on parcel­ lation trends between 1963-1970 was negligible owing to the short time lag between its execution and the period *a _< 0.05; Critical Value ■ + 2.00; N ■ 30; df » 58. actual projected 2,500,000 , ACRES OF LAND 2 000,000 o (A o» 1,500,000 CD CO CM J^ !!!SlLIiorS^liLl9~Acr< Parcels 1,000,000 (Urge Porcclsi I0.I-I0.9 IQ*J^creParceh 500,000 1975 I960 YEAR 1985 1990 1995 FIGURE 5-1 TRENDS IN NONAPPROVED PARCELS, 1963, 1970, 1977 AND PROJECTED TO 2000 BY SAMPLE 2000 109 parcellation data were collected. division Control Act la g g e d The effect of the Sub­ but after 1970 the Act began to effectively change parcellation trends in the state; partic­ ularly, the 10 and 10.9 acre parcels were more affected than the others. Figure 5-1, showing trends in all categories of parcel­ lation is based on the quasi-experimental design model dis­ cussed in Chapter Four. The Subdivision Control Act, introduced in 1967 and effective in 1968, is observed to have affected all categories of parcellations between 1970 and 1977. This is reflected in the abrupt change in the slope of the trend graphs. A simple linear extrapolation projection is adopted to project trends to the year 2000. The test of hypothesis of no trend support the graphi­ cal depiction of trends in Figure 5-1. The Subdivision Control Act had induced effect on demand and supply of parcels. It c ou ld,however, be argued that in general parcellation was increasing during the period of study but the increase accelerated as a result of the policy variable introduced in 1967, thus explaining variations in parcellation over the period of study. Table 5-4 reports the numerical and percentage changes in the amount of parcellation, which occurred during the 14 year period and among parcellation categories. Total all categories increased by 45,616 acres or 30 percent between 110 1963 and 1970. Between 1970 and 1977, about 147,291 acres of new parcels or 74.5 percent of the total amount of par­ cels in 1970 were added. TABLE 5-4 Increases in Parcellation Acreages by Size Unit for 30 Counties in Michigan Size Unit 1963- 70 Numerical P e r ­ cent Acres 1970- 77 Per­ Numerical Acres cent 1963- 77 Numerical Per­ Acres cent ASPAS (10-) 28,385 31.5 70,521 59.5 98,906 109.6 AMPAS (10) 11,340 21.4 66,720 100.4 80,060 150.7 3,891 45.2 10,650 85.2 14,541 168.9 LAGPAS (10-11) 17,231 27.9 76,770 97.2 94,001 152.3 AASPAS (11- ) 45,616 30.0 147,291 74.5 192,907 127.0 ALPAS (10+) New parcellation acreages more than doubled during this second period. Such a significant increase in the amount of new non-platted parcels synchronized with the period of the Subdivision Control Act of 1967. Thus, the rate of increment in parcellation rose significantly higher after the Act, (Figure 5-1). Trend analysis indicated that the SCA of 1967 explains about 62 percent or 18,000 acres of total 11- parcels and about 80 percent of the 10-11 acre parcels between 1970-77. The remaining 38 percent and 20 percent may be attributed to normal increasing demand Ill associated with several socio-economic factors such as increased income, desire for larger secluded homesites with rural environment, increasing demand for small part- time farms (sometimes as hobbies) and non-platted sub­ divisions, especially in the northern Lower Peninsula in anticipation of future boom in recreational and tourist activities. Between 1963-1977 a total of 192,907 new u n ­ p la tted parcels less than 11 acres were created in the 30 counties. This represented an average yearly p ar c e l ­ lation of 13,850 acres or 462 acres per county per year. Trends in Parcellation Categories Table 5-4 reports trends in categories. "Small” parcels increased by about 28,385 acres or 31.5 percent between 1963 and 1970. A relatively moderate increase (compared with other size units) of 59.5 percent acres) occurred during the second 7-year period. (70,521 A total of 98,906 acres of land were withdrawn into "s ma ll ” parcels of under 10 acres in the 30 counties. This represented an average yearly small parcel developments of 7,065 acres or 235 acres per county per year. Large parcels between 10 and 11 acres increase by 17,231 acres or 27.9 percent during the first interval period. Spectacular increases occurred in this size unit during the second interval period. Acreage figures rose by 76,770 acres or 97.2 percent over the 1970 totals. Thus 112 "large" parcels more than tripled in amounts. parcels component of this size unit cent The 10 acre increased by 21.4 p e r ­ in 1963-70 period and 100.4 percent during 1970-77 period, and the 10+ acre component increased by 45.2 p e r ­ cent during 1963-70 period and almost doubled to 85.2 p e r ­ cent. These remarkable gains in large parcels, are partly associated with the Subdivision Control m in imum acreage provisions. Between 1963-77, a total of 94,001 acres of new unplatted large parcels (10 and 10+ acres ranges) were developed in the 30 counties. an annual This figure represented large parcellation of 6,714 acres or 244 acres per county per year, just as much as small parcels, despite the fact that the latter category contained over 80 percent of all parcellation holdings. Based on data presented in Table 5-4, and trend analysis of Figure 5-1, as well as the results of t-tests of no d i f ­ ference in mean parcellation, nificant it is concluded that no sig­ increase in parcellation (all size units) occurred between 1963 and 1970, but increases were highly signifi­ cant for the period between 1970 and 1977 after the e f fe c­ tive date of the Subdivision Control A c t . Of major interest to the study is the relative proportions of total parcellation acreages and relative trends in the proportions for various size units. Table 5-5 reports the proportionate shares of the total amounts of parcels for each period by parcel categories. 113 In 1963, parcels under 10 acres constituted about 59 percent of total acreage of all parcels under 11 acres. It Increased its share very slightly (by about 0.6 percent point) in 1970 and its shares declined to 54.8 percent. Projected estimates (page 108 of this chapter) indicates that by the year 2000, "small'' parcels would constitute about 50 percent of total acres of land under 11- acre parcellation; this size unit would make up about 75 per­ cent of all holdings. TABLE 5-5 Percentage Shares of Parcellation Categories of Total Amount of Parcels Per Period Share Percent of Period Total Size Unit 1963 1970 1977 ASPAS (10-) 59.4% 60% 54. 8% AMPAS (10) 35.0 33.6 38.6 5.7 6.3 6.7 40.6 40.0 45.2 ALPAS (10+) LAGPAS (10-11) Large parcels of 10 acres or more (but less than 11 acres) constituted a little over 40 percent of the total acreages in 1963 and 1970; but its share increased to a little over 45 percent in 1977. Projection indicates an increasing trend to about 50 percent by year 2000. "large" parcel category, Of the 10 acre units made up 35 percent in 1963, dropped slightly in 1970, and again increased its 114 share to about 38 percent. gained steadily Since 10+ acre parcels also (in share) during the study period, any gains in proportionate share observed in "large" lot parcels must have been at the expense of small parcels. This implies that between 1970 and 1977, the rate of large parcel developments exceeded the rate of small parcellation. TABLE 5-6 Percent Share Distribution of Total Increments in Parcellation Acreages Between "Small" and "Large" Parcels Percent Shares Size Unit 1963-70 1970-77 1963-77 Small Parcels (10-) 67. 9% 47.9X 51. 3X Large Parcels (10-11) 32.IX 52. IX r• 00 Of probable significant interest is the amount of parcels each category of parcellation contributed to total increments during the period of study. Table 5-6 reports size unit shares to incremental changes in amount of pa r c e l ­ lation. Small parcels accounted for about 68 percent of the total increment 1963-1970. 1977 period. in the amount of parcels created between This share dropped to 48 percent during 1970On the average, during the 14-year period of study from 1963 to 1977, small parcels contributed about 51.3 percent to the total of all new parcels measured in acres. Large parcels increased in share from 32 percent 115 in the first 7-year period to 52.1 percent during the second period, thus outstripping small parcels in its contribution to the total. On the average, about 49 pe r ­ cent of the total acres of new unplatted parcels created between 1963 to 1977 was made up of 10-11 acre units. This is clearly reflected in the t-value of the two parcellation categories. The t-scores for "small" parcels (ASPAS), b e­ tween 1970 and 1977 was 2.43 and that of "large" parcels (LAGPAS) was 3.09, all significant at less than a 5 p e r ­ cent level of probability. Summary on Time Trend Analysis of Parcellation 1. 1970-1977 was a period of significantly higher parcellation activity in Michigan. The period coincided with the Subdivision Control Act of 1967 (about 3 years-lag). Time series analysis indicate that the Act contributed to the spectacular increase in the amount of parcellation which occurred after 1970 and t-tests of near differences revealed that increases in parcellation (all categories) were not statistically signifi­ cant for the interval period between 1963 and 1970, but were highly significant for the period between 1970 and 1977. The Subdivision Control Act explained about 40 percent of the total variation which occurred in the amount of parcellation over the 14 year period. The rest was due largely to increasing demand for larger homesites, associated with increasing a f f l u e n c e , demand for environmental resources and quality of life and demand for second and third recreational homes. After 1970, 10 and 10+ acre parcels rose in importance as taker of land. The acceleration can be associated with the impact of the minimum lot provisions in the Subdivision Control Act of 1967. The proportion of large parcels (10-11) rose faster than that of small parcels and by 1977 the two broad categories of small lot parcellation, (ASPAS and LAGPAS) were accounting almost 50-50 to the total acreage under parcel­ lation, despite the facts that LAGPAS consti­ tuted only 30 percent of all holdings. Between 1970 and 1977 alone, about 150,000 acres of new parcels under 11 acres were created in the 30 counties and over half of these were 10-11 acres. There is a tendency for this category to increase in relative importance. 117 5. These findings confirm the hypotheses that (1) large parcels are increasing in amount and in holdings in the state and (2) that the Sub­ division Control Act has contributed to the upward trend. Estimated Acreages of Parcellation in Michigan In Chapter Four estimation procedure was explained and weighted mean parcellation acreages were derived. Statis­ tical tests about the means indicated that the weighted mean parcellation figures obtained fell within the signifi­ cant confident intervals at a - 0.05. The weighted mean parcellation figures for acreage values were 8,104 acres in 1963, 9,877 in 1970 and 16,093 acres in 1977. Based on these mean values, state total acreage figures for each time period are obtained and pro­ jected to year 2000. Table 5-7 reports the resultant est imates. TABLE 5-7 Estimated Total Nonapproved Parcels in Michigan by Period of Study (1963, 1970, 1977) Estimates of Acreage and Holdings 1963 1970 1977 Number of Holdings 125,330 153,882 253,,316 Acreages of Parcels 672,632 819,791 1,170,,348 118 Figures in Table 5-7 indicate that corresponding estimates of holdings and acreages in 1963 for the whole state were 125,330 and 672,632, respectively. About 28,552 holdings or 147,159 acres of land under 11 acres were developed between 1963 and 1970. Between 1970 and 1977, an additional 99,434 holdings or 515,928 acres were affected by 11- acre parcellation. total new About 50 percent of the nonplatted parcels was made up of large parcels in the size range of 10-11 acre units in 1977. Table 5-8 reports the straight line projection estimates for acreages from 1980 to 2000, based on the estimated figures for 1963, 1970 and 1977. TABLE 5-8 Quinquannial Projection Estimates of Amount of Parcellation Under 11 Acres in Michigan (1980-2000) Amount of 11- Acre Parcels Year Holdings In Acres 1980 268,928 1,335,719 1985 314,637 1,407,864 1990 360,346 1,643,981 1995 406,056 1,880,798 2000 451,765 2,117,615 A straight line projection will result in total holdings and acreages of 451,765 and 2,117,615 in another 20 years, 119 an addition of around 200,000 holdings and about 800,000 acres. It must be emphasized that the parcellation categories analyzed here, constitute only a part of the total parcellation which is occurring in the state. Parcels of 11 acres and more and approved subdivisions which require platting by law of the state are excluded in these estimates. Projected estimates for each size unit is provided in Appendix 5-F for 1980 to 2000. The average rate of increment during the 20 years would be about 2.6 percent per annum. A total of about 781,896 acres of land would be added to the 1977 figures during the 23 years of projection. This implies that about 34,000 acres of land would be affected annually. Of the annual increment in parcellation amount, about 16,000 acres or almost 48-50 percent would be accounted for by parcels in the range of 10-11 acres. Spatial Distribution of Land Parcellation in Michigan Concerned citizens are not only interested in knowing the amount of parcellation that is occurring in the state, or about the level and trends over time but also in where parcellation is taking place and why. It is important also to know how much land is being swallowed up by large lot parcellation and where the specific category of parcellation is occurring. In this section, an attempt is 120 made to show where parcellation is occurring in the state, generally, and where large-lot parcellation specifically is most severe. Also trends in the pattern of distribution are examined. To make acreage values comparable among counties and regions, counties are grouped into three categories accord­ ing to their weighted parcellation density scores.^" Counties in each density classification are arranged by density rank­ ing, and their corresponding acreage figures for the two parcellation categories being analyzed are also provided. Acreage figures cover 1977 counts. The acreage data of 1963 and 1970 are used mainly for spatial trend analysis. The main objective of this section is to determine current distribution of parcellation, over the state and trends in such distribution over time (trend in spatial distribution). Table 5-9 provides a summary of the 11- acre weighted parcellation density scores for 1963, 1970, p rojected values of 2000 AD. 1977 and for Corresponding acreage figures for 1977 are also provided county by county. Nineteen sixty- three is considered the base year and the Weighted Density scores are calculated over 1963*s density coefficient. In this way, the scores are standardized like any index number and can be compared over counties and over time. Thus, 1963 1 In Chapter Four, the derivation of the W.P.D. scores was e xplained and discussed and tables of the scores are provided in Appendix 5-G for each time period; 1963 is the base year. 121 TABLE 5-9 1963, Weighted Parcellation Density Index for 1970, 1977 and Projected Estimate for Year 2000 County Acreage of 11-Parcels 1977 1963 1970 1977 2000 Year 1 2 3 4 5 Allegan Alpena Antrim Bay Barrien 16,314 6,194 8,913 11,568 24,499 1.47 .30 .13 5.47 14.24 1.91 .44 .26 6.50 16.73 3.20 1.03 1.05 11.48 20.62 4.98 2.10 2.35 19.57 27.41 6 7 8 9 10 Calhoun Cheboygan Claire Clinton Crawford 13,659 9,005 8,214 12,770 11,593 4.36 .14 .10 .80 .15 5.88 .28 .46 2.06 .19 6.34 .60 1.00 3.21 .58 9. 30 1.24 .70 7.57 1.01 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 8,543 3,813 7,418 6,610 4,935 .20 .07 .38 .47 .07 .25 .07 .88 .70 .13 .22 .10 3.13 1.27 .30 .67 .17 7.05 2.39 .64 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 4,754 5,422 3,482 27,501 31,785 .10 .21 .05 .45 1.98 .15 .43 .07 .96 3.73 .42 .86 .06 7.33 16.09 .91 1.63 .07 17.24 35.88 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 3,021 17,792 10,295 4,969 17,658 .02 32. 57 .42 .06 5.60 .03 59.39 .70 .13 8.33 26 27 28 29 30 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 11,029 20,610 22,995 7,598 1,900 .45 .54 5.57 1.18 .02 .70 .79 8.65 1.55 .02 Sample Average 11,495 1 .00^ 1.46* .09 .05 88.75 148.07 1.29 2.53 .48 .20 19.75 12.59 1.72 1.65 11.86 2.76 .02 2 .89 + 3.43 2.98 34.14 4.60 .03 5. 3 2 * ♦The W.P.D. scores for the whole sample for each time period were derived by using mean parcellation amounts and the formula in Chapter Four. 122 density score is 1.00 or 100 percent. All other indexes are compared with the base year score. County scores for 1963 ranged from as low as 0.02 (Schoolcraft) to as high as 32.5 (Macomb). Most of the low density scores are reported for Upper Peninsula counties. In 1970, density scores ranged from 0.02 (Schoolcraft) to a maximum of 59.4 (Macomb). Compared with 1963 base year, all counties showed increases in parcellation density. Using the 1977 scores, counties are grouped into high, moderate and low degree of parcellation. Table 5-10 sum­ marizes the density classification of counties. The first group of counties are generally found in the southern half of the Lower Peninsula. They have very high to high Weighted Parcellation Densities and they are designated in this report as High Weighted Parcellation Density (HWPD) group. Their WPD scores are above 10.0. Group 2 counties include almost all counties from the southern LP except those in group 1. Both group 1 and group 2 contain all the sample counties south of the Oceana-Bay line of counties. They are characterized by moderate parcellation density scores. 1.0 to 10.0. WPD ranges between They are designated as MWPD (Table 5-10). 123 TABLE 5-10 Description of Each of the Groups of Counties Categorized by WPD County Group Number Group-1 Description Very High to high Counties that have esti­ Density WPD > 10.0 mated parcellation density (HWPD) of 10.0 and over. Some have very high density over 50.0 Group-2 Moderate Density Counties with estimated WPD 1.0 < WPD < 10.0 ranging between 1.0 to 10.0 (MWPD) (excl.) The threshold figure is WPD = 1 . 0 Group-3 for 1963 Very low to low Counties with estimated WPD Density; less than or equal to 1.0. WPD < 1.0 This group includes counties with WPD as low as 0.02. Parcellation pressure on land in such counties is insignificant. The third group of counties constitute the majority of the 30 counties studied; they are described as very low to low WPD areas and are designated as LWPD. All the UP counties and most of the NLP counties are included. 124 Parcellation density scores were calculated from 1977 county acreage figures. Areas of high amounts of parcellation need not necessarily coincide with areas of high parcellation density and vice versa, because of the population and land area factors in the estimation of the density scores. TABLE 5-11 Acres of Land Under Parcels Less Than 11 Acres and Large Parcels (10-109) with Corresponding Parcellation Density for Each Group of Counties Based on 1977 Data Total Group Density County Group 1-High Density N-30 6 WPD Score 168.4 Amount of Parcellation 11- Acre 10-10.9 Acre Parcels Parcels 126,297 50,316 2-Moderate Density 10 31.70 133,805 61,926 3-Low Density 14 6.21 84,757 43,474 Total 30 206.31 344,859 155,716 Table 5-11 reports the total acres of land in the two size categories for each group of counties with their c orresponding aggregate parcellation densities. Figures in the Table 5-11 reveal interesting facts about the nature of the distribution of land parcellation among the study counties and in the state generally. Six out of the 30 counties fell within the group category of 125 high parcellation density, 10 in the category of moderate parcellation density and the remaining 14 were designated as low parcellation density. The skewed distribution of density scores over counties closely reflect the population distribution of the state. The implications of the above figures are revealing. Despite the extremely skewed distribution of density scores, the amount of parcellation is fairly spread among the three groups. The first group contained about 36.6 p e r ­ cent of the 11- acre parcels and 32 percent of the largelot parcels. The second group contained larger proportions of the two categories of parcellation 40 percent respectively). (39 percent and Even the low density group together contained about 25 percent and 28 percent of the total acreages of the two categories. This implies that the high density areas are not necessarily those with large tracts of land in parcellation as defined by this study. A corollary implication is that areas with high parcellation density scores are characterized by high level parcellation and intensive land use so that only a few large parcels can be added each year; they are gradually approaching saturation point. Conversely, areas with low parcellation density tend to have fewer population and/or large land areas so that larger parcels could be created. 126 Six counties are classified as High Density areas, most of these counties are found In the southern Lower Peninsula where urbanization Is advanced and land use Is highly Intensive. For most urban counties, there would be few raw lands left for large lot parcellation; competi­ tion among various uses would limit extensive withdrawal. The NLP and UP still remain relatively moderate to low pressure regions. This is apparent from tables which report the parcellation density values by county and by group. TABLE 5-12 Density and Acreage by Group and C o u n t y , 1977 Group-1 HWPD Counties in Group-1 HWPD 1 2 3 4 5 6 Macomb Berrien Ottawa Livingston Bay Monroe Total 1980 Density 1977 Parcellat ion Acreaee 11- Acre Population 11- Acre 10-10.9 Acre Parcels Persons Per Parcels Parcels WPD Square Mile 1447.3 295.1 277.8 175.0 268.0 240.0 88.8 20.6 18.8 16.1 11.5 12.7 17,792 24,499 22,995 31,785 11,568 17,658 5,697 10,858 7,775 17,667 2,484 5,835 M ___ 126,297 (136.6%) 50,316 (29.8%) Table 5-12 reveals that, even though, Macomb recorded the highest WPD among all the counties studied, it did not account for the greatest proportion of acreage totals for 127 both categories. Livingston, which ranked fourth in density contained the largest amount of large-lot parcels as well as of total acreage of all categories studied in 1977. Parcellation densities in counties such as Monroe, Bay, Ottawa and Livingston are relatively lower than that of Macomb and possibly other continguous counties such as Wayne and Oakland, despite the fact that all these counties may be classified as High Parcellation Density. This sug­ gests that relative land use pressure in the former group-1 counties must be lower than that of the latter; a continued parcellation activities in the former counties and a slow­ ing down to negligible large-lot parcellation activities in the latter group-1 counties should be expected. Large-lot parcellation in counties such as Wayne, Genesee, Washtenaw, Oakland and Ingham, which may be included in group-1 density classification, may not be practicable in the future if it is not so now. These counties as the main hub of intense » land use with Livingston and Lapeer forming the center of current large-lot parcellation activities. However, most of the very high to high density counties are moderate to low large-lot parcellation areas. These findings confirm the hypothesis that: Large-lot parcellation is more likely to occur in areas of relatively low socio-economic activitiesrural and fringe counties. Conversely, counties • 128 with intense land use and socio-economic activities are likely to experience least large-lot p ar c e l l a ­ tion. 1 A simple correlation analysis of the various parcellation categories and selected independent variables fails to 2 reject this hypothesis. All the six counties classified as High Parcellation Density are found in the Southern Lower Peninsula (SLP); four of them in the eastern section and the remaining two in the western section. In general, the south eastern section of the state is under greater parcellation pressure than the west, and the whole South together is much more pressured by parcellation activities than the North. Table 5-13 reports figures for group-2 counties. There were 10 counties in this group described as moderate parcellation density. Most of the counties in the group are again found in the Southern Lower Peninsula. *A later section of this Chapter attempts at providing statistical test results of the relationships between parcellation categories and selected independent variables. 2 See page 139 for correlation analysis. 129 TABLE 5-13 Parcellation Density and Acreages Group-2 MWPD Counties in Group-1 HWPD 7 8 9 10 11 12 13 14 15 16 Lapeer Calhoun Clinton Allegan Grand Traverse St. Joseph M ontcalm Newaygo Hillsdale Manistee 1980 Density Population 11- Acre Parcels Persons Per WPD Square Mile 1977 Parcellation Acreage 11- Acre 10-10.9 Acre Parcels Parcels 106.0 198.1 94.1 98.1 118.1 7.3 6.4 3.2 3.2 3.1 27,501 13,659 12,770 16,314 7,498 12,763 4,354 4,354 7,794 3,577 111.1 7.1 41.2 70.1 42.0 2.7 1.7 1.6 1.3 1.3 7,598 11,029 20,610 6,610 10,295 3,230 4,506 13,039 3,000 5,337 *._ 133,805 (38.8%) Total 65,926 (42.3%) Among the 10 counties in this group, Lapeer recorded the highest Parcellation Density score (7.3) and the largest amount of 11- acre parcellation. lot parcellation. It was second in large- Newaygo, Hillsdale and Manistee recorded the least Parcellation Densities in the group (1.6-1.3). Newaygo, however, recorded the highest amount of large-lot parcellation (13,039) and ranked second after Lapeer in 11- acre parcellation. This negative relation between p a r c e l ­ lation density and amounts is due to the fact that p o p u l a ­ tion density in these counties is low compared to a county such as Calhoun, which ranked second on the density scale with a total parcellation amount of 13,659 acres. It should 130 further be recognized that counties such as Newaygo and Hillsdale are outliers with several lakes that are an attraction for second home developments. Demand for parcels in these outlying counties is predominantly recreationally motivated rather than employment related. The 10 counties contained about 39 percent and 40 percent of the amounts of parcellation in the two parcellation categories reported. The following conclusion can, therefore, be drawn. The extent of parcellation, both large-lot and 11acre parcels, is relatively higher in the Moderate Parcel­ lation Density areas, especially for those counties in the Southern Lower Peninsula than either the High Parcellation Density areas or other Moderate Parcellation Density areas. Most of the counties in the group are described as 'fringe1 in most population studies of the state. The moderate density region includes southern counties such as St. Joseph, Branch and Hillsdale and other counties of the Southern Lower Peninsula between the east and west High Parcellation Density areas. The moderate region then branches east and westwards to include counties such as Arenac, Tuscola and Sanilac to the east and (Newaygo, Oceana, Mason)1 and Grand Traverse with all other adjacent counties. the group-2 counties are, therefore, Most of found west of the ^These counties are not necessarily contiguous. They be­ long to Muskegon— Norton Shores— Oceana S.M.S.A. Area. 131 Michigan Meridian, south of the Oceana-Bay county line. With the exception of some counties of the Thumb, most counties of the Southern Lower Peninsula may be described as high to moderate parcellation density and high to moderate amount and total parcellation area. It may, therefore, be concluded that large-lot parcellation is still taking place in the south, specifically, at a moderate rate in counties of Moderate Parcellation Density and at relatively low rate in the High Parcellation Density counties. Projected esti­ mates of the parcellation density acres indicate that the western portion of the state, from the southern boundaries to the northern tip of the Lower Peninsula would be most affected by large lot parcellation. In general, the whole Lower Peninsula would be affected by very high to high large-lot parcellation if the process continues at current rate. Fourteen counties fell within the group-3 Low Parcel­ lation Density category. these counties. Table 5-14 reports figures for Five counties in the group were scored almost zero parcellation density. This may imply that, either population densities are low, or amounts of parcel­ lation small or both. 132 TABLE 5-14 Parcellation Density and Acreages Group-3 LWPD 1980 Density 1977 Parcellation Acreage Population 11- Acre 11- Acre 10-10.9 Acre Persons Per Parcels Parcels Parcels WPD Square Mile Counties in Group-2 LWPD 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Alpena Antrim Cheboygan Clare Iosco Crawford Delta Houghton Huron Gogebic Iron Mackinac Menominee Schoolcraft Total 57.0 34.0 29.0 42.0 52.1 17.2 33.0 37.0 44.0 18.8 12.5 10.7 24.6 7.0 1.0 1.0 0.6 1.0 0.9 0.6 0.2 0.3 0.4 0.1 0.1 0.1 0.2 0.0 6,194 8,913 9,005 8,214 5,422 11,593 8,543 4,935 4,754 3,813 3,482 3,021 4,969 1,900 2,394 5,942 5,459 6,169 3,124 2,468 2,468 2,554 1,343 1,323 1,362 1,480 1,571 805 — __w 84,757 (24.6%) 43,474 (27.9%) An examination of both the parcellation density and acreage figures reveal that all the five counties that scored near zero parcellation density are characterized by both low population densities and small amounts of parcel­ lation. In 1980, the mean population density of Michigan was 162. The maximum density among the bottom group-3 counties was 24.6 reported by Menominee. Also, the mean amount of parcellation (11-) for 1977 was 11,495 and the maximum amount obtained in the group is 4,969, Menominee. again by All five counties are in the Upper Peninsula. 133 Alpena topped the list of the group by density ranking though Crawford contained larger amount of pa rcellaction. The group contained about 28 percent of total large-lot parcel in the sample, and 25 percent of all parcels. In general, the whole Upper Peninsula Region is characterized by low parcellation-low density. The most important con­ trolling factors are ecological and remoteness. All the 7 counties from districts 5 and 6 in the extreme north of the Northern Lower Peninsula are also found in the third density group. This implies that the extreme north of the Lower Peninsula and the Upper Peninsula are still relatively least pressured by parcellation. If other conditions are favorable, more increasing large-lot parcellation in these areas should be expected than any other part of the state in the future. It should be noted that, counties such as Crawford, Clare, Cheboygan and Antrim have quite a bit of 11- acre parcels and Antrim, Cheboygan and Clare have al ­ ready experienced some degree of large-lot parcellation; this may indicate that other conditions in that locality favor future large-lot parcellation. This will become apparent in later sections when spatial variation in parcel­ lation is analyzed using multiple regression. Summary on Generalized Parcellation Distribution in Michigan Distribution of amount of parcellation extent in the state in 1977 suggests a division of the state into three 134 regions, the southern half, below Oceana-Bay county line, the northern half of the Lower Peninsula above the line and the UP. Generally, parcellation is far advanced in the south, decreasing northwards in amount and spatial coverage. The UP is, as yet, unaffected significantly by any type of parcellation. The parcellation density analysis confirms the following findings: 1. Small parcel fragmentation (partitioning of tracts of land into parcels less than 11 acres) exist 2. in Michigan and is on the increase. Parcels in the size category 10 to 11- acres are increasing in number and in amount faster than those in the range less than 10 acres. 3. The proportions of 10- acre parcels have been declining over the years, while that of 10 to 11have been increasing; projections reveal that the two categories are converging to a 50-50 ratio. 4. Despite the scant time series data, it is ap­ parent that relationship exists between the SCA and large-lot parcellations as well as all cate­ gory parcellation; the SCA has obviously contrib­ uted to the accelerated creation of larger parcels and observation and information from other sources indicate that the creation of some of these largelot parcels were speculative. In some NLP counties, such parcels are becoming more and more tax delinquent. (Responses from County Extension Di re c t o r s . ) Even though the proportion of 10- parcels in on the decline (acreage measure), total amount of 11- acre parcels is still increasing and pulling all categories along. It is expected that some 2.5 million acres would be in this kind of parcel­ lation in 2000, this would constitute about 7 percent of the state land. Current estimates of farm land conversion indicate that over 50 percent of this amount of parcellation would be carried on farm lands. This implies that in year 2000, this particular parcellation would have contributed a cumulative total acreage of 1.2 million acres to agricultural land with­ drawal; the remaining 1.3 million would have come from other land uses. Parcellation is most advanced in the southeastern Lower Peninsula than any other region, followed by the western LP. The NLP is only moderately affected and the process is still insignificant in the UP except isolated areas. Parcellation density is very high in the ESLP than any other area in the state. Its 136 distribution follows closely the population distribution of the state. Parcellation density declines northwards. 8. "Large" lot parcellation is more extensive in areas of low parcellation densities, especially 9. the western half of the LP. The UP is unique in the following sense: a. parcellation density is very b. parcellation amount and level are also low low but and c. the rate of parcellation is also low d. there is little indication to suggest that the region w ould significantly be affected by any kind of parcellation on an extensive scale. Parcellation activities that would occur in the region would be localized. 10. Large-lot parcellation activities would not be practicable any more in the ESLP. Most of the parcellation that would occur in that area would involve re-partitioning of lot sizes in the range of 10-20 acres or more into still smaller lots. 11. The western SLP still has room for large lot parcellation however, total amount of parcellation would increase only slowly and a great proportion of the new parcellation activities would involve re-subdivisions of relatively larger lots. 137 12. The NLP would be the region to watch closely. Overall parcellation would increase at relatively fast rates and more and more large lots could be created if other conditions become favorable. Statistical Analysis of Variations in Parcellation Several factors were selected a priori and discussed in Chapters Three and Four as contributing to the spatial variations in parcellation. In Chapter Four, an attempt was made to provide a general statistical formulation showing the relationships between the dependent variable (parcellation) and the independent variables. Both correlation and multiple regression analyses con­ firm that parcellation of land is closely related to socio­ economic and bio-physical factors as well as institutional or policy parameters. Simple correlation matrix was used to identify highly intercorrelated variables prior to the multiple regression analysis. It also provided a measure of the degree of association between several' pairs of key v ar ia bl e s. Three sets of correlation matrices were obtained; one set showed relationships between pairs of dependent variables (not reported), a second set showed correlation between dependent and independent variables (Table 5-15) and the third set reported relationships between pairs of 138 TABLE 5-13 Correlation Coefficient Matrix of the Dependent Variables and Independent Variables for the years 1963, 1970, and 1977; df - 28. Middle and Last values are the coefficients for 1970 and 1977 respectively. Dependent Variables Independent Variable AAPAS (11-) ASPAS (10-) LAGPAS (10-10.9) AMPAS (10) ALPAS (10+) PAV .591* .778* .383* .643* .827* .521* .466.638* .195 .414* .587* .138 .587.699* .390* TOTPOP .591* .526* .393* .641* .571* .535* .468.416* .199 .419* .363* .140 .584* .514* .405* MOHU .636* .543* .345 .679* .589* .491* .519* .429* .156 .467* .374* .096 .633* .533* .373* PHPI .643* .658* .506* .675* .694* .694* ,542.548* .232 .499.484* .178 .625* .658* .402- LACO -.286 -.353 -.423* -.287 -.345 -.406* -.260 -.331 -.378- -.253 -.329 -.381* PURB -.254 .158 -.051 .327 .260 .193 .128 .002 -.294 .007 -.035 -.350 SPNM -.481* .330 .337* .489* .296 .095 .415* .346 .621* .377* .353* .062* .495* .275 .230 PUBREC -.415* -.493* -.518* -.408* -.510* -.616* -.389* -.424* -.330 -.365* -.376* -.279 -.426* -.302* -.453* .544* .537* .448* .551* .560* .539* .487* .454* .281 .455* .407* .228 .539* .530* .431* AQUA * i .05 r* • .36 -.255 -.291 -.240 .219 .109 .053 139 Independent variables (Table 5-16). Critical 'r* values were set at .36 and .70. Even though all categories of parcellation were highly correlated (critial *r' for pairs was greater than .90 in most cases), all categories were used in regression equations. Further, all categories were significantly related to most of the independent variables for the three study periods. Weakest relationships were noted for only 1963. Correlation Analysis Almost all the size units of the dependent variables were significantly correlated with most of the socio-economic factors such as TOTPOP, NOHU, PHPI and PAV, especially for 1970 and 1977 (Table 5-13), and at 5 percent level of significance. (Significant 'r* between dependent and in­ dependent variable is .36, while 'r 1 for inter-correlation between independent variables is .70.*) correlation coefficients, Table 5-15 includes for the three study periods and all parcellation categories with the 9 selected independent variables. *To determind the critical used: ’r', the following formula was t .05,df * r/^( 1-r2 )/(n-2 ) . For a full discussion of the above formula, see R.G.D. Steel and J.H. Torrie, Principles and Procedures of Sta ti st ic s: A Biometric A p p r o a c h ; 2nd ed. McGraw-Hill Book Co. 1980- 140 On broad regional basis, however, considerable vari­ ations exist among the South and North with regard to significant correlations, (Table 5-23, page 164). Where as socio-economic factors such as percent net migration (SPNM), PHPI and PAV continued to show strong association with some of the parcellation categories in the South, in the North, physical factors such as LACO and PUBREC in­ creased in importance in addition to the socio-economic factors. In general, however, all the factors selected were significantly related to the study variables for each study period most of the time. The strength of relation­ ship either increased or weakened with time and sometimes, the sign changed with time. For example, the correlation between NOHU (Number of Housing Units) and AAPAS (Parcels Less than 11 Acres) declined from r * .636 in 1963 to r ■ .345 in 1977, whereas, the correlation between PHPI (Per Head Personal Income) and ASPAS (Parcels Less than 10 Acres) hardly showed any change (r - .675 in 1963 to r * .649 in 1977). Thus while some factors gained in importance in their relationships with the dependent variables others lost their importance over time. Such variability in relationship is also observed with respect to spatial distribution of the parcellation phenomenon. These relationships have no connotations for causality. Increasing or decreasing relationship does 141 not correspond to Increasing or decreasing contribution to total variations in the dependent variable. The correlation coefficients reflect trends in association over time. Of interest are the signs of association. Almost all factors selected were positively correlated to all the categories of parcellation. The only explanatory variables that had negative correlation coefficients consistently over time, and with all parcellation cate­ gories were LACO and PUBREC. This was expected. Counties with Large Land Area (especially in the UP) are generally associated with few parcellation amounts; also, most of the counties with public recreational lands are associated with few amounts of parcellation of all categories. Land Area, however, had insignificant relationship with most of the parcellation categories and for most of the time PURB is not significantly correlated with any of the parcellation categories. In general, PURB had been positively correlated with all categories of parcellation (however insignificant) during the early periods. Highly urbanized counties had been associated with high amounts of parcellation. However, during the later periods of study, the sign changed to negative for some parcellation categories (albeit insignificant amount). For example, the correlation between PURB and AAPAS, LAGPAS and 142 AMPAS, changed from positive to negative Implying that, with time, counties with highly urbanized population b e ­ came associated with fewer parcellation amounts. This phenomenon has been described in Chapter Four as the ceiling effect. As county population becomes more and more urbanized, there is considerable amount of parcel­ lation initially, but with intensive land use parcels become small and the process of parcellation declines, since only a few new parcels can be created without ex­ cessive costs. Counties as Oakland, Wayne, Genesee and Kent exhibit such ceiling effects. Parcellation in highly developed areas may have peaked and reached a limit where only re-subdividing of existing parcels may be feasible. It is therefore expected that some amount of smaller-lot parcellation activities are still going on in the very High Parcel­ lation Density areas as the correlation coefficients s ug ge st . Table 5-16 reports the correlation coefficients for pairs of all the selected independent variables. Coe f­ ficients of correlation between PHPI and PAV, TOTPOP, and NOHU exceeded .70, the chosen critical correlation. 'r' value of inter­ It was generally noted that the relationship between pairs of socio-economic factors tended to be stronger than among pairs of physical factors, or between pairs of 143 TABLE 5-16 Correlation Coefficient Matrix for the Eight Independent Variables for the Years 1963, 1970 and 1977; df ■ 29. Middle and Last Values are the Coefficients for 1970 and 1977 Respectively. Starred Items are Significant at 5% Confidence Level or Less 1. PAV 2. 70TP0P 3. MOHU 4. PHPI 5. LAC0 6. PURB 7. SPN1I 8. PUBREC *5 0.05 r* ■ .70 fc — a a 4 8c « a s £ 6 U a a a a e. a z c. Partial R“ Corr. Coefs. Deletes Beta ■eights Part iul K2 Corr. Coeffs. Deletes .158 .126 .290 .143 .091 .232 .487 .493 .469 .362 .106 .123 .299 .110 .134 .572 .605 .603 -1.345 .320 -.214 .231 .473 .469 .541 .125 .387 .087 .541 .607 -.396 -.160 -.374 2.357 -.097 -.099 -.297 .307 .492 .492 .449 .445 -3.223 .332 -.409 3.568 -.160 .311 -.380 .175 .600 .568 .544 .598 Partlal Beta ■eights Corr. Coeffs. R2 Deletes .162 -.068 .358 .706 .712 .671 -1.631 -.065 -.263 -.053 .692 .713 6.965 .481 -.271 -4.970 .417 .410 -.264 -.336 .653 .656 .692 .677 .157 -.063. .323 .5000 .6103 .7136 .066 .101 .001 151 LAGPAS were explained by the variations in the selected predictors. This implies that about 40 and 50 percent of the variations around the means in the two dependent vari­ ables were left unexplained. The remaining percentage in variations may be attributable to other factors (specified in the general model as U) such as cultural, personal, etc. which could not easily be quantified. However, over time the selected factors increased their predictive power and in 1977, over 80 percent and 70 percent of the variations in ASPAS and LAGPAS respectively were explainable in terms of the variations in the selected variables. The level of significance also improved in all cases. Physical factors such as air quality (index of Environmental Amenities) and public lands (index of Recre­ ational Amenities) played only minor roles in predicting variation. However, as their predictive power declined with time, relative to ASPAS (small parcels) their impor­ tance increased with time with regards to LAGPAS. example, For the beta weight of AQUA in the ASPAS equations declined from .308 in 1963 to .216 in 1977, while for LAGPAS, it increased from .240 in 1963 to .323 in 1977. Thus, even though the impacts of ecological factors on statewide land parcellation is still weak, these factors contributed positively to land parcellation and in the case of largelot parcels, their impacts were increasing. Buyers of parcels were increasingly taking into consideration 152 environmental amenity factors such as air quality, water quality, forest resources, recreational facilities and the like. Economic factors, in general, have played an important role in land parcellation. However, for both ASPAS and LAGPAS the predictive power of PHPI declined over time whereas the contribution of PAV increased correspondingly, especially in the case of LAGPAS. The variations in LAGPAS was closely associated with assessed land values. In general, the higher the assessed value of land, the fewer the large-lot parcels created. In 1970, the relationship between PAV and LAGPAS was positive (Table 5-19), but was not as important as that of 1963 and 1977. beta weight suggest a reversed impact. The 1970 positive Parcellation must have increased in areas of low assessed valuation as indi­ cated by the sign of the beta weight in 1963. increasing parcellation, increased so that But with land values must have gradually in 1970, one finds that parcellation was increasing in areas with increasing land values. less, Neverthe­ in 1977, PAV contribution to R 2 was still too small to be operationally important. However, after 1970, the trend reversed again. Social factors such as population, and number of hous­ ing units still remained the strongest predictors of vari­ ations in both ASPAS and LAGPAS, increased with time. and their importance 153 It, however, appears that migration and percent urban population have remained weak predictors. But the signifi­ cance of these two factors to the study lies in their signs. PURB has consistently maintained negative coefficients. This confirms the hypothesis that highly urbanized areas would generally exhibit declining parcellation process since parcellation would have reached advanced stage and ceiling effect would affect future parcellation. impact of ceiling effect In fact, the is borne out by the fact that the beta weight becomes smaller and smaller with time. The contribution of the predictor declines over time. In 1963, SPNM had negative contribution and, was a stronger contributor than TOTPOP. predicting ASPAS, LAGPAS. in fact, It ranked third in though it was not very important for Its sign changed to positive in 1970 for both ASPAS and LAGPAS and, improved over the 1963 values. This implies that while in 1963, areas that had large influx of migrants experienced less amounts of parcellation than those with few influx, the trend reversed in 1970; areas with high influx of migrants experienced large amounts of pa rc e l ­ lation. This clearly is explainable in terms of state migration trends. Between 1960 and 1965, migration was towards urban centers where new parcellation was very limited. Between 1965 and 1975, movement of population was towards contiguous 154 counties and this was associated with increasing parcel­ lation since most of these counties had not already experi­ enced heavy parcellation. The positive sign for 1970 changed to negative again by 1977 in the case of ASPAS as small lot parcellation in contiguous counties declined, and again, since most such counties (such as Oakland, Washtenaw and Macomb) began to reach parcellation satura­ tion. However, for LAGPAS, the sign remained positive be ­ cause such parcels were created in rural areas (e.g. NLP) which were experiencing in-migration from urban areas of the SLP. Analysis of AMPAS (10 Acre Parcels) and ALPAS (10+ Acre Parcels) A comparison of the beta weights in Tables 5-20 and 5-21 for AMPAS and ALPAS with those in Table 5-19, page 150, for LAGPAS reveal a very close similarities in the signs and magnitudes of the coefficients. Factors which contrib­ uted most to the variations in R 2 for LAGPAS remained the most important predictive variables for AMPAS and ALPAS. Thus, PAV, TOTPOP and NOHU remained the most important predictive variables for the three periods of study. This implies that the analysis for LAGPAS is applicable to that of AMPAS and ALPAS. further discussion. They, therefore, do not require TABLE 5-20 (LS) Multiple Regression Statistics fur AMPAS (10) as Dependent Variables and LACO, PUBREC, AQUA, PAC, PHPI, TOTPOP, SPNM, PURB. and NOIIU as In­ dependent Variables for the Years 1963, 1970, and 1977; df = 20 Explanatoryr Variables 1961 1970 * Partial Beta R2 Heights i Corr. Coeffs. Deletes PHYSICAL LACO PUBREC AQUA PAV PHPI uUL 1AL» TOTPOP SPNM PURB NOHU Significance Level 1977 R2 Beta Partial Weights Corr. Coeffs. Deletes •> Beta H“ Partial Weights Corr. Coeffs. Deletes .131 .123 .220 .112 .087 .202 .428 .431 .411 .279 .143 .075 .215 .135 .075 .505 .520 .526 .078 .002 .251 .079 .003 .279 .699 .701 .676 -1.432 .288 -.215 .197 .408 .412 .605 .055 .392 .035 .443 .528 -1.501 -.055 -.238 -.044 .683 .701 -.401 -. 152 -.360 2.403 -.093 -.089 -.272 .296 .430 .430 .390 .380 -1.747 .307 -.410 2.069 -.079 .266 -.344 .093 .523 .492 .465 .524 7.095 .449 -.311 -5.248 .416 .380 -.295 -.346 .639 .151 .673 .661 .43-19 .5283 .7014 .152 .043 .001 TABLE 5-21 (L3) Multiple Regression Statistics for ALPAS (10+) as Dependent Variables and LA00, PUBREC, AQUA, PAV, PHPI, TOTPOP, SPNM, PURB, and NOHU as Independent Variables for the Years 1963, 1970 and 1977; df = 20 Explanatory Variable Beta Heights PHYSICAL LACO PUBREC AQUA ECONOMIC SOCIAL 1970 1963 Partlal R2 Corr. Coeffs. Deletes Beta Heights 1977 Partlul R2 Corr. Coeffs. Deletes Beta Heights Partial R2 Corr. Coeffs. Dele! .233 .121 .279 .238 .104 .306 .601 .620 .585 .555 -.020 .248 .523 -.027 .329 .672 .761 .733 .462 -.388 .554 .370 -.290 .475 .512 .540 .456 PAV PHPI -.876 .393 -.163 .319 .614 .581 .273 .318 -.261 .270 -.744 .743 -1.687 -.087 -.226 -.059 .556 .577 TOTPOP SPNM PURB NOHU -.327 .169 -.375 1.905 -.093 -.120 -.340 .289 .621 .613 .575 .589 7.258 .358 .371 7.520 -.418 .411 -.431 .431 .712 .713 .707 .707 4.108 .470 -.001 -2.091 .218 .341 .000 -.123 .558 .523 .578 .572 B2 Significance Level .6240 .007 .7618 s0.0005 .5787 .old 157 Spatial Variations in Parcellation Trends This section examines variations in parcellation rates and trends by districts and regions. land partitioning are examined, viz: Three modes of (1) 11- acre parcels, (2) 10-10.9 acre parcels (large parcels) and (3) approved subdivisions. Acreage totals are used in preference of holdings. In general, in the state. isolated parcellation has been increasing However, the rate of increase varies from county to county, district to district and region to region. Tables in Appendix 5-H summarize acreage parcellation figures for the districts and regions and for 11- acre and 10-10.9 acre parcels. The trend figures indicate that total parcellation increased in all districts. Sharper increases occurred after 1970, after the Subdivision Control Act (P.A. upward. 288, 1967) and the trend continued monotonically Districts 3, 4, 5 and 6 experienced sharper shocks; D-7 and D-8 in the Upper Peninsula were hardly affected. Large-lot parcellation exhibited similar trends. Table 5-22 reports the average annual rates of large-lot parcellation for the 30 counties, by districts. Again, the second 7-year period experienced the greatest rates but there is a general tendency of declining rates in the future. However, Districts 3, 4, 5 and 6 showed higher TABLE 5-22 Trends in Average Annual Rates of Change in (10-11) Acre Parcellation, Actual and Projected by District District 1963-1970 1970-1977 1963-1977 1980-2000 %(r) %Cr) %(r) %(r) 1 S.E.S.L.P. 4.5 9.8 7.1 3.0 2 S.W.S.L.P. 1.2 4.4 2.8 2.6 3 C.E.S.L.P. 5.9 15.0 10.4 3.5 4 C.W.S.L.P. 3.2 11.0 7.0 3.1 5 E.N.L.P. 5.7 16.0 10.8 3.6 6 W.N.L.P. 7.1 11.9 9.5 3.4 7 E.U.P. 1.3 8.2 4.7 2.5 8 W.U.P. 2.3 8.6 5.6 2.6 3.6 10.2 6.8 3.1 Total ♦Annual Rates (?) is calculated by the formula: log( 1 +r) = log *" - .lcg x° yn from the equation: (l+r)n = 159 rates than all the others. These four districts consti­ tute a contiguous region which lies north of the OttawaUacomb County boundary line. This implies that large-lot parcellation should occur more rapidly in the northern two-thirds of the Lower Peninsula than in the southern one-third or in the U.P. With the exception of SWSLP district in the Lower Peninsula, the U.P. districts recorded the lowest rates of change both actual and projected. Table 5-22 reveals these variations. It appears that be­ tween 1980 and 2000, the mean annual rate of parcellation in the state would be around 3 percent. Variations in Parcellation Among Study Regions Figure 5-2 displays a cuboid locational graph, intented to show trends in regional parcellation totals (1 1 - and 10-10.9). Projected amount of parcellation for 2000 AD is also shown. The map clearly demonstrates both spatial distribution of amount of parcellation and of relative parcellation trends among regions. More parcellation would still occur in Region I by the year 2000, Region II and then Region III. However, followed by in terms of rates of increase, Region 111 clearly would experience the highest rate, followed by Region I. The rates of increase in amount of parcellation in Region II and IV are upwardly steady but relatively low compared with the z ? 6 280,000 240.000 200.000 CO Ul 160,000 160 tr u 120,000 80JD00 £ 40,000 FIGURE 5-2 TRENDS IN NONAPPROVED PARCELS 1963-1977 AND PROJECTED TO 2000 BY STUDY REGION 161 other two. the E.S.L.P. Two regions to be watched by policy makers are and the N.L.P. are, however, different. The underlying causal factors It was hypothesized that the main causal factor in the E.S.L.P. is population concentra­ tion associated with socio-economic activities and hence demand for employment oriented homesites while in the N.L.P. the obvious factors are ecological, recreational homesite demand and land availability at relatively cheap value. Multiple Regression Analysis based on two broad regions of Southern and Northern Michigan, attempted to reveal the regional differences in the impacts of factors that contrib­ uted to parcellation in the state. Broad Regional Analysis of Non-Platted Parcels Previous analysis of the spatial variations in parcel­ lation based on the whole state indicated that Michigan can be divided into two broad regions for comparative study (Figure 5-3, page 162). In general, statewide analysis of variations in the early sections tended to be too generalized and to gloss over regional differences. It was pointed out that two types of demands for parcels existed in the real estate market of small lot parcellation. Job oriented demand for homesites was particularly important in the southern half of the state where most of the socio­ economic activities are occurring; recreational and amenity 162 Northern Michigan (NM) Southern Michigan (SM) FIGURE 5-3 MAP SHOWING BROAD STUDY REGIONS FOR STATISTICAL ANALYSIS 163 related demands for parcels are particularly noted in the northern half of the state where tourism is becoming an important industry. The dichotomy in the demand for small parcels less than 1 1 - acres associated with the predominant activities in the different regions called for the specification of multiple regression equations for each region. Two sets of regression equations for all parcellation categories and for 1977 were developed for the two broad regions-Southern and Northern Michigan. Each equation contained all the selected independent variables. Tables 5-23 and 5-24 report the regression statistics. Regression equations for all categories (except ASPAS, which was excluded) of parcellation for 1977 for the broad regions revealed that in the Southern counties TOTPOP (17.58), NOHU (-14.76) and PAV (-2.73) were the most important predictive factors which contributed the most to explain variations in parcellation (AAPAS) as well as to all other categories. The same factors played major roles in the Northern counties with beta weights of (-5.65, 5.53, and .91) respectively but with corresponding signs reversed. Further, in the North, physical factors contributed signifi­ cantly to the variations more than they did in the South. For example, LACO (-. 6 6 ) and PUBREC (.71) were quite important in the North; in the South their corresponding TABLE 5-23 (LS) Multiple Regression SLatislies for AAPAS (11-), LAGPAS (10-11), AMPAS (10) and ALPAS (10+) as De­ pendent Variables and LACO, PUBREC, AQUA, PAV, PIIPI, TOTPOP, SPNM, PUKB, and NOIIU us Independent Variables tor the Year 1977, for SOUTHERN MICHIGAN AAPAS (11- ) Dependent Variable (1977) LAGPAS (10-11) AMPAS (10) All*AS1 (10+) Explanatory Beta Weights PHYSICAL LACO PUBREC AQUA ECONOMIC PAV PIIPI SOCIAL TOTPOP SPNM PURB NOUU Partial R2 Betu Corr. Deletes Weight Coeffs. Part ial R2 Corr. Deletes Coeffs. lietu Weights Partial R2 Beta Corr. Deletes Weights Coeffs. ■> Partial R“ Corr. Del eli Coe ffs. .059 .685 .245 .078 .608 .315 .791 .671 .770 .091 .693 .289 -.129 .643 .390 .583 .175 .471 .097 .659 .219 .135 .615 .301 .812 .702 .797 .036 .648 .543 .033 .454 .461 .583 .475 .471 -2.730 .833 -.435 .46G .744 .735 -2.348 .659 -.410 .412 .563 .561 -2.316 .609 -.398 .398 .780 .780 -1.788 .520 -.218 .226 .563 .561 17.576 -.317 -.713 -14.757 .707 -.225 -.361 -.683 .586 .782 .762 .611 9.805 .187 -.549 -7.426 .517 .146 -.307 -.454 .578 .565 .582 .582 10.609 .132 -.603 -8.248 .538 .101 -.327 -. 484 .740 .813 .793 .758 2.841 .402 -.116 -1.098 .113 .203 -.044 -.049 .578 .565 .582 .582 R2 .7927 .8232 .8150 .5832 Signifloanee Level .211 .154 .168 .651 TABLE 5-24 (I.S) Multiple Degression Statistics for AAPAS (11-), LAGPAS (10-11), AMPAS (10) und ALPAS (10+) us De­ pendent Variables and LACO, PUOKEC, AQUA, PAV. TOTPOP. SPNM, PUltll, and NOIIU as Independent Variables for the Year 1977, NORTUEHN Ml<111GAN Dependent Variable (1977) AAPAS (11-) Explanatory Variable Beta Weights LAGPAS (10-11) AMPAS (10) ALPAS (10+) 2 2 2 2 Partial R Beta Partial R Beta Partial R Beta Partial R Corr. Deletes Weights Corr. Deletes Weights Corr. Deletes Weights Corr. Deletes Coeffs. CoefIs. Coeffs. Cueffs. PHYSICAL LACO PliBREC AQUA -.661 .714 .166 -.497 .637 .224 .758 .693 .808 -.729 .455 .031 -.788 .730 .088 .883 .905 .955 -.714 .457 -.018 -.739 .685 -.046 .872 .891 .942 -.129 .023 .271 -.107 .-26 .341 .801 .802 .777 ECONOMIC PAV PHPI .907 -.188 .430 -.120 .776 .815 .551 -.605 .506 -.620 .940 .928 .834 -.569 .614 -.545 .907 .917 -1.493 -.237 -.602 -.145 .691 .799 SOCIAL TOTPOP SPNM PURB NOHU -5.673 .948 -.478 5.527 -.654 .617 -.444 .689 .680 .705 .773 .653 -3.630 1.175 -.267 3.390 -.748 .892 -.490 .764 .899 .783 .942 .894 -3.890 1.065 -. 186 3.303 -.726 .843 -.324 .710 .878 .800 .935 .883 1.158 .670 -.459 .695 .168 .471 -.416 .114 .797 .747 .762 .800 Signlf1ounce Level .8175 .9557 .9422 .8029 .164 .007 .013 .191 166 beta weights were LACO (.06) and PUBREC (.69) for AAPAS. Similar differences were noted for other parcellation categories. In general, all the selected variables explained over 80 percent of the total variation around the mean in all parcellation categories in the North (Coefficients of Multiple Determination, R 2 , were: AAPAS (.818), LAGPAS (.956), AMPAS (.942) and ALPAS (.803)). In the South, the variables explained over 70 percent of total variations in AAPAS ( R 2 =.793), LAGPAS (=.823) and AMPAS (=.815) and less than 60 percent of variations in ALPAS both North and South, (R2=.583). In selected variables showed their weakest predictive power in relation to ALPAS (10 acre p a r c e l s ). Reversal of the beta weight signs is noteworthy. The negative beta weight for LACO in the Northern equation (but positive for the South) implies that counties with bigger land areas are still experiencing relatively few parcellations. However, the beta weight for PUBREC is positive which suggests that recreational and natural amenities do positively contribute to variations in pa rc el ­ lation in that region. PUBREC is positive Since correlation between LACO and (.438), it can be inferred that e v e n t u a l ­ ly, large counties with low intensity of land use will eventually attract more parcellation if other factors are 167 favorable in the statewide analysis where LACO had positive beta weights for all categories of parcellation. This assumption was borne out in the South, the two physi­ cal factors had positive beta weights (.059 and .685) but the predictive power of LACO was extremely weak. This may be attributed to the fact that the southern counties are relatively homogeneous with respect to land area more than the northern counties. Similar explanation can be provided for the signs of TOTPOP, and NOHU. The positive beta weight of TOTPOP for the South suggests that increasing population pressure in this region calls for increasing parcellation, especially since demand for parcels here is associated with employment. On the other hand, the negative sign for the North confirms earlier analysis that the demand for parcels in that region is closely linked to second and third homes for recreational purposes so that absentee ownership is common. A closer examination of the regional regression tables reveal further that PHPI had positive and relatively signifi­ cant beta weight (.833) for the Southern equation but negative and very small beta weight for the North (-.188). Thus while incomes seemed to play a role in parcellation in the South (especially since land for parcellation is rela­ tively scarce and land values are high), personal incomes did not significantly influence parcellation in the North 168 failing to reject the hypothesis that recreational demand for homesites in the North is not necessarily income gener­ ated. Both lower and upper income classes are equally involved. Another variable of interest is percent net migration (SPNM). The beta weights of SPNM for AAPAS were -.317 (South) and .948 (North). Migration was therefore, an important contributor to parcellation in the North but was not a significant for large lot parcellation (LAGPAS). Analysis of migration in Chapter Three indicated that the North is experiencing considerable influx of migrants lately while the South is losing population. The negative sign for the South probably reflects the fact that out­ migration releases pressure on land and hence permits additional parcellation. It should be observed that even in the South, positive beta weights were obtained for the large parcel categories. This is because outmigration in the region may be intra-regional so that those southern counties acting as destinations are experiencing large-lot parcellation. Parcellation in counties such as Lapeer and Livingston and Clinton are the result of population spill over from the adjacent cities (Detroit, Flint and Lansing). However, this hypothesis could not be tested by the analysis provided in this study. It should also be pointed out that the use of R 2 as a sole 169 measure of goodness-of-fit presents a problem in that confidence regions may still be large. Nevertheless, regional analysis has conclusively con f irmed t h a t : 1. the demand for parcels less than 11 acres is dichotomous according to the two broad regions in the state; 2. variables operating on parcellation differ in importance regionally; 3. the signs of variables vary with region, and with parcellation categories, 4. time and amenity factors are becoming more and more important in land parcellation. The next section examines, briefly, trends in platted and approved parcels. Approved Subdivisions Trends in Approved Subdivisions Table 5-25 reveals virtually no trend except a slight downtrend over time. Subdivision developments appear to have been irregular over the years, strict pattern. following hardly any However, a critical study of the various components reveal intriguing cyclical pattern of develop­ ment . TABLE 5-25 Subdivision Statistics for 30 Selected Counties and the State of Michigan 1969-1979 Number of Subdivisions Year 226 219 213 294 263 187 151 131 190 176 640 639 624 670 695 756 536 515 579 636 211 Cumulative Total Mean Annual Source: State Sample 6,644 7,349 7,267 7,523 6,686 4,448 3,511 2,890 4,132 4,199 4,971 State Sample State 15,300 21,760 17,000 16,840 19,070 13,115 11,525 10,425 13,118 10,074 10,599 11,035 10,444 8,738 6,108 4,842 4,980 7,910 5,991 7,588 34,600 33,940 24,100 26,229 28,340 19,638 15,850 17,145 21,951 — — 2,261 5,651 59,619 138,153 88,309 221,793 206 628 5,419 15,350 8,028 24,644 (a) Michigan Department of Treasury, Plat Office, Plat Files, Collected May 1980; (b) State of Michigan: Annual Report of the State Treasurer, Research Section, Oct. 1977 to Sept. 1978. Table 33, page 65* 170 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 Sample Number of Subdivisions Acres of Subdivisions FIGURE 5 - 1* TRENDS IN APPROVED SUBDIVISION PLATS WITH FITTED CURVE, SAMPLE AND STATE, (1969-1979) '■j t I NUMBER OF SUBDIVISION PLATS 173 Between 1970 and 1978, a total of 5,651 new subdivisions were platted and approved in the state. This represents an annual mean subdivision development of 628 plats. Corre­ sponding acreage and lot figures were 138,153 acres, and 221,793 lots or mean annual figures of 15,350 acres and 24,644 lots. During that same period, the 30 sample counties created 2,261 plats or 59,619 acres and 88,309 lots correspondingly. These figures constituted about 32.3 percent, 34 percent and 31.9 percent of the totals, respectively. The mean annual figures for the sample were 206 plats, 5,419 acres and 8,028 lots per year. The almost non-apparent declining trend is also evident in the sample figures. Figure 5-4 shows the line graphs portraying trends in subdivisions for both state and sample counties, based on plat figures. The graph also shows the Least Square Linear trend lines. The actual figures hardly reveal any trend; however, the fitted trend lines reveal very slowly declining trend in the number of new subdivisions platted and approved. are negative, The slopes of the two trend lines (about -10.23 for the state and -4.89 for sample). Acreages Under Subdivision Of more importance and greater interest than the sub­ division plats are the acreage figures, since acreage 174 36,000 r 32,000 i- NUMBER OF SUBDIVISION ACREAGES AND LOTS 20,000 r 24,000 - 20,000 h 6,000 12,000 ACRES (state) LOTS (sample) 8,000 4,000 FIGURE 5-5 TRENDS IN SUBDIVISION ACREAGES AND LOTS, SAMPLE, STATE, (1969-1979) 175 measure is more revealing than number of plats (compare with holdings and parcellation acreages). Figure 5-5 portrays trends in both acreage and lots of subdivision developments in the state of Michigan from 1969 to 1979 (Sample and State). Declining trends in acreages are much steeper than the number of plats reflected. simply implies that This individual subdivisions are getting smaller and smaller in area, and are containing fewer and fewer lots. Acres per plat declined from about 25 to 210 and may stabilize around 24-20 acres per plat. spondingly, the number of lots per subdivisions had been declining from over 50 to about tween 35-30. ly. Corre­ 30 and may stabilize b e ­ The downward trend cannot continue indefinite­ Logarithmic projection indicated a stabilization level around 20 acres and 35 lots per subdivision after 1985 or about 0.6 of an acre per lot. Trends in Approved Subdivisions by District and Region Since subdivision data covered a continuous period of 10 years trend equations based on district subdivision acreage figures are provided along with regional trend graphs in Appendix 5-1 and 5-J. Statistical equations for Districts 2, 3, 7 and 8 were not significant at 5 percent and hence w ould not be used for predictive purposes. How­ ever, their coefficients reveal the direction of trend. 176 Three districts, S.E.S.L.P., E.U.P. positive coefficients, and W.U.P. showed implying possible increasing trend. The correlation coefficient of the first district, S.E.S.L.R, was quite high (r»0.8) and significant at 99 percent level. However, the latter two U.P. districts had very low r- c oefficients (r“0.1 and 0.3) and were not significant at a ■ 5 percent. This implies that generally, subdivision developments in the U.P. over time. the rate of has remained constant The rest of the 5 districts had negative c o ­ efficients wit h their r's ranging between -0.5 to -0.8, and significant at over 90 percent. With the exception of W.S.W.L.P. District, all the other districts which showed declining trends in approved subdivision developments, on the contrary, showed highly increasing trends in unplatted parcellation. concluded that: It may be (1) some districts which were experiencing decline in approved subdivisions were, on the other hand, exp er ie nc i ng positive and increasing isolated parcellation. There appears to be a substitution between approved and n on-approved subdivisions. C.W.S.L.P., W.N.L.P. These districts are C.E.S.L.P., and E.N.L.P. They are all in the northern two-thirds of the L.P. (2) The U.P. districts are experiencing virtually no trend in approved subdivisions and very little increasing trend in isolated parcellation. 177 (3) The S.E.S.L.P. experiencing very minimal and S.W.S.L.P. districts are upward and downward trends. They constitute the southern third of the L.P. and parcel­ lation activities are reaching a constant rate. It may be generalized that the most of land fragmentation, currently, important mode is not approved sub­ division but isolated land parcellation and the main area in the state mostly affected is the north of Ottawa-Macomb boundary lines, i.e. mid and northern Lower Peninsula regions. Figures in Appendix 5-J show trend graphs of sub­ division acreage for the four regions. In general, sub­ division declined steeply in Region III (N.L.P.) where isolated parcellation was noted to be increasing at increasing rate. Region II (W.S.L.P.) showed a gradual declining trend while Regions I (ESLP) and IV (UP) again revealed slight increasing trend. However, trends were not very clear owing to the periodic fluctuations. Conclusion The Multiple Regression analysis have clearly shown that the most important determining factors in parcellation are economic and social. Economic factors such as personal income and land values played a major role in explaining variations in parcellation categories. The effects of 178 income, declined towards the terminal period of the study. Social factors such as population, number of housing units and percent net migration remained as highly important factors and their effects increased with time. Physical factors did not appear to contribute significantly to the variations in any of the parcellation categories, overall state-wide analysis. for the However, when the state was divided into two broad regions of North and South, the impacts of physical factors became apparent Further, in the North. it is still premature to judge the impacts of physical factors since they more or less increased in importance as contributors towards the terminal year of study. Such a trend suggests that in the future, physical factors are going to play considerable role in influencing land parcellation in Michigan. Since most of the regressions were significant, it can be concluded that the selected predictors do have con­ siderable influence on land parcellation and they do help explain variations in the spatial distribution of all cate­ gories of parcellation. For the most p a r t , approved subdivisions showed de ­ clining trends for the whole state. vealed, however, Regional analysis re­ that the decline is not universal. The East Southern Peninsula and the UP still exhibited increas­ ing subdivision trends. The declining trend in subdivision 179 noted for the NLP was, nevertheless, associated with in­ creasing non-platted parcellation, especially for parcels in the size range of 10-11 acres. This presupposes some kind of substitution between platted and non-platted p ar c e l ­ lation in the region. The concluding chapter of this study attempts to relate these findings to selected land use policies and then based on these findings, certain policy rec om me nd a­ tions are made. CHAPTER SIX SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary Increasing awareness about both physical and economic scarcity of land resources relative to population growth and multiplying demands and doubts about the effectiveness of the "free market" to optimally allocate and distribute land to meet broad social welfare goals, have generated public concerns over the utilization of its basic resource— land. Gone are the frontier days when Michigan land re­ sources were considered as inexhaustible. Governments are increasingly being requested by the general public to monitor and regulate future uses and distribution of private lands, not only for the benefit of current users, but also for the security and survival of future generations. Land is one of the most important possessions of mankind which Adam Smith's "Invisible Hand" has visibly failed to allocate optimally among competing ruses and to distribute efficiently intra- and inter­ generations . Parcellation (non-platted subdivisions) has been one major unregulated method of land allocation and distribution 180 181 in the state and in many parts of the nation that is based on "free market" operations. Platted subdivisions have remained the other side of the coin where governmental in­ tervention in the "free market" has been channelled to direct and control private land use and development. As the government is called upon more and more to control land use and its allocation, there is going to be a need for more and more information and data about who owns land and for what; about how much land is owned and how land is being partitioned and transferred into other uses and to different users so that appropriate and e f f e c ­ tive land use policies can be initiated and executed. This study has attempted to provide information about how mu ch of Michigan land has undergone small tract parcel lation over the 15-year period and what probable future trends in this method of land distribution are to be ex ­ pected. Attempt has been made to provide explanation for both the spatial and time trends in non-platted parcels less than 11 acres. It was hypothesized that land parcellation is in­ creasing in the state and that several socio-economic and institutional factors have been contributing to this in­ crease. Institutional factors such as the Subdivision Control Act and its 10 acre mi n i m u m lot provision have diverted developers from smaller lots less than 10-acres to lot sizes in excess of 10 acres. 182 The general framework of the study has been to collect data on parcels less than 11 acres and subject the infor­ mation to statistical analysis, using multiple regression and time series. This enabled the research to locate areas most affected currently by the process of parcellation as well as areas that would experience increasing parcellation process in the future. Spatial Distribution The process of non-platted parcellation was found to be most advanced and intense in the southern half of the state, especially in the Lower Peninsula, which has been the hub of the state's socio-economic activities. Approved subdivision activities have been much more intensive in the south, too, than in the north. However, spatial trend analysis indicated that the Northern Lower Peninsula re­ mained the frontier region for intensive parcellation activities in the future. Parcels in excess of 10 acres were becoming the most popular non-approved subdivisions in the region. Unlike the south where the demand for parcels had been associated closely with employment, the demand for land parcels in the north has been motivated by environ­ mental amenities and recreational activities. Little parcellation activities were noted in the Upper Peninsula. 183 Time Trend Parcellation leading to land fragmentation is a c o n ­ tinuing process in the State of Michigan. There is a t e n ­ dency towards increasing trend over time. Between 1963 and 1977, about 700,000 acres of land were parcelled out into lots less than 11 acres. This represented a doubling of the total acreages of parcels that existed in the state in 1963. Trend analysis indicated that more new non-platted parcels were created between 1970 and 1977 than between 1963 and 1970. Projected estimates indicated that another 800,000 acres of land wo uld be parcelled out in the next 20 years, or some 1,000,000 acres of new parcels wo uld be created between 1977 and 2000. While the extent of 10 and 10+ acre units may not at the moment be as large as the fears of some individuals, groups and public agencies suggest, growing phenomenon, it still remains a involving problems that merit policy considerat i o n s . Multiple regression analysis revealed that scattered homesites have been one of the major factors contributing to land parcellation in the state. This calls for policies that related to demand for homesites. It is to be r ecog­ nized that subdivisions once created lock up land w hich cannot be retrieved for non-urban uses without considerable economic and social costs. 184 Conclusions Parcell&tion definitely involves "important" rural lands and "critical" physical resources that Michigan would like to preserve in or for uses other than residential or other urban land use. The proportion of relatively large parcels (10-11) is increasing over time and spread in g to several frontier regions. Time series and multiple regression analyses indicate that two non-policy factors have been contributing to the spatial and time trends of (10-11 acre) parcellation and one policy factor accounts for a greater proportion of the sudden accelerated increase in the relatively larger p a r ­ cels between 1970 and 1977. The first non-policy factor relates to demand which involves job-oriented homesites. This is occuring in rural and semi-rural counties contiguous to urban centers. In­ creased personal incomes and quick transportation systems that permit commuting to work have contributed to reinforce this demand. Differences in land values between urbanized and non-urbanized counties appear to have contributed to such demand too. The second demand occurs mostly in the hinterland of Michigan in the Northern Lower Peninsula where physical environment as landscape, water bodies, and recreational resources such forests and so on, have att ra c­ ted extra regional demand for second and recreational 185 homesites. Again land values and tax burden in these fron­ tier regions tend to be lower than the southern heavilyurbanized regions. Neither of these demands is necessarily closely related or associated with local demand. Parcellation of land, in the range of 10-11 acre size units are already advanced in most counties in the southern third of the Lower Peninsula. Zoning and other land use regulatory policies can hardly have any further effect on the fragmentation process and trends in these counties (e.g., Oakland, Wayne, etc.). It is probably inappropriate to talk of "process" in such counties. term "process" is still applicable and appropriate. is available for large lot parcellation, equal. In the north, the Land all other factors Projection indicates that the north of the Lower Peninsula is the main hub of future large-lot development activities. The rush by real estate agents, developers and subdividers, to the north in response to expected boom in the recreational industry is likely to generate irres­ ponsible land transaction activities— speculation, p r e ­ mature developments, leap-frogging, narrow strip de v e l o p ­ ments and eventual idle lots. Independent studies have shown that residential uses will continue to be the predominant taker of land in urban areas and of the new parcels that are being created in the rural contiguous and fringe areas, as well as in remote 186 up-country. Single-family housing still consumes most of the residential land though in high-density urban areas, multi-family housing has exceeded single-family in the number of dwelling units. Mobile and modular homes are also increasing in number each year and contributing to the rapid consumption of prime lands and the increasing number of small-tract parcellation. The trend continues to increase over time and to spread spatially from the south toward the north. As the process proceeds, and land use conflicts increases (between urban and rural land uses), public concern increases and anxiety about Michigan's future rural industry resource base mounts. The third factor is a policy variable. The Subdivision Control Act which exempts lots greater than 10 acres did significantly contribute to the acceleration of the creation of 10+ acre parcels. Subdivision Regulation; Subdivision Control Act of 1967 This study found that the Subdivision Control Act has contributed significantly to the increasing trend in large lot parcellation in the state. The exemption of land over 10 acres from platting provided by the Plat Act has in­ creased the demand for and supply of small tract holdings slightly in excess of 10 acres. The side effects of the 187 Plat Act on rural land resource base may not have been anticipated. The Plat Act has tended to create the problem of haphazard developments more than before with 5-split limitation within 1 0 -year p e r i o d » in many counties where zoning regulations are not strictly enforced. In their responses to letters sent them about the issue, several County Extension Directors reported that the 10-acre lim­ itation in the Plat Act is positively related with the increasing trends in the creation of larger tracts for residential purposes. Donald Jud's study (1980)1 on the effects of Zoning on Single-Family Residential Property Values in the City of Charlotte, N. Carolina, concluded that: 1. Buyers of residential housing seek uniformity in neighborhood (communities) land use. Where such uniformity is provided by a residential zoning classification, consumers are willing to pay a premium for it. 2. A decrease in per acre cost of homesites reduces the fixed cost component of residential construc­ tion and hence the total average cost of subdivi­ sion development. The tendency is to increase the supply and reduce the price of large-lot res­ idential homesites. 1 Jud, G. D., 1980, (ibid.), p. 152. 188 3. Increased supply of large-lot tracts tend to lower prices for large-lot housing relative to small-lot housing units on per unit basis, and hence demand for such housing units increases. The 10-acre limit which has encouraged the creation of lots slightly in excess of 10 acres has actually generated demand for and supply of large-lot residential lands in the state. As one extension director pointed out, "Real estate people have made costs of acquiring land fairly attractive with requirements of low down payments and low monthly p a y m e n t s ." 1 Subdividers reduce acquisition cost by avoid­ ing platting requirements and costs. They are, therefore, able to sell residential sites at a lower cost per acre. This has encouraged younger and middle-aged individuals to seek more secluded areas. The impact of the Subdivision Control Act of 1967 on publicly approved subdivisions, however, was not conclusive. The declining trends in the number of plats and acreages of subdivided land is a reflection of increasing platting costs, uncertainties in the real estate market in the mid1970's when many developers lost money and the natural tendency of the market to move towards equilibrium. Responses of County Extension Directors confirm this ar­ gument. See Appendix 3-B, especially Lapeer and Antrim. 189 The Impact of the Subdivision Control Act is best re­ flected in isolated non-platted parcels. Loopholes in both Michigan Land Sales Act and the Subdivision Control Act allowed developers to partition land into more than five, but less than 25 parcels, each parcel over 10 acres in size without meeting any legal or policy requirements. It is here concluded that zoning and subdivision reg­ ulations in Michigan have contributed to large lot parcel­ lation, but not to publicly approved subdivisions and the process of creating these large lots is still increasing over time and spreading spatially to most rural lands where land values are relatively low and recreational and environ­ mental amenities tend to attract homesite consumers. Wit h ­ drawal of prime rural lands into such secluded large lot residential units will continue if no further action other than current land use control mechanisms is taken. The process of large-lot fragmentation favors devel­ opers who reap economies of size. Certain zoning regula­ tions also favor farm owners whose lands are protected and preserved at the expense of society, but who have the option to eventually sell land for capital gains, the protection and preservation having contributed to the appreciation of the value of the land. Cost is imposed on all taxpayers who must pay for the provision of public services and util­ ities to these scattered and secluded residential units or for the protection program. 190 Recommendations A. Recommendations Stemming from the Study Findings In view of the above analysis, the following recom­ mendations are made: 1. The minimum lot size provision in the Subdivision Control Act should be removed from the definition. The state may still retain the 5 split limitation within the 10-year period. States such as Alaska do not have any acreage limitations, but m ai n­ tain split minimum. Alaska does not allow any divisions of property before platting. Illinois, Massachusetts and Ohio also do not have any m i n i ­ mum splits before platting. Michigan, California, Idaho, Wisconsin and a few other states allow 4 splits. So far the number of allowable splits before platting range from 0-4. Michigan appears to be the only state that has the * 10-acre minimum specified in the Act. have 5-acre minimum (Alaska, Ohio, Oregon). Rhode Island, Idaho, Most states Illinois, for example, has only 1-acre minimum and Wisconsin has li-acres minimum. It is suggested that a comparative study about the impacts of these different m i n i ­ m u m lot size provisions be made so that Michigan 191 can draw from them. On the other hand, the sub­ division definition may be made more flexible by requiring p latting of any kind of small lot s plit­ ting, size. irrespective of the number of splits and In this case, market forces will eventually establish these threshold limits. It is to be recognized that pa rt itioning of land into splits less than 25 parcels each of w hich is larger than 10 acres is virtually uncontrolled. The plat act should require the official platting of all subdivisions of under 40 acres when the intent is to use the subdivided parcels for res i­ dential purposes. Some eas in g of platting res­ trictions may be wit h plats of 4 or less units. In this respect, more detailed research is needed to ascertain the effects of such easing off and w hich types of the provisions to be relaxed. 2. A state level land use policy coordinating agency should be established to coordinate the actions of local land use agencies, to organize a detailed study of tax laws, zoning regulations and all other public policies that independently and in their various jointness combine to infringe on private land use and resources. Confli ct in g and counter balanc in g effects that tend to neutralize 192 the possible positive impacts on these regulations should be sorted out and corrected, if any exists. This has implications for the next recommendation. 3. Public policies which contribute to urban sprawl— ranging in single or in combinations— must be studied and screened or reshaped to meet with the increasing public demands for limit to urban expansion and growth. 4. This study has revelaed that for certain counties large-lot parcellation is already extremely ad­ vanced and the process has reached its limits. Such counties as portrayed by density index in Appendix 5-C may require different types of policies which will promote land consolidation and recombination. B. Recommendation for Future Research This study has raised more questions than it answered. This result was expected and indeed had been an implicit objective of the study. Land parcellation is a social, economic, political, personal and psychological issue which takes its root in the fundamental institution of land ownership and natural resource allocation. Research topics are never exhaustive in this area of human activity especially as society develops and human problems multiply in complexity and in dimensions. 193 An area of great interest for future research may have to deal with the nature and scope of land transactions in the state— who sells and who buys, why, for what, what are the impacts on the economy— social set up, power structures, and so on. How many times does a single tract change hands (study of turnover rate in land parcellation). Another p arcellation category worth studying is the 1 1 - 2 0 acre parcels. As farmlands increase in size, will 11-20 acre plots be economically feasible? What kind of farming and forestry can efficiently be practiced on such limited tracts? Another area of research might be a study of the im­ pacts of local interest pressures on the administration of land use control regulations. What are the benefits and costs of such local impediments? the costs? Who gains and who bears Why are land use regulations not adopted and/or enforced in all localities, counties, and municipalities? Land parcellation, as pointed out, is a process by w hich land is allocated among persons and among uses. cannot be checked under a normal process of market It forces. Sellers and buyers alike derive private satisfactions— either economic, actions. social or psychological, However, from land t r an s ­ the process can be directed and efforts channelled to meet current as well as future goals and as­ pirations of the people in the state. 194 Final Comments Planning for the Future The future is for the one who plans ahead; planning does not necessarily imply governmental interference in private rights or socialization of private ownership. It is simply looking ahead into the future and letting the past and present provide guidance into the unknown future. It involves careful and reflected anticipation and calcu­ lated choices among competing alternatives*, therefore, there is always a cost. Planning contributes to weighing and balancing competing ends within the constraint of available resources; making appropriate choices that would maximize current and future welfare of society at little cost to individuals in the society. Such an optimization process requires the input of all citizens as they provide the needed mandate to their political representatives and further cooperate with the various institutional set ups for the successful and effective implementation and working of planned policies. In this respect, the citizens of Michigan have a major responsibility and role to play in determining the proper use, administration and development of their base resource— land. 195 An African wise saying about land ownership has It that: "Land belongs to the numerous d e a d . It is the trust of the living few and is the property of uncountable m illions yet u n b o r n ." This philosophy sums up all traditional land use policies in indigenous Africa. In this respect, probably, the advanced nations have something to learn. The notion of land as a "commodity" may eventually have to give way to land as "common trust". APPENDICES APPENDIX 2-A Sample of Land Atlas and Plat Book Map of Portage Township, Houghton County, Upper Peninsula 196 APPENDIX FIGURE 2-A S a m p l e of L a n d A t l a s an I P l a t Township, Houghton County, B o o k M a p of P o r t a g e Upper Peninsula “ fgffH6A5T PORTAGE T 5 2 NrR.34 9Stiff*****•***•• . 1 's .x~’ “— r * . "ll *• •I‘ f c V *4 **■»«?J C •• ’ Y.“^ r Z 'i £ ! L /*■'»**•*#•#-ir*« H4/A* ^ #/r.MEN1 O F M i H I C U L I U K I: u i n l ' l K A T I N I j N m u m I n i D c \c lv p m e n r R esources Ih n id in g February 10, 1981 Mr. Lawrenco M. Stebblns Ottawu County Extension Director County Building Cranii Haven, Ml 49417 Dear Larry: Ttie Du par m e n t of Resource Development at Michigan State University is currently studying the problems an sue ia ted with the parcellation of land tracts of LI acres and less. Your county is one of a sample of 30 Michigan counties that has been sc Lee ted for study. Our analysis to date Indicates tliat your cuunty has experienced .in above average amount of parcellation when degree of parcellation is measured in terns of area parcelled weighted by county population density and area. We realise tliat factors of which we arc not aware may explain the parcellation trends in your county. To help us better understand what has been going on during the last 13 yearn, wc would greatly appreciate it if you would take a few moments to give us your perceptions about trends in the creation of small tract holdings - must particularly lots in the 10 to I! acre size category, the farces that arc favoring small lot parcellation, and the problems (if any) that this may be causing for local government. Please send your coirnncnts to .Ither Dr. Raleigh Burl owe or Mr. Francis Arthur of the Department of Resource Development. Your cooperation in responding to this request will ba greatly appreciated as it will provide nveded perspective to our overall study. Thank you. Yours sincerely. Raleigh Barlowe Raleigl Extension Specialist F.xten* Land and Natural Resources Policy RB/js APPENDIX 3-B-2 Selected Responses of CEDs 198 E5 COOPERATIV" EXTENSION SERVICE JL viChkun state university •u s department of agriculture i genesis county cooperating w PiMtm 40504 February 23, lJrtl 1 Mr. Francis Arthur Resource Development Department ill Natural Resources Michigan State University Last Lansing. MI 43824 Dear Mr. Arthur: I am respondim; to vciir survey of our countv regarding the parcellation of land. Several Factors cause jicoplc to buy a 10 acre t£ct of land. !. 2. 3. 4. lies ire for space Consider the rtiral area a better nlacc to raise a family A desire to pet hnck-to-the-earth type of living. Attempt to supplement their income • gardening, raising small fruit and scllint: their nrrxluctioii. t6. l.ack choice to huy smaller rract (less chan JO acres) dexLl*C15 offer 10 n lus icres to avoid nlot jet. P. N ; m c buyer doesn't real ice the extent of space purchased - later they become frustrated and use only a fraction of the acreage. ". High wages and salary in area provide enough income to support the more extragnnt life-styleH, Ihev move to avoid urban social problems - crime, racial, ect. 0. Land sjieculatoi's bought farms with the sale intention of dividing it into 10 plus rr.rcels. 10. farmers real ice return on their land b y selling off the frontage on roads. Problems for local government: I. Increase cost of services for roads, bridges, fire and police protection. 2 . Demand for waste disposal svstcins - sewage ami solid waste. 3. School class room snace increase demand 4. Conftict of values with the "natives". Problems with a mure global perspective: Increasedcansimption Leo W. Dorr County F.xtension Director C i i ’ fv d ' i i . * a . '* # " t o 1*1 0 9 C 9 W 4 h a t 9 « < H lT n n a i h i '1 C * COW *tyr*<1 rm ia o r ig in Of M s of 1‘ncrgy. 199 COOPERATIVE^ EXTENSION SERVICE MICHIGAN STATE UNIVERSITY . u 3 DEPARTMENT OF AGRICULTURE A COUNTIES COOPERATING LAPEER COUNTY EXTENSION OFFICE • 1STS SUNCREST DRIVE • LAPEER. MICHIGAN 4«44« (313) 687-0341 February 26, 1981 Raleigh Barlowe Extension Specialise Land 6 Natural Resources Policy 313 Natural Resources Michigan State University East Unsing, MI 48821 Dear Raleigh: Why is Lapeer County being "parcel 1i2ed" into 10 ta 11 acres plots, and what problems does this bring? 1. L'p to now a person could purchase 10 acres for little mare than an improved lot in a subdivision in Oakland and Cenesee Counties (510,000.00) 2. FT'.IIA and Federal U n d Bank have provided ready financing - in the case of HMHA the rural FHA subdivision, has featured more folks from Flint and Pontiac. (This created more pressure from those who could afford it, to isove out to get some land around them.) 3. We have had a large number of real estate people covering Lapeer County promoting the splits. 4. Land contact financing favors splits - easy quick commission no red tape. 5. '.(any of our earlier splits are being split again now that the 10 years is up (under the pl3t act) 6. Thu Lapeer County Health department has "caved in" under pressure and now will approve alsiost any 10 acre site if people want to spend the money for a septic system builtto their specifications. 7. The extension of M-53 (VanDyke) and the extension of M-21 from Lapeer to Port Huron has caused more speculation. 3. Most of our buyers now are third home owners and retired folks that want to move to their "final” dream home. (We still do have some river frontage, and some woods left.) ■Cooperative F ..rn .in n Servier P-nrmtTi. mre ptvp m ill wiihnni t m n l m tw * cpIp*. VI 200 2. Berlowe -2/26/81 9. Gardner Real Estate h''ve 4 rfici>o surrounding Lapeer County, and they specialize in splits 10. The Glaciers came and went five tines through Lapeer CountyMuch of our land base is "alfalfa" - animal agriculture type land - and you know what the beef Industry has gone through. Our dairy numbers aro holding at 218 farms, down from 250 In 1968. The area south of K-21, and west of US-24, and the four townships surrounding Lapeer are where moat of the splits are occurlng. 11. The back to the earth urge is getting stronger and stronger. A husband and wife with two incoaws can out bid a farmer or retired land owner for land. Ue still have $850.00 per acre farm land, and $30.00 per acre land rent. 12. The impacts of the out migration is pressure on schools, lack of adequate roads, and road maintenance. Garbage removal and solid waste management are problems. Rural crlsM, substance abuse - (probloms amoting young people coming from rc-locted families are frequent). In response, church growth, volunteeriAn1 community Involvement is building. See attached 4-H growth rates 1980-81 enrollments are now over 2000. James M. Hutchinson County Extension Director JMH:mg Attached is: 1. Land in farms by township 1940 to 1978 2. "Are you Concerned" - a slide tape script 3. Lapeer County Demographics taken from our solid waste project now in^progress. 4. U.S. Metropolitan Area Projections 5. 4-H Enrollment 6. Raleigh Barluwo letter 7. Comprehensive Development Plan Susssary a c.c. Helen Ulllis Adger Carroll 201 *.Cooperative Extension Service M ICHIGAN STATE UNIVERSITY ............ - ■ - - ■— A N T R IM CO UNTY BUILDING P.O. 9 o * 427 BsMsee. Mi. 49615 1616) 533-6607. E*t. 31 U.S. O ep stim e n i o f A griculture and A n trim C ounty cooperating February 23. 1981 Dr. Raleigh Barlowc Excenaion Specialise Land and Natural Resources Policy 313 Natural Resources Michigan State University East Lansing. MI 488 24 Dear Raleigh, This is in response to your February 10 letter in which you requested feedback on your study relative to the parceling of land traces of eleven acres and less. These comments have been nut together as a result of a meeting be­ tween myself, Warren Studlcy Soil Conservation Service District Conservationist, and Karl Larson, retired Antrim Countv Extension Director and presenc part-time Kalkaska County Extension Director. I. Perceptions about trends in creation of small tract holdings. A. We see the trend continuing at a declining race. We expecc the rate to decline as transportation costs increase. Also associated with this is an adverse economic climace at the presenc time. Of course this may change. At the present time there are many small tract parcels available for sale that are not moving which could be due to a varie ty of reasons. However, we also sec further evidence of small tract hold­ ings still being developed. II. Forces favoring small acre parceling. A. Wo effective controls: The Plat Act, as you know, provides an exemption of land over ten acres. It's obvious that small tract holdings slightly in excess of ten acres have been developed simply to avoid the Plat Act. Tn fact, in some ways it has created a problem because it has caused the development to be more haphazard than it was before with the 5 split limitation within the 10-year period. 9p«fr«m •« A«r>cuitur«, M*ro««f*9. Natural R i w c n Pu«i>c Policy. Family bivin« 4 6 m voura 202 Dr. Raleigh Barlow* Feb. 23, 1981 * page 2 B. Terms of acquisition: Real Estate people have made costs of acquiring land fairly attractive with requirements of low down payments and low monthly payments. Again, as the economy is tightened more people arc probably out of the market at the present time chan have been in the past. C. A desire to own land, especially in the north: There seems to be a feeling that owning land gives one proprietary rights over Che great out-doors. Particularly with regard to recreational use and the feeling of being able co flex one's muscle a bit. D. Lack of planning and zoning: The lack of this emphasis has enabled development of land in an undesirable manner. There is evidence to suggest chat the zoned townships have goccen a much better handle on the situation. In fact some people have purchased land in zoned areas because of the protection afforded. C o u n t y zoning was defeated twice in Antrim County and there is still not universal support for zoning as evidenced by a majority of townships in the county not yet zoned. E. Geographic relationship to metropolitan areas: Antrim County is four or five hour's drive from Michigan's metropolitan areas as well as similar areas in Illinois, Ohio and Indiana. This has made it easy for people to spend weekends or longer in the area. It is unclear whether the increased transportation cost will In­ fluence this aspect or not. F. Comparatively low land values have promoted parceling: Normally land values in northern Michigan are lower chan comparable land in southern Michigan or many ocher parts of the country. It has also been noted that there are differences in land values on a micro level within the county with more parceling occurring in the lower valued land. G. There has been a difference noted in the types of people purchasing secluded ten acre parcels versus those who purchase in sub-divisions which have also been developed in recent years. An observation was made that younger and middle-’aged people tend to buy the more secluded areas for reasons stated above, whereas the older, possibly retired buyers are more concerned about proximity to services and amenities. III. Problems caused by parceling. A. The most serious is one of which you are well aware. Services, utilities, roads, law enforcement, fire protection and waste disposal are all areas which are adversely impacted by parceling. B. Loss of high value agricultural land has occurred in some areas. This is particularly true of fruit sites which, unfortunately, are generally also attractive building sites. This has helped co drive up the costs of fruit land. Of course, high cherry prices some years didn't serve much tc keep it down either. 203 Dr. Raleigh Barlow* Feb. 2 3 , 198L page 3 C, Expectations of "imigrants" relative to those of "natives" appear co be somewhat higher as they view excepcable levels of services offered. D. We have not noted an increased pressure on schools. What has happened apparently is chat as school enrollments are declining in many parts of the country they are more or less remaining in a static condition in this area. I expect that in-migration is compen­ sating for natural reduction, Related to this, however, would be an expanded need for school bus routes which is related co A. above. One very serious problem which could result has to do with some­ thing chat has noc happened yet, but which we feel is worthy of commenc. That has to do with a proposed change now being considered for Public Act 96, "Commercial Forest Act". We understand that there is a political movement afoot co alter the eligibility of land to include parcels smaller chan Che current 60 acre limitation. We feel that this would be very undesirable for several reasons. One, it would serve to increase demand for smaller trace parcels, chereby compounding the current problem. Two, it could drastically reduce the tax base leaving the local units in a much worse condition chan they are currently. Philosophically, I personally feel that there should be some reasonable cost associated with owning land and re­ ducing the eligibility requirements will certainly increase market demand in favor of small tract parcels especially for the n o n ­ economic sized parcels of which we are discussing. I hope that these comments will help in your study. Sincerely, Burton J Stanley fl County Extension Director BJS/Jh cc: Warren Scudley, SCS District Conservationist Karl Larson, C.E.D., Kalkaska County Dave Twining, County Planner 204 INTRODUCTORY THOUGHTS AiJUUT IRlihDS, FORCES, AND PROBLEMS OF SMALL LAND PARCELS by Wayne h'icrman, Oakland County County Extension Director Trends: A major portion of the land area has already been divided and sold in 10, 11, 11, or IS acre parcels. It isno longer in process, for the most part it has happened. Forces: Most people wishing to dissolve ownership of large parcels (1U0 acres or morcj find the prospective buyers tobe few in number. A hundred acre parcel is too small to farm as a single farm and too big for most city people seeking life in the country. By breaking the parcel? into 10-acre increments, higher per acre value and greater numbers of prospective buyers become available. Dividing the acreage into small 10-acre parcels further provides for increasing the value per acre while attracting still larger numbers of interested buyers. To take the initial ton acres, subdivide, plat, construct roads, sewers, and streets for subdivision development requiring an added period of time, doesn't appeal to most sellers. It involves added investment, compliance with state and local statutes, and entails an extended period of time for reclaiming one's initial investment. There is also the added element •• "Risk." T think most would soo this alternative as not being worth the anguish. Problems: Most land parcels larger than two to three acTes have no useful purpose for the city of suburban family settling in the country. They simply don't know what to do with the six or seven remaining acres. Some elect horses and some livestock. For most, the vegetation at the time of purchase remains unmanaged throughout one's ownership. Transportation routes, schools, fire protection, emergency medicai service, and policy protection are soon over committed and require replanning and added revenues for expansion and modernization. Rather than consolidating population growth near cities in regulated subdivision development, ten-acre paroels have distributed populations farther form business'and occupation locations. It would seem that this distribution requires greater demands for energy in the form of transportation, communications, and electrical service. For local governments, it means responding to the special interests and concerns of more people. It moans hearing more complaints, pro­ cessing more tax statements, added puhlic hearing? on community issues, and making decisions which have long-term economic and social conse­ quences for the expanding constituency. Often small parcels arc sold at unusually high prices when contrasted to surrounding property. These increases are not fondly received by long-time permanent residents, thus often placing the burden of blame 205 on the assessor. The blame should be placed on the buyer who purchases laud above its existing market value. However, it is not easy for him to know what the fair market value should be either. Maybe the blame should go to the seller who over stated his price? On the other hand, tho price was set in a "free market place." Who can say who is right? Kells, se p t i c systems, and w a s t e disposal add burdens to local govern­ ments. We have witnessed contamination of chemical waste disposal in landfill sites. Increasing population to rural areas makes it in­ creasingly hazardous to bury and dispose of materials in landfill sites for fear it will penetrate the well water supplies. Landfill sites are becoming increasingly difficult to find, not to mention expensive. It is local government who must locate and monitor these facilities. February ism APPENDIX 4-C Frequency Histograms with Fitted Smooth Curves for Categories of Parcellation Data 206 IO r S No ol Acres 10 r L0 123 230 373 300 623 750 9731000 1125 1250 (3731500 0 23 50 73 100 123 ISO 175 200 225 230 273 300 A c re a g e in O's ■Or u Acreage in O's to r 0 23 30 73 100 123 ISO 173 200 223 230 273 300 325 Acreage in OO's APPENDIX 4-D Descriptions of Variables used in Regression and Correlation Models 207 APPENDIX 4-D Description of Variables Used In Regression Equations II Variable Acronyms Description and Measure____ Parcellation PAS Dependent Variable (in a c r e s ) 1. All Parcels AAPAS Amount of all parcels less than 11 acres 2. "Large" Parcels LAGPAS Amount of parcels in the size range 1 0 - 1 1 acres 3. "Small" Parcels ASPAS Amount of parcels less than 10 acres 4. Minimum Lot Parcels AMPAS Amount of 10 acre parcels (Minimum-Lot provided in the Sub­ division Control Act of 1967) 5. Large-Lot Parcels ALPAS Amount of parcels in excess of the minimum-lot parcel (1 0 + acres) Physical Determinants (P) 6. LACO Land area per county in square miles 7. PUBREC Public land in rec­ reational and forest activities as percent of total land area per county (1974) APPENDIX 5-E Statistics on Non-Platted Parcel Table E-l to Table E -8 209 APPENDIX 5 TABLE E-l Holdings of Parcels Under 11 Acres by County and by Period 1963, 1970, 1977 Number Holdings of 11- Acre Parcels County 1963 1970 1977 1,904 494 239 1,326 3,278 2,180 707 395 1,482 3,649 3,114 1,259 1,307 2,980 4,243 Calhoun 7 Cheboygan 8 Clare 9 Clinton 10 Crawford 2,224 516 218 777 1,259 2,264 937 451 1,867 991 3,104 1,492 2,905 1,695 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 1,118 393 229 459 211 1,353 436 505 713 460 1,882 779 1,372 1,097 857 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 232 553 488 400 1,147 361 682 507 725 2,252 1,234 956 641 4,976 5,293 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 188 1,979 636 285 2,209 348 2,357 514 3,456 1,796 1,301 3,369 26 27 28 29 30 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 921 1,592 1,916 860 263 1,249 2,029 2,392 305 2,444 3,265 3,265 1,536 342 28,316 37,067 65,392 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 Total 1,112 781 2,556 1,021 1,200 210 APPENDIX 5 TABLE E-2 Distribution of Total Acres of Land Under 11- Acre Parcellation for 30 Selected Counties in Michigan 1963, 1970, 1977 Acreages of 11- Acre Parcels Selected Counties 1963 1970 1977 10,476 2,138 1,723 6,155 19,344 11,783 2,781 2,797 6,679 20,720 16,314 6,194 8,913 11,567 24,499 Calhoun Cheboygan Clare 8 9 Clinton 10 Crawford 9,528 3,020 1,692 4,102 5,706 10,303 5,300 2,611 8,385 5,297 13,659 9,005 8,214 12,769 11,593 11 Delta 12 Gogebic 4,799 13 Grand Traverse 14 Hillsdale 15 Houghton 1,478 2,943 1,303 5,775 2,432 2 ,940 4,082 2,396 5,543 3,813 7,418 6,610 4,935 Huron Iosco Iron Lapeer 20 Livingston 1,178 2,298 2,681 2,784 6,631 1,178 3,047 1,803 4,793 12,517 4,759 5,422 3,482 27,501 31,785 1,299 11,196 4,147 1,645 11,281 1,990 13,222 6,492 3,384 13,172 3,021 17,792 10,297 4,969 17,658 3,896 9,880 10,837 4,333 1,438 5,384 12,227 12,915 5,064 1,654 11,029 20,610 22,995 7,598 1,900 151,952 197,568 344,859 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 16 17 18 19 21 Mackinac 22 Macomb 23 Manistee 24 Menominee 25 Monroe 26 27 28 29 30 Montcalm Newaygo Ot t awa S t . Joseph Schoolcraft Sample Total 2,121 211 APPENDIX 5 TABLE E-3 Trends in Parcels Less Than 10 Acres For 30 Counties in Michigan 1963, 1970, 1977 Acreages of 10- Acre Parcels County 1963 1970 1977 5,601 1,528 633 4,378 10,818 6,185 2,170 959 4,790 12,253 8,520 3,800 2,971 9,084 13,641 Calhoun Cheboygan 8 Clare 9 Clinton 10 Crawford 6,885 1,478 436 2 ,475 4,193 7,003 2,485 3,841 5,578 3, 393 9,305 3,546 2,045 8,444 4,113 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 3,710 1,370 698 1,555 661 4,578 1,501 1,388 2,330 1,488 6 ,075 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 788 1,448 1,578 1,333 3,575 1,076 1,855 3,411 2,298 1,686 2,120 2,351 6,791 14,738 14,118 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 478 7,026 1,838 763 7,628 1,120 8,323 3,115 2,481 8,825 1,541 12,095 4,958 3,378 11,823 26 27 28 29 30 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 2,666 4 ,218 6,838 2,603 848 3,498 5,281 8,258 3,038 983 6,523 7,571 15,220 4,368 1,095 90,237 118,622 129,143 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 Sample Total 2,490 3,841 3,610 2,381 212 APPENDIX 5 TABLE E-4 Trends in 10 Acre Parcels For 30 Counties by Study Period Acreage of 10 Acre Parcels County 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien Calhoun 7 Cheboygan 3 Clare 9 Clinton 10 Crawford 6 1963 1970 1977 4,360 510 1,090 1,630 6,730 4,810 590 1,820 1,710 6,430 6,240 1,890 5,700 1,950 1,500 1,140 1,490 1,460 2,260 2,720 1,440 2,250 1,820 3,430 5,270 5,560 3,370 7,070 910 730 770 1,050 910 1,510 1,290 750 1,880 1,270 3,430 2,170 610 1,160 1,070 4,300 1,290 3,040 1,320 12,080 14,800 860 3,670 3,230 840 3,160 1,470 4,080 4,950 1,350 4,260 4,790 3,810 1,320 580 1,770 5,780 4,810 1,480 650 4,180 11,170 7,050 2,180 700 53,110 66,450 133,170 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 400 840 1,050 1,430 2 ,310 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 800 3,130 26 27 28 29 30 M o ntcalm Newaygo Ottawa S t . Joseph Schoolcraft Total 1,020 610 2,120 6'50 2,760 1,220 2,200 2,200 8,170 1,000 213 APPENDIX 5 TABLE E-5 Trends in 10+ Acre Parcels For 30 Counties in Michigan 1963, 1970, 1977 Acreages of 10+ Acre Parcels County 1960 1970 515 788 0 0 21 21 147 1,796 179 2,037 Calhoun Cheboygan 8 Clare 9 Clinton 10 Crawford 693 42 116 137 53 1,040 95 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 179 147 21 10 21 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 26 27 28 29 30 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 Total Sample 210 557 84 368 32 42 462 158 0 10 95 32 53 10 21 242 1,428 746 1977 1,554 504 242 284 2,688 924 189 609 956 410 588 53 147 839 1,554 53 84 42 683 2,867 21 10 10 1,040 189 32 893 1,229 147 63 1,187 1,617 387 10 872 189 410 116 1,166 347 540 10 21 326 1,869 725 1,050 105 8,605 12,496 23,146 221 1,575 214 APPENDIX 5 TABLE E-6 Trends in <10-11) Acre Parcels For 30 Counties 1963, 1970, 1977 County Acreages of 10-11 Acre Parcels (Large Lots) 1963 1970 1977 4,875 510 1,090 1,777 8,526 5,598 611 1,841 1,889 8,467 7,794 2,394 5,942 2,484 10,858 Calhoun Cheboygan 8 Clare 9 Clinton 10 Crawford 2,643 1,542 1,256 1,627 1,513 3,300 2,815 1,650 2,807 1,904 4,354 5,459 6,169 4,326 7,480 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 1,089 750 780 1,388 642 1,197 931 1,552 1,752 908 2,468 1,323 3,577 3,000 2,554 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 400 850 1,103 1,451 3,056 705 1,192 1,080 2,442 5,728 1,343 3,124 1,362 12,763 17,667 870 4,899 3,377 903 4,347 1,480 5,697 5,337 1,531 5,835 1,886 671 4,506 13,039 7,775 3,230 805 78,946 155,716 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 22 23 24 25 21 Uackinac Macomb Manistee Menominee Monroe 821 4,170 2,309 682 3,653 23 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 1,230 5,662 3,999 1,730 590 27 28 29 30 Total Sample 61,715 6,946 4,657 2,020 215 APPENDIX 5 TABLE E-7 Trends in Number of Holdings of 10- Acre Parcels For 30 Counties 1963, 1970, 1977 Number of Holdings of 10- Acre Parcels County 1963 1970 1977 1,419 495 130 1,149 2,434 1,624 646 2,342 211 1,294 2,812 714 2,733 3,170 Calhoun Cheboygan 8 Clare 9 Clinton 10 Crawford 1,963 362 93 615 1,108 1,939 656 287 1,589 801 2,673 947 586 2,477 949 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 1,010 318 151 322 147 1,234 343 350 540 370 1,638 647 1,015 801 609 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 192 468 378 255 845 291 563 399 482 1,687 1,100 644 505 3,703 3,540 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 106 1,567 406 217 1,848 261 1,873 775 691 2,077 366 2,838 1,264 1,145 2,793 26 27 28 29 30 Montcalm Newaygo Ottawa St. Joseph Schoolcraft 798 1,030 1,517 689 204 1,061 1,340 1,928 821 238 1,995 1,970 4,208 1,218 263 22,186 29,183 49,871 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 Total Sample 1,022 216 APPENDIX 5 TABLE E-8 Trends In Number of Holdings of Parcels (10-11) Acres for 30 Counties 1963, 1970, 1977 Number of Holdings of (10-11) Acres County 1963 1970 485 51 109 177 844 556 61 184 188 837 772 237 593 247 1,073 Calhoun Cheboygan 8 Clare 9 Clinton 10 Crawford 261 154 125 162 151 325 281 164 278 190 431 545 614 428 746 11 12 13 14 15 Delta Gogebic Grand Traverse Hillsdale Houghton 108 75 78 137 64 119 93 155 173 90 244 132 357 296 248 16 17 18 19 20 Huron Iosco Iron Lapeer Livingston 40 85 70 119 108 243 566 134 312 136 1,273 1,753 21 22 23 24 25 Mackinac Macomb Manistee Menominee Monroe 82 412 230 87 484 337 90 429 148 622 532 156 576 26 27 28 29 30 Montcalm Newaygo Ottawa S t . Joseph Schoolcraft 123 562 399 171 59 188 689 464 87 449 1,295 774 318 80 6,130 7,835 15,521 1 2 3 4 5 Allegan Alpena Antrim Bay Berrien 6 7 Total Sample 110 145 302 68 361 200 1977 APPENDIX 5-F Quinquannial Projection Estimates of Amount of Parcellation in Michigan by Size Unit (1980-2000) 217 APPENDIX 5 TABLE F-l Quinquannial Projection Estimates of Amount of Parcellation in Michigan by Size Unit (1980-2000) , . . Projected Year Size Unit and Parcellation in Acres 10- Acre Parcels 10 Acre Parcels 10+ Acre Parcels 10 k 1 0 + Parcels 1980 644,862 448,243 80,754 528,997 1985 764,090 540,351 102,723 643,074 1990 882,818 631,289 123,290 754,587 1995 1,000,585 733,511 146,702 880,213 2000 1,120,218 830,105 165,174 995,279 APPENDIX 5-G Summary Data For - Parcellation Density Scores for 1963, 1970 and 1977; Projected Estimates for Year 2000 Table G-l to Table G-4 218 APPENDIX 5 TABLE G-l Derivation of Weighted Parcellation Densities For 30 Counties (1963 * 100) Land Area x 640 (Ai) Acres County 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Allegan 826 565 Alpena 476 Antrim Bay 447 580 Berrien 709 Calhoun Cheboygan 721 Clare 571 Clinton 572 Crawford 561 Delta 1 ,177 Gogebic 1 ,107 Grand Traverse 462 Hillsdale 600 Houghton 1 ,017 Huron 819 Iosco 544 Iron 1 ,171 Lapeer 658 Livingston 572 Mackinac 1 ,014 Macomb 480 552 Manistee Menomi­ nee 1 ,034 Monroe 557 Montcalm 712 Newaygo 849 Ottawa 563 St. Joseph 506 School­ craft 1 ,181 Sample Average 704 Pop. Density 1960 (Di) Persons Per Square Mile 11-Acre Weighted Parcels D. 1963 1963 (PAi)Acres 1.47 .30 .13 5.47 14.24 4.36 .14 67 9 29 10,476 2,038 1,723 6,155 19,344 9,528 3,020 1,692 4,102 5,706 4,799 22 2,121 .07 72 58 35 41 30 14 64 103 .38 .47 .07 844 34 1,478 2,943 1,303 1,178 2,298 1,681 2,784 6,631 1,299 11,196 4,147 32.57 .42 24 179 50 28 175 83 1,645 11,281 3,896 9,880 10,837 4,333 .06 5.60 .45 .54 5. 57 1.18 8 1,438 .02 82 5,065 1. 0 0 70 50 22 240 258 196 20 20 11 .10 .80 .15 .20 .10 .21 .05 .45 1.98 .02 Coefficient ®1063 * 59 1'059 Parcellation Density: WPP * 6 Di PAi ♦Area figures are not repeated for 1970, 1977 and 2000 AD. 219 APPENDIX 5 TABLE G-2 Weighted Parcellation Density Index by By County and for 1970 (1963 ■ 100) Population Density (1970) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Allegan Alpena Antrim Bay Berrien Calhoun Cheboygan Clare Clinton Crawford Delta Gogebic Grand Traverse Hillsdale Houghton Huron Iosco Iron Lapeer Livingston Mackinac Macomb Manistee Menominee Monroe Montcalm Newaygo Ottawa St. Joseph Schoolcraft Sample Average *1963 " 1 - 0 5 0 0 W P D 1963 - 1 - 0 0 81 54 27 263 283 200 23 29 85 12 31 19 84 62 34 42 46 19 80 103 10 11- Acre Parcels (1970) 11 2 2 6 20 10 5 5 8 5 5 2 2 4 2 1 3 2 4 12 1 783 781 800 679 720 303 300 491 385 297 775 432 940 082 396 781 047 766 793 519 990 1,303 36 24 213 56 33 228 94 7 13 2 2 2 6 492 3 384 13 172 5 384 12 227 12 915 5 058 1 654 94 6 ,586 Weighted Parcellation Density (1970) 1.91 4.40 .13 6.50 16.73 5.88 .28 .46 2.06 .19 .25 .07 .88 .70 .13 .15 .43 .07 .96 3.73 3.00 59.39 .70 .13 8 .33 .70 .79 8.65 1.55 .02 1.46 220 APPENDIX 5 TABLE G-3 Derivation of Weighted Parcellation. Density Index For■ 30 Counties 1977 (1963 ■ 1 0 0 ) County 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Allegan Alpena Antrim Bay Berrien Calhoun Cheboygan Clare Clinton Crawford Delta Gogebic Grand Traverse Hillsdale Houghton Huron Iosco Iron Lapeer Livingston Mackinac Macomb Manistee Menominee Monroe Montcalm Newaygo Ottawa* St. Joseph Schoolcraft Sample Average ®1963 “ 1*0590 W P D 1963 * 100 Population Density (1980) 98 57 34 268 295 199 29 42 97 17 33 18 118 70 37 44 52 11- Acre Parcels (1977) 7 16,314 6,194 8,913 11,568 24,499 13,659 9,005 8,214 12,770 11,593 8,543 3,813 7,418 6,610 4,935 4,754 5,422 3,482 27,501 31,785 3,021 17,792 10,295 4,696 17,658 11,029 20,610 22,995 7,598 1,900 107 11,495 12 106 175 10 1,447 42 25 240 67 41 279 111 Weighted Parc. Density Index (1977) 3.20 1.03 1.05 11.48 20.62 6.34 .60 1.00 3.21 .58 .22 .10 3.13 1.27 .30 .42 .86 .06 7.33 16.09 5.00 88.75 1.29 .20 12.59 1.72 1.65 18.86 2.76 .02 2.89 221 APPENDIX 5 TABLE G-4 Weighted Parcellation Density* Index for 2000 AD (Projected) (1963 * 100) County 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Allegan Alpena Antrim Bay Berrien Calhoun Cheboygan Clare Clinton Crawford Delta Gogebic Grand Traverse Hillsdale Houghton Huron Iosco Iron Lapeer Livingston Mackinac Macomb Manistee Menominee Monroe Montcalm Newaygo Ottawa St. Joseph Schoolcraft Sample Average 11- Acre Parcels WPD index 2000 4.98 25,368 12,577 19,885 19,731 32,567 20,015 18,600 5,757 26,991 20,147 14,395 6,414 16,673 12,403 10,661 10,246 10,283 4,372 64,658 70,880 5,793 29,685 20,157 11,941 27,702 22,055 37,232 41,635 12,661 2,654 .67 .17 7.05 2.39 .64 .91 1.63 .07 17.24 35.88 .09 148.07 2.53 .48 19.75 3.43 2.98 34.14 4.60 .03 21,138 5.32 2.10 2.35 19.57 27.41 9.30 1.24 .70 7.57 1.0 1 *WPD for 2000 AD is based on 1980 population density. APPENDIX 5-H Trends in District and Regional Parcellation Agreages Table H-l to Table H-4 APPENDIX 5 TABLE H-l Trends in District Parcellation Amounts (11- Acres) Actual and Projected For 30 Study Counties Year District 1977 1980 1985 1990 1995 2000 1 S.E.S.L.P. 32,051 42,995 13,845 19483 94,410 109,336 124,262 139,189 2 S.W.S.L.P. 43,681 41,864 62,070 64,340 70,908 77,475 84,043 90,610 3 C.E.S.L.P. 14,219 21,638 56,592 61,083 76,216 91,349 106,482 121,616 4 C.W.S.L.P. 26,305 36,017 62,848 67,825 80,877 93,928 106,979 120,030 5 E.N.L.P. 13,062 16,425 32,214 34,247 41,087 47,927 54,767 61,607 6 W.N.L.P. 7,348 12,232 26,626 29,172 36,057 42,942 49,827 56,712 7 E.U.P. 7,536 9,419 13,464 14,374 16,491 18,608 20,725 22,843 8 W.U.P. 7,750 10,948 17,199 18,715 22,090 25,464 28,839 32,214 151,952 197,568 344,859 369,237 438,136 507,029 575,924 644,821 Total 222 1970 1963 APPENDIX 5 TABLE H-2 Trends in District Large-Lot (10-10.9) Parcellation Amounts - Actual and Projected 3,970 1977 1980 1985 1990 1995 2000 1 S.E.S.L.P. 12,267 16,726 32,199 34,634 41,753 48,872 55,998 63,109 2 S.W.S.L.P. 17,774 19,385 26,236 30,905 35,976 41,046 46,116 51,187 5,255 7,843 20,916 22,524 28,118 33,711 39,304 44,897 4 C.W.S.L.P. 12,147 15,139 31,489 33,407 40,315 47,223 54,131 61,039 District 3 C.E.S.L.P. 1963 4,415 6,522 18,457 19,828 24,843 29,858 34,873 37,888 6 W.N.L.P. 4,179 6,770 14,858 16,230 20,044 23,858 27,672 31,486 7 E.U.P. 2,500 2,738 4,753 4,939 5,743 6,547 7,351 8,156 8 W.U.P. 3,178 3,822 6,810 7,198 8,459 9,792 11,089 12,386 78,945 155,718 169,665 205,287 240,907 276,534 312,148 Total 81,715 223 5 E.N.L.P. APPENDIX 5 TABLE H-3 Trends in Regional Parcellation Amounts (11- Acre Parcels) Actual and Projected for 30 Selected Counties Region 1963 1970 1977 1980 1985 1990 1995 2000 1 E.S.L.P. 46,270 64,633 130,437 140,566 170,626 200,685 230,744 260,805 2 W.S.L.P. 68,294 78,390 116,704 123,334 140,625 157,914 175,203 192,492 3 N.L.P. 22,102 34,142 67,054 72,250 88,304 104,358 120,413 136,467 4 U.P. 15,286 20,367 30,663 33,089 38,581 44,072 49,564 55,057 APPENDIX 5 TABLE H- 4 224 Trends in Regional Large Lot (10-10.9) Parcellation, Actual. and Projected for 30 Selected Counties Region 1963 1970 1977 1980 1985 1990 1995 2000 1 E.S.L.P. 17,522 24,569 53,115 57,158 69,871 82,583 95,302 108,006 2 W.S.L.P. 28,665 32,874 51,556 57,798 68,002 78,225 88,449 98,673 3 N.L.P. 9,850 14,492 39,484 41,572 53,176 63,760 74,343 84,927 4 U.P. 5,678 6,560 11,563 12,137 14,238 16,339 18,440 20,552 APPENDIX 5-1 Subdivision Trend Equations by District APPENDIX 5-1 Trends in Acreage Under Subdivision by Districts (1969-1979) Using Predictive Equation of the Form A = bQ + b^T + E A District D-01 S.E.S.L.P. D-02 S.W.S.L.P. D-03 C.E.S.L.P. D-04 C.W.S.L.P. D-05 E.N.L.P. D-06 W.N.L.P. D-07 E.U.P. D-08 W.U.P. A = Acreage Predictive Equation 847.1 + 90.5T A 01 645.0 - 22.5T A02 473.4 - 29.6T A03 A04 A05 A06 = 1417.0 - 8 6 .4T = 1470.2 - 133.8T = 2610.4 - 218.0T A07 A08 T = Trend 152.9 + 1.7T 138.8 + 9.2T Statistics Zero Year r = .08 s = 395.8 A x = 139.0 r = -0.5 s = 157.2 A 2 = 509.4 r = 0.5 s = 187.6 5 3 = 295.5 r = -0.7 s = 402.2 54 = 898.5 r = -0.8 s = 573.6 A 5 = 667.5 r = -0.8 s = 945.6 A6 = 1302.2 r = 0.1 s = 85.1 A 7 = 162.9 r = 0.3 s = 119.0 A 8 = 193.9 (1959) 99X (1977) 851 (1985) 851 (1985) r-Significant at 951-991 225 Code (1980) 991 (1981) 991 (1977) <801 (1954) <801 APPENDIX 5-J Subdivision Trend Graphs by Region ACRES OF SUBDIVISION PLATS IN THOUSANDS ro I » M O ro -tk o O i -1 o --T 7 0 7 2 7 4 n i 30 * cn (A (~ T> 7 « 7 0 to to IMS O ro to o O » I c: -4 m -*4 a* CM O SELECTED BIBLIOGRAPHY SELECTED BIBLIOGRAPHY Ackoff, R.C. , Gupta, S.K. and Minas, J.S. Scientific Methods Optimizing Applied Research D e c i s i o n . John Wiley and Sons, Inc.; 1962. Agricultural Zoning H a n d b o o k . Staff Publication-West Michigan Regional Planning Commission, June 1978. A h l , J.G. "Use-Value Assessment in Macomb County, Michigan: Simulation Effects on Township Finances in Five Town­ ships in Rural-Urban Fringe, 1960-1969." Ph.D. D i s ­ sertation, Michigan State University, 1971. American Psychological Association Publication M a n u a l , 2nd ed. Webster's New College Dictionary, 1974. Babbie, E.R. Survey Research M e t h o d s . Co. Inc., Belmont, C a l ., 1973. Wadsworth Publishing Babcock, R.F. The Zoning Game: Municipal Practices and P o l i c i e s . U.W.P., 1966. Bair, F.H.Jr. and Bartley, E.R. The Text of a Model Zoning O r d i n a n c e , A.S.P.O., 1966. Barlowe, R. "Use-Value Assessment Legilation in the United States." Land E c o n o m i c s . Vol. XLIX, no. 2, 1973. Barlowe, R. Land Resource Economics: The Economics of Real E s t a t e . 3rd ed. Prentice-Hall I n c . , Englewood Cliffs, New Jersey, 1973. Barlowe, R. and Hostetler, J.E. "Subdivision Trends in South­ western Michigan, 1944-1958." Quarterly Bulletin of the Michigan Agricultural Experimental S tation. M.S.U., Vol. 42, no. 2, pp. 373-385, Nov. 1959. Barlowe, R. Project *80: Rural Michigan Now and in 1 9 8 0 . Land and Water Resources Research Report, no. 52. Natural Resources Dept., MSU Agric. Exp. Station and Co-operative Ext. Service, 1965. B a u m o l , W.J. and Oates, W.E. Economics, Environmental Policy and the Quality of L i f e . Prentice-Hall In c . Englewood Cliffs, N . J . , 1977. 227 228 Beegle, J.A. Net Migration by Age and Color for Michigan Counties, 1 9 5 0 - 1 9 6 0 - 1 9 7 0 . Rural Sociology Studies, No. 6, MSU Agricultural Experimental Station, 1978. Bentick, B.L. "Capitalized Property Taxes and the Viability of Rural Enterprise Subject to Urban Pressure." Land Economics 56:4 (1980):451-456. Beurscher, J.H. Land Use: Cases and M a t e r i a l s . American Case Book Series, West Publishing C o . , 1969. Borg, W.R. and Gall, M.D. Educational Research: An Intro­ d u c t i o n . 2nd e d . , David McKay Co. Inc., N e w York, 1963. Bosselman, F. The Taking Issue: An Analysis of the C o n s t i ­ tutional Limits of Land Use C o n t r o l . 1973. Callies, D.L. "State Initiative on the National Scene: The Quiet Revolution in Land Use Control" Paper Prepared for the Environmental Leader's Forum on Land Use Policy Sponsored by The New York State College of Agriculture Life Science, Feb. 1973. Case, K.E. Property Taxation: The Need for R e f o r m . Cambridge Mass. Ballinger Publishing Co. 1978. Catton, B. Michigan: A H i s t o r y . W.W. Norton and Co. Inc. N.Y. American Association for State and Local History, Nashville, 1976. Chou, Ya-Lun. Statistical Analysis with Business and Economic A p p l i c a t i o n . Holt, Rinehart and Winston, Inc., 1969. Clark, W.A.V. The Impact of Property Taxes on Urban Special D e v e l o p m e n t s . The Institute of Government and Public Affairs. The University of California, Los Angeles, 1974. Clawson, M. Suburban Land Conversion in the United States: An Economic and Governmental P r o c e s s . Baltimore and London, The John Hopkins Press, 1971. Conklin, H.E. Report on Preservation of Agriculture in an Urban R e g i o n , Northeast Regional Research, Project 90. New York's Food and Life Science Bulletin: 86 (1980). Cunningham, R. "Public Control of Land Subdivision in Michigan: Description and Critique," Michigan Law Review, 66:1, (Nov. 1967). 229 Campbell, D.T. and Stanley, J.C. Experimental and Quasi Experimental Design for R e s e a r c h . Rand McNally College Publishing Co., Chicago, 1963. De Busscher, J.F. (ed.). "Limits to Division and Sale of Land." The Michigan Surveyors N e w s l e t t e r . 10:14, (1975) Lansing, Michigan, 2-3. "Farm: On Urban Fringe, Farmers are an Endangered Species." Detroit Free Press, (Opinion), March 10, 1980, pg. 14A. "Farm Tax: How to Harvest the Profit, not the Crop." Detroit Free Press, June 17, 1980, pg. 8A. "Farmland: The Switch from Soybeans to Subdivision." Detroit Free Press, February 25, 1980, pg. 14A. Eginton, C.W. "Impacts of Federal Tax Policies on Potential Growth in Size of Typical Farms*" American Journal of Agricultural E c o n o m i c s . 62:5 (1980):929-939. Flinn, W.L. and B u t t e l , F. "Sociological Aspects and Social Consequences of Scale in Agriculture." American Journal of Agricultural Economics (Proceedings): 62:5 (1980) 946-953. Frazier, D.N. "Locational Analysis of Participants in Michigan's Farmland and Open Space Preservation Programs." MA Thesis, Michigan State University, 1979. Freilich, R.H. and Levi, P.S. Model Subdivision Regulations: Text and C o m m e n t a r y . ASPO, 1975. Frey, T.H. "Major Uses of Land in the United States: USDA Agricultural Economic Report: 440, (1979). 1974." Fletcher, J.E. "A Systematic Approach to the Analysis of Land Sales Regulation Programs: A Case Study of Michigan Land Sales Act of 1972." Ph.D. Dissertation, Michigan State University, 1978. Haris, P.E. "Land Ownership Restrictions of the Midwestern States: Influence on Farm Structure." American Journal of Agricultural Economics: 56:5, (1980), 940-945. Hayemi, Y. and Ruttan, V.W. Agricultural D e v e l opment: An International P e r s p e ctive. John Hopkins Press, 1971. 230 Heady, E.O. and Whittig, L.R. "Rural Development Problems and Potentials" in Rural Development and Land Use P e r s p e c t i v e . Soil Conservation Society of America, Feb. 1974. Herly, R.G. and Rosenberg, J.S. Land Use in the United S t a t e s . 2nd ed. RFF. John Hopkins University Press, Baltimore, 1972. Huemoeller, W.A. (and others). Land Use: Ongoing Developments in the North-Central R e g i o n . ISU Center for Agricultural and Rural Development, 1976. Huck, S . W . , Cormier, W.H. and Bounds, W.G. Jr. Reading Statistics and R e s e a r c h . Harper and Row, Publishers, N.Y., 1974. Kimball, W.J. (and others). "Michigan Citizens Speak Out on Community Problems, Preferences and Government Spending: Results of the Michigan Public Opinion Survey." Development and Public Affairs Research R e p o r t , 378; (1979). Jud, D.G. "The Effects of Zoning on Single Family Residential Property Values: Charlotte, North Carolina.” Land Eco­ nomics 56:2, (1980) 142-154. Kimball, W.J. "Land Use in Michigan: Achanging World." Extension Bulletin, 6 1 0 . Natural Resources Series, Co-operative Extension Service, MSU, (Jan. 1969). Kmenta, J. Elements of E c onometrics. Macmillan C o . , N.Y. 1971. L e h n e r t , D. "Farm Land and Open Space Preservation." Michigan F a r m e r . Feb. 2, 1980. Libby, L.W. "Land Use Policy Implications for Commercial Agriculture." American Journal of Agricultural E c o n o m i c s ; 56:5 (1974) 1143-1152. Lin, W. "U.S. Farm Numbers, Sizes, and Related Structural Dimensions: Projection to Year 2000." USDA Technical B u l l e t i n . 1625, (1980). Luttrel, C.B. "Our Shrinking Farmlands: Mirage or Potential Crisis?" Federal Reserve Bank of St. Louis R e v i e w . 62:8 (Oct. 1980). 231 Michigan State University, Graduate School of Business Administration, Division of Research, Michigan Statisti­ cal A b s t r a c t s , 1965, 1974, 1979. Mandelker, D.R. R e p o r t , no. "Transferable Development Rights." ASPO 304. Planning Advisory Service, 1975. Mentius, F.S. "Venice Township: A Study of Land Fragmen ta­ tion by 10.1 Acre Parcels, 1968-1978." Preliminary Research R e p o r t . Meshenberg, M.J. "The Language of Zoning: A Glossary of Words and Phrases." ASPO R e p o r t , no. 322, 1976. Michigan F a r m Bureau, Co-op. Ext., MSU. The Use of Zoning to Retain Essential Agricultural L a n d s , 1976. Michigan Department of Agriculture. Lansing, July 1979. Agricultural S t a t i s t i c s . Michigan Community Development Cabinet. "An Agricultural Preservation Strategy for Michigan." Report to Governor Milliken, Nov. 1980. Michigan House Legislative Analysis Section. Amended Plat." House Bill 5115, 11-2-79, Re. Steve Andrews. "Subdivisions: Sponsored by ____________________. Change of Definition of Subdivision, Bill 5567, 3-19-80, Sponsor: Rep. John Kelsley. House Michigan Laws Relating to P l a n n i n g . State Board of R e g i s t r a ­ tion for Professional Community Planners, 1978. Michigan State University. "Options in Land Use Policy for Michigan." Conference P r o c e e d i n g s , April 8-9, 1974. Michigan State University. "Towards an Effective Land Use Policy for Michigan." Conference P r o c e e d i n g s , May 17-18, 1973. National Association of County Research Foundation. D i s a p ­ pearing Farmlands: A Citizen's Guide to Agricultural Land P r e s e r v a t i o n . 1979. Nelson, R.H. Zoning and Property Rights: Analysis of the American System of Land Use R e g u l a t i o n . Cambridge, Mass., The MIT Press, 1977. 232 Neter, J. Fundamental Statistics for Business and Economics, (Abridged 4th e d . ) Allyn and Bacon, Inc., 1966. Nino, R.F. "A Comparative Analysis of Michigan's Subdivision Control Act and Critique." M.A. Thesis, MSU, School of Urban Planning and Landscape Architecture, 1970. O'Hare, W. "Michigan Population Growth and Distribution Changes." Research Report no. 331, Development and Public A f f a i r s . MSU Ag. Exp. Station Publication, 1977. Patterson, T.W. Land Use Planning; Techniques of Implementa­ tion . Lipton Educational Publ. Inc., 1979. Rathge, R.W. and Beegle, J.A. "Urban and Rural Population Change in Michigan Counties, 1960-1975." Rural Sociology Studies, no. 7, MSU Ag. Exp. Station, 1978. Santer, R.A. Michigan, Heart of the Great L a k e s . Kendall Huns Publ. Co., 1977. Scott, R.W. Management and Control of Growth: Iss u e s . Techniques, Problems. T r e n d s . Vols. 1"^ 2"J 3"! U L I , 1975. Scott, R . W . , Brower, D . J . , and Miner, D . D . , eds. Manage­ ment and Control of Growth; Issues, Techniques, Problems, T r e n d s . Vol. 3 (1975), 577-595. Snyder, J.H. "A New Program for Agricultural Land Use Stabilization: The California Land Conservation Act, 1965." Glannini Foundation Paper no. 226, Re­ print in Land E c o n o mics. Vol. XLII, Feb. 1966, no. 1. Sofranko, A.J. and Williams, J . D . , eds. Rebirth of Rural America: Rural Migration in the M i d w e s t . North Central Regional Center for Rural Development. Iowa State University, Ames, June 1980. Sommers, L . M . , ed. Atlas of M i c h i g a n . Michigan State Univers­ ity Press, 1977. State of Michigan. Annual Report of the State T r e a s u r e r . October 1, 1977 to September 30, 1978. Steel, R.G.D. and Torrie, J.H. Principles and Procedures of Statistics: A Biometrical A p p r o a c h . 2nd ed. McGraw-Hill Book Co., 1980. The Council of State Governments. A Taskforce Report on L a n d . "State Alternatives for Planning and Management." 1975. 233 Toner, W. "Saving Farms and Farmlands." American Society of Planning Officials. A Community Guide Report no. 3 3 3 . Planning Advisory Service. July 1978. Tweeten, L. and Brinkman, G.L. Micropolitan Development: Theory and Practice of Greater Rural Economic Develop­ ment . Iowa State University Press. Ames. 1976. USDA. Census of Agriculture. Preliminary Figures, 1978. United States Department of Commerce. Bureau of the Census. 1969 Census of A g r i c u l t u r e , vol. 1, part 13. Michigan County Data. USDA. Perspectives on Prime L a n d s . Background Papers for Seminar on Retention of Prime Lands, 1975. USDA. Recommendations on Prime L a n d s . From the Seminar on Retention of Prime Lands, July 1975. Ward, R.M. and Barnath, R.K. "Keeping Farmers on the Farm: Assessment of Farmland Preservation Legislation in Michigan: PA 116 of 1974." The Michigan A c a d e mician, pp. 307-319. V e r w a y , D.I. "Michigan County Income During Recession." Michigan State Economic R e c o r d . 19:6, (1977). Wright, K.T. "Michigan's Agriculture, Location and Changes in Major Products, Number of Farms and Income, County and State Data." Extension Bulletin 785. Farm Science Series. MSU Co-op. Ext. Service, Oct. 1974.