It's Not Roughly Zero
In theory and in practice.
If someone askes me, “Jason, how large is the fiscal multiplier?”, my answer will be simple. First, it will depend on a few factors: What is the method and timing of financing? What is the state of the economy and what does stratified data tell us? What is the outlook of the monetary authority? What is the duration and type of fiscal spending?
Ramey (2011) tells us that a deficit-financed, temporary fiscal injection, most likely will have a multiplier of ~ .8 - 1.5. Perhaps, one could interpret multipliers of < 1 as a standard crowding out story. Scott Sumner claims something different; that the fiscal multiplier is roughly zero. In an AS/AD model, this claim is defensible. In fact, it is “correct” in that it makes sense with standard models, but it is “wrong” when addressing the reality of fiscal multipliers and what they ought to be.
Why has the effect of fiscal stimulus been so meager in recent years? After all, interest rates in the United States have been close to zero since the end of 2008. The most likely explanation is monetary offset, a concept built into modern central bank policy but poorly understood. We can visualize monetary offset with the Keynesian aggregate supply and demand diagram used in introductory economics textbooks. If fiscal stimulus works, it’s by shifting the aggregate demand (AD) curve to the right. This tends to raise both prices and output as the economy moves from point A to point B, although in the very long run, only prices are affected.
Monetary Offset exists, but it’s not automatic - it is not a constant phenomenon empirically or even fully occurs in an “optimal policy” world.
Woodford (2011) shows that in a *shudders* RBC economy, the fiscal multiplier is generally not 0. Fiscal policy, a fiscal shock, will actuate responses in the private sector that influences a list of variables: output, consumption ages and labour hours. Said multiplier might not be large, but it is non-zero. One can contrive a model in which the fiscal multiplier was roughly zero or exactly zero, but that would be a problem - one would have to contrive a model.
Connecting back to the previously linked Ramey paper, Ramey & Zubairy (2014) use VARs to estimate fiscal multipliers with a military news shock identification technique. Again, there are varying levels of multipliers and many are contestable, but what is clear again is that the fiscal multiplier is not roughly zero. Consistent with *shudders* even NK models, multipliers can be as high as 1.5 at the ZLB.
Ramey’s macro shock paper also shows a list of estimated multipliers from various papers, consistent with the view that the fiscal multiplier is not zero and higher during periods with slack. Again, monetary offset is real and consistent with an inflation targeting regime, but when you approach the question of how this works in practice and in theory, you must be more languid about what is changed and will be changed. Current data does not support the notion that fiscal policy will be negated by monetary policy during a recession and models in which a zero fiscal multiplier is true must be contrived. There is overwhelming econometric evidence to suggest that, empirically, monetary policy makers do not fully offset fiscal shocks. Even in a normative sense, it would be hard to come to this conclusion. There are little to no models where optimal monetary policy would generate a multiplier of roughly 0. In the normative sense, optimal monetary policy regimes would not be inclined to fully offset fiscal shocks.
What role does this leave for fiscal policy? What would an effective fiscal stimulus look like? It turns out that fiscal policy could play a role, but only through supply-side channels. Return to the AS/AD diagram discussed above. If policymakers were able to increase aggregate supply, then the Fed would be under no pressure to offset the effects with tighter monetary policy. That’s because supply-side tax cuts actually tend to lower the inflation rates and raise growth. A good example is a cut in the employer-side of the payroll tax, which would encourage hiring but would not boost wages or prices. Indeed, the cost of labor from the firm’s perspective would decline, whereas workers would see no change in take home pay. Some economists believe that cuts in taxes on investment income might also boost aggregate supply.
In Sumner’s defense, he doesn’t claim that fiscal policy, fiscal shocks, will not affect real output. From his NGDP lens, the multiplier is about fiscal policy’s effect on nominal GDP and/or inflation. But if such is the case, then there isn’t any point to the discussion. Is the phrase “fiscal multiplier” synonymous with the phrase “the effect of an exogenous change in real government purchases on NGDP” or is it synonymous with the phrase “the effect of an exogenous change in real government purchases on RGDP” - I believe, the answer is the latter. This can be confusing during normal discourse when people claim that fiscal deficits might be inflationary, but that is irrelevant in the context of a multiplier. In terms of long-term fiscal sustainability? Sure, but that’s a different discussion.
There’s still a solid defense to the standard MM view that I agree with/sympathize with. “Monetary-only” will make sense if NGDP doesn’t fall - what the monetary authority does has a great influence on chance of a recession to happen and/or the severity of it. But, once you’re in the recession and once you think about the mediums in which outcomes can be achieved and the general income distribution, is a “monetary-only” response warranted? Given current data and modeling? I don’t think you can confidently say that. (Note: I’m still very much on the “monetary does the heavy lifting” side) One reason why we are fine with fiscal action now, is that we know monetary benefits aren’t immediately passing through to people - it seems plausible that this is true in literally every recession. In a “normal” recession such as 08, it seems plausible that a robust targeted fiscal response is warranted when addressing all possible economic outcomes even if it’s still much smaller relative to the size of the fiscal response.
Last paragraph written with help of Trevor Chow.