American Political Science Review Vol.81 in countries with different factor endow- costs,for example,is indistinguishable in ments.Moreover,I shall suggest that its impact from an across-the-board these implications conform surprisingly decrease in every affected state's tariffs well with what has been observed about (Mundell 1957,330);so is any change in patterns of cleavage and about changes in the international regime that decreases the those patterns in a great variety of coun- risks or the transaction costs of trade.The tries during four periods of global change converse is of course equally true:when a in exposure to trade,namely the "long" nation's external transport becomes sixteenth century,the nineteenth century, dearer,or its trade less secure,it is the Depression of the 1930s,and the years affected exactly as if it had imposed a since World War IⅡ. higher tariff. Nonetheless,what I present here The point is of more than academic remains conjectural and preliminary.The interest because we know,historically, evidence I shall be able to advance is sug- that major changes in the risks and costs gestive rather than conclusive.It is prin- of international trade have occurred: cipally the clarity of the logical case that notoriously,the railroads and steamships seems to me to justify further refinement of the nineteenth century brought dras- and testing. tically cheaper transportation (Landes 1969,153-54,196,201-2;Hobsbawm The Stolper-Samuelson Theorem 1979,Chap.3);so,in our own genera- tion,did supertankers,cheap oil,and In 1941 Wolfgang Stolper and Paul containerization (Rosecrance 1986,142). Samuelson solved conclusively the old According to the familiar argument of riddle of gains and losses from protection Kindleberger (1973)and others,inter- (or,for that matter,from free trade). national hegemony decreases both the They showed that in any society protec- risks and the transaction costs of inter- tion benefits-and liberalization of trade national trade;and the decline of hege- harms-owners of factors in which that monic power makes trade more expen- society is poorly endowed,relative to the sive,perhaps-as,according to this inter- rest of the world,as well as producers pretation,in the 1930s-prohibitively so. who use the scarce factors intensively.1 Analyzing a much earlier period,the Conversely,protection harms-and lib- Belgian historian Henri Pirenne (1939) eralization benefits-owners of factors attributed much of the final decline of the the given society holds abundantly rela- Roman Empire to the growing insecurity tive to the rest of the world,and pro- of interregional,and especially of Medi- ducers who use the abundant factors terranean,trade after 600 A.D.3 intensively.2 Thus,in a society rich in Global changes of these kinds,it labor but poor in capital,protection follows,should have had global conse- would benefit capital and harm labor;and quences.The "transportation revolu- liberalization of trade would benefit labor tions"of the sixteenth,the nineteenth, and harm capital. and scarcely less of the mid-twentieth cen- So far,the theorem is what it is usually tury must have benefited,in each affected perceived to be:merely a statement,if an country,owners and intensive employers important and sweeping one,about the of locally abundant factors and must have effects of tariff policy.The picture is harmed owners and intensive employers altered,however,when one realizes that of locally scarce factors.The events of the exogenous changes can have exactly the 1930s should have had exactly the oppo- same effects as increases or decreases in site effect.What,however,will have been protection.A cheapening of transport the political consequences of those shifts 1122American Political Science Review Vol. 81 in countries with different factor endow￾ments. Moreover, I shall suggest that these implications conform surprisingly well with what has been observed about patterns of cleavage and about changes in those patterns in a great variety of coun￾tries during four periods of global change in exposure to trade, namely the "long" sixteenth century, the nineteenth century, the Depression of the 1930s, and the years since World War 11. Nonetheless, what I present here remains conjectural and preliminary. The evidence I shall be able to advance is sug￾gestive rather than conclusive. It is prin￾cipally the clarity of the logical case that seems to me to justify further refinement and testing. The Stolper-Samuelson Theorem In 1941 Wolfgang Stolper and Paul Samuelson solved conclusively the old riddle of gains and losses from protection (or, for that matter, from free trade). They showed that in any society protec￾tion benefits-and liberalization of trade harms-owners of factors in which that society is poorly endowed, relative to the rest of the world, as well as producers who use the scarce factors intensively.1 Conversely, protection harms-and lib￾eralization benefits-owners of factors the given society holds abundantly rela￾tive to the rest of the world, and pro￾ducers who use the abundant factors intensive1y.l Thus, in a society rich in labor but poor in capital, protection would benefit capital and harm labor; and liberalization of trade would benefit labor and harm capital. So far, the theorem is what it is usually perceived to be: merely a statement, if an important and sweeping one, about the effects of tariff policy. The picture is altered, however, when one realizes that exogenous changes can have exactly the same effects as increases or decreases in protection. A cheapening of transport costs, for example, is indistinguishable in its impact from an across-the-board decrease in every affected state's tariffs (Mundell 1957, 330); so is any change in the international regime that decreases the risks or the transaction costs of trade. The converse is of course equally true: when a nation's external transport becomes dearer, or its trade less secure, it is affected exactly as if it had imposed a higher tariff. The point is of more than academic interest because we know, historically, that major changes in the risks and costs of international trade have occurred: notoriously, the railroads and steamships of the nineteenth century brought dras￾tically cheaper transportation (Landes 1969, 153-54, 196, 201-2; Hobsbawm 1979, Chap. 3); so, in our own genera￾tion, did supertankers, cheap oil, and containerization (Rosecrance 1986, 142). According to the familiar argument of Kindleberger (1973) and others, inter￾national hegemony decreases both the risks and the transaction costs of inter￾national trade; and the decline of hege￾monic power makes trade more expen￾sive, perhaps-as, according to this inter￾pretation, in the 1930s-prohibitively so. Analyzing a much earlier period, the Belgian historian Henri Pirenne (1939) attributed much of the final decline of the Roman Empire to the growing insecurity of interregional, and especially of Medi￾terranean, trade after 600 A.DS3 Global changes of these kinds, it follows, should have had global conse￾quences. The "transportation revolu￾tions" of the sixteenth, the nineteenth, and scarcely less of the mid-twentieth cen￾tury must have benefited, in each affected country, owners and intensive employers of locally abundant factors and must have harmed owners and intensive employers of locally scarce factors. The events of the 1930s should have had exactly the oppo￾site effect. What, however, will have been the political consequences of those shifts