Monday, December 29, 2014

Indirect and Incorrect Tests of Ricardian Equivalence and Supply-side Economics

The latest garbage from Stephen Moore on what Paul Krugman allegedly said and on tax revenues during the 1980’s got me thinking oddly about an alleged test of Ricardian Equivalence by Paul Evans . But let’s start with this whooper from Moore about Art Laffer’s cocktail napkin:
It was 40 years ago this month that two of President Gerald Ford’s top White House advisers, Dick Cheney and Don Rumsfeld, gathered for a steak dinner at the Two Continents restaurant in Washington with Wall Street Journal editorial writer Jude Wanniski and Arthur Laffer ... This was the first real post-World War II intellectual challenge to the reigning orthodoxy of Keynesian economics, which preached that when the economy is growing too slowly, the government should stimulate demand for products with surges in spending.
I guess Moore has never heard of the Friedman-Phelps model and how combining it with rational expectations started the New Classical challenge to Keynesian economics. This all started well before this dinner and Gerald Ford’s stupid WIN buttons. But our current post is about the 1980’s when some classical economists were predicting that the Reagan tax cuts will lead to less national savings and higher real interest rates thereby crowding-out investment and lowering long-term growth. Proponents of Ricardian Equivalence, however, suggested that tax cuts not associated with permanent reductions in government spending would be entirely saved, which meant there would be no effect at all. Evans ran a regression of interest rates and the Reagan deficit to argue that the traditional effect was not consistent with empirical evidence. While some argued that non-fiscal macroeconomic factors could be lead to a negative relationship (recessions lowering interest rates and tax revenues), Evans noted he was indeed capturing a period of strong fiscal stimulus. But note during this period, consumption did rise as did real interest rates. What Evans captured was a decline in nominal interest rates from the steep decline in inflation. In a way Moore’s latest commits the same two sins inherent in the Evans regression equation. First it is only an indirect test- which we will get back to. Secondly, it confuses nominal with real as Paul explained:
I have a suspicion that the Post forced him to include the inflation-adjusted number, rather than let him get away with the gee-whiz nominal number, which is, um, inflated by the relatively high rate of inflation that prevailed even in the later Reagan years. In any case, however, Moore offers no context, leaving the impression that this was an extraordinary achievement. So I looked at real federal receipts over a longer period
In an update, Paul noted:
I actually should have used total federal receipts, not just taxes. When you do this the pattern is weaker but basically the same: Real revenue growth 36 percent in the 8 years before Reagan, 26 percent under Reagan, 28 percent in the years following.
In a way, this mea culpa may have been unnecessary. Moore wanted us to believe that the income tax cuts led to more tax revenues but he included in numbers the receipts from the increase in the payroll tax. I noted the importance of this:
The usual line is that Federal tax revenues almost doubled from $517.1 billion in 1980 to $1032.0 in 1990. The inflation-adjusted part comes from the fact that the GDP deflator rose by 50.3% over this period so in real terms revenues rose by 32.8% over the entire decade. But there is another serious problem with this that anyone who followed the various tax policy changes during the Reagan years should know. Yes income tax rates were cut in 1981 but there were various tax rate increases that followed including a significant increase in payroll tax rates in 1983 ... Payroll taxes rose from $157.8 billion in 1980 to $380 billion in 1990. Yes, a 140.8% nominal increase and a 60.3% increase in real terms when this tax rate was increased. All other Federal taxes therefore rose by only 20.8% in real terms over the decade.
But isn’t supply-side economics about the alleged benefit to output growth rather than tax revenues? To put this in context, had there been no damage to economic growth from the 1981 tax cut, output would have grown by 3.5% per year or by more than 40% during the decade. However, we got something on the order of 3% output growth per year or less than 35% during the decade, which is why income taxes rose by less than 21%. Payroll taxes rose by 60% precisely because that tax rate was increased. But maybe it is best to simply turn the microphone over to Dean Baker:
Moore likes to have the world begin in 1982. This was the trough of the steepest downturn in the post-war era. Economies typically bounce back from steep downturns with steep upturns (not in the most recent one, because that was the result of a collapsed bubble). For this reason the recovery is primarily a measure of the severity of the downturn. The more honest way to measure an economy's performance is comparing it to the prior business cycle peak. By this measure, the 1980s had slower growth then high tax days of the 1970s and much worse growth than the higher tax days of the 1960s. The world looks a bit better if we start at 1982, but that is not a serious way to assess the Reagan performance. This reminds me of a time when I was on a radio show with Moore. Then too he was touting the wonders of the Reagan boom. I pointed out that the 1970s had better growth than the 1980s and offered Moore a $100 bet on the topic. Moore accepted and then touted the 1982 to 1989 growth rate. When I pointed out that the 1980s began in 1980, Moore got upset. Unfortunately for Moore and other Laffer-Reagan backers, the 1980s still begin in 1980.

No comments: