Ian Fletcher: Natural Strategic Tariff

TradeReform.org ^ | Friday, 10 October 2008 | Stumo

[rmlew]

Ian Fletcher is an economics consultant in private practice in San Francisco.  He was previously head of government relations for the American Engineering Association.  Fletcher outlines an analysis of a Natural Strategic Tariff as a corollary to the Gomory-Baumol multiple equilibrium model of international trade.

Yes.  Yes.  I know.  Econo-speak. 

Gomory-Baumol probably have replaced, on the merits, Ricardo's more simplistic ideas about trade.  Of course, Ricardo is cited by the pseudo-academics that like to avoid thought.

The Natural Strategic Tariff is a controversial idea that would certainly be shouted down with "protectionist" accusations.  But take a look and feel free to write comments.  And Ian Fletcher welcomes comments to him directly.

The Natural Strategic Tariff

Ian Fletcher

 
Abstract:  The Gomory-Baumol (2000) multiple-equilibrium model of international trade has overthrown the neoclassical Ricardian contention that global free trade will produce optimal outcomes for all participating national economies and the world.  But it leaves unanswered what trading regime would be more advantageous, for any given nation state, than free trade. Most previous solutions have indicated that only a sophisticated industrial policy can do this, making repudiation of free trade vulnerable to the known political difficulties of achieving such a policy.  This paper argues that, in the case of a large advanced economy, a flat tariff will have a naturally strategic effect, due its different effects on different industries. It will favor the “retainable industries” the Gomory-Baumol model indicates are the key to exceptional national economic performance.  It will tend to more strongly protect, and thus drive the economy towards, industries where capital investment, not direct labor, constitutes the largest part of production costs.  It will tend to more strongly protect industries that are on the early part of concave declining cost curves, thereby naturally favoring industries which are retainable, due to their increasing economies of scale, but not yet retained by another national economy, and thus stil available.

 
In trade theory, one major recent theoretical breakthrough has been that of Ralph Gomory and William Baumol (hereinafter GB), who have propounded a multiple-equilibrium model of world trade in their book Global Trade and Conflicting National Interests.
 
The GB model holds that if one assumes increasing economies of scale, then the distribution of industries among nations under free trade will exhibit not one equilibrium, as in the classic Ricardian model, but multiple equilibria.  Each such equilibrium will be locally optimal (more efficient than any similar equilibrium) but may or may not be globally optimal (more efficient than any possible equilibrium).  Which equilibrium the world economy settles on will be an historical accident, driven by such things as which nation entered which industry first. Therefore, any actually existing equilibrium can be sub-optimal, both from the point-of-view of maximizing world output, and the output of any given nation.
 
In the GB model, the outcomes that maximize the output of any given nation are those in which that nation has captured a disproportionate share of what they call the world economy’s “retainable industries.”  These are industries that, once established in a nation, are susceptible to becoming entrenched and difficult for foreign competitors to dislodge, even under free-market conditions.  They become entrenched due to their increasing economies of scale, which give an established high-volume competitor, simply because it is high-volume, an innate advantage over challengers trying to break into the industry from scratch, which by definition begins at a volume of zero. Such industries, because they are sheltered from the full blast of international competition, can earn exceptional profits and pay exceptional wages.

 
The natural question, if one assumes self-interested nation states, is what kind of trade policy will acquire any given nation these retainable industries.  Unfortunately, there is no easy answer here, because in order to win these industries, the state must, naturally, know which industries they are – or it will be incapable of directing its tariffs, subsidies, and other policy measures to their appropriate targets. (And this is not even considering the fraught problem of how to effectively favor the right industries, once they have been identified.) 
 
While there is some evidence that fabled institutions like Japan’s Ministry of International Trade and Industry (MITI) have successfully done this in the past, there are also vast empirical and theoretical grounds for supposing that this is extremely difficult to do, indeed probably beyond the ability of most contemporary governments. Because the cost of trying and failing is potentially very high, this is (not entirely unreasonably) taken as a dispositive argument against protectionist measures aimed at winning retainable industries.
 
There is, however, one possible loophole in the above problem.  The above reasoning presumes that the state must know which industries are potentially retainable. That is, it presumes that the only possible means to impose strategic protectionism is central planning based upon ex ante industry-specific knowledge.  But what if this were not so?  What if, instead, it were somehow possible to have a effective strategic tariff without such hard-to-obtain knowledge?  What if, that is, there existed some simple rule for tariff policy which, when applied to the complex empirical conditions of the economy, had the desired complex effect?  If such a simple rule exists, we should call it a “natural strategic tariff.”  It is my purpose in the remainder of this paper to vindicate the proposition that such a thing does exist.

 
The formula for the natural strategic tariff (hereinafter NST) I shall examine is simplicity itself:
 
A flat tax on all imported goods and services, without exception or variation.
 
Prima facie, such a tariff is, of course, strategically meaningless, because it protects, and thus promotes domestic production in, all industries equally.   Whatever merits it may have with respect to other critiques of free trade, with respect to the GB model, it is useless.
 
Unless, of course, the prima facie view of a flat tariff is mistaken.  I believe it is.  The key intuition underlying the NST is this:

 

A flat tariff, on a bumpy economy, isn’t flat in its effects.

 
The key here, is that a flat tariff affects different industries differently.
 
Before analyzing why this NST  is naturally strategic, it is important to remind ourselves what a strategic tariff is, under GB assumptions. (Remember we are just asking here, at this point in the logic of the argument,  for the definition of the thing, and are strictly speaking agnostic, at this point, whether an NST exists, or could be successfully imposed.) By definition, an effective strategic tariff, under GB assumptions, is any tariff that produces an assignment of industries among nations that is more advantageous, to the nation imposing it, than the assignment that would otherwise have occurred.  
 

GB assumptions indicate that this will happen if the tariff causes the nation to garner more “retainable industries” than it otherwise would have (up to a point, after which the nation is “too greedy” and would be better off shedding some industries to its trading partners.)   These industries, as previously noted, are characterized by increasing economies of scale.
 
It follows from the above that a tariff will be strategically effective if it pushes the nation’s economy into industries which will turn out to be retainable.  Now we know what characterizes such industries (increasing economies of scale), and where increasing economies of scale come from (high fixed capitalization.)  Therefore, a tariff that pushes the nation’s economy into industries with high fixed capitalization will tend to increase its capture of retainable industries.

 
Note here that “high fixed capitalization” does not only mean physical plant paid for by the owners of the industry.  It also includes investments in human capital made by the industry’s employees, investments in infrastructure made by the government, and difficult-to-monetize investments like the accumulation of productive knowledge through learning-by-doing. It means anything that cashes out in a declining cost curve.

Also note that a beneficial tariff here will be beneficial on net.  It will obviously have some countervailing negative effects, which are well-known in the case of tariffs generally. Harberger losses to consumers, and other problems, will not vanish, they will merely be outweighed if the tariff is strategically effective.
 
The argument below is framed with the United States as the hypothetical test case. Therefore, parts of it depend upon empirical facts which may not be true of other nations.  For them, there may be no natural strategic tariff, or the natural strategic tariff may be different than that described below.

 
Within GB assumptions, there are not one but two reasons why the NST under discussion is naturally strategic in the American case:
 
Different Cost-Competitiveness of Different Industries
 
The United States is more cost-competitive, or closer to being cost-competitive when it is not, in highly-capitalized industries.  As a result, a flat tariff will have a greater impact on these industries than on others.  Crudely put, a 30% tariff would not cause the relocation of the clothing industry into the US from abroad, because the difference between American and foreign labor costs is too large for a 30% premium to tip the balance in America’s favor.  But it would be likely to cause the relocation of high-technology manufacturing like flat-panel displays, where direct labor constitutes only a small percentage of production cost.
 

These capital-intensive industries are precisely the ones that the GB model indicates it is advantageous to capture, because they are likely to exhibit the increasing economies of scale that cause retainability and thus excess returns.  This fact makes a flat tariff natural strategic.
 
Note that a flat tariff would not have this effect in every nation that could conceivably impose it.  If Costa Rica, for example, imposed one, it would not push the Costa Rican economy towards high-tech manufacturing, because Costa Rica, unlike the US, is not closer to being cost-competitive in high-tech industries than in other industries.  Therefore, there would be no differential impact in favor of such industries.  The NST thus only has a beneficial effect by means of its interaction with the existing competitive strengths of the national economy in question.   In effect, the NST privatizes the key decisions concerning which industries and companies to favor, avoiding the need for analytically puzzling and politically tricky choices.
 
Different Tariff Effect on Different Parts of the Declining Cost Curve

 
The second reason that a flat tariff is naturally strategic, is more subtle than the first.  It concerns the difference in effect of a tariff impacting industries that are at different points on their declining cost curves.
 
Not all cost curves decline, of course, or have the shape shown below.  But crucially, the cost curves of the retainable industries of the GB model do tend to have this shape.  These cost curves tend to be concave, like the example illustrated below, because of the mathematics of how they are generated.  (A concave average-cost curve is guaranteed if we assume that the production cost consists of an investment in capital plant plus an incremental cost for each unit of production.)  As a result, these curves tend to have steeper slopes in their earlier than later stages.
 

 
 
It is this difference in slope that is the key.  When a tariff is imposed upon an industry that is in the early stages of its declining cost curve, costs fall rapidly with relatively small increases in output.  Therefore, when the domestic industry is given tariff protection, and its sales increase as a result, it will enjoy a large cost decline.  See the graph below:
 

 
This induced cost decline will, naturally, improve the industry’s cost advantage over its foreign competitors even more.  This will result in further sales increases at foreign expense, further cost declines, and so on.   A virtuous cycle is created, so a fairly small percentage tariff can trigger a much larger cost advantage for the domestic producer, which will outlast tariff protection.  In the extreme case, this virtuous cycle only ends when the domestic industry has wiped out its foreign competitor and become globally dominant.

 
On the other hand, when a tariff is imposed on a mature industry, i.e. one that is in the late stages of its declining cost curve, the slope of the average-cost curve is relatively flat, so fairly large increases in sales will not shift the producer’s costs by very much.  Therefore, tariff protection will not trigger the aggressive “virtuous cycle” described above, and the impact of the tariff will be mild.
 
The net effect, of this combination of aggressive and mild impacts, is that a flat tariff will have a bias towards stimulating industries that a) have declining cost curves, and b) are on the early part of those curves.  These are precisely the industries that the GB model implies a nation should wish to stimulate, because they are both susceptible to retainability (because they exhibit increasing economies of scale) and not yet retained by some other nation (because they are still in the early part of their declining cost curves) and thus out-of-reach.
 
This all resembles, of course, the classic “infant industries” case for tariffs – without, of course, the contentious problem of deciding which industries qualify as infant, or for how long.   (A classic Hamilton-List-MITI “infant industries” tariff program would have the same effect as the flat tariff described here, only more strongly, as it would not “waste” any tariff incentive on industries that were on the later part of their cost curves.)

 
Does this imply America should impose a tariff?
 
Yes and no. Unfortunately for opponents of free trade, the above reasoning does not, on its own, prove that the US would be better off with a tariff.  Although the above logic does explain how such a tariff would bring a benefit, it does nothing to quantify this benefit against the well-known expected costs.   Such quantification would require a major econometric study, and because the GB model used above is not the only valid critique of free trade, this study would have to incorporate these other theoretical models as well.  This essay should only be taken as a contribution to the underlying theory that makes such studies possible.

[rmlew]

To often, free trade partisans will slander their opponents as economically illiterate. They will misquote Smith, or quote Ricardo and Hayek as doctrine. It doesn't fly. Given our crippling balance of trade deficit and loss of economic wealth, it is time to address this honestly.

[rmlew]

trade ping

[steve8714]

Economic models don’t predict future outcomes. At all. Ever.

[steve8714]

You never know the elasticity or the ranges involved, so it never makes much sense to pursue policies based on this kind of analysis.
Only weather and climate rival economics when it comes to incorrect forecasting.
Joseph P. McKenna taught me that.

[dennisw]

Donald Trump is the best chance for tariffs and to bring us back to balance and sanity on trade.