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.
I hear that the US fiscal stimulus contains restrictions on buying imported goods. One could perhaps argue that the Canadian fiscal stimulus also concentrates demand on non-traded goods, and so does sort of the same thing, but in a manner which is less provocative (or more devious, if you prefer). This is dangerous. If all countries follow a "buy domestic" policy, the best that could happen is that we end up with the same level of total output, but a much worse composition of output (we would all rather consume more imported and fewer domestic goods). And it could be much worse than that.
I don't have anything new to say about the collective dangers if all countries end up restricting trade as they did in the 1930's.
But I do have something to say about the benefits to an individual country of following this policy. Even if no other country retaliates, it is irrational for an individual country with flexible exchange rates to follow this policy. I am repeating a point I made before, and before, but it has suddenly become much more relevant and important, so it needs repeating. If I can convince people that I am right (OK, I don't think much of my chances either, but I can only try), then the dangers of ending up in the 1930's are smaller, because countries will not want to follow these policies.
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.
But these are NOT normal times. An expansionary fiscal policy will not cause an increase in the rate of interest. Central banks won't let it happen. The perfectly interest-elastic demand for money won't let it happen either.
An expansionary fiscal policy will not cause the rate of interest to increase. It will cause total spending to increase, and some proportion of this increased total spending will be on imports. But an increased demand for imports will cause an excess demand for foreign exchange on the current account of the balance of payments. Nothing happens on the capital account, because interest rates don't change. The excess demand for foreign exchange causes the exchange rate to depreciate (the exact opposite of the normal case). The exchange rate depreciation causes net exports to rise. This will fully offset the increased imports from increased income and spending (it has to, if the capital account stays the same). So, in the new equilibrium, all of the increased spending goes to increased demand for domestic goods. None leaks out to increased demand for foreign goods.
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.
My argument does not apply within the Eurozone of course. The countries share a common currency and so there are no exchange rates to adjust. Similarly, it does not apply to provincial fiscal policy within Canada. And it does not apply to countries on fixed exchange rates, where the foreign country will intervene in the foreign exchange market to prevent its currency appreciating against our currency.
Quick question. The media and various other people keep claiming that a "buy domestic" policy is protectionism. Is it?
Posted by: Robert McClelland | January 30, 2009 at 08:23 PM
Pretty much, yes. Both policies are designed to encourage people to substitute away from buying imports to buying domestically-produced goods.
Posted by: Stephen Gordon | January 30, 2009 at 08:32 PM
Does that still apply though if it's just the government substituting away from buying imports.
Posted by: Robert McClelland | January 30, 2009 at 09:06 PM
Hmmm. I think TypePad lost my last comment. Trying again:
Robert: yes, I think so. It's exactly as if someone imposed a tariff or quota on government purchases of imports.
Posted by: Nick Rowe | January 30, 2009 at 10:16 PM
The goal isn't to erase the trade deficit, it's to create jobs in our own country. China isn't going to start demanding pricy American products out of the blue while their own economy is in the dumpster, so we have to do it ourselves. Once Americans have jobs again they can resume spending their income on goods from China and live happily ever after.
Posted by: anonymous | January 31, 2009 at 05:34 AM
anonymous: but if I'm right, a "buy domestic" policy won't do either. It won't change the trade deficit, and it won't create jobs. Fiscal policy would create the same number of jobs regardless of whether it contains a "buy domestic" provision.
Posted by: Nick Rowe | January 31, 2009 at 08:27 AM
Politicians (as many people) need to learn that words mean things. "Buy American" can be a solidarity thing, but it really means something else and creates certain expectations. Looks like Mr. O did not learn much from the "we will re-negotiate NAFTA" incident.
Posted by: learning james | January 31, 2009 at 02:12 PM
Excellent discussion but at first I thought this stood in stark contrast to what Dani Rodrik has being saying about the use of mercentalism to increase the Keynesian multiplier. But then Dani's posts assume fixed exchange rates while you assume floating rates. More here:
econospeak.blogspot.com/2009/01/does-protection-have-no-impact-on.html
Posted by: pgl | January 31, 2009 at 04:05 PM
How will buying steel manufactured in the US instead of steel manufactured in China not create more jobs in the US? The increased cost is all in labor, not raw materials, and that is exactly what we want to create, more jobs and higher wages in the US. Buying the steel from China would simply create fewer jobs (immediately) in the US.
This is not a trade embargo. This is the US government selectively choosing the best way to increase US employment in the short-term.
Posted by: anonymous | January 31, 2009 at 04:18 PM
If you think that perhaps shipping steel from China creates more jobs in the transport process than manufacturing it in the US, then we can always replicate this effect by requiring projects in New York to buy steel from manufacturers in California :-).
Posted by: anonymous | January 31, 2009 at 04:21 PM
pgl: Thanks! I enjoyed reading your post.
This is one thing that pgl says which I should have said, but didn't:
"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." Agreed.
I need to think more about the comparison between the 1930's, under the gold standard, and today. One difference is fixed exchange rates in the 1930's, and that's obviously important to my argument here. But a second difference is that international reserves are fixed by the quantity of gold. Today, they are not. Feeling another post beginning to well up inside me...
anonymous @8.14:
"This is not a trade embargo. This is the US government selectively choosing the best way to increase US employment in the short-term."
It is very much like a smaller-scale trade embargo (which only applies to some not all purchases). And, if my argument above is correct, it is NOT the best way to increase US employment in the short-run. It won't work.
Posted by: Nick Rowe | January 31, 2009 at 05:36 PM
If this is a small-scale trade embargo, then so is any conscious consumer choice to buy American-made products. But of course in a free market people are supposed to be free to make decisions as to who they buy from, and factor any information they want into the decision of which product is more valuable to them, including where the product is made.
FWIW, the "buy american spending" probably represents only a few billion dollars of the stimulus _at most_. The vast majority is either tax cuts (which, if spent, will largely be spent on foreign goods), aid to state governments, and other costs not under the "buy american" restriction.
Posted by: anonymous | January 31, 2009 at 06:03 PM
Such a generous response at Misty...but I am sidelined with "unacceptable data" (but Econbrowser recognizing me...as harmless)...I'm mostly concerned about major companies, like GM, but any "multi-national" which really trades above not with, nations...anso "transnationals" really the missing element, the makers and shakers of fx and, following Kasriel, those TARP packages bein used for transnational benefits..."shareholders" if you like but don't confuse that with most of the public.
Ok does this fly at WCI or am I thoroughly grounded?
Posted by: calmo | January 31, 2009 at 06:26 PM
anonymous: yes, it's sort of the same. It is the same in its effects. There is a conflict between national/individual consumer sovereignty and "no discrimination" rules. (Note the racial/national origin of goods parallel). It worries me. But if my argument were accepted, there would be less need for those rules, because there is less incentive for national preference.
But look, the US is NOT the only guilty party here. And maybe as you say, it's not a big percentage. But symbolically, it's very provocative for international retaliation.
calmo: I got the same "data unacceptable" message when I tried to post on Economists' View just now. Must be a TypePad problem?
Posted by: Nick Rowe | February 01, 2009 at 07:25 AM
Wonks Anonymous (no relation) reiterates my point, albeit with a bit of added sarcasm.
The bottom line is 1) this particular "trade embargo" is tiny, and therefore probably won't make a big enough wave to have the effects you argue here, and 2) it's not actually a trade embargo, so as soon as we've provided these jobs to US industries those workers will be going out to Wal-Mart to buy Chinese-made TVs and to their local Toyota dealership to buy Japanese-designed Priuses. So no real reason to worry about protectionism here.
In any case, looks like it might be nixed anyway.
Posted by: anonymous | February 01, 2009 at 03:48 PM
[cross-posted from Economist's View, where I'm also having Typepad problems]
I had a feeling this was coming from Paul Krugman. But I think he could do us the favour of expanding 'buy American' to 'buy any country with a stimulus package'. See my Paul Krugman prediction.
Nick - does your analysis still work if short-run exchange rates are not determined efficiently by trade flows, but are somehow sticky or driven by expectations? Or, for example, if there is unsatisfied demand for US Treasuries which soaks up all the dollars received by exporting countries? Sorry if I'm not being clear - I will think this through and try to make it a bit more lucid. I think the Econospeak link above says the same thing better.
p.s. hasn't it been demonstrated that import tariffs, in most models, do more damage to the importing country than the exporter? This seems like another reason why the clause isn't individually rational.
p.p.s. even though the House bill only contained a 'buy America' clause for iron and steel, it seems that the Senate version is expanding it to other goods too. Good to see Obama is at least trying to do something about this.
Posted by: Leigh Caldwell | February 02, 2009 at 12:28 AM
Leigh:
Paul Krugman's model implicitly assumes totally fixed exchange rates. Mine assumes flexible exchange rates.
Regarding expectations, I made the standard assumption of static exchange rate expectations in the analysis of the capital account of the balance of payments (i.e. expected depreciation = 0). If we replace this assumption with rational expectations, we get some interesting dynamics when fiscal policy is expected to be changing.
But this is all second- or third-order stuff. The basic disagreement is on a first-order question. Paul Krugman took a model he originally used to look at fiscal policy coordination inside the Eurozone. Since the Eurozone has a common currency, it's OK to assume fixed exchange rates. But it is not OK to use that same model to look at the world, which is what he did.
Most models of tariffs assume full-employment (or at least assume that output is not constrained by aggregate demand. So they don't apply here. But even in a demand-constrained world, there are costs to interfering with trade. You produce and consume the wrong mix of goods. I took those costs as understood, because I wanted to tackle the effect of protectionism on aggregate demand.
Posted by: Nick Rowe | February 02, 2009 at 07:33 AM
This is pretty off-topic, but a question that struck me when thinking about the opposite of the situation you describe. Does this all hold in the reverse with a contractionary fiscal policy?
That is, in normal times, a contractionary fiscal policy will put downward pressure on interest rates as the demand for money decreases with lower income. Or the central bank lowers interest rates to offset the decreased demand to keep inflation on target. An decrease in domestic interest rates will cause a capital account outflow, which causes the exchange rate to depreciate. The exchange rate depreciation will cause net exports to rise. The rise in net exports offsets the contractionary 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 decreased demand from an contractionary fiscal policy will be gained due to a increase in net exports. Some or all of the reduced demand just leaks out to foreign countries.
So, in normal times, the government could run substantial surpluses to depress exchange rates, pay down debt, and suffer only a portion of the drag on the economy due to fiscal policy. Seems to me that this might be a good strategy to use once the economy recovers to prepare for the demographic time bomb we're facing with the retirement of the boomers. It also seems to be the strategy we were using in the late '90s/early oughts. Of course, we're pretty far from that situation now, running a 10% federal deficit this fiscal year.
Did I miss something?
Posted by: Andrew F | February 02, 2009 at 03:22 PM
Andrew F: It should be perfectly symmetric, in normal times. And there shouldn't be any drag at all, as long as the Bank of Canada can use monetary policy to keep AD the same (and keep inflation on target), regardless of the degree of capital mobility.
Of course, if all countries try to do the same thing at the same time, we get the closed economy version, where all the adjustment has to come from interest rates, savings and investment, and none from net exports.
Posted by: Nick Rowe | February 02, 2009 at 03:40 PM
Then this suggests if your political goal is to increase wages in the manufacturing sectors of the economy, one should not have a flexible exchange rate regime or something else must give.
Posted by: Mandos | February 04, 2009 at 10:59 AM
Mandos: sorry, but I don't follow you. This may (or may not) be relevant:
If there are asymmetric shocks (that hit countries differently), something has to give (i.e. adjust). And it adjusts either up or down depending on whther it's a positive or negative shock. Under flexible exchange rates, it's the exchange rate that adjusts. Under fixed exchange rates, it's the general level of prices and nominal wages that adjusts, but since prices and wages adjust slowly, the immediate adjustment is employment.
Posted by: Nick Rowe | February 04, 2009 at 11:19 AM
Let me turn it around. A great deal of the USA is "broken". It's "broken" because of the (relatively rapid, in terms of the history of humanity) destruction of the US manufacturing sector as revealed in the employment and wages. Do we at least agree on this?
In fact, it is so broken, some of the USA is kind of Mad-Maxian at the moment. You know, post-apocalyptic urban and rural landscapes and social-scapes? I've seen it. It even possible that at some point, guillotines can be involved. This does happen, historically. The rulers of the USA have managed to avoid it better than others, but still. I'd rather guillotines not be involved, as there's a chance (as someone who spends most of the year in the USA) that I could be on the receiving end of a guillotine.
So my wish is to increase employment and wages for those whose skillsets and upbringings and geography and commitments place them in the manufacturing (and increasingly other non-service) sectors of the US economy. Retraining them is not a solution, people are not computers on which you can just install new software, and in any case computer programming can be done in India too.
What policy tool do you recommend under present circumstances to do this?
I would choose the protectionism hammer. According to you, this wouldn't work under conditions of flexible exchange rates. Then, by implication, it suggests to me that we should not have flexible exchange rates. If by abandoning flexible exchange rates, then we would have either increased employment or increased wages, and maybe eventually de-mad-maxify increasingly apocalyptic landscapes in the USA, then I'm all for it.
But if this not a good solution, then what is? I have never seen an economist suggest an instrument that can both fly politically AND work AND not involve guillotines. The best I've seen is the protectionism hammer, which mainstream economists oppose.
Posted by: Mandos | February 04, 2009 at 11:55 AM
This might apply between the US and Canada where exchange rates are allowed to fluctuate with market forces. It might even be a good idea to modify the "buy US" provision in the stimulus package to include Canada and Mexico "buy North American".
Still I must dispute your assumption that exchange rates respond to market forces. In particular China is not a market economy but a centrally planned economy with some market features. To suppose that the exchange rate between the Dollar and the Yuan is decided by anyone but the government of the PRC is simple naivete.
It's that sort of naivete that gives economists a bad name.
Posted by: Wonks Anonymous | February 04, 2009 at 04:26 PM
Wonks: Yep, China doesn't have a free float. But it may not be fixed either. It is not obvious whether and how the government of the PRC would change the exchange rate in response to a change in US fiscal policy. We had a good discussion on this in the comments of the "fixed exchange rate" post. But only about 11% (IIRC) of US trade is with China.
Mandos: Not being American, I don't follow the US economy as closely. But I think I would support most US economists recommendations of monetary and fiscal expansion (with perhaps very aggressive non-orthodox monetary policy, coupled with a commitment to price-level path targeting). I would also place much more faith in Americans' ability to recover. They have done so many times in the past, and will do so again.
Posted by: Nick Rowe | February 04, 2009 at 07:00 PM
Recover from what, to what? I am not talking about this financial crisis now. This is merely the echo of a very long period of stagnant wages. Recovery (from that) requires restoring the ability of the American worker to bargain for higher wages.
How to do that, is my question. It's a priority.
Posted by: Mandos | February 04, 2009 at 07:49 PM
Longer run? Education, training, technology, savings and investment,: all the hard, boring things that slowly improve productivity over time.
Posted by: Nick Rowe | February 04, 2009 at 08:21 PM
I don't see how the US can bargain for higher wages until wages in China and India catch up. What proportion of manufacturing jobs (or other jobs) are really so high skill that they can only be done in the US/Europe? We've lost so much manufacturing already, and China is just starting to move up the food chain. Wait until they start exporting cars to North America. And as for high skill levels, look at the progress Huawei is making. Long-haul DWDM equipment isn't exactly low tech.
Posted by: Richard Vieregge | February 05, 2009 at 11:06 PM