A Note on Gold Production and Additions to International Gold Reserves

THE PURPOSE OF THIS NOTE is to survey trends in the production of gold and the flow of gold to monetary reserves, with the intention of estimating the probable additions to international reserves during the next decade.1 The discussion is based on the assumption that the U.S. price of gold will remain at $35 per ounce.


THE PURPOSE OF THIS NOTE is to survey trends in the production of gold and the flow of gold to monetary reserves, with the intention of estimating the probable additions to international reserves during the next decade.1 The discussion is based on the assumption that the U.S. price of gold will remain at $35 per ounce.

THE PURPOSE OF THIS NOTE is to survey trends in the production of gold and the flow of gold to monetary reserves, with the intention of estimating the probable additions to international reserves during the next decade.1 The discussion is based on the assumption that the U.S. price of gold will remain at $35 per ounce.

Additions to the total gold reserves of the countries outside the Soviet bloc will depend upon the amount of gold initially made available by new production, the sales of gold by the Soviet bloc to the rest of the world, the net amount of gold used for industrial and artistic purposes, and the amount of gold bought privately for saving or hoarding. Gold acquired for this last purpose may be in the form of gold coins, semi-processed gold, or jewelry and other fabricated objects.

Gold Production

World gold production, excluding that in the Soviet bloc countries, increased in every year in the postwar period, except 1951 and 1953, and by 1957 its value had reached $1,020 million, compared with $736 million in 1945 (Table 1). The average compound rate of increase over the entire period was 2.7 per cent per year; in the five years 1953–57, it was 4.7 per cent per year, because of the expanding production in South Africa. Maintenance of, or an increase over, the present level of production during the next decade appears to depend mainly upon the course of South African production, whose comparative importance is suggested by the data in Table 2.

Table 1.

Value of World Gold Production, Excluding Countries in Soviet Bloc, 1910–56

(Valued at US$35 per fine ounce; in millions of U.S. dollars)

article image
Sources: Data for 1910–40 are based on figures in Board of Governors of the Federal Reserve System, Banking and Monetary Statistics (Washington, November 1943), Table 159, adjusted to exclude the U.S.S.R., China, and Rumania; the 1910–33 figures in the source have been revalued from $20.67 per fine ounce to $35. Data for 1941–44 are from Federal Reserve Bulletin (Washington) and exclude the U.S.S.R. Those for 1945–56 are from International Financial Statistics (Washington) and exclude the Soviet bloc. Figure for 1957 is preliminary.
Table 2.

Value of World Gold Production, Excluding Countries in Soviet Bloc, 1948–57

(Valued at US$35 per fine ounce; in millions of U.S. dollars)

article image
Source: International Financial Statistics (Washington).

The increase in South African production during the past few years has been due to new mines in the West Rand and the Orange Free State. The output of these mines rose from 100,000 ounces in 1951 to 3.1 million ounces in 1953 (27 per cent of total South African output) and 5.3 million ounces in 1955 (38 per cent of the total). 2 Total output has continued to rise. In 1957 the value of the output of all mines was $596 million, about 7 per cent higher than in 1956; in the third quarter of 1957 it was at the rate of $611 million per year, and in the fourth quarter at a rate of $599 million. Production in the rest of the world has been approximately constant, with increases in some countries offsetting decreases in others. A brief examination of the gold mining industry and its problems will be of assistance in suggesting the extent to which these trends are likely to continue.

Economic problems of gold mining

Gold mining has been, and it is assumed will continue to be, subject to the combination of a fixed dollar price3 for gold for official uses (though the hoarding demand for gold kept its “free market” price above the U.S. official price for a considerable number of years) and increasing costs of labor and materials. Increasing costs arising from higher prices and wages apparently cannot be completely offset by increasing efficiency of operations. The amount of gold production that can be directly attributed to subsidies is relatively small, though it is significant for marginal producers or for marginal production in certain countries. New discoveries require a number of years of development to come to maximum output.4 Major gold discoveries occur infrequently, and recent discoveries have required the solution of many difficult technical and social problems and the long-term investment of substantial amounts of capital.5

After World War I, and particularly until 1934—when the United States raised its gold price from $20.67 to $35 per fine ounce—there were repeated forecasts that gold production would decrease. These forecasts, including those made for the Gold Delegation of the League of Nations, were almost universally too pessimistic.6 Gold production increased slowly during the 1920’s: by 7 per cent from 1920 to 1925 and by 7½ per cent from 1925 to 1930. Between 1929 and 1933, under the stimuli of unemployment and falling costs, but without any change in the dollar price of gold, production increased by 22 per cent. After the U.S. price of gold was raised in January 1934 to $35 per ounce, production increased faster. By 1940 it was more than 60 per cent higher than in 1933. Production fell sharply during World War II, largely because wartime requirements diverted resources to other ends. By the end of the war, production had fallen to the 1933 level. In the postwar period, gold mining has been faced with an unchanged official dollar price but with substantially higher costs.7 Though production in 1957 was 40 per cent more than at the end of the war, it was still 22 per cent less than the all-time peak reached in 1940.

The fact that the official price of gold in terms of dollars has been $35 per ounce since 1934 has not ruled out changes in the free market price of gold. Sales of new gold at premium prices have helped, during most of the postwar period, to increase the average price realized from new production. For example, the average premium realized on South African sales (at both official and premium prices) was 10 per cent in 1952 and 5 per cent in 1953.8 The current premium is minimal.

The unchanged official price of gold also has not ruled out changes in domestic prices, as measured in local currencies, in various countries. From this point of view, the economic conditions under which gold is produced may be classified in two groups, depending upon the behavior of the domestic price. The United States and Canada are the two countries where the price of gold in local currency now is the same as, or lower than, the price in 1934. The U.S. price has been unchanged since 1934. The Canadian price, in Canadian dollars, has fluctuated between 11–15 per cent above, and 6 per cent below, the price in 1934. In 1957 the price of gold in Canadian dollars was about 4 per cent below that at the beginning of 1934. Neither country protects the domestic market for gold, but gold production is a much more important industry, both absolutely and relatively, in Canada than in the United States. In all the other important gold producing countries, the price of gold in local currency is now higher than in 1934. It has increased by 80 per cent in Australia and South Africa and 250 per cent in Colombia and Mexico. In these and other countries, therefore, the cost-price squeeze of the past few years is attributable only in part to the price term of the relationship. Price advantages initially contributed by devaluation have been more than offset by rising domestic wages and other factor costs. It is possible that a devaluation, by modifying the terms of trade, may lower aggregate real income. This may be the case whatever the export commodity is. If the advantage of the devaluation is not lost through domestic cost increases, however, production will be stimulated, whether the export commodity is gold or anything else.9 In a number of countries, gold production, along with other export industries, has been the victim of increases in domestic prices.

Gold production has also been adversely affected during the postwar period by the effects of sustained prosperity and full employment. The boom in all other kinds of mining, and the increased demand for mining labor, would by itself have raised gold mining costs, though this need not be the case where gold is a by-product of copper and other metals.10 Passing of the boom, or advent of a depression, with its attendant unemployment and its falling prices and costs, would make gold mining—with its unlimited market at stable prices—relatively more attractive. It was this factor that accounted for the 22 per cent increase in gold output between 1929 and 1933. Without this built-in stability, gold mining would have been affected more than it has been in the postwar period. It has been observed that, at least in Canada, “the steady employment which the gold mines offer is a major attraction, especially to heads of families. In the gold mines, something like a ‘guaranteed annual wage’ has long been a reality … the peculiarities of gold mining as an occupation suggest that the supply of labour may remain larger than it would be in other occupations in the face of similar trends in production.” 11

In many countries, the conception of a cost-price relationship as applied to gold mining is more complicated than it would at first appear to be. Where gold production is a relatively small part of the total production in the economy, the gold industry must buy its labor and other production factors at market prices over which it has little or no influence. In such cases, it is likely to be treated uniformly with other comparable industries in regard to taxation, assessment for social benefits, etc. Under these conditions, the term “costs” has a fairly specific meaning which is related to the national structure of costs. This is the situation, for example, in the United States and Canada. However, where the gold mining industry is relatively important it will probably be singled out for special treatment. Costs may then contain important elements of profit sharing in the form of taxes, differential exchange rates, or the like. In addition, such elements may be made to bear differently upon domestically owned mines and upon foreign-owned ones. These elements can be and often are increased if production becomes more profitable; they can be decreased if it becomes less so. As Hardy stated in 1936 with respect to South Africa, “the present level of costs is shot through with items which really constitute a distribution of the net income of the industry … rather than the payments of costs that really have to be met in order to maintain a high level of output. … The costs which the industry has to incur in the period of its prosperity are not a good criterion of the necessary costs involved in keeping it going. …”12

The tax treatment of gold mines in South Africa illustrates three of the ways used by governments to take account of changing profitability: (1) adjusting the tax rates applicable to taxable income; (2) modifying the definition of taxable income; and (3) treating gold mines as a special class for tax purposes. As to the first, South Africa increased the tax burden of the gold mining industry by the Gold Mines Excess Profits Duty Act, 1933, which imposed a tax additional to the normal income tax on the difference between gold mining profits at the new price of $35 per ounce, established in that year, and the profits that would have been received if the gold price had remained unchanged.13 After sterling was devalued in 1949, the tax formula was changed in order to increase tax rates.14 Later, as costs increased, the formula was changed from time to time to produce lower rates. Tax relief of the order of £700,000 was granted for 1956, equivalent to a reduction of 4¾ per cent for those mines paying the formula tax.15 By 1956, South Africa had returned to the same tax schedule it had used prior to the devaluation of 1949.16 The government revenue collected from the gold mining industry in recent years is shown in Table 3.

Table 3.

Gold Production, Working Profit, and Government Revenue from Gold Mining in South Africa, 1946–56

(In millions of South African pounds)

article image
Source: Transvaal and Orange Free State Chamber of Mines, Sixty-Sixth Annual Report, Year 1955 (Johannesburg), pp. 72–73, and Sixty-Seventh Annual Report, Year 1956, pp. 80–81 and 93.

The second method of recognizing profitability is to vary the amounts and kinds of permissible deductions and the ways of meeting specific industry charges. Extended discussion of this highly technical subject is beyond the scope of this paper; however, it may be noted in passing that different kinds of mining are given special treatment in many countries, especially with respect to depletion. Nevertheless, two examples will illustrate the point. In South Africa, gold mines are not regarded as having any taxable income until all capital expenditures to bring a mine to production have been written off. This provision was recently extended by a special concession for new deep-level mines, which may have unusually long preproduction periods. Their permissible write-off consists not only of the capital invested but, in addition, of an amount equal to 5 per cent per year compound interest on the balance to be charged off.17 The taxation liabilities of gold mines were also affected by the responsibility accepted in 1956 by the South African Government for the payment of about £500,000 a year for 30 years in connection with silicosis legislation.18

The third method, i.e., to treat gold mines as a special class for tax purposes, is perhaps the most important. The tax rate applicable to each gold mine is determined by a general formula applicable solely to gold mines. The tax rate for each mine depends upon its own ratio of taxable income to gross income. The more profitable the mine, the higher the applicable tax rate. In contrast, the tax rates imposed on other kinds of mines, and on nonmining enterprises, do not depend upon the profitableness of the enterprise; for each type of mine, and for non mining enterprise in general, there is a separate, uniform tax rate. The tax rates applicable to the more profitable gold mining companies are higher than those applicable to other companies, mining or nonmining; on the other hand, the situation may be the reverse for the least profitable companies.19 For a rich mine, i.e., one with a high proportion of net to gross income, the income tax plus profit sharing may come to as much as two thirds of net income.20 The emergence of uranium as a byproduct, or as a coproduct, may raise some interesting questions with respect to taxing gold mines.21

There have been complaints for many years in South Africa that the taxation of gold mines is discriminatory and unduly burdensome. It has also been argued that “graduated tax rates may justifiably be applied to the incomes of individuals but it is difficult to defend their application to the incomes of public companies, whose shareholders include both rich and poor individuals.”22 In a recent defense of the tax system, the Government pointed out that the taxing formula had been devised by the Gold Mining Taxation Commission in 1945, that the results achieved by the formula were just, and that low profit mines did not have to pay tax.23

This brief survey of some of the outstanding economic problems of gold mining suggests that the industry operates subject to a number of special factors. There is little doubt that an increase in the price of gold would, after a time, and to a greater or lesser extent, increase the production of gold. But since it is assumed that the price of $35 per ounce will not be increased, the only question with which this paper is concerned is what the course of gold production is likely to be if the price continues unchanged and there is no serious depression. A closer examination of the situation in the major producing countries will help to answer this question.

Union of South Africa

The Union of South Africa in 1957 was responsible for 58 per cent of the gold production of the world, excluding countries in the Soviet bloc. It is the only country that in recent years has benefited by the discovery of substantial new gold fields. It is likely in the next decade to be the major determinant of the world level of gold production. Within South Africa, the production of the newer mines, principally in the Orange Free State, will continue to increase for some time. The future course of South African production appears to depend, first, on the extent of this increase and, second, on the prospects of production from the older mines. The following points bear on this question:

1. The ultimate capacity of the newer mines “is still unknown. It depends partly on geological factors, of which the amount of the accumulated water still under the ground and the amount of faulting are the most obvious.” 24 Many of the newer mines are not yet at capacity production, and it has been estimated that the monthly milling rate of the existing Orange Free State mines will rise over the next few years to more than twice the rate attained at the end of 1955.25 In addition, the gold content of milled ore has been improving and is likely to continue to improve.26 The increase in the milling rate, in combination with an improvement in the grade of ore, would suggest a peak output of about 5–6 million ounces from the new mines already in production in the Orange Free State, compared with 3.8 million ounces in 1957. In addition, several new mines are being brought to the stage of production, and others are in a more or less advanced planning stage.27

2. Output on the Witwatersrand, comprising the remainder of South African production, declined in the decade which followed the record year of 1941, but it has been increasing since 1953. The 1957 output was 5 per cent greater than in 1956 and was exceeded in only a few earlier years.28 The increases in an area where output was for some time declining have been brought about mainly by the contribution of new producers and the concentration on mining richer grades of ore.29

3. Uranium has become an important by-product of the gold mining industry, and it provides a substantial support for gold production. The extent of this support is suggested by figures on the distribution between the two metals of the working profits for all mines.30 These figures show that the percentage of total profits that was due to uranium increased from 17 in 1954 to 36 in the third quarter of 1957.

The importance of uranium to gold mining is even greater than the figures suggest. As of December 1957, 22 of the 53 operating gold mines were combined gold and uranium producers, and it is expected that this number will increase to more than 30. These combined producers were responsible for one half of the gold output in South Africa. About half the working profits of these producers came from gold, and the other half came from uranium, based upon operating results for 1956 and the first nine months of 1957. These data have two interesting aspects. First, though a few mines can produce uranium without producing gold, i.e., by reworking the tailings from which gold has been extracted, none of the newer mines can do this, since they produce uranium from newly mined ore. Uranium production is covered by long-term contracts. These provide for complete amortization of processing facilities during the lifetime of the contracts. Second, in interpreting these profits data it must be kept in mind that, for costing purposes, uranium is treated as a by-product, and all costs of mining and treating the ore until after the gold has been extracted are charged to gold. Uranium is charged only with the direct costs of processing the tailings.31 With such cost allocation, gold might economically be produced at a loss. Indeed, in the third quarter of 1957, 6 combination mines showed a loss on gold. In any event, production of the two metals can be mutually supporting as long as the profits from both are adequate. And a change in the tax system, which would apply different rates to uranium and gold, now under discussion, might make net profits after tax larger.32

4. There is no reason to believe that the process of shutting down older and less efficient mines will not continue. Some mines will close because their ore is almost exhausted. Neither an increase in the price of gold nor a decrease in working costs can materially affect this. Other mines with low-grade ore and narrow profit margins will be seriously affected by rising costs. Although some of these mines will be able to postpone closing by operating more efficiently and by concentrating on better yielding ores, others will undoubtedly have to close.33 In June 1957, the President of the Chamber of Mines put the problem in the following perspective: 34

Of the 56 mines in production in 1956, 18 had been in production for 30 years or more and in no fewer than nine of them, production began in the last century and has been maintained for some 60 years. This is a remarkable achievement. …

The problem of the continued existence of these older mines, which are particularly vulnerable to a rise in working costs, has received a great deal of publicity in recent months. This has created the impression that the problem has suddenly arisen. Nothing could be further from the truth.

And he went on to say:

… every mine inevitably reaches the stage when, under existing conditions, it can no longer work profitably. It is clearly not in the interests of shareholders that mining operations should be continued beyond this stage. When this stage is reached, the interests of shareholders are best served by closing the mine, selling the assets and making a return of capital which could be available for reinvestment in new and potentially more profitable mining ventures.

Nor, indeed, is it in the interests of the country to continue, by one expedient or another, the production of gold by a mine that can no longer do so at a profit, if alternative employment is available for that mine’s employees in profitable enterprises. There is also a real danger that whatever expedient is adopted may further prejudice the national economy by placing additional burdens on it. The problem we are faced with is essentially that of moving men from one job to another with as little disruption as possible.

… We have expressed our willingness to assist the Government in attempting to arrive at a workable solution which might postpone for a period the inevitable closure of the present ‘vulnerable mines’ while the newer mines are progressing towards full production. In the interests of shareholders and of the country generally, such postponement cannot be unduly prolonged and the main aim should be to encourage the establishment of new mines to replace those that must in any circumstances close, be it this year or in the years to come.

According to one estimate, shutdowns by marginal mines, plus shutdowns by those mines whose ores will become exhausted, could not affect more than 750,000 tons of mill throughput per month by 1960.35 This would be equivalent to a maximum reduction of 1.8 million ounces of gold output per year. Such a reduction would be a net loss to the mines concerned, but not necessarily to total gold production in South Africa.

5. For many years, there has been a scarcity of all classes of mining labor. As a result, few mines have been able to mill at capacity in recent years. Thus, given present mining techniques, an increase in the total labor force is one condition for larger output; and this in turn would tend to reduce unit working costs. The prospects of remedying the state of affairs resulting from the over-all shortage of labor are problematical.36 It is difficult to disagree with a 1956 analysis which concluded that “over the next five years, and probably for longer, the demand for labour on the gold mines of South Africa was likely to be uncomfortably in excess of supply.”37 It is, therefore, more important than ever to provide workers with more capital equipment, to develop more efficient operating methods, and to classify, train, and upgrade workers as effectively as possible. Considerable progress is being made along these lines. The introduction of new methods has increased the speed and reduced the cost of development. A great deal has been accomplished by applying modern techniques of aptitude testing, hiring, wage administration, and training. But, as has been noted by many observers, both inside and outside the industry, there is a limit to what can be done, because of the bounds imposed on the mining industry by law and by custom in the utilization of non-European labor.38 Under these conditions, the transfer of labor from less to more productive workings is an important element in the expansion of South African gold production.

Native labor is apportioned by quota to the various mines on the basis of each mine’s scheduled mill capacity and a number of other factors and variables. Transfer of labor thus implies a change of quotas. Making labor available for transfer involves reducing the scale of operation of some mines by concentrating on their richer ores, i.e., overmining. Since South Africa has enforced a general policy prohibiting overmining, on the ground that it involves a loss of the country’s mineral resources, transfer of labor requires governmental as well as private action. Whether the national interest would be served by relaxing this policy depends upon the over-all scarcity of native labor and the extent to which this scarcity adversely affects the expansion of the newer and richer mines.39

6. Since the end of World War II, working costs per ton of ore have risen by 70 per cent, or from 26s. in 1946–47 to 43s. in 1956 and 44s. 5d. in the first quarter of 1957.40 The sharp postwar increase in working costs appears even sharper in the perspective that working costs in 1946–47 were almost the same as they had been in 1920 and in 1902. Present costs are about 130 per cent higher than in 1934, when the price of gold was last increased.

Working revenue per ton has also increased in the postwar period. There was a large increase resulting from the currency devaluation of 1949. Since 1950, working revenue has increased by 40 per cent (from 42s. 7d. to 60s. 8d.) by reason of the mining of richer ore.

In 1950, immediately after the devaluation, working profit per ton was 17s. 4d., the highest point in this century. It declined sharply in the next three years, and was 11s. 11d. in 1953. Since then the increasing number of new mines has raised the average working profit. In 1956 it was 14s. 6d., and in the first quarter of 1957 it was 16s. 0d. The latter margin was almost the same as it had been in 1902 (16s. 6d.), the second highest point in this century, and almost twice that in the late 1920’s.41

There is a wide range in profit margins among mines. Those producing both gold and uranium averaged 24s. 7d. in the first quarter of 1957, and only one, an extremely small one, had a profit margin of less than 10s. per ton. On the other hand, the 33 straight gold producers averaged 11s. 2d.; 14 of these had a profit margin of 5s. or less per ton, and 4, of 5s. to 10s. Mines with a profit margin of 5s. or less presently contribute about 18 per cent of South African production, and those with a margin of 5s. to 10s., almost 7 per cent.

The considerations that have been described thus point to different directions in the future, and it is not easy to strike a balance among them. They appear to suggest a steadily rising South African output that by 1960–62 may reach perhaps 19–21 million ounces per year, valued at $660–730 million, compared with $596 million in 1957. However, estimates involving so many variables—economic, geologic, social, and political—can be subject to large margins of error and considerable differences of opinion. It is understandable that the official South African position, as stated in September 1956, was that “total production is likely to show an increase in 1957, but this is solely due to the expected large increase in the output of the new mines. The output of these new mines may also be sufficient to sustain total production, though probably at a slightly lower level, after 1957.”42

United States

The value (at $35 per ounce) of gold production in the United States rose from $89 million in 1933 to $210 million in 1940; by 1951, however, it had fallen to $66 million. It has changed little since then, and in 1957 it was $64 million. In the process, many gold mines have ceased operation, but the major units have continued to produce gold. Some of these produce gold as a by-product. In 1953, for example, 31 per cent of all gold production was a by-product of large-scale mining of low-grade copper ore, and 8 per cent was a by-product of other types of base metal mining.43 In general, 35–40 per cent of gold production in the United States depends upon the production of base metals, principally copper. Though some of the remainder is associated with the production of silver or other metals, it may be considered, for the purposes of this paper, as depending upon the profitability of gold mining. This is the case, for example, with the largest mine in the United States, which produces about 28 per cent of total U.S. output. Output in 1957 for the Homestake Mine, the fifth largest in the world, was almost 555,000 ounces. This was the largest in any postwar year and may be compared with its highest figures of 570,000 ounces in 1939 and 560,000 ounces in 1936 and 1941. The output of recent years was achieved despite large increases in factor costs. Profits and dividends are substantially lower now than in the bonanza period 1934–41, but dividends in the postwar period have continued substantially unchanged, as indicated by the data in Table 4. The earnings shown in the table are after allowances have been made for percentage depletion; and these depletion allowances apply to a mine that has operated for more than 70 years and produced more than $400 million of gold. Whether or not such depletion is, in large part, more properly considered as tax exempt income, the company as a matter of policy has for many years paid dividends in excess of net income.

Table 4.

Production, Earnings, and Dividends of the Homestake Mining Co., 1925–56

(Production in thousands of ounces; earnings and dividends in millions of U.S. dollars)

article image
Source: Moody’s Investors Service, Moody’s Manual of Investments, American and Foreign, Industrial Securities (New York), and Standard & Poor’s Corporation, Standard Corporation Descriptions (New York). Production data through 1946 were calculated from reported bullion receipts, except that 1933 data could not be so determined. Earnings are after percentage depletion.

The record of the Homestake Mine makes clear that, at least under certain conditions, production can be maintained at high levels despite sharply increased factor costs which continue to increase. It is interesting to note that plans are going forward at Homestake for an expansion of production at deeper levels, and that these plans involve substantial additional investment.

While the United States produced $64 million, or 6 per cent, of the world’s gold in 1957, there is no expectation that this amount will increase. If anything, output may decline somewhat.


Gold mining in Canada differs, or has differed, from that in the United States in two important respects. First, for half the period since 1934, or roughly during the years 1939–46 and 1949–51, the price of gold was higher in Canadian than in U.S. dollars, by reason of the discount on Canadian dollars. This further tended to stimulate gold production, as did some sales of gold at premium prices in the international market in the postwar period. Indeed, it has been argued that in retrospect the stimulation of the gold industry in the 1930’s was uneconomic and that, by encouraging mines to enlarge their operations to work lower-grade ores, it increased costs of production.44 Especially as regards lower grades of ores, this enlargement has become more burdensome with the increase of costs, the appearance of the premium on Canadian dollars, and lower premium prices in international markets. Since 1951 there has been a premium on Canadian dollars, so that the price of gold in Canadian dollars has been lower than it was in 1934. Secondly, unlike the United States, Canada has had a subsidy program since 1948. During the period 1948–55, the value of gold output was $1,350 million. Subsidy payments during the period totaled almost $100 million, and thus increased the gross income from sales of newly mined gold by approximately 7½ per cent. Assistance costs in 1956 are estimated at $8.9 million, or 5.8 per cent of gold production of $154 million.45 The Assistance Act has been extended through 1958.

The gold mines eligible for assistance produce more than 3 million ounces per year. With a gold price of $35 per ounce, and parity between the Canadian and U.S. dollars, their output would be valued at more than Can$105 million. During 1957 the exchange rate premium on Canadian dollars averaged about 4 per cent, so that the average price per ounce realized by Canadian producers was about Can$33.65. Their revenue was therefore about Can$4 million less than it would have been at dollar parity, and the loss of revenue is equal to about half their estimated assistance payments for the year.

As in the United States, some gold is produced as a by-product of base metals; about 13 per cent of gold production appears to be associated to a significant extent with the production of other metals. The great expansion of mining in Canada since the war has exerted a strong pull on available resources of labor and capital. Factor costs, including wages, have risen sharply in response to prosperity in general and the expansion of mining in particular.

Gold production has responded to these conflicting factors by holding its own, though the number of mines has fallen by one half since 1948. The value of production (at US$35 per ounce) rose from US$104 million in 1933 to US$187 million in 1940. In the eight years 1950–57 it was relatively stable, averaging US$155 million per year. Production is likely to continue at this level for some years, but if full employment and rising costs are assumed expansion is unlikely.

Other gold producing countries

South Africa, Canada, and the United States in 1957 produced gold valued at US$815 million. The value of output in all other countries was $205 million, with the largest amounts produced in the countries listed in Table 5. It is impossible to characterize, in a limited survey, the different production situations in these and other minor gold producing countries. Some have more or less extensive subsidies. Some permit producers to sell gold output in external premium markets. Some have domestic markets for gold which are insulated and where, in effect, local production may be sold above the world price. In the Philippines, for example, nonresident holders of blocked pesos may acquire dollar exchange by buying currently mined gold at the domestic free market price and selling it to the Central Bank at $35 per ounce. A number of countries have recognized that the payment of subsidies is an alternative to permission to sell at prices higher than the official price, so that a gold producer cannot be eligible for both. Some countries vary the share of profits that they obtain, and the profits that they permit gold mining companies to earn, through a multiple rate structure or differential taxation on the basis of size and ownership (domestic or foreign). In none of these countries is gold of major significance, either as a part of total production or of exports. Among the countries in the group, gold production is of greatest importance to the Gold Coast (Ghana), where it constituted 10 per cent of exports in 1956. In no other country (excluding South Africa) does gold constitute as much as 5 per cent of exports. Under these circumstances, the cost of subsidizing gold production can never be very large, while countries often consider that the advantages of continued production are great. These considerations, in the light of the conditions under which gold is produced, suggest that production is unlikely to fall below current levels.

Table 5.

Value of Gold Production in Selected Countries, 1933–56

(Valued at US$35 per fine ounce; in millions of U.S. dollars)

article image
Source: Data for 1933 and 1940 are based on figures in Board of Governors of the Federal Reserve System, Banking and Monetary Statistics (Washington, November 1943), Table 159. Data for 1948 and later years are from International Financial Statistics (Washington). Some of the data for 1957 are preliminary.

Includes Nigeria and Sierra Leone.

Not available.

Outlook for gold production

In summary, the major determinant of changes in world gold production in the next decade is likely to be production in South Africa. If output there remains at present levels, and production in all other countries does not decrease, world production in the next decade is likely to be of the order of $1,050 million per year. If South African production increases significantly in the years to about 1961, and slowly thereafter, and production in other countries does not decrease, total production in the next decade is likely to average $1,150 million per year. An increase in South Africa and a decrease in the rest of the world are likely to result in an average production of at least $1 billion per year. A depression in this period would increase these figures.

Additions to Gold Reserves in the Past

The proportion of gold production that has gone into monetary gold stocks has varied considerably from year to year. The proportion in any one year is not very significant (and in any case is very difficult to measure accurately), but the average proportion added to monetary stocks over long periods has great significance for any study of gold reserves. For the period 1835–1952, it has been estimated that the increase in monetary gold was equivalent to 70 per cent of gold production.46 During 1931–45, gold production was $14,455 million, and reported receipts of gold by monetary authorities from the U.S.S.R. were at least $365 million. Official gold reserves increased by $15,670 million, equivalent to 108 per cent of new production and 106 per cent of estimated new supplies. Thus, during this period there was a net outflow of gold from private channels of $850 million.47

The data for 1946–57 show that gold reserves increased then by $5.4 million, equivalent to 52 per cent of estimated production (Table 6). Additions were particularly small in 1951 and 1952; for the two-year period, they were equivalent to slightly more than 25 per cent of production. In 1954 and 1955, however, they were equivalent to 70 per cent; in 1956, to 50 per cent; and in 1957 to at least 70 per cent.

Table 6.

Value of World Gold Reserves and Production, Excluding Countries in Soviet Bloc, 1945–56

(In millions of U.S. dollars)

article image

Estimated world total, excluding the U.S.S.R. Source: Federal Reserve Bulletin (Washington).

Source: International Financial Statistics (Washington).

It is not unreasonable, at least as a first approximation, to assume that in the next decade gold additions to monetary reserves may average from one half to two thirds of gold production. The percentage has rarely been less than 50 over any substantial period, and it may be two thirds, or more, under certain conditions, which it may be helpful to consider briefly.

The variability of additions to gold stocks, in any year or in comparatively short periods, depends very much more on demand than on supply factors. The variability of gold production (leaving aside trend) is not great. On the assumption that there is no net dishoarding of gold on private account, the only other short-term variable factor affecting supply is sales of gold by the U.S.S.R. These sales are unpredictable, and they have never been measured with any accuracy. Incomplete tabulations for 1931–43 show that gold shipments by the U.S.S.R. to official recipients, excluding sales in gold markets, averaged $70 million per year. Such shipments were equivalent to 7 per cent of average annual world gold production.48 Estimates of U.S.S.R. shipments in postwar years vary widely, and none can be accepted as authoritative. The estimates that have been published suggest that sales in some years were negligible; that in a number of years they were of the order of $150 million; and that in 1957 they reached a postwar high of about $250 million. Sales since 1950 have been much larger than in the years immediately following World War II. It is not unlikely that for the 12-year period 1946–57 they may have come to $900–1,300 million.49 These figures are equivalent to 9–13 per cent of world gold production during the period.

Any forecast of gold sales by the U.S.S.R. in the next decade can be nothing but a guess, but such a guess would have to take account of these principal factors: (1) During the last 25 years, the U.S.S.R. has sold gold more or less continuously, but in differing amounts, to meet its trade deficit. (2) During the postwar period, the U.S.S.R. has commonly been considered the second largest producer of gold in the world. For the period ended 1956, its production, according to one estimate, was at least $150 million per year and was possibly $350 million; according to other estimates, its production was at least $420 million per year and was possibly $560 million. Estimates for 1957 are considerably higher; they place the U.S.S.R. production at $600 million, approximately the same as South Africa’s output.50 Its gold holdings may be $7 billion or more, and it cannot have any significant amount of short-term liabilities, as do the United States and the United Kingdom. Average yearly sales of $80–110 million per year (the average of the postwar period), or of $200 million per year (which may perhaps be the level of the last two years), would still permit a significant increase of Soviet gold reserves. Gold has the advantage for the Soviet Union that it can be sold in large amounts at guaranteed prices, and that it can be produced by labor generated in large part by internal political conditions. Given the role of gold in the Soviet economy—a war chest and a way to pay for foreign goods—and the chronic shortages of goods, particularly producers’ goods, there is little reason to expect Soviet sales to decrease. It is probably fair to say that the U.S.S.R. has seldom been able to buy as much from the West as it would have liked. (3) There are a number of factors that might lead to larger gold sales. A decrease in the tension between the Soviet bloc and the Western countries might make it possible for the Soviet Union to buy more of the goods it wants. An increase in the economic strains in the Soviet satellite states might make it necessary for the Soviet Union to finance larger imports for them. Increased economic aid to underdeveloped countries in which the U.S.S.R. wishes to play a larger role might make it necessary for the Soviet Union to finance imports of capital goods by these countries, some of which may have to come from the West. A stepping up of Soviet propaganda activities or subsidies might require selling more gold. None of these factors necessarily assumes any increased eagerness on the part of the West to increase trade with the U.S.S.R.

These considerations suggest that average sales of $100 million per year may not be unlikely in the next decade. Such a guess may well be too low. Sales of this magnitude would be lower than those in the past few years and, with allowance for changes in the prices of goods, much lower than in the 1930’s.

The demand for gold flows from its three uses: for money, for industry and the arts, and as a vehicle for saving or “hoarding.” Of these, in all probability, bona fide industrial and artistic demands are relatively the most stable. As shown in Table 7 (last column), net usage in the United States during the ten years 1946–55 averaged $80 million per year, but in four of these ten years it was less than $50 million per year.51 The latter figure may be a better approximation of bona fide uses than the former, since it was hardly possible, given the active premium gold markets in some years, to prevent all illegal attempts to acquire and export gold.52 On the assumption that the United States accounted for perhaps two fifths of the gold used for world industrial and artistic purposes, world demand would be of the order of $125–200 million per year.53 The expected growth of real income during the next decade should lead to some increase in this demand, but it is unlikely that this would be proportional to the growth of income. The increase is more likely to be of the magnitude of 3–5 per cent, implying an average demand for industrial and artistic purposes in the next decade of $130–210 million per year.54

Table 7.

Issues of Gold for Private Use in the United States, 1946–55

(In millions of U.S. dollars)

article image
Source: Annual Reports of the Director of the Mint (Washington).

The most variable component in the demand for gold is for the purpose of saving or “hoarding.” To some extent, this demand may be for jewelry or other fabricated gold products and thus be indistinguishable from bona fide industrial and artistic demands. It is difficult to measure with any accuracy the effect of saving or “hoarding” in any short period, and it is even more difficult to estimate its future course. This will depend largely upon the threats of war, with their effects upon prices and the stability of currencies; the prospects of monetary disorders (inflation and devaluation), if peace is maintained; the continuation of prosperity or the prospect of depression; the efforts of governments to eliminate, or at least prevent the growth of, private gold holdings; and such longer term factors as the spread of savings institutions and the tempo of development in underdeveloped countries.

On the whole, the period of the 1930’s was one during which private holdings of gold were reduced, the reduction being considerable in the Middle East and the Far East. There was, however, some increase in European holdings.55 In the postwar period, private holdings increased substantially: purchases in the Middle East and the Far East reconstituted holdings that had been reduced during the 1930’s; those in Europe further increased private holdings there. It would appear that in the entire period 1931–53 private holdings in a few countries of Western Europe increased substantially.56 In the last four or five years, the scale of hoarding appears to have been substantially lower than it was immediately after World War II and during the Korean war. It is perhaps not too hazardous to guess that in the postwar period the amount of gold that annually went into private hoards ranged from a low of about $200 million (1954–55) to a high of about $520 million (1951–52). For the 11 postwar years, the average was probably about $300 million per year.

The total amount of gold in private hoards is clearly very large. According to one estimate it is more than $12 billion, that is, approximately the same in value as world monetary holdings outside the United States and the U.S.S.R.57

Some of the private demand for gold must be regarded as normal, partly because it is a traditional form of saving and partly because many people are not accustomed to making any large use of savings institutions. This demand will continue. On the other hand, a substantial amount of the hoarding of the postwar period was associated with the hostilities in Korea, and the demand for gold from the mainland of China seems to have been sharply reduced.58 The satisfaction of the private demand for gold in underdeveloped countries with important development programs, e.g., India, is unlikely to be facilitated, and it may be further restricted legally. The possibility of inflation in many countries, especially underdeveloped ones, encourages hoarding, regardless of whether this inflation results from unduly ambitious development programs, or from spending to cope with international uncertainties, or from other domestic programs inconsistent with price stability. In addition, the private demand for gold in some Western European countries is substantial, and it appears to vary with the degree of political uneasiness and the fears of inflation and devaluation.

A substantial private demand for gold must therefore be expected to continue. Most of this private demand is met by smuggling or by other activities which are legally frowned upon, but some of it is satisfied by purchases which are officially sanctioned.59 It has been facilitated by private minting in Italy and official or “commercial” minting in France, Austria, Switzerland, Germany, and the United Kingdom. On the whole, it does not appear unreasonable to suggest that the movement of gold into private hoards in the next decade might be no larger than in the past one ($300 million per year), and that it might indeed be smaller. The probable limits of such a movement would appear to be from $200 million to $450 million per year.

Expected Additions to Gold Reserves, 1956–65

The preceding discussion has made it clear that there can be no categorical answer to the question, What increase of monetary gold stocks may be expected in the next decade, on the assumption that the present official price of $35 per ounce is continued? The best that can be done is to determine a range within which the actual result is likely to fall.

Average annual production of gold in the next decade may be assumed to fall within the limits of $1,050 million and $1,150 million. Additions to gold reserves may be assumed to be equivalent, on the average, to from one half to two thirds of gold production.60 These two sets of assumptions delimit the maximum and minimum additions to monetary gold stocks during the next decade as follows:

article image

The minimum figure—$525 million per year—represents an increase of 22 per cent over the average annual additions to gold reserves, about $430 million, during 1946–56.

The question may also be approached by combining the individual demand and supply elements already described to yield a minimum-maximum range. This is done in the following summary, where values are stated as annual averages for the decade:

article image

These figures add up to a minimum addition to monetary gold stocks of $500 million per year, and a maximum addition of $920 million per year. A probable value within this range would be of the order of $730 million, equivalent to about two thirds of gold production. Since average annual additions were $450 million in 1946–57, these figures for the next decade represent, as a minimum, an increase of 10 per cent over the past rate; as a maximum, a doubling of the past rate; and, as a probable value, an increase of 60 per cent over the past rate.61

At the end of 1957, gold reserves of all countries except the Soviet bloc were $37 billion; excluding the United States, they were $14 billion. Reserves of gold and official holdings of exchange (largely dollars and sterling) were, respectively, $53 billion and $30 billion. During the next decade, it may be expected that gold will increase by perhaps $7.3 billion, and that in any event the actual increase will fall within the range of $5.0 billion and $9.2 billion. These are equivalent to the following percentage increases in total country reserves during the decade:

article image

If the assumption is made that U.S. gold reserves will not increase during the next decade—or that any increase will be matched by an increase in official dollar balances held abroad—the percentage increases in reserves of the rest of the world will be as follows:

article image

Mr. Altman, Advisor in the Research and Statistics Department, is a graduate of Cornell University and the University of Chicago. He taught economics at Ohio State University and was on the staff of the National Resources Planning Board and the French Supply Council. He was Director of Administration of the Fund until 1954. He is the author of Savings, Investment and National Income and of a number of papers published in technical journals.


The longer the period, the more hazardous is the estimate. Yet a period of two or three years would not be meaningful for the present problem. Most of the writings on gold production agree that a decade seems to be as long a period as can be dealt with significantly. It does not seem necessary, for the purposes of this paper, to discuss whether the only “satisfactory” way to increase international reserves is to increase gold reserves, on the ground that increases in reserve holdings of foreign exchange (dollars, sterling, etc.) by some countries are accompanied by larger liabilities of others, so that exchange holdings cannot increase reserves on a “net” basis. This appears to be what Harrod had in mind when he said, “While these ad hoc expedients may serve a useful purpose for the time being, there is no doubt that in due course it will be desirable to devise a radical cure for the gold shortage.” (Roy F. Harrod, “Imbalance of International Payments,” Staff Papers, Vol. III, No. 1, April 1953, p. 5.) There is a discussion of this question in B. Goodman, “The Price of Gold and International Liquidity,” Journal of Finance (Chicago), Vol. II (1956), pp. 15–28.


The Economist (London), June 9, 1956, p. 1009.


The assumption, made in this paper, that the U.S. price of gold will continue to be $35 per ounce does not in any way suggest that an increase may or may not be desirable. The United States has stated many times that it does not intend to increase its buying price of gold. (See, for example, International Monetary Fund, Summary Proceedings of the Eleventh Annual Meeting of the Board of Governors, Washington, September 1956, p. 69.) A recent Canadian study noted, “In considering whether the United States Government is likely to raise the price at which it buys gold it should be remembered that the gold standard is a gold-price-fixing system and that the only change in the price of gold which has been made in the last century and a half was that made by President Roosevelt in 1933–34 when the United States had for the time being abandoned the gold standard. He hoped that by raising the price of gold a rise in the price of other commodities might be brought about also, thus easing the severity of the price deflation. It is conceivable that in circumstances as serious as those of 1933 such a manoeuvre might once more be tried by an harassed United States Government. To hope for a rise in the price of gold in a period of prosperity is quite another matter. … Although the continued purchase of gold by the United States Government at the present price can be relied on, it would be most unwise to count upon an early rise in the Washington price of gold.” (Report of the Committee of Inquiry into the Economics of the Gold Mining Industry, Gold Mining in Ontario, Toronto, 1955, p. 24.) Essentially the same point was made by the Bank for International Settlements in its Seventeenth Annual Report (Basle, 1947), p. 96. South Africa has argued on many occasions that the price of gold should be increased (e.g., IMF, Summary Proceedings, op. cit., pp. 62–66). Nevertheless, in a Report of the Transvaal and Orange Free State Chamber of Mines, it was stated, “Although many of us hold the view that the time will come when there will be an increase in the price of gold, I firmly believe that the Industry, in its current operating decisions, should continue to avoid assumptions regarding such a rise in price.” (Sixty-Sixth Annual Report, Year 1955, Johannesburg, p. 58.)


After surveying the gold discoveries since 1814, Kitchin concluded that “Averaging all these occurrences we have four alluvial discoveries, reaching their maximum in four years with £8 million each, four reef discoveries reaching their maximum in seventeen years with £5 million each, and one banket discovery [The Rand, 1886] which has lasted forty-three years without having reached the top and has so far attained £43,500,000.” (Joseph Kitchin, “Gold Production,” The International Gold Problem, London, 1931, p. 60.) The time needed to bring important mines into operation is strikingly illustrated in a press announcement dealing with a proposed ultra-deep mine, Western Deep Levels, Ltd. It was estimated that production from the shallowest level would start 6–7 years after the project was started; from the middle level, 9 years; and from the ultra-deep level, about 12,500 feet, 18 years. (The South African Mining and Engineering Journal, Johannesburg, August 2, 1957, pp. 1504–5.)


These points are suggested by the Anglo American Corporation of South Africa Limited in its 39th Annual Report (Johannesburg, 1955, pp. 18–19) which noted that “in a little under ten years, 11 of the 12 mines projected in the original Orange Free State development programme have come to production. Of these 11 producing mines, six are administered by the Anglo American Corporation. The magnitude of the achievement can best be illustrated by a number of globular statistics. During this development programme, 27 shafts have been sunk to an aggregate depth of over 22 miles and over 430 miles of underground tunnelling and development have taken place. In order to allow underground development to proceed on this scale, it has become necessary to pump water from underground and provide facilities for its disposal by means of evaporation at a rate of 4,000,000,000 gallons a year. The weight of water pumped to the surface in 1955 was 22,000,000 tons or nearly three times the tonnage of ore crushed during that year. Approximately £18,000,000 has been spent on housing for Europeans alone. The total population occupying the townships thus created and the accommodation provided on the mines is now over 130,000, of whom 8,000 Europeans and 50,000 Natives are in the direct employment of the mines. The mines themselves have so far involved a total capital investment of the order of £167,000,000, some £95,000,00 being for the mines of the Anglo American Corporation Group.” As the mines go deeper, total investment increases substantially, and new technical problems appear. The proposed ultra-deep mine, Western Deep Levels, Ltd., has an estimated capital cost of over £20 million, compared with an average cost of the present Orange Free State mines of about £14½ million. (The Economist, July 27, 1957, pp. 336–37.)


C. O. Hardy, when discussing the most important forecasts of gold production made for the 1920’s and the 1930’s, concluded that they were unduly conservative, partly for the reason that “estimates made by mining engineers on a basis appropriate for corporate purposes are inherently more conservative than the kind of estimates that are needed as a basis for economic policy.” (C. O. Hardy, Is There Enough Gold? Brookings Institution, Washington, 1936, Chap. 3, p. 61.) Kitchin’s estimates of production in the Rand were too low, as Hardy noted. Kitchen himself pointed out in 1930 that “past estimates of the Rand’s later production have been much too low”; he found that the average of ten estimates made between 1913 and 1926 of the value of 1929 production came to £30 million. Actual production was valued at £43.5 million, or 40 per cent more than the average estimate. (Kitchin, op. cit., pp. 60–61.) Kitchin as late as 1926 had estimated that South African production in 1930 would be £75–80 million, compared with £80 million in 1925. (Royal Commission on Indian Currency and Finance, Appendices to the Report of the Royal Commission on Indian Currency and Finance, London, 1926, p. 519.) In fact, production increased by 10 per cent. It should be noted that during these years the official price of gold remained the same but costs and prices increased substantially. In 1930, two estimates of South African production for 1930–40 were published. One was official; the other, by Kitchin, was about 10 per cent higher. Both were too low. Yet Kitchin had “the advantage of special knowledge” and calculated separately “the lives, outputs and prospects of thirty-nine mines (including six not yet producing).” (League of Nations, Interim Report of the Gold Delegation of the Financial Committee, Geneva, 1930, p. 60; see also p. 89.) A more recent illustration of the same difficulty in estimating production is the testimony in April 1954 of the President of the Homestake Mine, the largest in the United States. He testified that, despite the new mines in the Rand and the Orange Free State, he doubted that South Africa “would do much more than hold its own.” (Gold Reserve Act Amendments, Hearings Before a Subcommittee of the Committee on Banking and Currency, U.S. Senate, Washington, 1954, p. 264.) Since then, the rate of South African production has increased by more than 40 per cent. All of this underscores the difficulties and the hazards of making any estimates, including those made in this paper.


The net increase in the costs of gold mining reflects increases in domestic prices (which in most cases have involved devaluations), changes in productivity, changes in the grade of ore, and other factors. Canadian costs, which were affected to only a minor extent by changes in the external value of the Canadian dollar, changed as follows:

article image

In the postwar period, the grade of ore mined has remained relatively unchanged, and the variable costs per ton of ore and per ounce of gold recovered have tended to move together. This was not the case in the earlier period, when the variable costs per ounce increased faster than the variable costs per ton. Furthermore, in 1954 the difference between the selling price and variable cost was Can$9.66, or Can$3.72 less than in 1946. Of this, Can$1.03 was due to higher variable costs and Can$2.69 to a change in the Canadian dollar relative to the U.S. dollar. (Report of the Committee of Inquiry, op. cit., pp. 107 and 110.)


South African Reserve Bank, Annual Reports, 1952–57 (Pretoria).


Thus, the Chairman of the Transvaal and Orange Free State Chamber of Mines noted in 1955 that in 1949 “our gold producers received a higher price for their gold in terms of sterling. The leaders of the Gold Mining Industry … emphasized that, welcome as the relief to the Industry was, everything must be done to prevent rising working costs from absorbing the benefits of this relief. Yet, despite this warning, cost factors beyond the Industry’s control absorbed these benefits all too soon and the Industry is again faced with the dilemma of rising costs against a fixed price for its product.” (Sixty-Sixth Annual Report, Year 1955, pp. 57–58.)


Silver is associated with gold to some degree in gold deposits and is a co-product of gold-placer mining and lode-gold mining. Crude platinum is often present in gold placers; at some areas in Colombia, enough is present that it becomes a coproduct. Other platinum-group metals are present in very small quantities in some lode-gold ores. Uranium is present in small quantities in South African ores. Uranium and thorium bearing minerals are present in some placers. (Department of the Interior, Bureau of Mines, Mineral Facts and Problems, Bulletin 556, Washington, D.C., p. 5.)


Report of the Committee of Inquiry, op. cit., p. 44. As the South African Reserve Bank stated in 1953, “… a world recession or depression would benefit our gold mines by lowering their production costs, if not also by raising the price of their product, and to the extent that contraction and unemployment should take place in other local industries, the gold mines would be able to obtain more labour and power and thus to produce more in view of their substantial surplus productive capacity.” (Report of the Thirty-Third Ordinary General Meeting, pp. 21–22.)


C. 0. Hardy, op. cit., pp. 67–68.


John W. Shilling, “An Outline of Gold Mining Taxation, South African Law Journal (Cape Town and Johannesburg), May 1950, p. 152.


Bureau of Census and Statistics, Official Yearbook of the Union of South Africa (Pretoria), No. 27 (1952–53), p. 605, and No. 28 (1954–55), pp. 402–3.


The South African Mining and Engineering Journal, May 11, 1956, p. 714, and May 25,1956, p. 781.


The nature of the tax formula and the tax rates resulting from it are discussed in footnote 19.


The South African Mining and Engineering Journal, July 26, 1957, p. 1497.


Ibid., May 25, 1956, p. 781.


The tax formula is of the type y = a-b/x, where a and b are constants and x is “the ratio, expressed as a percentage, which taxable income derived from mining for gold bears to the income derived therefrom.” The formula for 1956 is y = 60–360/x. With these formula constants, the rate applied to all the taxable income of a company whose ratio of net to gross is 30 per cent will be 60 - 360/30, or 48 per cent. The rates of income tax as the formula constants were changed from time to time are shown below, and applicable tax rates for other kinds of business enterprises are shown for comparison.

article image

The minimum of 15 per cent shown in the above tabulation was applied to mines with profits of £7,500 and more per year; that of 20 per cent, to those with profits of £20,000 and more. Below these profits, i.e., for small or unprofitable mines, the formula is applied but a maximum tax rate is specified. Gold mines in South Africa are subject to a number of other payments, including one equivalent to a royalty or leasehold payment. This has been in effect for many years. The payment rates are determined separately for each mine on the basis of the same type of formula that is used in determining income tax rates. The leasehold payments formula is negotiated at the time the mine begins operations. The constants in the formula are such that, for many years, royalty payments have been substantially smaller than income tax payments. Such payments are apparently considered as a cost in determining taxable net income. In the case of the proposed Western Deep Levels, Ltd., the leasehold formula is y=l5–90/x. If the ratio of net to gross income should be (say) 50 per cent, the leasehold payment would be 13.2 per cent of net income before income tax. (The system of leasing gold areas is described in the Official Yearbooks of the Union of South Africa; see No. 28, 1954–55, pp. 583–85.)


Anglo American Corporation of South Africa Limited, Chairman’s Statement, 1957 (Johannesburg, 1957), p. 11.


The income derived from mining gold includes, by statutory definition, income derived from silver, osmiridium, or other minerals which may be won in the course of mining gold, or which result directly from mining gold. In recent years, the most important other mineral has come to be uranium; and about half the gold in South Africa comes from gold-uranium producers. Should profits from uranium, when this is the dominant product, or a major coproduct, be taxed at the gold mines rate, or at the rate for base metals mines? Apparently, in one recent case, uranium profits from a gold mine whose gold deposits had become uneconomic were taxed at the base metals rate of 30 per cent.


South Africa Social and Economic Planning Council, Economic Aspects of the Gold Mining Industry, Report No. 11 (Pretoria, 1948), pp. 12 and 40. For selected references to the element of discrimination, see First Report of the Committee of Enquiry Into the Income Tax Act (Pretoria, 1951), and the review of this report by J. Lavine, “Some Aspects of Enquiry Into Income Tax Act,” South African Journal of Economics (Johannesburg), June 1952, especially pp. 138–39. Also, Report of the Committee on Gold Mining Taxation, which notes “the long-standing practice of taxing gold mining companies at a higher rate than other companies” and considers it justified (Pretoria, 1945, par. III; see also pars. 125–37). An older but very interesting discussion is in W. J. Busschau, The Theory of Gold Supply (Oxford, 1936), especially Chaps. VI and VII.


See comments by the Minister of Finance reported in The South African Mining and Engineering Journal, May 25, 1956, p. 781. The Minister stated that, of the 46 mines quoted on the Johannesburg Stock Exchange, 21 paid a tax on profits and 25 did not. The Report of the Committee on Gold Mining Taxation (op. cit., par. 163) said that “the ratio of profit to recovery is the sound basis for a sliding scale. … It is the rich mines which should contribute most, not only to the normal cost of governmental services, but also to the replacement in more permanent form of the country’s patrimony which is gradually being exhausted.”


The Economist, June 9, 1956, p. 1011. This point was strongly underscored when a new mine that started operations in March 1956 suspended operations seven months later because of flooding. Operations had not been resumed in November 1957.


Anglo American Corporation of South Africa, op. cit., p. 19. Also reported in The Economist, June 9, 1956, p. 1010, and Barron’s (New York), June 11, 1956, p. 9. The South African Mining and Engineering Journal, March 30, 1956, p. 415, reported that 19 new mines, all either milling or due to start in 1956, were milling 1.1 million tons per month, compared with an expected maximum of 2.4 million. The Standard Bank of South Africa reported that “there has been a sharp rise in production in the Orange Free State, and it would appear that the limit of the upsurge is still far distant” (The Standard Bank Review, London, September 1956, p. 12).


Data compiled from The Standard Bank Review, March 1957 and September 1957, and from Barron’s, June 11, 1956, p. 24, indicate the following changes:

article image


Certainly the most dramatic prospect announced to date concerns the proposed Western Deep Levels, Ltd. According to the announcement, the mine will have a capacity output of 1,440,000 ounces (about $50 million) per year, or more than 8 per cent of present South African output. It is expected that the life of the mine will be unusually long, about 60 years. (The South African Mining and Engineering Journal, August 2, 1957, pp. 1504–5; The Economist, July 27, 1957, pp. 336–37; Anglo American Corporation of South Africa, Chairman’s Statement, 1957, p. 6.)


Output declined from 13.8 million ounces in 1941 to 11.0 million ounces in 1953. It thereafter increased gradually to 12.8 million ounces in 1957.


The Standard Bank Review, February 1956, p. 13.


The following figures have been compiled from data of the Transvaal and Orange Free State Chamber of Mines, published in Annual Reports, 1954–56, and Summary, Third Quarter of 1957.

article image


“The extraction of uranium in South Africa is at present conducted as a by-product of gold mining, for the concentration of uranium in gold-bearing ore is low and the most economic way to produce it is to have gold bear the cost of crushing and mining the ore.” (Naomi Eliovson and Beelah Eliovson Long, Industrial South Africa, Johannesburg, 1956, p. 27.) An article, “Yardsticks for Uranium Shares,” in The Economist, January 19, 1957, p. 231, noted that “with those qualifications the profit per ton, even after amortisation of the plant, is high—comfortably over £110s. per pound of oxide. It is doubtless that margin of safety plus the fact that uranium is a by-product with few direct mining costs of its own that gives the directors the courage to forecast a continued market for South African uranium when the industry, with its plant written down to nothing but without a guaranteed market, has to go out and find a buyer.” This procedure for allocating cost can, of course, be modified depending upon the relative profitability of gold and uranium. This might have the effect of reducing the calculated cost of producing gold.


The South African Mining and Engineering Journal, June 14, 1957, p. 1145.


In any case, this would be the result of the normal mining practice of mining the average grade of ore reserves. When costs rise, ore which can no longer be mined economically is dropped from reserves. The average grade of ore reserves therefore rises, and the mine works toward this higher average. On the contrary, “In 1950 the higher price for gold enabled the industry to include a large amount of previously unpayable reef in its ore reserves, with a consequent reduction in the average grade of these reserves and, in conformity with the industry’s practice of mining to the average value of payable ore available, the yield fell to 3.759 pennyweights per ton. This yield was maintained until 1952 when, owing mainly to the exclusion of marginal ore through increased costs, it began to rise.” (Transvaal and Orange Free State Chamber of Mines, The Mining Survey, Johannesburg, September 1955, p. 28.) It is generally considered that any gold mine is “vulnerable” to further increases in costs if its profit margin is less than 5s. Od. per ton of ore milled. In June 1957 there were 14 mines in this position. (The Standard Bank Review, September 1957, pp. 12–13.)


Transvaal and Orange Free State Chamber of Mines, Annual Report, Year 1956, pp. 56 and 58.


The South African Mining and Engineering Journal, March 30,1956.


The Standard Bank Review, February 1956, p. 13. Another expression of this view is that “a shortage of manpower … has forced the industry to work below capacity and further increased the unit costs of production. This problem is not confined to gold-mining. It is a general industrial problem in South Africa but, as far as gold-mining is concerned, the effects of the shortage of trained men are aggravated by the expansion of operations being undertaken by the industry. There is also a shortage of Native labor, although recent action taken to meet this latter shortage has produced encouraging results.” (“The Union of South Africa,” Statist, London, February 1956, p. 16.)


The South African Mining and Engineering Journal, March 30, 1956, p. 415. There is a discussion of labor efficiency and productivity in the issue of July 27, 1956, pp. 121 ff. The labor situation in South Africa is complex. More than 60 per cent of the native labor in 1956 came from neighboring countries; and these countries impose quotas on the number of laborers who can be recruited. Labor turnover is high, and this not only affects efficiency but also involves additional costs for recruiting, testing, training, and transportation. (South Africa Social and Economic Planning Council, op. cit., pp. 25–34.) The dimensions of these problems are suggested by the fact that native laborers from outside South Africa are usually engaged for 12 months, after which they must return to their homes unless they are permitted to re-engage for further periods, usually not exceeding 6 months. The extent of native turnover is suggested by the fact that, although the average labor force in 1954 was less than 315,000, recruitment totaled 178,000 from the Union, Swaziland, Basutoland, and South Bechuanaland, and 125,500 from other areas. (Official Yearbook of the Union of South Africa, No. 28, 1954–55 pp. 571–72.)


For example, the use of native labor in the mining industry is limited by law to unskilled work. (Transvaal and Orange Free State Chamber of Mines, Annual Reports, 1954, pp. 59–61; 1955, pp. 59–65; and 1956, p. 61.) A recent article stated that “racial and craft barriers, both formal and informal, are likely to be the chief obstacles to be overcome” in attaining greater mechanization and automation. (The South African Mining and Engineering Journal, December 28, 1956, p. 1107.)


A recent article noted that “in practice the system undoubtedly results in the labour position being tighter at some mines than at others, and indeed there have recently been cases where one or two of the older mines have reported actual labour surpluses. However, in the long term these discrepancies get adjusted and there seems no reason to think that the quota system is giving any widespread dissatisfaction.” However, “one or two [declining] mines are now trending towards a small scale basis of operation, and it might well be advantageous for others to do so. …” (The South African Mining and Engineering Journal, March 30, 1956, pp. 413 and 419.)


The data in this subsection are from Transvaal and Orange Free State Chamber of Mines, Annual Report, Year 1956, and First Quarterly Report, 1957.


In terms of dollars, the working profit margin is approximately the same now as it was in the second half of the 1920’s, but only about 55 per cent of what it was in 1902. Adjusted for changes in both prices and the rate of exchange, the external purchasing power of the profit margin is now about half of what it was in the late 1920’s and less than one fifth of what it was in 1902.


See International Monetary Fund, Summary Proceedings, op. cit., p. 64.


Department of the Interior, Bureau of Mines, Minerals Yearbook 1953 (Washington), pp. 1–13; Mineral Facts and Problems, op. cit., pp. 1–12; and preprint from Minerals Yearbook, 1954, p. 11.


“Usually the mines cannot return to the former smaller-scale, higher-grade operations as carried on before 1934. The expansionary process cannot easily be reversed.” (Report of the Committee of Inquiry, op. cit., p. 16.)


Northern Miner (Toronto), June 14, 1956, August 29, 1957, and November 25, 1957. Selected data on the assistance program are as follows:

article image


W. J. Busschau, “Some Notes on Gold Production and Stocks,” Shall We Return to the Gold Standard—Now? (National Industrial Conference Board, New York, 1954), p. 163. Data for 1930–39 are given in his The Measure of Gold (Cape Town, 1949), p. 92. Kitchin estimated that the addition to monetary stocks during the period 1835–1929 was equivalent to approximately 56 per cent of production, and that additions in selected subperiods were equivalent to the following percentages of production: 1835–89, 50 per cent; 1890–1929, 58 per cent; 1920–29, 54 per cent. Joseph Kitchin, op. cit., p. 49. Data for the period 1930–52, given by Busschau, suggest that the increase of monetary gold was equivalent to 90 per cent of gold production, and that in the subperiod 1930–39 it was greater than gold production.


“The Private Demand for Gold, 1931–53,” Federal Reserve Bulletin (Washington), September 1954, p. 938.




There have been many estimates of Soviet gold sales, of which only a few can be noted. Samuel Montagu and Co., Ltd. (London) discussed them in their Annual Bullion Reviews for 1953–56. The Bank for International Settlements reported various estimates in its Twenty-Sixth Annual Report (Basle, 1956). Franz Pick’s estimates are published in a number of places, including the daily press.


These estimates for the period ended 1956 are from Northern Miner, November 29, 1956, and The Gold Mining Record (London), January 1956, p. 21. Other estimates include those of Samuel Montagu & Co. of “probably more than 10 million ounces” (Annual Bullion Review, 1956), or $350 million, and a German estimate of 12 million ounces, or $420 million (The Financial Times, London, February 11, 1956). The figure of $420 million was also reported by the South African Mining and Engineering Journal (April 6, 1956, p. 469). The Wall Street Journal said that output was at least $450 million (New York, November 20, 1957). Samuel Montagu & Co. estimated production in 1957 at 17 million ounces, or almost $600 million. (Annual Bullion Review, 1957.) Probably the last official Russian report of production was $190 million for 1937. As reported August 15, 1957, the new economic chief of the Yakut region declared that by 1960 he hoped to increase gold output sharply, mostly concentrated in the former Lena fields of Yakutsk. (Northern Miner, August 15,1957.)


In contrast, estimates of the U.S. Bureau of the Mint indicate that during 1931–37 returns of scrap gold, etc., exceeded new issues for industrial purposes by $15 million per year. For the world as a whole, industrial consumption during that period was estimated at $566 million and returns from scrap at $799 million (Federal Reserve Bulletin, August 1937, p. 704). The South Africa Social and Economic Planning Council, (op. cit., p. 42), citing the Bank for International Settlements, Eighth Annual Report, stated that “between 1931 and 1936, however, industrial consumption fluctuated between 2 and 4 million fine ounces annually … and the demand could usually be met by the return of coin and scrap gold.” In the late 1920’s, net industrial and artistic consumption was estimated at about $100 million per year ($165 million at $35 per ounce) by Loveday (League of Nations, op. cit., p. 90).


For example, during periods of active gold buying for private use, there were reports from time to time of operations to smuggle gold out of the United States. The New York Times, September 30, 1951, reported that U.S. authorities had cracked a smuggling ring, consisting of 50–60 persons, engaged in shipping $100 million of gold a year to black markets in Uruguay, Tangier, Bombay, and Rotterdam. It had reported several cases of smuggling the year before (issues of September 26 and 28, 1950).


There is one estimate that the prewar ratio was 50 per cent (Federal Reserve Bulletin, September 1954, p. 937). Estimates of world industrial uses for each year from 1947 through 1955, published by the Bank for International Settlements in its Twenty-Sixth Annual Report, show a range from $120 million in 1947 to $210 million in 1955, and an average of $170 million. The U.S. Bureau of the Mint publishes data on consumption of gold for industrial and artistic purposes based on questionnaires to a large number of governments. The reported data admittedly reflect a wide range of definitions, methods of estimation, and accuracy. The data for 1953, one of the more complete recent years, show consumption of $165 million for 33 countries, of which the United States is charged with 45 per cent. (Annual Report of the Director of the Mint for the Fiscal Year Ended June 80, 1955, Washington, D.C., p. 110.) Corresponding data for 1955 show $180 million for 29 countries, with only 26 per cent attributed to the United States. (Annual Report, 1956, p. 84.) On the other hand, the data for 1952 show $160 million for 32 countries, with 60 per cent attributed to the United States. (Annual Report, 1954, p. 120.)


Kitchin noted, in 1935, that industrial consumption of gold had not grown as rapidly as output of gold. (Article on “Gold,” Encyclopedia of the Social Sciences, Vol. VI, 1935, p. 691.) The BIS estimates referred to above (footnote 53) show a small upward trend. All estimates must be qualified in the light of the great difficulties of definition and estimation.


For example, an unpublished study by I. G. Patel shows liquidation of private holdings in India each year from 1931 through 1941, with the general conclusion that during the years studied, 1925–42, the demand for gold was highly responsive to changes in the relative price of gold.


“The Private Demand for Gold,” Federal Reserve Bulletin, September 1954, pp. 935 and 938, and an earlier article in the Bulletin of August 1937. There was a net increase of $2,350 million in private holdings, excluding those in Eastern countries but including non-U.S. industrial uses. If these industrial uses were of the order of $900 million, i.e., the maximum it could have been in the United States, the net increase in Western European private holdings during 1931–53 was of the order of $1.5 billion.


This estimate by Franz Pick distributes the estimated $12.4 billion of private hoards in 1957 by areas as follows (in billions): France, $3.9; other Europe, $1.8; Asia, $2.9; Africa, $1.4; Western Hemisphere, $1.6; all other, $0.8. (Northern Miner, November 14, 1957.)


Information on the movement of gold into “hoards” is necessarily conjectural. The comments by Samuel Montagu & Co. in its Annual Bullion Reviews are of interest: In 1951, the principal consumer markets were on the European continent, “where the demand is stimulated by the recurrent fears of devaluation,” and in the Far East, “where the demand is due in part to political fears connected with the Korean War” (Review, 1951, p. 6); in 1952, the main support of the free market demand was “the French hoarding demand,” but the same year witnessed the “obituary of the former active Chinese gold markets whose insatiable demands only a few years ago forced the free market prices to over $50 per ounce” (Review, 1952, pp. 3–4); and in 1955, the purchases of gold for hoarding were over 9 million ounces ($315 million at mint prices), of which Europe took 5 million (“the greater part of this gold eventually found its way into France”), and the Middle East took 2 million, largely for Indian account (Review, 1955, p. 4).


As to the latter, German banks in April 1957 started selling gold bars and coins to private individuals and others. The law permitting private purchase and ownership was one move to help absorb domestic purchasing power. Such sales in the past year are reported to total $100 million, and 90 per cent of the turnover was in coins. Coins sold at premiums ranging from 15 per cent to 50 per cent, depending upon their scarcity. Gold bars were reported as being sold close to their bullion value. There was also a gold sales tax of 4 per cent. (The New York Times, March 4, 1958.)


It is interesting to note that the Gold Delegation in 1930 estimated the proportion of gold production that would be available for monetary purposes in 1930–32 at about 55 per cent. (League of Nations, op. cit., p. 16.)


All of these estimates assume that the world will continue to increase its hoards of gold. If the course of events should bring greater confidence in the stability of prices and of exchange rates, gold hoardings might possibly be reduced in some years and gold reserves might grow by more than gold production. Furthermore, a serious depression would probably result in dishoarding of gold.