Kieran Healy links to a paper by Pierre-Olivier Gourinchas and Olivier Jeanne in which a calibrated growth accounting model is used to show that the gains from unrestricted capital mobility are likely to be of the order of 1 per cent of GDP. Gains from risk sharing aren’t mentioned but other papers are cited to say that these are of a similar magnitude.
Those who listen to the general pronouncements of economists might be surprised by the modest size of the estimated gains. But for those who have looked at similar exercises in the past there is no surprise here. One of the better-kept secrets of economics is the fact that most studies suggest that the replacement of a typical high-tariff regime (say Australia’s in the 1960s) will yield long run benefits of about 3 per cent of GDP.
Those who raise questions about this point are likely to be brushed off with a reference to supposed dynamic gains, not captured in this ‘static’ analysis. This brings us to an even better-kept secret. These ‘dynamic gains’ have about as much basis in neoclassical economic theory as the Tooth Fairy.
To complicate matters a bit further, there is a theoretically respectable category of dynamic gains, arising from the removal of distortions in intertemporal resource allocation, but these are even more modest than the static gains. In fact, the gains looked at by Gourinchas and Jeanne.
The last line of defence is the idea of X-efficiency, or the ‘cold shower’ effect of competition. As Chicago stalwart George Stigler was the first to point out, this idea is based on the fallacious assumption that additional work effort is costless. This fallacy is hard to kill, but anybody who’s experienced 1990s-style ‘workplace reform’ knows it for what it is. I’ve been hammering away on this point for at least a decade, for example here and here (PDF), but with very little impact.
John Quiggin 
Is that 3% annualised?
I believe the 3% is the cumulative long-run impact. The level of GDP is raised by 3% in the long run. It may come about by growth being raised by 1% (OK, 0.9902%) in each of the first 3 years or by 0.3% in each of the first 10 years with no impact thereafter.
Are there no respectable models in which trade can raise the long-run growth *rate*? Some kind of thing with an endogenous pace of technology improvement that is increased by trade somehow?
I know nothing about this, so please excuse any stupidity in the question.
A very sensible question, Daniel. An endogenous growth model ought to yield a growth rate effect for free trade vs autarky. On the other hand, it’s unlikely to show free trade as the optimal policy – it will almost certainly be dominated by policies that subsidise the accumulation of knowledge capital.
John
Paul Romer wrote a paper a few years ago in which he showed that the gains from trade should be measured as rectanges not triangles. (Apologies to those who have no idea what I’m talking about.) I think these were dynamic gains, and they were big (in theory), much bigger than the triangular removal of distortions.
Romer is no carpet bagger for an ideology. Whether this is neoclassical economics or not is entirely semantic.
A fairly simple bit of modelling shows that we should only expect very small harm from tariffs at low levels, which means conversely only small gains going from what we had to free trade.
Consider not tariffs but negative tariffs, or subsidies: these will also cause harm, essentially by moving things away from a optimum (including costs of funding them, just as we should also allow for beneficial tax cuts or revenue gains elsewhere when looking at the overall effects of tariffs). So we must have some non-linear function to describe levels of harm in terms of the size of the distorting feature, maybe a square law, but anyway something with a broad plateau of near optimal results. I omit the case of a narrow spike shaped function, as laissez faire could never stay on the “sweet spot” of an optimum anyway – free trade gains would require positive management to stay on the optimum and ipso facto that wouldn’t be free trade.
But all that means that small levels of tariffs, say for revenue purposes, will not cause proportionate levels of harm but much less.
Is that taken from the standpoint of a national economy or the global economy?
Japan benefited from not having free trade but it hasn’t helped them in the long run, mainly through non-economic factors.
Joe, can you give me the Romer reference. I certainly don’t mean to define neoclassical in the kind of way that would exclude Romer. Broadly speaking, anyone who talks triangles and rectangles fits my definition.
Clearly, I overstated my position a bit. Papers giving large estimates of dynamic trade gains within neoclassical economics are more common than papers purporting to prove the existence of the Tooth Fairy. But no such paper has gained widespread acceptance and the standard neoclassical model still shows all the gains in small triangles.
Also, as I observed in an earlier comment, any model giving large gains from free trade relative to autarky will almost certainly give larger gains from some other policy.
In case readers hadn’t spotted it, Brad de Long has picked up on this blog entry.
One point he makes is that a few percent is not to be sneezed at. I have commented there that it is, given that it is so easily swamped by other small cumulative effects that come from allowing open foreign capital inflows. I forgot to spell out that these are only harmful when there is no matching real investment, but that we already allowed for that. It would be double counting to claim that the benefits of those investments should be added as the few percent gains already contain any gains from outside investment, and the decline of Australian ownership already factors in any acquisitions for nominal investment done with US fiat dollars (their export of inflation via middlemen who acquire our revenue streams).
“any model giving large gains from free trade relative to autarky will almost certainly give larger gains from some other policy”
Yes, but do these other policies provide a stable political equilibrium? If not, they are theoretically interesting but in practice irrelevant because they can’t be arrived at and maintained.
One of the free traders’ arguments is that free trade is a more stable position because government creation of rents leads to self-funding lobbies for other rents. In other words, once you start handing out favours, even ones where the allocative costs are small, you will be pressured to keep handing out new ones – as the Bush administration has discovered. These new favours may cost much more than the initial ones, so you should be very careful about creating rents in the first place. In Australia the superannuation (pension) industry’s push for 15% compulsory contributions is a contemporary example. Tariff policy under Black Jack McEwen is an historic example.
Whoops, sorry for the problem with bolding above.
DD, I think the Bush admin example illustrates the weakness of the free trade claim. The problem of political irresponsibility is essentially orthogonal to the policy regime. Free-trade governments can be thoroughly corrupt and silly, so can protectionists.
“The problem of political irresponsibility is essentially orthogonal to the policy regime.”
But my point denies just this. You don’t have to be a paid-up member of the Virginia school to understand that the more government dabbles the more vested interests will try to influence that dabbling – especially where the initial dabble has actually created the vested interest de novo (see the examples I gave). In the absence of large direct welfare gains by a particular dabble the government should therefore avoid future ‘political irresponsibility’ by current restraint.
It’s the same argument small government people put elsewhere – they believe small government is clean government. Though I believe that in some of these other fields there is the prospect of large direct welfare gains to offset the risk.
DD, you’ve responded by assertion rather than evidence. Where is your evidence of correlation between avoidance of systematic intervention (e.g. capital controls) and avoidance of pandering to vested interests. Would you say that vested interests have less influence over the Howard government than over the Menzies government for example.