1. The major fluctuations in gold production in this century can be explained in substantial part by economic influences. Outside South Africa, where the combined output of all other gold producing countries now accounts for only about one fourth of the world total (excluding Russian output), the dominant influence has been the movement in the general level of world prices in relation to the international gold price. Response to this relative price influence in South Africa has been far weaker and has been subject to long time lags, with the main impact on investment decisions. In the medium term, movements in relative prices have influenced the profitability of South African gold output with less marked effects on the level of output. New discoveries of gold in South Africa have been important since World War II; economic influences have played some part in the pace of their exploitation. The underlying influence on South African gold output has been its relatively high productivity in relation to other South African industries, even after the almost continuously adverse movement in the “real,” or relative, gold price in terms of domestic currency since 1950 and in terms of U.S. dollars since the mid-1930’s.


1. The major fluctuations in gold production in this century can be explained in substantial part by economic influences. Outside South Africa, where the combined output of all other gold producing countries now accounts for only about one fourth of the world total (excluding Russian output), the dominant influence has been the movement in the general level of world prices in relation to the international gold price. Response to this relative price influence in South Africa has been far weaker and has been subject to long time lags, with the main impact on investment decisions. In the medium term, movements in relative prices have influenced the profitability of South African gold output with less marked effects on the level of output. New discoveries of gold in South Africa have been important since World War II; economic influences have played some part in the pace of their exploitation. The underlying influence on South African gold output has been its relatively high productivity in relation to other South African industries, even after the almost continuously adverse movement in the “real,” or relative, gold price in terms of domestic currency since 1950 and in terms of U.S. dollars since the mid-1930’s.

Summary and Conclusions

1. The major fluctuations in gold production in this century can be explained in substantial part by economic influences. Outside South Africa, where the combined output of all other gold producing countries now accounts for only about one fourth of the world total (excluding Russian output), the dominant influence has been the movement in the general level of world prices in relation to the international gold price. Response to this relative price influence in South Africa has been far weaker and has been subject to long time lags, with the main impact on investment decisions. In the medium term, movements in relative prices have influenced the profitability of South African gold output with less marked effects on the level of output. New discoveries of gold in South Africa have been important since World War II; economic influences have played some part in the pace of their exploitation. The underlying influence on South African gold output has been its relatively high productivity in relation to other South African industries, even after the almost continuously adverse movement in the “real,” or relative, gold price in terms of domestic currency since 1950 and in terms of U.S. dollars since the mid-1930’s.

2. The depressive impact of the declining relative gold price on gold production outside South Africa has been checked in part by subsidies. In the medium-term future, the impact of subsidies on gold output might be expected to wane at a number of points; at the same time, the increase in the proportionate share of gold that is mined as a by-product, which has accompanied the decline in gold output in these countries in the past, may be expected to continue and to provide some cushioning against a further decline. Between 1960 and 1966 gold output outside South Africa declined by an annual average of 3 per cent. A continued decline seems likely. This assessment is made on the basis of an international gold price of $35 a fine ounce, and assuming a continuation of a gradual increase in world prices as a whole.

3. In South Africa, the impact of the falling relative price on profitability was countered until 1966 in the major section of the industry by the improved availability of high-grade ores. The waning of this favorable “resource” influence has increased the pressure for the impact of the lower relative gold price to be absorbed in other ways. These have included some reshaping of fiscal arrangements, and also moves toward improved disposition of the labor force. The fiscal alleviations have thus far been selective, comprising relatively modest subsidy payments aimed at maintaining production at the hardest pressed mines, on the one hand, and valuable tax relief for major development expenditures, on the other. The effect of rising cost levels in increasing the minimum grade of ore that companies find it economic to develop in a new mine can be, and has been, offset at least in part by these fiscal alleviations. Additional such devices, and perhaps also some general alleviation of gold-mining taxation, could conceivably be found economically advantageous in the years ahead. On the basis of present fiscal arrangements, and assuming continued flexibility in the application of fiscal policy with a view to maintaining existing production and encouraging development expenditures, South African gold output may be maintained at or somewhat above the 1966 level for the next few years. This assessment corresponds roughly to the projection of the official development plan.

4. Over the longer term, projections made by the mining industry and by the South African authorities point to a sharp decline in South African gold output. These projections are based on known ore reserves of existing mines, on the economy of working those reserves on the basis of forward projections of a continued modest increase in cost levels at an unchanged selling price, and on the basis of existing fiscal arrangements. Quantitative projections of future output must necessarily rest on a given framework of conditions such as these that are partly external to the industry. Actual future output will depend, inter alia, on the degree to which these conditions remain as assumed. Thus, in the absence of a change in these conditions that had the effect of halting or reversing the decline in the net revenue of South African gold producers, a substantial and sustained fall in South African gold output is to be expected. To put the matter another way, avoidance of a substantial decline in South African gold output may be dependent on an increase in the net revenue of South African gold producers. This could in principle occur indirectly through additional government support,2 or through a marked reduction in costs, or directly through an increase in the rand price of gold.

5. The pressure for deliberate action to be taken by the South African authorities in one or more of these directions will be governed essentially by the degree of success of other South African industries in fortifying the trade balance. The bulk of South Africa’s merchandise exports consists of primary products, which in the past 15 years or so have on balance experienced slightly less, rather than more, favorable price trends than gold at its fixed price. On the basis of the present export-import structure, and recent trends in it, there might be little reason to assume that the comparative advantage of gold mining is declining to any marked degree; and in this measure, even a continuing increase in South African gold production, in response to stimuli such as those noted above, should not be ruled out altogether. While South Africa has made important strides in diversifying its economy internally, the net contribution of manufacturing to the trade balance has been relatively disappointing heretofore. Considerable efforts are, however, being made to improve this contribution, and South Africa’s industrial base may now have grown to the scale at which the external competitiveness of its industries can be expected to increase. A further significant fall in the relative price of gold might therefore be expected to erode the comparative advantage of gold mining in the future, even though adverse relative price movements may not have had this effect hitherto. On this reasoning, a long-run decline in South African gold production is unlikely to be avoided, given the present international gold price. But the decline may not be as sharp as is sometimes predicted. For world gold production as a whole (always excluding the U.S.S.R. and other nonreporting countries), this prognosis implies approximate stability for the short term, with a probable decline after that. But even this general prognosis is subject to a wide margin of uncertainty, relating particularly to resource availability in the world at large and to general political and economic influences on the South African economy.

6. World gold production as a whole in the years ahead seems likely to be rather less sensitive to changes in the relative gold price than it has been in the past. Outside South Africa, this sensitivity has been very strong in the past, though it has been modified somewhat in recent years by subsidies and in future may be similarly affected by the increased role of by-product gold. Moreover, as the share of non-South African production has declined, from over half the world total in both the mid-1930’s and the early 1950’s, to one fourth currently, the proportionate weight of this more sensitive sector has declined correspondingly. In South Africa, for a number of reasons related in large part to the importance of gold mining as a source of external earnings, gold production may continue to be markedly less responsive to changes in relative gold prices, both in an upward and a downward direction, and as a result of changes in either the gold price itself or the general price level. This outcome is far from assured: the critical factor is likely to be the availability on an economic basis of replacements for gold production as a support for the balance of payments. Thus, if past efforts to prepare for the eventual anticipated running down of the gold-mining industry bear fruit in the emergence in a smooth way of the necessary additional exports or replacements for imports, then a steady reduction in the level of gold production, reaching substantial magnitudes over a decade or so, is to be expected in response to a falling relative gold price; any attempts to check the running down of gold mining in these circumstances would clearly be of questionable economic merit. But equally, insofar as loss of gold output at the margin would necessitate alternative means of supporting the external balance more costly in terms of real resources than the forgone gold production even at the lower real gold price, the volume of gold production that could rationally be judged as economically justified might not fall substantially for many years, even at a falling real international gold price. Correspondingly, any improvement in the real international gold price could have its major impact, at least in the initial period, on reducing the need for supporting the profitability of gold mining by indirect and local methods. Over the short term and perhaps also the medium term, therefore, there can be no assurance that a change in the real international gold price in either direction would have significant effects on the volume of South African gold output. Projections based on a static conception of surrounding economic conditions, i.e., assuming that changes in gold-mining factor costs will have no consequent indirect effects on gold-mining net revenue, are of doubtful applicability for an economy in which the role of gold mining is central. Over the longer term, however, this special influence is likely to lose force, and the more general influence of relative price movements on gold output is then more likely to predominate.

Influences on Gold Production

Attempts to project the future trend of gold production have a notorious record of inaccuracy; there has been a persistent tendency to underestimate.3 This paper focuses on general influences on gold production rather than putting primary emphasis on a comprehensive quantitative projection, though the assessment is in quantitative terms wherever this appears meaningful. The major influence considered is the depressive impact on gold production of a fixed international selling price at a time when other prices are rising, with the additional effects on gold-mining profitability of subsidies, indirect assistance, and government policy in taxation and other relevant spheres. Analysis of the past response of gold production to relative price changes does, however, provide certain elements relevant to hypothetical calculations of the future response of gold production to any increase in the international gold price. These elements are considered briefly at a few points; but they are incidental to the central theme, which is the outlook for gold production at the price of US$35 a fine ounce.4 No assessment is made of Russian gold production, which in the terms of this paper involves the double uncertainty of the conditions of production and of the weight given to the various factors determining it.

I. General Influences: Introductory Considerations

Over the medium term, the level of gold production is determined by four main factors: (1) The discovery or continued availability of gold-bearing ores, changes in their gold content (grade), and physical ease of mining—a composite factor that will be termed “availability” throughout this paper; (2) the state of mining technology; (3) the level of gold-mining factor and material costs, i.e., abstracting from changes in technology and availability as just described; (4) net revenue, governed by the gold price in local currency, as modified by indirect revenue through special government assistance, on the one hand, and certain forms of differential taxation, on the other. An additional influence is the net revenue obtainable from commodities produced in joint supply with gold.

Broadly speaking, since the world gold-mining industry entered its modern phase 80 years ago with the exploitation of the South African Rand, the influence of the first two factors independently of the latter two “financial” influences has waned. The major fluctuations in gold production in at least the first half of the twentieth century can be explained in terms of “relative profitability” on the crudest global measure, i.e., of fluctuations in the general level of world prices (as an influence on relative gold-mining factor and material costs) in relation to the “international” gold price (as an influence on net revenue).5 In Chart 1, to be discussed in more detail below, this relationship is expressed in terms of a “real dollar price of gold,” i.e., the official price of gold in U.S. dollars divided by an index of U.S. wholesale prices, or strictly the price of gold in terms of wholesale commodities in the United States.6 This indicator is seen to have moved fairly closely in line with world gold output from the eve of World War I to the aftermath of World War II. The statistical relationship admittedly disappeared and even reversed itself in the most recent phase, the major expansion in world gold output since the mid-1950’s. The main development expenditures on which this expansion has been based—the opening up of new mines in new gold producing areas of South Africa—were set in train in the 1950’s, when relative profitability in terms of local currency had been boosted by the devaluation in 1949, though the major factor was a highly favorable resource influence in the availability of extensive new deposits of high-grade ores. This special resource influence, together with the time lag in development, represent wasting economic assets. While the world gold price is assumed to remain stable in a period when other prices—and therefore gold-mining costs for any given level of availability and technology—are expected to rise, there is a presumed tendency for future financial influences on gold production to be contractionary.

Chart 1.
Chart 1.

Gold Production and a Real Dollar Price, 1910–66

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

Sources: Tables 16 and 17.

In assessing the force of this general financial influence, however, one major distinction can usefully be made. There are prima facie reasons to believe that this influence will be strong where gold mining is one industry among many. It is convenient to define this condition in negative terms, i.e., where either (1) the role of gold mining in the domestic economy is not so predominant that its factor prices, material costs, and contributions to government revenue are to a significant degree endogenously determined, or (2) its role in export earnings is not so significant that its own level of output could become a decisive influence on the level of the exchange rate or on government measures designed to produce some of the effects of exchange rate adjustments.

Where, by contrast, gold mining is a large element in the economy, the situation is less clear cut. Two general influences may then break the “normal” link between relative profitability, in this global sense used above, and gold output. First, not only gold mining but all the economic entities associated with it may initially be earning “excess” revenue, with some of the qualities of economic rent. That is, net revenue from the mining of gold at its fixed price may provide a surplus over the opportunity cost of the resources used. In economies in which the role of gold mining is small, this surplus will ordinarily show itself directly in exceptionally high profits or royalties. But where gold mining is a large element in the economy, such a surplus may be distributed in a number of ways: in taxation and other payments to government, in wage rates notably above those that would be set by marginal productivity in other industries, in differentially high payments for materials or other charges such as freight rates, as well as in super profits. Now, whatever the proportions in which this surplus has been retained by the industry or shared with its suppliers, workers, and the government (in accordance with relative economic and political strength), the super profits, actual or latent, can be expected to provide a cushion in the event of a decline in profitability.7 And to the extent that super profits have been built in to the industry’s cost level, costs will in some degree have become a function of profits. A fall in profitability may then partly reduce costs, and quasi costs such as taxation, before it has an influence on production by raising marginal cost above marginal revenue.8

These are examples of market influences that can be regarded as independent of specific government action, though such action may be needed in greater or less degree to reduce taxation in line with profitability, and to eliminate any differentially high rate of taxation, depending on the tax formula used. In addition, government policy may, of course, be invoked to stimulate these market influences, or to reinforce them. In particular, if gold mining contributes importantly to the balance of payments, the government is likely to weigh the opportunity costs, in terms of foreign exchange, of countenancing a major fall in this contribution. Such an assessment could, conceptually, suggest an alleviation in taxation beyond the mere elimination of any earlier excess taxation in relation to other industries. This might then be tantamount to a shadow premium on foreign exchange earnings from gold mining, represented in a differential subsidy to gold mining, which might replace earlier differentially high taxation tantamount to an effective “export tax.”

II. Recent Trends in Gold Output

These general considerations, as will be noted in Sections III and IV, are of considerable practical importance in interpreting both past experience and future prospects of gold mining in South Africa and other producing countries. In recent years, South Africa has been the only country in which gold production has played a significant proportionate role in the national economy, with the possible exception of Ghana. Table 1 lists the nine largest producers of gold in 1966, accounting for 95 per cent of estimated gold production outside the U.S.S.R. and other nonreporting countries in that year. In absolute terms, gold output in South Africa roughly doubled between 1955 and 1965 (from 1953, when the big postwar rise began, to 1966 it increased by 2.6 times), while the combined output of all other producers between 1955 and 1965 fell by 13 per cent. As a result, South Africa’s share of “world” output9 rose from roughly one half to three fourths in these 10 years. At the same time, gold output became a generally still larger part of both the national economy and the export economy in South Africa, and a still smaller one in the other producing countries. In South Africa, the share of gold production in relation to national merchandise exports, including all the gold production as a potential export,10 was about 50 per cent before World War I, reached a peak of 70 per cent in the 1930’s, dropped to just over 30 per cent in the mid-1950’s, and recovered to more than 40 per cent in the next decade.11 In 1967, the ratio dropped to 36 per cent. Of the other producers, Ghana has had a ratio fairly stable at about 9 per cent in the past decade. No other producing country showed a ratio above 5 per cent in 1965, and, with the exception of Rhodesia, the ratios were all tending to decline in this 10-year period (Table 1).

Table 1.

Selected Countries: Gold Production and Its Relation to Exports and to National Income

article image
Sources: Gold production: see Chart 1. Gold/export ratios: United Nations, Monthly Bulletin of Statistics; Board of Trade, Statistical Abstract for the British Commonwealth, No. 69 (London, 1947). National income: International Monetary Fund, International Financial Statistics.

F.o.b. values, as recorded in International Financial Statistics, plus gold production, i.e., not adjusting for gold consumption.

Gold valued at $35 a fine ounce.

Exports of Zambia and Malawi included in denominator.

Excluding U.S.S.R., Rumania, and China.

Gold output in South Africa has been subject throughout to notably different influences from those elsewhere, as a result partly of geological formation and industrial organization but also of market influences and government policy that can be linked to gold’s large proportionate role in the economy. It is useful therefore to examine world gold output in two categories, South Africa and what may for shorthand be termed the Rest, i.e., all other producers outside the postwar (World War II) nonreporting countries. Aggregate output of this residual group was estimated at $371 million in 1965 and $361 million in 1966; about one third of this was in Canada, and a similar portion in the United States, Australia, and Ghana combined.

The contrast between South Africa and the Rest is particularly striking in the response of gold output to relative price changes. Chart 1 plots gold output on this basis for the years 1910-66, together with the index of the “real” price of gold in terms of U.S. dollars referred to above, representing the U.S. official price in each year divided by the official index of U.S. wholesale prices.12 Two main features, which to some degree are interconnected, can be observed:

(1) With exceptions at the beginning of the period (1910-15) and at the end (after 1950), output in South Africa and the Rest move in the same direction; but in the major cycles the swings are notably stronger for the Rest. Thus, between 1915 and 1920 output fell by 10 per cent in South Africa and by 37 per cent in the Rest; in the 1920’s, with no very marked trend, output rose fairly steadily in South Africa and stagnated in the Rest; from 1929 to 1940, under the impetus first of the Depression and then of the rise in the price of gold in one currency after another, output rose by 35 per cent in South Africa13 and by 195 per cent in the Rest. Between 1940 and 1948, output fell by 18 per cent in South Africa and by 53 per cent in the Rest.14 Between 1948 and 1950, output rose by less than 1 per cent in South Africa and by 15 per cent in the Rest. Between 1950 and 1965, when the trends diverged, output rose by 162 per cent in South Africa and declined by 16 per cent in the Rest. From 1963, the rise in South African output decelerated. In 1967 there was a slight decline, but a modest upward movement was resumed in 1968 (Table 2).

Table 2.

South Africa: Trend of Gold Production, 1960–681

article image
Source: International Monetary Fund, International Financial Statistics.

These figures, based on those of the South African Reserve Bank, diverge slightly from those in Table 9, derived from the Transvaal and Orange Free State Chamber of Mines.

(2) Over the whole period, fluctuations in gold output for the Rest moved fairly closely in line with the real price index in dollars, whereas South Africa showed at best a weak relationship and, especially since the mid-1950’s, an inverse one (Charts 1A and 1B). It should be stressed that this index relates directly to the commodity price of gold in the United States, and only indirectly to that in other producing countries. The effect of expressing this real gold price in terms of movements in the local gold price and local wholesale prices is, however, surprisingly similar, as will be seen below.

Chart 1A.
Chart 1A.

Gold Production and a Real Dollar Price, 1910–66: South Africa

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

Chart 1B.
Chart 1B.

Gold Production and a Real Dollar Price, 1910–66: Rest of the World

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

Chart 1C.
Chart 1C.

Gold Production and a Real Dollar Price, 1910–66: United States

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

The results of a correlation and regression analysis of movements in gold output against movements in the gold price index in terms of U.S. dollars between 1910 and 1965 on the basis of figures for 1910, 1915, 1920, and each subsequent year to 1966 are shown in Tables 3A and 3B. A simple statistical analysis of this kind is, of course, far from conclusive in itself; it is nonetheless interesting that it bears out general relationships that might be expected deductively.

Table 3A.

Gold Output and Real U.S. Dollar Price: Coefficients of Determination, Correlation Coefficients, Regression Coefficients, and Elasticities of Gold Supply Equations, 1910, 1915, and 1920–661

article image

Not statistically significant from zero at the 5 per cent level.

Data are shown in Appendix II.

The coefficient of determination gives the proportion of variation in the dependent variable (gold output) explained by variation in the independent variable (the real gold price).

The regression coefficient gives the average absolute change in the dependent variable (gold output, shown here in millions of U.S. dollars) resulting from an increase of 1 percentage point in the real price of gold.

Percentage change in gold output divided by percentage change in real price index.

Table 3B.

Gold Output and Real U.S. Dollar Price: Coefficients of Determination, Correlation Coefficients, Regression Coefficients, and Elasticities of Gold Supply Equations, 1915 and 1920–66, with Lagged Response1

article image

Not statistically significant from zero at the 5 per cent level.

Gold price lagged one period.

The coefficient of determination gives the proportion of variation in the dependent variable (gold output) explained by variation in the independent variable (the real gold price).

The regression coefficient gives the average absolute change in the dependent variable (gold output, shown here in millions of U.S. dollars) resulting from an increase of 1 percentage point in the real price of gold.

Percentage change in gold output divided by percentage change in real price index.

For the period as a whole, the analysis shows no statistically significant correlation for total world output. This is the resultant of a negative correlation for South African output and a fairly strong positive correlation for output of all other producers combined (”the Rest”). The coefficient of determination for the Rest for the whole period was 0.47, indicating that 47 per cent of the variation in gold output was apparently “explained” by variations in this real gold price (Table 3A). If the period 1941-49 is omitted (to remove the special influence of the temporary, wartime element in the 1940-45 fall in output and the subsequent postwar recovery), the coefficient of determination for the Rest becomes 0.51. For the United States, the correlation was stronger still, with 63 per cent of gold output in this period apparently explained by the real price movement. For South African output, a marked positive correlation is found if the period is terminated in 1945. The coefficient of determination is then 0.46; for world output as a whole the coefficient is 0.72 for this period.

The tables also use the data in logarithmic form to show the elasticity of gold output in response to price changes. The elasticity for South Africa is negative in the whole period between 1910 and 1966, and modestly positive in the period to 1945; for the world total, elasticity is positive but not statistically significant in the whole period; in the period to 1945, it is 0.68. For the Rest, elasticity is 0.53 for the whole period, indicating that each percentage change in price was accompanied on average by a change in output in the same direction of just over half the percentage price change; in the period to 1945, with an elasticity of 0.94, the response was almost one for one. For the United States, the elasticity is 0.68, and is as high as 0.84 if one omits the years 1941–49. Table 3B shows the results on the basis of lagging the gold price one period, i.e., correlating the price in 1920 with output in 1921, etc., to allow for delayed response. These results generally show slightly higher correlations, but the broad picture is the same.

Tables 3A and 3B also show the results of simple linear regressions indicating the changes in absolute amounts of gold output associated with each percentage point of increase in the real price index, on the basis of 1926 = 100. For the Rest, this coefficient for the whole period between 1910 and 1966, without price lag, shows a change of $2.03 million in gold output for each percentage change in the real price, just over one fourth of this change being attributable to the United States.

Great care would obviously be needed in attempting to apply these results to predicting the effect on future gold output of a continued gradual fall in the real gold price as a result of the inflation of other prices, let alone any theoretical calculations of the impact of a hypothetical increase in the international gold price. Influences other than the real gold price appear to have become of generally increased importance in the most recent period, in part as a result of the relative decline in this price itself, and these extraneous influences may in some degree have cushioned the sensitivity of future movements in gold output to relative price changes. These influences, operating primarily through subsidies and other forms of official assistance, as well as through the increased importance of gold mined as a by-product, are discussed extensively in the ensuing sections. At this stage, one general point may be noted from this statistical record.

The response of gold output in the Rest to changes in the real gold price as here defined in terms of U.S. dollar prices was almost the same in the period of the sharply increased price in the early 1930’s as that indicated by the average elasticity of 0.53 over the period as a whole (Table 3A). The effect of lagging the price by 1 year is much the same as for the whole period—slightly higher responses within the same general range. But the response in the Rest to this increase of 114 per cent in the real price in the 5 years 1929-34 appears to have continued for the ensuing 6 years when the real price showed no decided trend. In the period 1929-40 as a whole, with a rise in the real price of a net 104 per cent, output of the Rest rose by almost 3 times. South African output in the same period rose by about one third (Chart 1 and Appendix II). Thus, even allowing for delayed responses, output in the Rest was far more sensitive to movements in the real gold price.

III. The Rest: Subsidies and the Influence of Relative Profitability

This gold price index in terms of constant U.S. dollars is not, of course, a full measure of relative profitability, even aside from the “nonfinancial” influences of availability and technology. Additional influences include, on the revenue side, (1) any divergent movements in the gold price in local currency, which will normally correspond to movements in exchange rates against the dollar; (2) indirect increases in revenue through special government assistance and subsidies; and, on the side of costs, (3) divergences in wholesale price movements from movements in the United States—these will normally offset the exchange rate movements, and the combined influence will be limited to divergences in the wholesale price of local goods relative to U.S. goods; and (4) internal divergences in gold-mining factor and material costs from wholesale prices. The extent to which these four influences have been mutually offsetting is reflected in Tables 3C and 3D. These show that the correlations of gold output in South Africa and Canada against the real gold price in terms of local currency and local wholesale prices are broadly similar to the correlations shown for these countries in Tables 3A and 3B in terms of U.S. dollar prices; a similar close relationship would be shown for Australia. Another manifestation of this offsetting character of the local currency influences is that after excluding the United States and Canada from the Rest (last column of Tables 3A and 3B), the coefficients in terms of U.S. dollars are not only high but actually higher than with the United States and Canada included. This reflects the statistically insignificant correlation for Canada when the period after 1945 is included.

Table 3C.

Gold Output and the Real Local Price: Coefficients of Determination, Correlation Coefficients, Regression Coefficients, and Elasticities of Gold Supply Equations, 1910, 1915, and 1920–661

article image

Not statistically significant from zero at the 5 per cent level.

Based on the local gold price, as shown in Chart 2, divided by index of wholesale prices. Data are shown in Appendix II.

Table 3D.

Gold Output and the Real Local Price: Coefficients of Determination, Correlation Coefficients, Regression Coefficients, and Elasticities of Gold Supply Equations, 1915 and 1920–66, with Lagged Response1

article image

Not statistically significant from zero at the 5 per cent level.

Gold price lagged one period.

In the United States the first three “local” factors mentioned above have been absent,15 and the scope for the fourth is limited by the very small weight of gold mining in the domestic economy. So it is perhaps not remarkable that, as noted on page 422, the correlation between U.S. gold output and this indicator of relative profitability is particularly strong when the effects of World War II are omitted. In all the other main producing countries, the gold price in local currency has diverged from the dollar price during this period (Chart 2). Thus, the price of gold in terms of Canadian dollars rose with the devaluation in 1939, fell in line with the Canadian revaluation in 1946, rose with the Canadian devaluation in 1949, fell again with the appreciation of the Canadian dollar in the 1950’s, and rose again after the devaluation in 1962—though in each case the movements were relatively modest. Prices for producers in Australia, Rhodesia, and Ghana, as well as South Africa, broadly followed the fluctuations in exchange value of the pound sterling before 1967, involving a temporary rise in gold prices during 1919-25, with further increases in the early 1930’s ahead of the rise in the dollar price in 1933-34;16 sustained increases not “overtaken” by the dollar price took place only with the sterling devaluations in 1939 and 1949. These movements are shown in Chart 2, which also shows movements in local market prices and the effect of sales in premium gold markets where these have been significant elements in producers’ revenue.

Chart 2.
Chart 2.

Gold Prices, in Selected Currencies, Per Fine Ounce of Gold, 1910–67

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

Alongside these divergences in local gold prices from the U.S. dollar price, general price levels as reflected in wholesale prices have diverged from those in the United States, and assistance programs for gold mining have become widespread.17 Clearly, however, these different sets of “local” influences have tended to offset each other. Devaluations against the dollar have been at once a result and a further cause of relative increases in wholesale prices against the United States. Relative increases in wholesale prices can also be offset by higher subsidies; and subsidy payments have sometimes moved inversely with the local gold price as influenced by exchange rate movements.

Thus, subsidies introduced by Australia in 1930 were reduced in 1931 and ended in 1932, under the influence of the stimulus available from devaluation. The further rise in the local gold price following the devaluation in line with sterling in 1939 was accompanied by a partly offsetting Gold Tax; this was repealed at the end of World War II. Renewed subsidies introduced by Australia in 1948–49, and at the same time by Rhodesia, were discontinued after the devaluations in 1949. A new Australian subsidy was introduced in 1954 after a particularly sharp increase in the price level; assistance rates were increased in 1961 and again in 1965. Rhodesia reintroduced a subsidy in 1963.

In Canada, subsidies were introduced in 1947–48 following the additional pressure on gold producers from the 1946 currency appreciation. The subsidies were reduced following the boost from the 1949 depreciation of the Canadian dollar. This was shortly followed by appreciation of the Canadian dollar on abandonment of the fixed exchange rate in 1950. Under this regime, the price paid by the Canadian Mint was adjusted weekly on the basis of the week’s average market exchange rate of the U.S. dollar in Canadian funds, which was applied to the U.S. official gold price of $35 an ounce. A number of modifications were made in the assistance arrangements, which on balance had the effect of significantly increasing the rate of assistance per ounce of gold (Table 4, columns 4 and 5), though not to a sufficient extent to offset the reduction in the Canadian gold price (Table 4, column 7). In 1955, assistance scales were revised in order to channel a greater portion to the high-cost producers, and to reduce the total. Thus, in 1957 these payments totaled Can$9.7 million on gold production that at the year’s average Canadian Mint price of Can$33.55 was worth Can$148.8 million. This was some Can$22 million below what these receipts would have been at the gold price of Can$38.50 which prevailed during 1939-46 and in 1949-50. In this sense, the assistance payments could be regarded as only a partial offset to the exchange appreciation. In 1958, in recognition of this factor as well as of higher costs, the amounts payable were increased by 25 per cent. In the period 1948-65 as a whole, in which the annual average of the Canadian exchange rate was about 1.4 per cent below U.S. parity and the average gold price in Canadian dollars was $35.50, total assistance payments totaled Can$216 million, increasing gross income from sales by 8.0 per cent.

Table 4.

Canada: Gold Production, Assistance Payments, and a “Real” Canadian Gold Price, 1948–66

article image
Sources: Canada, Department of Energy, Mines, and Resources, Report on the Administration of the Emergency Gold Mining Assistance Act, 1967; International Monetary Fund, International Financial Statistics (for general index of wholesale prices).

Column 5 plus column 6.

Column 7 divided by index of Canadian wholesale prices, as percentage of figure for 1955.

Total payable.

Under Canada’s current legislation, the position is as follows. Assistance payments are not available to mines producing gold as a by-product of base metal, which accounted for 17 per cent of total Canadian gold production in 1965, against 10 per cent in 1948. Eligible recipients are gold mines with costs of production of Can$26.50 an ounce or more: the rate of assistance is on a sliding scale, based on production costs, of two thirds of the amount by which average production costs exceed this sum, subject to a maximum rate which is reached when average cost of production rises to Can$45 an ounce of gold produced.18 In 1965, three fifths of output eligible for assistance, and more than one third of total Canadian gold output, was produced at costs per ounce in excess of Can$35 (US$32.38). In 1966 the share of this high-cost output rose to three fourths of output eligible for assistance and to two fifths of total output.

Between the early 1950’s and 1965, there was no decline in “relative profitability” on a comprehensive Canadian reckoning for marginal producers, i.e., on the basis of the (increased) Canadian gold price plus assistance payments per ounce of eligible production, this sum being deflated by the index of Canadian wholesale prices (Table 4, column 8). This helps to explain the relative resilience of Canadian gold output, at least until the early 1960’s, in face of the fall in relative profitability (Chart 1D). By 1967, however, output had fallen by more than one third below its level in 1960.

Chart 1D.
Chart 1D.

Gold Production and a Real Dollar Price, 1910–66: Canada

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

Comparisons between gold subsidies in different countries are complicated by the different bases on which these schemes are drawn, and further by what may be important differences in conditions external to the schemes, in taxation arrangements, and the like. Within these limitations, an attempt at a comparison of effective rates of assistance in five countries is made in Table 5A; Tables 5B and 5C summarize certain other characteristics of these schemes.

Table 5A.

Gold Subsidy Schemes. A Comparison

(Data relate to calendar year 1965, unless otherwise stated)

article image
Sources: Canada, Department of Energy, Mines, and Resources, Report on the Administration of the Emergency Gold Mining Assistance Act, 1967. Australia, Gold Mining Industry Assistance Act, Twelfth Annual Statement; Commonwealth Bureau of Census and Statistics, Statistical Bulletin: Minerals and Mineral Products, No. 2, 1965, and Statistical Bulletin: The Gold Mining Industry, Australia, December 1966 and Year 1966; and data supplied by Australian Government. Southern Rhodesia, Gold Mining (Financial Assistance) Act, 1963. See also textual references.

Salient characteristics of Canadian and Australian schemes are described in the text, and those of Rhodesian and Philippine schemes in footnotes 9 and 11 to this table; see page 450 and footnotes 4–6 to this table on South African scheme.

Assistance payments may relate partly to production in earlier years; these figures, based on output in the same year, should therefore be taken only as indications of broad magnitudes.

Gold valued at $35 an ounce.

Estimate based on aggregate figures for 1963–66 and 1964–66, mentioned on page 450 of the text. The annual rate in mid-1967 was $5.6 million.


Relates to loan payments to cover losses up to 10 per cent of revenue; excludes effect of government reimbursement of pumping costs of certain marginal mines.

Year ended June 30, 1966; includes payments to producers in Papua and New Guinea, but excludes development allowances equivalent to US$75,000 in this period.

On scheme for large mines; small mines are eligible for a flat-rate subsidy of US$6.72 an ounce after June 30, 1965.

Before allowing for market premium over parity price equivalent to $35 an ounce. Until November 1965, Rhodesian gold production was usually sold on the London market where the average gold price in 1965 was about $35.13. Since the Unilateral Declaration of Independence, Rhodesian gold production has been sold wholly to the Reserve Bank authorities in Salisbury. Scheme allows payments in respect of “gold won” of discretionary amounts not exceeding either (1) amount of working loss in quarterly period (account may be taken of profits earned in earlier periods), or (2) £R 3 an ounce produced in the period.

As from September 1966.

The scheme combines a flat-rate subsidy equivalent to $12.82–15.38 with payment of 65–70 per cent of the positive difference between the cost of production and the official price, to the maximum indicated. The lower scales apply to mines producing more than 100,000 ounces a year, of which there was one in 1965. Mines producing gold as a by-product receive only a flat-rate subsidy (of $12.82 an ounce).

Table 5B.

Gold Subsidy Schemes: Selected Aspects of Coverage

article image
Sources: See Table 5A.

Sales of gold bullion above official buying price of Mint or central bank.

Output sold in the form of ore or concentrate is eligible for assistance when sales are either to a foreign smelter or to a domestic smelter whose operator certifies that an equivalent number of ounces of refined gold have duly been sold to Canadian Mint.

Under an amendment of the legislation proposed by the Australian Government in August 1968, the required deduction from subsidy payments of any premiums received would be reduced from 100 per cent to 75 per cent, i.e., subsidized producers would effectively retain 25 per cent of the premium.

See footnote 11 to Table 5A.

Assistance extended under 1966 Act until such period as the present official world price of gold is increased by international treaty or convention to at least US$70 an ounce.

Table 5C.

Gold Subsidy Schemes: Total Payments, 1948–66

(In millions of U.S. dollars)

article image
Sources: See Table 5A.

Including payments to producers in territories of Papua and New Guinea; excluding development allowances.

Total payable.

Australia’s assistance arrangements for large producers (small producers receive a flat-rate subsidy of $A 6, equivalent to US$6.72, an ounce) are on the same general principle as the Canadian arrangements. Under the Gold-Mining Industry Assistance Act as amended in 1965, these producers receive from the Government three fourths of the difference between their production costs and $A 27 (US$30.24) an ounce, subject to a maximum payment, which was raised in 1965 to $A 8 (US$8.96) an ounce. This increase and accompanying relaxations noted below led to a sharp rise in the total of this assistance, from $A 1.7 million (US$1.9 million) in the year ended June 30, 1965 to $A 2.4 million (US$2.7 million) in 1965-66.19 Payments in each year relate in substantial part to production in earlier years (about three fifths of payments in 1965-66 were in this category), but as annual changes in total output have been relatively modest in the past few years, the figures for average assistance per ounce of gold produced, shown in column 5 of Table 5A without adjustment for this lag, give an indication of the broad magnitudes. This average subsidy rose by about 50 per cent between 1964-65 and 1965-66, to the equivalent of US$2.90 an ounce. This remained distinctly below the corresponding figure in Canada for the calendar year 1965; but the portion of output benefiting from this assistance, while increasing to 54 per cent,20 remained below that in Canada, so that payments per ounce of assisted output were more closely comparable (Table 5A, columns 4-6). Mines become eligible at a lower level of cost in Canada than in Australia (column 7). But for very high-cost mines, where output is most affected by assistance payments, at least in the short term, the payments are roughly the same in both countries (columns 8 and 9).

The bulk of Australian gold production is now mined from low-grade ore. The two largest producing companies, Gold Mines of Kalgoorlie (Australia) Limited and Lake View and Star Limited—accounting in 1965 for about 35 per cent of the country’s total gold output—showed grades of about 4 pennyweights (dwt), i.e., one fifth of one ounce of gold, per ton of ore treated,21 comparable to the low grades of the old Rand in South Africa (see page 457). Among the eight major Australian producing companies, the highest grade was 10.5 dwt., at Central Norseman Gold Corporation N.L. This company accounted for 11 per cent of Australian gold output in 1965, but showed exceptionally small proved reserves: at 634,000 tons of ore on June 30, 1966, these were less than three times the amount of ore treated in the previous year.22

Australia’s subsidy arrangements are more liberal than Canada’s in relation to gold sales at a premium over the official parity on world markets (Table 5B). These sales are permitted freely under special arrangments consistent with the sale of all newly mined gold to the Reserve Bank in the first instance.23 From 1962 through 1964, some 98 per cent of Australian exports of gold bullion, including re-exports, went to Hong Kong (Table 6); shipments to the United Kingdom, which were substantial in 1961, have not been recorded on any significant scale since then, in either Australian or U.K. statistics. Even so, the average premium in 1964 was very small indeed, the equivalent of eight U.S. cents an ounce; the aggregate addition to receipts at the official price was only 2 per cent as large as receipts from subsidy payments.24

Table 6.

Australia: Gold Production and Disposal, 1960–66

(In thousands of ounces)

article image
Sources: Commonwealth of Australia, Year Books. Reserve Bank of Australia, Statistical Bulletin. The Australian Mineral Industry, Quarterly Review. A. J. Gourlay, “Gold,” The Australian Mineral Industry 1965 Review, 1966, pp. 133–41.

Slight differences from data in International Financial Statistics owing to conversion and rounding, as well as revisions in Australian statistics.

Residual figure of line 2, and changes in holdings of official and banking institutions both on their own behalf and on behalf of nonresidents, together with gold content of change in stocks of minerals awaiting refining, as shown in Commonwealth of Australia Year Books.

Line 1 — (2 + 3 + 4); also equivalent to line 7 — 6.

More significant additional relief for the Australian gold-mining industry as a whole is total exemption from income tax, including tax on dividends in the hands of shareholders. The 1965 amendments to the Gold-Mining Industry Assistance Act removed the earlier provision under which the amount of subsidy payable to large producers was reduced by any amount by which profit after payment of subsidy would have exceeded 10 per cent of the capital used in the production and sale of gold. The amended Act also provided for inclusion in large producers’ production costs for subsidy purposes of approved diamond drilling carried out by the producer elsewhere than on his mining property.25 At the same time, these extensions and increased assistance rates were accompanied by a lapsing of the Gold Mines Development Assistance Act of 1962, under which increases in approved development expenditure were reimbursable in full, subject to an upper limit related to current ore production. The increase in proved ore reserves in respect of which the allowance was payable was also subject to a “reasonable limit,” generally an additional 2 years’ supply beyond the number of years’ supply at June 30, 1962.26 Relief under this Act excluded recourse to the subsidy. In the fiscal year ended June 30, 1965 payments of development allowances and advances totaled $A 141,000; payments in 1965-66, expected to be the last under the Act, totaled $A 63,000. Arrangements of this kind are apt to involve administrative difficulties, involving arbitrary judgments or inadequate policing, and the Australian authorities preferred to absorb some of the features of the scheme in the 1965 extension of the Gold-Mining Industry Assistance Act.

Among the most extensive assistance arrangements are those in the Philippines, where in addition to a flat-rate bonus of P = 50 or P = 60 (US$12.82–15.38) paid on all gold output sold to the authorities at the official price, including by-product gold, additional payments are made on 65–70 per cent of the excess of the cost of production over the official price. Maximum aggregate remuneration (i.e., official price plus assistance) under these schemes runs up to the equivalent of more than US$66 an ounce. Subsidy arrangements were instituted by the Philippines in 1954; the amendment of the legislation in 1966, which increased the level of assistance, extended the assistance until such period as the present official world price of gold of US$35 an ounce is increased by international treaty or convention to an amount equivalent to at least US$70 an ounce.27 In some other countries, special assistance to gold mining has itself been in the form of access to (or compensation determined by) the free exchange market or free gold market. Thus, in Colombia the central bank buys domestic gold production at prices reflecting the free market rate of exchange, under the conditions prevailing early in 1967.28

In many instances, the subsidies have been aimed not so much at gold production as at employment. In Canada, and to some extent also in Australia, the aims have been social and “strategic,” to maintain the existence of small frontier communities; it has been considered that the collapse of such communities not only would lead to social hardship but also might prejudice development of other resources, including other minerals, which might follow on from the presence of the gold-mining communities. This was an important consideration in Canada when the subsidy program was initiated late in the 1940’s. The renewal of the program at the end of 1963 effectively confined assistance to existing gold-mining communities, a clear indication of the social emphasis.29 The purpose of the restriction was “to prevent the establishment of new gold-mining communities which would be dependent upon government assistance from the outset and which would add to existing social problems.” 30

Concern with both the level and the conditions of employment has been a major factor in government support for gold mining in Ghana, though recently the focus has shifted in part to securing a recovery in output. In 1956, before independence, the Government of the Gold Coast agreed to make grants to four companies mining low-grade ore to compensate for a substantial wage increase and avert mine closures. In 1961, after the proposed cessation of government assistance had brought the imminent prospect of closures, the Ghana Government acquired the capital of five of the seven British-owned gold-mining companies in the country, then producing some 60 per cent of the country’s total gold output. Subsequently, the Government supported these mines directly, by meeting the losses of the State Gold Mines Corporation.

Despite this support, the gold output of this nationalized sector in 1965 was half the comparable level in 1960. In the same period, gold output in the higher grade and still privately owned Ashanti Goldfields Corporation rose by about a third, to two thirds of the country’s total, as a result of a substantial investment program. On September 30, 1966, the company calculated ore reserves of 3.67 million tons at the high grade of 19.81 dwt., or nearly one ounce of gold, per ton. Further expansion of output, however, appeared to be dependent on additional investment, which was inhibited by heavy taxation. In 1965, more than 72 per cent of the company’s pretax profit was paid in Ghana taxes, and the amounts available for dividends also became affected adversely by the new U.K. corporation tax.31 The devaluation of the new cedi by 30 per cent in July 1967 should provide a general boost to gold output.32 Output in the nationalized sector may also be aided by improved availability of imported machinery and materials, as well as by discovery of new deposits in the Konongo area, estimated to add to gold output by $1.4 million a year for 10 years.33 Earlier in 1967, Ghana began to refine gold locally, in a refinery built with assistance from the U.S.S.R.; the resulting addition to the country’s gold revenues is unlikely to be more than ¼ of 1 per cent.

The United States has persistently opposed subsidy or assistance payments on the grounds that these could lead to uncertainty about the official gold price and thereby undermine confidence in the dollar. Recently, the United States has mounted a modest official program of resource development and research designed in effect to reduce producer costs, and aimed specifically at increased output. In 1961, gold was added to the list of minerals eligible for the exploration loan program of the Geological Survey, and in the next 5 years 37 contracts relating to gold were made, totaling $2.2 million. Studies by the Geological Survey identified a new type of gold deposit, which was developed by private exploration into the Carlin mine in Nevada. The opening of this mine in 1965 was responsible for an upturn in U.S. gold output, which was carried further in 1966 (Table 1).34 In 1966, a more extensive Heavy Metals program was launched jointly by the Geological Survey and the Bureau of Mines. Early results were claimed, including identification of a gold-bearing zone in the Cortez area of Nevada, similar to that of the Carlin mine, and suggesting the existence of other deposits.

The working basis of the official U.S. exploration and research effort is that discovery of new sources and development of new or improved mining and metallurgical techniques offer better promise of increased output than the reopening of known mines; though it is hoped that the lives of existing operations may be extended by improved technology. Attention is also being paid to marine exploration, which requires improved drill rigs; two different types of rigs are to be tested in the Bering Sea off Alaska.35

The spread of assistance payments outside South Africa has slowed the pace of mine closures and thereby has averted a significant fall in output from a growing number of marginal and submarginal mines. The proportion of gold output from mines with reported average production costs above the equivalent of US$35 an ounce was perhaps 30 per cent in Canada in 1965, and some 45 per cent in Australia in the year ended June 30,1966;36 in the same category, it may be guessed on less complete information, the share was perhaps about one fourth of the gold output in Ghana, the bulk of output in the Philippines, and a substantial portion of output in Rhodesia. In all, among non-South African gold producers, gold production equivalent to at least $80 million and possibly to $100 million or a little more, representing 22–27 per cent of total output of these countries and 5½–7 per cent of output including South Africa, may have been made economic by assistance payments; without these payments, the bulk if not the whole of this output would be expected to cease.

In this way, it can be said broadly, in terms of Chart 1B, that this assistance, along with local currency devaluations, has raised the output curve above the “real” dollar price curve. Even so, notwithstanding these considerable and increasing “local” influences in the field of price movements, exchange movements, and subsidy payments, no continuing divergence in trend between the output curve and this indicator of external relative profitability has emerged.

At least in some degree, then, the relatively close correspondence between movements in gold output outside South Africa and in the “real” dollar price of gold has been due to a self-correcting tendency for these local financial influences to cancel out. Looking ahead, there is no obvious reason to expect these offsetting tendencies to disappear, at least while relative profitability continues to decline (though one future counterpart to the lift given to the output curve by the subsidy arrangements at a time of declining relative profitability could be to deaden part of the stimulatory influence of any future improvement in relative gold prices; this point is touched on below). Four specific influences may be noted.

(1) In none of these countries does gold production play a sufficient role in external earnings for exchange rates to be influenced significantly by a further decline in gold output or by a potential recovery in response to a higher local gold price. Devaluations induced by other influences should, however, in themselves make gold production more profitable in the devaluing country, subject to counterinfluences from increased local prices and factor costs and to any offsetting reduction in subsidies and other assistance payments. Equally, gold production would tend to be affected adversely by any upward revaluations in these countries.

(2) Subsidies and other indirect additions to gold-mining revenue are in general likely to continue and may in some instances increase, but the primary importance attached in some important cases to maintaining existing gold-mining communities, to the exclusion of new ventures, may involve a gradually lessened impact on gold output. The “insurance premium” element in gold subsidies in relation to a possible major increase in the international gold price might also appear more questionable insofar as the prospect of such an increase recedes. Equally, to the extent that official assistance is motivated by a desire to maintain output and to safeguard employment for existing resources—whether on economic, social, or political grounds—rather than to encourage the absorption of new resources in this sphere, the impact of these subsidy payments could be expected to diminish over time.37 The same factor would tend to mitigate the stimulatory influence both of local devaluations and, theoretically, of any increase in the world gold price.

(3) Since the proportionate role of gold mining in these economies is tending to decline still further, factor costs and material costs are unlikely to diverge significantly from the general internal level, with resultant pressure on costs. Because of differences both in geological structure and in policy, gold producers in these countries do not always follow the South African practice of mining to the average grade of the ore (discussed on page 445 below). Avoidance of the effect of the cost squeeze on output is therefore likely to require either improvements in technology or discovery of new payable ore, with little possibility of increasing yields by exploiting richer seams in mines already developed.

(4) Minor modifying influences on these trends that are external to gold mining may stem from what may be summarized as by-product influences. Three categories may be distinguished.

(a) Perhaps 6–7 per cent of the world’s total gold output, and 20–25 per cent of gold output outside South Africa,38 is produced as a direct by-product of mining operations in which the recovery of other metals (uranium in South Africa, base metals and silver elsewhere) yields greater revenue than the recovery of gold. In these instances, gold output is determined primarily by the profitability of mining the main product or products. Demand for base metals and for silver has been strong in recent years, and the amount of gold extracted as a by-product of base metals has shown a stronger trend than “direct” gold production among the Rest. In the United States, the share of by-product gold, predominantly from copper ores, in the total was 40 per cent in 1965, having shown a slightly rising trend over the previous decade. In Canada, the upward trend was more marked, to 17 per cent in 1965. The share of by-product gold has also been showing a generally rising trend in Australia, reaching 21 per cent in 1966. In Japan, by far the larger part of gold production is mined as a by-product. Because this category of gold production has been little affected by the adverse relative price trend for gold, there has been a natural tendency for its proportionate weight in gold production to increase. This by-product gold may set an ultimate floor to the decline in gold production in response to declining relative profitability of gold itself, involving in the meantime some deceleration in the proportionate decline in gold production as the relative share of gold as a by-product increases.

(b) Gold may also be considered as an indirect by-product where its costs of production benefit significantly from overhead or infrastructure expenses undertaken primarily on account of mining operations for other metals. Hitherto, few if any such cases have occurred: the high value of gold in relation to most other metals has produced rather the opposite conjuncture, and, as has been noted, prospecting for gold has been seen by governments as possibly leading on to the discovery and development of other mineral resources. A reversal of this pattern could conceivably be seen in Australia, where large tracts of previously unexploited territory are being opened up for new mineral production. Aside from the possibilities of new gold discoveries being made economically viable in this way, development of nickel resources in the neighborhood of the gold-producing district of Kalgoorlie could presumably relieve certain financial strains arising from the maintenance of communities that are now dependent solely on gold mining. It should, however, be stressed that, as of 1967, no firm indications of any significant effects in this direction have yet appeared.

(c) Where other metals are extracted as a by-product of gold, the effect is to increase the net revenue of these gold mines accordingly. Silver is always associated with gold-bearing ore in some measure. But the most important by-product of gold in recent years has been uranium in South Africa, where uranium is produced entirely by reworking gold-bearing ore after the gold has been extracted. These operations provided substantial additional revenue in the 1950’s;39 in a few mines classed as “primary uranium producers” and considered under category (a) above, revenue from uranium still exceeds revenue from gold. In the first half of the 1960’s, world demand for uranium fell, and the significance of these operations in South African gold mining has declined. In 1966, South Africa’s uranium revenues turned upward after 5 successive years of decline, and the improved prospects are expected to continue. However, an advance in recovery techniques, which in some instances makes recovery profitable from ores of very low gold content as well as from old tailings as in the past, may weaken the link between uranium production and gold production in South Africa.

Recent trends and apparent prospects in gold output among the Rest may be summarized as follows. In Canada, output has fallen in each year since 1960, despite the buoyancy of by-product gold. At least one new mine may have considerable potential, and, since large areas have been far from completely surveyed for minerals, new discoveries remain a real possibility. The present trend of output, however, is clearly downward. In Australia, where gold output reached its peak in 1903, the long decline is expected to continue. The main present uncertainties on gold output outside South Africa relate to the United States, where output is normally buoyed by the particularly high ratio of by-product gold, and to Ghana, where the decline in output in 1965 and 1966 partly reflected inflationary strains and disturbed economic conditions. Some short-term revival in Ghanaian gold output may be possible, and the tight foreign exchange position may encourage it. This could also be an influence encouraging special efforts to increase gold output in Rhodesia. And if the new techniques of exploration adopted in the United States were to achieve major successes, they would no doubt be emulated elsewhere. In sum, however, in the absence of major new discoveries of gold deposits that are either exceptionally rich or exceptionally economical to mine, and on the basis of the present international gold price, the recent downward drift in gold output outside South Africa may be expected to continue. Gold output in these countries has responded in a fairly direct way to changes in the international price of gold in relative terms, and within certain limits the same broad response may be expected in the future.

IV. South Africa: The Influence of Comparative Costs

Influences on gold output in South Africa are more complex, essentially reflecting the far larger weight of gold mining in the economy. In the broadest terms, as noted above, the course of output hitherto can be explained partly in terms of relative profitability, taking into account the effects on major development expenditures that influence output for a decade and more ahead. But this broad influence can be modified or entirely offset by a complex of factors that have affected output in the short term or medium term in the past, and that could conceivably become more important in the future.

General Influences on South African gold output

Geological structure

The gold-bearing reefs of the Witwatersrand system are far more extensive than any other gold-bearing deposit discovered hitherto.40 The system occurs in a large basin and reefs have been traced in a great arc, roughly semicircular and about 300 miles in length. These were discovered and exploited in the mid-1880’s in the Johannesburg area, in what is now known as the Central Rand. Extensions followed to surrounding areas in the East and West Rand, with a second burst of new discoveries before World War I. The stimulus of increased profitability in the 1930’s encouraged further intensive exploration, along the “West Wits Line,” which opened up the major fields of the Far West Rand and the Klerksdorp area; these areas became significant producers after the end of World War II. Exploration in the 1930’s had also led to the discovery of gold in the Orange Free State, but development of this area was entirely a postwar phenomenon, further stimulated by the devaluation in 1949. Production in significant quantity began early in the 1950’s. This period has also seen the opening up of the Evander field, located on the eastern extremity of the arc and regarded as geologically separate from the Rand rather than an extension of it. Two points in this development of the South African gold fields may be noted. First, it spread outward fairly consistently on either side of the original center (see map). Second, the existence of the “arc” has been apparent only in retrospect; and while knowledge of the mineral structure grows progressively more thorough, there can of course be no definitive assurance that the present geological “map” will be the final one.

The South African gold reefs are often severely faulted and rarely exceed 6 feet in thickness. They occur at depths usually below 4,000 feet and down to 15,000 feet or more below the surface. The maximum depth of mining has steadily increased. Before 1945 the limit was about 7,500 feet; in the 1950’s, mining at 10,000 feet became practicable, and Western Deep Levels was planned to touch 13,000 feet.41 The maximum depth worked in the mid-1960’s was about 11,000 feet. In the older areas the average gold content of the ore was notably lower than the main gold deposits that previously had been worked in other countries: the Rand became payable on a large scale only with new techniques of treatment (the cyanide process) and of extraction. At the same time, this grade of ore varies considerably, even on the same reef horizon. The practice soon developed, and was given the force of law, of “mining to the average grade.”


south africa: location and age structure of gold mines

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

The precise meaning of this practice is not always entirely clear. Mines operating under state leases, which include all the mines in the Orange Free State, are subject to the general provision that operations are carried out to the satisfaction of the Minister, and that “the average value of the ore milled is to be in reasonable accord with the average value of the ore reserve.”42 This ore reserve comprises blocks of ore in the developed portion of the mine, of varying grade or gold content; but calculation of both the tonnage of the ore reserve and the average grade of that reserve takes in only the payable ore, and this only in the developed portion of the mine (which may well be the most promising portion). The normal accounting rule seems to be to include in the calculation ore down to the lowest grade of ore that can be profitably treated on current calculations of marginal costs. As a result, when increased working costs bring a corresponding rise in the pay limit, as previously marginal grades of ore become submarginal, the average value of the (payable) ore reserve also rises and the average grade of ore worked will rise accordingly. The counterpart to this is a depletion in the tonnage of ore and gold in the calculated reserve.43

The feasibility of extracting the richest ore in a mine ahead of the lower grades is limited in part by the physical formation. Because of narrow stoping widths (i.e., working areas at the face) it may not be possible or economical to return to former stoping areas, particularly using the recent technique of “concentrated mining.” This involves more rapid advance of the stope face, with support by removable props, and complete mining of all ore in each working area, including some ore that would have been left unmined in circumstances where selective mining was practicable even within the policy of mining to the average grade.44 Concentrated mining should therefore increase the total amount of ore extracted, in part by reducing the pay limit through savings in working costs. The practice does, however, speed up the rate of extraction and thereby shorten the life of the mines. And since concentrated mining is less selective at each particular working area, it reduces the scope for day-to-day adjustments in the average grade of ore extracted from the mine as a whole.45


These geological characteristics—extensiveness of ore bodies, thinness and depth of seams, relatively low average grade, and wide variations round this average—have contributed to two distinctive economic and organizational characteristics of the South African industry.

(1) Mining is a large-scale activity, requiring massive capital outlays before development and heavy outlays even after development. From the outset therefore the producers have been joint-stock companies, which tend to take a relatively long view on maximization of profits and may also be guided by corporate, social, and political influences, of the kind discussed below.46

(2) Because of the special consistency and “adjustability” of the gold-bearing resource, response to changes in profitability is in the first instance through an increase or decrease in the amount of payable ore. Thus, a reduction in costs or an increase in revenue will involve in the first instance a reduction in the pay limit.47 Under the practice of mining to the average grade, this will involve a fall in yield that may initially outweigh the increase in tonnage, and thereby result in an actual fall in gold output (as occurred in 1933 and 1934). Equally, the length of time needed for development may involve a particularly long lag in the response of output to increased profitability.

The usual volatility of gold output, particularly in response to changes in relative profitability, is reduced in South Africa by both geological and economic influences. But, in addition, these influences have from the outset been reinforced by government policy, as well as by deliberate policy of the industry itself. The industry has traditionally put a special premium on short-term stability and long-term growth. A number of considerations are involved: consciousness of the special role of gold mining in the South African economy; a desire to provide maximum assurance of future employment; and the particular rigidity of the supply of native labor. This is drawn almost entirely from tribal areas far afield (and for the most part from outside the territory of South Africa) on short-term contracts, of 4 to 18 months, and distributed to the mines by two central recruiting organizations.48 Another element is the desire among mine managements and investors to avoid fluctuations (and specifically reductions) in dividends. But this preference for a degree of dividend stabilization would not in itself necessarily militate against selective mining, since variations in returns from gold mining could be ironed out by varying the pay-out ratio and investing retained profits in financial securities.49

Government policy and taxation

Government policy has had two broad influences on South African gold mining. First, it has in effect skimmed off excess profits of the industry, partly through a special and differentially high tax scale but also through the terms of leases, through freight charges on the state railways, and through import duties on important mining stores. The amounts skimmed off in these ways have varied in an important degree with the course of total profits from gold mining.50 In addition, and sometimes in association with these general levies, government policy has influenced the behavior of the industry in three main directions.

First, both tax and lease liabilities have been so framed as to encourage maximum output in the long run, with inducements at any one time to mine ore of a lower grade than would otherwise be profitable. The basic formula has been to increase the effective rate of tax or lease payments with the rate of profitability, i.e., ratio of profits to gross revenue, and thereby provide a financial deterrent to selective mining.51 The tax is payable only if the profits/sales ratio exceeds 6 per cent. At a profit ratio of 40 per cent the rate of this “formula” tax exceeds 50 per cent under the basic rates (i.e., before allowing for the surcharges noted in footnote 50, below); the maximum rate of the formula tax exceeds 55 per cent. Mines established after August 1966 benefit from a lower scale. Companies other than gold or diamond mines in South Africa are subject to a flat rate on profits, which after the 1967 budget was 40 per cent, including loan levy; 2 years earlier the rate had been 30 per cent. Thus, high-profit gold mines in South Africa are taxed considerably more severely than nonmining companies, while low-profit mines secure differentially favorable treatment. Altogether, the element of gearing in relation to profitability in gold-mining taxation helps to make receipts from this source notably more volatile than other government revenue (Chart 3).

Chart 3.
Chart 3.

South Africa: Gold-Mining Taxation—Two Comparative Measures

In per cent

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

1 Column 4 of Table 7.2 Column 3 of Table 8.

Second, with a view to future gold production, the gold mines have been granted favorable capital allowances. Capital expenditures incurred to bring a mine to production have been allowable in full as a charge to taxable income, so that no tax is payable until after such expenditures have been written off.52 A number of additional reliefs have been granted in recent years.53

Third, and increasingly in recent years, in an effort to safeguard both present and future production, certain special payments have been made available to less profitable mines. Relief for these mines has come partly under the first two counts—thus, in the mid-1960’s, only about half the 50 or so mines paid tax. But certain special assistance payments have also been made. In 1963, the Government agreed to pay certain marginal gold mines the cost of pumping the water that seeped in from abandoned underground workings. This assistance was granted on an annual basis, but has been renewed regularly. By June 1966, 7 mines had received about R 1.4 million ($2 million) under these payments. In addition, since April 1964, 10 mines had received R 6.8 million in loans to cover working losses that were up to 10 per cent of revenue. In mid-1967, assistance at an annual rate was R 4 million. The R 12 million paid since the inception of the schemes had been responsible for aggregate additional gold production of 4 million ounces, valued at R 100 million.54 The proportion of South African gold output produced at a working loss has in fact been declining in recent years, to 2.2 per cent in 1966 (Table 13); the special factors involved are discussed in Appendix I.

The Deputy Chairman of Rand Mines said in May 1967 that 3 mines in the group (City Deep, Consolidated Main Reef Mines, and Crown Mines) would have closed “some time ago” without state assistance and efforts to increase productivity; that another mine, Durban Deep, had applied for assistance; and that East Rand Proprietary Mines might soon have to do so. These 5 mines, all in the Johannesburg area of the Central Rand, were currently producing almost 1.7 million ounces valued at R 42 million, about 5½ per cent of total South African production.55

This special assistance was designed broadly to be available to mines that otherwise might close but which would be restored to profitability by an increase in the international price of gold. The payments involved were relatively small (a comparison with subsidies in other producing countries is shown in Table 5A), and the Government was pressed to extend them.

New arrangements to this end were announced by the Minister of Finance in his Budget Speech of March 27, 1968. Direct assistance to mines in connection with the pumping of water, which amounted to R 1.9 million in the financial year 1967/68, is to continue as in the past. But the loan scheme, which absorbed R 3.3 million in this financial year, was replaced by a new scheme directed specifically to gold or uranium mines that “are likely to close down within eight years if not assisted, and with State assistance can show a significant potential increase in life and in [production] ….”56 Such mines may qualify for a reduction in tax or for an actual grant (”negative tax”). On being classified as an “assisted mine,” a mine will immediately lower its operating pay limit, as normally calculated, by approximately 20 per cent and conduct its future mining operations in accordance with the (reduced) average grade of ore reserves determined on the basis of the lower pay limit. The new scheme was expected to involve an average cost of some RIO million a year for the next 8 years, including in this estimate both grants and loss of tax and lease considerations.

An important indirect subsidy for marginal mines in South Africa has developed through government tolerance, and promotion, of mergers and working arrangements between profitable and unprofitable mines, both within the same group and between mines in different groups. Because of the particular progression in the rate of taxation of profits, and the substantial benefits that can be secured from the application of tax losses, such mergers involve substantial savings.57 This technique has been extended to encourage the opening of a large new medium-grade deep mine being developed at Vaal Reefs South. The capital expenditure necessary to bring the mine to commercial viability, seen as the point at which profits can finance future capital expenditure, was estimated at R 60 million, on the basis of which the formation of a new company to acquire the mining lease in the usual way was judged to be uneconomic. Instead, special arrangements were made to enable the new mine to qualify for tax treatment as part of Vaal Reefs, so that capital spending on the new mine becomes allowable against Vaal Reefs’ taxable income from its existing mine. Savings by this company in tax and lease payments should provide approximately half the capital costs of the new mine.58 This technique opens out major new possibilities, in the direction urged by leaders of the gold-mining industry, of switching the objectives of government fiscal policy for the gold mines from the extraction of maximum revenue to the encouragement of maximum production.59 Mr. H. F. Oppenheimer, in the statement cited in footnote 58, stated that by agreeing that the Vaal Reefs’ extension should be mined as part of the operations of the existing mine

the Government was in effect accepting a reduction in the taxation payable by the Vaal Reefs mine over the next few years in order to secure higher gold production and higher tax receipts in the more distant future. … It may well be that it is through the negotiation of special arrangements of this sort that the Government will best be able to encourage the expansion and development of the gold mining industry which is so vital to the future of the South African economy.

Fiscal encouragement of new mining ventures could presumably also be achieved through acceptance by the state of commitments for certain “infrastructure” expenditures hitherto borne by the mining companies. Thus, the exceptional expense of opening up entirely new mining areas has resulted partly from the provision of social capital in the form of housing, hospitals, and schools, which could appropriately be provided by the state when prospective mining profits are no longer sufficient to justify such capital expenditures. A limited precedent exists in the acceptance by the Government, from 1956, of responsibility for payments under legislation concerning silicosis, which had previously cost the industry some R 1 million a year. Such assistance, of course, does not involve any differential boost to development of new output.

Fiscal alleviations in this direction could go a long way before they would neutralize the differential burden of taxation borne by South African gold mining, let alone amount to a net subsidy. Between 1955 and 1965, tax and lease payments due to the Government from gold-mining companies rose by 3.7 times, to R 126.5 million (Table 7), and government revenue from gold mining rose from 5.4 per cent of total revenue to 10.7 per cent (Table 8 and Chart 3). In the same period, in which gold output doubled, the industry’s working profits rose by 2.7 times (profits from gold alone, excluding uranium, rose by 3.5 times). Thus, government revenue as a percentage of the working profits of the gold-mining companies rose from 27 per cent to 38 per cent (in 1966, the ratio reached 41 per cent); since dividends rose roughly in line with profits, the pinch was applied to retentions, which, at a time when future prospects were in any case clouded, may have caused a particular constraint on the financing of future expansion. Thus, the average effective rate of tax paid by the gold-mining industry as a whole in 1965 was well in excess of the nominal tax rate of companies outside the gold and diamond mining sector, then 30 per cent; the effective rate of tax paid by these companies was of course notably lower.

Table 7.

South Africa: Gold-Mining Profits and Taxation, 1955–66

article image
Sources: Transvaal and Orange Free State Chamber of Mines, Annual Reports, and 1966 Quarterly Statements.

Tax payable by gold mines inclusive of government share of profits under lease arrangements.

Column 3 as percentage of column 2.

Preliminary estimate.

Table 8.

South Africa: Governmet Revenue from Gold Mines and Total Revenue, 1913–66

article image
Sources: 1913–55: Bureau of Census and Statistics, Union Statistics for Fifty Years, 1910–1960 (Pretoria, March 1960); 1956–65: Bureau of Statistics, Statistical Year Book, 1966 (Pretoria); 1966: Bureau of Statistics, Monthly Bulletin of Statistics (Pretoria, March 1967).

Costs, Yields, and Profits

With these particular characteristics in mind, the relationship of South African gold output to relative profitability may now be examined more closely. Chart 4A compares the “real” price of gold in terms of U.S. dollars with the “real” price in South African rand, i.e., adjusting the actual rand price60 to the South African wholesale price index. It will be seen that the dollar and rand lines moved closely together until South Africa’s devaluation in line with sterling in 1949. Part of the resulting increase in the local gold price was, however, quickly offset by sharp consequential inflation in South Africa. This was in part a direct consequence of the increased profitability of gold mining, which prompted demands by the (white) miners’ trade unions for a share in the windfall through special bonuses.61


South Africa: Gold Output, Relative Prices, Grade, and Tonnage

Citation: IMF Staff Papers 1968, 003; 10.5089/9781451969160.024.A002

1 Rand price of gold divided by index of South African wholesale prices (see column 6 of Table 17).2 As in Chart 1.

But while South African gold output correlates little better with the real rand price than with the real dollar price (Chart 4B), there is a general inverse correlation between the real rand price and average grade of ore milled (Chart 4C). The grade rose in response to reduced profitability between 1915 and 1920, fell sharply with high profitability in the 1930’s, and fell momentarily in response to the devaluation in 1949, before beginning its latest and steepest rise. This has more than compensated reduced relative profitability, both on this general measure of the real rand price and by the measure of actual mining costs in relation to revenue, so that profits have increased in absolute terms and in relation to output (Table 9). Thus, the average grade in 1966, at about 7¾ dwt. (just over one third of a troy ounce of fine gold per ton extracted), was comfortably above its level through the 1920’s—even though the real rand price of gold was also quite substantially higher (Chart 4C). Until the latest phase, which began in the mid-1950’s, movements in grade also showed an inverse correlation with tonnage of ore milled (Chart 4D) as a result of the same general influence (increased profitability involving (1) an extension in the amount of payable ore, (2) an increase in tonnage, and, with mining to the average grade, (3) a fall in grade). In this way, until this latest phase, the inverse movement of grade and tonnage was a moderating influence on output (Chart 4E), and largely explained the lack of correlation between relative profitability and output noted on pages 415-22. In terms of tonnage milled (i.e., “ex-grade”), the relationship was fairly strongly positive from about 1915 to 1951 (Chart 4F).62

Table 9.

South Africa: Gold Production, Costs, and Profits, 1910–66

article image
article image
Sources: Transvaal and Orange Free State Chamber of Mines, Annual Reports. Leo Katzen, Gold and the South African Economy (Cape Town, 1964).

All producers.

Members of Chamber of Mines. Figures for 1961–66 differ slightly from those cited in text, which exclude primary uranium producers; this category was shown separately only from 1961.

Members of Chamber of Mines; from 1961, excludes primary uranium producers. See footnote 91 on page 478.

Column 8 divided by gold output of producers covered therein: i.e., by column 3 less gold output of nonmembers of Chamber and of primary uranium producers after 1961.

Average for period April 1 to December 31, 1922.

Between the early 1950’s and 1965, this earlier stabilizing influence entirely disappeared. Grade and tonnage milled rose together, and, with the grade rising more than sufficiently to offset the increase in working costs and tonnage rising despite it, output on both counts rose out of line both with earlier cost and relative price relationships in South Africa and with concurrent movements in other gold producing countries (Chart 1). In 1966, however, this unusual relationship was broken, as tonnage dropped and a continued rise in grade brought only a small increase (1.1 percent) in output.

The increase in South African gold output during 1955-65 was therefore exceptional in character as well as in size. The dominant influence was the development of mines in the new areas in the Far West Rand, Klerksdorp, and Evander, as well as in the Orange Free State. These have generally been higher yielding than the older mines in the Central and East Rand, but the gold-bearing seams also generally occur at greater depths, increasing the cost of mining at any given level of factor prices and technology. The old Rand was characterized by unusually low-grade ore together with unusually low costs of extraction in relation to the volume of ore available. The new Rand has ore of a grade that is no longer exceptionally low by current international standards, but that has distinctly higher costs of extraction, particularly in initial capital costs. (The relatively low-grade, low-cost mines of the Evander field are an exception, but this field is regarded as geologically separate from the Rand rather than an extension of it.) Thus, quite aside from any general “cost squeeze,” the grade of ore worked in these newer areas would not be expected to fall to the very low levels traditional in the old Rand, with an average sometimes below 4 dwt. a ton.63

In practice, the approximate doubling in grade of South African output as a whole between the mid-1950’s and the mid-1960’s represents three main influences: (1) The increasing weight of output from the newer mining areas, which by 1965 accounted for more than three fourths of total output. (2) Some increase in grade within the newer group attributable to fuller development of important high-yielding mines. (3) In both the newer and the older groups, the normal “economic” influence of reduced relative profitability. It is only in this case that increases in grade are a direct effect of rising factor costs. In the first two cases they can be regarded as a (probably minor) cause, insofar as increased availability of high-grade ore increases the industry’s ability to concede increases in factor costs.

In sum, there can be little doubt that the first two influences have preponderated; this explains the supporting, rather than inverse, movement in tonnage milled. This has also been reflected in profit margins: until 1964, there was a steady increase in profits per ounce of gold extracted, as well as in profits per ton of ore milled (which rise directly as a result of an increase in grade of ore); in 1965 and again in 1966 profits per ounce declined (Table 9). A “defensive” rise in grade simply in response to a rise in the pay limit, with exactly offsetting reductions in tonnage of ore milled, would normally leave both the working profits in terms of ounces of gold (i.e., revenue) and the total output of gold unchanged, as revenue per ton rose proportionately to costs per ton. (Working profits per ton would rise, as a smaller tonnage would produce the same revenue.) From 1955 to 1964, revenue per ton rose substantially faster than costs per ton, with a corresponding substantial net rise in working profits per ounce.

The rise in costs per ton itself reflects in part the increased cost of mining to greater depths in the new mines, and in that sense is linked to the rise in revenue from higher grades, quite aside from any accompanying increase in factor prices. Such increases became particularly marked in 1965, when costs per ton rose by a full 25 cents (South African), or by 4½ per cent, compared with an average annual increase of 15 cents from 1961 to 1964. A further increase in these working costs, at about the same rate as in 1965, took place in 1966; there was a marked deceleration in 1967. The Anglo American Corporation in its 1964 Annual Report cited a calculation that each increase of 5 per cent above the existing level of average working costs reduced potential gold production by approximately 11 million ounces, or by $385 million. In the spring of 1967, the Gold Producers’ Committee of the Chamber of Mines declared that the effect of a continuing rate of inflation of 4 per cent a year would be as follows: of an estimated 618 million ounces worth $21.6 billion “remaining to be extracted” at current cost levels, some 27.5 per cent (worth nearly $6 billion) would be written off.64 On the same assumption, gold production in the mid-1980’s would be reduced to one sixth of the current level.

Forecasts of this kind involve familiar difficulties. These relate both to the limited economic significance of ore reserves currently recorded (see footnotes 3 and 22) and to the fact that increases in factor costs, important as they are in determining total costs and profit margins, may also have significant influences on other cost and revenue elements. Specifically, increased factor costs, as noted below, may increase pressures for improved labor utilization and productivity, on the one hand, and for increased government fiscal assistance or eventually for exchange rate adjustment, on the other. Equally, the effect of inflation in raising the flotation pay limit for new mines (i.e., the average grade of ore necessary to make development of a new mine an acceptable financial venture) can be offset by special fiscal alleviations of the kind noted on pages 451-53.65 The impact of inflation on gold production in an economy where the level of gold production is a significant element in national economic performance is therefore not a direct or simple one: it is likely to redound in the medium term on the conditions influencing gold production itself. This does not, of course, make the impact of inflation less serious. In the same measure, indeed, the very importance of gold production in the economy makes the impact of inflation especially far reaching.

Working costs per ton doubled between 1949 and 1964, while wholesale prices rose by about two thirds. But, as indicated above, there is limited significance in such comparisons of working costs over a period in which grade has also changed substantially.66 On a longer comparison of 1965 with the 1920’s, when the grade was broadly similar (Chart 4E) and the money price of gold in terms of rand had roughly trebled (Chart 2), working costs per ton had roughly doubled, while wholesale prices had risen between 2 and 3 times. The level of factor and material costs increased by more than wholesale prices, with particularly sharp increases in earnings of white employees (Table 10); but the effect of this was evidently absorbed by improvement in technology and productivity. In terms of daily tonnage raised per employee (Table 11), this improvement was in fact most rapid during the 1920’s, with a reduction of one third for white employees; though comparisons with the slightly slower pace of improvement in the 1950’s must again take into account the increased grade (and increased average depths worked) in the later period. Between 1960 and 1964, a new phase of rapidly increasing productivity on this measure appears to have begun. As a result, almost for the first time in the history of the South African industry, shortage of labor ceased to be a limiting factor on production. The major expansion in output in the first half of the 1960’s was achieved with a slowly declining labor force. In 1965, mines belonging to the Chamber employed 44,098 Europeans, 10.3 per cent fewer than in 1961, and 375,329 non-Europeans, 5.9 per cent fewer than in 1961.67 These trends were continued in 1966.

Table 10.

South Africa: Annual Average Cash Earnings in Mining

(In rand)

article image
Sources: All mines: D. Hobart Houghton, The South African Economy (Oxford University Press, 1964), p. 238. Gold mines: Transvaal and Orange Free State Chamber of Mines, Annual Reports, 1965 and 1966, adjusted in columns 3 and 4 by South African consumer price index (Bureau of Statistics, Statistical Year Book, 1966, Table 1-16). Figures cover gold mines that are members of the Chamber.
Table 11.

South Africa: Gold-Mining Employees at Work per Thousand Tons Hoisted per Day

(Large producing mines)

article image
Sources: Annual Reports of the Government Mining Engineer (as quoted by Leo Katzen, op. cit., p. 26, for 1910–60).