Currency regimes and currency crises: What about cocoa money?

Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

Currency regimes and currency crises: What about cocoa money? Mark S. LeClair ∗ Department of Economics, 1073 North Benson Road, Fairfield University, Fairfield CT 06824, USA Received 23 May 2005; accepted 13 August 2005 Available online 13 September 2005

Abstract Since the mid-1990s, there has been a significant shift towards floating exchange rate regimes by developing nations, primarily due to the lack of a viable alternative. Hard-peg systems, which both eliminate independent monetary policy and, if not credible, are subject to speculative attack, are increasingly viewed as a poor choice. This paper explores, theoretically and empirically, the potential benefits and drawbacks of an alternative: commodity-backed money. While proposals for commodity-based money in the industrialized world date back to 1934, with the seminal work by Benjamin Graham, this work analyzes its application to the developing world and its key commodity products. © 2005 Elsevier B.V. All rights reserved. JEL classification: F33 Keywords: Currency regimes; Commodities; Commodity money

1. Introduction Recent currency crises have called into question the advisability of utilizing certain currency regimes, at least in developing nations. The most notable, the Asian crisis of 1997–98, demonstrated the problems with pegged, inter-dependent currencies. This was followed by the Ecuadorian Sucre crisis of 2000, which ultimately resulted in the dollarization of that economy. Most recently, the collapse of the dollar-standard in Argentina illustrates the drawbacks of this regime. Since the demise of the fixed exchange rate system in 1973, developing nations have used ∗

Tel.: +1 203 254 4000x2295; fax: +1 203 254 4074. E-mail address: [email protected].

1042-4431/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.intfin.2005.08.004

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

43

a variety of means to avoid floating their currencies. Pegged systems have been the most popular, with countries fixing their currencies against the U.S. dollar or an alternative, such as SDRs. In Africa, pegging has taken the form of a currency zone, with contiguous nations pegging against the French Franc (until its replacement with the euro). This search for stability has exposed the inherent problem with most fixed regimes: their susceptibility to speculative attack. The result has been a recent trend towards floating exchange rates—not because of their inherent attractiveness, but simply to avoid the drawbacks of pegged systems. The recent instability in currency markets has led to renewed interest in alternative regimes. One such alternative is the utilization of commodity-backed money, a concept that dates back to 1937, starting with proposals by Benjamin Graham to adopt a currency backed by a basket of commodities (Graham, 1997). As recently as 1999, the issue of securing currencies with some form of tangible asset was raised once again (see Haussmann, 1999). Although Graham originally favored the adoption of a commodity-backed currency in the industrialized world, the developing world’s poor experience with existing currency regimes suggests this would be a more fruitful place to examine commodity money. In addition to bringing about currency stability, a commoditybased currency would offer a solution to one of the problems that has plagued developing nations since the 1960s; commodity price instability. This paper will examine, theoretically and empirically, the advantages and disadvantages of adopting a commodity-backed currency within certain nations in the developing world. Although a number of commodities could potentially act as the backing for such a currency, cocoa has been chosen because of the unique characteristics of this particular market (see Section 7). The intent will be to demonstrate the feasibility of such a system, yet expose some of its inherent flaws.

2. Alternative currency regimes—advantages and disadvantages As will be demonstrated in the section that follows, there has been a general movement towards floating exchange rate regimes, whether fully flexible or lightly managed. This is partly due to the advantages of flexible rates: The ability to utilize monetary policy, insulation from international economic shocks, and freedom from the threat of currency crises. Undoubtedly, it is also attributable to the poor experience developing nations have had with fixed exchange rates. The disadvantages of floating rates, volatility that impacts foreign trade and bias towards excessive monetary growth, have been overshadowed by the advantage of not facing a currency crisis. Actively managed (“dirty”) floats are designed primarily to eliminate exchange rate instability, not to direct the path of currency valuations. A successful managed float requires significant international reserves, as the monetary authority will be intervening regularly to counteract shortterm market fluctuations. Nations that use managed floats are unlikely to be subject to currency crises, since there is no commitment to maintaining a currency’s value. As with flexible exchange rates, control over monetary policy is maintained. Pegged systems, Crawling and Pegged with Bands all reduce the uncertainty of exchange rate movements, with the latter performing nearly as well as fixed exchange rates. All pegged systems are, however, vulnerable to currency crises, and all require the holding of significant international reserves to achieve target exchange rates. Nations utilizing pegged systems can apply monetary policy only sparingly, and international disequilibria will be corrected through real sector adjustments. In the past, the attractiveness of pegged systems in the developing world resulted from a search for both stability, and a means of taming inflation that resulted from imprudent monetary policy.

44

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

Hard pegging systems, either in the form of a Currency Board or so-called “Dollarization”, require that the exchange rate be maintained at a fixed level against a designated foreign currency. The recent situations in Ecuador and Argentina have resulted in divergent responses to currency crises, with the former adopting dollarization, and the latter forced to abandon a Currency Board system. Hard peg systems virtually eliminate monetary policy, both in terms of managing economic conditions, and even in terms of being able to ensure adequate liquidity in the financial sector. The quantity of the domestic currency issued is dependent upon the holdings of foreign reserves, indicating that disequilibria will be eliminated through variations in domestic economic activity. Hence, after a 10-year experiment with a Currency Board system, Argentina was forced to drop its 1:1 pegging to the dollar when economic conditions became too severe. The advantages of maintaining a hard peg arise primarily from its ability to provide credibility to a currency in the aftermath of a rapid inflationary period. As these two regimes completely eliminate monetary discretion, no apprehension is left in world markets about excessive future monetary growth (except, of course, when it becomes clear that a nation is about to abandon its hard peg). In addition, both regimes prevent currency crises of the type seen in Asia in 1997. Since the local currency is backed by hard currency, it is not subject to attack by speculators. This one advantage of a hard peg system may overwhelm all of its drawbacks, and induce nations to adopt such a regime. 3. Evaluations of currency regimes The literature that addresses the advantages and disadvantages of various exchange rate regimes is far too extensive to cite in its entirety; so recent examples will be noted. Eichengreen (1994) analyzes the various exchange rate regimes, and suggests which are more suitable under evolving economic conditions. Yagci (2001) provides an overview of the various arrangements open to developing nations, expanding upon earlier work by Edwards and Savastano (1999). These authors provide a thorough analysis of the potential drawbacks of each system. Currently, focus has been on the “disappearing middle”, observing that more nations are choosing either hard-peg or floating rates (see Fisher, 2001 and Frankel et al., 2001). This tendency was also noted by Eichengreen (1994) and Summers (2000). In light of the rising number of developing nations utilizing currency boards, Kopcke (1999) analyzes what this type of regime can and cannot accomplish. He concludes that currency boards are a good starting point for many nations, but should eventually be replaced to allow for independent monetary policy. Eichengreen et al. (1998) notes that conversion to a less rigid system should occur during a period of rising or stable exchange rates, to insure a peaceful transition. The economic consequences of choosing a particular regime are explored by Carramazza and Aziz (1998) and Levy-Yeyati and Sturzenegger (2000). Research has also been focused on the problem of the “holy trinity”, simultaneously achieving independent monetary policy, stable exchange rates and capital mobility (see Rose, 1996). Most authors have concluded that no exchange rate system is perfect for all nations at all times. 4. Why a new currency—trends in the utilization of various regimes Recently, there has been a significant increase in the number of countries that freely float their currencies.1 As of June, 2001, 40 nations (22% of those reported by the IMF) claimed independently floating currencies, and an additional 23.8% conducted a managed float of one form or 1 Although authors such as Calvo and Reinhart (2002) argue that monetary intervention is frequently extensive enough to label these as “dirty” floats.

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

45

another. Hence, nearly half of the nations listed by the IMF had chosen market determination of exchange rates. Those nations that did utilize a pegged system tended to choose hard pegs (including currency boards and adoption of another nation’s currency). Countries with arrangements of this type numbered 87 in 2001, or just over 48% of the sample. Finally, only 11 nations chose a variety of crawling peg system. These can be compared to IMF figures for December of 1990. Of the 153 nations listed by the Fund in that year, 25 (16.3%) floated their currencies, while 23 (15.0%) conducted a managed float. Hence, there has been a significant rise over the last decade in the number of countries that have departed from pegged systems. This has little to do with the attractiveness of the floating exchange rate regime, but is simply a reflection of frustration with the alternatives. As noted above, floating rates expose a country’s currency to excessive volatility, and also permit the central bank to carry out excessive monetary growth. This has to be balanced against its major advantage, which is that currency shocks are not translated into output shocks. 5. Commodity backed money A variety of different methods have been suggested for establishing commodity-based money. Yet, the plan envisioned by Graham remains the soundest foundation for such as regime. Fig. 1 illustrates the mechanics of the Graham plan. It is based on the maintenance of a stockpile comprised of key commodities, according to their contribution to international trade. Currency units (CUs) would then be issued based on a fixed quantity of this composite commodity, with the price of the CUs determined by the long-term average price of these commodities. As illustrated in Fig. 1, if the market price of these raw materials, Pc , rises above the established price, P ∗ , arbitragers will exchange currency units for commodities, increasing the world supply of raw materials. This will continue until the market price returns to P ∗ , and the arbitrage opportunity is eliminated. This results in both a reduction in the world price of commodities (back to their

Fig. 1. Graham’s commodity reserve plan.

46

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

long-term value) and a stabilization of the value of the CUs. As shown in Fig. 1, the reverse would occur if the imbedded value of the commodities rises to their long-run value. The potentially fatal flaw in Graham’s plan was the unintended monetary “policy” that commodity price swings might bring about. A rapid rise in the market value of the commodities imbedded in a CU would result in the redemption of a significant quantity of currency. The resulting drop in the money supply could be large enough to affect macroeconomic activity. Although changes in real sector activity are naturally imbedded in any fixed exchange rate system, the past volatility of commodity prices suggests that the swings could be destabilizing. Graham’s response was that the drop in money supply brought about by the redemption of CUs would lead to a generalized disinflation, resulting in a drop in Pc , reversing the effect of the initial price rise. The literature written in response to Graham’s original plan is extensive. Supporters included Graham (1938, 1941), Hayek (1943), Simmons (1945), and Rosensen (1948). Critics, mostly arguing about the practicality of the plan, included Anderson (1938), Beale et al. (1942), and Friedman (1951). Keynes (1942, 1943) also weighed in on the Graham proposal. Finally, Bennett (1949) conducted an empirical test of Graham’s proposal. In general, both supporters and critics acknowledge the political and practical difficulties of establishing commodity-based money in an industrialized nation. More recently, there was a revival in the academic discussion of commodity-backed currencies, with Hall (1982) and Greenfield and Yeager (1983) supporting commodity-based money, and White (1984, 1986) against. 6. Commodity backed money in the developing world The frustration with existing pegged systems results from their lack of flexibility in the presence of macroeconomic instability. The monetary authority lacks the ability to address recessions, resulting in popular frustration and potentially political unrest. Argentina’s rather abrupt departure from a dollarized economy illustrates the potential dangers. With declining economic output and vocal street protests, the political leadership was forced to abandon its currency board. Clearly, departure from a soft-peg system would have been less disruptive. Yet, few developing nations have chosen this type of regime, since it gains them neither stability nor control of monetary policy. A commodity-backed currency, designed a la Graham, would involve issuing currency that was backed by a fixed quantity of a particular commodity, say coffee. Potentially, such a scheme would simultaneously achieve two critical goals for the developing nation issuing the currency: stabilizing the value of its currency, and preventing price swings in the coffee market. If the price of coffee rose above its established price, P ∗ , holders of CUs would exchange them for beans, returning the price to P ∗ . Initially, the value of the CUs would tend to rise, as the value of the coffee imbedded in each unit exceeds the face value of the currency. As the market price of coffee is returned to P ∗ , however, the target exchange rate would be achieved also. All of the above is consistent with the original formulation of Graham’s proposal. A difference arises, however, in the final mechanism by which such a plan would achieve monetary stability. As outlined above, the over- (under-) issuance of CUs is offset by the onset of inflation (disinflation) and the effects of the feedback loop in Fig. 1. Although consumers in coffee-producing countries tend to absorb a great deal of coffee, it is not clear that a generalized change in the price level would necessarily spill over into the market for what is primarily an export product. Hence, a group of nations adopting a coffee-based currency could end up experiencing destabilizing monetary

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

47

Table 1 Production of, and dependency on, cocoa exports, major producers, 2000 Country

Production (000 metric t)

Percentage of world (%)a

Dependency ratio (%)

Brazil Cameroon Ghana Indonesia Ivory Coast Nigeria

128 120 440 396 1300 165

4.4 4.1 15.0 13.5 44.3 5.6

0.2 7.3 34.6 0.6 32.7 0.8

Source: export figures are taken from IMF, International Financial Statistics. Cocoa production figures are from CRB, Commodity Yearbook. a The remaining 13% is due to minor producers.

policy. The rejoinder to this argument is that, if the required extractions and infusions of currency into the economy are small, then the impact on the macroeconomy might be negligible. The destructive effects of currency instability (or lack of policy options under hard-peg systems) compounds the problems faced by developing nations that are single-commodity exporters. Although scholarly interest in export dependence has waned recently, a significant minority of developing nations remains vulnerable to commodity prices swings. As the focus of this analysis is on the cocoa market, dependency ratios (exports of key commodity as a percentage of total exports) are presented for the major producers of just this commodity (see Table 1). The adoption of a commodity-backed currency could eliminate the difficulty with price instability, as the redemption/issuance of CUs would necessarily result in the elimination of price swings. Thus, such a plan provides a means by which to deal with two of the most pressing economic issues facing trade-dependent developing nations. This is probably the key argument in evaluating the choice of a commodity-based currency, versus simply turning to a currency board like so many small nations have done. Both systems provide exchange rate stability, and both have the acknowledged “drawback”2 of eliminating activist monetary policy. Yet, commodity market instability is an issue that developing nations have been trying to find a solution to for decades. Both commodity money and currency boards would provide a high degree of credibility. Clearly, for those nations that are not dependent on a commodity (or group of commodities), currency boards would make more sense. 7. Why the cocoa market? A number of commodities are potentially suitable as collateral for a new currency. It is essential that the chosen commodity fulfill the following conditions: - Homogeneity. - Non-perishability—must be storable in order for a buffer stock to be maintained. - Low price elasticity—reduces the extractions from market supply that will come about when excess supply prevails.

2 Many would argue that the loss of control over monetary policy is an advantage, given the dismal record of monetary authorities in many developing nations. Both currency boards and commodity money prevent the excessive monetary growth that was characteristic of the Latin American economies in the 1980s.

48

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

- Barriers to entry in production—the new-found value of the commodity in terms of backing a currency will result in an incentive to initiate production, when possible, potentially reducing the price of the commodity. Although coffee was used for illustration purposed in the previous section, it is imperfectly suited to act as backing for a commodity-based currency. Coffee fails the homogeneity requirement, making it difficult to establish what exactly would be backing currency units issued by producing nations. In particular, coffee is commonly graded into two categories, Arabica and Robusta, with the former selling at a significant premium in the market. Within the Arabica classification, there is a breakdown into numerous other “premium” categories. Adoption of a coffee-based currency would result in an enormously complicated system of valuations for these various grades. Conversely, cocoa appears to be an ideal choice. Production is confined to nations with a tropical climate, and the long lead-time (in excess of five years) between planting and the first crop also discourages entry into the market (Table 2 presents current production figures). There is no viable substitute for cocoa, resulting in a price elasticity of approximately 0.09 (International Cocoa Organization, 2002). Cocoa is non-perishable, at least in the near-term, and cocoa beans are regarded as relatively homogeneous in quality. It should be noted that the characteristics of cocoa outlined above (low elasticity, homogeneity, barriers to entry, etc.) are very similar to those for gold. The very long history of using gold as a basis for a fixed exchange rate system demonstrates the workings of such systems. The difference, of course, is that gold is not consumed in the same manner that a commodity like cocoa is consumed, and this was one of the primary thrusts of Graham’s proposal; that the mechanism

Table 2 Cocoa supply, demand and resulting stocks, 000s metric t, 1981–2000 Year

Beginning stocks

Production

Consumption

Change in stocks

1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

510 603 698 655 463 538 617 692 871 1192 1384 1486 1338 1334 1326 1181 1409 1321 1137 1211

1649 1709 1535 1526 1923 1946 1981 2177 2446 2396 2435 2300 2360 2502 2386 2907 2693 2644 2803 2937

1556 1614 1637 1718 1848 1867 1906 1998 2125 2204 2333 2448 2441 2485 2507 2650 2753 2802 2758 2924

93 95 −102 −192 75 79 75 179 321 192 102 −148 −81 17 −121 257 −60 −158 45 13

Source: International Cocoa Organization.

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

49

provide stability in the commodity market(s) as well as the currency market (see details in next section).3 8. Modeling a commodity-backed currency One of the most attractive features of Graham’s original plan was its simplicity. A commodity “dollar” could be redeemed for a fixed quantity of the commodity basket. If the price of a particular commodity rose, redemption of commodity dollars would occur automatically, since the value of the commodities now exceeded the value of the currency. This would automatically increase the quantity of the commodity in circulation, driving back down its price. Consequently, the value of the currency is never altered, and the price of the commodity is stabilized at its target level. The application of Graham’s plan to a single commodity, such as cocoa, is far simpler. Suppose, the cocoa market can be described as: -

Pt = market price in absence of the buffer stock P ∗ = target price for cocoa Qt = output in year t Dt (Pt ) = demand in year t, exclusive of buffer stock εc = elasticity of cocoa demand St = stockpile in year t δ = carrying cost per unit Cu = number of units of cocoa in each unit of currency P• = percentage change in price (market versus target).

Under conditions of excess supply, Pt < P ∗ , and the value of currency unit is less than its target value: Pt Cu < P ∗ Cu. Arbitragers will then trade cocoa for cocoa currency, reducing the available supply until equilibrium in the market is reached. As the value of the cocoa currency does not change, currency stability is achieved. The opposite exercise can be carried out for situations where demand exceeds supply. One of the fundamental criticisms of Graham’s plan was the significant cost of maintaining the stockpile. The low elasticity of demand for cocoa (0.09) indicates that the required purchases of cocoa would be quite small. If, as indicated above, the long-term buffer stock holdings are St , then the carrying cost before any increase in stocks is: δSt . The automatic increase in the quantity held by the buffer stock is equal to: {εc P • Dt (Pt )}

(1)

3 Although the immediate reaction might be that cocoa money, with its similarities to gold-backed currency, might be open to speculative attacks, such as those that occurred against the U.S. dollar in 1971 and 1973. Yet, cocoa is not gold, and it inconceivable that speculators would seek to accumulate vast stockpiles of such a commodity in anticipation of some form of devaluation, which could technically only occur if the nations participating in the regime literally ran out of cocoa.

50

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

Where variables are as defined above (since εc P• = the percentage change in demand).The carrying cost in the subsequent year would then be: δ [St + {εc (P • )Dt (Pt )}]

(2)

The low elasticity of demand results in carrying costs that are likely to be small. The other major concern is the financial resources that must be available for the swapping of cocoa for CUs. In the long run, if the target price is properly set, these should be minor. In the short run, however, the buffer stock must be able to absorb significant quantities of cocoa during years when production markedly exceeds demand. The excess production is defined as: Qt − Dt (P ∗ ). (Note that demand has been reduced/increased from its observed value due to the alteration of the price through the currency arrangement.) The exchanging of cocoa for currency, at a fixed price of P ∗ will “cost” the buffer stock: Ca =

n 

(P ∗ qi )

(3)

i−1

where Ca = cost of acquiring excess cocoa, qi = quantity exchanged for currency in period i and

n 

qi = {εPDi (Pi )}

i=1

from Eq. (1) above In the short run, the minimum financial resources of the buffer stock would have to equal the cost of the stockpile (acquired at the target price), plus the cost of additionally acquired cocoa and the expense of carrying both the stockpile and any acquisitions:  PSt +

n 

 ∗

(P )qi

(1 + δ)

(4)

i=1

where δ = annual per unit carrying cost. In contrast to a standard buffer stock plan, no profit would accrue due to the acquisition and sale of cocoa. The established price of cocoa is fixed at P ∗ , so purchases and releases of cocoa would take place at an identical price. This increases the cost of such a plan, as no profit is accrued on the divergence between target and market prices. An important aspect of commodity-based currency plans, is that the nation(s) involved would not have to hold sufficient quantities of cocoa to back all issued currency; any more than countries under the old gold standard had enough gold to redeem all issued paper currency. The monetary authorities would simply have to agree to back any exchange of CUs for cocoa, and vice-versa. As with any currency that has intrinsic value, this could open CUs to speculation if it was rumored that a devaluation or revaluation was imminent due to a shortage or surplus of cocoa (as is the case with any hard peg system). It would be up to the monetary authority to avert such a situation by securing the financial resources to back the system.

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

51

Table 3 Market price of cocoa, 1972–2000, Brazil (pre-shipment) Year

Dollars per metric t ($)

Year

Dollars per metric t ($)

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986

577 1067 1614 1245 1694 4038 3378 3096 2356 1859 1052 1828 2317 2089 2025

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 1999 2000

1847 1599 1251 1080 1046 990 978 1230 1330 1396 1603 1652 1170 1170 919

Source: International Cocoa Organization.

9. An illustration The most challenging aspect of the plan outlined above would be determining the appropriate price of cocoa. As noted, Graham suggested a 10-year average would be a reasonable choice, although this was based on the assumption that long-run prices were stable, a condition that does not necessarily hold for the cocoa market. Table 3 provides price data for cocoa from 1972–2001. It is obvious from these figures that cocoa prices exhibit a long-term decline, beginning in 1978. Hence, it may be necessary to establish a pricing mechanism that takes into account cocoa’s declining terms of trade. Three simulations will be carried out, each utilizing a different method for determining the appropriate price of cocoa. First, a perfect foresight model (average of the 1991 through 2000 price) will be used, primarily to demonstrate how, if price is properly set, market stability is the natural result. Second, the simple mean of cocoa prices during a base period (1980 through 1990, as suggested by Graham) will be used to establish the base value of currency units during the subsequent period (1991 through 2000). Noting the recent downward trend in prices in the 1990s, the third estimation will utilize the long-term price for cocoa (1964–90). It is believed this is a better measure of the equilibrium price of cocoa, since observations in the 1980s continued to reflect the commodity price inflation of the 1970s. What will not be used are the logical (but, in the end, wrong) methods of forecasting prices. Preliminary testing on forecasting models (versus a time trend or against models utilizing factors of demand) produced poor results. In general, these methods resulted in: - Forecasted prices than were less than zero in the later years of the estimation. - Severe problems with autocorrelation. Once a target value is determined for each experiment, cocoa demand will be estimated at that price using the known demand elasticity of 0.09. The quantity that must be extracted from

52

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

Table 4 Results of simulations, actual and projected surplus, and cost to maintain fixed value of CUs, 1991–2000, in millions of dollars Simulation method

Price utilized

Year

Surplus (+)/short-fall (−) (metric t)

Surplus/short-fall @ target P

Cost @ target price

Perfect foresight

1231.4

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

102 −148 −81 17 −121 257 −60 −158 45 13

133.6 −104.8 −35.9 17.3 −139.1 225.1 −134.7 −244.1 57.4 79.7

164.6 m −129.0 −44.1 21.2 −171.3 277.2 −165.9 −300.6 70.7 98.2 −179.1

Total 1981–90 average

1739.7

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

102 −148 −81 17 −121 257 −60 −158 45 13

185.8 −53.0 15.2 82.5 −67.9 304.1 −40.5 −145.3 126.3 137.1

323.2 −92.2 26.4 143.6 −118.1 529.1 −70.4 −252.7 219.7 238.6 947.1

1460.3

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

102 −148 −81 17 −121 257 −60 −158 45 13

161.6 −77.0 −8.5 52.2 −100.9 267.5 −84.2 −191.2 94.3 110.5

235.9 −112.5 −12.4 76.3 −147.4 390.6 −123.0 −279.2 137.7 161.3

Total 1964–90 average

Total

327.4

the market is then the known supply minus the projected demand. In each simulation, it will be assumed that the existing stock of cocoa becomes the basis for the currency. This would ensure the adequate issuance of CUs, and eliminate the chance that the excess stock would unravel the plan right from the beginning. An aspect that was missing from Graham’s original plan was an allowance for monetary growth. Even the most conservative of monetarists would argue that a minimum of 2.5–3% must be added to the money supply annually to allow for real economic growth. Hence, the last two simulations will be conducted a second time, with a 3% increase in the quantity of CUs built into demand. The results of these simulations are presented in Tables 4 and 5.

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

53

Table 5 Re-estimation of cost with 3% money supply growth, in millions of dollars Simulation method

Price utilized

Year

Change in stocks to raise Ms (metric t)

Required additional changes (metric t)

Cost @ target price

1981–90 average

1739.7

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

41.5 42.8 44.0 45.4 46.7 48.1 49.6 51.1 52.6 54.2

58.8 70.6 71.4 33.4 17.4 8.5 −26.2 −5.8 45.9 95.5

102.3 122.9 124.2 58.2 30.3 148.1 −45.5 −62.2 79.9 166.1

Total 1964–90 average

Total

592.6 1460.3

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

41.5 42.8 44.0 45.4 46.7 48.1 49.6 51.1 52.6 54.2

29.9 42.0 43.0 −3.0 −22.5 −35.5 −78.6 −90.9 7.2 63.4

43.6 61.3 62.8 −4.3 −32.8 −51.8 −114.7 −132.7 10.5 92.6 −65.5

10. Analysis As noted, simulation 1 utilized a perfect foresight price for cocoa ($1231.4/metric t) for the period 1991–2000. The figures in Table 4 demonstrate the resulting withdrawals (+) or injections of cocoa that are necessary to maintain the target price (the actual surplus or shortfall in each year is given to provide a basis for comparison). In the early part of the period (1991–4), excess production of cocoa exceeds what would have occurred at actual market prices (133.6 metric t versus 102 metric t in 1991, for instance). This is offset by the rapid deaccumulation of cocoa in 1997 and 1998. The maximum pay-out to maintain the target of $1231.4/metic t is in 1996 ($277.2 million). These resources are recaptured in 1998 when $300 million in CUs would be turned in for cocoa. Overall, the accumulated “cost” to the buffer stock is −$179 million.4 Hence, if prices could be perfectly forecast, the resources necessary to make a commodity-backed currency of this type function would be minimal. Simulations 2 and 3 provide a more realistic assessment of a commodity-backed currency, as the target price in each instance is based upon a medium- or long-term average of past prices. 4 Although it may appear that over the entire period the net expenditure should be zero, this is not the case. Prices tend to only partially reflect the size of the surplus/shortfall in any given year. Hence, a large shortfall may be accompanied by only a modest rise in prices. The redemption of CUs would then be small, despite the relatively large variance between supply and demand.

54

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

The average value of cocoa over the 1981–90 period (the method suggested by Graham) was $1739.7/metric t. Not surprisingly, since this exceeds the average price in the 1991–2000 period by approximately $500 per t, this results in a sizable surplus in several years, one in excess of $500 million. The releases of cocoa from the buffer stock, particularly in 1998, are insufficient to offset this, and the resulting accumulation is quite large. By 2000, the buffer stock would have expended nearly $950 million, and accumulated an additional 544,000 metric t of cocoa. It is unlikely that resources of this magnitude could be devoted to a plan such as this by small developing nations. Simulation 3, which utilizes the long-term price of cocoa ($1460.3/metric t), is far more realistic. The largest single outlay of currency units is in 1996, when $406.3 million in additional CUs would have to be issued. Yet, this is nearly offset by redemptions in 1998 of $266.7 million. During the ten-year period, only $278.0 million in accumulated expenditures would be necessary; a figure that is manageable. This simulation demonstrates that, with properly chosen prices, both currency stability and commodity price stability can be achieved through such a plan. Table 5 demonstrates how the situation changes if monetary growth is built into the model. As noted, a conservative change of 3% per year has been utilized. Using a target price of $1739.7 (as in simulation 2 above), column four (change in stocks to raise Ms ) indicates the amount of cocoa that would have to be withdrawn, with compounding, to maintain a 3% rise in the money supply. Column 5 presents figures for the additional withdrawals that would be necessary to achieve the target price, which are now predictably smaller. The largest release of CUs is now only $148.1 million (1996), and the accumulated cost has been reduced to $592.6 million; still a somewhat problematic sum. In contrast, a re-estimation of simulation 3 (target price of $1460.3) demonstrates that the system would experience a net release of cocoa over the period, suggesting that carrying costs are now the only concern. Thus, in the presence of mandated monetary growth, the system is potentially workable. 11. Resources The simulations presented above illustrate the workings of a commodity-backed currency, and at the same time, assess the resources necessary to support such a plan. Clearly, it would be difficult to support Graham’s original contention that a ten-year average be utilized as the basis for the target price. The annual commitments of resources are large, and the accumulated commitment over ten years is unmanageable. Conversely, simulation 3 (without monetary growth) resulted in a required outlay that is easily within the means of a consortium of developing nations. With the introduction of controlled monetary growth, even simulation 2 is potentially viable. The cumulative outlay of nearly $600 million must be placed in the context of the expenditures now being spent on developing nations through debt forgiveness and IMF lending programs. As these recurring debt crises are at least partly attributable to instability in commodity markets, it would seem to be in the best interests of the developed world to provide at least some resources to make such a plan work. This is particularly true if it could be asserted that future payments problems would likely be avoided in the presence of stable currencies and stable commodity prices. As noted above, one of the practical criticisms of Graham’s plan is the burden of the carrying costs of the accumulated cocoa. Estimates of these costs ranged from 3% (Hart et al., 1963) to an extreme value of 6% (Grubel, 1966) of the value of the stockpile in any given year, a not insubstantial figure. In the perfect foresight model, the net accumulation is at its highest value in the first year, $164.5 million. Using an estimated range of 3–5% (assuming Grubel’s figure is

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

55

probably somewhat too high), the carrying cost would range from $4.9 to $8.4 million (similarly, in 1996, the only other year with substantial accumulations, the figures are $3.6 and $5.9 million). Estimation 3, which utilizes the average price from 1964 through 1990, the comparable figures for 1996 are $15 and $25.2 million. Hence, in the absence of imbedded monetary growth, the cost of maintaining the stockpile becomes somewhat problematic. Conversely, with monetary growth, and the utilization of long-run prices, the carrying costs are no longer a concern (simulation 5). The largest accumulation is in 1993, with $167.7 million in cocoa having been added to the stockpile. This translates into $5 million (3% carrying cost) to $8.4 million (at 5%), figures that are not large enough to automatically derail such a plan. 12. Monetary instability As noted, the perfect currency regime would provide stability, the opportunity for independent monetary policy and an environment conducive to international capital flows. Although a commodity-backed currency achieves two of these goals, it does not permit independent monetary policy. In fact, if production of cocoa were to become erratic, the resulting variability in the redemption/issuance of CUs would result in destabilizing changes in the money stock. In simulation 2, one could certainly regard several of the injections/withdrawals are excessively large. It should be noted that Graham’s rejoinder to such criticism does not apply here. Graham noted that in the presence of excessive monetary growth, inflationary pressures would drive up the imbedded value of the CUs, leading to their redemption. This would reduce the monetary stock, reversing the trend. Cocoa prices are established in international markets, and are largely independent of the monetary stock in producing countries. Thus, there is not comparable mechanism for bringing back stability. For simulations 3 and 5, the most plausible of those presented, the exchange of cocoa for CUs might lead to a rise in the money supply as great as $400 million for the former, and a fall in money supply of $133 million for the latter. The combined money supply in the three nations that would most likely engage in the adoption of cocoa CUs (Ghana, Cameroon and Cote D’Ivoire) was approximately four billion U.S. Dollars in 1990. The $400 million rise in the money supply under simulation 3 is clearly too great, although it would be somewhat offset by the 3% growth in the money supply that would be desired anyway. Conversely, the maximum change under simulation 5 does not pose a problem, as $133 million is just 3.3% of the existing money stock. 13. Conclusion The concept of commodity-backed money is one that will continue to be raised during periods of monetary instability, whether during the high inflationary decade of the 1970s or during the more recent international currency crises in the developing world. As nations in Latin America, Africa, and the Pacific Rim abandon fixed rates in search of a regime that provides stability, commodity-based currency deserves examination. This paper provides an analytical look at one possible form that such a regime could take. While Benjamin Graham’s original work foresaw the use of commodity money in the developed world, the attractiveness of flexible exchange rates, particularly the ability to conduct independent monetary policy, suggests that this is unlikely. Yet, for small, commodity-dependent countries, flexible rates are now simply the best of a set of bad choices. The model developed here utilizes the basic framework set up by Graham to describe a currency unit that could be appropriate for these developing nations.

56

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

It has been demonstrated that a properly designed currency unit, backed by a product like cocoa, could simultaneously provide both currency and commodity price stability. The resources required to adopt and manage such a system are within the means of the principal cocoa producers, and, as argued in Section 12, could also be provided by the industrialized world at a fraction of what is now being spent on debt-forgiveness. As noted, commodity money does not possess all of the positive traits that should characterize an ideal currency (the holy trinity). The most troubling failure would be the potential for erratic and destabilizing monetary policy. The numbers presented here suggest that it is unlikely that such a difficulty would develop if the target price is appropriately chosen, as the resulting swings in the money supply are relatively small. References Anderson, D., 1938. Review of Graham’s storage and stability. Journal of Political Economy 46, 435–437. Beale, W., Kennedy, M., Winn, W., 1942. Commodity reserve currency: a critique. Journal of Political Economy 50, 579–594. Bennett, M., 1949. International Commodity Stockpiling as an Economic Stabilizer. Stanford University Press, Stanford. Calvo, G., Reinhart, C., 2002. Fear of floating. Quarterly Journal of Economics 117, 379–408. Carramazza, F., Aziz, J., 1998. Fixed or flexible: getting the exchange rate right in the 1990s. IMF Economic Issues 13. IMF, Washington, DC. Edwards, S., Savastano, M., 1999. Exchange rates in emerging economies: what do we know? What do we need to know? NBER Working Paper 7228. Eichengreen, B., 1994. International Monetary Arrangements for the 21st Century. Brookings Institution, Washington, DC. Eichengreen, B., Masson, P., Bredenkamp, B., Johnson, B., Haman, J., Jadresic, E., Orker, I., 1998. Exit strategies: policy options for countries seeking greater exchange rate flexibility. IMF Occasional Paper 68. Fisher, S., 2001. Exchange rate regimes: is the bi-polar view correct? Speech given at the American Economic Association annual meeting, New Orleans, January 6, 2001. Frankel, J., Fajnzylber, E., Schmukler, S., Serven, L., 2001. Verifying exchange rate regimes. Journal of Development Economics 66, 351–386. Friedman, M., 1951. Commodity reserve currency. Journal of Political Economy 59, 203–232. Graham, B., 1997. Storage and Stability: A Modern Ever-Normal Granary. McGraw-Hill, New York (reprint). Graham, F., 1938. Review of Graham’s storage and stability. American Economic Review 28, 575–577. Graham, F., 1941. Transition to a commodity reserve currency. American Economic Review 31, 520–525. Greenfield, R., Yeager, L., 1983. A laissez-faire approach to monetary stability. Journal of Money Credit and Banking 15, 302–315. Grubel, H., 1966. The case against an international reserve currency. Oxford Economic Papers 17. Hall, R., 1982. Explorations in the gold standard and related policies for stabilizing the dollar. In: Hall, R.E. (Ed.), Inflation: Causes and Effects. University of Chicago Press, Chicago. Hart, A., Kaldor, N., Tinbergen, J., 1963. The case for an international commodity reserve currency. UNCTAD, Geneva. Haussmann, R., 1999. Should there be five currencies or one hundred and five? Foreign Policy 116, 65–79 (fall). Hayek, F., 1943. A commodity reserve currency. Economic Journal 53, 176–184. Keynes, J., 1980. On commodity control. In: Moggridge, D. (Ed.), Collected Writings of John Maynard Keynes, Vol. XXVII. MacMillan, London. Keynes, J., 1943. On the subject of price stability. Economic Journal 53, 185–187. Kopcke, R., 1999. Currency boards: once and future monetary regimes. New England Economic Review (FRB Boston), 21–37. Levy-Yeyati, E., Sturzenegger, F., 2000. To Float or to Trail: Evidence on the Impact of Exchange Rate Regimes. CIF Working Paper No. 01/2001. Universidad Torivato Di Tella, Buenos Aries. Rose, A., 1996. Explaining exchange rate volatility: an empirical analysis of ‘the holy trinity’ of monetary independence, fixed exchange rates, and capital mobility. Journal of International Money and Finance 15, 925–945. Rosensen, A., 1948. International commodity reserve currency reconsidered. Review of Economics and Statistics 30, 135–140.

M.S. LeClair / Int. Fin. Markets, Inst. and Money 17 (2007) 42–57

57

Simmons, H., 1945. Review of Benjamin Graham’s world commodities and world currency. Journal of Political Economy 53, 279–281. Summers, L., 2000. International financial crisis: causes, prevention, and cures. American Economic Review 90, 1–16. White, L., 1984. Competitive payment systems and the unit of account. American Economic Review 74, 699–712. White, L., 1986. Competitive payments systems: reply. American Economic Review 76, 850–853. Yagci, F., 2001. Choice of exchange rate regime for developing countries. World Bank-African Region Working Paper Series 16, Washington, DC.