Build more stadiums! Build them! Or…think again

In 1999 the building costs of the Millenium Stadium in Cardiff were around £130 million. And now a new report found that the economic punch for Wales is £130 million a year and sustains more than 2,500 permanent jobs, a new report has revealed. That would suggest that building more stadiums is a fantastic return for a society.

Or?

Or should we be suspicious on such fantastic figures? Or should we do better economics?

Actually, both: we should be suspicious for such fantastic figures and do better economics.

Why should we be suspicious?
Correcting for inflation the costs were about £175 million and the annual return on investment is 130/175, or 75% for society. If a bank would suggest such a return, we would all be very suspicious. However when a government or a stadium comes up with such figures we believe and copy them happily. Returns which sound too good to be true, are often indeed just that.

How should the economic analysis be improved? Economic Impact Analysis underlying such big number has a couple of flaws. (These flaws are not present in social cost benefit analysis).
One, costs are often seen as a benefit. Every spending in an EIA has a positive effect, even if the spending is mostly a costs. For example, often construction costs end up positive in an EIA, ignoring the alternative uses of the money.
Second, markets are often ignored. For example in the Millennium Stadium case, do we really believe that if there was no stadium, that those 2500 workers would be unemployed, or do we think that (labor) markets work and create alternative jobs? The latter is more likely. Without the stadium those 2500 would be working in other jobs, probably less well paid. So if you want to know the contribution of the stadium you should also include the production in those other jobs that is no longer produced because people are working for the stadium (directly or indirectly). And a stadium than only creates a welfare gain of a few percentage of the initial estimate. And the employment benefit is zero, that is in the long run (like in the study cited), no more people will be working (so no difference in unemployment in Wales), but they will be working in different jobs.

So I think the numbers in the article and the report are way too positive.
(And yes, even with much lower numbers the investment in the stadium might have been wise, but the numbers presented here do not show that).