The price of imported oil in the US doubled between summer 2003 and summer 2005, reducing consumers’ purchasing power by more than 1 per cent of gross domestic product. Nevertheless, the economic slowdown that was widely expected never occurred. Consumers kept spending and businesses kept investing. … The continued strong growth contrasts sharply with the economic weakness that occurred after almost every previous significant rise in the oil price. How do we explain this remarkable difference?
Data source: BEA Table 2.1, line 34
The key to the economy’s strength in 2004 and 2005 was that household saving declined dramatically while the price of oil rose….The primary cause of this dramatic shift was the fall in interest rates and the resulting rise in mortgage refinancing. Homeowners who refinanced their mortgages took out cash and reduced their monthly payments at the same time. Much of the cash obtained by refinancing was spent on consumer durables, home improvements and the like.
The graph at the right displays the drop in personal saving to which Feldstein was referring, which was certainly dramatic. However, I would like to make a couple of observations about Feldstein’s conclusions.
The first point worth emphasizing is that shaving 1/2% per year off income growth would not really have been that noticeable. The relatively modest size of this number even for a doubling of oil prices has to do with a fact that I’ve long emphasized: given the relatively small dollar share of oil in total GDP, even a doubling in oil prices is not enough to produce a recession if the only effects were the dollar flow consequences to which Feldstein refers. Instead, in the earlier historical episodes which Feldstein mentions, GDP dropped by considerably more than this direct effect. In my opinion, this resulted from abrupt changes in the composition of spending and overall uncertainty about the future. Consumers could have maintained higher spending in these earlier episodes, too, had they been confident about the future, but chose not to.
We certainly saw the first stages of such an effect in the fourth quarter of 2005, when real GDP grew at an anemic 1.1% annual rate. In terms of accounting, one can attribute this entirely to difficulties in the auto sector– had motor vehicle sales and parts held steady, real GDP would have grown at a 3.2% rather than 1.1% rate. However, the recognizably temporary nature of the hurricane-related disruptions may have succeeded in arresting the usual oil shock macroeconomic dynamics. The typical consumer is now perhaps reasonably confident that all those problems are behind us.
A second observation I would like to make about the relation between saving and the 2005 oil shock is that, as Feldstein has of course been emphasizing throughout much of his career, the drop in saving is not something unique to 2005 but is instead the continuation of a longer term trend. The graph at the left projects the trend that I’d earlier fit to data through 2004. The decline in saving that actually appeared in 2005 is exactly what we would have predicted from extrapolating that trend. Although the mortgage refinancing and oil shock are unquestionably part of the mechanics of what produced this additional drop in saving in 2005, one also has to think that the long run trend played a very important role as well.
But none of this is to dispute the validity of the overall conclusion that Feldstein draws from all this:
The powerful effect of mortgage refinancing on consumer spending was a very happy coincidence for the American economy at a time when oil prices were depressing consumers’ real incomes. If oil prices were to rise again in 2006 or 2007, the adverse effect on consumers’ real incomes would not be offset by increased mortgage refinancing. Mortgage refinancing has now peaked and is declining. The Federal Reserve is raising interest rates again to counter the inflationary pressures that remain from the rise in energy costs. And individuals no longer have the large amounts of household equity against which to borrow.
I agree with that conclusion, in part because I would also add that the 2005:Q4 experience has already left consumers and the economy a bit unsteady. I would be quite concerned, for example, if current tensions led to a significant disruption of oil shipments from Iran… or Nigeria… or Iraq….
Markets this week seem to be shrugging off these concerns in the light of prospects of additional supplies. Let’s hope the market has this right.