Thursday, August 9, 2012

Laffernomics strikes again

In a previous post, I provided what I believe to be a serious critique of Professor Arthur Laffer's economic analysis in his book Return to Prosperity: How America Can Regain Its Economic Superpower Status (co-authored with Stephen Moore of the Wall Street Journal).  Now Dr. Laffer has recently written a new op-ed article for the aforementioned daily newspaper.

In the article, Professor Laffer attempts to demonstrate that fiscal stimulus is not only not very effective, but that it actually depresses economic growth, even in the short-run.  There are some ways that similar results could be forthcoming, but I'm not sure any of them are convincing, especially having no short-run stimulative effect.  As I address his argument, let me demonstrate that I am not constructing a straw-man and, thus, perpetrating an unpardonable sin in academic discourse.  Professor Laffer sums up his position by saying, "[M]assive spending programs have hurt the economy and left us with huge bills to pay." That seems clear enough: stimulus spending is a horrible failure that not only did not help but actually hindered recovery, and it adds to the national debt, to boot.

How does he arrive at these conclusions?  It has to be in his model.  Most of it is an old-timey Ricardian Equivalence: 

For every additional government dollar spent there is an additional private dollar taken. All the stimulus to the spending recipients is matched on a dollar-for-dollar basis every minute of every day by a depressant placed on the people who pay for these transfers. Or as a student of the dismal science might say, the total income effects of additional government spending always sum to zero.

To start off, his definition of stimulus throughout the article is very misleading, since he talks about it as an increase in government spending (which could be accompanied or unaccompanied by an increase in taxes), but there is a substantial difference between a balanced budget increase in spending and a deficit-financed increase in spending.  Because I don't want to get bogged down in semantics, I take his position that an increase in government spending is a sign of stimulus.  Realistically, however, what has happened in the United States and many other countries is primarily deficit-spending, as tax collections have not only failed to increase, but have generally declined.  Under consumption smoothing, there should be a short-run increase in aggregate demand due to a growing deficit (particularly when confronted with a liquidity trap), although it would have to be paid for by slightly lower demand in forthcoming years, so I do not find the idea that there is a dollar-for-dollar matching decline in aggregate demand by the private sector for every deficit-driven increase in government spending very believable.

Professor Laffer further states that there are substitution and price effects that will also depress aggregate demand when there is government "stimulus".  There may be substitution effects, I suppose, but they would probably be more than fully offset by the short-run increase in spending.  As for price effects, I suppose he is suggesting that there would be intensified inflation which would depress real aggregate spending.  I have two things to say about that: first, inflation would also decrease the real value of debt that is currently a significant drag on the economy and, second, it is hard to imagine a significant problem with growing inflation when interest rates are stuck around 0% for medium-to-long-run government bonds.

Another factor which Professor Laffer completely ignores is that the United States and many other advanced economies in his list are at what some economists call the zero lower bound (where government securities have nominal interest rates of 0%, so they can't be pushed lower).  This means that the next best option for an investor to do with their money is to stuff it under their mattress, put it in a shoe-box or a mayonnaise jar, or bury it in their backyard.  I wonder how preventing money from being put to such a use by issuing additional treasury debt can actually decrease output.

Professor Laffer's article also includes a chart where he puts two pieces of data for each country (which have faced some criticism by Professor Glasner).  The first piece is the change in government spending as a percent of GDP and the second is the change in real GDP growth from 2006-7 to 2008-9.  With this data, Dr. Laffer purports to show that bigger stimulus leads to slower growth rates.  Leaving aside what this data means for a moment, I came across a blog post not long ago by Nick Rowe about "Milton Friedman's Thermostat" (not the title of the post).  In it, he explains a problem with econometrics: namely, that it can be pretty much impossible to disentangle business cycles and policy instruments, if they are applied decently.  Here is my take on how this relates to the data presented: If interest rates are the primary mechanism by which policymakers attempt to counteract the business cycle and they are stuck a 0%, then a rational fiscal policy maker would try to counteract the business cycle through expansionary fiscal policy as the economy slows down (when the growth rate slows, as happened in the period under consideration).  Even without this, as Professor Glasner pointed out, many policies function as automatic stabilizers, like unemployment insurance and even eligibility for certain low-income assistance programs.

Does Professor Laffer know that his data is problematic?  He states so explicitly, while trying to make it sound like he has avoided the problem:

In many countries, an economic downturn, no matter how it's caused or the degree of change in the rate of growth, will trigger increases in public spending and therefore the appearance of a negative relationship between stimulus spending and economic growth. That is why the table focuses on changes in the rate of GDP growth, which helps isolate the effects of additional spending.

He basically says, yes, negatively changing rates of growth trigger automatic increases in public spending, but I have accounted for this by comparing changing rates of growth to changes in public spending.  He identified the problem with his data, then used it anyway and acted like it proves something!


  1. This comment has been removed by the author.

  2. You give him too much credit. He was graphing spending *as a percent of GDP*. Meaning any change in GDP, is a change in the denominator of his measure of spending, so of course declines in GDP go along with increases in (spending/GDP). That's some first rate hackery.

    Am I the only one who finds it ironic that Arthur Laffer, the guy responsible for giving conservatives the idea that tax cuts are so good for the economy that you can cut taxes while you're running a deficit and they will pay for themselves, is now writing an op-ed invoking Ricardian Equivalence, the idea that running deficits and raising taxes have the same effect on the economy?

  3. Eric,
    I agree with you on the first point (although his measure he used is the change in the rate of change in real output). On the other point, I have to point out that Laffer's primary position on tax rates is that higher tax rates discourage increased output (for lack of better words), which I think is quite distinctive from the Keynesian views of deficit-financed expenditures. That being said, I'm not so sure how his model could apparently treat increased spending so unfavorably even when financed through borrowing, while borrowing to allow lower taxes could be so favorably treated by the same model.