This paper offers a novel explanation for urban blight and endogenous divergence in the overall quality and wealth of neighborhoods and simultaneously derives the salient features of actual urban renewal and other aid programs from optimizing government behavior based on collective public preferences. These features appear when the objective of such public aid programs is to restore the ex ante distribution of wealth or property values within a blighted neighborhood, while equilibria exhibiting deficient levels of private investment and blight itself can arise when residents accurately anticipate the potential provision of public aid to an affected neighborhood and ex ante investment in private insurance diminishes neighborhood eligibility for such aid. Examples of antipodal equilibria in which urban renewal programs entirely crowd out local private investment or in which neighborhood residents invest in efficent levels of private mitigation illustrate these results, which stand in direct contrast to both traditional explanations of urban blight and to the new social-interaction models of neighborhood divergence.

Although crime rates have long been thought to influence residential housing prices, no previous study has measured the effect of crime rates on rates of default on residential mortgages. Using a standard model of default in which crime rates can affect both the value of property and the liquidity of mortgageholders this paper empirically measures the effect of state--level crime on the frequency of residential mortgage default. Specifically, regression analysis, based on FBI data on both violent and property crime rates, is used to analyze default rates over a pooled sample of residential mortgages for all U.S. states and the District of Columbia during a 14--year period (1981-94).It is found that crime, possibly acting as a proxy for more general socioeconomic neighborhood deterioration, significantly affects the rate of mortgage default. This effect, somewhat surprisingly, is most important for conventional mortgage loans. Violent crime has a more delayed impact than does property crime in increasing defaults. Other factors which are found to strongly influence default are state--level personal income growth and state unemployment rates. These results support the assertion that the economic costs of crime are more pervasive and subtle than often discussed.

Significant variation in the terms and volume of lending across classes of borrowers distinguished only by qualities independent of credit risk is often interpreted as evidence of inefficient or inequitable discrimination in credit markets. Increasing accuracy in the measure of credit risk renders common theories of lending discrimination and credit rationing based on lender preferences or asymmetric information increasingly implausible. We consider a traditional model of lending with complete markets, in which equilibria may exhibit disparate loan terms or access to credit across such classes of borrowers, despite common knowledge about the parameters describing the market. Rather than evidence of inefficient equilibria owing to discrimination, however, such equilibria can arise solely from the influence of asset price volatility on participants strategically exercising the options embedded in standard debt contracts. Extending substantially different loan terms or even rationing credit to different classes of borrowers can be a rational response by value-maximizing lenders when borrower classes are correlated with the degree of price volatility exhibited by the otherwise similar assets being financed by members of each of these classes. We discuss our results in the context of actual credit markets.

Maintaining a trend observed since the first Presidential declaration of a major disaster in 1953, federal disaster declarations are increasing at an increasing rate and the annual average federal expenditure on domestic disaster aid has grown at an annual average rate of two hundred fifty-five percent (255%). Owing to the discretionary authority of the incumbent President in granting gubernatorial requests for such declarations, moral hazard as well as actual losses may influence executive approval of such requests and the corresponding magnitude of aid awarded. We test, within a single empirical model of recursive choice, the hypotheses that the sequential executive decisions to grant disaster declarations and the conditional amount of aid allocated are affected by political incentives. Using a unique dataset for the period 1969-2003 that combines expenditure and approval data from FEMA, state-level demographic, fiscal, presidential and gubernatorial election data, we find, after accounting for the severity of the disaster, an incumbent president is more likely to grant gubernatorial requests for disaster declarations when facing reelection, Democratic presidents are more likely to award larger volumes of federal aid to affected states than their Republican counterparts and all presidents tend to award a larger volume of federal aid per capita to those states with a larger number of electoral votes and to those states which suffer disasters affecting a relatively greater proportional measure of their population.