A thread on @mmoralmo's amazing job market paper, one of the best crime/social policy papers I've seen.

Suppose you're mayor. You learn that a program--maybe it's drug treatment, lighting, social services, or policing--drops crime by a big amount. How much should you invest?
Most research can't tell us. Crime studies mostly look at reported offenses or arrests. Those are incomplete, underreporting crime. And a Mayor has no idea how to translate that into a budget investment.

Miguel's answer: Use housing prices to see what people are willing to pay.
WTP isn't going to price in everything, but it's going to incorporate a lot:

- Mental health benefits

- Expected growth in business & investment

- Some (not all) of the adverse effects from mistrust in police or excessive force

A huge step ahead of crime & arrest counts
One trouble is you need incredibly fine grained crime and price data. You also need a large-scale intervention with a plausible control group.

Miguel gathers block-level data from 3 cities in Colombia. And he finds a huge intervention: the building of 100 mini-police stations.
Multi-city evaluations are extremely rare, but important. In this case, the 3 are incredibly different.

Wealthy, no organized crime Bogota

Medellin, with its hierarchical organized crime

Cali, with its more freewheeling youth gangs and everyday disturbances
Miguel has several ways to estimate the localized crime impacts of the stations: a broad differences-in-differences; inner versus outer rings; etc.

Let's focus on his preferred one: grid cells that receive a station late compared to those that got one early.
The impacts are huge: a new mini station deters violent crime by 12% and property crime by 22% in the immediate vicinity, with no crime displacement nearby
Before we even get to the main aim of the paper (WTP), this is already a big finding: a multi city study with an estimate of the effects of a large and sustained, highly localized, crime reduction intervention. With big effects. Causal.

Most students would have stopped there.
Miguel estimates a hedonic pricing model, instrumented with the crime impact of the stations.

The crime reduction leads to a 5% increase in property values — a gain of $3.5 million for households directly affected.

That's massive.
Most people still would have stopped there.

Miguel takes it one step further. The natural experiment gets you a local treatment effect on the people who were willing to live in a (previously) unsafe neighborhood.

That's not the policy parameter the Mayor needs.
So Miguel estimates a model that accounts for residential sorting, to try to get the willingness to pay for crime reductions for the average person in the cities.

It's a sophisticated and structural way of re-weighting treatment effects to be more representative of the pop'n
Miguel is on the economics and public policy market!

Here is his website: https://sites.google.com/view/morales-mosquera/research?authuser=0

DM or email me if you want my (effusive and detailed) recommendation letter
You can follow @cblatts.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled:

By continuing to use the site, you are consenting to the use of cookies as explained in our Cookie Policy to improve your experience.