It’s not about the multiplier
Mar 21st 2011, 18:50 by G.I. | WASHINGTON
[Greg Ip] HOW you feel about fiscal stimulus generally comes down to how big you think the multiplier is. Does a dollar of stimulus produce two dollars of GDP, one dollar, or none? Stimulus sceptics generally assign a lower number to the multiplier. Conservatives usually assign a lower multiple to government spending than to tax cuts.
An extreme example would be that of House Republicans who backed last December’s tax deal, which both extended existing tax cuts and enacted new ones, on the grounds that it would avert great harm to the economy. That implied a large, positive multiplier on tax cuts. They argue, however, that their proposed cuts to spending this year will actually raise employment, implying a negative multiplier for government spending. That claim has been backed by John Taylor.
I find it hard to believe that the multiplier is zero, much less negative, in an economy sporting interest rates jammed up against zero and a reserve currency. But rather than rehash that debate, let’s take a look at an intriguing new case being made against stimulus via government spending. Proponents of this case agree the government spending multiplier is positive, but maintain that the thing that is being multiplied, GDP, is the wrong policy target. Policymakers should be concerned with welfare, not GDP, and those aren’t the same thing. If government spends the money badly, it might not raise anyone’s welfare.
Greg Mankiw and Matthew Weinzierl elegantly lay out this argument in a paperpresented at the Brookings Papers on Economic Activity on March 17. [Note: See my digression at the end.] Their model assumes that in equilibrium, society has as much government spending as it wants; an additional dollar of government spending produces exactly as much welfare as an additional dollar of private spending. Thereafter, more government spending makes society worse off because it hogs dollars that people would rather spend themselves. Suppose the ideal level of government spending is 25% of GDP. It then rises to 26% as private consumption falls to 74%. GDP would be the same but people would be worse off because the welfare from that increased government spending didn’t make up for the welfare lost through lower private consumption.
Suppose then a shock comes along that pushes down private consumption; government spending will rise above 25% of GDP. The best policy for pushing GDP back to its potential, then, is one that also returns private consumption back to its ideal share of total output. In their model, the best way to do this is through conventional monetary policy, that is, lower short-term rates. If short-term interest rates can’t go lower because they’re stuck at zero, the second best way is through unconventional monetary policy, that is, for the Federal Reserve to commit to driving up inflation (and thus pushing down real interest rates). The third best way is through a tax cut that mimics monetary policy by acting like a negative interest rate. Mankiw and Weinzierl propose an investment subsidy. Only after that possibility is exhausted, should policymakers turn to the fourth option, higher government spending. Their model finds that policies that maximise welfare often have the lowest multipliers.
What are the flaws in their argument? BPEA has a silly edict that its proceedings remain off the record, so I present these points without citation. An investment subsidy may not work if no one will invest because the housing market has collapsed and credit has been cut off. Second, boosting investment could distort the composition of private consumption, hurting welfare. For example, the homebuyer’s tax credit boosted investment in houses, perhaps above its equilibrium level. Third, Paul Krugman argues, in the real world the Fed can’t sufficiently commit to unconventional monetary policy because of political blowback.
I don’t agree with Krugman. The Fed seems to have amply demonstrated it is willing to pursue unconventional policy, including calling for higher inflation, in spite of the political blowback. It appears that central bank independence works symmetrically, providing the means to pursue both unpopular easy monetary policy and unpopular tight policy. But suppose for a moment we assume that unconventional monetary policy is so politically unpopular that it is out of the question. In that world, logical forms of fiscal stimulus would be even more unpopular and thus unlikely.
To me, the biggest practical shortcoming of Mankiw’s and Weinzierl’s model is that it requires that we discard GDP as our measure of welfare. It was Keynes who implicitly equated the two by arguing that in a liquidity trap we may as well bury bottles of money and pay people to dig them up. Nick Rowe has argued, and Mr Krugman concedes, that literally useless government spending of this sort does not add to welfare, even if it adds to GDP. But in the real world, everyone equates GDP (or proxies of GDP, such as employment) to welfare. A given policy, whether conservative or liberal, will be judged on its contribution to GDP. To do otherwise would allow combatants to defend or oppose policies based on radically different definitions of welfare. Joseph Stiglitz, like Mankiw and Weinzierl, is sceptical of GDP as a measure of welfare but I suspect he draws very different policy implications from that.
[Digression: Leading edge economic research centers like BPEA these days produce mostly empirical research that studies real world problems with real world data. This year’s papers deal with subjects like the financial fragility of households, the effects of quantitative easing on bond yields in the 1960s, German unemployment and unemployment duration in America. Such papers are relatively easy for non-PhDs like me to understand. They have Greek letters but you can skip them and go straight to the conclusions. Mankiw and Weinzierl break from that tradition by producing a purely theoretical paper, which means you skip the Greek letters at your peril. I have done my best to interpret them but still find myself wishing I’d pursued that PhD.]
The original post is linked here.