[Editor's Note: Contrary to myth, Austrians are not opposed to quantitative or empirical assessments of history in light of economic knowledge. The confusion may arise from the fact that Mises regarded this not as economics, strictly speaking, but as economic history, or the work of historians. Nonetheless, discussions of measurements of economic phenomena are often informative. In this article, Julian Adorney examines some Keynesian methods of measuring the economy.]
Keynesian economists claim that fiscal and monetary stimulus help cure recessions, but their evidence is largely correlational. They pump money into a recession, and if the situation improves they claim the credit. If it doesn’t improve, they argue that the picture would look much worse without their intervention. In their defense, we can’t make a control economy to test the effects of Keynesian policies versus a laissez-faire approach.
So Keynesian correlational evidence makes some sense. But what if we examined, not just 1 recession, but a history of 100 years’ worth? When you do that, it turns out, the evidence just doesn’t back up Keynesian claims. How do we assess this? Christy Romer put out a paper in 1999 in which she created a metric to measure how deep a recession was: percentage-point months of industrial production lost until previous peak is regained (PPM). Using this, she quantified every recession since 1886. The chart she created is reproduced below.
Now, we have to add 3 addendums to this data. First, Romer doesn’t include the 2009 recession in her chart. In order to be absolutely fair to the Keynesians, I’ll borrow Paul Krugman’s number for this recession of 455 PPM rather than calculate my own. He actually says 455 is a little low, but since high numbers would hurt his argument (for reasons that will become clear below), I won’t adjust it. Second, Romer doesn’t include data on the 2000-2001 recession. Using industrial production data, I figure that recession clocks in at around 63 PPM. My number’s pretty rough, though. Third, Romer analyzes the entire Great Depression at 3120 PPM, but she doesn’t break it down by year. Disaggregating the data by year is essential for our purposes, because stimulus spending on a large scale occurred in 1932 (under President Hoover) and really took effect in 1933. Fortunately, Randall Parker has an article on the data of the Great Depression where he breaks down industrial production loss year by year. That isn’t precisely the data Romer uses to arrive at her metric (she uses a combination of unemployment, industrial production, and GNP), but it presents enough parallels to be useful. Using the data, for instance, we can see that 49.7% of the industrial output loss from the Great Depression occurred from 1933 to 1937. We then multiply that 49.7% by Romer’s aggregate number of 3120. That tells us that, by Romer’s metric, about 1551 PPM of damage during the Great Depression came during or after 1933. Not the prettiest estimate, but it works.
So, in addition to Romer’s data, we now have 3 more data points. The Great Depression’s post-1932 number of 1551 PPM, the 2000-1 recession number of 63 PPM, and the 2009 recession number of 455 PPM (which Krugman gave us, and which again he admits is low). We can break all of this data up into three periods: that in which the Austrian business cycle theory mostly dominated (1887-1913), that in which the US government practiced monetary but not fiscal stimulus (1913 to 1932) and then the period in which the United States practiced both monetary and fiscal stimulus (the full Keynesian) to cure recessions (1933-2013). The second period, as a hybrid period, is useless for our purposes. So let’s study Period 1, characterized by an Austrian laissez-faire approach to recessions, and Period 3, characterized by Keynesian policies: an expansionary monetary policy and a spike in federal spending to drive up demand.
If you add up the total output loss for every single recession in Period 1, you come up with 1106.7 PPM. If you add up the output loss of all the recessions during the Keynesian years, you get 3869 PPM. But because the Keynesian period (80 years) was so much longer than the Austrian period (27 years), we have to further break it down by year. So let’s look at average output loss per year across both periods, or essentially the average economic drag produced by all the recessions in Period 1 vs Period 3. That will give us an idea of how much recessions have hurt the economy during each period. When you divide by year, you get a yearly output loss of 40.98 PPM (per year) for the Austrian years, and 46.11 PPM (per year) for the Keynesian years.
So while Keynesian stimulus creates fewer recessions (one every 5.7 years, compared to one every 3.9 years in the Austrian years), these recessions are deeper. And, taken as a whole, they’ve hurt the US economy a little worse than the frequent but generally mild recessions of the Austrian years.
This is significant. Looking at a stretch of 107 years, we’ve proven that Keynesian policy during this time has done a mathematically worse job of coping with recessions than the Austrian approach did. Stimulus creates recessions that are less frequent but much deeper than does a laissez-faire approach to handling recessions. So even if printing money and spiking government spending were free, it would be a bad idea. It prolongs and deepens recessions and causes (slightly) more economic damage over a long period than the Austrian approach. But as we know, spending is not free. In fact, advocates like Krugman might say that’s the point; with a size measured in dollars, a good fiscal stimulus package needs to be enormous to counteract large recessions. But what Keynesians often breeze past is that the American people will be left footing the bill of any stimulus.
The financial cost of just fiscal (not monetary) stimulus in the 1930s was about $542 billion (in 2013 dollars) for the New Deal. The stimulus passed by Presidents Obama and Bush to cope with the 2009 recession totaled over $1.1 trillion. Presidents from Eisenhower to Kennedy to Carter have proposed and passed fiscal stimulus measures. However, the cost of these all told is difficult to estimate, because government spending tends to expand even after the recession ends. The high stimulus budget becomes the new baseline.
The Mercatus Center, in a report on fiscal stimulus, explores this phenomenon.
(The red line is what Keynes would have advocated: a sharp rise in spending early on, followed by a decline. The blue line, which shows a steadily expanding government, is what actually happens. Stimulus spending often opens the lid on Pandora’s Box)
And that's just the cost of fiscal stimulus. When you add in an expansionary monetary policy (essentially a tax on saving), the costs rack up fast.
There is another cost of fiscal stimulus spending as well: it misallocates resources and encourages bubbles. When government takes it upon itself to replace consumer spending, the end result is a few government agents deciding which industries deserve public funding and which do not. For instance, President Obama’s 2009 stimulus program included $51 billion in green energy investment, designed to help this industry to succeed. When government picks winners and losers by industry, they create economic distortions and help noncompetitive industries to edge out competitive ones. An expansive monetary policy creates invisible taxes on savings and investment. These costs are difficult to put a dollar sign to, but they are many.
This is the legacy of Keynesian policies, the combination of fiscal and monetary stimulus. It balloons the deficit and misallocates economic resources in the name of preventing or fixing recessions. But, looking at the numbers, it does a poorer job of handling recessions than a laissez-faire Austrian approach.
