I’m responding (as usual with a long delay) to a post by Chrish Bertram over at Crooked Timber on the Baumol effect and also to discussions more recently on this blog about public spending on health and education. The reference is to a paper from the 1960s: Baumol, W. (1967), ‘Macroeconomics of unbalanced growth: the anatomy of the urban crisis’, American Economic Review 57(3), 415–26.

To simplify drastically, the Baumol effect arises when productivity growth is more rapid in the goods-producing sector than in the service sector, and particularly in the provision of ‘human services’, including health, education, culture and recreational services (including the subject of Chris’ post, restauarants). Since labour and capital are mobile between sectors in the long run, wages grow at much the same rate in both sectors, so the price of services has to rise relative to the price of goods.

In addition, we need to assume that, for broadly defined categories of goods or services, demand is inelastic. That means that, when the price goes up, we consume less, but not so much less that our total expenditure falls. (This won’t be true in cases where goods with declining prices, such as vacuum-packed cassoulets in Chris’ example, are close substitutes for services with rising prices).

Given all this, we can expect to see an increasing share of income being spent on services like health and education, even though the real supply of these services is increasing only modestly (in the extreme case of zero productivity growth in services, supply will actually decline). Conversely, we will consume substantially increased quantities of manufactures even as employment in the manufacturing sector declines.

Since productivity is rising in at least part of the economy, we expect incomes to rise as well. But it’s useful to observe that the relative price effects would arise even if there were no income growth at all.

Once we understand the role of relative prices, it’s easy to understand things that puzzle a lot of quite smart people. For example, in the last few years we’ve seen a book called Myths of Rich and Poor by Cox and Alm asserting that even the poor are much better off than they used to be, and another called The Two-Income Trap , by Warren and Tyagi asserting that even the middle classes are finding it hard to pay for the bare necessities. These books are reminiscent of the story of three visually challenged persons examining an elephant.

In essence, Cox and Alm focus on goods while Warren and Tyagi focus on services. Thus Cox and Alm note that “many ‘poor’ people enjoy home and car ownership, air conditioning, cable television, and an entire collection of things that most truly poor people would never even dream they could own” while Warren and Tyagi note that, in two-income middle class families ” both incomes are almost entirely committed to necessities, such as home and car payments, health insurance and children’s education costs”. With the exception of cars and houses, the split is perfect. A correct analysis would score a win for Cox and Alm on cars (cheaper in quality-adjusted real terms) and for Warren and Tyagi on housing (more expensive).

As we’ve already seen, when we use a standard CPI adjustment, Cox and Alm are dead wrong about poverty in the US. However, it’s important to note that the situation of the poor in the US is substantially affected by the availability and quality of publicly provided health (Medicaid) and education. In states where the poor have access to Medicaid and good quality public education, the 1963 poverty line indexed by the CPI allows for an increase in absolute consumption. Where access to Medicaid is tightly restricted and schools are funded by local taxes (that is, poor neighborhoods get lousy schools) the reverse is true.