Evan Soltas
Apr 7, 2015

The Pain in Spain, Part 2

All of Spain's housing bubble, I showed last week, can be explained by the boom-and-bust in mortgage debt. Had there been no bubble, real home prices would have risen 30 percentage points less than they did at their peak and would not have declined at all since the peak.

Today, we look at another major piece of the fallout of Spain's housing bubble -- unemployment -- and the distortionary effects that mortgage debt had on the Spanish labor market.

In February, 23.2 percent of the Spanish workforce was unemployed, down from a high of 26 percent in 2013, and the highest of any Eurozone economy except Greece.

Those numbers are astonishingly high, at least from an American perspective. It is important to remember, however, that a "normal" rate of unemployment for Spain is high. The average unemployment rate from 1985 to 2007 was about 15 percent. That doesn't make 23-percent unemployment any less painful, but it does give a more realistic view of the kind of output gap it implies.

The data I use, to review what was in the first post, come from several Spanish government agencies and are measured at the province level. Importantly, the data for employment only go back to 2002 and on sector-level employment to 2008, so intuitively, most of the analysis will be driven by the bust rather than the boom. (If earlier data exist, please email me.)

Here's what I find.

First, the effects of mortgage debt on the labor market are large. For every 10-percent increase in monthly mortgage debt issuance, the employment rate rises 0.2 percentage points and the unemployment rate drops 0.2 percentage points. (For reference, mortgage debt issuance rose about tenfold during the boom.)

These labor-market effects are distortionary. A 10-percent increase in mortgage debt issuance leads to a 5-percent increase in construction employment, a small increase in services employment, and small decreases in agricultural and industrial employment, as workers reallocate between industries in response to relative demand for labor.

With the labor market -- especially construction -- so dependent upon mortgage debt, it's easy to see that the collapse of the Spanish housing bubble plays a key role in explaining high Spanish unemployment.

Had had the Spanish housing bubble never popped, I estimate the unemployment rate would currently be 14 percent, about 10 percent lower than it is today. Put differently, Spain's housing bubble fully explains why its downturn was so much worse than those of, say, France or the United Kingdom.

How can we say that? We've already seen mortgage debt affected the labor market through construction employment. We can use that knowledge -- it helps to know what instrumental variables (IV) are here -- to estimate a counterfactual scenario in which Spain's housing bubble didn't pop. The first counterfactual, labeled "OLS", is more crude -- it tries to look at the direct link between mortgage debt and employment.

Now, it's naive to think that Spain's problems would be solved had the bubble not popped. It would have had a gigantic, distortionary housing bubble! The better idea is not to have one at all.

What would have unemployment looked like in that case? Due to the limitations of my data, I'm not able to run the IV estimate in green any further back. From the OLS estimate, we can see that the portion of housing boom between 2003 and 2008 probably pulled down the unemployment rate. Notably, Spain's unemployment rate declined substantially between 1995 to 2008, which is the time period of the housing bubble.

So we can't really say where Spain would be in terms of unemployment had the bubble never happened. But the bust had a devastating effect.