The Welfare Cost of Subsidizing Fiat

Following the collapse of car sales in January (-32.6% on an annual basis), on Friday 2/6 the Berlusconi government  passed a bill introducing state-aid for the automobile sector. Each purchase of a new car (in exchange for an old one) will be subsidized by 1500 Euros, with the declared objective of stemming  job losses in the sector (these are forecast around 300 thousand by the Entrepreneurs Association, Confindustria). CSP, a specialized research centre in Bologna, estimates that the subsidy will apply to 500 thousands sales, and will induce 300 thousands new sales (Source:

Basic microeconomic considerations  suggests that a subsidy will entail the following effects on the sector: 1. it will reduce the price paid by consumers, increasing demand, 2. it will raise the price for, the supply of and the profits of, producers; 3 it will generate a disbursement of public funds (equal to the grant multiplied by sales). The well known result is that the amount paid by the state will exceed the benefits obtained by consumers and firms. A subsidy wastes resources (the “Harberger triangle”), because it reduces the price paid by the consumer (his monetary evaluation of the good) below the (marginal) cost of producing the good. We must also take account of other “general equilibrium” effects: 4.The subsidy generates new tax revenues (from extra profits and new sales), 5. The demand for other durables falls (you buy less flat-screen TV’s or washing machines) and so does the tax revenue from these sources; 6. The future demand for cars falls (the subsidy is temporary); 7. Part of the subsidy benefits foreign firms and jobs; 8 lobbies of other industries have good reason for claiming aid and declaring themselves damaged from the auto subsidy.

What are, approximately, the orders of magnitudes involved (I leave aside six, seven, and eight)? Suppose that, due to excess capacity, firms are able to increase production without incurring in increasing (marginal) costs, so the auto supply is flat, and suppose we  accept the (rather optimistic) estimates of CSP, that the bonus raises demand by 300 thousand  units. It then turns out (for detailed calculations check my blog ) that the subsidy implies a direct budget cost of 600m Euros and benefits buyers of new cars by 420m. Then add the additional VAT revenue on new purchases and deduce the fall of tax revenues from those other goods (especially durables) that are substituted away. The literature suggests that for every 100 Euros of extra-spending on a new (small-medium) cars, the outlays for other goods (starting from washing machines, hi-fi’s etc) fall  between 25 and 90 Euros (see for example Berry, Levinsohn e Pakes, “Automobile Prices in Market Equilibrium”, Econometrica, 1995, pp841-890). In first case of low substitution, which is very unlikely because of the crisis and the ongoing restrictions of consumer credit, the net burden for the budget would be low, 15m Euros, and there would be a net welfare gain of 405 million (= 420 -15 ). In the second, much more likely, case of high substitution, the burden on the budget would be considerable, 522 million and the society would suffer a dead-weight loss of 102 million Euros (522-420).

– It remains the argument that not subsidizing the automobile sector (e.g. Fiat) when all other countries are favouring domestic producers may be harmful. It is the same argument used to support protectionism, and would require much space. In short, wasting public resources is not advisable even if other countries are so doing. The most serious problem of the subsidy, however (and I will not mention congestion, pollution etc.) is of a “political economy” nature: the measure will open the Pandora Box of sectoral state aids. Aids whose effects will likely cancel out, and which will severely undermine the fragile sustainability of the Italian public debt.

3 Responses to "The Welfare Cost of Subsidizing Fiat"

  1. Guest   February 8, 2009 at 2:43 pm

    Well, there must be something wrong with this analysis because it refers to the “marginal.” Didn’t Sraffa put a bullet through Marshall decades ago?

  2. Paolo Manasse   February 8, 2009 at 3:11 pm

    Looking at modern micro and macro, I guess one should conclude that Marshall did survive Sraffa…Anyway, I am just taking a flat supply curve (average and marginal costs constant),for simplicity.

  3. Guest   February 9, 2009 at 11:12 pm

    It wasn’t until I got to the end of this article that I realized thatthe Fiat _car_ company (not the fiat monetary system) was beingwritten about. You got me pretty good that time, thanks..