Latin Hypercube Sampling and the Identification of the Foreclosure Contagion Threshold

Abstract

Over the last several years, the U.S. economy has experienced a significant recession brought on by the collapse of the residential real estate market. During this downturn, the number of real estate foreclosures has risen drastically. Recent studies have empirically demonstrated a reduction in real estate values due to neighboring foreclosures, termed the foreclosure contagion effect. The foreclosure contagion effect impacts healthy neighboring properties that surround the foreclosed property as a function of both time and distance. We mathematically specify a precise equation that identifies the foreclosure contagion threshold, that is, the boundary that separates surviving markets from those that crash. Using a new technique to our field known as Latin Hypercube Sampling, we presents the results of a large scale sensitivity analysis to find that beyond the foreclosure discount and disposition time variables, the percentage of adjustable rate mortgages (ARMs) and the foreclosure distance discount weight are the two secondary most contributing causes to a market collapse.

Keywords:

• Emergent behavior
• Foreclosure contagion threshold
• Agent-based models
• Latin hypercube sampling

Notes

1. By similar results, we mean that both methods produce a similar distribution of the output variables.

2. Using a torous grid environment, this means that up to 351 homes are considered each time an appraisal is performed.