Valuing Laws as Local Amenities
I recently started a paper which argues that the value of a law should be judged by the extent to which it raises housing prices and lowers wages. This may seem an odd way to assess the welfare effect of a law. After all, higher housing prices and lower wages are thought to be bad outcomes, not good ones. But the proper way to understand these changes is as signals of positive outcomes, not positive outcomes themselves. They indicate that something good has happened in the community. Housing prices go up because more people want to live there. Wages go down because more people want to work there. Phrased more formally, higher housing prices and lower wages are how markets ration an attractive local amenity. Indeed, the increase in housing prices combined with the reduction in wages provides a measure how much people are willing to give up to enjoy the amenity. Conventional economic thinking recognizes this when it comes to estimating the social value of a new park or a better school. The same logic, I argue, applies when the amenity is anything from a better tort system to smarter rules regarding capital punishment.
This is, of course, not the standard practice. Under the conventional approach, the welfare effect of a law would be measured by evaluating the law’s effect on specific, related behaviors. For example, a three-strikes law would be evaluated by its effect on homicides; a unilateral divorce law by its impact on rates of domestic violence or divorce; and a tort reform by its impact on insurance premiums and accidents. These are certainly sensible metrics for judging the laws at issue. But none is as effective at measuring the welfare effect of a law as its impact on housing prices and wages.
First, the housing and wages approach employs a more direct proxy for welfare. The conventional approach tells use how much, e.g., the felony-murder rule reduces robbery, but it does not tell us how much people value that reduction in robbery. Yet that is the very strength of my proposed approach. The increase in housing prices and the loss of wages reveals how much the marginal resident who moves to a community is willing to pay – in terms of lower non-housing consumption – to be subject to a new law in that community.
Second, the conventional approach often provides an incomplete picture of any given law. Frequently, relevant implications are too hard to measure or are unexpected, and are therefore left out of the empirical analysis. For example, a typical study might ignore the expressive benefits of an anti-discrimination law or the placebo effects of corporate governance reforms because these consequences are so hard to quantity. With respect to unexpected outcomes, until recently scholars studying abortion rights overlooked the important effect of abortion rights on crime rates. The conventional approach also tends to ignore the enforcement costs of laws, whether direct (higher property taxes) or indirect (reduction of other government services). The housing and wages approach does not suffer these omissions. It provides a measure of the net benefits of a law, accounting for intangible benefits, unintended consequences, and enforcement costs.
To be clear, I do not contend that the housing and wages approach offers a perfect measure of welfare. It has important limitations. From a normative perspective, it gives disproportionate weight to individuals with greater income. It ignores individuals – such as prisoners and military personnel – who do not participate in the housing market. And there is some leakage when evaluating, for example, laws which convey benefits or impose costs on other jurisdictions. But, for the reasons given above, it is a better second-best than the conventional approach to valuing the within-jurisdiction benefits of a law, as well as competing methods for estimating the willingness-to-pay for public goods. Moreover, so long as the limitations inherent in my approach affect all applications equally, it can still be used to conduct relative welfare analysis or rank different legal reforms.
Skeptics will surely wonder whether there is too much noise in housing and wage data to identify the (likely small) effects that any individual law has on those outcomes. But this is an empirical question and the paper offers an empirical answer. It examines the effect of six types of laws (tort reforms, abortion access laws, no-fault automobile liability, unilateral divorce laws, capital punishment, and health insurance mandates) on local housing prices and wages. Data on housing prices and characteristics are drawn from the American Housing Survey. Data on wages are from the Current Population Survey. Finally, data on laws are from recent studies by Jonathan Klick, Thomas Stratmann, Paul Rubin, Joanna Shepherd, Leora Friedberg, and RAND. My preliminary results suggest that tort reform may reduce local welfare and that executions and diabetes coverage mandates may raise local welfare. (I stress, however, that these findings have not been demonstrated robust and should be taken as a proof-of-concept for my methodology rather than as policy recommendations.)
The purpose of this post is to ask readers two questions. First, can you tell me your criticisms of my argument? My draft (which you can find at http://www.law.virginia.edu/home2002/pdf/malani/amenity.pdf, esp. at pp. 15-19) already has a number of criticisms. I’d like to know if you have others. Second, and more importantly, can you think of examples or cases where someone has moved from one jurisdiction to another based, in part, on the laws in those jurisdictions. Examples where people at least know or inquire about the laws in either jurisdiction will help. For instance, has your realtor ever told you about a local law when you looked at a home? The main criticism I get is that people are skeptical that legal rules/decisions play a role in peoples’ decisions to move. My empirical evidence, however, suggests there is a correlation between various laws and housing prices and/or wages. I want to figure out if what I’m finding is merely spurious correlation. Thanks in advance for your feedback.