The "Costs" of a Gold Standard

Roger W. Garrison
 

I. Introduction 
The term "gold standard," whether used in an historical or a theoretical context, can mean many different things. And for each meaning of this term, a reference to the "costs" of a gold standard will not be unambiguous. Yet it is commonly believed, by economists and laypersons alike, that cost considerations eliminate gold as a viable medium of exchange in modern-day economies.
        The purpose of this paper is to examine the arguments against the gold standard which are based on considerations of costs. Section II identifies the benefits of a gold standard in order to put the discussion of costs into proper perspective. Section III compares two conflicting views of the gold standard and of the resources devoted to maintaining it. Section IV deals with actual estimates of the resource costs of gold. Sections V and VI employ the more broadly conceived concept of "opportunity costs" to argue the irrelevance of "resource costs" to the comparison of alternative monetary institutions. Section VII calls into question the assumed equivalence of monetary stability and price-level stability. This assumption, which underlies many of the cost estimates, has clouded some fundamental issues in ways that have prejudiced both monetary theorists and policymakers against the gold standard. Section VIII provides a summary assessment. 

II. The Gold Standard: Costs and Benefits
Any discussion of the costs of a gold standard and of the controversy that surrounds this issue is, by its very nature, a one-sided discussion. The comparison of alternative standards on the basis of costs will not be meaningful unless the corresponding benefits are brought into view. Spelling out the particular type of gold standard being discussed and identifying its benefits—in comparison to a paper standard—puts the cost comparisons into proper perspective and goes a long way toward justifying the costs.
        The term "gold standard" in the present paper is used to denote the outcome of a market process. Using the term in this way serves to consolidate at least three propositions—based on both economic theory and historical insight—about the nature of markets and about the nature of money. (1) Left to its own devices, a market economy will give rise to medium of exchange.⁽¹⁾ (2) The commodity that emerges as the medium of exchange will be one that possesses a certain set of characteristics.⁽²⁾ (3) This set of characteristics has its clearest and most pronounced manifestation in gold.⁽³⁾ So conceived, the gold standard—at least in its purest form—neither requires nor permits the State to exercise control over the money supply. And as argued below, the absence of centralized, discretionary monetary control constitutes the primary benefit of the gold standard.
        The perception by the layperson that the costs of a gold standard are "too high" is not difficult to understand. Under a gold standard suppliers of goods or of labor services exchange their supplies for gold (or for bank notes redeemable in gold) not because the gold standard is seen as having great merit, but because gold is the customarily accepted medium of exchange. To each market participant gold per se has no particular benefits, although the custom of accepting some specific thing does. When consciously pondering the nature of money, the layperson is likely to see the custom in a different context and to see the value that others attach to gold—or that "the market" attaches to gold—as "irrational," as being based on superstition or mythology. Gold in the layperson's view is a "barbarous relic" (to use Maynard Keynes's phrase). Yet individuals in modern economies continue to devote resources to securing this shiny yellow metal. Could not some other custom have the same benefits without having such high costs?
        Market-oriented economists adopt a third view of the gold standard—one that differs from the views of both the market participant and the layperson pondering the gold question. The economists see the difficulties—and costs—of replacing an evolved custom with a designed system. The differences among such economists stem from the differing estimates of the nature and magnitude of these difficulties.
        Economists who oppose the gold standard may recognize what has to be achieved in order to make a centrally controlled paper standard superior to a decentralized gold standard. Milton Friedman poses the key question: "[H]ow ... can we establish a monetary system that is stable, free from irresponsible tinkering, and incapable of being used as a source of power to threaten economic and political freedom?"⁽⁴⁾ How, in other words, can we design a system that we cannot tinker with? While Friedman goes on to suggest how such a system might be designed, economists who support the gold standard argue that this objective is self-contradictory and operationally impossible. Any monetary institution that is designed and implemented by a central authority can be abused by that central authority.
        The proponents of gold are not suggesting that irresponsible tinkering is inevitable—whatever the nature of the monetary system; they are instead making the sharp distinction between a designed institution and an evolved institution. A monetary institution that has been consciously designed is much more subject to tinkering than one that simply emerged as a consequence of market processes. F. A. Hayek points to the positivist slogan that "what man has made he can also alter to suit his desires."⁽⁵⁾ The positivists were correct so long as they were referring to consciously and deliberately designed institutions such as a paper standard. Of course, it is government officials (not "man") who design the paper standard, and it is government officials who can (and do) tinker with it. Hayek goes on to point out the limits of the positivists' view. The slogan is a "...complete non-sequitur if `made' is understood to include what has arisen from man's actions without his design."⁽⁶⁾
        A gold standard—one that has emerged as a consequence of market processes—cannot easily be altered to suit the State's purposes. It would be an overstatement (and a matter of historical inaccuracy) to claim that the State cannot even in the long run interfere with the operation of the gold standard. What is true (both theoretically and historically) is that the State can supplant a spontaneously evolved monetary system with a centrally controlled system only after a prolonged struggle in which it must slowly and gradually overcome (through propaganda and the use of coercion) the market's reluctance to abandoned gold. It is the gold standard's substantial immunity from State manipulation and tinkering, and not the associated superstition and mythology, that recommends gold as a monetary standard. In the words of Ludwig von Mises, "the advantage of the gold standard ... is due solely to the fact that, if once generally adopted in a definite form, and adhered to, it is no longer subject to specific political interference."⁽⁷⁾ In the judgment of the proponents of the gold standard, the benefits of gold, immunity from State intervention and the resulting monetary stability, outweigh the resource costs of gold—and any other costs that might be associated with the gold standard—by a comfortable margin. 

III. The Resources Devoted to Gold: Too Few and Too Many
Discussions of the gold standard typically gravitate toward a consideration of the amount of resources used up in the maintenance of it. Well recognized market processes will devote a certain amount of resources to the gold-mining industry, sometimes more resources, sometimes fewer, depending upon market conditions in the rest of the economy. Changing market conditions have both price effects and quantity effects that come into play. Consider, for instance, an increase in the demand for money brought about by a desire on the part of market participants for greater liquidity. This demand shift puts downward pressure on prices. Because the actual adjustment in prices is not immediate, the increased monetary demand will have a temporary effect on quantities as well. Excess supplies of goods and of resources—both labor and capital—will develop. In general, the more rapidly the prices adjust, the less pronounced the temporary adjustment in quantities, and conversely, the more slowly they adjust, the more pronounced the adjustment in quantities.
        The adjustment process is facilitated in part by changing market conditions in the gold-mining industry and in supporting industries. In these markets, movements in prices and quantities are opposite in direction to the movements in markets for goods that exchange against money. Downward movements of prices in general mean an increased value of the monetary commodity; excess supplies of labor and capital mean an increased availability of resources for mining gold. Both the price and the quantity effects stimulate the production of the monetary commodity and in the process relieve the pressure that gave rise to the stimulation. The final result is that the increased demand for money is accommodated in part by an actual decline in prices and in part by an increased quantity of the monetary commodity. The relative size of the two accommodating factors depends upon the supply conditions in the gold-mining industry.
        Much of the dissatisfaction with the gold standard stems from dissatisfactions with the quantity of resources devoted to the extraction and processing of gold. Paradoxically, some opponents of gold believe that too few resources are involved for the gold standard to be viable, while other opponents believe that too many resources are devoted to the mining of gold. Not surprisingly, these opposing opponents of gold are reasoning in markedly different ways.
        The first line of reasoning is based on the assumption that prices are extremely "sticky," and hence that all adjustments to changing market conditions are quantity adjustments. An increased demand for money means a decreased demand for goods. Since the goods cannot all be sold at existing prices, surpluses pile up, production is curtailed, and workers become unemployed. An economy-wide depression sets in. The only excess demand in the economy is for the monetary commodity. But because of the nature of gold—its relative scarcity—the gold-mining industry can absorb only a small fraction of the unemployment. The demand for money cannot be fully met at the existing level of prices. If the gold-mining industry absorbed the same amount of resources that were unemployed as a result of the increase in the demand for money, then the gold standard would perform admirably in this view and would constitute an automatic countercyclical device. Employment could shift from goods to gold or from gold to goods, but the level of employment would remain unchanged. Unfortunately, the gold-mining industry does not employ enough labor and capital resources to provide for such economic stability.⁽⁸⁾ This line of reasoning has even caused one monetary reformer to advocate the abandonment of gold and the adoption of the common clay brick as a monetary standard.⁽⁹⁾
        The other line of reasoning considers the alternative of a centrally directed system of paper money that can mimic the countercyclical effects of a clay-brick standard but without devoting any resources at all to the production of clay bricks or to the mining of gold. Each increase in the demand for money could be met with a costlessly produced increase in the quantity of money supplied. An economy whose transactions are facilitated by such a managed paper money would never experience an economy-wide downward pressure on prices that could result in resource idleness. Thus, the economy could devote all its resources to the production of real (non-monetary) output. With this possibility in mind the allocation of any of the economy's resources to the production of gold is seen as wasteful, and as constituting too many resources.⁽¹⁰⁾
        Proponents of the gold standard should not feel called upon to argue in the context of either of these two lines of reasoning that the quantity of resources actually devoted to gold is just enough but not too much. Practically any quantity would at the same time be too little and too much—depending upon the opponent's particular point of view. Both viewpoints, however, can be called into question by an examination of the meaning and relevance of key concepts used by each. Particularly critical to the issue of the gold standard are the concepts of costs, resource costs, price stability, and monetary stability. These and related concepts provide a focus for the remainder of the present paper. 

IV. The Costs of Gold and the Costs of a Constant Price Level
Estimates of the resource costs of gold depend critically upon the assumed rate of extraction. The actual rate of extraction, as indicated above, would be determined by market conditions. In an expanding economy with given supply conditions for gold, an increasing demand for money would cause additional resources to be committed to gold-mining operations. If competitive forces in the banking industry have given rise to the circulation of redeemable bank notes, the actual shift in the demand for gold caused by the expansion would be significantly reduced. The additional quantity of resources committed to gold would depend upon the elasticity of supply and the magnitude of the demand shift. Gold's relative inelasticity of supply would ensure that the dominant effects of the increase in the demand for gold, whatever its magnitude, would be a price effect rather than a quantity effect. That is, the value of gold would rise, or conversely, the prices of other goods would fall with respect to gold. There would be some increase in the quantity of gold supplied, but due to the price effect, this increase would be small in comparison to the increase in demand. The resource costs of extracting the additional gold would be correspondingly small.
        Unfortunately, the most commonly cited estimates of the resource costs are based on the assumptions that there is no circulation of bank notes and that there is no price effect at all. Further, the supply of gold is assumed to be perfectly elastic.⁽¹¹⁾ Increases in the demand for money, under these assumptions, are met in full with increases in the quantity of gold supplied. The rate of gold extraction, in other words, is assumed to be sufficiently large to offset totally the downward movement of prices that would otherwise be necessary in an expanding economy. The fact that the supply of gold is actually inelastic is simply brushed aside. The resulting estimate of the resource costs, then, is not an estimate of the costs of a gold standard at all but rather an estimate of the costs of maintaining a constant price level by adopting an elastically supplied commodity money.
        Not surprisingly, actual estimates that are based on these assumptions show that the costs of a commodity money are quite high. Neglecting changes in the velocity of money, Friedman calculated that for the first half of the twentieth century, the resource costs of a pure gold standard would have amounted to about one-and-one-half percent of national income, or about one half of the annual growth rate of output. (Velocity considerations would increase these figures to two percent of national income and two-thirds of the annual growth rate.)⁽¹²⁾ These particular estimates are three decades out of date, but the estimating procedure is in current use. Allan Meltzer cites the Friedman estimate and then updates the figures to reflect the current ratio of money to income. The new calculations indicate that the costs are down from fifty percent of the annual growth rate to something like sixteen percent. But the cost of a gold standard in Meltzer's own judgment "remains high."⁽¹³⁾
        The estimating procedure adopted by Friedman and more recently by Meltzer is flawed on both positive and normative grounds. The positive analysis makes use of the classical long-run perspective in which all supply curves are perfectly elastic. Friedman notes explicitly that his cost estimate is independent of which commodity is used as the monetary standard.⁽¹⁴⁾ Perfect supply elasticity is a particularly inappropriate assumption when the gold standard is at issue. The supply of gold is inelastic in the short run because of the increasing marginal costs of extraction and inelastic in the long run because of the natural scarcity of this particular element. Friedman's and Meltzer's calculations fail to take into account these particular supply considerations which help to qualify gold as a monetary commodity. Indeed, they fail to make any distinction whatsoever between gold and all other commodities.
        The normative judgment upon which their cost calculations are based is the judgment that the maintenance of a constant price level over time is an undisputed desideratum and the appropriate basis for evaluating alternative monetary arrangements. The significance of a constant price level in this regard is the focus of the penultimate section of the present paper. The following two sections distinguish between two different cost concepts and question the use of resource costs as a criteria for choosing among alternative monetary institutions. 

V. Costs, Resource Costs, and the Gold Standard
It was demonstrated above that commonly cited estimates of the resource costs of gold are based on untenable assumptions about the supply conditions in the gold-mining industry and about the desired behavior of the price level. The present section is concerned with the relevance of any estimate of resource costs to the comparison of gold and paper standards. It is argued that the resource costs are doubly irrelevant in assessing the relative merits—and the relative costs—of the two alternative standards. The critical issues are likely to be overlooked if there is a failure to distinguish between (1) the resource costs of gold and (2) the costs of a gold standard. The two cost concepts are totally dissimilar despite the similarity in the verbiage. This section deals with the costs of the gold standard and of paper standards over and above the narrowly conceived resource costs; the following section puts the resource costs of gold into proper perspective.
        So-called resource costs are an inadequate proxy for total costs or opportunity costs—unless the former term is defined in such a way as to make it synonymous with the latter two, in which case the modifier "resource" becomes redundant and misleading. The inadequacy is especially pronounced when the issue is the relative costs of alternative institutional arrangements.⁽¹⁵⁾ A penal system that segregated convicted criminals from the rest of society may involve more "resource costs" than one that only slapped the criminals' wrists and turned them back into society. But it would be difficult to argue that the total costs, which would have to take into account the subsequent crimes perpetrated by convicted criminals, are greater for the former institutional arrangement than for the latter.
        There is a similar difficulty in the argument that a gold standard costs more than a paper standard. Comparing the resource costs of gold to the resource costs of paper does not settle the issue. The true costs of the paper standard would have to take into account (1) the costs imposed on society by different political factions in their attempts to gain control of the printing press, (2) the costs imposed by special-interest groups in their attempts to persuade the controller of the printing press to misuse its authority (print more money) for the benefit of the special interests, (3) the costs in the form of inflation-induced misallocations of resources that occur throughout the economy as a result of the monetary authority succumbing to the political pressures of the special interests, and (4) the costs incurred by businessmen in their attempts to predict what the monetary authority will do in the future and to hedge against likely, but uncertain, consequences of monetary irresponsibility. With these considerations in mind, it is not difficult to believe that a gold standard costs less than a paper standard. The judgment that a gold standard is the less costly reflects the wisdom in a simile attributed to Alan Greenspan: Allowing the State to create paper money is like putting a penny in the fuse box. The resource costs of the penny may be lower than the resource costs of the fuse, but the total costs, which take into account the likelihood of a destructive fire, are undoubtedly higher.
        Some proponents of a paper standard base their counter arguments on their perception of the costs of a gold standard over and above the narrowly defined resource costs. But it is difficult to produce a laundry list of costs that will rival the list that was easily produced for the paper standard. The one cost most commonly cited stems from the fact that the supply of gold is not perfectly elastic, that gold production chronically falls short of real economic growth.⁽¹⁶⁾ This circumstance requires that the price level be continually adjusted downward as the ratio of money to real output steadily declines. And the market process by which individual prices become so adjusted can be time-consuming and costly. Monetary disequilibrium, in effect, gets translated into disequilibrium in all markets throughout the economy.⁽¹⁷⁾ Until equilibrium is eventually restored, the disequilibrium prices will result in the misdirection of some resources and the idleness of others. Under a paper standard, the monetary authority can eliminate the need to continually adjust prices by continually increasing the money supply to keep pace with the economy's real growth.
        At one level of abstraction, the opponents of the gold standard have an appealing, if not compelling, argument. But when the analysis penetrates beneath the issue of the changing price levels associated with the gold and paper standards, the argument all but disappears. At least three considerations are relevant here.⁽¹⁸⁾ First, individual prices in a market economy are changing all the time in response to changing market conditions. Some prices are increasing, some decreasing. If the supply of gold is not increasing as fast a real output, the pattern of individual price changes will be altered. Prices that are increasing will not increase quite so much; prices that are falling will have to fall a little further. Some prices that would have had to be increased a little will not have to be increased at all; some which would have remained unchanged will have to be slightly decreased. As a result of the altered pattern of price changes, the price level, the weighted average of all prices, will be lower. It is misleading, though, to associate the cost of price changes with a changing price level. Prices would have had to be changed in any event—though in a slightly different pattern.
        Second, even if we allow ourselves to abstract from individual price changes and think in terms of price levels, an elastically supplied currency will not eliminate the need for costly price adjustments. Consider, for instance, a growing economy in which the real rate of interest is declining. Which price level should the monetary authority keep constant: the consumer price level, the factor price level, or the general price level (which includes prices of both consumer goods and factors of production)? If the consumer price level is kept constant, then factor prices will have to be continually adjusted upward as the rate of interest falls; if the factor price level is kept constant, then the prices of consumer goods will have to be continually adjusted downward; if the general price level is kept constant, then the prices of both factors of production and consumer goods will have to be continually adjusted so as to reflect the declining interest rate. There is no price level whose constancy will eliminate the necessity for economy-wide adjustments in individual prices.
        Third, the alleged price-adjustment costs of a gold standard are identified by comparing the gold standard as it actually operates with a paper standard as it ideally operates. Such comparisons never provide a sound basis for choosing between alternative institutional arrangements. The comparison assumes away all the relevant costs of a paper standard. If paper standards were administered by angelic monetary authorities whose sole objective was to minimize money-induced disequilibrium, the choice between a gold standard and a paper standard would be much less consequential than it actually is. But actual paper standards have price-adjustment costs too. And as history teaches, the magnitude and costliness of upward price adjustments under a paper standard dwarf the magnitude and costliness of downward price adjustments under a gold standard.⁽¹⁹⁾

VI. The Unavoidable Resource Costs of Money
In the preceding section the contention was made that the resource costs are doubly irrelevant to the issue of alternative monetary standards. Total costs, which are poorly proxied by resource costs, are the appropriate bases for comparison. This section establishes the double irrelevancy by showing that while resource costs are only a fraction of total costs, they constitute a part of total costs that the economy incurs whether on a gold standard or a paper standard. That is, the resource costs of gold constitute a part of the costs of both standards, but all of the costs of neither standard. These costs, then, cannot be costs that influence the choice between the two monetary standards.
        The effectiveness of the resource-cost argument against the gold standard rests on the popular perception that the activities of mining gold, refining it, casting it into bars or minting it into coins, storing it, and guarding it are collectively wasteful activities and the implicit assumption that if the gold standard were supplanted by a paper standard, these activities would cease. But making the implicit assumption explicit is enough to demonstrate its falsity. The imposition of a paper standard does not cause gold to lose its monetary value. To believe otherwise is to hold the naive view that the State can repeal the laws of economics. Gold continues to be mined, refined, cast or minted, stored, and guarded; the resource costs continue to be incurred. In fact, a paper standard administered by an irresponsible monetary authority may drive the monetary value of gold so high that more resource costs are incurred under the paper standard than would have been incurred under a gold standard. Market processes operating since antiquity have identified gold as the premier monetary commodity, and until the market's adoption of an alternative standard causes the value of gold to fall to a level that reflects only the non-monetary uses of gold, these resource costs cannot be avoided.
        There is a certain asymmetry in the cost comparison that turns the resource-cost argument against paper standards. When an irresponsible monetary authority begins to overissue paper money, market participants begin to hoard gold, which stimulates the gold-mining industry and drives up the resource costs. But when new discoveries of gold are made, market participants do not begin to hoard paper or to set up printing presses for the issue of unbacked currency. Gold is a good substitute for an officially instituted paper money, but paper is not a good substitute for an officially recognized metallic money. Because of this asymmetry, the resource costs incurred by the State in its efforts to impose a paper standard on the economy and manage the supply of paper money could be avoided if the State would simply recognize gold as money. These costs, then, can be counted against the paper standard.
        As suggested earlier, resource-cost comparisons that favor paper over gold are comparisons between real-world gold standards and fictitious paper standards.⁽²⁰⁾ Typically, the alternatives considered are strictly noncormformable: They consist of a market process that gives rise to the use of gold as the medium of exchange and an outcome that no known process can bring about. Wouldn't the world be a better place to live if there were no monetary value attached to gold (or to silver, copper...) and if the monetary authority were constitutionally bound to increase the issue of paper money at a relatively slow, fixed, and foreknown rate? Wouldn't the world be an even better place to live if there were some other monetary commodity, a commodity which was relatively scarce, which could not be extracted by any known mining technique, but which was costlessly coughed up by nature at a slow and steady rate in locations that were experiencing economic growth? These worlds can be imagined to look just like the one that we actually live in—minus the resource costs of gold. Such imaginations may provide the basis for bad science fiction, but they are no basis at all for devising monetary theories or for choosing among alternative institutional arrangements. 

VII. A Constant Price Level vs. Monetary Stability
The assumption of a constant price level has a history that is both long and wide. Over the years theorists representing diverse schools of thought have invoked this assumption in their effort to abstract from monetary influences on the course of economic activity, and have adopted as a self-evident truth the notion that a constant price level is the hallmark of monetary stability. The significance of a constant price level for both theory and policy has been taken to be so obvious and self-evident that the literature is virtually devoid of attempts to defend these common practices. Yet a sampling of the many writers who do not question this assumption—and of the few who do—exposes the assumption as the Achilles' heel of the popular stance against the gold standard and of many other theoretical pronouncements and policy prescriptions.
        Hayek noted in the early 1930s that an unaccountable preeminence of the constant price level characterized the writings of such monetary notables as Gustav Cassel and A. C. Pigou.⁽²¹⁾ That a country should regulate its currency so as to achieve a constant price level appeared to Cassel as the "simplest assumption." If a country's currency were so regulated, money would exert no influence of its own, according to Pigou. The idea that an equality between economic growth and monetary growth is "natural" and that money whose growth rate satisfies this equality is "neutral" had become commonplace by the end of the twenties. The general acceptance of this idea eliminated the need for a theoretical justification.
        The assumed relevance and desirability of a constant price level are incorporated in later decades in the writings of American economists. In the early 1950s Clark Warburton included in his list of assumptions that underlie monetary theory the need for monetary growth to accommodate real economic growth. "[A]s a result of [population growth, technological developments, and increasing labor productivity] and of the stability of customs (such as the periodicity of income payments) which affect the rate of circulation of money, the economy needs for equilibrium a continuous increase in the quantity of money."⁽²²⁾ Following suit, Friedman assumed "for convenience" that a stable price level of final products is the objective of policy.⁽²³⁾ (It is both revealing and disquieting to note that Friedman's estimate of the resource costs of a gold standard discussed in Section IV above depends critically upon an assumption that was made for the sake of convenience.)
        In the late sixties Friedman reaffirmed that he "simply took it for granted, in line with a long tradition and near-consensus in the profession, that a stable level of prices of final products was a desirable objective."⁽²⁴⁾ The purpose of the article that contains this statement was to replace the assumed optimum of a constant price level with a theoretically derived optimum. After identifying costs and benefits of a changing price level, Friedman, following standard microeconomic procedures, set the marginal costs equal to the marginal benefits and solved for the optimal, or welfare-maximizing, rate of change in the price level. It turned out that with an assumed rate of economic growth of three or four percent per year, a decline of prices of four or five percent per year would maximize economic welfare.⁽²⁵⁾ At this rate of price deflation, the marginal gains associated with the deflation-induced increase in real-cash holdings would just be offset by the nearly negligible marginal costs of increasing the supply of money. (These results apply to an economy using fiat currency. If gold were used instead, the marginal costs of extraction would cause the optimal rate of price deflation to be somewhat higher.)
        From the outset Friedman failed fully to persuade even himself of the merits of his theoretically derived optimum. He ended the article with "A Final Schizophrenic Note" in which he teetered between endorsing a monetary rule which would optimize welfare as suggested by his theory and endorsing a monetary rule which would maintain a constant price level. In retrospect Friedman's calculations can be seen as a curious and contrived exercise in the application of marginalism. But today his arguments ring hollow. The unquestioned assumption of the desirability of a constant price level has regained its former status in discussions of monetary policy.
        Economists of the Austrian school have always held the minority view that stable money and a constant price level are two different things.⁽²⁶⁾ At root their case is a very simple one. It requires only the most cursory consideration of what goes on behind the aggregates and the averages of the more orthodox monetary theory. It is true that productivity gains increase the level of output and thereby exert downward pressure on the level of prices. An offsetting increase in the total quantity of money can exert upward pressure and thereby preserve a constant price level. But productivity gains are themselves not neutral with respect to the composition of output. Economic growth does not consist in an across-the-board increase in the quantity of goods produced. It consists instead of increases in the quantities of some goods and decreases in the quantities of other goods, improvements in the quality of some goods, and the introduction of new goods. Growth-induced changes in the pattern of output are accompanied by corresponding changes in the pattern of prices. The fact that the price level calculated on the basis of the new pattern is lower than the price level calculated on the basis of the old pattern is strictly incidental. To the extent that each individual change in the pattern of prices can be attributed to non-monetary factors, the issue of monetary non-neutrality does not arise despite the fall in the price level.⁽²⁷⁾
        The Austrians go on to point out that if an increase in the supply of money is brought about so that the economic growth can be "accommodated," the effects of the monetary injection on prices will be compounding rather than counteracting. Economic growth coupled with monetary growth may allow for a constant price level, but the pattern of prices will be affected in one way by the economic growth and in some other way by the monetary growth. Although it can be imagined that the increase in the supply of money affects only the price level, this lone effect cannot, because of the very nature of money, be actualized. Actual monetary injections, whether in the presence or the absence of economic growth, are always non-neutral.⁽²⁸⁾ They always have their own relative-price effects which, in turn, have effects on the pattern of output. A constant price level, then, is neither an appropriate assumption for devising monetary theories nor the most appropriate goal of monetary policies.
        In current debate the goal of a constant price level enjoys a certain popularity for two reasons. Both have at least some merit, but neither constitutes a telling case against the gold standard. The first reason has to do with political feasibility. Some may argue that the prospects of persuading the central bank to adopt a constant price level as its goal are better than the prospects of persuading it to surrender totally to the dictates of a commodity money. The adoption of such a goal would at least be a step in the right direction, and it would not preclude a further step to a commodity standard if such a step were to prove desirable and feasible. But those who now favor the gold standard do not expect that the central bank will adopt a goal of a constant price level. In fact, they believe that the central bank's unwillingness to do so—or otherwise to behave responsibly—goes a long way towards proving the desirability of a commodity standard. And they believe that the question of which sort of monetary institution is the most desirable should be kept separate from the question of the political feasibility of bringing about the needed institutional change.⁽²⁹⁾
        The second reason for the popularity of this goal derives from money's role as a unit of value and its relationship in this regard to other units such as units of length and units of weight.⁽³⁰⁾ The analogy between the need for invariant units of length and weight and the need for an invariant unit of value is appealing. Carpenters would not fare well in their trade if they had to use measuring devices that expand and shrink on their own; truck drivers would experience an increased dread of weigh stations if they had to wonder how heavy a pound is today. The images conjured up by examples of this sort drive the point home. To be serviceable, units of length, weight and value must be invariant over time. The analogy is persuasive and may be just the right medicine for those who advocate inflation or who advocate artificially cheap credit even if the ultimate result is inflation.
        But for the advocates of sound money, there is more to be learned from the sense in which the analogy does not hold than the sense in which it does. Invariance can be achieved for units of length and weight but not for units of value. Modern attempts to discover or create an invariant unit of value (in the form of multiple-commodity standards, indexation schemes, and the like) represent a throwback to the old pre-marginalist, pre-subjectivist classical economics. They require that we unlearn the lessons implicit in Ricardo's fruitless search.⁽³¹⁾
        This point can be driven home with an analogy of a different sort. (It takes an analogy to beat an analogy.) A monetary commodity is more like a reference commodity, a base point, or bench mark, than like a measuring unit. An immutable reference value for gauging all other values has as its physical analog an immutable reference point in the cosmos. Some might argue that the Earth cannot serve as such a reference point, because the Earth is revolving around the Sun, which is revolving around the center of the Milky Way Galaxy, which is moving through the Universe. An immutable reference point has to be independent of all these movements. Different schemes for locating such a point, which take into account all the relative locations of all the heavenly bodies, might be proposed. But reflection will reveal that the immutable reference point is as useless as it is elusive. The most relevant reference point is the point where cosmic developments have put us. And so it is with the reference value. The most relevant reference commodity is the monetary commodity that market processes have given us. Once gold emerged as the world's monetary commodity, it became irrelevant that certain prices or prices in general may be "unstable" with respect to some other reference value or some index of values. If undisturbed by political schemes, gold should be regarded as a stable money until the market process itself, for whatever reasons, begins to favor some other commodity as a value reference.⁽³²⁾
        The different opponents of the gold standard have radically different reasons for wanting to reject gold as money. Some want to harness the monetary forces and put the reigns in the hands of government; others want to nullify the monetary forces that are inherent in any commodity standard. The former like to think of monetary stability as those monetary arrangements that result in full employment; the latter like to think of monetary stability as those monetary arrangements that result in a constant price level. Proponents of the gold standard hold that neither full employment nor a constant price level is an appropriate goal of government policy. Nor is either of these goals consistent with monetary stability. And achieving the goal of stable money, which may well result in both a fuller employment and a more nearly constant price level than would otherwise be possible, requires only that the government refrain from interfering with the commodity money chosen by the market. 

VIII. Concluding Remarks
Opponents of the gold standard calculate the costs of gold in dollars and cents and report their calculations as a percentage of the economy's output. The intended interpretation is clear: But for the costs of gold, the economy would have had an output that much greater. Proponents of the gold standard would be ill-advised to respond with a cost figure of their own. If the true costs of a gold standard could be calculated at all, it would have to take into account the monetary instability associated with alternative standards and the consequent loss of output. But incorporating these considerations would undoubtedly cause the cost figures to turn negative. The gold standard has net benefits, not net costs. An appreciation for these benefits, but not a precise quantitative estimate, can best be gained by comparisons of historical episodes which are illustrative of economic performance under a gold standard and economic performance under a paper standard. The superiority of the former in comparison to the latter constitutes the net benefits of the gold standard.⁽³³⁾
        Ultimately, the cost of any action, commodity, or institution is the alternative action, commodity, or institution forgone. The opportunity cost is the only cost that counts. The cost of one institution is forgoing some other institution; the cost of the gold standard is forgoing a paper standard; the cost of sound money is forgoing unsound money.
 

Notes

* I would like to acknowledge the helpful comments and criticisms offered by Don Bellante, Don Boudreaux, and Leland B. Yeager of Auburn University and Gerald P. O'Driscoll, Jr. of the Federal Reserve Bank of Dallas.

1. Carl Menger, "On the Origin of Money," Economic Journal, vol. 2 (June, 1892), pp. 239-55. Also see Menger, Principles of Economics, trans. and ed. by James Dingwall and Bert F. Hozelitz (Glencoe, IL: Free Press, 1950), pp. 257-62.

2. One early list of such characteristics includes (1) utility and value, (2) portability, (3) indestructability, (4) homogeneity, (5) divisibility, (6) stability of value, and (7) cognizability. William Stanley Jevons, Money and the Mechanism of Exchange (New York: D. Appleton and Company, 1882), p. 31. It might be noted that characteristics 1 and 6 of Jevons' list are strongly reinforced as the particular commodity so characterized emerges as the medium of exchange.

3. Ibid. p. 41. Jevons lists in order: gold, silver, copper, tin, lead, and iron. Also see Ludwig von Mises, The Theory of Money and Credit, trans. by H. E. Batson (New Haven: Yale University Press, 1953), pp. 30-34.

4. Milton Friedman, "Should There Be an Independent Monetary Authority?" in Leland B. Yeager, ed., In Search of a Monetary Constitution (Cambridge: Harvard University Press, 1962), p. 228.

5. F. A. Hayek, "The Results of Human Action but not of Human Design," in Hayek, Studies in Philosophy, Politics and Economics (New York: Simon and Schuster, Publishers, 1969), p. 104.

6. Ibid. Language and money are probably the two most striking examples of social phenomena that have "arisen from man's actions without his design."

7. Ludwig von Mises, On the Manipulation of Money and Credit, trans. by Bettina Bien Graves and ed. by Percy L. Graves, Jr. (Dobbs Ferry, New York: Free Market Books, 1978), p. 80. Also see Mises, Human Action, 3rd rev. ed. (Chicago: Henry Regnery Company, 1966), pp. 471-76.

8. John Maynard Keynes, The General Theory of Employment, Interest, and Money (New York: Harcourt, Brace and World, Inc., 1964), pp. 229-36. Keynes is very clear on this point: "It is interesting to notice that the characteristic which has been traditionally supposed to render gold especially suitable for use as a standard of value, namely, its inelasticity of supply, turns out to be precisely the characteristic which is at the bottom of the trouble." This statement follows the more fanciful account of the trouble with gold: "Unemployment develops, that is to say, because people want the moon;--men cannot be employed when the object of their desire (i.e. money) is something which cannot be produced and the demand for which cannot be readily chocked off.

There is no remedy but to persuade the public that green cheese is practically the same thing and to have a green cheese factory (i.e. a central bank) under public control." Ibid. pp. 235-36.

9. For an account of the clay-brick standard first proposed by C. O. Hardy, see James M. Buchanan, "Predictability: the Criterion of Monetary Constitutions" in Yeager, In Search of a Monetary Constitution, pp. 155-83.

10. Milton Friedman, "Commodity-Reserve Currency," in Essays in Positive Economics (Chicago: University of Chicago Press, 1953), pp. 204-50. In his discussion of the supply elasticities of alternative monetary commodities, Friedman points out the disadvantages of an elasticity greater than zero (pp. 209-10) as well as the disadvantages of an elasticity less than infinity (pp. 210-14). The idea that the use of a monetary commodity whose supply elasticity is greater than zero involves "waste" has a long history and was endorsed by the classical economists including Smith and Ricardo.

11. Ibid. p. 210. Also, see Milton Friedman, A Program for Monetary Stability (New York: Fordham University Press, 1959), p. 5.

12. These estimates are from Friedman's "Commodity-Reserve Currency," which was originally published in 1951. The high resource costs were to become even higher by the end of the decade when Friedman offered a new estimate in his Program for Monetary Stability. On the basis of a slightly higher rate of economic growth, he estimated that about two-and-a-half percent of the economy's output, or $8 billion dollars per year, would have to be devoted to the procurement of additional quantities of the monetary commodity.

13. Allan H. Meltzer, "Monetary Reform in an Uncertain Environment," Cato Journal, vol. 3, no. 1 (Spring, 1983), pp. 93-112. Meltzer's updated estimate is found on page 105. This particular estimating procedure, in which the estimate of the resource costs of gold is always some multiple of the economy's growth rate, gives rise to an interesting paradox. If mismanaged paper money causes so much economic discoordination that the growth rate drops to zero, the estimated resource costs of maintaining a gold standard would also drop to zero; but if the adoption of the now-inexpensive gold standard creates economic stability and fosters new growth, maintaining the gold standard would once again become too costly.

14. Friedman, "Commodity-Reserve Currency," p. 210; A Program for Monetary Stability, p. 5.

15. This point has been emphasized by Ronald Coase: "It would seem desirable to use [an opportunity-cost] approach when dealing with questions of economic policy and to compare the total product yielded by alternative social arrangements." Coase, "The Problem of Social Cost," Journal of Law and Economics, vol. 3 (October, 1960), p. 43.

16. It would be a gross misrepresentation, of course, to add the costs that are incurred as a result of an assumed inelasticity of the supply of gold to the resource costs that are based on an assumed perfect elasticity.

17. Arguments of this sort have their roots in the writings of Clark Warburton. See Warburton, "The Monetary Disequilibrium Hypothesis," in Warburton, Depression, Inflation, and Monetary Policy, Selected papers, 1945-1953, (Baltimore: The Johns Hopkins Press, 1966), pp. 25-35. Also, see Leland B. Yeager, "Stable Money and Free-Market Currencies," Cato Journal, vol. 3, no. 1 (Spring, 1983), pp. 305-26.

18. A more thorough discussion of the relevance of a constant price level is offered in Section VII below.

19. For historical perspectives on gold and paper, see Murray N. Rothbard, What has Government Done to Our Money? (A Pine Tree Publication, 1963), pp. 27-49; Ron Paul and Lewis Lehrman, The Case for Gold: A Minority Report of the U. S. Gold Commission (Washington, D. C.: Cato Institute, 1982), pp 17-142; and Alan Reynolds, "Why Gold?" Cato Journal, vol. 3, no. 1 (Spring, 1983), pp. 211-32.

20. Again, it was Ronald Coase who sensitized the profession to "the usual treatment of [problems of social cost in which] the analysis proceeds in terms of a comparison between a state of laissez faire and some kind of ideal world." He went on to point out that "very little analysis is required to show that an ideal world is better than a state of laissez faire unless the definitions of a state of laissez faire and an ideal world happen to be the same." Coase, "The Problem of Social Costs," p. 43. Harold Demsetz has elaborated on the differences between the "nirvana approach" (which is implicitly adopted by many critics of the gold standard) and the "comparative-institutions approach" (which is adopted in the present paper). Demsetz, "Information and Efficiency: Another Viewpoint," Journal of Law and Economics, vol. 12 (April, 1969), pp. 1-22.

21. F. A. Hayek, Prices and Production, 2nd ed. (New York: Augustus M. Kelley, Publishers, 1967), p. 107.

22. Warburton, "The Monetary Disequilibrium Hypothesis," p. 28. Warburton notes that this assumption "pervaded so much economic literature of the nineteenth century and early part of the twentieth that supporting documentary references seem superfluous." Ibid., p. 29.

23. Friedman, "Commodity-Reserve Currency," p. 210.

24. Milton Friedman, "The Optimum Quantity of Money," in The Optimum Quantity of Money and Other Essays (Chicago: Aldine Publishing Company, 1969), p. 48.

25. Ibid., p. 46.

26. Hayek, Prices and Production, pp. 105-31; Mises, Human Action, pp. 219-28, and passim; Mises, On the Manipulation of Money and Credit, pp. 1-49; and Mises, Theory of Money and Credit, pp. 108-69.

27. Mises, Theory of Money and Credit, p. 123. Mises maintains a first-order distinction between "two sorts of determinants of the exchange-value that connects money and other economic goods; those that exercise their effect on the money side of the ratio and those that exercise their effect on the commodity side." Only the effects of the first-mentioned sort are relevant to the issue of monetary neutrality.

28. Mises, Human Action, p. 418. "The notion of a neutral money is no less contradictory than that of a money of stable purchasing power. Money without a driving force of its own would not, as people assume, be a perfect money; it would not be money at all." The fact that money is not neutral is the common denominator of the monetary theories of Mises, Hayek, and others who have written in the Austrian tradition.

29. Friedman has enunciated this maxim but has not always abided by it. "The role of the economist in discussions of public policy seems to me to be to prescribe what should be done in the light of what can be done, politics aside, and not to predict what is "politically feasible" and then to recommend it." Milton Friedman, "Comments on Monetary Policy," in Essays in Positive Economics, (Chicago: University of Chicago Press, 1953), p. 264. William Hutt, who is critical of the Friedman maxim, recommends that the economist take political considerations into account, but only if such considerations are made explicit. This, he argues, is the business of old-style "political economy." William H. Hutt, Politically Impossible...? (London:Institute of Economic Affairs, 1971), pp. 22-27.

30. Yeager, "Stable Money and Free-Market Currencies," pp. 305-308.

31. "The Search for a stable unit of account is ultimately the search for an invariant standard of value, the quixotic goal of classical political economy." Gerald P. O'Driscoll, Jr., "A Free-Market Money: Comment on Yeager," Cato Journal, vol. 3, no. 1 (Spring, 1983), p. 328.

32. Economic calculation does not require monetary stability in the sense in which this term is used by the champions of the stabilization movement. The fact that rigidity in the monetary unit's purchasing power is unthinkable and unrealizable does not impair the methods of economic calculation. What economic calculation requires is a monetary system whose functioning is not sabotaged by government interference." Mises, Human Action, p. 223. It should be noted that the dramatic fluctuations in the value of gold in recent years do not imply that gold is no longer a viable monetary commodity. Quite to the contrary, volatile shifts in the demand for hard money are a reflection of the instability of our present monetary institution. The paper money is losing its viability. See Mises, On the Manipulation of Money and Credit, pp. 76-77.

33. Historical studies are cited in note 19, above.