Question for help
ItSaulGoodman
Chapter11powrr.pptxChapter 11
Implicit Market Measures of the Benefits of Crime Control
Introduction
Economic theory suggests that the value of an additional item can be estimated by observing the market price for a good (assuming there are no externalities).
If no externalities exist for a good, then the private and social benefits are equal to the price.
2
Introduction
Crime does not fit this valuation model since there is no direct market for crime (usually) and, even if there were, transaction prices wouldn’t accurately reflect externalities.
This chapter explores other valuation models, such as the implicit market model and the hedonic regression estimation, to estimate the social cost of crime
The Nature of Implicit Markets
Implicit markets arise when good with several characteristics are only sold as bundles.
Implicit market techniques require that individuals reveal their preferences through market purchases
The Nature of Implicit Markets
Consider a restaurant that offers 10 different sets of 4-course meals at 10 different prices
The only information from choices made would be about the relative utility of the meals, but we could say nothing about the utility from individual items.
However, if there are a sufficient number of restaurants, the set menu is not an impediment to uncovering the marginal valuation each course.
Hedonic Model
The “hedonic regression” estimates the prices of characteristics of goods sold in bundles.
This model was first suggested by Andrew Court (1939) to estimate the value of automobile characteristics
= weight, = wheel base, = Horsepower
Sidenote: Taking the natural log of the outcome variable changes the interpretation of the parameter. For example, if = 0.04, I would mean that a 1 pound increase in weight leads to a 4% increase in price.
Hedonic Model and Crime
Thaler (1976) later used the hedonic model to estimate the benefits of crime control in Rochester, NY
is the sale price, is a vector of housing characteristics, is a vector of neighborhood characteristics, and is the neighborhood property crime rate
Hedonic Model and Crime
is the coefficient of interest and estimates how much a one unit increase in crime lowers the value of a home or how much the average home buyer is willing to pay for a one unit reduction in crime rates
Hedonic Model and Crime
Thaler’s (1976) coefficient estimated that a standard deviation increase in the crime rate is associated with a decline of $430 in home value or $2,429 in 2012 dollars
Assuming there are 100,000 houses in Rochester and a 5% annual rate of return, the external cost of a standard deviation increase in property crime in Rochester is $12,145,000. This represents the money the citizens of Rochester are willing to give up in order to avoid this crime.
Hedonic Model and Crime
Gibbons (2004) investigates the effect of burglary rates on home prices. He points out a potential problem with the OLS hedonic model.
The problem: More valuable homes may be bigger targets for burglars. This would cause reverse causality bias in the estimated parameter since more valuable homes cause more burglary activity.
Hedonic Model and Crime
The solution: Use instrumental variables!
Gibbons uses non-residential burglary rates as an instrument for residential burglary rates. He argues that this is a valid instrument since non-residential burglary is not caused by residential home price, and non-residential burglary is correlated with residential burglary.
He finds that using OLS there is a strong positive relationship between burglaries and home prices, but that this relationship becomes statistically insignificant using IV.
He also included criminal damages (such as vandalism) in his model. His results showed that criminal damages had a negative effect on home prices.
Megan’s Law
This federal law, which passed in 1996, requires state and local governments to record and disclose the residential location of convicted sex offenders.
This information is available online and is easily accessible to homebuyers.
Megan’s Law
Hedonic model could be used to estimate the effect of living within 0.1 miles of a registered sex offender on home values.
What is a potential identification problem?
Sex offenders may tend to live in neighborhoods with characteristics that are correlated with house prices.
Solution? Use a difference in difference model!
Megan’s Law
Linden and Rockoff (2008) used DID to find that living within 0.1 miles of a sex offender lowered property values by 4%. This is a $220,000 reduction in home values from a single sex offender
Megan’s law
Pope (2008) used similar model, but allowed for the exit of sex offenders.
He found that these effects were symmetrical in that house prices within 0.1 miles of the sex offender were 2.3% lower, but that this effect disappeared within a year after the sex offender left.
Practice Problem
Assume that Tweedle Dum and Tweedle Dee are identical apartment complexes with 1,000 units each. The physical units are the same and accessibility to jobs and shopping is the same. Actually, there are two differences between Dum and Dee. The violent crime rate is 0 in Dee and 4 per year in Dum, and the rent is $120 per year lower in Dum.
As an Economist, what conclusion regarding the cost of violent crime might you draw from this example?
Assume that Dum is considering putting in a security system that cost $80,000 per year and would reduce violent crime from 4 to 0 per year. Would you advise Dum to do and why?
