Published June 29, 1993
Almost all industrial countries today are suffering from permanent unemployment, and most are suffering from currency instability. The two problems are related: an expansive economic policy which leads to a currency crisis is usually intended to reduce unemployment. This report considers how both problems might be solved at the same time — and more specifically, the circumstances in which currency devaluation will be necessary, useful or ineffective.
Weak Currencies and Unemployment: Causes and Cures
Several months ago, we examined the reasons for the rash of devaluations within the European Monetary System (“The Lessons of Europe’s Money Muddle: The ‘Reserve Currency Curse’ Visits Europe,” LBMC Report, November 30, 1992). We showed that a chronically weak currency is essentially a monetary problem — an excess supply of money, which is reflected in an equal excess of payments over receipts for nonmonetary wealth. The typical cause is excessive expansion of domestic credit to finance government deficit spending. This causes a loss of official foreign exchange reserves, and sooner or later, a currency depreciation.
The cure is to sufficiently curb the expansion of domestic credit. There is nothing technically difficult about this. But practically speaking, it may require prohibiting the central bank from financing the government’s deficit spending, to assure the central bank’s de facto independence.
More recently, we considered the reasons for the sharp rise in permanent unemployment in industrial countries over the past two decades (“Transfer Payments: Cause of Permanent Unemployment,” LBMC Report, May 26, 1993). We showed that a rising burden of transfer payments (and minimum wage levels) has created a surplus of labor — in other words, unemployment — by putting an above-market floor under wages. The evidence indicated that this is the chief cause of the rise in permanent unemployment in industrial countries.
The only way to end this permanent unemployment is to restore flexibility in wages. This can be done in two ways. One is to cut the transfer payments or minimum wage in nominal terms, restoring downward flexiblity in money wages. The other is to engineer a one-time increase in the price level, while holding the minimum wage or transfer payments constant, thereby restoring flexibility in real wages.
If transfer payments are too high and indexed to wages or prices (or provided in kind), unemployment will remain no matter what happens to prices, because real wages will be inflexible.
A Question of Priorities
Which combination of policies to use, and in which sequence, depends on a country’s priorities.
A country whose highest priority is price stability (for a large country) or exchange-rate stability (for a small country) can achieve that goal simply by restricting domestic credit sufficiently. But then it must either accept the permanent unemployment caused by its transfer payments, or else cut transfer payments and the minimum wage in nominal terms to restore wage flexibility and reduce unemployment.
A country whose highest priority is reducing unemployment must either cut transfer payments in nominal terms while pursuing a noninflationary monetary policy, or else de-index transfer payments while engineering a one-time increase in the price level.
How this one-time increase in prices might be brought about depends on whether the country is large or small. In a closed economy, an easier monetary policy will affect the price level directly. But this analysis applies in practice only to a large country like the United States whose currency is widely used as official reserves by other countries.
A small country cannot affect the price level without changing its exchange rate, because it cannot affect the world price level. If a small country tried to run an expansive monetary policy while maintaining a fixed exchange rate, it would lose foreign exchange reserves until it was forced to depreciate anyway. A more sensible approach would be simply to re-fix the currency. (For a small country, refixing is better than floating: since financial markets adjust quickly and goods markets slowly, a floating exchange rate tends to “overshoot,” creating an additional adjustment burden.)
In short, whether devaluation will be necessary or effective in reducing unemployment depends on the degree to which wages are flexible. As one writer on the subject (W.M. Corden) succinctly put it: “To make devaluation necessary money-wages have to be inflexible, and to make it effective real wages have to be flexible.”
A devaluation will be both necessary and effective if wages are downwardly rigid in nominal terms — for example because of transfer payments or the minimum wage — but the wage floor is not indexed to prices. In this case a devaluation, by raising prices on a one-time basis, will make it profitable to hire previously unemployed workers. Thereafter, full employment can be maintained indefinitely — without inflation — as long as domestic credit remains in check and transfer payments are not raised again.
In all other cases, devaluation will be either unnecessary (if wages are flexible) or ineffective (if the wage floor is indexed to prices) in reducing permanent unemployment.
Real World Application
Despite its simplicity, the analysis presented above captures the essence of the economic policy problem facing most nations in today. (In an appendix we add some more complicated assumptions, but they do not change the basic conclusions.)
What practical consequences can we draw by applying the analysis to recent experience?
1. Several European currencies have been devalued recently. But only Italy simultaneously abolished widespread indexing — the “scala mobile.” Therefore casual observation suggests that, other things equal, Italy should benefit more in permanently reducing unemployment than other countries which devalued their currencies.
2. Of those European currencies which have not been devalued against the Deutschemark, only the Netherlands appears to have taken drastic steps to cut transfer payments in nominal terms. Therefore casual observation suggests that, other things equal, the Netherlands will have more success in reducing permanent unemployment than other countries which have not devalued against the Deutschemark.
3. Reserve-currency countries like the United States and Germany could reduce permanent unemployment either by cutting transfer payments in nominal terms, or by de-indexing them while domestic monetary policy permitted the price level to rise. (In Germany’s case this might require a decline of the Deutschemark — and all currencies tied to it — against the dollar.) But neither has done so. Therefore we should not expect any permanent reduction in average unemployment, though there may be cyclical variations.
Appendix: Nuts and Bolts of Devaluation
The model presented in the text captured the essential dynamics of monetary policy and permanent unemployment. But it contained a number of simplifying assumptions, and was silent on certain issues, such as the effect on trade or capital flows. We can adopt more complicated and realistic assumptions, though this does not change the main conclusions.
Tradeable and Nontradeable Goods
Our discussion implicitly assumed the “law of one price,” meaning that prices everywhere in the world are the same, adjusted for exchange rate differences. And we ignored any discussion of international trade or capital flows.
Let us suppose, more realistically, that each country produces two kinds of goods: tradeable and nontradeable. (The main difference is that the international transport costs of nontradeable goods are very high in relation to their value; one doesn’t fly from the U.S. to Germany to get a haircut, for example.) We will assume at first that the price of all tradeable goods, exportable and importable alike, is the same. The price of tradeable goods is the world price times the exchange rate. The price of nontradeable goods is determined by domestic supply and demand.
Let us also suppose, at first, that there are no international capital flows — no private lending or direct investment. In this case, a balance of payments deficit is the same as a trade (or to be more precise, a current account) deficit. A country with a payments deficit must be losing foreign exchange reserves to pay for an excess of imports over exports. This is sometimes described by saying that the country’s “absorption” — spending on goods for consumption and investment, both public and private — exceeds its income. (Income is determined, in turn, by national output.)
If the excess money which caused this deficit in payments comes from people reducing their cash balances (and thus purchasing other forms of wealth), the deficit will stop when their portfolio adjustment is complete. But if the deficit is due to continuous excessive credit expansion by the central bank, the deficit will continue. To cure the deficit, it is necessary to tighten domestic credit, thus reducing the excess of total payments over receipts. Part of the decline in spending will fall on tradeable goods, part on nontradeable goods.
Since the price of tradeable goods is set on the world market, reduced spending on tradeables causes a decline in imports and releases goods for export, without affecting the price of tradeables. But the price of nontradeable goods is set by supply and demand. Reduced spending on nontradeable goods at first causes unsold goods and incipient unemployment; what happens next depends critically on whether wages and prices are flexible.
The Crucial Role of Wage/Price Flexibility
If wages and prices are flexible, the price of nontradeables declines. This causes a shift in the relative price of tradeable and nontradeable goods. So people switch some of their demand from tradeables to nontradeables; and some labor and capital is switched from producing nontradeables to producing tradeable goods for export. The increase in demand for, and reduction in supply of, nontradeable goods restores full employment. Thus, with flexible wages and prices, the exchange rate does not have to change to restore payments balance and full employment.
But now let us suppose wages and prices are not flexible downwards — for example, because transfer payments or a minimum wage put a high nominal floor under wages. If wages or prices in nontradeable industry cannot fall in response to reduced demand, there is no shift in the relative price of tradeables and nontradeables. This means a larger reduction in spending is necessary to correct the payments deficit in tradeable goods. And in the face of the fall in spending, the downward wage/price rigidity causes unemployment in nontradeable industry.
In this case, a devaluation is necessary to bring about the “switching” of expenditure and productive resources that was automatically brought about by flexbile prices. The devaluation raises the price of tradeable goods, while the price of nontradeables does not rise, or does not rise to the same degree, in response to the decline in spending. Once again, shifts in supply and demand between tradeable and nontradeable goods are induced to restore full employment while payments balance is simultaneously restored.
But what if wages are not only rigid downward but also indexed to rises in the general level of prices? In this case wages are inflexible not just in nominal but in real terms. Automatic indexing raises the wage floor by the same amount as prices, so the unemployment remains. At best there may be a temporary reduction in unemployment if there are adjustment lags. But attempts to exploit a temporary improvement will lead to a state of chronic inflation with chronic unemployment.
In each case, two instruments are necessary to meet the two economic goals — payments balance and full employment. To use terms long familiar in this field, the payments deficit is corrected by expenditure reduction (relative to income); while full employment is maintained by expenditure “switching” — the response to the shift in relative prices of tradeables and nontradeables.
Misconceptions About Devaluation
Note that, with or without devaluation, a cut in absorption is still necessary to improve the balance of payments. However, a cut in real spending will usually be indirectly initiated by the devaluation. This is because, by raising the price level, the devaluation reduces the real value of the existing money supply. In order to rebuild their cash balances to the desired level, people must buy less or sell more.
As long as the central bank refrains from expanding domestic credit, this will improve the balance of payments and increase foreign exchange reserves on a one-time basis, until the domestic money supply expands by the desired amount. But if the central bank expands domestic credit, the money supply will increase to the desired amount without first improving the balance of payments; so the payments deficit will remain despite the devaluation. Nothing will have changed, except that the price of both tradeable and nontradeable goods will have risen in proportion to the devaluation.
A devaluation, then, will not improve the balance of payments,except insofar as it absorbs some of the excess money that caused the deficit in the first place. The devaluation is essentially a “switching device” intended to prevent unemployment resulting from the cut in spending that does improve the balance of payments. And whether the devaluation will be necessary or effective in this role depends on the behavior of wages and prices.
If wages and prices are flexible in nominal terms, no devaluation is necessary to maintain full employment. (In fact, there can be full employment even if the currency appreciates.) If wages and prices are downwardly rigid in nominal terms, a devaluation will be necessary to restore flexibility in real wages and restore full employment. But if wages and prices are downwardly rigid and indexed to rises in the price level, a devaluation will be ineffective in reducing unemployment permanently, since real wages are fixed.
International Capital Flows
When we take account of international lending and investment, the overall balance of payments means net payments on the trade and capital accounts combined. It is therefore no longer possible to predict (at least without more information) whether, say, the trade balance will improve in response to restrictive domestic credit combined with a devaluation. But we can be confident that the overall balance of payments will improve — either because of an improved trade balance, or else because of an inflow of capital as the country becomes a more attractive place to invest, or some combination. The conclusions of our simple model still apply to the overall balance of payments. But to the extent that there is a net capital inflow, a smaller cut in current absorption will be necessary to improve the overall balance of payments.
Reserve Currency Countries
Another simplifying assumption was that we are dealing with a small, “open” economy. In a closed economy, or for the world as a whole, there is no external exchange rate or external payments balance. The larger a country, the more closely it approximates a closed economy, and the greater the power of its economic policy to affect the world level of prices. A large country may also be able to affect its own “terms of trade” — the ratio of import to export prices. To the extent that the terms of trade change, the necesary change in absorption may be larger or smaller than if the terms of trade had remained constant. However, the small-country assumption is basically realistic for any country whose currency is not an official reserve currency.
The major practical exceptions are the United States (since the dollar is the world’s reserve currency), and to a lesser extent Germany (since the Deutschemark is Europe’s regional reserve currency). As we showed in our report on the EMS, reserve currency countries have little control over their exchange rates, which are essentially determined by other countries, and any increase in the other countries’ official foreign exchange reserves imposes a balance of payments deficit upon the reserve-currency country.
The idea — popular with the Clinton Administration — that a decline of the currency, by itself, would improve the U.S. trade balance is not supported by theory (nor, on the whole, by experience). Among other things, this view ignores the existence of international capital flows and the dollar’s reserve currency status.
The model’s other assumptions, such as the lack of economic growth or continuous inflation in world prices, are easily adaptable without changing the qualitative conclusions of the model. For example, if the world price level is continuously inflating, e.g., because the reserve currency country is causing inflation for the whole system, de-indexing transfer payments will lower unemployment even without a devaluation by other countries. But this may take many years, depending on the rate of inflation and the size of permanent unemployment.