ON THE INDEPENDENCE OF THE CENTRAL BANK REFLECTIONS ON POST WAR US EXPERIENCE

Annamaria Simonazzi
University of Rome

1. Introduction

The cheap money policy applied in the United States during the Second World War, the policy of stabilization of interest rates in the '40's and the subsequent "liberation" of monetary policy (i.e., the return to flexible interest rates) had mainly been accounted for in terms of a conflict between the two institutions, the Treasury and the Federal Reserve, and the alternating prevalence of one over the other.1 According to this interpretation, the experiment went on too long and thus fueled post-war inflation: with the Federal Reserve committed to supporting government securities, monetary policy lost control of liquidity, thus severely impairing its effectiveness in the control of the cycle. With nominal interest rates stabilized at an exceptional low in the history of the American economy and a high, variable rate of inflation, the actual result was negative real rates.

Recent studies have re-examined these developments in the light of the new classical macroeconomics. The policy of interest rate stabilization pursued by the Federal Reserve from 1942 to 1951 is reinterpreted as a credible commitment to a monetary rule, i.e., one that, by stabilizing long-term inflationary expectations, would allow for the maintenance of nominal interest rates.2 In their attempt to reconcile the data with the postulates of their theory, these studies have come up against two problems, i.e., the reasons that led the Federal Reserve to relinquish its independence in pursuit of a policy of interest rate stabilization, and the reasons that induced rational agents to accept negative real interest rates.

The paper takes issue with both the conflict hypothesis and the credibility hypothesis proposed by the new macroeconomics. It is argued that the policy of interest rate stabilization stood as a response to the risk of financial instability. As long as this was perceived as a real risk, the Federal Reserve faithfully collaborated with the Treasury. We consider next the concepts of credibility and monetary independence in the new macroeconomics and in the specific experience of the United States. The reappraisal of the debate of the '40's reveals a more complex view of the working of the economic system than that assumed by the new classical macroeconomics. Thus, we argue that the arguments in favour of the NMF concept of credibility as a principle underlying economic policy lose much of their force.

2. Expectations and financial stability

The United States came out of the war with a huge public debt, accounting for almost 60% of total financial liabilities, largely in the hands of the banks and representing over half their assets. In this situation, the main concern of both Federal Reserve and Treasury was how to guarantee that the financial assets in possession of the economy could be held onto. We can distinguish two reasons behind a liquidation of securities: speculative reasons, whereby fears of increases in interest rates lead to an increase in liquidity preference, and the need to finance expenditure. We argue that the period of interest rate stabilization can be roughly distinguished in two sub-periods: up to 1947, when financial stability questions prevailed; and from 1948 to 1951, when the control of liquidity in order to control expenditure, and perhaps inflation, prevailed.

In the immediate aftermath of the war, there was no obvious anchor to stabilize expectations. It was feared that this could result in extrapolative expectations: any factor creating uncertainty would have destabilized the market and lead to financial crisis. The stabilization of the structure of interest rates was considered, by both the Treasury and the Federal Reserve, as the only policy to provide an anchor to expectations and, by reducing uncertainty, stabilize the market.

With interest rates frozen, control of liquidity was assigned to direct controls3, and economic stability to the integration of monetary policy with a flexible fiscal policy, reinforced at critical times by direct controls over prices, wages, and supplies.

The actual result of the Federal Reserve's commitment to stabilizing interest rates was to transform the Treasury Bills into interest bearing reserves. At the same time, by eliminating the risk of capital losses on long-term bonds, it prompted a steady shift in portfolios towards these bonds with the Federal Reserve System banks buying up the Treasury bills.4 The control over the creation of these "reserves" was in the hands of the Treasury, which had shifted its deficit financing strategy towards short-term bonds. The Federal Reserve began to worry that the continued sale of Treasury bills to Reserve banks, associated with banks' purchases of bonds from nonbank sources, could lead to an indefinite expansion of reserves and deposits.

3. The unpegging of the short term rates

In July 1947, after more than one year of negotiations with the Treasury, the Federal Reserve abolished the buying rate for Treasury bills. We argue that this decision can be considered as the divide from the regime of interest rate stabilization and the new regime of flexibility. Certainly, this is the way in which the financial community interpreted the move. Despite continued reassurance by Federal officials that support for long term bonds would continue, the banks started to switch their portfolios from long to short term bonds.5 While the decision to free Treasury bill rates was taken owing to concern over continued inflation, the choice to free only the short term rates can be explained by the persisting concern over banks' losses rather than by opposition from the Treasury. A gradual dismantling of the support program was, at that time, in the interest of the whole financial community. The rapid follow up in the liberalization of other shorter maturities, however, gave the impression that it was only a question of time. We conclude that the end of the support program had already been discounted by the banking community since 1947.

For the time being, there was the problem of how to combine a flexible short term rate with a ceiling on long term rate. Although the experience of the thirties suggested that there was little osmosis between the short-term rates and the long-term rates, there was general acknowledgment that variations in the short-term rates could have significant psychological effects. This may account for the proposals for special reserve requirements advanced by the Federal Reserve, but fiercely opposed by the banks and eventually rejected by Congress. Supporters of selective controls argued that they did not substitute a flexible monetary policy, but were in fact a necessary condition for enhanced flexibility (Musgrave, 1948). Critics argued that segmentation of the market was impossible, that the rate structure was close-knit, and that stabilizing a part thus meant stabilizing the whole, no compromise between stability and flexibility being possible.

4. The debate on the efficacy of monetary policy

Between 1946 and 1950, the growth in nominal income and private liabilities greatly reduced the weight of government securities in total assets. Concern over the instability of the financial market was correspondingly reduced. While there was unanimous acknowledgment that the policy of interest rate stabilization implied a loss of control over money supply, it was not taken for granted, even at the Federal Reserve, that interest rate flexibility was a sufficient condition for control over monetary aggregates and that control over money supply meant guarantee of control over credit, expenditure and prices.

One result of the then dominant Keynesian approach had been to shift the emphasis from the demand for funds to the availability of credit. The Roosa Doctrine represents an ingenious attempt to reconcile different approaches: provide the needed nexus between short run interest rates flexibility and expenditure through effects on the supply of funds; reconcile flexibility on the one hand of the term structure with stability on the other extreme; reconcile flexibility of interest rates with stability of demand for securities. It did so by incorporating several institutional features of the financial markets:

a) the ownership distribution of government securities biased towards the speculative part of the market;

b) banks' dislike for capital loss realization and their consequent prudential behavior, matured in a context in which interest rate stabilization had produced static (or regressive) expectations;

c) mark up pricing and credit rationing due to a behavior based on customer relationship;

d) financial market imperfections making for little osmosis across interest rates.

Thus, the flexibility of short-term interest rates, by creating uncertainty, froze demand of securities (the "lock-in effect"), affected the supply of funds, and hence the financing of expenditure.

A common criticism of the lock-in effect was that it relied upon static expectations, themselves the result of the support program. Once the stabilization policy was gone, expectations would adapt to the new uncertain situation and the lock-in effect would no longer hold. But as Schlesinger (1960) pointed out, the debate raised by the Roosa Doctrine is a good illustration of the changed approach to the public debt - no longer seen as a threat to financial stability but as a lubricating element in the transmission mechanism. The existence of a highly developed government security market transformed the money stocks of the non-financial sectors into a pool that all could draw upon to finance expenditure. And it was on the effectiveness of monetary policy in these conditions that opinions differed. One line of argument, by rehabilitating the working of the transmission mechanism, restored the interest rate to its role of credit allocation. The other challenged this view observing that, with unlimited supply of reserves, money creation would be determined by demand. It was precisely disquiet over these points that inspired proposals designed to control the composition of banks' portfolios in order to prevent a rise in expenditure financed by an activation of the money stock. Criticism of proposals of direct controls stressed the ineffectiveness of any strategy designed to segment the market while the price support policy was still in operation.

Thus we come to the liberation of monetary policy in 1951 without any of the open questions in the debate on the efficacy of flexible interest rates being finally settled in favor of the monetary policy. As Tobin (1953: 461) pointed out, "neither the initial skepticism about money nor its recent rediscovery has been solidly grounded on empirical evidence". The outbreak of the Korean War in June 1950 fueled inflationary trends and provided the opportunity for a show-down between the Treasury and the Federal Reserve. With the "accord" of March 1951, support for long-term securities was withdrawn and the long-term interest rate was allowed to increase. With fears of financial instability gone, and interest rates free to float, the Federal Reserve abandoned the Treasury, ceased to sponsor requests for supplementary reserves and in fact opposed all further proposals for direct control.

5. Modern Interpretations

Recent studies have re-examined these developments in the light of the new classical macroeconomics. The policy of rate stabilization is taken by some to represent a credible commitment to a future squeeze (Toma, 1985 and 1991; Hutchinson and Toma, 1991) and by others the commitment to maintain an implicit price-level target zone (Eichengreen and Garber, 1990), which "effectively decoupled inflation from inflationary expectations". The hypothesis of credible commitment can be squared with non-conflictual interpretations of relations between the Central bank and the Treasury. Extending the bureaucracy theory to the monetary authorities, Toma (1985) thus answers the conflicting targets theory with a model based on collusion between the Federal Reserve and the Treasury in order to maximize profits accruing from seigniorage, while Eichengreen and Garber (1990) stress the financial stability objective to account for the commitment to interest rate stabilization.

Various factors have been put forward to justify the credibility of commitment to future deflation, e.g., the reputation built up by the United States (and the United Kingdom) in the period before the Second World War, or the role played by the return to the Gold Standard as an anchor for long-term expectations. Whatever the reason, it is more often assumed to have been the one condition compatible with the premises of the model and the hypothesis of rational agents. The presumed neutrality of money implies the existence of an ex ante (or expected) real interest rate determined exogenously (or at any rate by real factors alone), which guides the decisions of economic agents. If expectations are rational, then low nominal interest rates and exogenously determined real interest rates are only compatible when expected inflation rates are negative, and the expected rates of monetary expansion are therefore also small or negative.

Two lines of criticism have been advanced against the credibility model based on the new macroeconomics. A general criticism is that it consistently adopts a rather ad hoc macroeconomnic model and preference technology. Agents can be aggregated into a monolithic private sector, i.e., they have the same expectations and make their decisions on the basis of the same true model of the economy. This, in turn, implies that the technological constraints, i.e. "the feasibility and controllability of policy instruments; the feasibility of policy makers' objectives; and the accuracy and relevance of the economic theory that policy makers use" are all satisfied (Blackburn and Christensen, 1989).

A more radical criticism comes from the (old and more recent) interpretations based on widely divergent approaches to the whole question of uncertainty, and thus the formation of expectations, the behavior of economic agents, the influence exerted by the institutional context and the modus operandi of monetary policy. In a world of true uncertainty no information regarding future prospects exists today, the future is not calculable, money not a "mere counter" but performs the role of reducing uncertainty. This makes for the impossibility of a macro-analysis in real terms. Thus, it casts doubt on the plausibility of taking real interest rates as an objective variable deriving from the process of intertemporal optimization pursued by the economic agents. A model accounting for real interest rates as the ex post result of two partly independent variables - the nominal interest rate and the rate of inflation - would appear more realistic. In a world of uncertainty, the "credibility" of a policy has to do with persuasion, more than commitment and operates to reduce uncertainty over an unknown future, rather than thorough commitment to abstain from interfering with a deterministic and already known model.

In the immediate aftermath of the war the Keynesian message was taken mainly in the form of the possibility to manipulate policy variables, rather than beliefs, thus emphasizing the first rather than the second, more radical line of criticism. As we have shown, taking the non-neutrality of money seriously, the debate in the forties focused upon the modus operandi of monetary policy, the possibility to control money supply, liquidity and interest rates. It was based on theoretical premises very different from the one proposed by the new macroeconomics, where it is assumed that the central bank can choose the rate of inflation. Consideration of the second line of criticism will only strengthen our conclusions.

6. Conclusion

The US. experience showed the difficulty of stabilizing interest rates without losing control of credit, given the liquidity conditions of the post-war period, and the reluctance to rely on direct control over the banks' assets. The aim of pegging the interest rate was to maintain confidence in the securities market, by stabilizing expectations on long-term nominal interest rates, thus stabilizing the demand for securities and reducing the burden of the debt. The return to an independent monetary policy, though, left open the question of whether a flexible monetary policy could be a powerful tool: objections advanced on the basis of the availability of credit and the endogenous nature of money remained unanswered.

During the period of interest stabilization, the pressure of demand combined with specific circumstances (abrogation of direct control on prices, the Korean War) to determine a positive trend of inflation. The real interest rate thus turned negative. The explanation offered in this paper points to institutional and economic factors contributing to sustain the demand for government securities: the ownership distribution of the public debt, mainly in the hands of the banks and institutional investors, the lack of alternatives for private investors, the absence of an external constraint and a different conception of credibility that guided the behavior of economic agents and shaped expectations. According to our interpretation, where expectations are affected by the institutional and economic context, the outcomes of the interest rate stabilization policy may have been affected by the predominance of a theoretical model still stressing the possibility for the economic authority to control the economic variables.

1 Cf. for example Friedman and Schwartz (1963); Stein (1969) and Cagan (1980).

2 Toma (1985, 1991); Hutchinson and Toma(1991); Eichengreen and Garber (1990).

3 As early as 1940 the Federal Reserve had faced the problem of designing measures to freeze part of the securities on the market, applying to Congress for authorization to increase the reserve coefficient. It brought up the problem again in 1945 and in 1948.

4 From 1942 to 1946 the proportion of Treasury bills held by the banking system declined from 66 to 7%.

5 The share of long-term (over 5 years) securities fell from over half (53.5%) the bank portfolios, as the situation still stood at the beginning of 1947, to 38.7% in 1948 and 22.1% in 1949, with the Federal Reserve still engaged in compensative operations.