– From the introduction to Capitalism: Competition, Conflict, Crises –
Chapter 11 closes Part II of this book by applying the classical argument to international competition, that is, international trade balances and terms of trade (real exchange rates).
The theory of international trade is a critical part of modern debates about the costs and benefits of the globalization of production and finance. Neoliberalism portrays markets as self-regulating social structures that optimally serve all economic needs, efficiently utilize all economic resources, and automatically generate full employment for all persons who truly wish to work. Proponents of neoliberalism point to the indisputable fact that the rich countries are market-based economies that developed in the context of a world market. Critics of neoliberalism dispute all of these claims. They note that rich countries, from the old rich of the West to the new rich of Asia, relied heavily on trade protectionism and state intervention as they developed and that they continue to do so even now. They contend that the trade liberalization imposed on the developing world has actually led to slower growth, greater inequality, a rise in global poverty, and recurrent financial and economic crises in most countries. Most important, they generally argue that in any case orthodox free trade theory is irrelevant because free competition does not prevail even in the rich countries, let alone the poor ones—a standard trope among heterodox economists because they conflate competition with perfect competition (chapters 7 and 8). This chapter demonstrates that the theory of real competition has a very different set of implications for international trade. The conventional (Ricardian) theory of free trade does not follow in a competitive context and the very patterns to which heterodox economists point as evidence against (perfect) competition can be explained from real competition. From the latter perspective, globalization has worked as expected—favoring low-cost producers over the high-cost ones.
Section II examines two crucial premises of the theoretical foundations of orthodox trade theory: (1) that free trade is regulated by the principle of comparative costs; and (2) that free competition leads to full employment in every nation. The principle of comparative costs is eminently familiar, most often presented as the proposition that a “nation” would always stand to gain from trade if it were to export some portion of the goods it could produce comparatively more cheaply at home, in exchange for those it could get comparatively more cheaply abroad. It is implicit that trade will be balanced (i.e., that the value of imports will be equal to the value of exports). But this purely normative proposition has little significance unless it can be shown that free trade among market economies actually creates such outcomes. International trade is actually conducted by profit-driven exporting and importing firms. Therefore, whenever conventional trade theory seeks to appear more realistic, it switches to the positive claim that free trade will be regulated by comparative advantages and that the terms of trade will always arrive at a point which equates the values of exports and imports. No nation need fear trade due to some perceived lack of international competitiveness because, in the end, free trade will make each nation equally competitive in the world market. This conclusion requires that the terms of trade of any country will automatically and successfully move to eliminate trade deficits or surpluses. The assumption of universal full employment in rich and poor countries is equally critical: after all, who can say that trading exports for imports is a “gain” if that outcome is achieved at the expense of sustained job losses? The theory of comparative advantage then seeks to explain the determinants of comparative costs. For instance, on the dual assumptions that trade is ruled by comparative costs and that full employment always obtains, the Hecksher–Ohlin–Samuelson (HOS) model of comparative advantage claims that differences in national comparative costs are rooted in differences in national “endowments” of land, labor, and capital.
All three of the central propositions of orthodox trade theory have been vigorously disputed. The notion of universal full employment becomes a cruel jape in light of the fact that there were a billion people in the world who were unemployed or underemployed even at the height of the global boom preceding the 2007 global crisis. The claim that a fall in the terms of trade will eventually improve the balance of trade has long been dogged by the infamous “elasticities problem.” And the claim that a trade deficit will automatically lower the terms of trade until the deficit is eliminated is bedeviled by the simple fact that balanced trade simply does not obtain anywhere, not in the developing world, not in the developed world, not under fixed exchange rates, not under flexible exchange rates. On the contrary, persistent trade imbalances are the rule.
Section III traces the two dominant reactions to the empirical problems of standard trade theory. The first type focuses on the fact that balanced trade and/or Purchasing Power Parity (PPP) are only meant to hold in the long run, so that existing postwar data (now spanning seventy years or so) may not be long enough. Others have shifted ground by focusing on a host of short-run models that contradict each another and many elements of the reality they intend to explain. Despite the fact that many mainstream economists “readily admit their failure,” the underlying notion of comparative cost advantage continues to dominate textbooks and models and economic policy itself. The other major reaction has been to modify one or more of the standard assumptions by incorporating oligopoly, economies of scale, and various concrete factors such as the composition of trade, differential elasticities of demand, and differences in technology and in accumulated and/or institutionalized human knowledge. All of these give rise to particular exceptions to the standard results, which in turn provide some (limited) room for state intervention in strategic sectors and strategic activities such as R&D. The resulting models are extremely complicated, encompass multiple possible outcomes and provide “few unambiguous conclusions.” I argue that the real problem lies at the very root of these models, which is the Ricardian principle of comparative cost.
Section IV re-examines Ricardo’s principle of comparative cost in light of the theory of real competition. In real competition within a nation, firms constantly seek to cut their costs in order to be able to cut their prices and displace their competitors. Firms with lower costs tend to emerge more often as winners while those with higher costs are more likely to end up as losers. This is the central selection mechanism of capitalist competition. Smith emphasizes that “private profit is the sole motive” in the application of capital to domestic or international trade. Ricardo begins from this same point, seeking to show how international trade patterns arise from the actions of individual profit-seeking capitals in different countries. In order to bring out the stark logic of his argument, Ricardo begins by assuming that Portuguese capitals initially have lower cost-based prices in all commodities, so that they dominate both English and Portuguese markets. But then, as money flows into Portugal from England, Portuguese costs and prices rise and English costs and prices fall. We can imagine that as Portuguese goods become progressively more expensive and English goods progressively cheaper, the Portuguese commodity with the smallest absolute cost advantage over its English counterpart will be the first to switch from the winner’s column to the loser’s. From the English point of view, this will be the commodity with the smallest cost disadvantage. But unless trade becomes balanced, the process will continue and the Portuguese commodity with the second smallest advantage (the English one with the second smallest disadvantage) will switch columns, and so on. All of this obtains through the actions and reactions of individual profit-seeking producers in the two countries. When the Ricardian process comes to rest it will appear as if “Portugal” had chosen to specialize in producing the goods in which it had a “comparative cost advantage,” exchanging them for commodities of equal money value (since trade is balanced at the rest point) consisting of goods in which “England” had a comparative cost advantage. This allows Ricardo to jump from the argument that the behavior of individual profit-seeking firms will lead to the rule of comparative cost to the proclamation that countries should use comparative costs to determine their trade patterns. Neoclassical economics often skips the derivation altogether, resorting instead to the fictional representation of England and Portugal as individuals each of whom trades in order to “gain.” This has the ideological value of instilling the notion that the very purpose of free trade is to benefit all nations, rather than to make profits for their businesses. The section includes an extended treatment of the formal structure of the theory of comparative costs.
Ricardo’s conflation of the balance of trade with the balance of payments is extremely important to his construction. A country’s balance of payments is the sum of net inflows into the country: exports minus imports (the trade balance), direct investment in the country by foreigners minus investment abroad by domestic agents, short-term capital inflows such as private or business bonds purchased by foreigners (i.e., loans made by foreigners to domestic agents) minus similar financial transactions made in foreign countries by domestic agents, and so on. Ricardo proceeds as if commodity trade flows are completely separated from financial flows, so that a trade balance is synonymous with a payments balance. Money appears in his story as medium of circulation, but never as financial capital. This is extremely odd from a historical point of view, since the export and import of financial capital (international borrowing and lending) is intrinsically linked to the flow of funds arising from the export and import of commodities. It is equally odd from a theoretical point of view because it implies that trade and finance flows are completely divorced from each other. Both Marx and Harrod point to this as a critical weakness in Ricardo’s logic.
Section V develops the classical theory of absolute cost advantage. The Ricardian argument is really a story about the determination of international regulating capitals. When trade opens, Portugal and England each produce both wine and cloth, so there are two different regulating producers for each good, one in each country. Despite the fact that Portugal has the initially lower cost-based prices in both goods, the comparative costs argument says that international competition will end up selecting British firms as the regulating capitals for cloth leaving Portuguese firms with the regulating role for wine. In the theory of real competition, the price-leader (regulating capital) in any industry is the one with the lowest unit cost, the term “cost” now defined in the proper business sense as the sum of unit wages, materials, and depreciation. The first difficulty with the Ricardian story is that changes in the relative international prices of goods will also affect the relative costs of these same goods. This is the logical extension of Sraffa’s central point that prices and costs are inextricably intertwined (chapter 9, section XI). Then comparative costs may not change at all in response to any changes in the real exchange rate (nominal exchange rate and/or the relative national price level), leaving Portuguese capitals in charge of both industries and eliminating British ones. Even if comparative costs do respond to changes in real exchange rates, they may not respond sufficiently to displace Portuguese capitals, so once again British capitals are doomed. To put it differently, sufficiently large absolute costs advantages will not be overturned by real exchange rate effects. Worst of all comparative costs may change in the “wrong” direction (i.e., they may make the absolute cost advantage of Portugal country even greater). This means that even if the real exchange rate did automatically vary with the trade balance, as Ricardo supposes, comparative costs will not move in the Ricardian manner as long as real costs (real wages and productivity) are determined at the national level. It is formally demonstrated that for given real wages and specific industry efficiencies in each country, in a two sector model the comparative cost in any industry is a ratio of two linear functions of the international relative price and may fall or rise with the relative price depending on the coefficients. Moreover, the extent of any such a movement is itself limited by the relative structures of production. In the end, international competitiveness will be tied to differences in efficiency, real wages, and technical proportions, and there is nothing in free trade itself that will eliminate absolute cost advantages or disadvantages.
The second problem with the Ricardian theory is that real exchange rates need not change at all in the face of trade imbalances. Marx comments that a country with a trade surplus will experience an increase of liquidity which will lower its interest rate, while a country with a trade deficit will experience a tightening of liquidity and an increase in the interest rate—all through the normal functions of capital markets. Harrod comes independently to the same conclusion. With capital flows offsetting trade imbalances, the net effect on the balance of payments will depend on the relative magnitude of these two effects: the exchange rate may not change at all, or if it does, it may change in the “wrong” direction (i.e., the exchange rate of the trade surplus country may depreciate rather than appreciate).
In international real competition, the regulating capitals will essentially be those with the lowest integrated real unit labor costs. Assuming that countries export the goods in which they have the lowest costs (for given quality), the terms of trade of any country will depend on the ratio of the integrated real costs of its exports relative to that of the producers from which it gets its imports. The key point is that the terms of trade are pinned by national real wages and structures of production, so that they cannot also move to endogenously balance trade as in the Ricardian theory. The classical formulation can be extended to cover nontradable goods, which will affect input costs insofar as they enter into production and affect the money wage insofar as they enter the wage basket. Then the classical argument implies that the terms of trade (real exchange rate) is driven by two components: relative real regulating costs and the ratio of tradable/nontradable goods. A similar expression is developed for the common currency ratio of any two national price indexes, which immediately tells us that this ratio will be constant only if the two had the same overall composition in the sense of having the same composition of goods and the same ratio of nontradable to tradable prices. The classical argument therefore implies that PPP will not generally hold.
The application of real competition to the theory of international trade leads to several distinct propositions. First, industry comparative costs and terms of trade are determined by relative real wages, relative productivities of regulating capitals, and the effect of tradable/nontradable goods. Second, the direction of a nation’s trade balance is determined by its absolute cost advantage or disadvantage (a classical channel) while its size will also depend on relative national incomes (a Keynesian channel). Changes in the latter will affect the trade balance but will not permanently switch it from surplus to deficit unless they switch comparative costs. Third, trade imbalances will create payments imbalances which will affect interest rates and induce short-term international capital flows (a classical channel), and perhaps also change national income through their influence on investment (the Keynesian channel). The end result will be that countries with absolute cost advantages will recycle their trade surpluses as foreign loans while countries with absolute cost disadvantages will cover their trade deficits through foreign borrowing. All of this will arise through the workings of free trade and free financial flows, though, of course, policy measures may produce similar effects.
Section VI compares the standard and classical theories of free trade with the empirical evidence. The comparative cost hypothesis implies that the real exchange rate will vary so as to ensure that trade remains balanced while the international real competition implies that trade imbalances will be the norm. Trade data for fifteen major countries over the half-century from 1960 to 2009 makes is abundantly clear that trade does not generally balance. The orthodox PPP hypothesis posits that real exchange rates will be stationary over the long run, but the large empirical literature discussed in this section establishes that PPP does not hold. A chart of the real effective exchange rates in terms of producer prices for the United States and Japan shows both to be highly trended in opposite directions. The PPP argument can also be formulated as the hypothesis that nominal exchange rates will depreciate at the same rate as inflation (so as to maintain a constant real exchange rate). The US and Japan data makes it clear why this (relative) version of PPP is equally unsupportable as a general empirical proposition. However, in the particular case of high inflation, (relative) PPP does appear to hold. The classical theory of trade predicts both the trended nature of real exchange rates evident in the US and Japan data and also the correlation between nominal exchange rates and inflation rates observed in the case of high relative inflation. The classical hypothesis is that the real exchange rate er ≡ p · e/pf = (p/pf ) e, where p = the domestic price level, e = the exchange rate (foreign/domestic currency), and pf = the foreign price level, depends on relative real unit labor costs and the tradable/nontradable price ratio. Since the latter two terms change slowly from year to year, the real exchange must also change slowly (except for shocks). But the real exchange rate is the product of the domestic relative price level (p/pf ) and the exchange rate. Hence, when the relative price level rises sharply in the face of rapid domestic inflation, the nominal exchange rate must depreciate at roughly the same rate.
The preceding argument also implies that the real exchange rate will be linked to corresponding integrated real unit labor costs adjusted for the ratio of tradable/nontradable prices. Direct unit labor costs were used in the absence of data on integrated costs to construct adjusted real unit labor for the two countries relative to their trading partners and the corresponding charts show that each country’s real exchange rate does indeed track the classical fundamentals. On the econometric side, the actual and fundamental variables were found to be cointegrated with speeds of adjustment which are statistically significant and of the correct sign. Finally, it is shown that the deviations of the real exchange rates from adjusted relative real unit labor costs are stationary. Given the data limitations discussed and the large impact of the capitalflow and interest-rate shocks, it is remarkable how stable this actual/fundamental ratio is over the long run. Hence, the classical approach also provides us with a robust policy rule-of-thumb for the competitively sustainable level of the real exchange rate—a rule which is clearly superior to the widely used PPP hypothesis.