Empirics

1. Trailer Park EconomicsCaitlin Gorback Duke University, 2011 Durham, NC Durham, NC Disequilibrium Model: Risk-Sharing and Uncertain Growth Equilibrium Model: Bad Tenants Pure Rental System from Owner’s Perspective: Owns large tract of land upon which rental structures are placed. Goal: maximize profits. He cannot observe whether a renter is “good” or “bad”. Revenues depend on a flat rental rate across types of tenants. Costs are: renovation, park upkeep, eviction and are higher with bad renters. The owner recognizes he cannot maximize profit if he has “bad renters”, so he tries to “homogenize” renters, by requiring them to take on risk. Pure Ownership from Tenants Perspective: Tenants own both land and capital. A bad tenant creates negative externalities for the rest. The tenant maximizes utility subject to a budget constraint, which includes externalities. Neighbors can choose to hire a 3rd party regulator or not. Uirepresents utility, eiis effort it takes to be good, nei is negative externalities, Ceis cost of eviction. The Game without 3rd party actor, Nash Equilibrium in red: The Game with 3rd party actor as “Benevolent Dictator”. Introduction of eviction costs borne by bad tenants, Ce > ei Mixed Rental and Ownership System: Land owners rent only land, and tenants prefer to own only structure, but how does the system come into existence? By a contracting game and backwards induction. When both abide by contract: Owners gains U(rll), and tenant gains Ui-e.If tenant breaches contract, owner finds out with probability, P: owner gains P[U(rll-rlli)], tenant gains P(Ui-Ce). There is clear incentive for all to abide by the contract. Summary: It is in the owner’s best interest to require tenants to purchase their own housing unit, so he can avoid repairs to rented housing. It is in the tenant’s best interest to hire a 3rd party regulator to get rid of unsavory neighbors, even though this may mean sacrifice of some housing rights. Predictions: This model predicts that trailer parks will be very homogeneous within parks, but across parks will differ based on the contracting systems between the tenants and owners. Empirics Introduction General Model: We collected census block group level data (about 1,500 people per clock group) for 47 counties in North Carolina. As a way to empirically test each model, we developed a general empirical model: Ti = β0 + Di’β1 + X#i’β2 + PDiβ3 + εi , Where Xi is a vector of Case Variables, with # relating to (1,2,3,4) for each model, Di is a vector of Demographic Variables, PDi is the population density of the block group, and Ti is a variable indicating whether or not the block group is trailer park dense. The demographic variables include income, education, white/blue collar, race, and unemployment, and the case variables are those relating to each theoretical discussion. Dataset Adjustments: We used Principal Components Analysis to reduce the number of variables in Di and X#i, allowing us to move from variables such as race, unemployment, and income to components as “Non-Asian Minorities” or “Social heterogeneity”, which combined effects of multiple variables into components. Additionally, to correct for bias in data collection (data skewed away from trailer parks), we had to establish a control group. Using Propensity Score Matching, we were able to trim the dataset by first dividing the dataset into trailer park dense and trailer park scarce groups, then matching individual observations that had similar demographic characteristics across the groups. In this way, we can ensure that our sample has many similar block groups that only differ on whether or not they are trailer park dense. Risk Sharing and Uncertain Growth model: We used data from 2009 and 2010 on rural employment levels, rural business counts, urban employment levels, and urban business counts. We posit that the growth of manufactured housing should be highly correlated with growth in both industrial jobs and businesses, a key application of uncertain growth. Because the model measures the relation between trailer park growth and industrial growth, we take first differences on both the measures of trailer park density and on the aforementioned four industrial variables and alter the general regression: ΔTi(t-j) = β0 + Di’β1 + ΔX3i(t-j)’β2 + PDiβ3 + ε. The table below shows results, with * denoting 90% significance, &** 95%. The coefficient on urban industrial growth shows that with each unit increase in the industrial growth index, we expect trailer parks’ percentage of housing to grow by 0.11, for each unit increase in rural industrial growth index, trailer parks’ percentage of housing will increase by about 0.13. The higher and more significant coefficient on rural industrial growth is reassuring, as space limitations would favor trailer park growth in rural environment as opposed to urban environments. Since the first mobile camping unit was strapped to the back of a Ford Model T in the 1920’s, the concept of manufactured homes has been an important factor of the American housing market. In North Carolina alone, a little more than 16% of the housing stock is comprised of manufactured housing. With approximately 5,200 census block groups of around 2,000 people, 1,300 of which have at least 30% of their housing stock in manufactured homes, we can estimate that about 780,000 North Carolinians live in manufactured homes, and likely half of those live in trailer parks. So what does the trailer park literature have to say about this institution? Next to nothing, it turns out. Current trailer park literature is limited to observations on rent control in trailer parks or measuring how happy people are in trailer parks relative to other forms of low income housing. Nothing discusses why people choose to live in such an odd mixed rental-ownership system as a trailer park, instead of fully owning or renting their housing. We propose four models through which the development of trailer parks could be explained. The first two are equilibrium models, in which tenants and landowners contract while in a housing market equilibrium. The second two show that trailer parks arise in response to growth shocks in the economy in which housing becomes scarce or uncertain. Equilibrium models include 1) Bad Tenants, 2) Capital Constraints on both Investors and Occupants, while disequilibrium models include 3) Risk-Sharing and Uncertain Growth, and 4) Short vs. Long Term Urban Growth.* *Due to space limitations, only models 1) Bad Tenants and 3) Risk-Sharing and Uncertain Growth will be displayed. The landowner is risk averse and gets less utility from profits than he does disutility from losses. Because the owner has an uncertain future, he wants to minimize his future costs. He cannot minimize costs by scaling back the entire venture; uncertainty can still result in high losses if tenants follow work to another city. It is the owner’s best choice to share some of the burden of risk with the tenants, whom he views as flight risks This model strongly relates to factory towns. In towns with newly built factories, there is a high demand for new housing immediately. Manufactured housing is a low-cost and expedient way to provide housing to blue-collar factory workers. The developer fears the factory will close, leaving the new housing developments empty. It is most profitable for the landowner to rent parcels of land, but not invest in the housing units, instead having renters provide their own units. Owners: maximize utility U(π)=√π where preferences are Cobb-Douglas, satisfying the necessary risk aversion specifications. We will define π=(rk-ck)k+(rl-cl)l, with r’s as rents, and c’s as costs, with respect to k, capital, and l, land. Over two time periods, the owner will maximize his expected utility. We assume the owner has information about the growth and urban climate in the first time period, but is uncertain about the second period. We represent the probability of growth, as P, with 0<P<1. He must maintain the housing units and land over both periods, so cost of maintenance is unaffected by probability of growth. Utility from future profits is diminished by a factor of ß, showing that next period’s profits are not as valuable to the owner presently. The current period is represented by the subscript “0”, and future period by the subscript “1”. Thus, Summary: The owner maximizes profit with respect to k and l. As probability of growth goes to zero, the landowner must absorb all costs associated with maintaining both land and capital structures, without any profit. The owner realizes with uncertain growth, he can maximize profit only by cutting out k (providing housing no units for rent), thus minimizing losses in the second period. He cannot maximize by cutting out land because in order to rent out structures, he must place them on land. We end up with trailer parks, as the only mechanism by which he can reduce costs requiring tenants to provide their own housing unit. Tenant 1 Tenant 1 Bad Bad Good Good 1: U1 – e1 2: U2 – e2 1: U1 – e1 2: U2 – e2 1: U1 – e1 – ne1 2: U2 1: U1 – e1 – ne1 2: U2 1: U1 2: U2 – e2 – ne2 1: U1 2: U2 – e2 – ne2 1: U1 – ne1 2: U2 – ne2 1: U1 – ne1 2: U2 – ne2 Good Good Bad Bad Tenant 2 Tenant 2 Important Events in Trailer Park History The Ford Model T is introduced, Americans attach travel trailers The Great Depression hits. Travel Trailers become mobile homes. Conclusion WWII: Military jobs leads to housing crunch in defense cities, and trailer park solution. After noting the holes in housing literature with respect to American trailer parks, we developed four theoretical models to explain the development of trailer parks, both in a housing market in equilibrium, and in a housing market experiencing growth shocks. The equilibrium model, “Bad Tenants” provides a game scenario in housing market equilibrium, in which tenants choose to hire a third party regulator to ensure good neighbors. The model shows that across the trailer park market, trailer parks are very heterogeneous, but within parks, neighbors are homogenous. Our disequilibrium model, “Risk Sharing and Uncertain Growth” moves towards a model responding to housing shocks. We have shown, both theoretically and empirically, that growth in manufactured homes will arise when the landowner responds to the high need for housing in the current period, but is uncertain about profits in the next period. Empirically, we did find support in North Carolina that in times of industrial contraction, trailer park growth is negative. This supports our hypothesis as well as demonstrates the magnitude of the shifts in trailer park populations during an industrial contraction. Much groundwork still remains to build a solid foundation for trailer park economics. We believe that the above models and empirical results provide interesting and dynamic groundwork for this development. 1950’s-1060’s housing crunch from lack of building during Depression and WWII. Veterans and families move to trailer parks. 1970: President Richard Nixon supports trailer parks. Acknowledgements: My advisor, Dr. Charles Becker, is owed a huge amount of thanks and gratitude. Without his guidance and patience, I could not have accomplished this project. 1971: Federal Housing Administration expands mortgage insurance opportunities to manufactured homes 1974: Mobile Home Manufacturers Association changes name to Manufactured Housing Institute. Expansion of trailer parks due to 1970’s positive policies. References 1) Dixit, Avinash K., & Susan Skeath. (1999) "Collective-Action Games." Games of Strategy. New York: W. W. Norton & Company. 2) Hurley, A. (2001). Diners, Bowling Alleys and Trailer Parks. New York, NY: Basic Books. 3) Rosebaum, P.R., & Rubin, D.B. (1983). The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika, 70(1), 41-55. Retrieved from http://www.jstor.org/stable/2335942/ 4) Stiglitz, J.E. (1974). Incentives and Risk Sharing in Sharecropping. The Review of Economic Studies, 41(2), 219-255. Retrieved from http://www.jstor.org/stable/2296714/ Mid-1990’s: Push for single-family housing, increased access to easy credit for low-income homebuyers. Trailer park contractions. The “Providence” floor plan from Commodore Homes of Pennsylvania Contact Information 2001: Manufactured Housing Improvement Act Caitlin Gorback Trinity Class of 2011| B.S. Economics P.O. Box 96850 Durham, NC 27708 csg18@duke.edu 2007:SUBPRIME MORTGAGE CRISIS! General housing contraction Will we see trailer park growth in future with more restricted credit?