Back from the Urban Economics Association sessions that are organised as part of the annual NARSC meetings. As well as the session on cities and the public sector that I described last week, there were also a couple of great sessions on policy evaluation of Enterprise Zones. Covering US, French and UK experience, one common theme was that these schemes appear to need to be expensive to have much impact and, even then, the impact comes mainly in the form of displacement from nearby areas.
Matt Freedman (Cornell) looked at the effect of the New Market Tax Credit in the US. This scheme mostly provided support for commercial development and involved billions of dollars of expenditure, mostly spent on commercial development. Matt's paper gets at the effect by comparing eligible to non-eligible areas (what's known as a discontinuity design). My take on Matt's results is that the scheme (at least around the threshold) had no effect on home values, median income, unemployment rate or employment. There does seem to be some effect in reducing the poverty rate, but household turnover is up so this may well just be a composition effect. It's also not clear where this effect is coming from given the absence of effects on employment or house prices (which might conceivably drive this effect)
Matt's paper didn't explicitly consider the issue of displacement, but this is considered by Andrew Hanson and Shawn Rohlin. Their paper looks at areas just outside successful enterprise zones and compares them to areas just outside unsuccessful zones and find substantial evidence of displacement (areas just outside unsuccessful zones do relatively better because they don't experience the displacement). In preliminary work, Elias Einio and I show that the only effect of the UK Local Enterprise Growth Initiative was to move employment from an area approximately 1km outside the boundary of the scheme to 1km inside the boundary. Aside from this displacement we can (so far) detect no other net effects.
Florian Mayneris and co-authors look at the French EZ scheme which spends a big amount of money (something over 300 euro's per person if I remember correctly). They find that this scheme can have a significant effect on firms tendencies to locate inside scheme boundary, but no effect at the level of the municipality as a whole. That is, the scheme shuffles firms from the non-eligible to the eligible part of the municipality.
Of course, policy makers may want to claim that all this displacement of jobs is what they are trying to achieve -but it is certainly not consistent with the rhetoric that is usually used (which usually talks about 'creation' rather than displacement).
One final thought, the French paper is part of a series of four projects due to report early 2012 to help inform the French government's decision whether or not to renew the scheme. The problem? The scheme was already extended for five years in 2011. Another depressing example of evidence based policy making in action?