The Pennsylvania State University. The Graduate School. College of Earth and Mineral Sciences THE MYSTERY OF THE MOBILE HOME:

Transcription

1 The Pennsylvania State University The Graduate School College of Earth and Mineral Sciences THE MYSTERY OF THE MOBILE HOME: A GEOGRAPHIC INVESTIGATION OF MOBILE HOMES IN RURAL PENNSYLVANIA A Thesis in Geography by Destiny D. Aman 2008 Destiny D. Aman Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science August 2008

2 ii The thesis of Destiny D. Aman was reviewed and approved* by the following: Brent Yarnal Professor of Geography Thesis Advisor Deryck W. Holdsworth Professor of Geography Karl Zimmerer Professor of Geography Head of the Department of Geography *Signatures are on file in the Graduate School

3 iii ABSTRACT In rural Pennsylvania counties, mobile homes make up nearly 11 percent of the total housing stock and serve as a significant source of affordable housing. Despite the importance of this housing type, local and state-level policymakers seeking to address issues relevant to mobile home residents including land tenure, spatial restriction, and vulnerability to hazards are limited by data availability and accuracy problems. This thesis attempts to overcome these data limitations by developing and implementing a methodology to analyze thematic and geographic data on mobile homes collected from county tax assessment offices. Research findings indicate that mobile homes in rural Pennsylvania are typically aged structures in below-average condition on leased land. Ground-truthing observations lend texture to this analysis by identifying multiple uses for this housing type across the study counties. Based on these different uses, the thesis develops a taxonomy of mobile home contexts in rural Pennsylvania. In addition, the research identifies several issues affecting mobile home residents in rural Pennsylvania notably, many mobile home residents in the study counties may be vulnerable to housing insecurity, escalating energy costs, and hazards including flooding and fire. Since mobile homes are expected to be an important source of affordable housing in rural communities for years to come, rural policymakers can use this information to address the special needs of their communities.

4 iv TABLE OF CONTENTS List of Figures...vi List of Tables...viii Acknowledgements... x Chapter 1: Introduction The Relevance of Mobile Home Study The Mystery of the Mobile Home Study Area: Rural Pennsylvania Research Goal & Questions Thesis Structure Chapter 2: Methods Introduction Phone Survey of County Offices Analysis of Tax Assessment Data Geographic Data Analysis Recapitulation Chapter 3: Results Introduction Analysis of Tax Assessment Data Geographic Data Analysis A Taxonomy of Mobile Homes in Rural Pennsylvania Recapitulation... 77

5 v Chapter 4: Discussion Introduction The Nature of Mobile Homes in Rural Pennsylvania Issues of Land Ownership, Spatial Restriction & Hazards Vulnerability Data Accessibility Issues Recapitulation Chapter 5: Conclusions Introduction Summary Future Research References... 90

6 vi LIST OF FIGURES Figure 1.1 Rural Pennsylvania Counties Figure Rural Pennsylvania Counties Chosen for Tax Assessment Analysis Figure 2.2 Differing Tax Parcel Number Formats Figure 3.1 Lycoming County Mobile Home Density Based on Tax Parcel Data Figure 3.2 Adams County Mobile Home Density Based on Tax Parcel Data Figure 3.3 Indiana County Mobile Home Density Based on Tax Parcel Data Figure 3.4 McKean County Mobile Home Density Based on Tax Parcel Data Figure 3.5 Monroe County Mobile Home Density Based on Tax Parcel Data Figure 3.6 Venango County Mobile Home Density Based on Tax Parcel Data Figure 3.7 Wayne County Mobile Home Density Based on Tax Parcel Data Figure 3.8 Adams County Mobile Home Parcels Intersecting the 100-Year Floodplain. 59 Figure 3.9 Large Mobile Home Parks Clustered in Lycoming County Figure Air Photo of Lycoming County Mobile Home Parcels Figure 3.11 Lycoming County Mobile Homes Along Major Roadways Figure 3.12 Large Mobile Home Park Located Behind Factory in Lycoming County Figure 3.13 Timberend Estates in Lycoming County Figure 3.14 Small Mobile Home Park in Adams County Figure 3.15 Meadowbrook Mobile Home Park in Lycoming County Figure 3.16 Medium-Sized Mobile Home Park in Northern Lycoming County Figure 3.17 Mobile Home as Big Land Rural Housing in Lycoming County... 72

7 vii Figure 3.18 Mobile Homes as Big Land Rural Housing in Indiana County Figure 3.19 Mobile Home on a Farmstead in Lycoming County Figure 3.20 Fishing Cabins in Lycoming County... 76

8 viii LIST OF TABLES Table 2.1 Overview of Mobile Home Data in Rural Pennsylvania Counties Table 2.2 Data Costs for 12 Selected Counties Table 2.3 Selected Demographic Information on Counties Chosen for Tax Assessment Analysis Table 2.4 Tax Assessment Variables Relevant to Mobile Homes Available from 12 Selected Rural Counties Table 2.5 Percent of Mobile Home Tax Assessment Records that Did Not Relate to Parcel Map Table 3.1 Number of Mobile Homes Per County Tax Assessment Data Table 3.2 Year Built for Mobile Homes in Available Rural Counties Table 3.3 Acreage for Available Rural Counties Table 3.4 Mobile Home Class for Available Rural Counties Table 3.5 Mobile Home Condition for Available Rural Counties Table 3.6 Mobile Home Building Type for Available Rural Counties Table 3.7 Year Built for Single-wide Mobile Homes in Available Rural Counties Table 3.8 Year Built for Double-wide Mobile Homes in Available Rural Counties Table 3.9 Condition of Single-wide Mobile Homes in Available Rural Counties Table 3.10 Condition of Double-wide Mobile Homes in Available Rural Counties Table 3.11 Measures of Central Tendency for Mobile Homes with Acreage, by Type Table 3.12 Percentage of Mobile Homes in Rural Municipalities Table 3.13 Parcels with Multiple Mobile Homes... 56

9 ix Table 3.14 Percentage of Mobile Homes in Rural Municipalities Table 3.15 Mobile Home Parcels Intersecting the 100-Year Floodplain... 59

10 x ACKNOWLEDGEMENTS I would like to first thank the members of my thesis committee: Brent Yarnal, Deryck Holdsworth, and Marylee Taylor for their thoughtful support and encouragement throughout the research process. I am grateful to have been a part of an incredibly supportive graduate student community here in the Penn State Geography Department. Mamata Akella, David Fyfe and Darrell Fuhriman gave me invaluable technical advice on this thesis, and the CIRA crew including Tim Frazier, Kieron Branche and Jess Whitehead offered a virtually unending supply of moral and professional support. I owe many thanks also to the Penn State chapter of Supporting Women In Geography (SWIG) for building an inspiring and comfortable space to explore, celebrate, challenge and change the experiences of women in geography. I would like to give special acknowledgment to the Center for Rural Pennsylvania and the Penn State Department of Geography for providing financial support for this research. I also owe a huge debt to the many rural Pennsylvania county officials who took time out of their busy days to answer many more questions than they ever expected from me. Finally, I would like to thank Rhyena Laney and Dolly Freidel for helping me fall in love with geography and for inspiring me to follow my passion. I hope someday to give others what you have given me. I dedicate this thesis to Chris, for always giving me a soft place to land.

11 1 Chapter 1 INTRODUCTION 1.1 The Relevance of Mobile Home Study As the housing choice of nearly 20 million Americans nationwide, mobile homes 1.1 are an important but understudied feature of the American housing landscape (U.S. Census 2005a). Having grown steadily since World War II, new mobile home construction now accounts for one out of every seven new housing units in any given year (HUD 2002). This growing popularity has been attributed to mobile homes relative affordability, availability, and flexibility, as compared to traditional site-built housing (Genz 2001). Mobile home residents, however, also face unique challenges related to this housing type including land tenure, ownership, institutional barriers, and an increased vulnerability to hazards Affordability In a historically tight American housing market, housing advocates have recognized the mobile home as an increasingly important component of the unsubsidized affordable housing sector (HUD 2002). The relative affordability of mobile homes puts homeownership within reach of millions of households and is perhaps the single largest contributor to their increasing popularity (Genz 2001). For example, in 2005, the average 1.1 Mobile home was the official term used for a manufactured home built before June 1976, when the name formally changed to manufactured home. In 2005, the majority of residents living in this housing type referred to them as mobile homes. Only 16 percent called them manufactured homes. This work will refer to these dwellings as mobile homes (Foremost Insurance Group 2005).

12 2 price of a single-wide 1.2 mobile home in Pennsylvania was $38,900 and the average price of a double-wide was $63,600 (U.S. Census 2005b), while the average cost of a new, unoccupied conventional site-built dwelling was $165, (U.S. Census 2005a). Because mobile home owners on average have significantly lower incomes than owners of site built homes (Genz 2001), and affordable housing alternatives continue to dwindle, the relative importance of mobile homes in the affordable housing market can be expected to grow Availability & Flexibility The availability and flexibility of mobile homes also contributes to their popularity in the American housing market. Mobile homes can be shipped nearly anywhere in the 48 contiguous United States, including locations where it would be difficult or expensive to find a builder or supplies. In addition, due to their smaller overall square footage, mobile homes require less physical space than most newly constructed site-built homes. Moreover, because they rest on a chassis, mobile homes do not require a foundation or a basement meaning that it is possible to site them nearly anywhere permitted by building codes (NAHB Research Center 2000). Ironically, the same unique qualities that make the mobile home a popular alternative housing choice for 20 million Americans also create unique challenges for its residents. Mobile home owners face issues of land tenure, financing and ownership, 1.2 The terms single-wide and double-wide refer to the width of the mobile home, which in turn affects the area of the structure. Restricted by state limits for towable vehicles, early mobile homes were eight feet wide and of variable lengths generally no longer than 35 feet (Wallis 1991). Today, newly constructed single-wide mobile homes are typically 15 feet wide and can be as long as 76 feet. Double-wide mobile homes, as the name suggests, are merely two single-wide structures fastened together. The average area of a single-wide mobile home in the United States is 1245 square feet, and the average area of a double-wide mobile home is 1605 square feet (U.S. Census 2000). 1.3 It should be noted that while the cost of a site-built dwelling includes the cost of land, the cost of a new mobile home does not include associated land. If the home is sited in a mobile home park, the mobile home owner typically leases a space from the park. Mobile homes may also be placed on land owned separately by the mobile home owner.

13 3 spatially restrictive institutional barriers, and increased vulnerability to hazards not experienced by traditional site-built homeowners. Often these issues arise in part from the mobile home s historical classification as a travel trailer, the vestiges of which persist even though mobile homes have become a more permanent housing solution for millions of residents Land Tenure & Ownership Land tenure is one quality that distinguishes the mobile home from virtually any other type of permanent housing. Developed in the 1920s as recreational travel trailers, by definition the earliest mobile homes were designed to be transient, and land was not included in their purchase (Wallis 1991). Mobile homes were sited in campgrounds and parks designed for temporary use. However, over time, mobile homes have increasingly become less mobile. In fact, according to the U.S. Census, 60 percent of mobile home owners in 2005 stated that their mobile home had never been moved (U.S. Census 2005a). While mobile home owners may purchase land and subsequently place their homes upon it, roughly half of the mobile home owners in the United States lease the land for their mobile home (Foremost Insurance Group 2005). This unique land tenure situation produces distinctive challenges and vulnerabilities for mobile home owners. In Wheel Estate: The Rise and Decline of the Mobile Home, Wallis (1991) explains that as early as the 1950s and 1960s mobile home park tenants (who typically owned their home but rented their park lot) could expect to be required to comply with a specific set of park rules, without the protection of a lease or state laws to prevent eviction on frivolous grounds (Wallis 1991, p. 194). Extra fees and restrictions on visitors and utilities were

14 4 common. Wallis suggests that these abusive practices arose in part from the absence of the clear and significant rights afforded to traditional homeowners, and were often exacerbated by the scarcity of park space. For example, where there was high demand and few vacancies, a park operator could force prospective tenants to purchase a unit from the operator or from selected dealers. The enforcement of other rules could include eviction: exceptionally disruptive tenants might find their homes towed out of the park within a day (Wallis 1991, p. 195). Although landlord-tenant protections have been extended to mobile home owners since that time, land-leasing mobile home owners still encounter hardships related to land tenure. High demand for limited park space continues to be an issue for mobile home park residents today Record-breaking land values in the late 1990s to mid-2000s tempted many park owners to sell their land to developers, exacerbating the shortage of space nation-wide. Across the country, park sales forced mobile home owners to relocate their homes (often at their own expense) to other mobile home parks (Kramer 2007; Kreller 2007; Nielson-Stowell 2006; Saemann 2007). Complicating this issue are the regulations governing age and condition maintained by remaining parks: even if a resident s home is in moveable condition (many are not), the owner of an older mobile home may have trouble locating a park that is willing to accept it. In perhaps the most extreme example of the relative housing insecurity faced by mobile home owners, those that cannot move their homes for one reason or another are forced to abandon them (O Donnell 2008).

15 Institutional Issues & Barriers Institutionally, mobile home owners face unique issues in financing and real estate. As opposed to the traditional mortgages offered for site-built homes, financing procedures for most mobile homes today are similar to those for automobiles yet another holdover from the housing type s origins as a travel trailer (NAHB Research Center 2000). In what has been cited as a salient example of the poor pay more effect, this specialized type of lending results in interest rates up to five percentage points more than a regular mortgage (Genz 2001). Additionally, manufactured housing lenders specialize in subprime lending, which itself can be as much as three percentage points higher (Genz 2001; Renuart 2004). Mobile home residents are also disproportionately affected in the wake of the current U.S. housing crisis, in which the American housing market is experiencing high rates of foreclosure (RealtyTrac 2008). Mobile home foreclosure is a much more rapid process than traditional home foreclosure, and can be completed in as little as 30 days (Capozza et al. 2005). Traditional foreclosure by contrast typically takes several months, during which time a debtor may be able to halt the process by paying delinquent payments and fees. Early legislative responses by some states to address the foreclosure crisis have not engaged with the special needs of mobile home residents, leaving little relief for mobile home residents at risk of losing their homes (Davis 2008). Because mobile homes do not appreciate in value at the same rate as traditional site-built houses (in fact, two thirds of mobile homes actually depreciate, much like an automobile), mobile home owners are denied much of the financial flexibility and opportunity offered by traditionally financed site-built housing (Capozza et al. 2005;

16 6 Genz 2001). Although touted as an affordable alternative to site-built housing, owning a mobile home can put residents at risk for a number of other hidden costs. For example, Salamon and MacTavish (2006) found that high energy costs associated with older, poorly insulated mobile homes can quickly consume small household budgets. Poor quality structural features including fixtures, trim, and floor coverings lead to high costs of repair and maintenance for mobile home residents (Consumers Union 2002), and those living in the oldest mobile homes may face a losing battle trying to maintain homes designed for short-term use. Other institutional barriers encountered by mobile home owners have a distinctly geographic bent. Historically, mobile homes were often restricted to the campgroundlike parks that were established to serve them during their travel trailer days (Wallis 1991). By the end of World War II, mobile homes began to be perceived as a threat to both real estate values and to a community s moral character, and thus in many communities the parks themselves were limited to commercial and industrial areas (Wallis 1991). Over time, Wallis argues, unfavorable zoning regulations in cities pushed mobile home development to more rural locations. It is possible that mobile home owners are today faced with many of the same spatial and social limitations encountered by their historical counterparts. In a 2004 study using Geographic Information Systems (GIS), Wubneh and Shen (2004) found that all things being equal, proximity to a mobile home decreased the value of nearby site-built residential properties in North Carolina. According to MacTavish and Salamon (2006), social inequality continues to be an issue for mobile home residents, especially in the face of increasing economic stratification. In a 1998 survey of 199 North Carolina county and

17 7 municipal planning directors, Lowrey (1998) found that administrative districts can differ in the level of zoning restriction pertaining to mobile homes; municipalities tend to be much more restrictive toward mobile homes than counties. These restrictive zoning regulations can force mobile homes farther away from community services. In fact, in a rural context, Shen (2005) discovered that compared to other forms of housing, mobile homes were farthest away from all the community services included in the study including hospitals, health care clinics, and police and fire stations. For non-critical (but still essential) community service locations including banks, restaurants, shopping centers and daycare centers, mobile residents were forced to travel two to three times longer distances than households in single-family homes (Shen 2005). In addition to the social stigma associated with the housing type, from a strictly financial perspective it is easy to understand why local governments do not have incentive to create a mobile home-friendly environment. Because mobile home values are much lower than those of other forms of housing and often do not have associated taxable land, communities receive much less property tax revenue from mobile homes Increased Vulnerability to Hazards An important issue faced by mobile home owners is their increased vulnerability to environmental hazards. In their work on vulnerability, geographers and sociologists recognize that as part of the built environment, mobile homes offer comparatively little reliable protection in the face of hazards of every type (Borden et al. 2007; Cutter and Emrich 2006; Montz and Tobin 2005; Morrow 1999; Whitehead et al. 2000). In addition, mobile homes can in some cases be more exposed to certain hazards. Compared to other

18 8 forms of housing, for instance, a higher percentage of mobile homes can be found in flood zones (Shen 2005). This enhanced exposure to floods may date back to their historically mobile nature: since it was assumed that mobile homes could easily be moved should a flood occur, they were often permitted in flood plains (Wallis 1991). Mobile homes also experience a much higher rate of fire deaths than other home types, with older mobile homes at higher risk (Parker et al. 1993; Runyan et al. 1992). Many mobile home owners are vulnerable to hazards not only because of their physical exposure in structurally inferior housing and their proximity to certain hazards, but also due to their lower income status. Poverty has been identified as an important component of social vulnerability (Fothergill and Peek 2004), and since mobile home owners have lower incomes on average than owners of site-built houses, it follows that they are subject to a double whammy of vulnerability (Morrow 1999). Residents living in older mobile homes are particularly vulnerable because state and federal regulations requiring the use of special roof connectors and other devices to strengthen homes against hurricanes and other hazards have only recently been required and applied to mobile homes built after the regulations were implemented. In a recent deadly tornado outbreak in Florida, for example, all 21 fatalities were in mobile homes that were over 16 years old, predating the more stringent Florida tie-down regulations (Herald Tribune 2007). Although the increased vulnerability to hazards experienced by mobile home residents has been evident for some time in the academic community, there has been little progress on the part of local, state, and federal policies to mitigate this issue. Leslie Chapman-Henderson, president of the Federal Alliance for Safe Homes, recently called mobile homes one of the true ticking time bombs (Swider 2007). Important to this

19 9 research, community officials cite the lack of data on mobile homes as a prohibiting factor to further mitigative action (Tillman and Swider 2007). 1.2 The Mystery of the Mobile Home While mobile homes may have certain advantages over traditional site-built houses, they also have problems often related to their traditional roles as transient and temporary housing. Despite these issues, however, the role of the mobile home as an important form of housing in the United States shows no sign of deteriorating, creating a mounting need to understand the housing type. Despite the rich potential for academic study in this area, investigation of mobile homes and their residents is sparse in the academic literature. Recent work on mobile homes primarily includes historical overviews of mobile home development (Hart et al. 2002; Wallis 1991), ethnographic studies of community and human development in mobile home parks (Hackenberg and Benquista 2001; MacTavish 2001, 2006; MacTavish and Salamon 2006), and anecdotal references to mobile home residents as a population vulnerable to both natural hazards and housing insecurity (Borden et al. 2007; Cutter and Emrich 2006; Davis 2006; Morrow 1999; Whitehead et al. 2000; Williams et al. 2005). One explanation for the lack of academic research about mobile home residents could be the absence of accessible data. The three main sources of statistical data on mobile homes include the U.S. Census Bureau, the Manufactured Housing Institute, and the Foremost Insurance Group, each of which produce data at the national and state-level, but have distinct limitations. The U.S. Census Bureau collects basic data on mobile home residents in the decennial

20 10 census. However, the housing type category is found only on the census long form, where data is collected from one out of every six or seven households and is weighted to represent the entire population. This process introduces a certain amount of error, which can be particularly significant for rural areas where estimates can vary by up to 50 percent for even the most basic information, such as total number of mobile homes (U.S. Census 2007b). The Manufactured Housing Institute follows trends in the mobile home business community including the number of new mobile home shipments and overall market share, but they do not research demographic characteristics of mobile home owners (Manufactured Housing Institute 2007). The Foremost Insurance Group uses a market-research firm to survey a non-random panel of mobile home users, seeking data that reveal many characteristics of mobile homes and mobile home residents not sought by the Census Bureau especially those data that relate to the mobile home market (Foremost Insurance Group 2005). However, The Foremost Insurance Group does not release their data at scales smaller than the state. Additionally, since 2005 the survey has been conducted solely online. This methodology is problematic because it excludes mobile home owners who do not have access to the Internet, perhaps biasing the sample. It also reduces comparability between recent data and previous surveys, which were conducted via mail (Foremost Insurance Group 2005). Although each of these sources produce valuable information for their own purposes, they fall short of producing the type, quality, or geographical resolution of data required by policymakers to address many of the issues mentioned at the beginning of this proposal. Nevertheless, there is one very rich source of data on mobile homes that can be found at the county and local level. Each county in the United States collects

21 11 detailed data on mobile homes for local taxation and planning purposes. Mobile home owners are required to register their homes with the local tax office, so this data source provides comprehensive coverage of the mobile homes in any given county. Data collected can include information on year built, size, land tenure, value, and more. The quality of these data is limited in that they are unstandardized and vary from state to state and county to county in their quality, format, and accessibility. One reason for the lack of accessible information on this type of housing and its residents could be that mobile homes are notoriously difficult to categorize they have historically occupied a grey area between house and automobile. Mobile homes are built, sold, and financed like automobiles, but increasingly they are taxed and regulated like traditional homes (Hart et al. 2002). In addition, because many owners do not own the land beneath their mobile home, uncertainty often exists even as to whether a mobile home is private property or real property (Wallis 1991, pp ). These ambiguities make collecting data on mobile homes difficult, particularly at smaller spatial scales where, to establish laws and regulations for insurance, financing, zoning, and taxing, public and private sector organizations have resolved classification differences individually. The availability of this information, therefore, varies spatially and presents a challenge for researchers attempting to collect and analyze data on mobile homes. Academic investigation into the demographic and geographic characteristics of mobile home residents would increase understanding of how these characteristics change over space and of which issues affect mobile home residents. Only detailed analysis of this housing type and its residents can unearth this type of policy-relevant information,

22 12 but it is clear that such investigation must include a methodology for addressing data limitations. 1.3 Study Area: Rural Pennsylvania Pennsylvania has great geographic variety. For example, it is the sixth most populated, yet one of the most rural states in the country: sixty-four (64) percent of the Commonwealth s 2,576 municipalities and 72 percent of its counties are considered rural (Center for Rural Pennsylvania 2007). According to Lewis (1995), from the time of earliest settlement, Pennsylvania s geographic variety was a major asset to its people, providing a broad variety of economic opportunity, so that over time the state became a major national force in agriculture, mining, transportation, and manufacturing (p. 2). This geographic variety exists today in the counties of rural Pennsylvania (Figure 1.1), from the oil fields of Venango County on the Appalachian Plateau in the northwest of the state, to the rich agricultural soils of the Adams County Piedmont in the southeast. This variety fuels the livelihoods of 2.8 million people engaged in nearly every economic endeavor possible from coal mining and forestry to farming, manufacturing, tourism and services. Rural Pennsylvania is the canvas upon which the bulk of one of America s richest geographic mosaics is painted (Lewis 1995, p. 1). The topic of mobile homes is particularly relevant in rural 1.4 Pennsylvania, where mobile homes make up roughly 11 percent of housing stock and are second in quantity only to single-family detached houses (U.S. Census 2005a). Statewide, however, mobile homes make up less than 5 percent of housing stock, illustrating a rural-urban divide in 1.4 To determine whether a county is rural this thesis will use The Center for Rural Pennsylvania s definition in which a county or school district is rural when the number of persons per square mile within the county or school district is less than 274 (Center for Rural Pennsylvania 2007).

23 13 choice of housing. A rural context also provides an appropriate backdrop for the study of mobile homes, which have been described as a new rural community form (MacTavish and Salamon 2001, p. 487) and products of the rural ghetto (Davidson 1996, p. x). Mobile home residents of rural Pennsylvania are likely to face many of the same problems as mobile home residents in other states, such as land tenure and ownership issues and natural hazards (particularly with regard to flooding). Furthermore, mobile home data quality and accessibility are highly variable in rural Pennsylvania: some rural counties keep digital records while others still operate with paper systems. In sum, due to its great geographic diversity, its proportionally large mobile home resident population, this population s exposure to a variety of home ownership, land tenure, and natural hazards issues, and the varying degrees of mobile home data quality and accessibility, rural Pennsylvania is an appropriate stage for research on mobile homes.

24 Figure 1.1: Rural Pennsylvania Counties 14

25 Research Goal & Questions The purpose of this thesis is to develop the understanding of mobile homes in rural Pennsylvania needed to address policy considerations related to land and home ownership, spatially restrictive institutional barriers, and vulnerability to natural hazards. The goal of this thesis, therefore, is to describe the characteristics of the mobile homes in rural Pennsylvania and in the process to develop and implement a methodology that will allow others to build on this knowledge. To reach that goal, this thesis attempts to answer the following three research questions: 1. What is the physical nature of mobile homes in rural Pennsylvania? 2. What issues of land ownership, spatial restriction, and hazards vulnerability affect mobile home residents in rural Pennsylvania? 3. How might mobile home data accessibility issues be resolved? 1.5 Thesis Structure This thesis has five chapters. Chapter 1 discussed the relative importance of mobile homes in the American housing market as well as some of the unique challenges facing residents of mobile homes. It then reviewed the scarcity of literature regarding mobile homes and related it to the difficulty in categorizing this housing type and in obtaining accurate data about mobile homes and their residents. The chapter identified rural Pennsylvania as an appropriate stage for mobile home study and then concluded with the research purpose, goals, and questions guiding this thesis.

26 16 Chapter 2 addresses the analysis of thematic and geographic data on mobile homes obtained from rural county tax assessment offices. It outlines the variables used to determine the 12 rural counties featured in the tax assessment analysis and discusses the methods for obtaining, processing, and analyzing the data. It then introduces the seven counties available for geographic analysis and details the methods used to produce and analyze the geographic data. Chapter 3 presents the results from the analysis of thematic tax assessment data then focuses on the results of the geographic analysis, then Chapter 4 discusses the significance of the results as they pertain to the stated purpose and goal of the thesis. This chapter includes specific policy recommendations for addressing important issues regarding mobile homes and their residents mentioned in Chapter 1. Finally, Chapter 5 presents a summary of the thesis results and identifies potential data limitations. The chapter concludes by identifying areas where future research on this subject might be useful.

27 17 Chapter 2 METHODS 2.1 Introduction Most data on mobile homes are scarce and problematic. Hard-to-access data collected from county tax assessment offices, however, can provide a rich description of this housing type in some cases, down to the type of siding and foundation of each mobile home in the county. This chapter discusses the processes employed to collect and process data on mobile homes, beginning with a phone survey of county offices. It explains how the results of the phone survey were used in conjunction with socioeconomic and geographic information to determine the counties selected for the tax assessment analysis and the geographic analysis. The chapter goes on to describe the methods used to obtain and analyze thematic tax assessment data from the chosen rural counties. It then explains how these tax assessment data were added to a Geographic Information System (GIS) to produce a geographic analysis of mobile homes, all the while taking note of data limitations and discussing how such limitations were addressed. 2.2 Phone Survey of County Offices To determine the condition and availability of mobile home data at the county level, telephone inquiries were made to each of the 48 rural counties in Pennsylvania (Figure 1.1). The process required contacting multiple offices in each county to locate and obtain information about available data on mobile homes. Contact generally began with the county tax assessment offices and progressed to county mapping, Geographic

28 18 Information System (GIS), or Management Information System (MIS) departments. Rural county offices often keep complementary information in different formats, and some counties outsource certain data storage and output operations to private companies. This survey provided information on data categories, costs, data request procedures, and other pertinent information Obtaining Tax Assessment Data Offices were asked if they store their tax assessment information in a digital database. Most county offices keep tax assessment data in digital form, although a few still operate paper card systems. For those offices compiling digital tax assessment data, it was necessary to identify the data categories the county keeps for mobile homes and to establish how it codes mobile homes. The type of coding system used by the office determines whether it is possible to extract mobile home data separately from data for other residences. To be accessible for further study in this project, it was important that a given county could extract all of the mobile home data separate from data on other taxable dwelling types. The phone survey showed that nearly every county office uses a different coding system for their mobile home data. Some of these systems are easy to query particularly if the mobile home code is unique and located in a suitable alphanumeric field, such as a numeric code located in a building type or land use field. Some systems have multiple mobile home codes located in fields populated with other information, such as the word mobile or trailer in a description field, or a T as a suffix to the parcel number. Such systems are much more difficult to query and heighten the potential for extraction error.

29 19 Other data issues became evident during the phone survey. For example, the term mobile home varies in meaning from county to county both in interpersonal communications and in the coding systems themselves. In some counties, single-wide mobile homes are coded uniquely as mobile homes or trailers in the digital databases, but double-wide mobile homes are coded as regular residences. Other data variations include coding mobile homes on a foundation or on more than 10 acres as a regular residence or an agricultural property, respectively. Some counties admitted that their mobile home data are significantly flawed due to data entry error and the difficulties inherent in classifying mobile homes. Some counties explained that although all fields were available in digital form, some of the variables were stored in different digital databases and were not, in fact, extractable. These data quality issues excluded most Pennsylvania rural counties from further study. The double-wide issue was particularly significant, affecting nine counties that otherwise would be suitable for further consideration. Of the 48 rural counties, only 18 counties had tax assessment data in a form suitable for extracting mobile home information (Table 2.1). Having accessible tax assessment data was only one piece of this project, however. To enable geographic comparison, it was important that most of the counties selected for analysis also had accessible geographic tax parcel data Geographic Data The availability of geographic tax parcel data in the rural counties is limited. Whereas most rural counties have some type of digital tax assessment information on mobile homes (if not in a format that could be used for this study), roughly half of the

30 20 Table 2.1: Overview of Mobile Home Data in Rural PA Counties County Name Available Digital Tax Assessment Data Available Digital Geographic Data Comments on Incomplete Data Adams Y Y Armstrong Limited N MHs on the same property as a house are not uniquely coded Bedford Limited N DWs are coded as site-built houses Blair Limited N MHs on private property are coded as site-built houses Bradford Y Limited Tax parcel maps are outdated and do not agree with tax assessment data (split parcels, etc) Butler N N No digital data paper card system + paper maps only Cambria Limited N MHs on private property and MHs on more than 10 acres are coded as site-built houses Cameron N N No digital data paper card system + paper maps only Carbon Limited Y Digital data is available but county officials would not return phone calls to define assessment variables Centre Limited Y DWs are coded as site-built houses Clarion Limited N DWs are coded as site-built houses Clearfield Limited Y It is possible to extract all MHs in Co, but system can only output names and addresses Clinton Limited Y DWs are coded as site-built houses Columbia Limited Y Most tax assessment variables are stored in a separate database and are not extractable Crawford N N New computer system is not updated with all information yet Elk Y Limited This data set does not include MHs that are rented out (less than 1 percent of total MHs) Fayette Y N Paper maps only digital in progress (finished in approx 3-5 months) Forest Y N Paper maps only digital in progress (finished in approx 9 months) Franklin Limited N It is possible to extract all MHs in Co, but system can output limited information only (parcel #, name, address, location, property type) Fulton Limited Y Most tax assessment variables are stored in a separate database and are not extractable Greene Y Limited Insufficient quality. Huntingdon Limited Y MHs on more than 10 acres are coded as site-built houses Indiana Y Y Jefferson Limited N Output would not include MHs on private property Juniata Limited Y DWs are coded as site-built houses Lawrence Limited Y Some DWs are coded as site-built houses (if on basement or crawl-space) Lycoming Y Y McKean Y Y Mercer N N No digital data paper card system + paper maps only Mifflin Limited Y Some DWs are coded as site-built houses (if on a basement) Monroe Y Y Montour Y N Paper maps only

31 21 Table 2.1 Continued: Overview of Mobile Home Data in Rural PA Counties County Name Available Digital Tax Assessment Data Available Digital Geographic Data Comments on Incomplete Data Northumberland Limited N MHs on property with a house are not uniquely coded Some modular homes are coded as DW MHs, MH courts are not included in output file Perry Y Limited Insufficient quality Pike Limited Y MHs on foundation are coded as site-built houses Potter Limited N DWs are coded as site-built houses, land use code is not well-maintained Schuylkill Limited Y Conflicting information: output does not uniquely code MHs Snyder N N No digital assessment data paper card system County has internal access to digital parcel maps, but they are not available to the public Somerset Y Limited GIS data is not accurate Map # s do not match assessment data Sullivan Y N Paper maps only digital in progress Susquehanna Limited N 75 percent of DWs are coded incorrectly as manufactured Tioga Y Limited Insufficient quality Union Limited Y Many MHs are incorrectly coded as site-built houses Venango Y Y Warren Y N Paper maps only digital in progress (finished in 3-4 months) Washington N N No digital data paper card system + paper maps only Wayne Y Y Wyoming Y N Paper maps only digital in progress

32 22 counties do not have any accessible tax parcel data in GIS form. Thirteen counties are currently in the process of digitizing parcel maps into a GIS, but digitizing is a timeconsuming procedure often requiring several years to complete particularly for resource-strapped counties. Some counties have worked around these issues by digitizing their tax parcel maps in conjunction with the building of 911 emergency networks. A few counties have internal access to digital tax parcel data, but have not yet made that information available to the public, and some counties have inaccurate or outdated geographic data. Of the 48 rural counties, 20 have complete and accessible geographic tax parcel data (Table 2.1) Data Costs The cost of acquiring both digital assessment data and geographic data in the rural counties is extremely variable. Tax assessment data costs range from no charge to nearly $300 for essentially the same data. Geographic data have a much greater cost range, varying from no charge to $8000 for a single countywide tax parcel layer. Many county offices indicated that it is possible to request an educational or government discount when ordering the data, and several offered the data at no charge for Penn State University. For the 12 counties that had both accessible tax assessment and geographic data, most of them quoted cumulative costs of less than $300 for both datasets (Table 2.2).

33 23 Table 2.2: Data Access Costs for 12 Selected Counties, in 2007 Dollars Name Tax Assessment Data Geographic Data Total Adams Bradford Elk Greene Indiana Lycoming McKean Monroe Perry Tioga Venango Wayne

34 Digital Data Ordering While some county offices require specific paperwork (a data order form or a data licensing agreement) in advance, most counties only require a written data request that includes exactly what information is needed and how it will be used. Most county offices with digital information have a turnaround time of less than two weeks to fulfill a data request; most are flexible with regard to available digital output formats that is, they can provide data in Excel, comma delimited, or other general format Compilation of Socio-Economic Information A total of 12 rural counties had tax assessment and geographic data of sufficient quality to be included in the analysis for this thesis. To establish the most important demographic variables needed to represent rural Pennsylvania as accurately as possible, the Center for Rural Pennsylvania, a legislative agency that serves as a resource for rural policy by sponsoring research projects, collecting data, and publishing information about rural Pennsylvania (Center for Rural Pennsylvania 2007) provided key demographic data to this project. These data (Table 2.3) varied from population statistics (total population, percent change, population density, median age) to household characteristics (average household size, percent owner-occupied homes, median value of owner-occupied homes), educational dynamics (percent of population with a Bachelor s Degree or higher, school district expenditures per student), economics (median household income, percent of individuals below poverty level, total assisted rental units, average annual unemployment

35 25 rates), and municipality characteristics 2.1 (number of municipalities per 100,000 residents, percent municipalities with ordinances). These data suggest that the 12 counties represent an adequate variety of the multiple variables reported, including areas with varying levels of growth (or even negative growth in some cases), population density, household composition, education, economics, and municipality characteristics Geographic Distribution In addition to socioeconomic variety, the 12 counties are also geographically dispersed across the state (Figure 2.1). They represent each of the state s major geographic regions and specialize in a variety of economic activities including coal mining and oil drilling (Bradford, Greene, Indiana, Venango), forestry (Bradford, Lycoming, McKean), agriculture (Adams, Bradford, Greene), manufacturing (Elk, Lycoming, McKean), recreation and tourism (Adams, Bradford, Tioga, Wayne) and services (Monroe, Perry, Tioga). The broad geographic signatures of these twelve counties are summarized by geographers contributing to a 1995 study of the state (Abler 1995, Lewis 1995, Miller 1995). 2.1 According to staff at the Center for Rural Pennsylvania, the municipality information serves as a useful proxy for the level of public services available to a rural population (Johnson 2008). In addition, as mentioned in Chapter 1, recent literature indicates that the presence of county or municipal zoning ordinances can have a strong affect on the spatial restriction of mobile homes (Lowrey 1998).

36 26 Table Table 2.3: 2.3 Selected Demographic Information on on Counties Chosen for for Tax Tax Assessment Data Analysis County Name 2000 Population 1 Percent Change Estimated Population 2 Percent Change Population Density 2 Median Age 1 Average Household Size 1 Percent of Occupied Units: Owner-Occupied 1 Median Value Owner-Occupied Home 1 Percent of Pop with a Bachelor s or higher 1 School District Expenditures Per Student, Median Household Income 1 Percent of Individuals Below Poverty Level 1 Total Assisted Rental Units Per 1,000 Population 4 Average Annual Unemployment Rate Average Annual Unemployment Rate Number of Municipalities Per 100,000 Residents 5 Percent Municipalities With Municipal and/or County Zoning Ordinances 5 Adams $ % 3.5% Bradford $ % 5.0% Elk $ % 5.4% Greene $ % 6.4% Indiana $ % 5.6% Lycoming $ % 5.6% McKean $ % 5.7% Monroe $ % 5.3% Perry $ % 4.2% Tioga $ % 5.8% Venango $ % 5.7% Wayne $ % 4.9% Sample Mean $ % 5.2% Mean for 48 rural counties $ % 5.6 % U.S. Census U.S. Census 2007a 3 Pennsylvania Department of Education Pennsylvania Department of Community & Economic Development Pennsylvania Department of Labor & Industry 2007

37 Figure 2.1: 12 Rural Pennsylvania Counties Selected for Tax Assessment Analysis 27

38 Analysis of Tax Assessment Data Following the completion of the phone survey, contacts were made with the 12 rural Pennsylvania counties chosen for further study (Figure 2.1) to arrange to purchase tax assessment and tax parcel data. Requests were sent on university letterhead explaining how the data would be used and asking for all digital tax assessment data fields available for all mobile homes in each county, including single-wide and doublewide mobile homes on leased land and private property, on any type of foundation or basement (or no foundation or basement), and on any number of acres. Upon receipt of the data (which took 15 weeks to accumulate), the process of cleaning and recoding the data for analysis began. Support was required from each county to define their variables because it became apparent that the county offices (and sometimes even individuals within the same office) often use different definitions for any given variable. For example, most tax assessment offices keep two types of variables on age actual year built and effective year built. Effective year built is a figure based on the actual year built, minus the physical depreciation of the structure, which for mobile homes can be quite great. Because year built data were requested generally, several counties sent effective year built instead of the intended actual year built data. The same issue arose with square footage (heated square feet versus exterior building footprint), and value (building value versus improvement value, which also includes porches, garages, outbuildings, and even cabins on the property). Thus, statistical calculations for each county could not be completed until assessment personnel at each office could supply information about each variable, which often required multiple contacts.

39 29 After data recoding and cleaning, the datasets were imported into the SPSS (Statistical Package for the Social Sciences) computer program and basic statistics were derived about mobile homes in each county including actual age, heated square footage, building type, class, condition, and acreage (see Chapter 3 for explanations of these variables). As with the quality of data, the breadth of detail available on mobile homes in each county varies dramatically, from merely acreage and total value in Elk County to the flooring type and number of bathrooms for every mobile home in Greene County. Table 2.4 details the tax assessment variables available from each county that relate to mobile homes. Thematically, the number of cases available for analysis ranged from 7,200 to nearly 46,000 the total number of mobile homes summed over the 12 counties depending on the variable. 2.4 Geographic Data Analysis After deriving descriptive statistics for each county, the tax assessment data were connected to digital tax parcel maps obtained from these counties, thereby allowing a geographic analysis of these data. This connection takes place in a Geographic Information System (GIS) and uses the tax parcel number, an identification number common to both the tax assessment data and the tax parcel map. However, in making this connection it became apparent that there are three major problems with the way most of the counties store their data:

40 30 Table 2.4: Tax Assessment Variables Relevant to Mobile Homes Available from 12 Selected Rural Counties County # of MHs Building Type Age Square Footage Acreage Property/Land Use Type Building Value Land Value Improvement Value Total Assessed Value Condition Class Foundation Type Exterior Type Roof Type Heat Type A/C Type Sewer Type Road Type Water Source Adams X X X X X X X Bradford X X X X X X X X Elk X X X X X Greene X X X X X X X X X X X X Indiana X X X X X Lycoming X X X X X X X X X X X X McKean X X X X X X X X X X X Monroe X X X X X X X X Perry X X X X X X X X X X Tioga X X X X X X X Venango X X X X X X X Wayne X X X X X X X X X X

41 31 1. Tax parcel numbers are not stored in the same format. In many cases, tax assessment and mapping offices in the same county do not use the same standards for data storage, meaning that even if the parcel numbers themselves are the same, their formats are different enough to make the GIS unable to match the numbers. For example, Figure 2.2 shows the different formats of the parcel numbers as stored by (a) a county tax assessment office and (b) the same county s mapping office (b). To connect these numbers in the GIS for any given county, extensive recoding of the thousands of mobile homes is necessary. 2. Mobile homes on leased land are given a unique parcel number that does not refer to the parcel where they are sited. Mobile homes are a unique form of housing because they can be sited on land owned by someone other than the mobile home owner. Consequently, it is necessary to assess the taxes for the land and the structure separately, so the identification number given to mobile homes is not the same as the number on the parcel map where that mobile home sits. Some counties maintain a separate field for the parcel on which a mobile home is sited, but many do not, instead adding a suffix to the number or otherwise altering it. For each of these cases, the county was consulted to determine how to identify the parcels on which these mobile homes are sited and how to recode their identification numbers so as to tie them back to that parcel. This process required considerable recoding, and in some instances the county could not help the researchers make these numbers match. These

42 32 Figure 2.2: Differing Tax Parcel Number Formats a b

43 33 counties (Bradford and Elk) were therefore excluded from the geographic analysis. 3. Digital tax parcel maps were of insufficient quality to be used in analysis. Three counties provided digital tax parcel maps, but it is impossible to connect tax assessment data to these files due to the way they store individual parcels in their files. Thus, these counties (Greene, Perry and Tioga) were excluded from the geographic analysis. Once the data were recoded and cleaned, the tax assessment information was added to the GIS with varying degrees of success. For many reasons, anywhere from 0.01 to 7.6 percent of tax assessment records did not join to the parcel maps in any given county (Table 2.5). In the end, however, it was possible to include seven counties in the geographic analysis: Adams, Indiana, Lycoming, McKean, Monroe, Venango, and Wayne. After adding the tax assessment information to the tax parcel maps in the GIS, maps were created showing the geographic distribution of mobile homes in each of the seven study counties. Municipality and 100-year floodplain layers for the counties were used to determine whether mobile homes are located in rural 2.2 or urban municipalities, and whether they intersect floodplains. On the maps, proportional symbols represent the number of mobile homes on each parcel with the largest showing the locations of mobile 2.2 In determining a municipality s status as rural or urban, I applied the Center for Rural Pennsylvania s municipality definition, in which A municipality is rural when the population density within the municipality is less than 274 persons per square mile or the municipality s total population is less than 2,500 unless more than 50 percent of the population lives in an urbanized area, as defined by the U.S. Census Bureau. All other municipalities are considered urban (Center for Rural Pennsylvania 2007).

44 34 Table 2.5: Percent of Mobile Home Tax Assessment Records That Did Not Relate to Parcel Map County Percent MHs Did Not Relate Owned Land Leased Land Total Adams Indiana Lycoming McKean Monroe Venango Wayne

45 35 home parks. Where appropriate, inset maps show highly clustered areas in more detail. For these inset maps, dot density in which one dot is equivalent to one mobile home is used because proportional symbols can block the spatial patterns of individual mobile homes. The maps make it easy to see the spatial organization of mobile homes in each county and to spot the location of large mobile home parks. To facilitate a comparison of different distributional patterns in each county, descriptive statistics were produced including the percentage of parcels with multiple mobile homes, the average number of mobile homes on those parcels, and the number of parcels with more than ten mobile homes. After generating maps for each of the seven study counties, ground-truthing visits were made to three counties: Lycoming, Indiana, and Adams. County and municipal officials were contacted on an informal basis to answer questions and provide contextual information about geographic patterns exhibited in the maps. Historical air photos were consulted to provide background on the development of mobile home parks. 2.5 Recapitulation This chapter described the methods used to analyze thematic and geographic tax assessment data from rural county offices. It began by detailing a phone survey of rural county offices that determined the quality and availability of digital mobile home data in all 48 rural counties. Using the results from the phone survey along with various socioeconomic and geographic variables, 12 rural counties were selected to serve as a representative sample of rural Pennsylvania for the analysis of tax assessment data. Of

46 36 those 12 rural counties, seven would also be chosen for the geographic analysis, which was also described in this chapter. This chapter made special note of the limitations on data availability and quality encountered throughout this process. The following chapter will present the results of the thematic and geographic analysis of tax assessment data completed for this thesis.

47 37 Chapter 3 RESULTS 3.1 Introduction This chapter presents the findings that emerged from the analysis of thematic tax assessment data on mobile homes in 12 rural counties including their age, size, condition, and acreage. It also details the results of the geographic analysis, which include information on mobile home location, density, distribution, and proximity to floodplains for seven rural counties. The chapter discusses the impact of various geographical processes on the spatial patterns revealed, and proposes a taxonomy of mobile homes in rural Pennsylvania. 3.2 Analysis of Tax Assessment Data Where other forms of mobile home data give only blurry estimates, tax assessment data paint a vivid picture of mobile homes in the chosen rural counties. The number of mobile homes in each rural county is displayed in Table 3.1, showing a wide variation in numbers from 1202 in Elk County to over 5700 in Bradford and Venango Counties. When normalized for population, other patterns emerge, including the proportionally high number of mobile homes in Greene, Tioga, and Venango Counties and the relatively low number of mobile homes in Monroe County (Table 3.1). According to the tax assessment data, mobile homes in the 12 rural Pennsylvania counties are considerably older than state or national averages (Table 3.2). Nearly 36 percent of all mobile homes in the sample are over thirty-three years old, as

48 38 Table 3.1: Mobile Homes per Capita, Based on County Tax Assessment Data County Total Number of Mobile Homes per 10,000 Mobile Homes Estimated 2005 population 1 Adams Bradford Elk Greene Indiana Lycoming McKean Monroe Perry Tioga Venango Wayne U.S. Census 2007a

49 39 Table 3.2: Year Built for Mobile Homes in Available Rural Counties, Pennsylvania, and the United States Year Adams Bradford Greene Lycoming McKean Monroe Perry Tioga Venango Wayne Total Valid Percent Before or Percent PA n/a n/a later Total , ,563 Median Percent U.S Foremost Insurance Median of the Medians

50 40 compared with only 17 percent and 12 percent at the state and national levels, respectively. Over half (52 percent) of the mobile homes in the 12 rural counties are not sited on land owned by the mobile home owner (Table 3.3). Of the mobile homes located on property owned by the mobile home resident, 72 percent are on five or fewer acres; the median acreage of all 12 counties is less than two acres. Class and condition data on mobile homes is sparse in the rural counties, with only two counties (Lycoming and McKean) storing data on class, and three counties (Bradford, Greene and Wayne) keeping data on the condition of their mobile homes. Class is defined by the county offices as pertaining to the overall quality of a mobile home s construction, whereas condition describes the current state of the building. As seen in Table 3.4, most of the mobile homes (86 percent) in Lycoming and McKean counties fall in the standard or economy classes. Forty-seven percent of mobile homes in Bradford, Greene, and Wayne counties are considered of average condition, but perhaps more telling is the percentage of mobile homes (45 percent) deemed below average in these three counties (Table 3.5). Most mobile homes (78 percent) in the sample rural Pennsylvania counties are single-wides (Table 3.6), and in addition to being smaller, these single-wide mobile homes are much older than their double-wide counterparts (Tables 3.7 and 3.8). The median year built for single-wide mobile homes is 1976, whereas the median for doublewides is It follows, therefore, that single-wide mobile homes are also in much poorer condition than double-wide mobile homes 55 percent are in below-average condition, compared to only 12 percent of double-wides (Tables 3.9 and 3.10). Although there is not a large variation in the amount of acreage associated with single-wide and

51 41 Table 3.3: Acreage Associated with Mobile Homes for Available Rural Counties Acreage Adams Bradford Elk Greene Indiana Lycoming McKean Monroe Tioga Perry Venango Wayne Total Valid Percent No land , or more TOTAL ,

52 42 Table 3.4: Mobile Home Class for Available Rural Counties Class Lycoming McKean Total Valid Percent Economy Standard Deluxe Luxury Total Table 3.5: Mobile Home Condition for Available Rural Counties Condition Bradford Greene Wayne Total Valid Percent Very Poor Poor Fair Average Good Very Good Total , Table 3.6: Mobile Home Building Type for Available Rural Counties County Percent Single-Wide Percent Double-Wide Bradford Greene Lycoming McKean Monroe Perry Venango Wayne Total 78 22

53 43 Table 3.7: Year Built for Single-Wide Mobile Homes in Available Rural Counties Year Bradford Greene Lycoming McKean Monroe Perry Venango Wayne Total Valid Percent Before or later Total , Median Median of the medians

54 44 Table 3.8: Year Built for Double-Wide Mobile Homes in Available Rural Counties Year Bradford Greene Lycoming McKean Monroe Perry Venango Wayne Total Valid Percent Before or later Total Median Median of the medians

55 45 Table 3.9: Condition of Single-Wide Mobile Homes in Available Rural Counties Condition Bradford Greene Wayne Total Valid Percent Very Poor Poor Fair Average Good Very Good Total Table 3.10: Condition of Double-Wide Mobile Homes in Available Rural Counties Condition Bradford Greene Wayne Total Valid Percent Very Poor Poor Fair Average Good Very Good Total

56 46 double-wide mobile homes sited on resident-owned land (Table 3.11), single-wide mobile homes are more likely to have no resident-owned land associated with them than double-wide mobile homes: 54 percent of single-wides are not sited on resident-owned land, contrasted with only 24 percent of double-wide mobile homes (Table 3.12). 3.3 Geographic Data Analysis The spatial organization of mobile homes varies among the study counties. For example, in Lycoming County (Figure 3.1) most mobile homes are clustered in the southern municipalities, whereas in Adams County (Figure 3.2), they are distributed evenly across the county. Adams County, however, has many large mobile home parks in comparison to Indiana County (Figure 3.3), where there are not only fewer parcels with multiple mobile homes, but also fewer mobile homes on those parcels. McKean County (Figure 3.4) and Monroe County (Figure 3.5) also show clustering patterns, with much larger mobile home parks located in Monroe County. Venango County (Figure 3.6) has a fairly even distribution of mobile homes across the county, with many mobile home parks of all shapes and sizes. In Wayne County (Figure 3.7) the mobile homes are also distributed evenly throughout the county, with only a few moderately-sized and one very large mobile home park. Statistical analyses of the map data highlight these characteristics. Table 3.13 demonstrates that the percentage of parcels with more than one mobile home varies among the study counties, ranging from less than seven percent in Wayne County to more than 39 percent in Venango County. Lycoming County has the greatest number of parcels (41) with more than 10 mobile homes, suggesting more mobile home parks there.

57 47 Table 3.11: Measures of Central Tendency for Mobile Homes with Acreage, by Type County Mean Acreage Median Acreage Standard Deviation All MHs SW MHs DW MHs All MHs SW MHs DW MHs All MHs SW MHs DW MHs Adams Bradford Elk Greene Indiana Lycoming McKean Monroe Perry Tioga Venango Wayne

58 48 Table 3.12: Percent Mobile Homes with No Land by Type County Percent Single-Wides With No Land Percent Double-Wides with No Land Bradford Greene Lycoming McKean Monroe Perry Venango Wayne Total

59 Figure 3.1: Lycoming County Mobile Home Density Based on Tax Parcel Data 49

60 Figure 3.2: Adams County Mobile Home Density Based on Tax Parcel Data 50

61 Figure 3.3: Indiana County Mobile Home Density Based on Tax Parcel Data 51

62 Figure 3.4: McKean County Mobile Home Density Based on Tax Parcel Data 52

63 Figure 3.5: Monroe County Mobile Home Density Based on Tax Parcel Data 53

64 Figure 3.6: Venango County Mobile Home Density Based on Tax Parcel Data 54

65 Figure 3.7: Wayne County Mobile Home Density Based on Tax Parcel Data 55

66 56 Table 3.13: Parcels with Multiple Mobile Homes (MHs) County Percentage of Parcels with Multiple MHs Average Number of MHs on Parcels with Multiple MHs Number of Parcels with more than 10 MHs Adams Indiana Lycoming McKean Monroe Venango Wayne

67 57 One very clear geographical pattern showcased through this work is the percentage of mobile homes found in rural versus urban municipalities (Table 3.14). In the seven study counties, most mobile homes are located in rural municipalities, with an overall average of only 11 percent of mobile homes located in urban municipalities. According to the geographic analysis of tax assessment data, many mobile homes in rural Pennsylvania are located on parcels that intersect the 100-year floodplain (Table 3.15). The overall percentage of mobile homes intersecting the floodplain in each county ranges from around 12 percent in Indiana County, to almost 36 percent in Adams County. The number of parcels with multiple mobile homes intersecting the floodplain ranges from 17 in Monroe County to 120 in McKean County. Note that by including the parcels that intersect the floodplain, as opposed to only those that are completely within the floodplain, these numbers include some parcels where the mobile home itself may not be in the floodplain (Figure 3.8). The 100-year floodplain is by nature an estimate of probability, however, and does not represent a firm barrier to flood waters; it is simply a general guideline used by planners and insurers. Given the acknowledged structural vulnerability of this housing type to all types of hazards, intersection was determined to be an appropriate criterion for this analysis. Spot checks of randomly selected aerial photographs for each sample county showed that, as suggested by the tax assessment data, many mobile homes are indeed found in floodplains. Although by no means comprehensive, these figures provide one measure of the degree of physical exposure to flooding experienced by mobile homes in each of the study counties.

68 58 Table 3.14: Percentage of Mobile Homes in Rural Municipalities County Percentage of MHs in Rural Municipalities Adams 95.3 Indiana 86.0 Lycoming 92.2 McKean 83.7 Monroe 71.9 Venango 98.7 Wayne 93.6 Total 88.8

69 59 County Table 3.15: MH Parcels Intersecting the 100-Year Floodplain Percentage of MHs on Parcels Intersecting the 100-Year Floodplain Number of Parcels with Multiple MHs Intersecting the 100-Year Floodplain Adams Indiana Lycoming McKean Monroe Venango Wayne Figure 3.8: Adams County Mobile Home Parcels Intersecting the 100-yr Flood Plain

70 60 It is clear from the differences in mobile home density and distribution across counties that many geographical processes influence these patterns. Ground-truthing visits and extended GIS analysis for Lycoming, Indiana, and Adams counties revealed that some of these processes are specific to mobile homes as a housing type, while others are merely reflections of general housing patterns in each county. In Lycoming County, nine of the 12 largest mobile home parks in the county (containing over 750 mobile homes total) are clustered together in only three townships (Muncy Creek, Fairfield and Wolf), due east of Williamsport (Figure 3.9). Historic 1938 aerial photographs of this region show that mobile home park parcels are similar in shape to the parcels of farmland that occupied this area in the 1930s, suggesting that park developers purchased park land directly from farmers. It is also possible that lenient zoning regulations in these municipalities may have led to an increased number of large parks in this region. Inexpensive land found near major roadways likely contributes to the spatial pattern seen in Lycoming County (Figure 3.11), where mobile homes can be found following the north-south highways and along all major roadways. In fact, according to GIS analysis, nine of the 12 largest mobile home parks and over 60 percent of all mobile homes in Lycoming County are located within 300 feet of a major roadway 3.1. County-scale mobile home distributions are also influenced by the economic activities taking place in each county. In Indiana County (Figure 3.3), for instance, the even distribution of mobile homes throughout the county mirrors the even distribution of dairy farms, Christmas tree farms, and bituminous coal mines crisscrossing the county. The uniform distribution of individual mobile homes and small parks is similar to agricultural Adams County (Figure 3.2). But in contrast, large mobile home parks in 3.1 Major roadways provided by Lycoming County and determined by traffic flow (Bower, 2008).

71 Figure 3.9: Large Mobile Home Parks Clustered in Lycoming County 61

72 Figure 3.10: 1938 Air Photo of Lycoming County Mobile Home Parcels 62

73 Figure 3.11: Lycoming County Mobile Homes Along Major Roadways 63

74 64 Adams County are clustered in municipalities bordering neighboring urban counties. Adams County has recently experienced a high rate of population growth rates as retirees, commuters, and telecommuters from nearby urban areas have migrated into these areas (Mookerjee et al 2006). The resulting shortage of affordable housing may have contributed to increased demand for space in large mobile home parks. Lycoming County (Figure 3.1) exhibits very different patterns of mobile home distribution, with a low density of homes in the northern part of the county that can be explained by the many state forests in the region. Closer to the economic center of Williamsport, however, many large mobile home parks are found near the manufacturing, medical, and postsecondary educational institutions that employ most county residents (PA Dept of Community and Economic Development 2007). Conveniently for mobile home residents in the county, nine large parks and over 35 percent of all mobile homes in Lycoming County are within a half mile of a city bus route. It is likely that the county planned bus routes to service these communities, but it is also possible that some mobile home residents choose to site their home close to bus service creating an even more prominent clustering pattern in the county. 3.4 A Taxonomy of Mobile Homes in Rural Pennsylvania In addition to economics and transportation, the differences in the distribution of mobile homes in each county may also be related to differences in mobile home function. Far from a one-size-fits-all form of housing, mobile homes are used in a variety of different ways. Throughout the geographic analysis, I observed certain similarities in the

75 65 form and function of this housing type and compiled the following taxonomy of mobile homes for rural Pennsylvania. Large Mobile Home Parks: Rarely located in urban municipalities (due to zoning restrictions), but frequently found in municipalities bordering urban areas, these mobile home parks serve as workforce housing for local economies that support large concentrations of employees. In Lycoming County these economies are primarily the manufacturing and service sectors located in the Williamsport area. Figure 3.12 shows a mobile home park located directly behind Muncy Homes a moderately-sized factory that produces (ironically) modular homes. The large mobile home parks bordering urban municipalities in Adams County also fall into this category. There are few large mobile home parks in Indiana County probably because the main economic activities of the county are geographically distributed resource-based industries that do not require large enough concentrations of workers in any one area to support large mobile home parks. Most large mobile home parks are clustered together with multiple moderate and even small parks nearby they might be owned by the same company or person. Nearly all large mobile home parks are within 300 feet of a major roadway. They primarily contain singlewide mobile homes on small lots in very dense proximity to one another (Figure 3.13) a strategy that maximizes revenue for the park owner. Medium and Small Mobile Home Parks: In rural communities these parks provide workforce housing in proportion to the size of the industry or community.

76 Figure 3.12: Large Mobile Home Park Located Behind Factory in Lycoming County 66

77 Figure 3.13: Timberend Estates in Lycoming County 67

78 68 In many rural towns this is housing for agricultural workers (Figure 3.14), although some of these parks are specialized communities housing retirees, for instance (Figure 3.15). These parks are not always located near a large roadway (unless they are clustered with several other parks), although they tend to be centrally located to the communities they serve. Figure 3.16 shows a mediumsized mobile home park in northern Lycoming County. Big Land Rural Housing: Often double-wides, these mobile homes are located on their own large lots on rural roads interspersed with traditional stick-built houses. Most of the newest and best-maintained mobile homes fit into this category. Older homes may have multiple custom additions constructed to expand living space as the family grows over time it may be difficult to distinguish the structure as ever having been a mobile home. This is essentially the mobile home serving as a rural ranch style home (Figure 3.17). This category comprises most of the mobile homes in Indiana County, where there are relatively few mobile home parks. Figure 3.18 shows several large-lot mobile homes across the street from a Christmas tree farm. Farmsteads: Housing extended family or farm workers on agricultural land, this category includes mobile homes sited on large farmsteads (Figure 3.19). These are typically older single-wide mobile homes. GIS analysis showed that the majority (over 30 percent) of mobile home parcels in Lycoming County are

79 Figure 3.14: Small Mobile Home Park in Adams County 69

80 Figure 3.15: Meadowbrook Mobile Home Park in Lycoming County 70

81 Figure 3.16: Medium-Sized Mobile Home Park in Northern Lycoming County 71

82 Figure 3.17: Mobile Home as Big Land Rural Housing in Lycoming County 72

83 Figure 3.18: Mobile Homes as Big Land Rural Housing in Indiana County 73

84 Figure 3.19: Mobile Home on a Farmstead in Lycoming County 74

85 75 located in areas zoned for agriculture, so many mobile homes in the study counties may fall into this category. It is important to note that these are technically mobile homes on leased land. This taxonomy thus adds an important qualitative dimension to the statistical information presented thus far for mobile homes on leased land as extended family members, for instance, the residents of these mobile homes may experience far less housing insecurity than a mobile home resident leasing space from a park owner. Hunting & Fishing Cabins: With its proliferation of state parks and nature areas, rural Pennsylvania serves as a weekend or vacation getaway for many Pennsylvanians. This category includes mobile homes near state parks and gamelands. The homes are often found alongside streams and in wooded areas on very rural roads (Figure 3.20). It can be difficult to separate these homes from those serving as Big Land Rural Housing, except that they tend to be older structures on smaller lots. It is clear from the analysis presented here that the spatial patterns of mobile home density and distribution in rural Pennsylvania counties are affected by a variety of geographical processes including transportation networks, historic and current economic activities, and population trends. The flexibility of the mobile home as a housing type contributes to its multi-functional use from hunting cabin to farmstead housing. These uses in turn affect rural Pennsylvania as a whole cementing the mobile home as a significant feature of the rural landscape.

86 Figure 3.20: Fishing Cabins in Lycoming County 76