Abstract
THE BEARING of exchange rate systems on the relative effectiveness of monetary policy on the one hand, and of budgetary policy on the other, as techniques for influencing the level of monetary demand for domestic output, is not always kept in mind when such systems are compared. In this paper it is shown that the expansionary effect of a given increase in money supply will always be greater if the country has a floating exchange rate than if it has a fixed rate. By contrast, it is uncertain whether the expansionary effect on the demand for domestic output of a given increase in budgetary expenditure or a given reduction in tax rates will be larger or smaller with a floating than with a fixed rate. In all but extreme cases, the stimulus to monetary demand arising from an increase in money supply will be greater, relative to that arising from an expansionary change in budgetary policy, with a floating than with a fixed rate of exchange.
THE BEARING of exchange rate systems on the relative effectiveness of monetary policy on the one hand, and of budgetary policy on the other, as techniques for influencing the level of monetary demand for domestic output, is not always kept in mind when such systems are compared. In this paper it is shown that the expansionary effect of a given increase in money supply will always be greater if the country has a floating exchange rate than if it has a fixed rate. By contrast, it is uncertain whether the expansionary effect on the demand for domestic output of a given increase in budgetary expenditure or a given reduction in tax rates will be larger or smaller with a floating than with a fixed rate. In all but extreme cases, the stimulus to monetary demand arising from an increase in money supply will be greater, relative to that arising from an expansionary change in budgetary policy, with a floating than with a fixed rate of exchange.
The Model
Let us assume a simple Keynesian model1 in which (a) taxation and private income after tax both vary directly with national income, (b) private expenditure (on consumption and investment) varies directly with income after taxation,2 and inversely with the interest rate, (c) the interest rate varies directly with the income-velocity of circulation of money (the ratio of national income to the stock of money), (d) the balance of trade (exports less imports of goods and services) varies inversely with domestic expenditure3 and directly with the domestic currency value of foreign exchange, and (e) the balance of payments on capital account varies directly with the rate of interest. All magnitudes are expressed in domestic wage units, and wages are assumed to remain constant in domestic currency. No account is taken of any changes in the propensity to spend from real income changes that result from changes in the terms of trade. No account is taken, initially, of the effect of exchange speculation on capital movements.
Effects of an Expansionary Shift in Budgetary Policy
Let us first compare the effects of an expansionary shift in budgetary policy brought about by an increase in public expenditure, without any change in tax rates, under (a) a fixed exchange rate system and (b) a floating exchange rate system, respectively. (A decline in taxation, resulting from a reduction in tax rates, would have effects on expenditure, income, and the balance of payments similar to, though less powerful than, those resulting from an equal increase in public expenditure. No essential feature of the ensuing analysis would be altered if it had been concerned with the former rather than the latter type of budgetary expansion.)
Under fixed exchange rates, an increase in public expenditure will give rise to an increase in income which will be associated—if the economy was previously underemployed—with increases in employment and output.4 The increase in expenditure will lead to a deterioration in the balance of payments on current account, owing, notably, to a rise in imports. The increase in expenditure and income will also enhance tax revenues, though not to such an extent as to equal the initial increase in public expenditure.5
In order to isolate the effect of a change in budgetary policy, it is necessary to assume that monetary policy remains, in some sense, unchanged. In this paper, that is taken to mean that the stock of money is held constant.6 To keep the money stock constant while the increase in government expenditure is pushing up incomes will necessitate economy in the use of money which is possible only if the interest rate is raised or allowed to rise. The rise in interest in turn will result in (a) a check to the increase in expenditure and income, though some increase will remain,7 and (b) a favorable shift in the balance of payments on capital account, i.e., a decline in capital exports and/or an increase in capital imports.
Since the increase in public expenditure provokes an unfavorable shift in the current balance and a favorable shift in the capital balance, it is uncertain whether the balance of payments as a whole will deteriorate or improve. It is the more likely to deteriorate, and the less likely to improve, the higher is the marginal propensity to import and the greater the adverse effect on the value of exports as domestic expenditure increases, the less sensitive is the rate of interest to changes in money income and hence in the velocity of circulation, and the less sensitive are capital movements to changes in the rate of interest.8
To the extent that the increase in public expenditure gives rise to an improvement or a deterioration, respectively, in the balance of payments, the maintenance of a constant stock of money will call for a decline or an increase, respectively, in the rate of expansion of bank credit. More important is the fact that, if the policy of budgetary expansion results in a deterioration of the balance of payments, shortage of reserves may ultimately compel the authorities to abandon the policy and to renounce the associated expansion in income and employment.9
Suppose, now, that the increase in public expenditure takes place in a country where the balance of payments is kept in equilibrium through exchange rate adjustments. Then, if the parameters of our model—notably the sensitivity of capital movements to changes in the rate of interest—are such that a rise in public expenditure would have resulted, with a fixed exchange rate system, in a deterioration in the balance of payments, it will result, with a floating rate system, in a depreciation of the exchange rate, which will bring about a partial restoration of the trade balance. (This restoration will, in general, be only partial since some net deterioration of the trade balance, compared with the situation before the rise in public expenditure, must remain to offset the improvement in the capital balance.) To the extent that the current balance is restored, there will be an increase—over and above that discussed above—in expenditure, income, and output. In other words, the stimulus to income, output, and employment resulting from a given increase in public expenditure will be greater with a floating exchange rate than with a fixed exchange rate.10 If capital movements were entirely insensitive to the rise in the rate of interest, the exchange rate would depreciate to whatever extent was necessary completely to restore the trade balance, and the stimulus to income and output would be of the same order as would have occurred in a closed economy.
On the other hand, if a rise in public expenditure would, with a fixed exchange rate, have effected an improvement in the balance of payments, it will, with a floating rate, lead to an exchange appreciation; and, to the extent that appreciation intensifies the deterioration in the trade balance, the net stimulus to income, output, and employment will be less than in an open economy with a fixed rate.11 At first sight, the case in which a rise in government expenditure produces an exchange appreciation would appear to be an academic curiosum without practical significance. However, as is shown in a paper prepared by Mr. R. R. Rhomberg, expounding an econometric model of the Canadian economy, the responsiveness of international capital movements to changes in interest rates, and the responsiveness of interest rates to changes in money national income, have probably been sufficiently great in that country over a large part of the postwar period, relative to the marginal propensity to import, for a rise in government expenditure at a constant money stock to have tended to produce just such a result.
It is of interest to note that, if the flow of capital between the country and the outside world were infinitely elastic with respect to the interest rate, the appreciation of the exchange rate resulting from the inflow of capital would bring about a net deterioration in the current balance of payments large enough to offset completely the stimulating effect of the budget deterioration on national income. National income would not increase at all, and the interest rate would remain at the original level.12
Effects of an Increase in the Stock of Money
Now, let us compare the effects on income, output, and employment of increasing the stock of money (a) with fixed exchange rates and (b) with floating exchange rates, respectively.
An increase in the stock of money will entail a decline in the velocity of circulation and lead to a reduction in the rate of interest which will stimulate an increase in private expenditure on investment and consumption, both directly and via the Keynesian multiplier. The rise in expenditure will be associated, as before, with a (smaller) increase in income and output13 and a deterioration in the balance of payments on current account.14 The rise in income will moderate the decline in the rate of interest but not to the point of eliminating it; otherwise, neither investment nor income could increase.15 Since the monetary expansion, even after the rise in expenditure and income, lowers the interest rate, some deterioration will tend to occur in the balance of payments on capital account. In the case of a monetary expansion, therefore, by contrast with that of an increase in public expenditure, a deterioration in the balance of payments as a whole is bound to occur in all circumstances. It follows that the monetary expansion, and the associated expansion of income and output, could only be sustained indefinitely to the extent that in their absence the balance of payments would have been favorable.
It is easy to see that a monetary expansion must always exercise a more powerful effect on income and output when there is a freely floating rate of exchange than when the exchange rate is fixed. The initial tendency toward an adverse shift in the balance of payments will cause a depreciation of the exchange rate to whatever extent may be necessary to keep external transactions as a whole in balance. The favorable influence of the exchange depreciation on the trade balance must come to outweigh the adverse influence of the increase in income to whatever extent may be necessary to produce a net improvement in that balance equal to the deterioration in the capital balance. The stimulus afforded by the depreciation to the trade balance will also act, both directly and via the multiplier, as a stimulus to income, raising it above the level which would have prevailed with a fixed exchange rate.16
The expansive effect of a given increase in the stock of money under the floating exchange rate system will be the greater, the greater the responsiveness of the international capital flow to movements in the rate of interest. If there were no responsiveness whatever, the exchange rate would depreciate to the point at which, despite the monetary expansion, no change occurred in the current balance of payments. Income would expand to the same extent as in a closed economy. On the other hand, if the capital flow were infinitely elastic with respect to the interest rate, the exchange rate would depreciate to the point at which the balance of trade became so favorable, and income increased so much, that the rate of interest remained at its original level. This implies that money income would increase by the same percentage as the stock of money.17
Relative Effects of the Two Kinds of Financial Policy
It remains to show that the effect on income and output of a given monetary expansion relative to that of a given budgetary expansion will never be less, and will generally be greater, under a floating exchange rate than under a fixed rate, even where budgetary expansion has a tendency to cause a depreciation of the exchange value of the currency.18 The simplest way to demonstrate this is to compare an increase in the monetary stock (Policy A) and an increase in public expenditure (Policy B) such that, under a fixed exchange rate, the two policies have equal effects in the aggregate on income, output, and employment, and to show that, under a floating rate, the effect of Policy A will never be less, and will in general be greater, than that of Policy B.
Since we have supposed that under a fixed exchange rate the two policies have the same aggregate effect on income and output, they will bring about approximately the same adverse shift in the balance of trade.19 Since, with incomes the same under the two policies, the money stock will be greater and the velocity of circulation less under Policy A than under Policy B, the rate of interest will be less under the former than under the latter policy. If capital movements were totally insensitive to changes in the interest rate, the two policies would, under a fixed exchange rate, have the same effects on the balance of payments as a whole; and under a floating rate, they would require an equal exchange depreciation to restore external equilibrium. The consequent restoration of the trade balance and the associated further stimulus to income would be the same for the two policies. However, if capital movements respond in any degree to interest changes, the two policies will have different effects. Since Policy A reduces, and Policy B raises, the rate of interest, Policy A under a fixed exchange rate will occasion a more unfavorable capital balance than Policy B. It follows that under a floating rate, Policy A will require, to restore payments equilibrium, a deeper exchange depreciation, and will consequently bring about a greater improvement in the trade balance, and a greater stimulus to income and output, than Policy B.20 The superiority of Policy A over Policy B as a means of increasing income and output depends notably, as we have seen, on the sensitivity of international capital movements to changes in the rate of interest. At zero sensitivity, there is nothing to choose between the two policies. If the sensitivity is infinite, the level of income resulting from Policy A will exceed that resulting from Policy B in much the same proportion as the money stock under A exceeds that under B.
The nature of the exchange regime has an important bearing not only on the relative effectiveness in influencing income and output of the two types of financial policy—monetary policy and budgetary policy—but also on their relative practicability or sustainability. Thus, under a fixed exchange rate—except to the extent that the external accounts were originally in surplus—monetary expansion can be sustained only as long as reserves hold out, while budgetary expansion, if capital movements are sufficiently sensitive to interest rates, may be sustained indefinitely.21 Under a floating exchange rate, on the other hand, not only is monetary expansion, while it lasts, likely to generate more additional income than budgetary expansion, relative to what would happen under a fixed exchange rate, but both types of policies can be sustained indefinitely, so far as the balance of payments situation is concerned.
The Exchange Speculative Element in Capital Movements
The foregoing argument has generally assumed the absence of exchange speculation. Under a floating exchange rate, the influence on exchange speculation varies according to whether it is equilibrating or disequilibrating. If it is equilibrating—as was generally the case, for example, in Canada in the 1950’s—it will tend to mitigate the exchange rate variations resulting from variations in internal financial policy, whether that policy is budgetary or monetary in character. However, since the greater relative effectiveness which a floating rate gives to monetary policy, compared with budgetary policy, is attributable to the stronger influence that the former exercises on exchange rates, it is to be expected that equilibrating speculation, by damping down exchange rate effects, will tend to reduce the difference in effectiveness between the two kinds of policy. Disequilibrating speculation on the other hand, by exaggerating exchange rate variations, tends to accentuate this difference in effectiveness.22
APPENDIX
Politiques financières intérieures avec un système de taux de change fixe et avec un système de taux de change fluctuant
Résumé
Les comparaisons entre les systèmes de taux de change ne tiennent pas toujours compte des incidences de ces systèmes sur l’efficacité comparée de la politique monétaire d’une part, et de la politique budgétaire d’autre part, en tant que moyens techniques d’influencer le niveau de la demande monétaire de la production intérieure. Le présent article montre que l’expansion résultant d’une augmentation donnée de la masse monétaire sera toujours plus forte si le pays a un taux de change fluctuant que s’il a un taux fixe. Au contraire, il n’est pas certain que l’accroissement de la demande de production intérieure résultant d’une augmentation donnée des dépenses budgétaires ou d’une réduction donnée des taux d’imposition soit plus fort ou plus faible avec un taux fluctuant qu’avec un taux fixe. A l’exception des cas extrêmes, l’influence stimulante exercée par l’augmentation de la masse monétaire sera toujours plus grande, en comparaison avec celle qui résulte d’une politique budgétaire à caractère expansionniste, lorsque le taux de change est fluctuant que lorsqu’il est fixe.
Política financiera interna bajo sistemas de tipos de cambio fijos o de tipos de cambio fluctuantes
Resumen
Al establecer una comparación entre los sistemas cambiarios, no siempre se toma en consideración la influencia que éstos ejercen sobre la eficacia relativa de la política monetaria y la política presupuestaria como medios de influir en el nivel de la demanda monetaria de la producción interna. En este estudio se demuestra que cuando un país emplea un tipo de cambio fluctuante, el efecto expansionista de un aumento dado del medio circulante será siempre mayor que cuando emplea un tipo de cambio fijo. Por otra parte, no se sabe si el efecto expansionista que ejercería sobre la demanda de la producción interna un aumento dado de los gastos presupuestarios, o una disminución dada de las tasas tributarias, habría de ser mayor o menor con un tipo de cambio fluctuante que con un tipo de cambio fijo. Salvo en casos extremos, cuando el tipo de cambio es fluctuante, el estímulo que ejerce sobre la demanda monetaria un aumento del medio circulante será siempre mayor, en relación con el que provoca una modificación de carácter expansionista de la política presupuestaria, que cuando el tipo de cambio es fijo.
Mr. Fleming, Advisor in the Department of Research and Statistics, is a graduate of Edinburgh University. He was formerly a member of the League of Nations Secretariat, Deputy-Director of the Economic Section of the U.K. Cabinet Offices, U.K. representative on the Economic and Employment Commission of the United Nations, and Visiting Professor of Economics at Columbia University. He is the author of numerous articles in economic journals.
It is assumed that the private marginal propensity to spend will always be less than unity with respect to income before tax.
It is assumed that the marginal propensity for the balance of trade to decline as expenditure increases is less than unity.
Since the marginal propensity to spend out of income is less than unity and since a fraction of each round of expenditure leaks abroad in additional net imports, the increase in income and expenditure will be limited, though possibly large. See Appendix, paragraphs 3 and 4.
The rise in tax revenue could exceed the initial rise in government expenditure only if the marginal propensity to spend out of private income after tax were substantially greater than unity. See Appendix, paragraph 5.
The only clear-cut alternative would appear to be that of defining constancy of monetary policy as the maintenance of a constant rate of interest. In “Flexible Exchange Rates and Employment Policy,” Canadian Journal of Economics and Political Science (November 1961), Mr. R. A. Mundell has compared the effects of monetary policy (defined as interest policy), fiscal policy, and commercial policy in a flexible exchange rate system and a fixed exchange rate system, respectively.
It is uncertain whether private expenditure, stimulated by the rise in income and depressed by the rise in interest, will increase or decrease. But expenditure as a whole, like income, will increase, except where income velocity is entirely inelastic. See Appendix, paragraphs 6 and 7. In this extreme case, not only expenditure but also income and the balance of trade will remain unchanged.
It is assumed not only that the exchange rate will remain fixed but that there will be no resort to restrictions on international transactions.
Ibid.
See Appendix, paragraph 11.
A high sensitivity of the interest rate to changes in velocity of circulation, i.e., a low elasticity of velocity with respect to the interest rate, while it makes for a favorable balance of payments response to government spending, and while it therefore tends to make the income response smaller under floating than under fixed exchange rates, also tends to reduce the magnitude of that response under both exchange systems. If the velocity of circulation were completely inelastic, a change in government expenditure would have no net effect on income under either exchange system.
See Appendix, paragraphs 15 and 16. However, in the extreme case where velocity of circulation is completely inelastic, money income will rise proportionately to the money stock under either exchange system. See Appendix, paragraph 17.
To put the same thing in other words, the effect under a floating rate relative to the effect under a fixed rate will never be greater, and will generally be less, in the case of budgetary expansion than in the case of monetary expansion.
We have to neglect, as unknown, any effects on trade of the difference in the composition of expenditure under the two policies.
It should be noted, however, that the responsiveness of capital movements to interest rate changes is made up of two components: a relocation of existing capital and a shift in the location of the placement of new savings. Since the former component is nonrecurrent and the latter recurrent in character, it is probable that the sensitivity of capital movements to interest changes will be greater in the short run than in the long run. Consequently, the difference between the two policies with respect to effectiveness and sustainability is also likely to be less in the long run than in the short.
Exchange speculation has a bearing not only on the relative effectiveness, but also on the practicability and sustainability, of the two policies. Under exchange rates that are fixed and are expected to remain so, exchange speculation would be absent. But if confidence in the fixed rate were less than complete, the fear of arousing disequilibrating movements of capital would tend to limit the magnitude and duration of the expansionary financial policies, particularly of monetary policy, the effect of which on the balance of payments is in any case the more adverse than that of budgetary policy.