13—
Exchange-Rate Changes and Ratchet Effects: A Historical Perspective^[*]

I have long complained that discussion of the great depression in the United States typically assigns no role whatsoever to the large exchange-rate changes, or to some small ones for that matter, that took place in the period. Two episodes are worth attention: first, the depreciation of the currencies of Argentina, Australia, New Zealand and Uruguay in 1929 and 1930; and second, appreciation of the dollar in the fall of 1931, generally characterized as the depreciation of sterling. I argue that both were seriously deflationary, and that in depression, depreciation is largely neutral while appreciation is deflationary. On the other hand, in the 1970s, appreciation was neutral and depreciation inflationary. In short, it can happen that fluctuating exchange rates affect international prices through a ratchet that works to accentuate inflation in a period of boom, and deflation in periods of depression. It can happen, too, that a change in exchange rates can raise prices partway in the depreciating currency, and lower them partway in appreciating, with no ratchet operating. Economic analysis that supposes that exchange-rate changes always operate in the same fashion, must be modified in the light of historical experience.

The first depression episode perhaps should be classified as a group of small exchange-rate changes rather than large. Shortly after the stockmarket crash of October 1929, first Argentina, then Australia, New Zealand and Uruguay either floated their currencies or depreciated them substantially. There was no direct connection with the stock

[*] A paper presented to a conference held at Brandeis University, December 4–6, 1987, and published in Stephan Gerlach and Peter Petri, eds, The Economics of the Dollar Cycle , Cambridge, Mass.: MIT Press, 1989.

market. These countries normally borrowed capital in London and New York, but had been unable to do so since the spring of 1928 when the stock-market boom began, when investors turned from foreign bonds to equities, and when interest rates rose in response to speculator demands for call money to buy stocks on margin. In the usual case, such a cut-off of long-term borrowing induces a shift to short money, and this response took place in Germany. It happened, however, that the London capital market worked somewhat differently with respect to the dominions and such favored borrowers as Argentina. These countries typically ran substantial bank overdrafts which built up as the deficit on current account ran along and were paid down with the proceeds of long-term borrowing when the lending banks considered that the overdraft was reaching an appropriate limit. When long-term lending was cut off in 1928, the borrowing needed to reduce or wipe out the overdraft was impossible. Given the reluctance of the London banks to enlarge the overdrafts, the central banks of these "regions of recent settlement," as they were called in League of Nations publications, had no choice but to let their currencies depreciate. The fact that this occurred after the stock-market collapse in October 1929 is fortuitous, although if the central banks had been able to hold on a little longer, the recovery of interest rates in New York and London might have enabled long-term debt to be sold in the spring of 1930, when foreign-bond markets briefly recovered and in fact hit new highs. As it happened, Argentina let the peso go in December 1929 and the other three currencies began to depreciate in the first quarter of 1930.

The depreciations of the Argentinian and Uruguayan pesos, and of the Australian and New Zealand pounds are usually thought of in terms of the stated currencies, but can be regarded as a small appreciation of the US and Canadian dollars. They had the effect of softening world prices in dollars for wheat and wool, and the sterling price of butter in London. The price of wheat fell from $1.50 a bushel in June 1929 to $1.05 in June 1930 — and New Zealand butter from 157.0 shillings per hundredweight in London in February 1930 to 135.2 shillings in April of the same year. Both price declines were well in excess of the percentage of depreciation — close to 5 percent at this early stage — commodity prices were generally soft and the appreciation of the US and Canadian dollars and of sterling in terms of the currencies that dominated wheat, wool, and butter markets gave these commodities a push against which resistance was weak.

It should be noted that wheat had not been strikingly affected by the spreading deflation from the stock market to commodities imported into New York that was a serious cause of deflation early in the depression. Wholesale prices fell between August 1929 and September 1930 by 22

percent in Japan, 16 percent in Canada, 15 percent in Great Britain, 14 percent it Italy, and 12 percent in the United States and Germany. I ascribe much of this to the liquidity squeeze of New York banks as they struggled in the crash to meet the consequences of the decline in stock prices for brokers' loans and stopped lending elsewhere. At that time, commodities exported to the United States, and especially to New York, were for the large part shipped on consignment to be sold on arrival to brokers who operated with credit. If credit seized up, the brokers who could not get their usual loans could not undertake their normal buying. The prices of import commodities traded in this way fell between September and December 1929 by as much as 26 percent in rubber, 18 percent in hides, 17 percent in zinc, 15 percent in cocoa, 13 percent in coffee, 10 percent in tin and silk, and 9 percent in copper and lead. An import commodity normally financed by the processing corporation in the United States, sugar, fell only 7 percent, and the export commodity, wheat, only 4 percent over these months. Overall, to be sure, the liquidity squeeze emanating from the stock market was far more important in depressing commodity prices than the limited effective appreciation of the dollars of Canada and the United States, and of sterling. They none the less added a small push against an open door.

Of much greater importance was the appreciation of the dollar and the gold bloc (depreciation of sterling and the sterling bloc) in September 1931. The sterling rate went from $4.86 on September 20 to $3.25 in December at the high for the dollar, a decline for sterling of 30 percent, or an appreciation of the dollar (and gold) of close to 40 percent. Prices did not rise in sterling, but fell in dollars and French francs. Measured from September 1931 to March 1932, prices of internationally traded commodities expressed in dollars behaved as shown in Table 13.1.

Whether prices rise in the depreciating country or fall in the appreciating when substantial change occurs in exchange rates, or fall in between, with some rise in the depreciating and some fall in the

appreciating currencies, depends, of course, on the elasticities. The normal classical assumption of elasticity optimism is that a country is a price-taker for imports and a price-maker in exports, so that the terms of trade go against it when its currency is depreciated. The ultra-classical assumption of such an economist an Frank Graham is that a country is a price-taker for both imports and exports and that an exchange-rate change will leave prices in foreign exchange, and the terms of trade, unchanged. In the British case, however, the terms of trade improved with depreciation — an anti-classical outcome — as Britain in world depression was more of a price-maker in imports, the world depending on its market for primary products more than it depended on world markets for its manufactured exports. It is noteworthy that the elasticities are partly associated with the nature of the commodities concerned, but also very much affected by the state of world markets, whether they are poised on the edge of deflation which makes elasticities of both demand and supply low for price decreases (and high for price increases which were not involved).

It can be argued that the spread between dollar and gold prices, on the one hand, and sterling prices, on the other, was achieved more or less by a part rise in sterling and part decline in dollars and gold. A chart of Australian export prices in Kindleberger (1986, p. 86) shows Australian export prices rising in sterling and Australian currency at the end of 1931, falling in dollars. But the Australian commodities, wheat and wool, behaved differently from the bulk of primary product prices in Table 13.1. The large number of prices falling 23 to 34 percent in dollars makes the case for the ratchet on the down side in depression.

There is something of a mystery why United States prices turned up with the depreciation of the dollar in the spring of 1933, with sterling and gold prices steady rather than pushed further down. In this instance the exogenous depreciation of the exchange rate lifted an index of internationally-traded commodities (Moody's) and of the Dow Jones industrial stock price index at a time when the world economy had not recovered very far from the lows of June 1932. Some commodities doubled in price from March 1933 to June — corn and rubber for example, one an export and the other an import commodity. Hides went from 5.2 cents a pound to 12.2 cents, a rise of 135 percent, as the dollar went from $3.40 to the pound sterling to $4.80. The depreciation was undertaken in response to the Warren and Pearson (1933) study of the relationship between commodity prices and the agio on gold in the greenback period from 1863 to 1879, when again US prices varied with the rate of depreciation and world prices held steady. At that time, however, the United States was much less of a dominant power in the world economy than in 1933.

I lack a satisfactory explanation why British prices held steady when sterling depreciated in 1931 while US prices rose and world prices remained unchanged when the dollar depreciated two years later. One could hypothesize that the world economy had hit bottom via the investment multiplier with gross investment close to zero and net investment strongly negative. It would be hard to claim that the United States was more a price-taker in internationally-traded goods than Britain, although this explanation is probably valid for the period after the Civil War.

Some part of the explanation may lie in the fact that the US exchange rate at par was strongly overvalued — certainly against sterling and the yen — and that letting the exchange rate go lifted a deflationary pressure. Ragnar Nurkse (1949) once wrote that the United States was wrong to depreciate the dollar when the US current account in the balance of payments was in surplus. In early Keynesian analysis it was thought that if a country had a surplus, it meant that its foreign trade was giving a positive stimulus to national income and that it would be a beggar-thy-neighbor type of policy to depreciate. These strictures clearly applied in the 1930s to Japan, which consciously adopted policies of depreciating to undervalued rates to export its unemployment. In the US case, however, it can be argued against Nurkse that the dollar rate had been overvalued, but that the declines in national income and in imports through the national income multiplier were so severe that imports fell more than exports and produced a residual export surplus, with no implication that the exchange rate was undervalued. Depreciation, in other words, relieved strong deflationary pressure on income and prices from the foreign-trade sector, and imparted an upward pressure to prices. But this and other explanations are very much ad hoc , and the divergent behavior in prices between Britain in 1931 and the United States in 1933 remains to be understood.

If we leave aside the 1933 case of the US dollar, it can be said that depreciation in the 1930s left domestic prices in the depreciating country relatively unchanged and put strong downward pressure on prices in those countries where the currency appreciated, producing a ratchet in which an exchange rate that went up and down could be expected to produce a net decline in prices. The ratchet in the 1970s, on the other hand, moved in the opposite direction. If depreciation raises domestic prices and appreciation leaves them unchanged, a currency moving sinusoidally down and up will find prices raised on balance. This is what happened during the 1970s, starting with the depreciation of the dollar in 1971 and the adoption of floating in 1973. Dollar prices rose when the dollar depreciated, and remained unchanged when the dollar recovered.

The ratchet worked in the 1970s on the whole as a market pheno-

menon, based on the circumstance that the world economy operated as a sellers' market, as contrasted with the 1930s when there was a buyers' market. The 1970s were poised on the edge of inflation, whereas the 1930s were strongly deflationary. But the model can be seen with conscious direction behind it if one looks at the Organization of Petroleum Exporting Countries (OPEC) and the price of oil.

Figure 13.1
World-market commodity prices, oil and non-oil, 
compared with the effective exchange rate for the dollar, 
1972–82. (Source: Bank of International Settlements (1982, pp. 40 and 42.)

In November 1973, at the time of the Yom Kippur War, the Arabian members of OPEC embargoed all oil exports to the United States and the Netherlands, which were deemed to be partial to Israel, and reduced all other exports by 25 percent. The price rose sharply to $20 per barrel before settling down at $10. Thereafter if the dollar depreciated, OPEC raised the price in dollars to maintain it in other currencies, such as the Deutschmark, sterling, and the yen. But if the dollar appreciated, the dollar price of Arabian crude was left unchanged, while prices in other currencies necessarily rose as these currencies depreciated against the dollar. The ratchet worked intermittently and with varying degrees of timing. Figure 13.1 shows the two oil shocks of 1973 and 1979 and the inching up of the dollar price of Arabian crude between those episodes when the effective exchange rate for the dollar rose in 1975 and 1978, with no decline in 1974 or 1976 when the dollar fell. The ratchet is much less clearly seen in non-oil prices during the decade: the effective exchange rate for the dollar, measured as a percentage change over four quarters, moves in a sine wave up and down, while world non-oil commodity prices exhibit for the most part price increases over 12 months, except in the recession of 1974 and in 1978.

The belief that prices in general, traded-goods prices and exchange rates behave differently in relation to one another, depending upon whether economic conditions are inflationary or deflationary, has an analogue in the operation of Gresham's law. Most economists summarize Gresham's law as bad money drives out good, i.e. that people spend bad or overvalued money and hoard, melt down or export good or undervalued money. Roland Vaubel's (1977) recommendation of parallel currencies rules out this possibility on the ground that good money may drive out bad. In his opinion, sellers will demand good money and refuse to take money that has been overissued or otherwise become overvalued. The possibility must be admitted. In a study of mints in Spain and France at the beginning of this millennium, Thomas Bisson (1979) cites a case of mints competing to have their coinage accepted and circulated and mints with full-weight coins succeeding over those which issued coins underweight. The usual formulation of Gresham's law as bad money drives out good rests on the belief that the choice of which money is spent rests with the buyer, not the seller, and implies a background of buyers' markets dominating over sellers'. The issue may be complicated in particular cases by laws governing legal tender, money which the law decrees has to be accepted in payment of debts. The general point, however, is that in the circulation of money, as in the relationship between domestic prices and exchange rates, outcomes may depend upon particular circumstances, in the instant cases whether buyers' or sellers' markets prevail.

This brings us to the present. The price of oil collapsed at the end of 1986 as the cartel proved to be unable to agree on sharing production cuts during stable phases of the Iran–Iraq war. The Federal Reserve undertook tight monetary policy from 1982, leading to high interest rates, capital inflow and an appreciation of the dollar that continued unexpectedly, in view of the fundamentals — a substantial budget deficit and a substantial trade deficit — to February 1985. The appreciation of the dollar may have held back inflation in the United States, but it did not lead to deflation. United States prices declined to the end of 1982, rose mildly through mid-1984 and then declined again. Prices in the Federal Republic of Germany rose substantially while the Deutschmark was depreciating in 1983, held steady during 1984, and came down again as the Deutschmark appreciated after February 1985. West German prices seem to have moved up during Deutschmark depreciation and down during appreciation, with US prices relatively steady. There was no ratchet such as took place on the upside during the 1970s or the downside in the 1930s. The historical analysis seems not to have applied, probably because world prices were slipping consistently during the 1980s until early 1987. Commodity schemes such as that for tin and coffee collapsed. The Green revolution and the policies of the European Community in agriculture depressed world food prices. Such a commodity as copper was hurt by the development of optic fibers. With Third World and other primary product prices falling, depreciation of the Deutschmark to 1985 and of the dollar thereafter did not produce the inflationary conditions of the 1970s.

Before the stock-market decline in October 1987, there were signs that this was changing. Raw material prices rose substantially in many commodities from the fall of 1986 or the early months of 1987 through September. There was fear that if the dollar depreciated still further, it would be highly inflationary, perhaps restoring the ratchet of the 1970s on the up side. The stock-market crash of October 1987 seems to have differed from that of 1929, as I write, in that the liquidity squeeze in the stock market has not been communicated to commodities. Not only has the ratchet model of the 1930s and the 1970s been suspended, but the connection between share prices and commodity prices that played a major role in the 1929 depression is out of action — thus far — as well.

This experience leads me once again to emphasize that in the real world it is necessary to change models frequently. As economists we are trained to look for models which are general, and to distrust ad hoc explanations. At the same time, it is vital to recognize that differences in the initial conditions may require changes of models from those that have functioned well on earlier occasions.

References

Bank for International Settlements (1982), Fifty-second Annual Report, 1st April 1981 to 31st March 1982 , Basle.

Bisson, Thomas N. (1979), Conservation of Coinage, Monetary Exploitation and Restraint in France, Catalonia and Aragon , c.1000-1125 A.D. , Oxford: Clarendon Press.

Kindleberger, Charles P. (1986), The World in Depression, 1929-1939 , revised edn, Berkeley: University of California Press.

Nurkse, Ragnar (1949), "Balance-of-Payments Equilibrium," in American Economic Association, Readings in the Theory of International Trade , Philadelphia: Blakiston, pp. 1-25.

Vaubel, Roland (1977), "Free Currency Competition," Weltwirtschaftliches Archiv , vol. 113, no. 3, pp. 435-59.

Warren, George F. and Frank A. Pearson (1933), Prices , New York: Wiley.