The employment costs of subsidizing Fiat (Mr Berlusconi meets Stolper and Samuelson)

Following the debacle of car sales and production in Italy, as elsewhere, see here,  and the “calls for action” by Confindustria, the entrepreneurs association (according to which about  300,000 may be at are at risk) also the Berlusconi government has finally decided to pass new support measures for the automobile industry: a subsidy of 1,500 Euros for purchases of a new cars (the measures also contemplates tax deductions for furniture and white goods).

I have argued elsewhere that the dead-weight loss associated to this  subsidy is likely to be in the order of 100million Euros, about 1/5th of the fiscal cost of the plan. One would at least hope that the plan would alleviate the massive lay-offs coming down the road. Quite the opposite: the subsidy is likely to reduce employment even further. The reason is quite simple. Remember that consumption subsidies, to a first order, do not raise the level of households’ incomes and expenditures. They switch spending from other goods towards cars by reducing their relative price. Hence they primarily affect the composition, rather than the level, of spending. The problem is that the car industry is relatively capital (not labour)-intensive, compared to other industries in the manufacturing sector (and even more so, when compared to agriculture and services).

According to ISTAT (the Italian Statistical Agency: economic accounts of enterprises), the number of employees required for producing  one euro-worth-of-sales in the automotive industry is about one half of that for the average manufacturing sector. Hence, if the subsidy succeeds in switching expenditures toward cars, it will reduce the aggregate demand for labour (and increase the aggregate demand for capital): the classic Stolper – Samuelson effect of international trade theory.

How many jobs will be lost (if real wages do not fall)? Clearly it depends on the effect of the subsidy on the demand for cars, as well as on the cross-elasticity of substitution. In the unlikely case where the latter is very small (one dollar spent on cars reduces demand elsewhere by only 25 cents) back-of-the-envelope calculations (see by blog) suggest that the positive effect on the auto sector should prevail, so that the subsidy may raise overall employment, albeit only marginally (about 3600 units). In the more likely case where the cross-elasticity is large (one dollar spent on cars reduces demand elsewhere by 90 cents), the subsidy may reduce employment considerably (by about 8,800 units).

The general lesson is the following: in order to support aggregate employment, the use of industrial policies that favour capital intensive sectors is a really, really bad idea.