Gernot Köhler

School of Computing and Information Management
Sheridan College
Oakville, Ontario

Canada L6H 2L1

This study takes a global view of money. The term "global money" is appearing in recent discussions1 and there is the occasional literature reference to "world money"2. My thesis in this paper is that money has a global structure. Stated more precisely, I am contending that (1) the value of money is non-homogeneous throughout the world system (even after exchange rates have been taken into account); that (2) the value of a country's money within the world system tends to correlate with that country's economic status in the world system; that (3) the exchange rate system is one of the mechanisms by which high-wage countries extract value from low-wage countries, thus adding to other mechanisms of unequal development and polarization; and that (4) the currencies of low-income countries tend to be undervalued.


It is common knowledge that: (1) money is a measure of value; and (2) money itself has a value. To elaborate the first point: When we buy goods or services, we pay with money (except in barter situations). For example, one pound of bananas is worth x dollars or y rupees in a specific place at a specific time. Or, the entire physical output of a country's goods and services of a year is measured, not in pounds, tons or number of pieces, but in terms of a single money figure (GDP). In these situations we use money as a measure of value.
     To elaborate the second point: We know that money, our measure of value, does not have a constant value. The value of money may change over time ("inflation") or across space, from one currency zone to another ("exchange rate"). The value of money varies diachronically (longitudinally, over time) and synchronically (cross-nationally, from country to country). Of course, both aspects may be combined.


Longitudinal valuation of money is a well-known exercise. The measurement of inflation is in the public eye virtually all the time as, for example, when the media report the latest inflation rates. These inflation rates are based on scientifically established methods of inflation measurement. Typically, economists define a "basket of goods and services", measure its price at different points in time and calculate inflation rates from that. While there are disagreements about the finer points of inflation measurement, the longitudinal valuation of money (inflation measurement) is a well-established practice.


In contrast, the cross-national valuation of money is, partly, well-known and, partly, more obscure. In order to determine the relative value of one currency in comparison with another, two conflicting concepts and measurement procedures exist, namely:

(1) currency exchange rates between two countries (i.e., the rates at which units of one currency are exchanged for units of another currency; e.g., how many dollars do I get for 100 rupees; how many rupees do I get for 100 dollars? etc.); and
(2) purchasing power parity rates (PPP rates) between two countries (i.e., the ratio of the purchasing power of money in countries A and B; e.g., how much money do I need in order to buy a pair of shoes in country A; and how much money do I need in order to buy an equivalent pair of shoes in country B?).
The two concepts differ significantly, even though they seem to be similar on the surface. The numbers which one obtains from either method are highly divergent in many situations.

My thesis concerning the global structure of money arises from the difference between these two methods of cross-national valuation of money -- namely, exchange rate versus purchasing power parity rate (PPP rate).


Here are some figures which exemplify the difference which I am talking about (Table 1):


               ( G N P / capita, 1992 )

COUNTRY        METHOD 1:           METHOD 2:          COMPARISON
               actual              PPP rates
               exchange rates      (purchasing power
                                   parity rates)

               US dollars          "international

Japan          28 190              20 160             col. 1 > col. 2

USA            23 120              23 240             similar values
Germany        23 030              20 610             similar   "
UK             17 790              16 730             similar   "
Australia      17 260              17 350             similar   "

                                                      PPP rate is greater
                                                      by a factor of:
Brazil          2 770               5 250             (factor  1.9)

Russia          2 510               6 220             (factor  2.5)

China             470               1 910             (factor  4.1)

India             310               1 210             (factor  3.9)

Bhutan             62                 630             (factor 10.2)
Mozambique         60                 570             (factor  9.5)

Source: World Bank. World Development Report 1994, p. 162, p. 220

As the examples in Table 1 show, the value of one country's money in relation to another country's money can be calculated with two different methods and expressed in two different rates -- exchange rates and PPP rates. For example, the actual exchange rate between Indian rupees and U.S. dollars "shows" that Indian GNP per capita was 310 dollars in 1992. In contrast, scientific measurement of purchasing power parity (PPP) "shows" that Indian GNP per capita was 1210 dollars, i.e. four times as much as shown in terms of exchange rates. In contrast, OECD countries like USA, Germany, UK and Australia have exchange rates to the U.S. dollar which are very similar to the relative purchasing power rates of the currencies.

Which of the two rates (exchange rate versus PPP rate) is the correct one? When we ask the question: "What is the value of rupees in relation to the value of U.S. dollars?", we obtain two sharply different answers (e.g., "310" and "1210" for GNP per capita in the above). That constitutes a puzzle -- not for practitioners, because they use exchange rates, but for scientists. It can be observed that exchange rates are real, in the sense that they are being used by money traders and traders of goods and services. The PPP rate, on the other hand, is scientifically arrived at. Does that mean that PPP rates are not real? Such a conclusion would be untenable. On the contrary, PPP rates are real as well -- they are based on carefully established methods of measurement.

The situation is one of a dual standard or a double standard -- namely, we have two standards for evaluating the same object. The object to be evaluated is the relationship between money A and money B (the evaluandum) and the two conflicting standards are exchange rate (standard 1) and PPP rate (standard 2). In order to deal with this discrepancy in a comprehensive manner, we require some theory.


In order to address the problem as outlined, we require four broad types of theory, namely: 
(1) measurement theory -- what is it that we are measuring?

(2) empirical theory and historical explanation -- how can we explain what we see? What causes it? How did it come about?
(3) theory of value -- how can we evaluate what we see? Which standard is the correct one or the best?
(4) theory of policy (praxis) -- what should be done, if anything?

In this study I will not attempt a comprehensive treatment of these issues. Instead, I will focus on two contentions, namely: (1) the discrepancies between the two measurement methods are systematic (non-random) and correlate with the global center-periphery structure (empirical claim); and (2) this situation constitutes a form of exploitation (normative claim).


Those who question the validity of PPP measurement may be reminded of the fact that this methodology has been developed at the World Bank. Furthermore, PPP measurement has great similarity with inflation measurement. The method consists of the following: (a) a "basket of goods and services" is defined (as in inflation measurement); (b) prices for this "basket" are collected in the countries around the world -- e.g., in country 1 in rubles, in country 2 in rupees, in country 3 in yuan, etc.; (c) the prices for the "basket" in the different countries are compared and permit the calculation of purchasing power parity rates (PPP rates). The World Bank developed and refined this methodology during the past decades to a point where it is as solid as familiar inflation measurement.


This investigation may seem a bit abstract. It has great practical relevance, however. There is an entire chorus of economists who keep repeating that the exchange rates of poor countries are overvalued. The IMF, in particular, has, over the past two decades and as part of its "Structural Adjustment Programs" (SAPs), forced many countries to devalue their currencies, using the argument that the currency was "overvalued". This practice has important practical implications for the countries and people affected by SAPs and, furthermore, it reveals two things, namely: (1) the IMF (and the chorus of economists referred to) has a "theory of value" regarding exchange rates. This theory may be largely implicit, but it is there, or else they could not arrive at the evaluative judgment that "exchange rate of country X is overvalued". And: (2) value theory of exchange rates is of great practical relevance. Whereas most economists agree with the view that the currencies of low-income countries tend to be overvalued, if anything, I claim the opposite, namely, that the currencies of low-income countries tend to be undervalued.


A look at Table 1 above suggests several observations, namely:
(1) for the countries of the core of the world system (OECD countries) exchange rates (method 1) and purchasing power rates (method 2) differ not at all or relatively little.
(2) for the countries of the periphery and semi-periphery of the world system (non-OECD countries) exchange rates (method 1) and purchasing power rates (method 2) yield significantly different results. The difference can be stated in two ways -- namely, either: exchange rate values are lower than PPP values, or: the purchasing power value of the country's currency is greater than the exchange rate value.
(3) overall, for all countries, it can be observed that the discrepancy tends to be greatest for the poorest countries and least for the richest countries. My Table 1 shows only eleven countries. However, when all countries are examined, the same correlation emerges. Statistical testing of these observations leads to the following results:
(4) for all countries, there is a statistical correlation between the country's GNP per capita and the discrepancy between exchange rates and PPP rates. In other words, the poorer the country, the lower will be the exchange rate value of its currency in relation to the PPP value of its currency. Kravis and Lipsey examined this relationship for 34 countries for the year 1975 (including 12 OECD countries and 22 non-OECD countries) and found a very strong relationship between the discrepancy among the two rates and GDP per capita.3 Subsequent studies confirmed the existence of this relationship. My own calculations for 120 countries (based on World Bank data) yield the following observations for the year 1995 (Table 2):


           between Currency Value Distortion and GNP per capita, 1995
           (n=120 countries)

1. Correlation between the "distortion of currency value"
            and "income level" (GNP/capita).................r = - 0.63

2. Average distortion factor for 120 countries .........avg d =   2.65                        (standard deviation 1.5)
3. Maximum distortion factor (Mozambique)...............max d =  10.1             Minimum (Switzerland, Japan)................min d =   0.6
4. Classes: average distortion factor for:

            (a)low-income (43 countries)................avg d =   4.0
            (b)middle-income (52 countries).............avg d =   2.4
            (c)high-income (25 countries)...............avg d =   0.98

5. Average GNP/capita (at exchange rate)..................US $ 6025
            (standard deviation 9408)
6. Minimum GNP/capita (at exchange rate)(Mozambique)......US $   80
            Maximum (at exchange rate)(Switzerland).......US $40630

Source: World Bank data. Details, see endnote (4).


In the context of world system theory and related theories (structural theory, imperialism theory), the observed correlation means that a country's socio-economic status in the world system and the relative value of the country's money within the world system are related. Topdog countries tend to have "hard" or "strong" currencies (i.e. valuable currencies); underdog countries tend to have "soft" or "weak" currencies (i.e. less valuable currencies). The general power/wealth gradient in the world system can thus be found once more in the value structure of global money.

In terms of causality, it may be asked whether the global power/wealth structure determines the global money structure or vice versa. I assume that both causalities exist. The global power/wealth structure contributes to the global money structure and the global structure of money feeds back into the perpetuation of the global power/wealth structure.


I will now shift from empirical observation to normative reasoning (theory of value). How can we evaluate this situation?

The discussion can be organized around two diametrically opposed evaluative (normative) propositions, namely:

Proposition (1): The exchange rates of low-income countries tend to be overvalued.

Proposition (2): The exchange rates of low-income countries tend to be undervalued.

Let's examine both propositions.


The over-valuation argument is common and is being used by the IMF, as mentioned earlier. In order to arrive at a verdict of "over-valuation" of a currency, the judging mind must have two things, namely, (a) certain kinds of factual information and (b) a normative standard by which to evaluate the facts; or an evaluation method which implies a normative standard. What normative standard or "yardstick" is being applied by experts who reach the verdict of "overvalued"?

In the practice the "Structural Adjustment Programs" of the IMF, conditions tend to be imposed on various countries -- conditions which tend to include stipulations about currency. "Typical conditionalities might be: the borrower's budget or balance of payments deficit must be reduced by x percent within one year; ... the currency must be devalued by x percent within six months, etc."5 There is thus a great concern for balancing the government budget and balancing the balance of payments. The standard by which a currency is evaluated is the balancing of the balance of payments. The connection between this standard and the judgment of "overvaluation" is stated, for example, in the following:

     "The currencies of most developing countries are overvalued ... When      exchange rates are fixed, a surplus demand for foreign currency tends
     to develop, which must be controlled ..."6

The balancing of the balance of payments is thus a major "yardstick" with which poor countries' currencies are evaluated. (It should be noted that the same yardstick is being applied when the currencies of high-income countries are evaluated.) This may be a valid standard. However, I contend that the following is also a valid standard, even though that may seem paradoxical at first.


When a low-income country (with a structurally distorted currency value, see Table 1) trades with a high-income country, the high-income country gains a quantity of real value which does not show up in any account and the low-income country loses a quantity of real value which does not show up in any account. In colonial times, traders may have exchanged cheap glass beads for valuable ivory. Both sides agreed to the deal. Similarly, a low-income country may make a deal with a high-income country and the deal is balanced in monetary terms at the prevailing exchange rates. However, a quantity of real value has been extracted from the low-income country in this deal which does not show up in any account.

The normative standard in this argument is "real value". I measure "real value" in terms of purchasing power parity (PPP). On this basis it can be argued that the currencies of the low-income countries tend to be undervalued. Here is an example.

First, let's eliminate the theoretical possiblity of complete autarky of two countries. In this situation there is no trade and no international money exchange. PPP rates can be measured, but there are no exchange rates. This situation is of no practical interest.

Next, let us examine a situation of balanced trade between country LIC (low-income country) and country HIC (high-income country), with no financial investment across borders, just payments for traded goods and services. (Let's further assume that LIC has some features of India and HIC has some features of USA. I am using the distortion factor of 3.9 for India from Table 1; the rest in the following is "made up".)

Here is a 2-country scenario (hypothetical):
[LIC = low-income country
 HIC = high-income country]
1. LIC has a GDP of 1200 rupees

2. HIC has a GDP of 1010 dollars
3. The exchange rate of rupee : dollar is 20 : 1
4. The two countries trade with each other. The volume of trade between the two countries is 200 rupees = 10 dollars in each direction. In other words, LIC exports goods and services valued as 200 rupees (or 10 dollars) and HIC exports goods and services valued as 10 dollars (or 200 rupees). The trade is balanced. The balance of payments is balanced.

Now let us examine the implication of the distortion factor. The distortion factor in cross-national currency valuation is 3.9 (from Table 1). This means that the purchasing power rate (PPP rate) between LIC's rupee and HIC's dollar is not 20 : 1 but, rather, 20/3.9 : 1 = 5.13 : 1. I am rounding this off to 5 : 1 .

When we assume that the PPP rate reflects the true value of the currency, then it can be observed that:
(a) from LIC's point of view:
LIC exported goods and services worth 200 rupees. In exchange, LIC imported goods and services worth 200 / 20 = 10 dollars at exchange rate. However, LIC would have imported 200 / 5 = 40 dollars worth at PPP rate. The monetary distortion factor had the effect that LIC lost 40 - 10 = 30 dollars worth of imports due to the distortion factor. This loss of 30 dollars, at PPP rate, corresponds to 150 rupees.
(b) from HIC's point of view:
HIC exported goods and services worth 10 dollars. In exchange, HIC imported goods and services worth 10 * 20 = 200 rupees at exchange rate. However, HIC would have imported 10 * 5 = 50 rupees worth at PPP rate. The monetary distortion factor had the effect that HIC gained 200 - 50 = 150 rupees worth of imports due to the distortion factor. This gain of 150 rupees, at PPP rate, corresponds to 30 dollars.
(c) Summary:
Due to the monetary distortion factor between the two currencies, LIC lost 150 rupees worth of import value; HIC gained 30 dollars worth of import value. These losses and gains of value do not show up in the accounts, because the value of money is itself structurally deformed. The trade is formally balanced.

As a percent of GDP, the gains and losses are, as follows:
LIC has an "invisible loss" (unrecorded loss) of 150 rupees. With a GDP of 1200 rupees, the invisible loss of value is 150/1200 = 12.5% of GDP. HIC has an "invisible gain" (unrecorded gain) of 30 dollars. With a GDP of 1010 dollars, the invisible gain of value is 30/1010 = 3.0% of GDP.

Based on this reasoning, I conclude that the currency values of low-income countries tend to be undervalued. The effect is exploitative. The core countries appear to extract a large amount of value from the periphery countries through the clever monetary device called exchange rate system.
Babones has recently made a parallel observation with respect to global center-periphery relations, namely that: "Surplus value is not extracted solely through coercion, but also through the working of financial markets." (He is thinking of credit markets.)7


Global money has a center-periphery structure; high-income countries tend to have "hard" currencies (more valuable money) and low-income countries tend to have "soft" currencies (less valuable money). The distortion factor for a country's currency value can be as high as 1 : 10. The distortion (undervaluation) of a country's currency correlates with the income level of the country -- the lower the income, the greater the distortion (undervaluation). The result is exploitation through the exchange rate system. The currencies of low-income countries tend to be undervalued, not overvalued as many economists claim. This judgment is contentious and depends on the standard of value which is applied. The prevailing standard has a bias against low-income countries.


1. For example, a recent conference in Italy included the topic of "Global Money, Capital Restructuring and the Changing Pattern of Production". See, post by Riccardo Bellofiore, pkt mail archive, 13 Nov 1997,

2. E.g., in the following passage: "the circulation process of industrial capital is characterized by ... the existence of the market as a world market. ... the money functions here as world money." (K. Marx, Capital, vol. 2, ch. 4, p. 190 in Penguin edition 1978 [reprinted 1992])

3. I.B. Kravis and R.E. Lipsey, "Toward an Explanation of National Price Levels," Princeton Studies in International Finance, No. 52, November 1983, p. 21 (Princeton University, USA. Department of Economics.)

4. Data source: World Bank. World Development Report 1997, p. 214-215, Table 1. "GNP per capita" in the correlation is in 1995 US dollars (exchange rate value). The distortion factor is the quotient of GNP/capita "in 1995 international dollars" (i.e. the PPP value), divided by GNP/capita in "1995 US dollars" (i.e. exchange rate value).

5. R.S. Brown, "The IMF and the World Bank in the New World Order," in: P. Bennis and M. Moushabeck, eds., Altered States: A Reader in the New World Order. New York, USA: Olive Branch Press, 1993, p. 122.

6. W. Lachmann, Entwicklungspolitik (Muenchen, Germany: Oldenbourg Verlag, 1994), p. 196. My translation.