The Katz-Miyao Debate #6: Katz on Miyao

Richard Katz (The Oriental Economist Report)

This commentary originally appeared in the "Japan-U.S. Discussion Fourm" (http://lists.nbr.org/japanforum) on May 17, 2003: posted here with the author's permission.

To Takahiro Miyao:

I'm still waiting for your evidence. All we got was more logic and more assertions. First, you asserted that changes in income and consumption taxes (up or down) have had no effect on people's consumption. Second, you made the even more remarkable assertion that consumers would not spend a yen of increased income as long as housing prices are falling. In other words, even if consumer income increased by 10%, 20%, or 30%, it would make no difference.

The reality is quite the opposite. Give consumers more money and they spend more, whether the money comes from wages, tax cuts, interest on savings, etc. Consumption is very tied to income. From 1990 through 1997, real employee compensation increased 17% and real personal consumption increased 16%. Then, during 1997-2002, when wage growth slowed to only 2.6% over the whole five years, spending also slowed, to 4.3%. Statistical tests show a very high correlation between disposable income and spending (e.g. a 90% correlation between the percentage increase or decrease in　disposable income one year and consumer spending the next year).

Meanwhile, the price of urban housing land was falling throughout, a 20% fall from 1990 through 2002. It didn't stop people from spending. In fact, statistical tests show no significant role for housing land prices in spending patterns, once income is taken into account. I don't doubt that there is some "wealth effect" (e.g. housing prices) on consumer spending, but it's surprisingly marginal.

Of course, asset prices have an effect on the economy, such as their effect on bank lending. But they appear too have surprisingly little DIRECT effect consumer spending. They affect consumer spending only insofar as they affect consumer income. So, they can be countered by applying direct remedies to the income problem, such as though cuts in income and consumption taxes, and by providing some housing-oriented tax breaks.

On taxes:

From 1991 through 2001, there was substantial negative 39% correlation between the percentage change in taxes households paid one year and the percentage change in consumer spending the following year. As their taxes went up, they spent less; as their taxes went down, they spent more. If taxes didn't matter, the correlation would be close to zero. Statisitcal regressions show that, even when other income is taken into account, the ups and downs of taxes have an important effect on spending.

Okay, that's my evidence. Where's yours?