October 1999
WAS IT ALL IN OHLIN?
Let me begin with an embarrassing admission: until I began working onthis paper, I had never actually read Ohlin'sInterregional and InternationalTrade. I suppose that my case was not that unusual: modern economists,trained to think in terms of crisp formal models, typically have littlepatience with the sprawling verbal expositions of a more leisurely epoch.To the extent that we care about intellectual history at all, we tend torely on translators - on transitional figures like Paul Samuelson, whoextracted models from the literary efforts of their predecessors. And letme also admit that reading Ohlin in the original is still not much fun:the MIT-trained economist in me keeps fidgeting impatiently, wonderingwhen he will get to the point - that is, to the kernel of insight thatended up being grist for the mills of later modelers.
Moreover, one can argue that Ohlin actually gains something in the translation:Samuelson famously found implications in Ohlin's own view of trade thatthe great thinker himself, due to his "diplomatic style" (in Tjalling Koopman'sphrase), had missed. Ohlin seemed to say that while trade shifts the distributionof income against scarce factors, it nonetheless probably improves theirlot in absolute terms; Stolper and Samuelson showed that in a simple modelthe stark fact is that scarce factors lose by any measure. Ohlin definitelyviewed factor-price equalization as only a tendency, surely incomplete;Samuelson showed that under the assumptions of Ohlin's Part I, "Interregionaltrade simplified", it was quite possible that trade would in fact fullyequalize factor prices. So just as a modern student of evolution mightbe forgiven for preferring to get his Darwin courtesy of John Maynard Smith,a modern economics student might be forgiven for preferring to get hisOhlin via Samuelson, and indeed via Krugman-Obstfeld.
And yet what Ohlin disparagingly called "model mania" can lead to anarrowing of vision. Samuelson himself entitled his 1971 article expoundingwhat has since come to be known as the specific-factors model "Ohlin wasright", conceding that in a multi-factor model some of Ohlin's skepticismabout the full factor-price equalization and strong Stolper-Samuelson effectsthat arise in a two-by-two model turns out to be justified after all. Whatelse might Ohlin have been right about?
Some years back I gave a short series of lectures (the Ohlin lectures,as it happens, written up in my bookDevelopment, Geography, and EconomicTheory ) on the way that a growing emphasis on formal modeling ledeconomists to "forget" insights about the role of increasing returns inindustrialization and economic location, only to rediscover those insightswhen modeling techniques became sufficiently advanced. Was the same truein international trade theory? In particular, did Ohlin's informal expositionof a theory of interregional and international trade contain the essenceof what later came to be known as the "new trade theory" and the "new economicgeography"?
The answer, it turns out, is yes and no. Ohlin did indeed have a viewof international trade that not only gave a surprisingly important roleto increasing returns (surprising because in Samuelsonian translation thatrole disappeared), but also one that suggested a sort of "unified fieldtheory" of factor-based and scale-based trade that is a clear antecedentof the "integrated economy" approach that ended up playing a central rolein post-1980 trade theory. On the other hand, despite Ohlin's title andhis repeated suggestion that he was offering a unification of trade andlocation theory, there is little inInterregional and InternationalTrade that seems to point the way to the distinctive features of "neweconomic geography". And there were a number of insights in modern tradetheory that Ohlin did not, as far as I can tell, anticipate at all.
But let me start with my startling discovery: the extent to which Ohlinin the original anticipates a view of trade that the "new trade" theoristshad to rediscover some 50 years later.
1.Increasing returns as a cause of trade
What did international economists know and think about increasing returnsin trade circa, say, 1975? Certainly they were aware of the issue: R.C.O.Matthews' 1950 integration of external economies and offer-curve analysisshowed up as a supplemental reading in graduate courses, as did later paperssuch as Chacoliades (1970). But I think my description in an essay of afew years ago (Krugman 1996) still captures pretty accurately the stateof general understanding:
"The observation that increasing returns could be a reason for tradebetween seemingly similar countries was by no means a well-understood proposition:certainly it was never covered in most textbooks or courses, undergraduateor graduate. The idea that trade might reflect an overlay of increasing-returnsspecialization on comparative advantage was not there at all: instead,the ruling idea was that increasing returns would simply alter the patternof comparative advantage. Indeed, as late as 1984 many trade theoristsstill regarded the main possible contribution of scale economies to thestory as being a tendency for large countries to export scale-sensitivegoods. The essential arbitrariness of scale-economy specialization, itsdependence on history and accident, was hardly ever mentioned. To the extentthat welfare analysis was carried out, it focused on the concern that smallcountries might lose out because of their scale disadvantages.
1. Ohlin also stated, more or less clearly, the ideathat if the increasing returns industry is capital-intensive the countrythat gets it will shift to a labor-intensive mix of constant-returns goods,and conversely. True, he argues in terms of factor prices, missing theSamuelsonian point about factor price equalization. But it is still amazinglyclose to what I regarded as a fresh discovery when I finally worked itout.
2. Richard Feynman once told a reporter who wanteda 5-minute summary of his work that if it could have been summarized in5 minutes, it wouldn't have warranted a Nobel. Yet the Big Ideas in tradetheory probablycan be summarized in not much more than 5 minutes.
