When asked for proof of the effectiveness of Keynes’s prescription of deficit spending to cure an ailing economy, Keynesians usually point to World War II—government ran large deficits, GDP surged, and unemployment disappeared.
The economy certainly appeared to be healthy. The military industrial sector grew larger and boosted measured GDP. Conscription forced unemployment close to zero. But did government spending on the military stimulate the economy through the vaunted Keynesian multiplier? Government spending on the military didn’t stimulate private consumption—it crowded it out. Consumers suffered.
That doesn’t disprove Keynes—maybe a smaller increase would have done the trick of fixing the economy without reducing private consumption. World War II doesn’t prove Keynes was wrong. But it sure doesn’t prove him right, either.
A better test came at the end of war. Paul Samuelson, a prominent Keynesian who later won one of the first Nobel Prizes, worried that if the war ended suddenly and government spending contracted quickly, we would face “the greatest period of unemployment and industrial dislocation which any economy has ever faced.”
Federal spending fell by 40% in 1946 and ten million soldiers were suddenly looking for work. Despite the warning of Samuelson and others, nothing was done to cushion the expected blow to the economy. But there was no recession—unemployment stayed low.
A year later, federal spending was 60% below the 1945 level. Output grew 7% and employment grew 4.8% hitting record levels. That extraordinary natural experiment could have ended the fascination economists had with the work of Keynes. Instead, the Keynesians re-calibrated their models and looked to the future.
The failure of the Obama stimulus package of 2009 to have the predicted effects on the unemployment rate didn’t change many minds either. In January of 2009, Alan Blinder of Princeton, a highly respected Keynesian, argued that a $650 billion stimulus package would be big enough to fix the economy. Instead, we spent $820 billion and in return, we have the worst recovery in modern American economic history.
Of course it’s possible that these large natural experiments shouldn’t be treated as decisive. The world is a complicated place. Lots of things happen at once. Correlation is not causation. Maybe factors besides government spending account for the lethargy of the US economy today or the vitality of the US economy in 1946.
That is why I think it isn’t just Keynesianism that’s in trouble these days but macroeconomics in general, at least the way journalists and politicians expect us to practice it. When Nobel prize winners argue that the stimulus should have been $2 trillion while other equally illustrious economists argue it should have been zero and both have studies to back up their claims, one has to wonder how much science there really is in economics.
Perhaps our profession should admit that some of the questions people want us to answer simply cannot be answered. One of those questions is whether $820 billion of additional federal spending using borrowed money is a good idea or not. I think it’s a bad idea. But my reasons for thinking so are based on logic and philosophy not fancy statistical analysis.
But hasn’t the non-partisan Congressional Budget Office found that the stimulus created millions of jobs?
Sort of. At one point in 2011, the Congressional Budget Office estimated that the stimulus had created between 400,000 and 2.4 million jobs. Not so precise. Worse, it’s not even an estimate. It’s simply the forecast the CBO did before the stimulus passed, re-calculated using the actual amount of money that was spent.
This is equivalent to sending a space probe to Mars but being unable to verify its location. When asked by the President where it is, the scientists assure him that it’s on Mars. How do they know? Because their models says so. In modern macroeconomics, we make predictions than cannot be confirmed or verified. I don’t know what to call that, but it isn’t science.
Some argue that we haven’t had enough serious recessions to model them adequately. Give us 20 or 30 more, they say, and we’ll have enough data.
I’m more inclined to side with Hayek in his 1974 Nobel Prize address, “The Pretence of Knowledge.” Hayek argued that the economy was simply too complex to allow predictions of the precision we expect from say, physics. So while a physicist can tell us where Mars will be on January 1, 2035, and actually verify the prediction when the time arrives, we economists will always struggle to identify the path of the economy six months from now.
Economics isn’t rocket science; it’s a lot harder. We should admit as much and when asked to measure things we cannot measure, we should admit our ignorance.