IMF: World Economic Outlook: October 2012
The Macroeconomic Outlook: October 8, 2012: Slides

Multipliers and Self-Financing Expansionary Fiscal Policy One Last Time

The IMF's extremely welcome and praiseworthy marking-to-market of its views on the multiplier--that the multiplier is large when monetary policy is at its zero lower bound, and consequently that IMF counsel over the past three years that the costs of an immediate shift to fiscal austerity were small were wrong--provides an opportunity to restate the central conclusions of DeLong-Summers: "Fiscal Policy in a Depressed Economy".

First, consider Jonathan Portes, who writes:

Not the Treasury view...: What explains poor growth in the UK? The IMF thinks it's fiscal policy: [I]n a commendable display of self-criticism, the Fund has gone back and reanalysed the forecasts that it made (as well as those made by the OECD and EU). Its conclusion: 

In line with these assumptions, earlier analysis by the IMF staff suggests that, on average, fiscal multipliers were near 0.5 in advanced economies during the three decades leading up to 2009. If the multipliers underlying the growth forecasts were about 0.5, as this informal evidence suggests, our results indicate that multipliers have actually been in the 0.9 to 1.7 range since the Great Recession. This finding is consistent with research suggesting that in today’s environment of substantial economic slack, monetary policy constrained by the zero lower bound, and synchronized fiscal adjustment across numerous economies, multipliers may be well above 1

Not the Treasury view What explains poor growth in the UK The IMF thinks it s fiscal policy

That is, the Fund is saying: "Delong et. al. were right; we were wrong". They even have a helpful chart, showing that the bigger the fiscal consolidation, the worse growth has been relative to IMF forecasts - implying that the Fund was drastically underestimating the negative impact of fiscal consolidation. 

Why does this matter? Everyone agrees growth since 2010 in the UK has been very disappointing… was it cutting the deficit too quickly… the spike in inflation… the impact on confidence from the eurozone?  Here at NIESR, we have taken the view that it was a combination of all of these…. The IMF have now definitively sided with those who think that tightening fiscal policy quickly and sharply had a very large and negative impact. 

Once again, the Fund deserve praise for going back, looking at their forecasts, analysing what went wrong, and saying very clearly:

We thought the impact of fiscal consolidation on growth would be relatively small. We got it wrong.

Will our government, and those of the eurozone, do the same?

Second, consider a marginal expansion (or contraction) of this-year government purchases ΔY. It has a number of effects.

It has effects on production:

  • +μΔY : a Keynesian boost to (or fall in) current-year output determined by the Keynesian multiplier μ
  • +ημΔY : a boost to (or fall in) next-year's output determined by the Keynesian boost to (or fall in) this-year's output and the economy's degree of "hysteresis" η that determines how this-year's slack aggregate demand translates into next-year's reduced potential output aggregate supply.
  • Thereafter--because hysteresis effects on aggregate supply are persistent but not permanent--decay of the hysteresis shadow cast on potential output aggregate supply at a rate ρ.

And effects on the government budget:

  • +ΔY : a boost to (or fall in) this-year's government purchases
  • -τμΔY : a boost to (or fall in) this-year's tax collections from the Keynesian boost to output.
  • -τημΔY : a boost to (or fall in) next-year's tax collections from the "hysteresis" [ next-year's reduced potential output aggregate supply.
  • Thereafter--because hysteresis effects on aggregate supply are persistent but not permanent--continued increased (or decreased) tax collections from the hysteresis shadow decaying at a rate ρ. At a real discount rate of r, the infinite-horizon sum of these persistent-but-not-permanent effects adds up to -τημΔY/(r+ρ)

How to aggregate these effects? The natural statistic to calculate is the change in the effective long-run balance--the change in debt minus the change in the present value of expected future tax revenues. The increase in spending is:

  • +ΔY : a boost to (or fall in) this-year's government purchases

The change in the present value of expected future tax revenues is:

  • -τμΔY -τημΔY/(r + ρ)

The change in the effective long-run debt balance is:

  • ΔY{1 - τμ(1 + η/(r + ρ))}

For the IMF multiplier μ=1.5, for a marginal tax share τ=0.33, and for a real interest rate r for credit-worthy sovereigns that is zero, an increase in government purchases reduces the effective long-run debt balance--expansionary fiscal policy is self-financing--as long as:

  • η/ρ > 1

That is not a high hurdle to surmount: the way to bet is that right now--in depressed economies at the zero lower bound for credit-worthy sovereigns that can borrow at zero in real terms--for the medium term of 3-5 years there is no long-run budgetary cost to governments borrowing and buying stuff, and so no reason not to. This calculus will change when economies exist from the zero lower bound on interest rates or thing shift so that they find that they can no longer borrow on such extraordinarily easy terms, but such an exit or shift will be a sign that recovery is really here, and is to be welcomed.

A year and a half ago when I started to work on "Fiscal Policy in a Depressed Economy", I thought that the policy argument had been (unfairly and depressingly) lost, and that we were in the business of setting down a marker for the future, for I thought in early 2011 that by the end of 2012 surely the long-run would have started to arrive and economies would be exiting or looking forward to exiting the zero lower bound. Not so--merely the latest of my big analytical mistakes over the past five years. With forward guidance now suggesting up to three additional years at the zero lower bound, now is still the time for governments to undertake the additional rounds of purchases that they should have undertaken in 2010.

And, tantalizingly, all of a sudden it looks as though the policy argument is no longer definitely lost…