<%BANNER%>

An Analysis of Real Estate Transactions and Socioeconomic Variables on Agricultural Lands in Polk County, Florida

Permanent Link: http://ufdc.ufl.edu/UFE0041557/00001

Material Information

Title: An Analysis of Real Estate Transactions and Socioeconomic Variables on Agricultural Lands in Polk County, Florida
Physical Description: 1 online resource (106 p.)
Language: english
Creator: Drannbauer, Edward
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: agriculture, florida, gis, land, polk, socioeconomic
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Rural and agricultural planning has not received the attention that planning for the urban built environment has received. This is particularly the case for the State of Florida in the medium and long runs. While some of the population increase find their way to reside in existing housing, the shear population increase that Florida has experienced in recent years has an impact upon the non-urban environment. This study examines the impact of real estate transactions and other socioeconomic variables on agricultural lands. Polk County, Florida was chosen as the study area due to its importance to the agricultural economy, location between two major metropolitan areas, and high amount of real-estate transactions in recent years. The quantity of transactions was investigated as a predictor for determining if land was classified agricultural or non-agricultural. The socioeconomic variables explored were variables shown in previous studies as having an impact in agricultural land owner?s decision making, as well as variables that affect the macro-environment. These are population, median income, age, size of the household, proximity to infrastructure, and commute times to metropolitan areas. A Geographic Information System (GIS) was used to analyze the variables and process the data. Land was examined at the Public Land Survey System Level (PLSS), rather than the parcel or county level. A binary logistic regression model was developed to predict the likelihood that a PLSS grid would be predominantly agricultural or non-agricultural. The model determined that average transactions, average household size, and commute times were the most significant factors in determining if a PLSS grid was predominantly agricultural or non-agricultural. The model had an overall accuracy of 67.5 percent.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Edward Drannbauer.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Lehtola, Carol J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041557:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041557/00001

Material Information

Title: An Analysis of Real Estate Transactions and Socioeconomic Variables on Agricultural Lands in Polk County, Florida
Physical Description: 1 online resource (106 p.)
Language: english
Creator: Drannbauer, Edward
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: agriculture, florida, gis, land, polk, socioeconomic
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Rural and agricultural planning has not received the attention that planning for the urban built environment has received. This is particularly the case for the State of Florida in the medium and long runs. While some of the population increase find their way to reside in existing housing, the shear population increase that Florida has experienced in recent years has an impact upon the non-urban environment. This study examines the impact of real estate transactions and other socioeconomic variables on agricultural lands. Polk County, Florida was chosen as the study area due to its importance to the agricultural economy, location between two major metropolitan areas, and high amount of real-estate transactions in recent years. The quantity of transactions was investigated as a predictor for determining if land was classified agricultural or non-agricultural. The socioeconomic variables explored were variables shown in previous studies as having an impact in agricultural land owner?s decision making, as well as variables that affect the macro-environment. These are population, median income, age, size of the household, proximity to infrastructure, and commute times to metropolitan areas. A Geographic Information System (GIS) was used to analyze the variables and process the data. Land was examined at the Public Land Survey System Level (PLSS), rather than the parcel or county level. A binary logistic regression model was developed to predict the likelihood that a PLSS grid would be predominantly agricultural or non-agricultural. The model determined that average transactions, average household size, and commute times were the most significant factors in determining if a PLSS grid was predominantly agricultural or non-agricultural. The model had an overall accuracy of 67.5 percent.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Edward Drannbauer.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Lehtola, Carol J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041557:00001


This item has the following downloads:


Full Text

PAGE 1

1 AN ANALYSIS OF REAL ESTATE TRANSACTIONS AND SOCIOECONOMIC VARIABLES ON AGRICULTURAL LANDS IN POLK COUNTY, FLORIDA By EDWARD DRANNBAUER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Edward Drannbauer

PAGE 3

3 To my fiance, Carissa

PAGE 4

4 ACKNOWLEDGMENTS Over the course of my program and this work, I have been fortunate to receive guidance from a multitude of people. I would like to thank my committee chair Dr. Carol Lehtola for her encouragement and support through my doctorial studies. I would also like to thank Dr. Grant Thrall for introducing me to the fascinating field of Business Geography. Dr. Wendal Porter has provided me with much practical and sound advice throughout my program. Stanley Latimer has been invaluable in his help and assistance with GPS and GIS applications. I would also like to thank Dr. WonSuk Daniel Lee for his time and support I would also like to thank my fellow grad students, Bill Todd, Dr. Brian Peddie, and Jodi Modica for their wonderful support and friendship over the years. Last but not least, I would like to thank my lovely fiance, Carissa, whose understanding and support has helped make completing this work possible.

PAGE 5

5 TABLE OF CONTENTS ACKNOWLEDGMENTS .................................................................................................. 4 page LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 LIST OF ABBREVIATIONS ........................................................................................... 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 INTRODUCTION .................................................................................................... 13 Overview ................................................................................................................. 13 Objectives ............................................................................................................... 18 2 LITERATURE REVIEW .......................................................................................... 20 Florida Hist ory from a Land Use/Land Development Prospective ........................... 20 Early History ..................................................................................................... 20 The British and Second Spanish Periods ......................................................... 22 Florida in the United States .............................................................................. 22 Post Civil War Development ............................................................................. 24 Railroads and Florida Land Development ........................................................ 25 Floridas 1920s Land Boom and Bust ............................................................... 26 Florida in World War II ...................................................................................... 28 Post Wor ld War II Florida ................................................................................. 29 Florida from 1950 to 1970 ................................................................................ 31 The 1980s and the Savings and Loan Crisis .................................................... 32 Recovery and Stabilization in the 1990s .......................................................... 33 The 2000 to 2010 Boom and Bust ................................................................... 34 Theories of Agricultural Land Use Change ............................................................. 35 Von Th nen and David Ricardo ....................................................................... 37 Sinclair .............................................................................................................. 37 Bryant ............................................................................................................... 38 Bowler and Ilbery .............................................................................................. 38 Johnston and Bryant ........................................................................................ 38 Hart .................................................................................................................. 39 Bryant and Johnston ........................................................................................ 39 Beauchesne and Bryant ................................................................................... 40 Agricultur e at the Rural Urban Interface .................................................................. 41 Defining the Rural Urban Interface ................................................................... 41 Locating the Rural Urban Interface .................................................................. 42

PAGE 6

6 3 METHODOLOGY ................................................................................................... 50 Data Preparation ..................................................................................................... 52 Construction of Dependent and Independent Variables ......................................... 53 The Dependent Variable ................................................................................... 53 Independent Variables ..................................................................................... 55 Logistic Regression ................................................................................................ 59 4 RESULTS ............................................................................................................... 69 Descriptive Statistics ............................................................................................... 69 Model Formulation .................................................................................................. 71 5 S UMMARY AND RECOMMENDATIONS FOR FUTURE RESEARCH .................. 95 Summary ................................................................................................................ 95 Recommendations for Future Research ................................................................. 96 FLORIDA DEPARTMENT OF REVENUE CODES ....................................................... 98 LIST OF REFERENCES ............................................................................................. 100 BIOGRAPHICAL SKETCH .......................................................................................... 106

PAGE 7

7 LIST OF TABLES Table page 2 1 Urban Influence Codes (USDA ERS 2003) ........................................................ 45 3 1 List of variables. .................................................................................................. 61 4 1 Results of regression variables. .......................................................................... 74 4 2 Model Predicted versus Observed. ..................................................................... 75 4 3 Descriptive Statistics for Socio Economic Variables in NonAgricultural PLSS Grids ................................................................................................................... 91 4 4 Descriptive Statistics for Socio Economic Variables in Agricultural PLSS Grids ................................................................................................................... 92

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Sign posted on farm that borders a housing development. (Photo by author.) ... 19 2 1 Rural Urban Interface Counties (Clark 2008) ..................................................... 46 2 2 Farm Numbers in Florida 1910 2008 (USDA NASS 2009) .............................. 47 2 3 Number of Farms in Acres in Florida 1950 2008 (USDA NASS 2009) ............ 48 2 4 Average Farm Size in Acres in Florida 1993 2008. (USDA NASS 2009) ........ 49 3 1 Study Area .......................................................................................................... 62 3 2 Polk County PLSS Sections (Polk County Property Appraiser) .......................... 63 3 3 PLSS grids used for analysis. ............................................................................. 64 3 4 Agricultural Lands in Polk County (Polk County Property Appraiser 2009) ........ 65 3 5 Agriculturally Dominant PLSS Grids. .................................................................. 66 3 6 Polk County parcels with total transactions. ....................................................... 67 3 7 Shapefile of census blocks. (US Census 2008) .................................................. 68 4 1 Average Transactions by PLSS grid. .................................................................. 76 4 2 Total Transactions by PLSS grid. ....................................................................... 77 4 3 Population by PLSS grid. .................................................................................... 78 4 4 Median Age by PLSS grid. ................................................................................. 79 4 5 Male Median Age by PLSS grid. ......................................................................... 80 4 6 Female Median Age by PLSS grid. ..................................................................... 81 4 7 Males by PLSS grid. ........................................................................................... 82 4 8 Females by PLSS grid. ....................................................................................... 83 4 9 Nu mber of Housing Units by PLSS grid. ............................................................. 84 4 10 Household Size by PLSS grid. ............................................................................ 85 4 11 Median Income by PLSS grid. ............................................................................ 86

PAGE 9

9 4 12 Drive time (minutes) to Orlando by PLSS grid. ................................................... 87 4 13 Drive time (minutes) to Tampa by PLSS grid. .................................................... 88 4 14 Distance to City by PLSS grid. ............................................................................ 89 4 15 Distance to Major Roads by PLSS grid. ............................................................. 90 4 16 Polk County Real Estate Transactions 1980 to 2002 (Polk County Property Appraiser). .......................................................................................................... 93 4 17 Polk County Real Estate Transactions from 1980 to 2009 (Polk County Property Appraiser.) ........................................................................................... 94

PAGE 10

10 LIST OF ABBREVIATION S BANANA Build Absolutely Nothing Anywhere Near At All DOR Department of Revenue ERS Economic Research Service ESRI Environmental Systems Research Institute FIRREA Financial Institutions Reform, Recovery, and Enforcement Act FGDL Florida Geographic Digital Library GIS Geographic Information System GPS Global Positioning System PLSS Public Land Survey System NAR National Association of Realtors NASS National Agricultural Statistics Service NIMBY Not In My Back Yard OLS Ordinary Leas t Squares Regression PARID Parcel Identification Number PASW Predictive Analytics Software SPSS Statistical Package for the Social Sciences TND Traditional Neighborhood Design UIC Urban Interface Code USDA United States Department of Agriculture

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AN ANALYSIS OF REAL ESTATE TRANSACTION S AND SOCIOECONOMIC VARIABLES ON AGRICULTURE LANDS IN POLK COUNTY, FLORIDA By Edward Drannbauer May 2010 Chair: Carol Lehtola Major: Agricultural and Biological Engineering Rur al and agricultural planning has not received the attention that planning for the urban built environment has received. This is particularly the case for the State of Florida in the medium and long runs. While some of the population increase finds their way to reside in existing housing the shear population increase that Florida has experienced in recent years has an impact upon the nonurban environment. This study examines the impact of real estate transactions and other socioeconomic variables on agricultural lands. Polk County, Florida was chosen as the study area due to its importance to the agricultural economy, location between two major metropolitan areas, and high amount of real estate transactions in recent years. The quantity of transactions was investigated as a predictor for determining if land was classified agricultural or nonagricultural The socioeconomic variables explored were variables shown in previous studies as having an impact in agricultural land owners decision making, as well as variables that affect the macro environment These are population, median income, age, size of the house hold, proximity to infrastructure, and commute times to metropolitan areas. A Geographic Information System (GIS) was

PAGE 12

12 used to analyze the variables and process the data. Land was examined at the Public Land Survey System Level (PLSS) rather than the parcel or county level. A binary logistic regression model was developed to predict the likelihood that a PLSS grid would be predominantly agricultural or nonagricultural. The model determined that average transactions, average household size, and commute times were the most significant factors in determining if a PLSS grid was predominantly agricultural or nonagricultural. The model had an overall accuracy of 67.5 percent.

PAGE 13

13 CHAPTER 1 INTRODUCTION Overview Agriculture has always been a vital aspect of the Florida economy. Floridas agricultural production is ranked at or near the top for a number of agricultural commodities. In 2007, Florida was ranked first in the US in value of production for oranges, grapefruit, tangerines, sugarcane for sugar, snap beans, fresh market tomatoes, cucumbers for fresh market, squash, bell peppers, watermelons, and sweet corn (USDA, 2009) Floridas agricult ural and natural resource industries had a total output of $35 billion, employment of over 336,000 persons, and exports of $20.7 billion in 2000 (Hodges and Mulkey, 2003) However, increasing population and development has resulted in enormous pressures being placed on agricultural and natural resources throughout Florida. Such pressures are particularly evident in rural a reas (both remote rural and urban proximate locations). As a result, local extension agents and decisionmakers need to better understand how this rapid growth affects their environmental resources and local quality of life from the perspective of practic ing farmers. Similarly, these interests need to better comprehend the role that local community based efforts can play in environmental protection, land use planning, sustainability, community development, and natural resource management. In light of curr ent growth trends, it is vital that local communities be prepared to respond to changing needs and increasing demands for resources. Farms across the U.S. are facing these same issues, but Florida is in a more urgent situation. The recent economic recession has slowed Floridas population growth. However, for many years the population growth was significantly higher than the

PAGE 14

14 entire United States (Colburn and DeHavenSmith, 2002) Florida is known throughout the country as the premiere vacation spot and retirement location. Favorable climate, beautiful beaches, a lack of a state income tax, and available real estate has been an i ncentive for many to establish permanent residence in the state. Between 1990 and 2000 the United States population grew by 13.1%, whereas Florida experienced growth at 23.5% (Census Florida, 2005). Additionally, between 2000 and 2003 Floridas growth rat e continued to rise and was almost double that of the United States (Census Florida, 2005). Between 2007 and 2009, many counties that experienced high growth have started to see a population decrease, but for the state overall, population has still been increasing, although at a slower rate than during the housing boom (Keen, 2008) Between 2007 and 2008, Floridas population increased by 0.7% and ranks fourth in the nation for population size (Bureau, 2008) Floridas weather also has had a major impact on agriculture. Until recently, hurricane Andrew was known to have the greatest economic impact on Florida agriculture, resulting in damages of over $30 billion (Attaway, 1999) Damage from hurricane Katrina to New Orleans is estimated around $100 billion (Schneph and Chite, 2005) At the same time, agricultural landscape has been rapi dly decreasing. How much of this decreasing of agricultural land is due to urbanization, and how much is due to disasters is not known. However dealing with changing weather conditions is not something new to farmers. Similarly, the effect of a disaster on agricultural land values (as opposed to crop values) is unknown. In spite of a major recession in 2008, (what some economists are referring to as The Great Recession) Florida still remains a major destination for retirees. The

PAGE 15

15 external forces of popul ation growth and weather disasters have created enormous pressure on the Florida farmer. Disasters have been destroying crops, property and equipment. Urbanization has been taking farmland away. There are many reasons for concern about these external forces: Rapid increase in both U.S. and world population are generating corresponding increases in domestic and foreign demand for U.S. agricultural products. Conversion of agricultural land to urban development will make it more difficult for advances in agri cultural productivity to keep pace with demand. Once land is converted to urban use, it becomes very difficult and costly to convert back for nonurban use. An adequate supply of quality agricultural land is important to the economic, social and environmental wellbeing of the citizens of the U.S. and the world (Olson and Lyson, 1998). Important functions of agricultural land include not only commodity production, but air and water quality, wildlife habitat, waste assimilation, biodiversity, climate modifi cation, aesthetics, and community (Olson and Lyson, 1998). Prime agricultural land in Florida also happens to be some of the most desirable land for urban development. As this developed land attracts a greater population, the effects of a disaster are more severe when one does occur. The majority of community planning has been focused on the urban environment. Rural and agricultural planning in Florida has hardly been investigated, particularly in the medium and long run. Understanding how these forces affect opportunity costs is particularly important in agricultural land issues because choices made today will clearly affect the options available to future decision makers. A current resource owner will choose among relevant land use options based on expected returns through a reasonable planning horizon (Austin et al., 1998) Future returns are usually worth less than present returns because there is some amount of risk. On the other hand, those who do not own the land under consideration (e.g. land use planners, land speculators) will likely have different values for present and future land uses. Because the opportunity costs of

PAGE 16

16 foregoing present gains for future returns are less for nonowners than for owners and the potential costs of the loss of future land use options are higher, conflict between owners and non owners is inevitable (Furuseth and Lapping, 1999) Newcomers to the state, and even those that already reside here, may not be seriously considering the many issues and challenges that ac company the benefits that a Florida lifestyle affords them. However, the many opportunities that growth brings also put stresses on the existing infrastructure, environment, and the concerned residents trying to manage growth carefully (Mackin, 2006) In the recent decade, Florida has been undergoing a unique pattern of growth never seen before in U.S. his tory. The baby boom generation that started after the end of World War II is reaching retirement age. Florida, a state with a warm climate, has been a prime retirement area, as well as a place for a second (and even third) homes for many. Additionally, F loridas southern location has made it a prime immigration location for people from South America. These demographic changes have created a demand for Florida housing and placed increased pressure on agricultural lands. These changes have also caused unique patterns of growth. Traditionally, growth occurs around established urban clusters, such as a metropolitan area (Miles et al., 2000) Growth in some Florida counties has derived from very large single planned developments which border no such cluster. The Villages, located in Sumter County, is one such example. Rural communities on the urban fringe are facing pressures not typically encountered in the strictly urban environment. These challenges include the social nature of the area, the traditionalism of rural societies, and a lack of knowledge and

PAGE 17

17 experience associated with planning for new development patterns. Often these communities face a choice of trying to preserve a tight knit community who places more value on open space and the presence of natural amenities, or put more emphasis on the denser, intense growth of economic development (Brennan, 2004) Lost in this shuffle between land developers, politicians, and lawyers is the individual farmer ; the one person whose final decision determines if food or fiber will be produced. Florida agriculture attrition has developed in the following way (Molnar, 1986) : 1. Investor s and speculators, following demographic trends, estimate the future value of land. 2. This land is then purchased for development, or held for future appreciation to resell (i.e. flipping). 3. The developed land is appraised at a higher value than it was before, leading to higher tax assessments. 4. As urban development takes root, complaints concerning the noise, odors, and unsightly aesthetics of farming start to have an increased bearing on the politics of the area. 5. Farmers and large land owners, under increased pressure, eventually sell their land and move. 6. Realizing they dont have the market area to sustain operation, supporting industries of agriculture (e.g. the tractor dealer) will begin to downsize, eventually leaving. Businesses not directly related to agriculture also downsize. 7. Without the supporting industries, it becomes extremely difficult for any remaining farms to continue operation, or for new farms to start up. Although it has been long realized that increased development results in fewer farms, the actual point when this occurs and under what circumstances is little understood. The farmer, as a decision maker, has many challenges to face. One challenge is the increasingly prevalent political attitudes of NIMBY (Not in My Back Yard) and

PAGE 18

18 B ANANA (Build Absolutely Nothing Anywhere Near at All), as more people move into the farm areas. The photo shown in figure 11 is an example of this. Ironically, the people expressing sentiments such as NIMBY and BANANA are actually the very people who spawned development in the first place. During the later 1990s and early 2000s a strong economy, rising affluence, easy access to credit, and low gasoline prices helped spawn the demand for housing and commercial sites in the urbanrural fringe. Many farmer s themselves, taking the if you cant beat em, join em approach, subdivided their own land and developed lar ge home lots. Objectives Identifying the underlying socioeconomic variables that affect the farmers decision to continue to farm is the overall goal of this study. Specifically, this study has three main objectives: 1. Investigate the relationship between the quantity of real estate transactions and agricultural land. 2. Identify and investigate socioeconomic variables significant to agricultural land use change. 3. Develop a model to estimate agricultural land use change, and where that change is likely to occur.

PAGE 19

19 Figure 11. Sign posted on farm that borders a housing development. (Photo by author.)

PAGE 20

20 CHAPTER 2 LITERATURE REVIEW Florida Histor y from a Land Use/Land Development Prospective Rural sociologists and agricultural geographers have long argued that an areas unique history and culture have an important impact on rural areas (Luloff and Swanson, 1990; Brennan, 2004) These variables have often been ignored in the study of agricultural land use, or seen as unimportant next to tangible variables, such as soil type or distance to market (Grigg, 1984) Economists are quick to argue that farmers, like most people, are profit maximizers. Recent research suggests however, that history and demographics have more of an impact in farmer decision making as was previously thought, particularly in industrialized counties where survival is no longer the main objective (Brown et al., 2003; Brennan, 2004) For this reason, a short history of Florida from a land use/land development prospective is presented here. Early History Land development and agriculture in Florida is an interesting story. The typi cal American high school textbook generally marks the beginning of the United States with the landing of the English at Jamestown in 1607 and Plymouth Rock in 1620 (Maddox 2003) By the latter date Florida had long since been explored and settled and St. Augustine was already fifty years old. Florida has been part of Spain longer than it has been part of the United States and will not catch up until 2055 (Cresap, 1982) Since the first Spanish settlers arrived in Florida during the 15th century, agriculture has been vital to the economy of the state. A central part of the land development of Florida was its tropical climate, which was beneficial to land development i n agriculture almost as many times as it was a

PAGE 21

21 detriment (Knotts, 2003) Agriculture has thrived in Floridas warm winters and year round growing season, but has suffered huge losses due to frequent storms and hurricanes. Accounts from Spanish records are filled with accounts of storms wiping out fleets of ships carrying men and supplies from Florida (Attaway, 1999) Hurricanes and harsh weather in particular made early development of the state very difficult. Throughout most of Floridas first Spanish period (1565 1763) the population had to depend on an economic subsidy, known as the situado, sent each year from Mexico City (Cresap, 1982) This annual allowance consisted of specie (cash), food, and sup plies. Often however, this was pilfered or delayed. In order to make the region self supporting, King Philip II of Spain decided to appoint Pedro Menendez de Aviles the new governor, and hence the first land developer of Florida (Cresap, 1982) Under co ntract issued by the King, Menendez was to settle 500 men in Florida, one hundred of them to be farmers. Florida agriculture essentially started with these men, as well as 100 horses, 200 calves, 400 swine, 400 sheep and some goats, and other livestock the colonists wanted (Cresap, 1982) The contract also stipulated for 500 slaves, however, Menendez was besieged by so many men that wanted to go with him to the new land, that he was not able to bring the 500 slaves the contract stipulated. Menendez fou nded the city of St. Augustine on September 8, 1565. He named it St. Augustine, because it was on that saints feast day, August 28th, when he had sight of Florida (Maddox 2003) Despite some early success with establishing the first permanent European settlement and defeating French attacks, very little else went well for the Spanish. Farming in Floridas sandy soils proved difficult. Native Indians typically

PAGE 22

22 made at tacks on farmers in the fields outside of the fort walls. There was often fraud by corrupt minor officials and numerous mutinies by troops (Cresap, 1982) In spite of the numerous setbacks, there was some success. Franciscans arrived in 1573 and built a successful mission system about 204 years before the first mission in California (Cresap, 1982) These missionaries instructed the Apalachee Indians around present day Tallahassee in the cultivation of maize, wheat, beans, pumpkins, tobacco, and cotton. Tobacco and cotton were valuable trade crops (Cresap, 1982) A thriving cattle industry developed near present day Gainesville. Cattle driven from La Chua (now known as Alachua), crossed the St. Johns River by flatboat to St. Augustine where the m eat was sold for local consumption and the hides were tanned for shipment overseas (Taylor, 1995) The British and Second Spanish Periods Duri ng the French and Indian War (17541763), Spain, sided with France. England quickly and easily seized Havana, one of Spains richest ports in the region. In order to regain this rich port possession, Spain had to trade all of Florida (Dodd, 1957) France lost all of its North American territories, including Canada. Florida was now English, and the treaty provisions placed the western boundary of Florida at the Mississippi River (Cresap, 1982) Because of its vast size, Florida was divided into two separate colonies: East Florida, with a capitol at St. Augustine, and West Florida, with a capital at Pensacola (Cresap, 1982) The Apalachicola River was the boundary between the m. This made Florida the fourteenth and fifteenth British colonies (Maddox, 2003) Florida in the United States Florida became a territory of the United States in 1821 and the 27th state on March 3, 1845 (Maddox, 2003) St. Augustine and Pensacola were the two largest

PAGE 23

23 cities in the new territory and the territorial governor appointed a commission to select a site for a new capital midway between the two cities (Maddox, 2003) Tallahassee was established as the capital in 1824 (Cresap, 1982) A dvances in transportation technology had a profound impact on the development of agriculture in the state. The development of the steamboat and its 1827 arrival in Florida opened the interior for commerce and tourism (Cumming, 2006) The first east coast steamboat arrived from Savannah on the St. Johns River at the then small town of Jacksonville in 1831 (Cresap, 1982) When the Seminole Indians were defeated and pushed south into the Everglades after the Second Seminole Indian war (1835 to 1842), steamboats brought more pioneers and settlers to the interior (Cresap, 1982) Florida agriculture was important to the Confederate economy during the Civil War (Taylor, 1995) but toward the latter half of the 1800s, Floridas plantation system began to fail due to failing crops, declining prices, cash shortages, and the attack of the boll weevil (Miles et al., 2005) Many of these failed cotton plantations were purchased by Northerners (refer red to by Southerners as carpetbaggers) at low prices (Cresap, 1982) The purchase of the Louisiana Territory by the U.S. from France led to the hazardous Florida Straits bec oming one of the busi est shipping routes in the world (Taylor, 1995) Key West at Floridas most southern tip grew into Floridas largest town by 1850, as well as one of the richest towns in the U.S. on a per captia basis (Dodd, 1957) In 1862, Congress passed the Homestead Act, enabling settlers who did not a lready own sufficient land to be granted title to 160 acres for each adult in the family

PAGE 24

24 simply by living on and improving the homestead for a period of 5 years (Cresap, 1982) Post Civil War Development Few battles occurred in the state, but Florida played a vital role supplying Confederate forces (Taylor, 1995) After the war, a tourism industry began to develop in addition to agriculture. This early tourism industry created the unique situation of pioneers and settlers catering to tourists. In 1878, Hullam Jones glued a window to the bottom of a dugout canoe and invented the glass bottom boat and Floridas first theme park, Silver Springs (Cresap, 1982) According to Census records, Floridas 1880 population was 269,000 people that resided mainly in Jacksonville, Key West, Pensacola, Tallahassee, and Tampa. In 1880 the State of Florida was nearing bankruptcy and faced the loss of credit to keep the States business running. Florida did have one asset; 20,000,000 acres of sw amp and overflowed land granted to it by the Federal Government under the Swamp and Overflowed Lands Act of 1850 (Taylor, 1995) This act enabled the transfer of title of federally owned swampland to private parties agreeing to drain the land and turn it in to agricultural and other productive uses The availability of a large amount of cheap land led to a land boom in the 1880s. One purchaser of a large amount of land included Hamilton Disston a P hiladelphia saw manufacturer and financier who purchased 4,000,000 acres or oneninth the state for $0.25 per acre (Cumming, 2006) The bulk of Disston holdings ranged south from what is now Kissimmee to, and surrounding, Lake Okeechobee. Disston also owned about 150,000 acres in what is now Pinellas County (Cumming, 2006)

PAGE 25

25 Disston establi shed his field headquarters in Kissimmee and began a massive drainage operation in the Kissimmee River basin. New towns such as Kissimmee City, Narcoosee, Runnymede, and Southport were built on former swampland and other towns such as Ancolote, Fort Myers, and Tarpon Springs were also developed. Development was booming and in the winter of 1884 northerners arrived in the area including Thomas Alva Edison, who decided to bring a prefab winter house complete with south Floridas first swimming pool to Fort My ers. Longtime Edison friend Henry Ford followed with a house of his own next door (Cumming, 2006) Florida was quickly becoming known as a favorable location because of its climate and perceived health effects. By 1893 the Disston Land Company had an estimated 1,200,000 acres of reclaimed land and lowered the level of Lake Ok eechobee by about four and onehalf feet (Dodd, 1957) Railroads and Florida Land Development The development of the railroads brought the next wave of major changes to the Fl orida landscape. Commercial agriculture up until this point relied upon rivers such as the St. Johns and Suwannee (Cresap, 1982) Tracks could reach more places than could be reached by navigable waterways, making more land available for development. T he State of Florida wanted to encourage development and as an incentive offered large amounts of vacant land for each mile of railroad that was built (Dodd, 1957) Floridas two major railroad developers were Henry Flagler and Henry Plant Flaglers Florida East Coast Railway developed lines along the east coast, eventually reaching all the way to Key West, while Plant extended his railway along the west coast (Cumming, 2006) As the railroads moved further south down both coasts resort hotels

PAGE 26

26 were built and Florida grew rapidly as a destination resort. Notable hotels included the opulent Spanish Renaissance style Ponce de Leon Hotel, 1888 (now Flagler College), the Cordova, 1888 (now Casa Monica Hotel), the wood frame Royal Poinciana Hotel, 1894 (Palm Beach), and the wood frame Royal Palm Hotel, 1896 in Miami (Cummin g, 2006) West P alm B each was founded in 1894 as a community to house the construction crews working across Lake Worth on Palm Beach Island and eventually the servants working in Palm Beachs two grand hotels the Royal Poinciana and The Breakers both built in 1896(Cumming, 2006) Resorts built by Plant included the Port Tampa Inn in 1888, the Moorish style Tampa Bay Hotel in 1891 (now the University of Tampa), Hotel Belleview in 1897 (now Belleview Biltmore Resort Hotel in Belleair), as well as the Fort Myers Hotel, Hotel Punta Gorda, Kissimmee Hotel, Ocala House, and The Seminole Hotel in Winter Park (Cumming, 2006) Floridas 1920s Land Boom and Bust The 1920s were one of the most prolific economic periods in American history. Availability of installment credit for automobiles and other consumer products soared during this time. B anks offered Americans their first home mortgages, resulting in record numbers of singlefamily homes being built (Brennan, 2004) Florida entered a period of frenzied real estate activity and speculative excess (Cum ming, 2006) Not unlike the recent years of 20012006, s peculators began to buy land to sell for increasing profit within a few months. Real estate values began to inflate rapidly along with intense building activity This even led to the failure of s everal banks in Massachusetts due to numerous depositors withdr awing their savings to buy Florida real estate (Cumming, 2006)

PAGE 27

27 Real estate in the southeastern part of the state skyrocketed. Miami Beach land increased in value 1,800 percent and by 1926 boasted 178 apartment buildings, 8 casinos, 2 churches, 308 commercial b uildings, 3 golf courses, 56 hotels including 4,000 rooms, 3 movie theater s, 858 private houses, 4 polo fields, and 3 schools (Dodd, 1957; Cumming, 2006) Hollywood, Florida, known as the Atlantic City of the South included about 2,420 houses, 36 apartment buildings, 9 hotels both completed and under construction, 252 commercial buildings, 18,000 acres of land, and six andonehalf miles of oceanfront property. The city grew to around18,000 residents There was also the well known Hollywood Beach Hotel, which rose seven stories and included 500 rooms with private baths It contained the world's largest solarium, and boast ed a private wire connection direct to the New York Stock E xchange for use by hotel guests (Cumming, 2006) Fort Lauderdale, known as the, Venice of America reached a population of 16,000 by 1926 (Cumming, 2006) The Florida land boom ended after the arrival of the Great Miami Hurricane on September 17, 1926 (Attaway, 1999) This devastating category 4 hurricane caus ed significant damage in the Miami area, leaving only a few buildings standing. Normalized for inflation and wealth effects in 2005 dollars, the Great Miami Hurricane was the costliest in history, at $157 billion dollars (Pielke et al., 2008) This is greater than even Hurricane Katrina in 2005 ($100 billion), or Andrew in 1992 ($55.8 billion). As a result of this hurricane, the first building code in the U.S. was initiated, which was duplicated by more than 5 000 cities across the country (Alexander, 2000)

PAGE 28

28 Floridas troubles did not end there. The eighth costliest hurricane in U.S. history ($33.6 billion) hit Florida two years later in 1928 (Pielke et al., 2008) The Okeechobee hurricane, the first recorded hurricane to reach Category 5 status on the Saffir Simpson Hurricane Scale, was also one of the deadliest. Fatalities were estimated at over 4,075, more than double the estimated fatalities of Hurricane Katrina of 1,836 (Attaway, 1999) These events led to a swift and brutal end to Floridas land boom. The rail system also failed causing costly construction delays on the east coast and Floridas reputation was damaged by proliferating promoters and conm en (Cumming, 2006) Floridas citrus industry was also damaged by t he Mediterranean Fruit Fly (Cresap, 1982) These events led to the collapse of h undreds of construction projects and developments around the state, resulting in thousands of problem properties that needed to be objectively appraised (Cumming, 2006) Florida entered the Great Depression three years ahead of the rest of Ameri ca and would not recover until the start of World War II (Cumming, 2006) However, there was one exception, the southern tip of Miami Beach, was one of the few places in the world that actually experienced a building boom during the Great Depression (Cumming, 2006) During this period of time hundreds of small Art Deco style apartment buildings and hotels were built for which Miami would be famo usly known (Dodd, 1957) Florida in World War II During World War II Florida became one large military training camp with 172 military installations by the end of the war in 1945 (Cumming, 2006) Many hotels and large resort hotels were converted t o military training barracks, or hospitals (Ramsey, 1975) The Biltmore Hotel and Country Club (Coral Gables), Breakers Hotel (Palm

PAGE 29

29 Beach), Don Cesar Hotel (St. Petersburg), and the Hollywood Beach Hotel (Hollywood) among others were converted for use as military hospitals (Cumming, 2006) Many of Floridas current major airports, as well as a large number of small general aviation airports started out as U S Army Air Fields (Seese, 1998) These were: Bartow Air Base, Boca Raton Army Air Field (now Florida Atlantic University), Brooksville Army Air Field, Cecil Field Naval Air Station (Jacksonville), Dale Mabry Air Field (Tallahassee), Drew Field (now Tampa International Airport), Henderson Field (now Busch Gardens and an industrial park in Tampa), Homestead Army Air Field, Lakeland Army Air Field (Drane Field Lakeland Linder Regional Airport), Morrison Field (West Palm Beach), and the Venice Army Air Base (Venice Municipal Airport) (Seese, 1998) Agriculture, although critical to the war e ffort at this time, saw a decline in the number of farms as labor and transportation shortages made running agricultural operations difficult (Ramsey, 1975) As a result, many Americans planted Victory Gardens as many food items and canned g oods were rationed (Harris and Bronson, 2002) Post World War II Florida At the end of World War II, farm numbers in Florida did show a slight increase as population in the state greatly increased as soldiers returned from the war. Many people stopped planting Victory Gardens and there was a strong demand for fresh fruits and vegetables (Gannon, 1996) The demand for new households greatly exceeded the demand for new farms, and farm numbers started to decline sharply over the next 20 years (USDA, 20 09) New households on the other hand started forming at a rapid rate. This was fueled by U.S.

PAGE 30

30 government policy that made purchasing a new home easier (Gottmann and Harper, 1966; Thrall, 2002) Men and women serving in World War II, and shortly after in the Korean War, were entitled to low interest VA loans for purchasing a house (Hudgins and Cromley, 1997) Because the federal government was making the loans, the home had to conform to federal government specified building codes. Many of these codes were not in effect prior to WWII, and therefore much of the housing constructed before WWII did not conform to the new codes (Thrall, 2002) Mu ch of the housing stock built before WWII thereby became obsolete. The purchaser of such a home, even if not bought using government loans, would run the risk of not having a future buyer. The result was a decline in many older American neighborhoods, and an urban expansion into farmland (Thrall, 2002) Florida saw a great number of farm workers move to high economic growth regions such as California and the Great Lakes industrial belt (Zieger, 1997) As Florida moved through the late 1960s, 1970s, and 1980s there started to be greater changes (Platt and Macinko, 1983) The American middle and upper middle class family had changed from the Victorian era extended family to the stereotypical nuclear family of the 1950s, to the far more varied family of today (Zieger, 1997) Today a family is inclusive of single parent households, unmarried couples with or without children, and persons living alone (Thrall, 2002) A suburban couple with two children may become divorced and sell their suburban home. One spouse may move back with his or her parents, and the other may buy a smaller condominium nearer to work or downtown. With the loss of traditional family households, so goes the traditional family farm.

PAGE 31

31 Florida from 1950 to 1970 The next general real estate boom that began in the mid1950s and lasted until the early 1970s was a period of time often known as Floridas Golden Age of Land Scams. From 1954 to 1982 about onethird of Floridas farm land was lost to development (USDA, 2009) In th e 1960s more Florida land was subdivided into lots than in the rest of the country combined (Cumming, 2006) Large tracts of land were subdivided and aggressively sold by various large and small companies, utilizing the latest technology of the day to attract buyer s. Many of the land promoters offered free trips to Florida w here guests were treated to resort hotels that had the electronics of an embassy and the ambiance of a car dealership The hotel rooms often had secret intercom systems so that the salesmen could listen to prospects private conversations. The land sales mantra was wine, dine, and solidify (Cumming, 2006) Typically, lot s were sold on an installment basis, with contracts in favor of the seller, who was able to repossess a lot if the buyer missed a payment. Other key drivers that led to the growth in land sales in Florida included a rising level of affluence, widespread availability of air conditioning, as well as the interstate highway system (Morgan and Munton, 1972) T he Interstate Highway Bill of 1956 created a nationwide highway network of 47,000 miles This included Interstate 4, Interstate 10, Interstate 75, and Interstate 95 in Florida. S imilar to the impact of the railroads in the second half of the 18 00s the new highway system fueled an unprecedented w ave of development that transformed much of Florida (Gannon, 1996) Downtown shopping districts declined in the 1960s due to the growing use of the automobile and gave rise to suburban shopping centers with large convenient parking lots. T hese new enclosed shopping centers became super

PAGE 32

32 regional shopping centers with substantial ancillary commercial development (Hoyt, 1970) By the late 1960s rural land was being subdivided at an unprecedented rate and this continued until the end of 1973, when the industry encountered several problems This included the Arab Oil Embargo and rapidly rising energy prices, higher interest rates, an economic recession, increased development costs, more rigorous consumer and environmental regulations, massive overbuilding, and negative publicity. Th e 1960s space race led to rapid development around Cape Kennedy from Titusville south to Melbourne. During this time there was a major expansion in the tourism industry The Walt Disney Company began to acquire tracts of land southwest of Orlando (Gannon, 1996) Disney World s Magic Kingdom opened in 1971 ; EPCOT, 1982; Disney MGM Studios, 1989; and Animal Kingdom, 1998) which was followed by Seaworld in 1973 (Cumming, 2006) Ringling Brothers Barnum & Bailey Circus developed Circus World in 1974 at the southeast q uadrant of I 4 and U.S. Highway 27 in Haines City, Florida. Circus World, under new ownership, was redeveloped in 1987 as Boardwalk & Baseball, only to close in 1990 (Cumming, 2006) Eventually the last of the big three central Florida theme parks, Universal Studios opened in the late 1980s (Gannon, 1996) The 1980s and the Savings and Loan Crisis During the very late 1970s and early 1980s U.S. saw high inflation rates, and real estate was viewed as a good hedge against inflation (Cumming, 2006) De regulation of the savings and loan industry in the early 1980s led entities known as Sav ings and Loans to begin lending to commercial real estate developers which essentially gave them the same power as banks, without the same regulation banks had (Fite, 1986)

PAGE 33

33 During this period loans were giv en out freely, and a tremendous amount of over building occurred(Phillips, 1986) As a result of this over building, m any developers and investors could not repay their loans to the savings & l oans and troubled real estate projects proliferated. This decimated t he savings & loan industry with the failure of over 1,000 savings & loan associations (Cumming, 2006) Recovery and Stabilization in the 1990s The result of the 1980s building boom and subsequent bust was the S avings & L oan C risis. T he Savings and Loan Cris is was stabilized by a bailout by the Federal government in 1989 through the Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA). FIRREA led to the creation of the Resolution Trust Corporation, which liquidated the assets of the failed S avings and Loans (Moeti and Khalo, 2007) Gradually, t he commercial real estate market in general began to recover in 1995 and strengthened until 2000 at a moderate pace. During this time, New Urbanism and the Traditional Neighborhood Design movement (TND) that started in the 1980 began to gain momentum, with the first fully new urbanist town built in Seaside, Florida (Cumming, 2006) This development created an idyl l i c pedestrian oriented live work shopcaf casual beach front resort environment all within a mile radius of the town center Seaside and similar developmen t s strive to provide a sense of place. In 1997 The Truman Show was filmed in Seaside. New Urbanism, or TND developments are t ypically upscale niche neighborhoods. Examples include: Abocoa (near Jupiter), Avalon Park (near Orlando), Baldwin Park (Orlando), Celebration (Disney), Haile Village

PAGE 34

34 Center (Gainesv ille), Rosemary Beach (Seasides panhandle sister town), and Wellington (Palm Beach) (Cumming, 2006) This trend required large areas of land (Babcock e t al., 1985) The 2000 to 2010 Boom and Bust After the September 11, 2001 (911) t he Federal Reserve rapidly lower ed the Federal Funds Rate, sending mortgage interest rates into a nose dive dropping to a 40 year low Demand for houses, condominiums, r emodeling & renovation, and home furnishings began to spike sharply upward in 2002 (Murley et al., 2006) A large amount of the investment capital that would normally flow into the stock market was diverted to the real estate market having the greatest impact on the housing market, particularly the condominium market sector (Taylor and Brester, 2005) H owever still impacting the commercial and investment sector by driving capitalization rates down and increasing v alues with very limited increases in property net operating income (Cumming 2006) Flipping became a popular way to make money with real estate, which drove up the price of property, and led to a construction boom (McGinnis, 1982) Adding to the new construction boom was the massive conversion of rental apartments to condominiums. The conversion of rental units to condominiums contributed to an oversupply of the market. A study by the National Associat ion of Realtors (NAR) found that about 23% of all American houses bought in 2004 were for investment, or speculation. Another 13% of houses were bought as second homes. Investors were prepared to buy houses and condominiums and rent out at a loss, just bec ause they thought prices would keep rising (Miles et al., 2005) Hous ehold incomes have not kept pace with the runup in housing prices creating a fundamental imbalance as more consumers stretch to buy with aggressive financing

PAGE 35

35 techniques including adjustable rate mortgages (ARMs), interest only mortgages, and zero down payment mortgages (Cumming, 2006) Theories of Agricultural Land Use Change Understanding evaluation of land prices, particularly agricultural land, can often be a complex process. Technological changes, farmer demographics, and increased globalization has changed the landscape of agriculture, particularly over the past 50 year s. This is particularly true in industrialized nations. Land use change models have not kept up with technical changes in agriculture (Ilbery, 1985; Ahuja et al., 2002) The economists models have typically been concerned with the notion of profit or utility maximization. Traditional economic theory suggests that people make decisions based on the expected change in their level of wellbeing, where the technical term for well being is utility (Symons, 1979) A n increase in an individuals util ity may relate to any number of events, such as a pay r ai se, the purchase of a new stereo, an improvement in health or an improvement in local biodiversity. However, the effect of the same event on increases (or decreases) in utility varies between people and between times in an individuals life so not all events bring the same level of increase in utility every time they occur. Utility is a very useful concept, which enables economists to model behavior in a conceptual way O ne of the main assumptions of traditional economic models is that humans seek to maximize their levels of utility. However utility is extremely difficult to use in a more practical manner largely because it is very hard to measure in real world situations. Recognizing this challenge, economists assume that money can act, albeit imperfectly, as a measure of the extent to which the utility of individuals is affected

PAGE 36

36 by any given change in circumstance. This reasoning leads to the situation observed in many agricultural economic model s where farmers are assumed to maximize utility and where profit is used as a substitute for utility (Castle et al., 1987; Thrall, 2002) Underlying assumptions of these models include that decision makers are rational, have complete information, and the ability to use this knowledge. Such assumptions have often been criticized as being unrealistic (Castle et al., 1987) Farmers, like other decision makers, cannot make perfect economic decisions and must act based on perceived conditions in an uncertain environment. Conversely, satisficer models have been proposed as a more real world alternative. Such models attempt to take into account the motivations, aspirations and attitudes of farmers (Wolpert, 1964; Ilbery, 1985) Being able to understand and explain farmers adaptations to these changes would enable a better understanding of an individual farmers decision in a broader economic system. Ilbery (1985) characterized different approaches to modeling agricultural land change into four main categories: 4. Geographic determinism models these assume that it is the physical environment that acts in a deterministic manner and controls agricultural decision making 5. Economics determinism models these assume that the economic factors of market, production, and transport costs on groups of homogenous producers, who in turn react in a rational manner to them. 6. Socio personal determinism models these assume that there are further set s of influences which affect agricultural decisionmaking, including farmers values, aim s, motives, and attitu des towards risk. 7. Radical models these assume that new technology and the increase in agribusiness over traditional farms are not signs of agricultural progress. Such an approach advocates that the future of agriculture in post industrial societies is in the formation of economically viable farm communities.

PAGE 37

37 The approaches above described by Ilbery are also indicative of the how the focus of interest in the various approaches have changed from post World War II through today. These theories are vast and s ometimes opposing. Von Th nen and David Ricardo The theories of two 18th century economists, David Ricardo and Johann Heinrich von Th nen, are classic examples of the dichotomy of agricultural models. Ricardos economic theory was based upon the relative productivity of the land, or the original and indestructible power of the soil (Ricardo, 1817) Fertile land was more expensive than barren land. Ricardian land rent theory is commonly the conceptual framework used in real estate appraisal (Murray, 1940) On the other hand, Von Thnens theory focused on absolute and relative spatial location and hypothesized that the value of land arises from the bidding process of the market (Thrall 1987). A user of agricultural land will be a profit maximizer and will be able to pay rent dependent upon the farmers total revenue, total cost, and total transportation cost. It is interesting to note, although Von Thnens work was published in 1826, it was generally unknown to the English speaking world until it was translated from the original German in 1966 (Thunen and Hall, 1966) Sinclair Sinclair argued that for industrialized nations, it was rapid urban development, rather than transportation costs alone, that would bid up land prices. Competition for land led to more intensive agricultur al production near cities (Sinclair, 1967) Sin ce Sinclair, almost all agricultural land use models have tried to incorporation the forces of population growth and the urban environment.

PAGE 38

38 Bryant Bryant (1973) argues that only certain types of agriculture are negatively affected by urban pressure, depending on the time the farmer needs to get a return on the capital invested. He also recognized urbanization as not just the direct effect on the physical conversion of land, but also indirect effects of neighborhood changes (Bryant and Johnston, 1992) Bowler and Ilbery Research by Bowler and Ilbery (1985) argues that traditional agricultural geography models i gnored the political, economic, and societal structures in w hich agriculture operates. In other words, it is the geographical and historical uniqueness to the region that could be important to understand the spatial ramifications of agricultural restructur ing. Johnston and Bryant Johnston and Bryant (1992) recognize a range of factors that can influence farmer decisions. These include: attributes of the f arm operation, such as the existence of an heir or the skill set of the farm operator; attributes of the local community, such as availability of farmland or community concerns about particular farm practices; and off farm factors, such as commodity market prices. The model identifies three types of farmer adaptations One is called a positive adaptation, such as adding nontraditional enterprises or intensifying production on the existing land base. Another is called a normal adaptation such as the adoptio n of a standard agricultural technology (e.g. hybrid seeds ) The third is a negative adaptation, such as exit from farming or a reduction in production intensity in anticipation of a future sale of farmland to developers (Bryant and Johnston, 1992; Sharp and Smith, 2003)

PAGE 39

39 Hart Using a case study of a fifty mile radius of Times Square in New York City, the Hart (1991) model uses the analogy of a standing wave that forms in front of the bow of a ship as it moves though the water. In this model, the bow wave results from a host of decisions by individual farmers. As this wave pushes outward, least intensive urban land uses steadily displace most intensive urban ones (Hart, 1991) Some sell their land and use the proceeds to buy new farms farther out (Gottmann and Harper, 1966) Some take off farm jobs but keep their land and continue to farm it less intensively or rent it to neighbors. Others retain the land and develop cashcrop operations on plots rented from neighbors, from nonfarm landowners, and from speculators who are holding it for development. Bryant and Johnston Next, Bryant and Johnston (1992) introduced a much more complex geography model and a farmer decisionmaking model that specifically made entrepreneurialism possible. In t his model, Bryant and Johnston (1992) diversify potential adaptations by recognizing that different modes of production utilize land, capital and labor differently, under different management systems, and therefore have varied adaptations. Utilizing a general political economic approach, they consider individual farms as a system within a societal context. Their model explores how farms can enjoy links with the urban industrial complex, and suffer from it as well, making the case why post industrial entrepr eneurialism is greatest in the citys countryside (Clark, 2008) In their model, urbanization affects the form (land uses), function (reasons for land uses) and the structure (interrelationships).

PAGE 40

40 Beauchesne and Bryant Beauchesne and Bryants model (1999) focuses on the interaction of actors and builds in the concept of filters. Local and regional forces are filtered at all scales. The scales o f greatest influence depend on the place. In t heir model, the peri urban area is a social construction of the actors locally and regionally in the sociocultural, political, economic and biophysical contexts. Therefore, the effect of forces are not predictable because it depends on the interplay of the actors involved (Beauchesne and Bryant 1999). T he issue of adoption of new policies is fundamental to agricultural development. The general topic of adoption is not a new area of study, and over the last 50 years much practical and academic effort has been invested in understanding the interaction of factors which influence the adoption of new products, policies and technologies (Ilbery, 1985; Bryant and Johnston, 1992; Peddie, 2008) Within an agricultural context, work in this area has identified at least five sets of nonfinancial variables that impinge on the adoption decisions of farmers. These are: farmer characteristics, household char acteristics, farm structure, the wider social environment and the characteristics of the decision to be made. (Luloff et al., 2002) The characteristics of the farmer, which are known to be important in adoption decisions, include age, education, gender, at titude to risk and personality (Mulkey and Clouser, 1985; Pinder, 1998; Willock et al., 1999; Sharp and Smith, 2003) Characteristics of the farm household known to be important include stage in family cycle, and the work patterns of the spouse (Willock et al., 1999) Several key elements relating to the structure of the farm business are also known to have an important influence on decisionmaking, and these include farm type, farm size and debt to asset

PAGE 41

41 ratio (Jones and Wallace, 1986; Lulo ff and Swanson, 1990; Veeck et al., 2006; Varner and Otto, 2008) More recently the structure of the social environment has al so been identified as important in influencing adoption decisions (Wolpert, 1964; W illock et al., 1999; Wu and Gopinath, 2008) Important variables included in this set of factors include the level of extension, information flows, local culture, social capital, attitude of trusted friends, the policy environment and the structure and imp act of a range of institutions (Spedding, 1988; Weldon et al., 1993; Volpe and Lavoie, 2008) The fifth and final set of variables, which influence adoption, relates to the charact eristics of the product or policy to be adopted. Agriculture at the Rural Urban Interface A growing body of research in agricultural systems and agricultural land use change has to do with farms and agribusinesses at the edge of large and expanding urban c enters. Identifying where the city ends and the country begins has become an increasingly controversial debate among researchers, political, and environmental groups. The difficulty of identifying and deli n eating this area of intermingling landuses is of ten compounded based on the size of the urban center, and the severity of the processes at work. In Florida particularly, where the last decade has seen significant population growth and highly speculative real estate transactions, the future of agricultu re has been characterized by a sense of fatalism (Brennan, 2004; Deg raw, 2005) Defining the Rural Urban Interface Various overlapping terms and definitions have been used in the literature to define the rural urban interface. Some of the terms used include fringe, urban fringe, inner fringe, rural urban fringe, urban shadow, exurban zone, and peri urban (Ilbery,

PAGE 42

42 1985) An early definition saw the area as, the area of transition, between recognized urban landuses and the area devoted to agriculture (Wehrwein 1942; Grigg, 1984; Ilbery, 1985) A more detailed definition presented by Pryor (1968) called it, the zone of transition in landuse, social and demographic characteristics, lying between the continuously built up urban and suburban area of the city and the rural hinterland and characterized by an almost complete absence of nonfarm dwelling, occupations, and landuse. Bryant et al. (1992) portrayed graphically the continuum between urban and rural as four distinct zones. This is identified as the inner fringe, where the majority of the land is under construction or has been permitted to begin constructio n. Outside the inner fringe is the outer fringe, and is characterized by predominantly suburban land uses. This area is infiltrated with urban elements such as single family homes along roadways, stock yards, industrial parks, and the megashopping center. Together, the inner and outer fringes form the rural urban fringe. Many properties in the rural urban fringe belong to nonfarm people who acquire the land in anticipation of future urban expansion. The third zone is the urban shadow, where the urban inf luence is minimal, but where the urban presence is felt through commuting patterns of nonfarm residences, and the nonfarm ownership of land. The outermost zone is the rural hinterland. Some elements of urban influence can still be present in this area, s uch as second homes. Locating the Rural Urban Interface Although conceptually the Rural Urban Interface is easy to understand, identifying the various zones is more difficult to delineate in reality. Bryant et al (1992) points out that in reality, much more variation exits. The USDA Economic Research

PAGE 43

43 Service (ERS) developed several classifications to measure rurality and assess the economic and social diversity of rural America (USDA, 2009 ) One key classification developed by USDA ERS is the UrbanInterface Code. Urban Influence Codes divides the 3,141 counties, county equivalents, and independent cities in the United States into 12 groups. These codes are listed in Table 21. These codes are designed to categorize a count ys potential for economic opportunity by virtue of its size and access to larger economics. Urban Influence Codes were used by Clark (2008) to define counties that are in the Rural Urban Interface. Smaller counties adjacent to metro areas experiencing high growth, counties with UIC codes of 5 7 and counties whose growth between 1990 and 200 0 was greater than the national median of 13.15 percent were selected. Counties that were coded with a UIC of 1 and had over 50% of their county land base in agricultural use were also included. It is interesting to note that by these criteria, all of Floridas counties except Taylor and Madison (64 out of 67) are considered Rural Urban I nterface counties Rural Urban I nterface counties are shown in Figure 21 Figure 22 shows the number of farms in Florida as recorded by the USDA. The farm numbers correspond to the events laid out earlier in this chapter. Florida farms grew in numbers gradually until the 1930s. During the great depression, the numbers grew at an increasing rate as unemployment was high. The World War II years saw a decrease in the number of farms as labor was in short supply. There was a small increase at the end of the war, but numbers has been in a st eady decline over the next several decades

PAGE 44

44 The 1990s shows a small increase in the number of farms, followed by a drop, and then an increase again. Figures 23 and 24 helps explain this, and gives us a clearer picture of farmland change. Although the data shows there are more farms in Florida in 2009 than in 1989, total number of acres, as well as average farm size, are both less. This reflects a trend of few large farms breaking up into multiple small farms, with a net farmland loss. These graphs reflect the importance of the historical and social context in explaining agricultural land use change.

PAGE 45

45 Table 21. Urban Influence Codes (USDA ERS 2003) Code Description 1 In large metro area of 1+ million residents 2 In small metro area of less than 1 million residents 3 Micropolitan area adjacent to large metro area 4 Noncore adjacent to large metro area 5 Micropolitan area adjacent to small metro area 6 Noncore adjacent to small metro area and contains a town of at least 2,500 residents 7 Noncore adjacent to small metro area and does not contain a town of at least 2,500 residents 8 Micropolitan area not adjacent to a metro area 9 Noncore adjacent to micro area and contains a town of at least 2,500 residents 10 Noncore adjacent to micro area and does not contain a town of at least 2,500 residents 11 Noncore not adjacent to metro or micro area and contains a town of at least 2,500 residents 12 Noncore not adjacent to metro or micro area and does not contain a town of at least 2,500 residents

PAGE 46

46 Figure 21. Rural Urban Interface Counties (Clark 2008)

PAGE 47

47 Figure 22. Farm Numbers in Florida 1910 2008 (USDA NASS 2009) 0 10 20 30 40 50 60 70 80 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005Number of Farms in ThousandsYear -Source: USDA National Agricultural Statistics Service

PAGE 48

48 Figure 23. Number of Farms in Acres in Florida 1950 2008 (USDA NASS 2009) 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 Thousand AcresYear -Source: USDA National Agricultural Statistics Service

PAGE 49

49 Figure 2 4. Average Farm Size in Acres in Florida 1993 2008. (USDA NASS 2009) 100 120 140 160 180 200 220 240 2601993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 AcresYear -Source: USDA National Agricultural Statistics Service

PAGE 50

50 CHAPTER 3 METHODOLOGY The area for this study is Polk County, Florida. The location of Polk County is shown in Figure 31. Several conditio ns make Polk County ideal for this type of study: Polk County is an agriculturally important area. According to the Census of Agriculture, Polk County has the 2nd largest amount of farmland in the state with approximately $880 million in sales annually. Po lk County experienced a large number of real estate transactions during the last 10 years. Polk County is located in between two of Floridas largest cities, Tampa and Orlando, making the area an ideal place to live for people commuting to the metropolitan area, thus putting pressure on land use change in the county. Data used for the analysis was collected from several sources. Parcel data was used from the Polk County Property Appraiser. Socioeconomic data was collected from the US Census Bureau and the Florida Geographic Data Library (FGDL) located at the University of Floridas GeoPlan Center. The FGDL also provided a variety of base infrastructure data used for geographic analysis. Other supplemental, nonGIS based data was collected from the USDA Nati onal Ag Statistics Service (NASS) and the USDA Economic Research Service (ERS). Because the data coming from different sources was often in different formats, several software packages were used in data preparation, modeling, and analysis. Microsoft Access was used for converting large database files (sometimes millions of records), and converting to and from ASCII text, Excel and dBASE formats. Environmental Systems Research Institutes (ESRI) ArcGIS 9.2 was used for the majority of geographic analysis. For drivetime analysis, Microsoft MapPoint 2009 was used. S tatistical analysis and logit regression w ere performed using PASW Statistics 13 (formerly known as SPSS.)

PAGE 51

51 An important aspect of the study is the prolific use of Geographic Information Systems (GIS) to analyze data from different sources and in different formats. As a result, a significant amount of time and effort w ere needed for data transformation and cleaning. Data in different projections was standardized in a common coordinate system, which the researcher chose to match the projection system used by the FGDL: Albers Conical Equal Area. In any study using GIS, data preparation is the most time consuming and critical aspect of the process. Mistakes early in the data preparation could have a prof ound effect on the results of the analysis. For this reason, the steps for data conversion are clearly explained so that the process could be repeated or adapted by others. Another unique aspect of this study is the use of Public Land Survey System (PLSS) sections. Originally proposed by Thomas Jefferson, the PLSS began shortly after the Revolutionary War, when the government became responsible for land outside the original 13 colonies (USGS, 1997) The PLSS typically divides land into 6 mile square areas called townships. Townships are subdivided into 36 onemile square sections. Sections could be further subdivided into quarter sections, quarter quarter sections, or irregular lots. Previous studies examining land use change below the county level used individual parcel sizes. Land parcels, particularly agricultural land parcels, tend to change over time as land is sold, subdivided for smaller homes, or divided among heirs. The benefi t to using PLSS sections over other methods are that they are standardized by and do not change over time. It should be noted however, that not all states use the PLSS system (Ramskill and Lansing 1922).

PAGE 52

52 In cases where it is desired to repeat this study w h ere the PLSS is not used, a generic grid (sometimes called a fishnet) could be constructed by most GIS software programs over the study area, at whichever scale the researcher desires. The PLSS sections for Polk County are shown in Figure 32. Data Prep aration As stated earlier, PLSS sections are the unit of measurement for analysis. An ESRI shapefile containing PLSS sections for Polk County in vector format was obtained from the Polk County Appraisers Office. Before the PLSS could be used for analysis, it needed to be checked for accuracy. With the exception of sections that border large water bodies or other counties, PLSS sections have an area of roughly 1 square mile or 640 acres. The PLSS shapefile was opened in ArcMap 9.2 and reviewed for accuracy After projecting the shapefile to the Albers projection system, the PLSS sections were then sorted by ascending order based on area. Any sections with areas less than 1 square mile were examined. To eliminate the influence of small parcels on the model, any sections with an area of less than 200 acres were deleted from the shapefile using the Editor function in ArcGIS. The resulting PLSS grids are shown in Figure 33. Since land that is submerged cannot be used for urban or agricultural use typical to Fl orida, a shapefile of hydrological features was also imported into ArcGIS from the FGDL. Fortunately, the FGDL already uses the Albers projection system, so changing the projection system of the shapefile was not necessary. Overlaying the hydrological shap efile over the PLSS shapefile, a select by location was performed, and any PLSS sections that were submerged and contained less than 200 acres of land were removed from the shapefile.

PAGE 53

53 Other areas of Polk County were also excluded from the analysis. These areas were land that could not be used for either urban or agricultural use, due to some sort of restriction or easement on the land. This included areas such as the Green Swamp Wildlife Management Area in Northwest Polk County, as well as Avon Park Bombing Range in Southeast Polk County. These areas were removed from the PLSS shapefile using the method described above. To include the effect of roads and highways in the analysis, a shapefile containing the major roads in Florida was used. The shapefile was also available as part of the FGDL, so no projection was needed. Since the shapefile contained major roads for all of Florida, scaling the shapefile down to only include Polk County was necessary. This was accomplished by clipping the major roads shapefil e with the Polk County PLSS shapefile. This cookie cutter approach ensured that only the roads in the study area would be included in the analysis. Construction of Dependent and Independent Variables In order for any type of analysis to be performed dependent and independent variables need to be constructed. Additionally, since these variables are constructed from a variety of sources at different scales, they all need to be normalized at the PLSS scale, the level chosen for this study. The following sec tion explains the method for how this was accomplished. The Dependent Variable Unfortunately, there is no reliable data on the number of active farms in the U.S. The only national attempt to collect data on American farms comes from the USDA Cens us of Agri culture (USDA 2007). The Census of Agriculture is conducted every five years in years ending with 2 and 7. It is conducted via a mail or internet survey with

PAGE 54

54 some follow up by telephone or inperson contact. Data includes number of farms, land in farms, si ze of farms, value of land and buildings, amount of cropland, amount of pasture, land use characteristics, the amount of value of harvested crops, and the number and value of livestock and poultry. The Census does not collect any information on the exact l ocation of farms. Since no accurate data on the number of farms and their locations exist in Polk County, lands classified as an agricultural use from the Florida Department of Revenue (DOR) was used instead. DOR use codes are numerical designations used by property appraisers to classify each parcel of real property. The Florida DOR use codes current to Polk County are included in Appendix A. Parcels that had agricultural use codes were dissolved together. The result was one map of Polk County that showed lands that are in agriculture, or at the very least, where agricultural use was allowed. This map is shown in Figure 34 A spatial join was then performed on the PLSS shape file and any agricultural areas of at least 200 acres in size that was within a PLSS grid was considered an agriculturally significant PLSS grid. The 200 acres size was chosen because this is the average farm size in Florida as reported by the USDS ERS. Each cell in the PLSS shapefile then had the value of 1 for agricultural land, or 0 for non agricultural land, thus giving a dependent variable AG for use in further analysis. The resulting Figure is shown in Figure 35 As in the definition of rural urban fringe, the definition of what is agricultural land can vary greatly (Ilbery et al., 1997) For purposes of this study, agricultural land is defined as land that could be used for agriculture, both physically and legally. For

PAGE 55

55 example, although forest and park areas could potential ly be used as agricultural land, if they are conservation areas or under some sort of easement, then legally they cannot be used as such. These restrictions are reflected in the Florida DOR land use codes. Independent Variables A number of possible explanatory variables were constructed for this study using both vector and raster techniques. Vector data stores spatial information in the form of discrete coordinate points, in the form of points, lines, or polygons, with associated data stored in an attribute table. Alternatively, raster data consists of a grid of cells where each cell stores a unique value. As stated in chapter 1, a n objective of this study i s to investigate the relationship between the quantity of real estate transactions and agricultural land. An explanatory variable, TotalN, was cons tructed to indicate the total number of times land was bought and sold (otherwise known as flipping) since 1980. A database containing sales transactions by Polk County Parcel ID was imported into MS Access. The table listed property by each parcel number, as well as the date the property was sold. This data was then converted to show the number of times a single parcel was sold after 1980 by using the following Structured Query Language expression (SQL): SELECT Pol k.PARID, Polk.SALEDT FROM Polk WHERE (([ Polk]![SALEDT]>#12/31/1979#)) ORDER BY Polk.PARID; PARID is the unique ID number of the Parcel, and SALEDT is the date of the sale. The resulting table was then joined to a parcel shapefile based on unique parcel numbers. The result of this process is a sh apefile organized by parcel number with the

PAGE 56

56 total number of times a parcel sold. This can then be displayed in ArcGIS. The resulting shapefile is shown in Figure 36. As with the agriculture dependent variables, the TotalN variables were normalized by PLSS using a spatial join in ArcGIS 9.2. The spatial join function was able to produce two different independent variables. One was the sum of TotalN per PLSS, as well as the average of TotalN per PLSS grid. This variable was called Avg_N. An inc reasing number of studies from the agricultural geography and rural sociology literature have suggested that socio economic variables, rather than more tangible aspects like soil type, have a much greater influence on agricultural land conversion than prev iously thought This is particularly true in industrialized nations. The second objective of this study is to identify and investigate socioeconomic variables significant to agricultural land use change. In order to accomplish this, U.S. Census data was used to construct socio economic variables for the model Income, a ge, and h ousehold size were of particular interest to this study. Census data for median income, age, and household size was available for Polk County by Census block. Census blocks are the smallest geographic unit used the by the US Census Bureau. On average, there are about 39 blocks per block ground, but in rural areas, areas can be much greater. By importing census block groups into ArcGIS and performing a spatial join on the PLSS grids, independent variables for median age, median income, and median household size were created. The census shapefile is shown in Figure 37. A

PAGE 57

57 Unique to the location of Polk County is its location between two of Floridas largest cities, Tampa and Orlando. Studies in urban sprawl and location by researchers such as Von Th nen and Thrall indicate the affect large metropolitan areas can have on agricultural land conversion. One of the primary reasons Polk County has been so attractive to new residents and develop ers is the commuting times to these large cities. In order to account for this influence, drive times from each PLSS grid in Polk County to Tampa and Orlando were used. The central business districts (CBDs) for Tampa and Orlando were located using MS MapPo int. MapPoint has the ability to draw drive time polygons around each CBD. Polygons were constructed in 10 minute increments and were then exported as Keyhole Markup Language (KML) files. KML is an open source language schema for representing geographic annotation and visualization. Most commonly KML is used with Google Earth software. On ce the MapPoint drive time polygons were converted to a KML format, they could then be imported into ArcGIS for conversion to ESRI shapefiles. Once converted into a shapefi le format, it could then be spatial joined to the PLSS grid. Thus, each PLSS grid in Polk County contained an independent variable, DTOrlando, for drive time to Orlando, and DTT ampa, for drivetime to Tampa. The past 20 years also showed significa nt growth for Polk Countys 17 municipalities. Between 2004 and 2009 there was an increase of 16,426 parcels within city limits (Polk County 2009). Expanding city limits places a large pressure o n agricultural land owners. In order to account for this aff ect, a variable that measured the distance from the city limit to each PLSS grid, CITYD was constructed. This was accomplished by importing a polygon showing city boundaries into ArcGIS. Using the B

PAGE 58

58 Spatial Analyst extension within ArcGIS, a distance raster was created. This was a straight line distance from the municipal boundary outwards. A second raster shapefile of the PLSS grid was also created using the Feature to Raster command in ArcToolbox, using the Feature ID (FID) as the cell value. The FID is a unique identifier for each PLSS grid With the two raster shapefiles, the Zonal Statistics function in Spatial Analyst was used to create a table with distances to city boundaries for each PLSS grid. This table was then joined to the original PLSS vector shapefile, thus giving a drivetime value for each PLSS grid. The drive time and other independent variables are listed in Table 31. Other variables such as years of farming, and type of farming operation could also be useful to the model, but were not included due to the nature of the available census data. Education level is another possible variable and is available in census data, but was not used in this study. Due to the conversions involved and the nature of vect or analysis, a key step is to manually review the completed shapefiles and corresponding attribute tables and correct any errors (Goodchild et al., 1993; Todd, 2002) The first part was to compare the number of PLSS grids to the numbers of each variable. All variables should match to the total of 1692 cells. In this case some of the PLSS cells had extra lines and polygons that were not initially visible. These extra lines and polygons were removed using ArcEditor. Another step was to verify that each independent variable corresponded to the correct PLSS cell. For example, a PLSS cell that falls within a city polygon should have

PAGE 59

59 a CITYDIST value of zero. In a few cases, the PLSS cell took on the value of an adjoining cell. This was corrected in the attribute table using ArcEditor. Logistic Regression This study classifies Polk County PLSS grids a s either agricultural or nonagricultural (e.g. residential commercial or mining .) For this reason, a binary logit model was chosen over other models to reflect the actual use of land. A binary logit model also has the added benefit that it does not impose any strict assumptions regarding the distributions of the variables or the residuals, as in the case with ordinary least squares (OLS) regression (Menard, 2002) The generalized model takes the form: 1. Y = f (Income, Population, Household Size, Age, Distance to City Limit, Drive time to Orlando, Drive time to Tampa, Number of Transactions.) The binary logit model takes the form: 2. where, Y = 1 for agricultural PLSS grids and Y = 0 for nonis the The predicted probabilities of occurrence can be calculated using the equations for the m aximum likelihood estimators for the logit model: 3. Pr (Yi = 1) = and 4. Pr(Yi = 0) = The logistic regression model was estimated using the statistical software package PASW 18 (IBM 2009.) As stated earlier, the dependent variable for the model is AG which represents a PLSS grid being predominantly agricultural or not. PLSS grids that are classified as AG doesnt necessarily mean that those areas are currently active

PAGE 60

60 farms. Rather, these are areas that are classified by the Polk County property appraiser as an agricultural parcel, and could be used in farming. The purpose of the logisti c regression model is to meet the third objective of the study, develop a model to estimate agricultural land use change, and where that change is likely to occur.

PAGE 61

61 Table 3 1. List of variables. Section name This section contains POP MALES FEMALES Population Number of Males Number of Females MED_AGE MED_AGE_F MED_AGE_M Median Age Median Age Female Median Age Male HH_SZ HSE_UNITS Average Household Size Number of Housing Units in TOTALN Total number of transactions or flips MED_INCOME CITYDIST_M ROADDIST_M Median Income Distance to city bounry, in miles Distance to major road, in miles AVG_N The average number of flips TPATIME The driving time to Tampa MCOTIME The driving time to Orlando

PAGE 62

62 Figure 31. Study Area

PAGE 63

63 Figure 32. Polk County PLSS Sections (Polk County Property Appraiser)

PAGE 64

64 Figure 33. PLSS grids used for analysis.

PAGE 65

65 Figure 34. Agricultural Lands in Polk County (Polk County Property Appraiser 2009)

PAGE 66

66 Figure 35. Agriculturally Dominant PLSS Grids.

PAGE 67

67 Figure 36. Polk County parcels with total transactions.

PAGE 68

68 Figure 37. Shapefile of census blocks (US Census 2008)

PAGE 69

69 CHAPTER 4 RESULTS Descriptive Statistics An important aspect of this study was how real estate transactions influence agricultural land. In developing the model, average and total number of transactions per PLSS grid was derived. As part of this process, the number of real estate transactions in Polk County per year was also derived. The following graphs show real estate transactions from 1980 to 2002, and 1980 to 2009. The graph in Figure 41 shows real estate transactions from the standpoint of an investor or developer reviewing the data in year 2002, the beginning of the real estate bubble in Florida. From this perspective, it would be very easy to assume that real estate transactions will continue to increase in line with population growth, making the purchase of property an ideal investment. From the standpoint of the farmer, it may look as though competing uses of land would make agriculture a more difficult and expensive occupation. Figure 42 gives us a much different picture. Although transactions did increase from 2002 to 2005, there was a sharp drop in the years following. Despite almost double the county population, 2009 transactions are at the level they were in 1985, suggesting fewer people were moving homes or buying investment properties. Another important objective of this study w as to compare the socioeconomic differences between agricultural and non agricultural lands. Tables 43 and 44 lists the descriptive statistics for the relevant variables. As suggested by previous research (Rosenberger, 1962; Mulkey and Clouser, 1985; Spedding, 1988) median age is higher in agricultural PLSS grids than in nonagricultural grids. Agricultural PLSS grids also

PAGE 70

70 have a higher mean population, household size, and average transactions, suggesting that Polk County has grown out, rather than up, or suburbanization as opposed to gentrification. Mean median income was higher in agricultural areas, but nonagricultural areas had a high mode and larg er maximum value. This suggests that overall, people in agricultural PLSS grids enjoyed a high level of wealth, but the highest income earners were in nonagricultural areas. Using the mean and standard deviation values found by the analysis, we can then map these by PLSS grid. Figures 41 and 42 maps total and average transactions by PLSS grids in Polk County by standard deviation. Dark and light green areas show areas that are 0 .5 standard deviations or more below the mean, yellow areas are around the m ean, and orange and red areas are 0 .5 standard deviations or higher above the mean. Dots are shown on the map to represent which grids are agricultural. The result is a highly informative map that shows precisely which locations are more (or less) desirabl e than others. This map gives us a hot spot analysis similar to that of the Getis Ord Gi* statistic (Getis and Boots, 1978) Figures 43 through 48 show PLSS maps for the remaining variables. Cells with greater population are in the northern part of the county along the I 4 corridor, the major road connecting Tampa and Orlando. This includes the population for males and females, as well as overall. Median age, for males, females, and overall are slightly higher on the eastern part of the county, the side closest to Orlando. The number of housing units tend to be higher in the northern part of the county, however average household size tends to be larger in the southern half. This suggests larger home in the

PAGE 71

71 southern part of the county as compared to the northern part. Median income is higher within the city limits, as opposed to the more rural areas Drive times to both Tampa and Orlando correspond to access to major roads. Model Formulation A logistic regression model can only be obtained for PLSS grids that contain complete data for all variables. The first step was therefore to verify that there was no missing data in any of the grids. In this case, there was no missing data, so all PLSS grids could be used. In order to assure that only those variables that are considered significant in the model are included, a forward stepwise routine was used in the logistic regression model. Forward stepwise routines work by starting the model with only the intercept, and then running the model multiple times adding one variable at a time. If that variable proves to be statistically signi ficant, then it remains in the model. If the variable is not significant, it is excluded from the model. The results of the stepwi se regression are shown in Tables 41 and 42 Of the 12 original independent variables in the model, only six were shown to be significant. The variable for the average number of transactions per PLSS gird, AVG_N was significant and positively correlated to AG, meaning as the average number of transactions in a PLSS grid increases, the likelihood of that PLSS grid being ag ricultural also increased. The variables representing driving times to Tampa and Orlando are also positively correlated, indicating as driving time increases, the probability of the PLSS grid being agricultural also increases. The inverse would also hold true, the shorter the drive time to those metropolitan areas, the less chance of that grid being predominantly agricultural.

PAGE 72

72 Average Household Size was also positively correlated, indicating as household size increases, so does the probability of a PLSS g rid to be predominantly agricultural. This seems to contradict previous research that indicates farmers have a lack of an heir to continue farming (and therefore, should have a smaller average household size.) Sociology research is quick to point out howev er, that the predominant makeup of the household in urban areas is a single parent, while suburban and rural areas are more likely to have two parent homes. Distance from major roads and city limits were included in the model, not as a proxy for commuting time, but as variables to represent the amount of infrastructure available such as water, sewage, and electricity. Both the variables CITYDIST_M and ROADDIST_M were negatively correlated to the dependent variable, meaning the likelihood of a PLSS grid being agricultural decreases as one get s closer to the city limits or a major road. Farms typically are thought of to be far away from infrastructure, but this model implies the opposite. It is important to remember this model uses land that is classified as agriculture, not actual farms. An agricultural status is highly sought after by landowners for property tax exemptions. A negative correlation is indicative to the rise of hobby farms and or other recreational use of large areas of land. When consid ering model significance, most people are familiar with Pearsons R square value as a measure of how well independent variables explain variance in the model. R square values in logistic regression however are not analogous (Pohlmann and Leitner, 2003) A more accurate measure of goodness of fit is how well the model correctly identifies PLSS grids. In our model, nonagricultural parcels were correctly

PAGE 73

73 identified 49. 5 % of the time, whereas agricultural PLSS gr ids w ere accurately predicted 83.0 % of the time, giving an overall accuracy of 67.5%. This is the best model as shown in step 6 in Table 42 Table 42 gives us the results of the independent variables. The column B represents the log odds used in logistic regression. By itself, log odds is not a very useful number, but taking the exponent of that number gives us the odds ratio(Menard, 2002) Standard Error and significance level (denoted p) is also given in the table Average household size has an odds ratio of 1.158. In other words, for each change in household members (either increase or decrease) the odds of the PLSS cell being agricultural changes by 16%. Drive time to Tampa had a odds ratio of 1.019. For each minute drive time away from Tamp a, the odds of the PLSS cell being agricultural increases by 2%. A 10 minute drive time would increase the odds by 20%.Drive time to Orlando had an odds ratio of 1.009. For each minute drive time to Orlando, the odds of the PLSS cell being agricultural changed by 1%. A 10 minutes drive time would change the odds by 10%. The odds ratio for the average number of PLSS transactions was 1.083. For each transaction that occurred, the odds that the cell being agricultural increased by 8%. The odds ratio for the road distance was 0.768. For each mile closer to a major road the odds of the cell being agricultural decreased by 23%. City distance had an odds ratio of 0.918. For each mile closer to a city, the odds of the cell being agricultural decreased by 8%.

PAGE 74

74 Table 4 1. Results of regression variables. B Standard Error p Odds ratio Step 1 CITYDIST_M .084 .006 .000 .919 Constant .854 .069 .000 2.350 Step 2 tpatime .017 .003 .000 1.018 CITYDIST_M .094 .006 .000 .910 Constant .246 .221 .266 .782 Step 3 tpatime .016 .003 .000 1.016 mcotime .009 .003 .003 1.009 CITYDIST_M .091 .007 .000 .913 Constant .753 .281 .007 .471 Step 4 tpatime .017 .003 .000 1.017 mcotime .009 .003 .003 1.009 CITYDIST_M .093 .007 .000 .911 AVG_N .080 .031 .010 1.083 Constant 1.067 .307 .001 .344 Step 5 tpatime .019 .003 .000 1.019 mcotime .009 .003 .003 1.009 ROADDIST_M .235 .091 .009 .790 CITYDIST_M .086 .007 .000 .917 AVG_N .081 .031 .010 1.084 Constant 1.124 .308 .000 .325 Step 6 ave_hh_sz .147 .054 .006 1.158 tpatime .019 .003 .000 1.019 mcotime .009 .003 .006 1.009 ROADDIST_M .264 .092 .004 .768 CITYDIST_M .086 .007 .000 .918 AVG_N .080 .032 .012 1.083 Constant 1.441 .332 .000 .237

PAGE 75

75 Table 42. Model Predicted versus Observed. Observed Predicted Percentage Correct 0 1 Step 1 ag 0 402 382 51.3 1 189 719 79.2 Overall Percentage 66.3 Step 2 ag 0 383 401 48.9 1 170 738 81.3 Overall Percentage 66.3 Step 3 ag 0 381 403 48.6 1 163 745 82.0 Overall Percentage 66.5 Step 4 ag 0 389 395 49.6 1 160 748 82.4 Overall Percentage 67.2 Step 5 ag 0 383 401 48.9 1 155 753 82.9 Overall Percentage 67.1 Step 6 ag 0 388 396 49.5 1 154 754 83.0 Overall Percentage 67.5

PAGE 76

76 Figure 41 Average Transactions by PLSS grid.

PAGE 77

77 Figure 42 Total Transactions by PLSS grid.

PAGE 78

78 Figure 43 Population by PLSS grid.

PAGE 79

79 Figure 44 Median Age by PLSS grid.

PAGE 80

80 Figure 45 Male Median Age by PLSS grid.

PAGE 81

81 Figure 46 Female Median Age by PLSS grid.

PAGE 82

82 Figure 47 Males by PLSS grid.

PAGE 83

83 Figure 48 .Females by PLSS grid.

PAGE 84

84 Figure 49 Number of Housing Units by PLSS grid.

PAGE 85

85 Figure 41 0 Household Size by PLSS grid.

PAGE 86

86 Figure 411. Median Income by PLSS grid.

PAGE 87

87 Figure 41 2 Drive time (minutes) to Orlando by PLSS grid.

PAGE 88

88 Figure 41 3 Drive time (minutes ) to Tampa by PLSS grid.

PAGE 89

89 Figure 41 4 Distance to City by PLSS grid.

PAGE 90

90 Figure 41 5 Distance to Major Roads by PLSS grid.

PAGE 91

91 Table 43. Descriptive Statistics for Socio Economic Variables in NonAgricultural PLSS Grids Mean Standard Error Median Standard Deviation Minimum Maximum Population (persons) 369.99 18.58 142 539.70 0.00 3868 Males (persons) 185.04 9.31 69 270.54 0.00 1946 Females (persons) 184.95 9.28 73 269.69 0.00 1922 Median Age (years) 34.02 0.63 35.75 18.24 0.00 86.5 Male Median Age (years) 33.74 0.64 35.4 18.59 0.00 91.5 Female Median Age (years) 34.35 0.63 36.2 18.33 0.00 82.2 Average Household Size (persons) 2.24 0.04 2.555 1.09 0.00 6 Average Housing Units (homes) 187.75 9.87 61 286.69 0.00 2407 Med i an Income (Year 2000 dollars) 35285.09 242.80 34824.50 7053.76 22364.00 77727.00 Total Transactions (each) 2.94 0.05 2.99 3.14 0.00 50 Average Transactions (each) 2.27 0.11 2.00 1.40 0.00 11

PAGE 92

92 Table 44. Descriptive Statistics for Socio Economic Variables in Agricultural PLSS Grids Mean Standard Error Median Standard Deviation Minimum Maximum Population (persons) 404.30 20.42 164 0 617.39 0 Males (persons) 206.99 10.69 85 0 323.18 0 Females (persons) 197.77 9.99 83 0 301.93 0 Median Age (years) 35.40 0.52 36.1 0 15.78 0 Male Median Age (years) 34.68 0.54 35.5 0 16.21 0 Female Median Age (years) 35.84 0.54 36.5 0 16.30 0 Average Household Size (persons) 2.46 0.03 2.67 0 1.00 0 Average Housing Units (homes) 190.43 10.35 71 0 313.02 0 Med i an Income (Year 2000 dollars) 35923.99 221.76 36906.00 27378.00 6704.21 22364.00 Total Transactions (each) 3.19 0.05 3.14 3 1.28 0 Average Transactions (each) 0.66 0.04 0 0 1.60 0

PAGE 93

93 Figure 41 6 Polk County Real Estate Transactions 1980 to 2002 (Polk County Property Appraiser). 0 10000 20000 30000 40000 50000 60000 70000 Number of TransactionYear

PAGE 94

94 Figure 417. Polk County Real Estate Transactions from 1980 to 2009 (Polk County Property Appraiser.) 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000 Number of Transaction

PAGE 95

95 CHAPTER 5 SUMMARY AND RECOMMENDATIONS FOR FUTURE RESE A R CH Summary This study examined the socioeconomic variables tied to agricultural land use. To date, no study attempted to model agricultural land change in Polk County, especially after the historic proportions seen in the years spanning 2002 2009. The study identified the socioeconomic characteristics that correlated to land that remained predominantly agricultural. Polk County was an ideal choice for this study. The area is a major agricultural producer that is located between two of Floridas most rapidexpanding areas. Polk County was also not immune to the speculation and real estate frenzy that took the state, and most of the country, by storm. Another aspect unique to this study was the use of PLSS grids. Unless changed by rule of law, PLSS grids will remain constant over time, where parcels can be merged or subdivided. PLSS grids are easily understood by a wide variety of audiences, including planners, agricultural land owners, and utilities. Real estat e transactions for the past 20 years were analyzed, and a map showing hot spots of activity was produced. Other socioeconomic variables were then mapped and analyzed using binary logistic regression to identify those variables that correlated to land being agricultural or nonagricultural in nature. The variables that most likely influenced if land would be used for agricultural use were drive time to Tampa and Orlando, average household size, and average number of transactions. The model correctly predicted the use of land 67.5% of the time. From the results of this research, several trends can be observed:

PAGE 96

96 1. Drive time to major metropolitan areas plays a major role in farmland conversion. The shorter the drivetime to a major area (not necessarily the shor test distance), the more likely the conversion from agricultural to nonagricultural use. 2. Average transactions are higher in agricultural grids than urban grids. Areas with the highest number of transaction will see transition from agricultural to nonagri cultural uses first. 3. Areas with easy access to infrastructure will be converted to nonag use before those with less access. 4. Agricultural areas with larger household sizes will be more likely to be converted to nonagricultural use. 5. Farm sizes are generall y getting smaller. The greater the transactions, the more likely the decrease in farm size. Recommendations for Future Research All real estate is local is a phrase typically used in real estate. Markets differ based on climate, demographics, geography, and available amenities. This model, like all models, is based on certain assumptions and does require some judgment calls to be made. One judgment call was the decision to use 200 acres as the threshold to call a PLSS grid agricultural. While this number may work well in Polk County, it may not work as well in the northeast, where average farm sizes are much smaller, or somewhere in the Midwest, where farm sizes generally tend to be larger. Formalizing a procedure to derive this number would be greatly beneficial. Similar to the way many homeowners felt about their water supplies (i.e. it will always be there), access to credit was long thought of as a resource that would never go away. For the most part, this was true, but recent economic conditions and regulations on banks have made access to credit much more difficult, especially a high risk candidate such as a farmer. Unless the purchaser has access to a large cash reserve, buying real estate is all but impossible. Even USDA loan programs require the

PAGE 97

97 borrower to first get approval from another institution before becoming eligible for a loan from the USDA. Incorporating avai lability and access to credit would be one area that could improve this and similar models. Agriculture is increasingly facing more political pressure. Issues revolving around food safety, agriterrorism, biofuels, genetically modified foods, all play great er roles in the farmers life today. This model made no attempt to capture these aspects. It is recommended that future research makes an attempt to incorporate these issues.

PAGE 98

98 APPENDIX A FLORIDA DEPARTMENT O F REVENUE USE CODES CODE DE SCRIPTION CODE DESCRIPTION 0000 VACANT RESIDENTIAL 5200 CROPLAND SOIL CAPABILITY CLASS 2 0002 VACANT CO OP W/NO MOBILE HOME 5201 CROPLAND SOIL CAPABILITY CLASS 2 WITH IMPROVEMENTS 0004 VACANT CONDOMINIA 5300 CROPLAND SOIL CAPABILITY CLASS 3 0005 VACANT CO OP W/MOBILE HOME ON PERSONAL PROPERTY 5301 CROPLAND SOIL CAPABILITY CLASS 3 WITH IMPROVEMENTS 0100 SINGLE FAMILY 5400 TIMBERLAND SITE INDEX 90 AND ABOVE 0104 SINGLE FAMILY HOMES/CONDOMINIA 5401 TIMBERLAND SITE INDEX 90 AND ABOVE WITH IMPROVEMENTS 0108 1/2 OF A DUPLEX 5500 TIMBERLAND SITE INDEX 80 TO 89 0200 MOBILE HOMES 5501 TIMBERLAND SITE INDEX 80 TO 89 WITH IMPROVEMENTS 0204 MOBILE HOMES/ CONDOMINIA 5600 TIMBERLAND SITE INDEX 70 TO 79 0300 MULTI FAMILY 10 UNITS OR MORE 5601 TIMBERLAND SITE INDEX 70 TO 79 WITH IMPROVEMENTS 0400 CONDOMINIA IMPROVED 5700 TIMBERLAND SITE INDEX 60 TO 69 0500 COPPERATIVES 5701 TIMBERLAND SITE INDEX 60 TO 69 WITH IMPROVEMENTS 0600 RETIREMENT HOMES NOT ELIGIBLE UNDER 196.192 5800 TIMBERLAND SITE INDEX 50 TO 59 0700 MISCELLANEOUS RESIDENTIAL MIGRANT CAMPS, ETC. 5801 TIMBERLAND SITE INDEX 50 TO 59 WITH IMPROVEMENTS 0800 DUPLEX 5900 TIMBERLAND AND NOT CLASSIFIED BY SITE INDEX TO PINES 0801 TWO OR MORE HOUSES 5901 TIMBERLAND AND NOT CLASSIFIED BY SITE INDEX TO PINES WITH IMPROVEMENTS 0803 MULTI FAMILY MORE THAN 2, LESS THAN 10 UNITS 6000 GRAZING LAND SOIL CAPABILITY CLASS 1 0805 HOUSE PLUS A DUPLEX 6001 GRAZING LAND SOIL CAPABILITY CLASS 1 WITH IMPROVEMENTS

PAGE 99

99 0900 UNDEFINED RESERVED FOR D.O.R. 6100 GRAZING LAND SOIL CAPABILITY CLASS 2 1000 VACANT COMMERCIAL 6101 GRAZING LAND SOIL CAPABILITY CLASS 2 WITH IMPROVEMENTS 1100 STORES, ONE STORY 6200 GRAZING LAND SOIL CAPABILITY CLASS 3 1200 MIXED USE STORE & OFFICE, STORE & RESENDENTIAL COMBINATIONS, ETC. 6201 GRAZING LAND SOIL CAPABILITY CLASS 3 WITH IMPROVEMENTS 1300 DEPARTMENT STORE 6300 GRAZING LAND SOIL CAPABILITY CLASS 4

PAGE 100

100 LIST OF REFERENCES Ahuja, L., Ma, L., and Howell, T.A. 2002. Agricultural system models in field research and technology transfer Boca Raton, Fla.: Lewis. Alexander, D. 2000. Confronting catastrophe : new perspectives on natural disasters New York: Oxford University Pres s. Attaway, J.A. 1999. Hurricanes and Florida agriculture. [Lake Alfred: Florida Science Source. Austin, E.J., Willock, J., Deary, I.J., Gibson, G.J., Dent, J.B., Edwards Jones, G., Morgan, O., Grieve, R., and Sutherland, A. 1998. Empirical models of farm er behaviour using psychological, social and economic variables. Part II: Nonlinear and expert modelling. Agricultural Systems 58(2): 225241. Babcock, R.F., Siemon, C.L., and Lincoln Institute of Land Policy. 1985. The zoning game revisited. Boston, Mass .: Oelgeschlager, Gunn & Hain. Beauchesne, A. and Bryant, C. 1999. Agriculture and Innovation in the Urban Fringe: The Case of Organic Farming in Quebec, Canada. Journal of Economic and Social Geography 90(3): 320328. Bowler, I.R. 1985. Agriculture under the Common Agricultural Policy : a geography Manchester [Greater Manchester]: Manchester University Press. Brennan, M.A. 2004. IFAS Community Development: In Search Of A Common Understanding Of Community. Brown, D.L., Swanson, L.E., Barton, A.W., and Rur al Sociological Society. 2003. Challenges for rural America in the twenty first century University Park: Pennsylvania State University Press. Bryant, C.R. and Johnston, T.R.R. 1992. Agriculture in the city's countryside. Toronto ; Buffalo: University of Toronto Press. Bureau, C. 2008. Estimates of Resident Population Change for the United States, Regions, States, and Puerto Rico and Region and State Rankings: July 1, 2007 to July 1, 2008. In (ed. NST EST200803), Washington D.C. Castle, E.N., Becker, M.H ., and Nelson, A.G. 1987. Farm business management : the decisionmaking process 3rd ed. New York: Macmillan. Clark, J. 2008. The Relational Geography in Newly Restructured, Consumptive Spaces: Developing a Model for Agricultural Geography of US Peri Urba n Areas. In American Association of Geographers Annual Meeting.

PAGE 101

101 Colburn, D.R. and DeHavenSmith, L. 2002. Florida's megatrends : critical issues in Florida Gainesville: University Press of Florida. Cresap, I.K. 1982. The History of Florida Agriculture: T he Early Era. Gainesville: University of Florida. Cumming, J.B. 2006. A Brief F lorida Real Estate History. Tampa: West Coast Florida Chapter of the Appraisal Institute. Degraw, J.L. 2005. Perceptions of Florida beef cattle producers on preparedness for a n agroterrorism attack. In. University of Florida, Gainesville, Fla. Dodd, D. 1957. Florida, the land of romance. Tallahassee,: Dept. of Agriculture. Fite, G.C. 1986. The 1980s Farm Crisis. Montana: The Magazine of Western History 36(1): 6971. Furuseth, O.J. and Lapping, M.B. 1999. Contested countryside : the rural urban fringe in North America. Aldershot ; Brookfield, Vt.: Ashgate. Gannon, M. 1996. The new history of Florida. Gainesville, Fla.: University Press of Florida. Getis, A. and Boots, B.N. 197 8. Models of spatial processes : an approach to the study of point, line, and area patterns Cambridge [Eng.] ; New York: Cambridge University Press. Goodchild, M.F., Parks, B.O., and Steyaert, L.T. 1993. Environmental modeling with GIS New York: Oxford University Press. Gottmann, J. and Harper, R.A. 1966. Metropolis on the move; geographers look at urban sprawl New York,: Wiley. Grigg, D.B. 1984. An introduction to agricultural geography London ; Dover, N.H., USA: Hutchinson. Harris, W. and Bronson, C. 2002. Florida's agricultural heritage. 25th anniversary ed. Tampa, Fla.: Florida Agricultural Hall of Fame. Hart, J.F. 1991. The Perimetropolitan Bow Wave. Geographical Review 81(1): 3551. Hodges, A.W. and Mulkey, W.D. 2003. Regional economic impacts of florida's agricultural and natural resource industries. In. University of Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, EDIS, [Gainesville, Fla.].

PAGE 102

102 Hoyt, H. 1970. According to Hoyt; 53 years of Homer Hoyt. Articles on law, real estate cycle, economic base, sector theory, shopping centers, urban growth, 1916 to 1969. [2d ed. Washington,. Hudgins, C.L. and Cromley, E.C. 1997. Shaping communities 1st ed. Knoxville: University of Tennessee Press. Ilbery, B.W. 1985. Agricultural geography : a social and economic analysis Oxford [Oxfordshire] ; New York: Oxford University Press. Ilbery, B.W., Chiotti, Q., and Rickard, T.J. 1997. Agricultural restructuring and sustainability : a geographical perspective. New York: CA B International. Jones, J.R. and Wallace, B.J. 1986. Social sciences and farming systems research : methodological perspectives on agricultural development Boulder: Westview Press. Keen, C. 2008. Economy slows Florida population growth to lower level in 30 years. In University of Florida News University of Florida Press, Gainesville. Knotts, B. 2003. Florida history Chicago, Ill.: Heinemann Library. Luloff, A.E., Krannich, R.S., and NetLibrary, I. 2002. Persistence and change in rural communities a 50year follow up to six classic studies Wallingford, Oxon, UK; New York: CABI. Luloff, A.E. and Swanson, L.E. 1990. American rural communities Boulder, Colo.: Westview Press. Mackin, A. 2006. Americans and their land : the house built on abundance. A nn Arbor: University of Michigan Press. Maddox, R.J. 2003. American history. In. Recording for the Blind & Dyslexic, Princeton, N.J. McGinnis, H. 1982. An evaluation of Florida's agricultural lands conversion problem [Tallahassee, Fla.]: Florida Office o f Planning and Budgeting. Menard, S.W. 2002. Applied logistic regression analysis 2nd ed. Thousand Oaks, Calif.: Sage Publications. Miles, M., Berens, G., and Weiss, M. 2005. Real Estate Development Principles and Process. Third ed. New York: Urban Land I nstitute. Miles, M.E., Berens, G., Weiss, M.A., and Urban Land Institute. 2000. Real estate development : principles and process 3rd ed. Washington, D.C.: Urban Land Institute.

PAGE 103

103 Moeti, K. and Khalo, T. 2007. Public finance fundamentals Cape Town, South A frica: Juta. Molnar, J.J. 1986. Agricultural change : consequences for southern farms and rural communities Boulder: Westview Press. Morgan, W.B. and Munton, R.J.C. 1972. Agricultural geography New York,: St. Martin's Press. Mulkey, W.D. and Clouser, R .L. 1985. An overview of agriculture in Alachua County and a perspective on the issue of agricultural land retention. Gainesville, Fla.: Food and Resource Economics Dept., Institute of Food and Agricultural Sciences, University of Florida. Murley, J.F., A lpert, L., Walker, L., Matthews, M.J., Burton, M., Bryk, C., and Kurry, M. 2006. Charting the Course: Where is South Florida Heading? No. Boca Raton: Center for Urban and Environmental Soultions Regional Initiatives. Murray, W.G. 1940. Farm appraisal; classification and valuation of farm land and buildings Ames, Iowa,: The Iowa state college press. Olson, R.K. and Lyson, T.A. 1998. Under the blade : the conversion of agricultural landscapes Boulder: Westview Press. Peddie, B. 2008. Grounded theory of Florida aquarium retailers' acceptance of the GloFish. In. University of Florida, [Gainesville, Fla.]. Phillips, S. 1986. Factors influencing agricultural land use change : a study of Alachua County, Florida.Thesis (M S ): University of Florida, 1986. Pielke, R., Gratz, J., Landsea, C.W., Collins, D., Sounders, M., and Musulin, R. 2008. Normalized Hurricane Damange in the United States: 19002005. National Hazards Review 9(1): 2942. Pinder, D. 1998. The new Europe : economy, society, and environment New York: John Wiley. Platt, R.H. and Macinko, G. 1983. Beyond the urban fringe : land use issues of nonmetropolitan America. Minneapolis: University of Minnesota Press. Pohlmann, J. and Leitner, D. 2003. A Comparison of Ordinary Least Squares and Logistic Re gression. Ohio Journal of Science 103(5): 7. Pryor, R.J. 1968. Defining the Rural Urban Fringe. Social Forces 47(2): 202215. Ramsey, D.M. 1975. Military installations in Florida during World War II [s.l.: s.n.]

PAGE 104

104 Ricardo, D. 1817. On the principles of pol itical economy, and taxation. London,: J. Murray. Rosenberger, S.E. 1962. A history of the Florida vegetable industry and state farmers' markets for vegetables.Thesis: University of Florida, 1962. Schneph, R. and Chite, R. 2005. U.S. Agriculture After Hur riance Katrine: Status and Issues. No. Washington, D.C.: Congressional Research Service. Seese, R.J. 1998. World War II facilities in Florida : includes list of war movies filmed in Florida [S.l.: s.n. Sharp, J.S. and Smith, M.B. 2003. Social capital an d farming at the rural urban interface: the importance of nonfarmer and farmer relations. Agricultural Systems 76(3): 913927. Sinclair, R. 1967. Von Thunen and Urban Sprawl. Annals of the Association of American Geographers 57(1): 7287. Spedding, C.R.W. 1988. An introduction to agricultural systems 2nd ed. London ; New York: Elsevier Applied Science : Sole distributor in the USA and Canada, Elsevier Science Pub. Co. Symons, L. 1979. Agricultural geography Rev. ed. Boulder, Colo.: Westvi ew Press. Taylor, M.R. and Brester, G.W. 2005. Noncash Income Transfers and Agricultural Land Values. Review of Agricultural Economics 27(4): 526541. Taylor, R.A. 1995. Rebel storehouse : Florida in the Confederate economy Tuscaloosa: University of Alabama Press. Thrall, G.I. 2002. Business geography and new real estate market analysis Oxford ; New York: Oxford University Press. Thunen, J.H.v. and Hall, P.G. 1966. Isolated state; an English edition of Der isolierte Staat 1st ed. Oxford, New York,: Pergamon Press. Todd, W.R. 2002. Using GIS and remote sensing to locate cattledipping vats in Alachua County, Florida.Thesis (M S ): University of Florida, 2002. USDA. 2009. In National Agriculture Statistics Service USDA. USGS. 1997. National atlas of t he United States. In. USGS, Reston, Va. Varner, T. and Otto, D. 2008. Factors Affecting Sales at Famers' Markets: An Iowa Study. Review of Agricultural Economics 30(1): 176189. Veeck, G., Che, D., and Veeck, A. 2006. America's Changing Farmscape: A Study of Agricultural Tourism in Michigan The Professional Geographer 58(3): 235248.

PAGE 105

105 Volpe, R.J. and Lavoie, N. 2008. The Effect of Wal Mart Supercenters on Grocery Prices in New England. Review of Agricultural Economics 30(1): 4 26. Wehrwein, G.S. 1942. The Rural Urban Fringe. Economic Geography 18(3): 217228. Weldon, R.N., Moss, C.B., and Erickson, K. 1993. The distribution of farm equity in the U.S. : an entropy approach. Gainesville, Fla.: Food and Resource Economics Dept., Institute of Food and Agricul tural Sciences, University of Florida. Willock, J., Deary, I.J., Edwards Jones, G., Gibson, G.J., McGregor, M.J., Sutherland, A., Dent, J.B., Morgan, O., and Grieve, R. 1999. The role of attitudes and objectives in farmer decision making: Business and envi ronmentally oriented behaviour in Scotland. Journal of Agricultural Economics 50(2): 286303. Wolpert, J. 1964. The Decision Process in Spatial Context. Annals of the Association of American Geographers 54(4): 537558. Wu, J. and Gopinath, M. 2008. What Causes Spatial Variations in Economic Development in the United States? American Journal of Agricultural Economics 90(2): 392408. Zieger, R.H. 1997. Southern labor in transition, 19401995. 1st ed. Knoxville: University of Tennessee Press.

PAGE 106

106 BIOGR APHICAL SKETCH Edward Drannbauer is a PhD. Candidate in the D epartment of Agricultural and Biological Engineering at the University of Florida specializing in a gricultural s ystems m anagement. His research interests include business geography, agricultural and occupational safety, biosecurity, and resource economics Born and raised in Daytona Beach, Edward completed a Bachelor of Science (BS) in a gricultural o perations m anagement from the University of Florida. After graduating in 1997, Edward moved to Tokyo, Japan and worked there for several years as a training consultant and scuba instructor. He also spent a year in Edinburgh, Scotland where he completed a Master of Business Administration (MBA) in 2003. After returning to Florida, he completed a Mast er of Science (MS) in f ood and r esource e conomics in 2007. Edward currently resides in Princeton, NJ with his fiance Carissa, and two shepherds Tiny and Dea.