Towards a General Theory of the Balance of Payments1
HARRY G. JOHNSON
The theory of the balance of payments is concerned with the economic determinants of the balance of payments, and specifically with the analysis of policies for preserving balance-of-payments equilibrium. So defined, the theory of the balance of payments is essentially a post-war development. Prior to the Keynesian revolution, problems of international disequilibrium were discussed within the classical conceptual framework of ‘the mechanism of adjustment’—the way in which the balance of payments adjusts to equilibrium under alternative systems of international monetary relations—the actions of the monetary and other policy-making authorities being subsumed in the system under consideration. While the Keynesian revolution introduced the notion of chronic disequilibrium into the analysis of international adjustment, early Keynesian writing on the subject tended to remain within the classical framework of analysis in terms of international monetary systems—the gold standard, the inconvertible paper standard—and to be concerned with the role and adequacy in the adjustment process of automatic variations in income and employment through the foreign trade multiplier. Moreover, the applicability of the analysis to policy problems was severely restricted by its assumption of general under-employment, which implied an elastic supply of aggregate output, and allowed the domestic-currency wage or price level to be treated as given, independently of the balance of payments and variations in it.
The pre-war approach to international monetary theory reflected the way in which balance-of-payments problems tended to appear at the time, namely as problems of international monetary adjustment. Since the war, for reasons which need not be elaborated here, the balance of payments has come to be a major problem for economic policy in many countries. Correspondingly, a new (though still Keynesian) theoretical approach to balance-of-payments theory has been emerging, an approach which is better adapted to postwar conditions than the ‘foreign trade multiplier theory’ and ‘elasticity analysis’ of the pre-war period in two major respects: it poses the problems of balance-of-payments adjustment in a way which highlights their policy implications, and it allows for conditions of full employment and inflation.
The essence of this approach, which has been termed ‘the absorption approach’, is to view the balance of payments as a relation between the aggregate receipts and expenditures of the economy, rather than as a relation between the country’s credits and debits on international account. This approach has been implicit to an important extent in the thinking of practical policy-makers concerned with balance-of-payments problems in post-war conditions. Its main formal development is to be found in the works of Meade, Tinbergen, and Alexander, though many others have contributed.2 The purpose of this chapter is to synthesise and generalise the work of these writers, and to use their approach to clarify certain aspects of the balance-of-payments policy problem.
Let us first summarise the traditional approach to balance-of-payments theory. The balance of payments must necessarily balance when all international transactions are taken into account; for imbalance or disequilibrium to be possible, it is necessary to distinguish between ‘autonomous’ international transactions—those which are the result of the free and voluntary choices of individual transactors, within whatever restrictions are imposed by economic variables or policy on their behaviour—and ‘induced’ or ‘accommodating’ international transactions—those which are undertaken by the foreign exchange authorities to reconcile the free choices of the individual transactors—and to define the ‘balance of payments’ to include only autonomous transactions. To put the point another way, balance-of-payments problems presuppose the presence of an official foreign exchange authority which is prepared to operate in the foreign exchange market by the use of official reserves so as to influence the exchange rate; and ‘disequilibrium’ is defined by changes in the official reserves, associated with imbalance between the foreign receipts and foreign payments of residents of the country, where ‘resident’ is defined to include all economic units domiciled in the country except the foreign exchange authority.3
The ‘balance of payments’ appropriate to economic analysis may then be defined as
B = Rf − Pf
where Rf represents aggregate receipts by residents from foreigners, and Pf represents aggregate payments by residents to foreigners. The difference between the two constitutes a surplus (if positive) or a deficit (if negative); a surplus is accompanied by sales of foreign currency to the exchange authority by residents or foreigners in exchange for domestic currency, and conversely a deficit is financed by sales of domestic currency by residents or foreigners to the authority in exchange for foreign currency. To remedy a deficit, some action must be taken to increase receipts from foreigners and reduce payments to foreigners, or increase receipts more than payments, or reduce payments more than receipts; and conversely with a surplus (though the rectification of a surplus is not generally regarded as a ‘balance-of-payments problem’).
The ‘balance of payments’ can, however, be defined in another way, by making use of the fact that all payments by residents to residents are simultaneously receipts by residents from residents; in symbols Rr ≡ Pr. Hence the balance of payments may be written
B = Rf + Rr − Pf − Pr = R − P.
That is, the balance of payments is the difference between aggregate receipts by residents and aggregate payments by residents. A deficit implies an excess of payments over receipts, and its rectification requires that receipts be increased and payments decreased, or that receipts increase more than payments, or that receipts decrease less than payments; and conversely with a surplus. In what follows, however, surpluses will be ignored, and the argument will be concerned only with deficits.
The formulation of a balance-of-payments deficit in terms of an excess of aggregate payments by residents over aggregate receipts by residents constitutes the starting point for the generalisation of the ‘absorption approach’ to balance-of-payments theory—what might be termed a ‘payments approach’—which is the purpose of this chapter. It directs attention to two important aspects of a deficit—its monetary implications, and its relation with the aggregate activity of the economy—from which attention tends to be diverted by the traditional sectoral approach, and neglect of which can lead to fallacious analysis. These two aspects will be discussed in turn, beginning with the monetary implications of a deficit.
The excess of payments by residents over receipts by residents inherent in a balance-of-payments deficit necessarily implies one or the other of two alternatives. The first is that cash balances of residents are running down, as domestic money is transferred to the foreign exchange authority.4 This can, obviously, only continue for a limited period, as eventually cash balances would approach the minimum that the community wished to hold and in the process the disequilibrium would cure itself, through the mechanism of rising interest rates, tighter credit conditions, reduction of aggregate expenditure, and possibly an increase in aggregate receipts. In this case, where the deficit is financed by dishoarding, it would be self-correcting in time; but the economic policy authorities may well be unable to allow the self-correcting process to run its course, since the international reserves of the country may be such a small fraction of the domestic money supply that they would be exhausted well before the running down of money balances had any significant corrective effect. The authorities might therefore have to take action of some kind to reinforce and accelerate the effects of diminishing money balances.
This last consideration provides the chief valid argument for larger international reserves. The case for larger international reserves is usually argued on the ground that larger reserves provide more time for the economic policy authorities to make adjustments to correct a balance-of-payments disequilibrium. But, as Friedman has argued in criticism of Meade,5 there is no presumption that adjustment spread over a longer period is to be preferred—the argument could indeed be inverted into the proposition that, the larger the reserves, the more power the authorities have to resist desirable adjustments. The acceptable argument would seem to be that, the larger the international reserves in relation to the domestic money supply, the less the probability that the profit- or utility-maximising decisions of individuals to move out of cash into commodities or securities will have to be frustrated by the monetary authorities for fear of a balance-of-payments crisis.
The second alternative is that the cash balances of residents are being replenished by open market purchases of securities by the monetary or foreign exchange authority, as would happen automatically if the monetary authority followed a policy of pegging interest rates or the exchange authority (as in the British case) automatically re-lent to residents any domestic currency it received from residents or foreigners in return for sales of foreign exchange. In this case, the money supply in domestic circulation is being maintained by credit creation, so that the excess of payments over receipts by residents could continue indefinitely without generating any corrective process—until dwindling reserves forced the economic policy authorities to change their policy in some respect.
To summarise the argument so far, a balance-of-payments deficit implies either dishoarding by residents, or credit creation by the monetary authorities—either an increase in V, or the maintenance of M. Further, since a deficit associated with increasing velocity of circulation will tend to be self-correcting (though the authorities may be unable to rely on this alone), a continuing balance-of-payments deficit of the type usually discussed in balance-of-payments theory ultimately requires credit creation to keep it going. This in turn implies that balance-of-payments deficits and difficulties are essentially monetary phenomena, traceable to either of two causes: too low a ratio of international reserves relative to the domestic money supply, so that the economic policy authorities can not rely on the natural self-correcting process; or the pursuit of governmental policies which oblige the authorities to feed the deficit by credit creation. In both cases, the problem is associated fundamentally with the power of national banking systems to create money which has no internationally acceptable backing.
To conclude that balance of payments problems are essentially monetary is not, of course, to assert that they are attributable to monetary mismanagement—they may be, or they may be the result of ‘real’ forces in the face of which the monetary authorities play a passive role. The conclusion does mean, however, that the distinctions which have sometimes been drawn between monetary and real disequilibria, for example by concepts of ‘structural disequilibrium’, are not logically valid—though such concepts, carefully used, may be helpful in isolating the initiating causes of disequilibrium or the most appropriate type of remedial policy to follow.
Formulation of the balance of payments as the difference between aggregate payments and aggregate receipts thus illuminates the monetary aspects of balance-of-payments disequilibrium, and emphasises its essentially monetary nature. More important and interesting is the light which this approach sheds on the policy problem of correcting a deficit, by relating the balance of payments to the over-all operation of the economy rather than treating it as one sector of the economy to be analysed by itself.
An excess of aggregate payments by residents over aggregate receipts by residents is the net outcome of economic decisions taken by all the individual economic units composing the economy. These decisions may usefully be analysed in terms of an ‘aggregate decision’ taken by the community of residents considered as a group (excluding, as always, the foreign exchange authority), though it must be recognised that this technique ignores many of the complications that would have to be investigated in a more detailed analysis.
Two sorts of aggregate decision leading to a balance-of-payments deficit may be distinguished in principle, corresponding to the distinction drawn in monetary theory between ‘stock’ decisions and ‘flow’ decisions: a (stock) decision to alter the composition of the community’s assets by substituting other assets for domestic money,6 and a (flow) decision to spend currently in excess of current receipts. Since both real goods and securities are alternative assets to domestic money, and current expenditure may consist in the purchase of either goods or securities, the balance-of-payments deficit resulting from either type of aggregate decision may show itself on either current or capital account. That is, a current account deficit may reflect either a community decision to shift out of cash balances into stocks of goods, or a decision to use goods in excess of the community’s current rate of production, while a capital account deficit may reflect either a decision to shift out of domestic money into securities or a decision to lend in excess of the current rate of saving.
The distinction between ‘stock’ and ‘flow’ balance-of-payments deficits is important for both theory and practical policy, though refined theoretical analysis has generally been concerned with ‘flow’ deficits, without making the distinction explicit. The importance of the distinction stems from the fact that a ‘stock’ deficit is inherently temporary and implies no real worsening of the country’s economic position, whereas a ‘flow’ deficit is not inherently temporary and may imply a worsening of the country’s economic position.
Since a stock decision entails a once-for-all change in the composition of a given aggregate of capital assets, a ‘stock’ deficit must necessarily be a temporary affair;7 and in itself it implies no deterioration (but rather the reverse) in the country’s economic position and prospects.8 Nevertheless, if the country’s international reserves are small, the economic policy authorities may be obliged to check such a deficit by a change in economic policy. The policy methods available are familiar, but it may be useful to review them briefly in relation to the framework of analysis developed here.
To discourage the substitution of stocks of goods for domestic currency, the economic policy authorities may either raise the cost of stock-holding by credit restriction or reduce its attractiveness by currency depreciation.9 Under both policies, the magnitude of the effect is uncertain—depreciation, by stimulating de-stabilising expectations, may even promote stock accumulation—while unavoidable repercussions on the flow equilibrium of the economy are set up. These considerations provide a strong argument for the use of the alternative method of direct controls on stock-holding, an indirect and partial form of which is quantitative import restriction.
To discourage the substitution of securities for domestic currency, the same broad alternatives are available: credit restriction, which amounts to the monetary authority substituting domestic currency for securities to offset substitution of securities for domestic currency by the rest of the community; devaluation, which affects the relative attractiveness of securities only through expectations and may work either way; and exchange controls restricting the acquisition of securities from abroad. Considerations similar to those of the previous paragraph would seem to argue in favour of the use of controls on international capital movements as against the alternative methods available.
In both cases, evaluation of the policy alternatives suggests the use of control rather than price system methods. It should be recalled, though, that the problem is created by the assumed inadequacy of the country’s international reserves. In the longer run, the choice for economic policy lies, not between the three alternatives discussed, but between the necessity of having to choose between them and the cost of investing in the accumulation of reserves large enough to finance potential ‘stock’ deficits. Also, nothing has been said about the practical difficulties of maintaining effective control over international transactions, especially capital movements.
In contrast to a ‘stock’ deficit, a ‘flow’ deficit is not inherently of limited duration. It will be so if the monetary authority is not prepared to create credit, but this is because its existence will then set up monetary repercussions which will eventually alter the collective decision responsible for it, not because the initial decision implied a temporary deficit. If the decision not to create credit is regarded as a specific act of policy equivalent to a decision to raise interest rates,10 it follows that the termination of a ‘flow’ deficit requires a deliberate change of economic policy. Further, a ‘flow’ deficit may imply a worsening of the country’s capital position, providing an economic as well as a monetary incentive to terminate the disequilibrium.11
In analysing the policy problems posed by ‘flow’ deficits, it is convenient to begin by abstracting altogether from international capital movements (other than reserve transactions between foreign exchange authorities) and considering the case of a current account deficit. In this case, if intermediate transactions are excluded, the balance of payments becomes the difference between the value of the country’s output (its national income) and its total expenditure, i.e.
B = Y − E.
To facilitate analysis by avoiding certain complications associated with the possibility of changes in the domestic price level, income and expenditure are conceived of as being valued in units of domestic output. A deficit then consists in an excess of real expenditure over real income, and the problem of correcting a deficit is to bring real national income (output) and real national expenditure into equality.
This formulation suggests that policies for correcting current-account deficits can be classified broadly into two types: those which aim at (or rely on) increasing output, and those which aim at reducing expenditure. The distinction must, of course, relate to the initial impact of the policy, since income and expenditure are interdependent: expenditure depends on and varies with income, and income depends on and varies with expenditure (because part of expenditure is devoted to home-produced goods). Consequently any change in either income or expenditure will initiate multiplier changes in both. It can, however, readily be shown that, so long as an increase in income induces a smaller change in aggregate expenditure, the multiplier repercussions will not be large enough to offset the impact effect of a change, so that an impact increase in output or decrease in expenditure will always improve the balance on current account.12
The distinction between output-increasing and expenditure-reducing policies may usefully be put in another way. Since output is governed by the demand for it, a change in output can only be brought about by a change in the demand for it; a policy of increasing domestic output can only be effected by operating on expenditure (either foreign or domestic) on that output. Given the level of expenditure, this in turn involves effecting a switch of expenditure (by residents and foreigners) from foreign output to domestic output. The distinction between output-increasing and expenditure-decreasing policies, which rests on the effects of the policies, may therefore be replaced by a distinction between expenditure-switching policies and expenditure-reducing policies, which rests on the method by which the effects are achieved.
A policy of expenditure-reduction may be applied through a variety of means—monetary restriction, budgetary policy, or even a sufficiently comprehensive battery of direct controls. Since any such policy will tend to reduce income and employment, it will have an additional attraction if the country is suffering from inflationary pressure as well as a balance-of-payments deficit, but a corresponding disadvantage if the country is suffering from unemployment. Moreover, since the impact reduction in expenditure and the total reduction in income and output required to correct a given deficit are larger the larger the proportion of the expenditure reduction falling on home-produced goods, and since different methods of expenditure-reduction may differ in this respect, the choice between alternative methods may depend on the inflationary-deflationary situation of the economy. Finally, since the accompanying reduction in income may lead to some reduction in the domestic price level, and/or a greater eagerness of domestic producers to compete with foreign producers both at home and abroad, expenditure-reducing policies may have incidental expenditure-switching effects.
Expenditure-switching policies may be divided into two types, according to whether the policy instrument employed is general or selective: devaluation (which may be taken to include the case of a deflation-induced reduction of the domestic price level under fixed exchange rates), and trade controls (including both tariffs and subsidies and quantitative restrictions). Devaluation aims at switching both domestic and foreign expenditure towards domestic output; controls are usually imposed on imports, and aim at (or have the effect of) switching domestic expenditure away from imports towards home goods, though sometimes they are used to stimulate exports and aim at switching foreigners’ expenditure towards domestic output.
Both types of expenditure-switching policy may have direct impact-effects on residents’ expenditure. Devaluation may result in increased expenditure from the initial income level, through the so-called ‘terms-of-trade effect’ of an adverse terms-of-trade movement in reducing real income and therefore the proportion of income saved. Trade controls will tend to have the same effect, via the reduction in real income resulting from constriction of freedom of choice.13 In addition, trade controls must alter the real expenditure corresponding to the initial output level if they take the form of import duties or export subsidies uncompensated by other fiscal changes; this case should, however, be classed as a combined policy of expenditure-change (unfavourable in the case of the export subsidy) and expenditure-switch.
Whether general or selective in nature, an expenditure-switching policy seeks to correct a deficit by switching demand away from foreign towards domestic goods; and it depends for success not only on switching demand in the right direction, but also on the capacity of the economy to make available the extra output required to satisfy the additional demand. Such policies therefore pose two problems for economic analysis: the conditions required for expenditure to be switched in the desired direction, and the source of the additional output required to meet the additional demand.
As to the first question, the possibilities of failure for both devaluation and controls have been investigated at length by international trade theorists, and require only summary treatment here.14 Export promotion will divert foreign expenditure away from the country’s output if the foreign demand is inelastic, while import restriction will divert domestic expenditure abroad if demand for imports is inelastic and the technique of restriction allows the foreigner the benefit of the increased value of imports to domestic consumers. Devaluation has the partial effect of diverting domestic expenditure abroad, via the increased cost of the initial volume of imports, and this adverse switch will not be offset by the favourable effect of substitution of domestic for foreign goods at home and abroad, if import demand elasticities average less than one half.
While the elasticity requirement for successful devaluation just cited is familiar, the approach developed in this paper throws additional light on what non-fulfilment of the requirement implies. From the equation
B = Y − E,
it is clear that, if direct effects on expenditure from the initial income level are neglected, devaluation can worsen the balance only if it reduces total world demand for the country’s output. This implies that the country’s output is in a sense a ‘Giffen case’ in world consumption; and that the market for at least one of the commodities it produces is in unstable equilibrium.15 Neither of these ways of stating the conditions for exchange instability makes the possibility of instability as plausible a priori as their equivalent, reached through sectoral analysis, in terms of elasticities of import demand.
The second, and more interesting, analytical problem relates to the source of the additional domestic output required to satisfy the demand for it created by the expenditure-switching policy. Here it is necessary to distinguish two cases, that in which the economy is under-employed and that in which it is fully employed, for both the relevant technique of analysis and the factors on which the outcome of the policy depend differ between the two.
If the economy has unemployed resources available, the additional output required to meet the additional demand can be provided by the re-absorption of these resources into employment: in this case the switch policy has the additional attraction of increasing employment and income. The increase in domestic output may tend to raise the domestic price level, through the operation of increasing marginal real costs of production, and conversely the foreign price level may tend to fall, thus partially counteracting the initial effects of the switch policy; but such repercussions can legitimately be analysed in terms of elasticity concepts, since under-employment implies that additional factors are available at the ruling price.
If the economy is already fully employed, however, the additional output required cannot be provided by increasing production; it can only be provided through a reduction in the previous level of real expenditure.16 This reduction may be brought about either by a deliberate expenditure-reducing policy introduced along with the switch policy, or by the inflationary consequences of the switch policy itself in the assumed full-employment conditions.17
If the increased output is provided by a deliberate expenditure-reducing policy, the nature of this policy will obviously influence the effects of the expenditure-switching policy, since the composition of the output it releases may be more or less substitutable for foreign output in world demand. Thus, for example, an expenditure-reducing policy which reduces domestic demand for imports and exportable goods will be more favourable to expenditure-switching than one which reduces domestic demand for non-traded goods. The analysis of the effects of an expenditure-switching policy supported by an expenditure-reducing policy must therefore comprise the effects of the latter in determining the composition of the productive capacity available to meet the increased demand created by the former, as well as the elasticity relations which govern the effects of the interaction of increased demand with increased production capacity on the prices and volumes of goods traded.
If the expenditure-switching policy is not accompanied by an expenditure-reducing policy, its effect will be to create an inflationary excess of aggregate demand over supply, leading to price increases tending to counteract the policy’s expenditure-switching effects. Inflation, however, may work towards curing the deficit, through various effects tending to reduce the level of real expenditure from the full employment level of output. These effects, which are familiar and have been analysed in detail by Alexander, include the effect of high marginal tax rates in increasing the proportion of real income absorbed by taxation as wages and prices rise, the possibility of a swing to profits increasing the proportion of income saved, and the effect of rising prices in reducing the real purchasing power of cash and government bonds held by the public, so reducing their wealth and propensity to consume. All of these effects, it may be noted, depend on particular asymmetries in the reactions of the sectors affected to the redistributive effects of inflation on real income or wealth, which may not in fact be present. The important point, however, is that these factors, on which the success of an expenditure-switching policy depends in this case, are monetary factors, and that the analysis required employs monetary concepts rather than elasticity concepts. As in the previous case, the elasticity factors are subordinate to the factors governing the reduction in aggregate real expenditure, in determining the consequences of the expenditure-switching policy for the balance of payments.
The argument of the previous paragraph—that in full employment conditions the success of expenditure-switching policies depends mainly on the effectiveness of the consequent inflation in reducing real expenditure—helps to explain both the prevalence of scepticism about, and hostility towards, exchange rate adjustment as a means of curing balance-of-payments disequilibria, and the fact that historical experience can be adduced in support of the proposition that devaluation is a doubtful remedy. The argument does not, however, support the conclusion frequently drawn from the analysis of devaluation in these circumstances, that import restrictions are to be preferred; this is a non sequitur, since import restrictions are equally an expenditure-switching policy. Rather, the proper conclusion is that expenditure-switching policies are inappropriate to full employment conditions, except when used in conjunction with an expenditure-reducing policy as a means of correcting the employment-reducing effects of the latter.
But what of the choice between devaluation and selective trade controls, to which reference has just been made? So far, it has not been necessary to distinguish between them, since from the point of view of the balance of payments both can be treated as expenditure-switching policies. It is from the point of view of economic welfare that they differ; and the arguments on their relative merits have nothing to do with the state of the balance of payments, except that if controls are preferable a deficit may offer an opportunity for introducing them with less risk of foreign retaliation than if trade were balanced.
The welfare arguments for controls on a country’s international trade may be divided into two groups, those centring on controls as a means of influencing the internal distribution of real income, by discouraging imports consumed by the rich and encouraging those consumed by the poor, and those centring on controls as a means of increasing the country’s gains from trade through exploiting its monopoly/monopsony power in foreign markets. The former are of doubtful validity, both because the ethics of disguising a real income policy as a trade policy are suspect, and because both the efficiency and the effectiveness of trade controls as instruments for governing real income distribution are dubious. The latter are valid, to the extent that the country has powers to exploit the foreigner and can use them without provoking sufficient retaliation to nullify the gains.
This is the familiar optimum tariff argument. Its application to balance-of-payments policy depends on the level of trade restrictions already in force, as compared with the optimum level of restrictions.18 If an expenditure-switching policy is required to correct a deficit, and the level of trade restrictions is below the optimum, restriction19 is preferable to devaluation until the optimum level is reached; in the opposite case, devaluation is preferable. But it is the relation of actual to optimum restrictions, and not the state of the trade balance, which determines whether restriction is desirable or not.
This concludes the analysis of alternative policies for correcting a ‘flow’ balance-of-payments deficit on current account. To complete the analysis of ‘flow’ disequilibria, it would be necessary to relax the assumption that international capital movements are confined to reserve movements between foreign exchange authorities, and to consider alternative policies for correcting a deficit on current and capital account combined. The central problem in this case is to determine the level of current account surplus or deficit, capital export or import, at which economic policy should aim. This raises two further problems too difficult to pursue here: the optimum rate of accumulation of capital for the community as a whole, and the degree to which it is desirable to discriminate in favour of investment at home and against investment abroad.
In conclusion, the argument of this chapter may be summarised as follows: formulation of the balance of payments as the difference between aggregate receipts and payments, rather than receipts and payments on international account only, has two major advantages. It brings out the essentially monetary nature of a deficit, which must be accompanied by dishoarding of domestic money or credit creation; and it relates the deficit to the operation of the economy as a whole. A deficit may reflect a ‘stock’ decision or a ‘flow’ decision by the community. The conditions which make a ‘stock’ deficit a policy problem indicate the use of direct control methods as against price system methods of correction. Policies for dealing with ‘flow’ deficits on current account may be divided into expenditure-reducing and expenditure-switching policies; in full employment conditions the latter must be supported by the former, or rely on inflation for their effect, which in either case cannot be analysed adequately in terms of elasticities. When capital account transactions are introduced into the analysis, the choice between policy alternatives requires reference to growth considerations not readily susceptible to economic analysis.
1 Reprinted from H. G. Johnson, International Trade and Economic Growth (London, George Allen & Unwin, 1958), 153–68.
2 See in particular J. E. Meade, The Theory of International Economic Policy. Vol. I: The Balance of Payments (London, 1951); J. Tinbergen, On the Theory of Economic Policy (Amsterdam, 1952); S. Alexander, ‘The Effects of a Devaluation on a Trade Balance’, International Monetary Fund Staff Papers, II, no. 2, April 1952, 263–78; also G. Stuvel, The Exchange Stability Problem (Oxford, 1951); A. C. Harberger, ‘Currency Depreciation, Income, and the Balance of Trade’, Journal of Political Economy, LVIII, no. 1, February 1950, 47–60; S. Laursen and L. A. Metzler, ‘Flexible Exchange Rates and the Theory of Employment’, Review of Economics and Statistics, XXXII, no. 4, November 1950, 281–99; R. F. Harrod, ‘Currency Depreciation as an Anti-Inflationary Device: Comment’, Quarterly Journal of Economics, LXVI, no. 1, February 1952, 102–16. The terminology of ‘absorption’ was initiated by Alexander; Machlup’s criticisms of Alexander’s argument (F. Machlup, ‘The Analysis of Devaluation’, American Economic Review, XLV, no. 3, June 1955, 255–78), though valid in detail, miss the main point of Alexander’s contribution, a point obscured by Alexander’s own emphasis on the contrast between the ‘elasticity’ and the ‘absorption’ approaches to devaluation and his attack on the former. The later argument of this paper attempts a reconciliation of the two approaches in a broader framework of analysis.
3 Where the central bank or other monetary authority also holds the foreign exchange reserves, it is necessary for the purposes of this paper to separate its functions conceptually into two parts, and to class its transactions as monetary authority (including those with itself as exchange authority) among transactions of residents.
4 Where monetary authority and exchange authority are one and the same institution, domestic monetary liabilities may simply be extinguished by sales of foreign exchange.
5 Milton Friedman, ‘The Case for Flexible Exchange Rates’, 157–203 in Essays in Positive Economics (Chicago, 1953), especially 186, n. 11.
6 With the community defined to include the monetary authority, a substitution of securities for domestic money can only be effected by drawing securities from abroad in exchange for international reserves.
7 A temporary deficit of this kind must be distinguished from a deficit which is ‘temporary’ in the sense that the causal factors behind it will reverse themselves, leading to a later compensating surplus: e.g. a deficit due to a bad harvest.
8 The deficit involves the replacement of international reserves by stocks of exportable or importable goods and/or by holdings of internationally marketable securities, the change being motivated by private profit considerations. For this to constitute a deterioration from the national point of view, the alternatives facing private asset-holders must be assumed not to reflect true social alternative opportunities, or private asset-holders must be assumed to act less rationally than the economic policy authorities, or the national interest must be defined so as to exclude their welfare from counting. If any of these assumptions is valid, it indicates the need for a remedial policy, but not one conditional on the existence of a deficit or to be applied through the balance of payments. This point is argued more fully below, in connection with import restrictions.
9 Stocks are built up by withholding goods from export or by increasing imports; depreciation makes both of these less attractive. A third policy might be increased taxation, either of stocks or of home-market sales of goods.
10 This assumption, which is slightly inconsistent with the argument above concerning the monetary implications of a deficit, is made here to avoid the necessity of repeating the analysis for the case where limited reserves prevent the authorities from allowing a deficit to solve itself.
11 Whether this is so depends on the use to which the finance provided by the deficit is put, which involves comparison with what would have happened in the absence of the deficit. If the deficit finances additional investment in productive domestic capital or income-yielding foreign assets the net effect on the capital position may be favourable; if it finances additional consumption it is likely to be unfavourable, though even additional consumption may sometimes increase productive capacity.
12 Differentiating the equation in the text, we obtain dB = (1 − e)dY + dE, where e is the marginal propensity to spend out of income, dY is the total increase in output (including multiplier effects) and dE is the autonomous decrease in expenditure. If multiplier effects through foreign incomes are ignored,
where dA is an autonomous change in demand for domestic output and m is the proportion of marginal expenditure leaking into imports. Splitting dA into two components, dO for output-increasing policies and − hdE for expenditure-reducing policies (where h is the proportion of expenditure reduction falling on domestic output), gives the result
Hence either an output-increasing or an expenditure-reducing policy will improve the balance, so long as e is less than unity. (Alexander has argued that since e includes induced investment it may well exceed unity; this possibility is ignored in the argument of the text.) Expenditure reduction will in fact improve the balance so long as multiplier stability is present.
13 These arguments conflict with the assumption, more frequently made in connection with trade controls than with devaluation, that the public will consume less because it cannot obtain the goods it prefers as readily as before. That assumption may well be valid in the case of a policy expected to be applied for a short period only, after which goods will become as available as before, or in the analysis of the short run during which the economy is adjusting to the change in policy; but it is invalid in the present context of flow disequilibrium, since it overlooks the effect of the policy change in reducing the future value of savings and hence the incentive to save. An example of this type of faulty reasoning is the assertion sometimes made that quantitative import restriction is particularly effective in under-developed countries because their economic structure allows little possibility of substitution for imported goods in either production or consumption.
One qualification to the argument of the text, which also applies to the final sentence of the paragraph, is that if the goods towards which domestic expenditure is switched are more heavily taxed than those from which expenditure is diverted (a type of complication which is ignored in the general argument of the text), real expenditure may fall rather than rise.
14 Impact effects on the level of expenditure from a given income level of the type discussed in the next-but-one paragraph preceding this one are ignored in this paragraph.
15 Cf. E. V. Morgan, ‘The Theory of Flexible Exchange Rates’, American Economic Review, xlv, no. 3, June 1955, 279–95. Morgan’s statement (285) that instability requires ‘very strong and perverse income effects’ is fallacious—all that is strictly necessary is a preference in each country for home-produced goods.
16 Recognition of this point may be regarded as the fundamental contribution of the absorption approach, though none of the authors cited seems to have appreciated all its implications: Meade, for example, analyses the case on the assumption that an appropriate expenditure-reducing policy is in effect, without examining the interdependence between the two policies or the alternative of inflation, while Alexander does not recognise that the effects of inflation on absorption could be achieved by policy.
17 For analytical simplicity, the possibility of both increased production through ‘over-full employment’ and of direct expenditure-reducing effects of a switch policy (discussed earlier in this chapter) are ignored here.
18 See S. Alexander, ‘Devaluation versus Import Restriction as an Instrument for Improving Foreign Trade Balance’, International Monetary Fund Staff Papers, I, no. 3, April 1951, 379–96, for a lucid and pioneering exposition of this principle.
19 Generally, optimum trade restriction entails restriction of both imports and exports; but if the country’s currency is over-valued it may imply subsidisation of some or even all exports, and if the currency is under-valued it may imply subsidisation of some or even all imports. (These conclusions follow from the fact that a devaluation is equivalent to an all-round export subsidy and import duty.)