And now on to something economists can really discuss

Well I really, really did not want to do this--because it is a curmudgeonly thing to do--but having wandered (without professional license) into the politics of trade policy, I cannot avoid the economics. Especially since so many of the comments around my recent posts leave the impression that the economics of trade liberalization is cut-and-dried, with only politicians and ignoramuses standing in the way. See for example the following (from Noah Yetter, via Marginal Revolution):

... peaceful exchange based on free consent benefits both parties and is therefore always good.... true FREE TRADE is the only correct policy.

So here is a straightforward economics question: under what conditions will trade liberalization enhance economic performance?

If you answered "under any and all," you flunk.  Here is the correct answer (adapted from here):

• The liberalization must be complete or else the reduction in import restrictions must take into account the potentially quite complicated structure of substitutability and complementarity across restricted commodities.
• There must be no externalities or microeconomic market imperfections other than the trade restrictions in question, or if there are some, the second-best interactions that are entailed must not be adverse.
• There must not be any increasing returns to scale, or else activities with scale economies must expand "on average."
• The home economy must be “small” in world markets, or else the liberalization must not put the economy on the wrong side of the “optimum tariff.”
• The economy must be in reasonably full employment, or if not, the monetary and fiscal authorities must have effective tools of demand management at their disposal.
• The income-redistributive effects of the liberalization should not be judged undesirable by society at large, or if they are, there must be compensatory tax-transfer schemes with low enough excess burden.
• There must be no adverse effects on the fiscal balance, or if there are, there must be alternative and expedient ways of making up for the lost fiscal revenues.
• The economy must not have a trade deficit that is already "too large," or else nominal wages or the exchange rate must adjust to compensate.
• The liberalization must be politically sustainable and hence credible so that economic agents do not fear or anticipate a reversal.

I could expand the list, but you get the point. And all of this is needed just to ensure static benefits. If you want dynamic (growth) benefits, we would have to add an even larger number of other prerequisites. (And just to be absolutely clear, the list above is no argument in favor of trade restrictions either.)

The point is that unconditional supporters of free trade take a whole lot for granted. Our professional training prepares us to be analysts who can make contingent statements.  Policy A is good if conditions X, Y, and Z are in place. Rule-of-thumb economists sweep all the caveats under the rug, and in the end, are not true to their training.

(I can see the next line of objections coming: Forget theory, some people will say. Look at the real world. Countries that follow open trade policies do so much better than those that don't.  Well, not so fast actually. But that has to wait for a later post.)   

Comments

[Charles]

Dear Professor Rodrik,

I am writing to commend you on an excellent blog. Your clear reasoning and balance have reinvigorated my interest in trade policy.

Although I have read some of your research papers previously, I have learnt more in a couple of weeks from your blog than perhaps everything after my first year of economics. Having had close contact with the trade research conducted in South Africa, your posts are particularly enlightening. I shall be sure to read your CID paper on the topic.

As enter the world of professional economics and thereafter, graduate school, I feel sure that trade shall now be one of my chosen specializations.

Regards,

Charles

[Dan]

I think you're making the wrong argument here; what matters is not that the circumstances under which trade "improves economic performance" (a vague enough phrase to raise some alarms), but whether trade actually does. It is, as Paul Samuelson has noted in many places, including his textbook, a useless banality to say "If production functions demonstrate constant returns to scale, if preferences are identical and homothetic, if there are no externalities, and if there is only one input in production, then trade is beneficial." That's just a tautalogy, true because it has to be. It is far more insightful to say, "trade is beneficial because it allows countries to specialize." Now I am quite convinced by your reserach that, in general, this need not be true; but that should be the end of the story, without reminding everyone that of all the various untested hypotheses that go into a theory.

[alex]

The economy must not have a trade deficit that is already "too large," or else nominal wages or the exchange rate must adjust to compensate.

There's one that the US fails big time. I've lost track of whether our CAD is closer to 6% or 7%/GDP these days. Let's fix that before we have any further trade liberalization. The "next trade agreement will help the CAD" line is getting to be like Lucy promising to hold the football while Charlie Brown kicks it.

In general I find your emphasis on stating the usually unstated assumptions and in looking at real-world evidence a refreshing change from the "free trade good", "free trade bad" level of discussion. But, no, I'm not going to buy the t-shirt.

[Ivo Staub]

The question what the consequences of trade are, is an empirical and not a theoretical one (there i said it).
A lot of statements by politicians and economists seem to be 'faith based'. Those who believe in Ricardo say 'free trade is great for everyone' and the others say it's evil. And both positions are justifiable with some economic theory (A few days ago (4/29) one of my favorite economists wrote: 'For almost any particular conclusion you want to arrive at, there is some economic model that will take you there').
The only way to find out what consequences trade has on growth, convergence, inequality and poverty is to do empirical research and to look what the data says. I'm no trade specialist but I think in most cases the answer is that trade has little to none positive or negative influence (e.g. your 'Institutions Rule'-paper). There are quite successful trade systems (like the European Union) and there are countries that had bad experiences with free trade, what I don't know is what exactly are the differences between successful and unsuccessful cases. Is it the products a country produces, the human capital, the specific institutions of the trade regime has (e.g. the structural adjustment program in the E.U.) etc.?

I like your blog a lot. Thank you for spending your much time on it.

(Sorry for the bad English. I'm not a native speaker).

[robert mcconnell]

It seems to me that one unquantified, but very important, effect of trade, free or otherwise, is the cost of invasive species carried in the ballast water of ships. This costs the U.S. billions of dollars a year in mitigation costs, and at present there is no mechanism to address this externality.
Naturally ships from U.S. ports have similar impacts worldwide. Comments?

[Gabriel M.]

Professor,

You're clearly not on the same page with the commenter from MR you contradict.

It's clear that you're operating with radically different visions of what's desirable/good.

Pareto efficiency is not a magical criterion beyond criticism, and I'm sad to report that it's politics.

Trying to pass a political criterion as apolitical, as "the economics [of it]", is somewhat fishy.

[conchis]

Gabriel,

I think your criticism falls flat here. Whether or not he's on the same page as the MR commenter (personally I'm not sure I'd want to be), Dani's pretty much just adopting the same standard as most of the non-psycho-libertarian) free trade crowd. The point isn't to elevate a contestible notion of the good as apolitical. It's to show that, without all the assumptions he sets out, free trade isn't guaranteed to meet the very standard it's proponents usually hold it too.

[Gabriel M.]

The cited MR commenter explicitly stated his standard. And free trade meets it.

The standard supported by the commenter is orthogonal to the "economic performance" standard employed by the author.

Therefore, this is a case of "my values are better than your values", a pseudo-rebuttal if you will.

[Jim Johnson]

Thank you! It is so very tiresome to listen to libertarians and neo-liberals and conservatives of various stripes talk about markets as though they are a panacea and operate attractively absent a (as you make clear) long and quite stringent set of initial conditions. That is why talking about the economics of trade liberalization (or nearly any other economic policy) without talking about the politics (hence political-economy) is simply silly. Markets may well be the best mechanism to use to coordinate some domain, but the decision to rely on them as opposed toone of a large number of competing institutional mechanisms is a prior political decision because (especially absent the requsite conditions) they are not always best.

[jdrietz]

Professor Rodrik, your intent, I assume, is to justify some level of trade restrictions when these conditions do not exist.

We can find scenarios in which some of these conditions are not present between states or regions within the U.S., or even between large metro areas. Do you believe it would be beneficial for governments at these levels to enact protectionist trade policies if your conditions are not met?

[jdrietz]

To clarify my first statement: while you state that these conditions are not an argument against free trade, you do state that they must exist for free trade to result in economic benefits. Therefore, the logical conclusion is that if one or more of these conditions does not exist, trade restrictions would at a minimum be justifiable from an economic perspective.

[Laurent GUERBY]

"The liberalization must be complete or else ..."

Well no so-called "free" trade agreement involving the USA has been without so-called "intellectual property" strenghtening measures.

An "intellectual property" is of course one of the worst protectionnist measure currently in existence. It distorts prices many more times than usual tariffs, not talking about behaviour of agents.

It must go. Immense gains for the world, including millions saved from death. More innovation and less broken windows in lawyers fees and government thought police.

But I don't see economists making the connection (excepted Dean Baker), why?

Any idea?

Why is there no debate? Is it all fake "free"?

References:

http://www.prospect.org/deanbaker/2006/11/trade_deals_the_washington_pos.html

http://www.prospect.org/deanbaker/2006/11/lets_get_the_story_right_the_c.html

http://www.dklevine.com/general/intellectual/against.htm

[itdepends]

The Professor is of course completely correct on his Economics -- and it's useful to see the theory of the second best summarized succinctly like this.
The concept of 'economic performance' needs to be evaluated with some care however. I presume Rodrik chose this 'performance' term loosely to stand in to mean something like 'welfare as judged by many social welfare criterion functions.'
I would point out however that implicitly the language of the post makes it sound like he (or readers) may really be talking about conditions where 'national economic performance (or 'welfare') may not be enhanced by partial trade liberalization or unfettered free trade.

I think it's fair to say however that there are many situations where national welfare can decline even as 'world welfare' (as defined by a spectrum of reasonable social weighting functions) could well increase. For instance US apparel workers clearly are harmed by more open trade and there may even be situations where "US economic performance" (variously defined) may be harmed, but once you add in the gains to Chinese workers, the final welfare balance may appear positive as judged by a similar welfare criterion which includes Chinese workers is added in.

[Michael Greinecker]

I want to comment only on the issues that can really be recast in the language of equilibrium theory:

"The liberalization must be complete or else the reduction in import restrictions must take into account the potentially quite complicated structure of substitutability and complementarity across restricted commodities."

Any opening of trade, with some ex ante lump some transfers will be at least as Pareto good as autarky in a standard model.

"There must not be any increasing returns to scale, or else activities with scale economies must expand "on average."

This is an issue for questions of existence of certain equilibria, but increasing returns to scale can be compatible with free trade being beneficial.

The argument underlying my position can be found here:

http://yetanothersheep.blogspot.com/2007/05/trade-efficiency-and-redistribution.html

It is basically based on a paper Grandmont and McFadden wrote in 1972.

[Joel P. Trachtman]

I enjoy the blog! But I agree with the tenor of Dan's comment. I would put it in terms of your language. You state a number of conditions as necessary (using the word "must"). They are only necessary to a conclusion that trade necessarily increases welfare. None of them is necessary to a conclusion that trade *may* increase welfare. In the world of the second best, policy-makers still have to make decisions.

[alex]

Michael Greinecker's assumption: some ex ante lump some [sic] transfers

More realistic economist: assume we have a can opener.

[Michael Greinecker]

"Michael Greinecker's assumption: some ex ante lump some [sic] transfers

More realistic economist: assume we have a can opener."

I didn't say it was realistic. But lump sum transfers underlie the reasoning that is usually used to justify free trade in economics.