Introduction:

The chief task of any modern Central Bank is to keep inflation under control; or, more properly, to keep as constant as possible the value of the monetary unit. Whatever the monetary policy technique used by the Central Bank, the consensus in the modern empirical literature on the longrun relationship between money, prices and output is clear. Money growth and inflation essentially display a correlation of 1; the correlation between money growth or inflation and real output growth is probably close to zero, although it may be slightly positive at low inflation rates and negative at high rates. Ever since Bodin, Cantillon and Hume, the quantity theory of money has never argued anything else. Policy techniques may differ, but the gospel of stable price-levels is rock solid. For the past fifty years, the dominant tradition in monetary theory and policy has been wrestling with this very same issue. In the United States, after nearly three decades of interest rate targeting, monetarists won briefly the political battle in the 1970s and early 1980s to finally yield most of the ground to inflation targeting in the 1990s. Even if the bedrock of German, and subsequently European, monetary policy has always been, and still is, under the spell of money growth targeting, the European Central Bank is clearly not immune to the winds blowing from Washington and London. Faced with such an overwhelming consensus, one could reasonably expect economic theorists to provide a very strong case to support such generally accepted anti-inflationary policies. As any schoolboy knows, this is not the case. It might not be the best kept secret in economic theory, but it is clearly one of the most massive gap in our understanding of the functioning of a monetary economy. In a nutshell, though Central Banks keep conducting anti-inflationary policy aimed at stabilising the value of money, none of our sophisticated models can provide a convincing demonstration of why, in the first place, money displays a positive value.

This paper intends to discuss some very broad aspects of a long story dating back to Walras and his general equilibrium model. With the help of a very large brush, this paper reflects on the logic of the successive attempts at providing a proper theory of the price of money similar to ? or even better, integrated with ? the theory of relative prices. Without delving into the technical apparatus of modern monetary theory, this paper tries to link some recent literature with the logic inaugurated by Walras and brought to fruition by Arrow, Debreu and Hahn. The main conclusion is that, despite 125 years of strenuous intellectual efforts, economists have not been able to provide a satisfactory solution to this central question. The best-developed modern Walrasian general equilibrium models cannot find room for money. The conceptual incompatibility of money with general equilibrium appears to lie in the structure of the Walrasian markets: in frictionless markets, there is no clearly defined role for money, i.e. the implicit technology of exchange renders it useless. It has been shown elsewhere (Bridel 1997), that, from its inception as a model of exchange between commodities, the very logic of general equilibrium theory barred money to play an essential role as a social institution allowing monetary exchanges among individuals. In particular, the purely static nature of Walras?s pure economics, and the simultaneous lack of a proper technology of exchange in which money as a means of exchange has an essential role to play, exclude ex definitione the integration of monetary and value theory.

All modern micro-foundations for the demand for money are thus inherently that of a non-Walrasian economy: money derives a positive value only from an attribute exogenous to general equilibrium theory. The logic behind recent non-Walrasian models attempting to provide a rationale to a positively valued fiat money is shown to rely purely on exogeneous ad hoc qualities attributed to money (qualities other goods do not have). Money is introduced to account for co-ordination failures foreign to the very nature of Walrasian equilibrium theory. Hence, the conceptual incompatibility of money with general equilibrium theory lies in the structure of Walrasian markets. In such frictionless markets, there cannot be, ex definitione, a clearly defined role for exchanges which call for the use of money. The central problem is clearly to question the ability of the results of general equilibrium theory to withstand the introduction of such rigidities which legitimate the presence of money.

Part One examines the pure logic of Walras?s initial monetary model. After briefly recalling the neo-classical indeterminacy of the price-level and Patinkin?s attempt at solving this issue with his real-balance effect, the bulk of Part Two examines the broad rationale behind six modern non-Walrasian attempts at specifying a positive demand for money which is necessary if, in equilibrium, money is to have a positive value. Some concluding remarks reflect on Walras vs non-Walrasian attempts at explaining money in a general equilibrium framework.