Saturday, January 31, 2009

Does Protection Have No Impact on Aggregate Demand?

Nick Rowe makes an argument against the “Buy American” provisions that I have made in the past:

Most (all?) economists agree that in a global recession, when each country wants to boost demand for the goods it produces, policies which steer demand to domestically-produced goods are individually rational (provided other countries don't retaliate), but collectively irrational when all countries do the same. I think most economists are wrong. It's not just collectively irrational, but individually irrational as well, at least for countries with flexible exchange rates ... In normal times, outside of a liquidity trap, an expansionary fiscal policy will put upward pressure on interest rates as the demand for money increases with higher income. Or the central bank raises interest rates to offset the increased demand to keep inflation on target. An increase in domestic interest rates will cause a capital account inflow, which causes the exchange rate to appreciate. The exchange rate appreciation will cause net exports to fall. The fall in net exports offsets the expansionary fiscal policy. Under imperfect capital mobility the offset will be partial. Under perfect capital mobility there will be full offset, for a small open economy. So in normal times, part or all of the increased demand from an expansionary fiscal policy will be lost due to a decline in net exports. Some or all of the extra demand just leaks out to foreign countries.

Nick then admits that if we are in a liquidity trap, the interest rate to capital account channel is cut off so a fiscal stimulus does not necessarily crowd-out net exports but then he writes:

A "buy domestic" policy will not shift demand towards domestic goods. If it did, so that imports fell and net exports increased, the current account surplus would merely cause the exchange rate to appreciate so that net exports fell to their original level. The current account must stay the same, because the capital account stays the same, because the interest rate differential stays the same, because interest rates stay the same.

Dani Rodrik, however, takes another view:

Yes it does. And not just in theory, but also in practice. The evidence comes from the 1930s, and from the work of Ben Bernanke himself (along with other scholars like Barry Eichengreen). The important finding is that countries that devalued their currencies by getting off the gold standard were able to recover more quickly, thanks in part to an increase in their net exports relative to countries that stayed on gold. Note that a currency depreciation amounts to a policy of combining import tariffs with export subsidies--hence the mercantilist intent and effect.

Dani also notes:

How much of a boost to economic activity will a fiscal stimulus provide? For those who believe that we have entered a Keynesian world of shortage of aggregate demand--me included--the answer depends on the Keynesian multiplier. The size of this multiplier depends in turn on three things in particular, the marginal propensity to consume (c), the marginal tax rate (t), and the marginal propensity to import (m). If c = 0.8, t = 0.2, and m = 0.2, the Keynesian multiplier is 1.8 (=1/(1-c(1-t)+m)). A $1 trillion fiscal stimulus would increase GDP by $1.8 trillion. Now suppose that we had a way to raise the multiplier by more than half, from 1.8 to 2.8. The same fiscal stimulus would now produce an increase in GDP of $2.8 trillion--quite a difference. Nice deal if you can get it. In fact you can. It is pretty easy to increase the multiplier; just raise import tariffs by enough so that the marginal propensity to import out of income is reduced substantially (to zero if you want the multiplier to go all the way to 2.8). Yes, yes, import protection is inefficient and not a very neighborly thing to do--but should we really care if the alternative is significantly lower growth and higher unemployment? More to the point, will Obama and his advisers care?

I guess Nick can come back and say Dani was assuming fixed exchange rates. So how is this all supposed to work out under fixed interest rates but floating exchange rates?

Our model is essentially:

Y = D(Y) + X(Y, e)

where Y = real GDP, D = domestic demand, X = net exports, and e is the real exchange rate. Let’s consider a hypothetical economy known as Obamia that has a domestic marginal propensity to spend = 0.8 and a marginal propensity to import = 0.2 and wants to increase real GDP by $1000 (think of America as one billion times the size of Obamia). Under fixed exchange rates and no trade protection, the multiplier is 2.5 so government purchases would have to be raised by $400 if no other policy tool was used. As real GDP rose increased by $1000, imports would increase by $200.

But suppose that the conservative part of Obamia balks at a large fiscal stimulus and its leaders reach some bipartisan compromise of having government purchases rise by only $200. Dani’s point is that if we adopt a mercantilist policy to increase the net export schedule by $200, then we can still achieve the real GDP goal.

Nick’s floating exchange rate version of the model, however, has the exchange rate automatically adjust such that the ultimate change in net exports is zero. In this case, the multiplier for fiscal policy is 5 and the multiplier for mercantilist policy is zero. In other words, a $200 increase in government purchases still achieves the goal of increasing real GDP by $1000. Lesson learned – floating exchange rates can achieve the same goal as Dani’s mercantilism. There is one difference, however, between the two approaches. Mercantilism often works by protecting the import competing sector. Under floating exchange rates, we are more likely to see increased employment in the export sector.


Nick Rowe said...


"There is one difference, however, between the two approaches. Mercantilism often works by protecting the import competing sector. Under floating exchange rates, we are more likely to see increased employment in the export sector."

Yes. Agreed. I missed that point in my post.

Dani Rodrick is assuming fixed exchange rates. That's one difference. But in the 1930's, under the gold standard, I think (but am not 100% sure on this), that if a country devalues against gold, that is equivalent to increasing its money supply. If ALL countries devalue against gold, that is equivalent to a revaluation of gold, and equivalent to an increase in international reserves, or an increase in the global money supply.

So, if I've got this right, the 1930's had an additional problem that, thankfully, we don't have now.

I am going to try to get my head around this, then maybe post on it.

Anonymous said...

Rowe's argument ignores China's peg. That makes it useless, in my view.

ProGrowthLiberal said...

Anon - if you believe that the China peg turns the current state of international macroeconomics into a Bretton Woods II (see Brad Setser) then we have Dani's model. If China ever drops its peg (hopefully) we have Nick's model.

Nick Rowe said...

Anon: even if China had absolutely fixed exchange rates with the US (and it doesn't), since China has only about 10% of US trade (imports+ exports) China would only make my conclusions 10% false. Not a big deal.

TheTrucker said...

The Chinese government is buying T-bills with excess dollars so as to support the dollar and keep it from falling. I fail to see that we have a floating exchange rate. In reality it is a dollar peg and the only way to defeat it is with import tariffs. The "Buy American" seems a mild and sneaky import tariff.

Anonymous said...

Anon here.
"Anon: even if China had absolutely fixed exchange rates with the US (and it doesn't), since China has only about 10% of US trade (imports+ exports) China would only make my conclusions 10% false. Not a big deal."

If China had 10% of America's current account deficit, your argument would hold.

Your argument does not hold because it considers trade as a whole, not the current account deficit. That latter is the problem.

Anonymous said...

"At the same time, they want to lessen dependence on China, which requires that Chinese policymakers stimulate domestic demand to a sufficient extent to allow for China to ease purchases of Treasuries and allow the Yuan to appreciate in a nondisruptive fashion. Seems like a steep expectation for the export-dependent Chinese, you are now faced with faltering growth rates. If the Chinese don’t cooperate, a portion of any US stimulus is lost to higher imports –always remember that the US doesn’t have much excess productive capacity in tradable goods. The excess capacity exists in China. And Congress would be less than happy to see US tax dollars supporting Chinese jobs."Tim Duy Threading the Needle

(Bold mine)

What is the point of worrying about "Buy America" when all the excess capacity in tradable goods is in China?

kevin quinn said...

I'm not sure that I agree with Nick's analysis here. If there are unemployed resources - even without a liquidity trap - then the positive effect of increased income on national saving makes a difference. So a fiscal expansion that initially reduces saving both raises interest rates and increases Y. In the new equilibrium there will then be less-than-complete-crowding-out of net exports. On the other hand, protectionism in an economy with unemployed resources can succeed by increasing saving along with income.