Thursday, June 16, 2011

The Laffer Curve, JFK, Tax Cuts, Supply-Side Economics, and Snake Oil

Continuing my reading of Supply-Side Economics: A Critical Appraisal, a 1982 collection of essays by authors ranging from Hazlett to Tobin published by GMU. One thing I gleaned from the book is that 1970s policymakers were having the same debates that think tanks are having today: What is the cause of the current deficit -- the Kennedy (Bush) tax cuts, or the (bipartisan) increase in government spending? Which mix of tax hikes and/or spending cuts is the best way to close that deficit? Many of the analyses in these essays could have been written today. But one proposal has been (hopefully) discredited-- that you can balance the budget simply by cutting taxes. This post will deal with the tax/fiscal policy debate. My next (much shorter) post will deal with the monetary policy debate of the time.

Since Art Laffer is considered the caricature of the supply-side movement, I was eager to read his essay "Government Exactions and Revenue Deficiencies" (free PDF at Cato) first. Laffer's theoretical argument about tax cuts is a bit more complicated than recognized, he presents a standard production function with capital and labor as inputs within a closed economy, and illustrates the effects on both variables given a change in taxes on one or both. He uses a standard microeconomic substitution and income effect analysis. His model draws five conclusions (P. 191):
1. Lower tax rates correspond to higher output.
2. Lower tax rates on either factor increase employment of both factors.
3. With government spending held constant, the constellation of tax rates affects output. How taxes are collected is important.
4. Lowered tax rates on any factor may or may not lower total revenue.
5. With revenue held constant, changes in the pairing of tax rates may shape the distribution of after-tax spending power, but only indirectly...when one factor's tax rate is raised and the other's is lowered, the second factor will end up in worse economic shape.

(I confess that even after reading his underpinnings #5 doesn't make much sense to me, particularly in a closed economy. )

Laffer then proceeds to analyze the "Kennedy" tax cuts of 1962-1965 to determine whether they boosted economic output and, crucially, to determine their effect on tax revenues. Laffer then compares the situation of 1962 to 1980 to make a policy recommendation.

The major differences between 1962 and 1980 (P. 194-195):
Defense spending 46.2 percent of the federal budget.
Transfer payments 25.2 percent of the federal budget.
1980 (1979):
Defense spending 21.3 percent of federal budget.
Transfer payments 41.2 percent of federal budget.

The major similarities between 1962 and 1980:
Capacity utilization estimated around 83 percent.
Elevated unemployment.

Laffer purports that the effective tax rate was higher in 1980 than the 1960s because of (1) inflation-induced "bracket creep," (2) inflation-induced understatement of true economic depreciation of capital and overstatement of capital gains, and (3) increases in the social security tax rate and wage base. Indeed, tax receipts constituted 36 percent of GNP in 1979 compared with 28.3 percent in 1963.

Ironically, the 1962 Kennedy tax cuts were promoted by Democrats and opposed by Republicans concerned with the deficit (as pointed out in Bruce Bartlett's essay in the book). Democrat Wilbur Mills gave the original supply-side speech on the house floor, arguing that the tax cuts would result in "the federal budget being balanced sooner than would be the case in the absence of these tax cuts... these lower rates of taxation will bring in at least $12 billion of additional revenue" (Pg. 278).

The Kennedy tax cuts consisted of investment tax credits, more favorable depreciation schedules, and lower corporate and personal income tax rates. There is little disagreement among economists today that the tax cuts spurred economic growth and boosted employment. No one disagrees that incentives matter, and taxing someone at a 91% marginal rate reduces their incentive to work. Laffer presents the changes in marginal personal income tax rates across all brackets (there were 25 personal income tax brackets in the 50's and 60's): (Pg. 197)

From 1963 to 1966, unemployment dropped from 5.6 percent to 3.8 pecent while real GNP grew at an annual rate of 5.7 percent (consistent with the trend since World War II, not pointed out by Laffer). The ratio of government spending to GNP fell slightly.

Laffer admits that "whether this expansion in economic activity and the general tax base...was sufficiently large enough to offset the tax rate reduction's negative revenue effects, is subject to considerable debate" (Pg. 198, emphasis mine). Laffer cites a Wall Street Journal article giving "anecdotal evidence" that tax revenue from the highest brackets increased every year from 1962 to 1965. However, Laffer admits that it appears taxpayers as a whole were on the wrong side of the "Laffer Curve," overall revenue decreased.

Laffer then doubles down-- he cites a study by Canto, Joines, and Webb using an ARIMA univariate time-series model that forecasts "what revenues would have been if the economy had continued to evolve along its normal path and the Kennedy tax cuts not been enacted." The authors found that revenues fell by almost $30 billion less than Kennedy's Treasury had projected: "These results... suggest that it is as likely that the federal tax cuts in 1962 and 1964 increased revenues as that they reduced them" (Pg. 199).

Evidence in hand, Laffer immediately pivots to 1980: "[T]he tax rate cut proposed by the Reagan administration (30 percent across-the-board, to be phased in over three years), is modest... The expansion of the economy that would occur... (and) the revenue feedback effects from across-the-board tax rate reductions are likely to be greater than those experienced fifteen years ago" (Pg. 201, emphasis mine).

Laffer ups the ante:
"It is reasonable to conclude (the tax cuts) would be self-financing in less than two years... By the third year of the tax reduction program, it is likely that the net revenue gains from the plan's first installment would offset completely the revenue reductions...Thus, the proposed Reagan tax cut has a far better chance of balancing the budget while restoring vitality to the American economy than programs attempted by the Carter administration" (Pg. 201, emphases mine). Note how this echoes Wilbur Mills remarks 20 years earlier.

David Henderson's essay (free PDF via Cato) takes Laffer to task for the sloppy empirics. He points out that the naive time-series forecast done by Canto, Joines, and Webb obviously doesn't control for other factors such as the expansion of the tax base due to the increase in Baby Boomers entering the workforce and the increased amount of international trade (Pg. 227). The omitted variables may be important, and probably don't exist in the context of 1980 (the low-hanging fruit was already picked).

It was a lot to sell an economic policy to Congress on (it was the Kemp-Roth bill). I noticed how in several essays, including Laffer's, newspaper articles were cited for support instead of academic papers. Thomas Hazlett wrote a critique of supply-side thinking from the Austrian perspective, and had this zinger: "The supply-side argument is unique: its 'theoretical' works have been written by journalists... supply-side economists have spent no time theorizing while the supply-side journalists have been breaking their backs with the heavy chore of model-building..." (Pg. 94).

Max Moszer's essay approaches the Laffer Curve from a Keynesian AD perspective. While he notes that he would certainly support tax relief, he feels that Laffer's revenue assumptions are theoretically unsound (Pg. 204). Moszer points out that there is no evidence the Kennedy tax cuts boosted productivity, which gets to the heart of the supply-side promise: the resulting increase in productivity will boost GDP and tax revenues while holding down inflation. From an AD perspective "For the supply effects of the Laffer Curve to work... the multiplier effect has to be large. A tax cut to an effective average rate of 10 percent would require a jump in real output to $2240 billion just to (hold tax revenue constant)... This yields a multiplier requirement as large as 10" (Pg. 222, emphasis mine). Most econometric models of the time (and today) had multipliers no larger than 2.

Moszer proposes an alternative: continuing with deregulation would likely yield better productivity gains and tax returns than the Reagan tax cuts. Henderson simply argues that "there are many other good reasons for cutting taxes. Taxation...denies our freedom... (and) civil liberties and reduces our material well-being. Taxation has made it easier for the government to wage war abroad..but the argument for tax cuts on the basis of the Laffer Curve is a castle made of sand" (Pg. 228, emphases mine).

Reihan Salam points us to a 1982 New Yorker piece on supply-side economics, where Laffer comes across "humble" about his assertions, and supposedly even supportive of Keynesian AD stimulus during a depression. But the point of his essay and his comments in 1980 are clear: The Reagan tax cuts would stimulate growth, tax revenues, and lead to a balanced budget even without a decrease in government spending.

And so the Kemp-Roth Bill became the law of the land-- tax brackets were consolidated and tax rates were lowered in 1982 (history here, free PDF):
The top marginal rate became 50%, the bottom was cut to 0%. $85,600-- the top bracket for married filing jointly -- is equivalent to $190,808 today. (Note that today we have a debate on whether to raise taxes on households married filing jointly earning $372,951 to a marginal rate above 35%. Our effective tax rate in 2009 was lower than at any time since 1950.)

From the National Review archives: Reagan's adviser, Paul Craig Roberts (who also has an essay in the book), submitted a white paper in February, 1981 projecting that the government would run a budget surplus by 1984.
Although federal revenues as a percentage of gross national product were projected to fall from 21.1 per cent in 1981 to 19.3 per cent in 1984, the dollar amount of budget receipts was expected to increase from $600.2 billion in the former year to $772.1 billion in the latter. In the same period, federal outlays were to rise from $654.7 billion to $771.6 billion, although falling in relation to GNP from 23.0 per cent to 19.3 per cent. In essence, this budget policy represented an effort to bring receipts and outlays in relation to GNP more nearly in line with the average postwar experience.
Unfortunately, tax revenue failed to rise as projected, and Reagan-era spending rose faster than projected:
Although the growth in federal outlays, in both nominal and real terms, slowed materially from fiscal 1980 through fiscal 1989, total outlays substantially exceeded those proposed in every Reagan budget. As a result, even had the revenues projected in the White Paper been realized, the budget would have failed to come into balance in 1984, when actual outlays of $851.8 billion were $80 billion more than had been contemplated.

Paul Krugman gives us this chart:
"(T)he revenue track under Reagan ...: a drop in revenues, then a resumption of growth, but no return to the previous trend. This is exactly what you would expect to see if supply-side economics were just plain wrong: revenues are permanently reduced relative to what they would otherwise have been."

You can get the data from the St. Louis Fed, but for time's sake Barry Ritholtz takes us to this chart:
And while it would be easy to blame the deficits on a Democratic Congress, there is some evidence to the contrary. From

But what about the real heart of the supply-side argument-- that productivity would increase as a result of the tax cuts -- just as Laffer believes happened from 1962-1965? As David Beckworth illustrates:

The Reagan tax policy's effect on productivity is certainly not obvious (or at least not until the Clinton administration, as some really strain to claim). If the Kennedy tax cuts boosted productivity within 3 years, and the Reagan policy cut an effective tax rate in 1980 that was higher than it was in 1962, then Laffer's hypothesis is apparently disproved by the lack of productivity growth in Reagan's eight years.

It is worth noting that Reagan later raised the payroll tax, the same sort of increase that Laffer considered a major reason effective tax rates had risen during the 1970s, and would later sign several other tax increases into law due to concern over deficits.

My next post will deal with the monetary policy debate of the time, since obviously the Fed-induced recession was a major factor in economic growth and tax revenues. Art Laffer comes out even worse on his monetary theories, quite frankly.

No comments: