The problem of asset price bubbles, and more generally of instability in the financial system, has been a matter of concern since the 1980s but has only recently moved to the center of the macroeconomic policy debate. Among the events contributing to concern about asset prices, the most notable have been the boom and bust in share prices, particularly those of technology stocks such as "dotcoms," and the subsequent boom in property prices. These asset price movements have been experienced, to varying degrees, in most OECD economies, including the United States, the United Kingdom, and (in the case of housing) Australia. Earlier episodes of boom and bust have affected East Asia, the Nordic countries, Mexico, Russia, parts of Latin America, and, most notably, Japan.
The main concern with bubbles arises when they burst, imposing losses on investors holding the bubble assets and potentially on the financial institutions that have extended credit to them. Financial stress might be limited to the failure of individual financial institutions that become overextended during the boom. Increasingly, however, the kinds of financial distress being encountered are systemic, with many institutions operating in a similar mode and simultaneously miscalculating and confronting difficulties. This implies trouble for a wide range of institutions with the strong potential for flow-on effects in the real economy, perhaps leading to a recession or debt deflation. There is no room in this paper to delve into the history of such crashes other than to note that the costs of dealing with banking crises through bailouts and recapitalization during the 1990s ranged from 5 percent to 40 percent of GDP, with even larger effects in terms of lost output (Macfarlane 1999, table 1).
As the frequency and severity of asset price fluctuations, including putative bubbles, has increased, there has been a corresponding increase in interest in measures that may prevent the emergence of bubbles or to seek gradual deflation of bubbles rather than catastrophic busts. Most attention has focused on the idea of making asset price stability a target of monetary policy, either in its own right or as a signal of incipient consumer price inflation. To deal with asset inflation Claudio Borio and Philip Lowe (2003) have called for a "subtle shift" in a policy paradigm based on inflation targeting. There has also been some consideration of a possible role for prudential policy (Schwartz 2002).
Subtle as these policy shifts may appear, they nevertheless involve a fundamental change in thinking about the role of financial markets. In the deregulated system, the task of allocating investment capital and consumer credit between individuals, firms, and nations is left to financial markets. As Jeffrey Carmichael and Neil Esho (2001) have observed, intervention aimed at changing asset prices and other financial market outcomes, such as "excessive" credit growth, are logically inconsistent with the "deregulated" framework of monetary policy and financial regulation that emerged in the wake of breakdown of the Bretton Woods system in the 1970s. This framework was based on the efficient markets hypothesis.
As Borio and Lowe (2003, 113) have observed, framing the debate in this way "can easily see it stray into almost ideological territory, unnecessarily pitching supporters and skeptics of 'market efficiency' against each other." Borio and Lowe regard this division as a source of artificial difficulties. In this paper, we argue on the contrary that the role of the efficient markets hypothesis is crucial and cannot be disregarded. It follows that the debate must have an ideological, as well as a technical, character. Any serious attempt to stabilize financial market outcomes must involve at least a partial reversal of deregulation.
This paper is organized as follows. We begin with a brief survey of the empirical literature on asset price bubbles and asset price volatility, with particular emphasis on the period following financial deregulation. This is followed by a survey of the theoretical literature on asset prices and the efficient markets hypothesis. Next, the recent debate on possible policy responses to asset price bubbles is critically assessed. In the main section of the paper, we develop the argument that no effective response to asset price bubbles is feasible within the current policy framework and consider possible alternatives. Finally, some concluding comments are offered.
Asset Price Bubbles and Asset Price Volatility
Borio and Lowe (2002) presented data collected by the Bank of International Settlements (BIS) on trends in asset markets across a range of countries since the early 1970s. Asset classes surveyed were equities, commercial and residential property, and a weighted aggregate measure of asset prices derived from these components.
Borio and Lowe discerned the following trends. First, equity prices tend to lead asset price upswings and are also the most volatile, followed, respectively, by commercial and residential property. Second, the aggregate asset data, in particular, reveal three broad cycles of asset inflation since the early 1970s, roughly corresponding to the early to mid 1970s, the mid 1980s to the early 1990s, and the mid 1990s to the present. Third, the amplitude and length of the cycles appears to be growing, with the latest upswing being driven mainly by equity markets.
In addition to these fluctuations in national asset markets, it is important to consider the behavior of exchange rates. When the Bretton Woods system of fixed exchange rates was abandoned in the early 1970s, it was expected that financial market transactions would act to stabilize exchange rates, eliminating the periods of severe overvaluation and undervaluation associated with fixed exchange rate regimes. In reality, the volatility of exchange rates increased substantially following the move to floating currencies.
Growth in the volatility of asset prices has been one of a number of inter-related developments associated with the end of financial repression which have served to increase both the gross volume of financial transactions and average levels of net indebtedness of households, business enterprises, and governments.
Borio and Lowe (2002) and others, such as Charles Goodhart (1999), point to the links between asset price inflation and credit growth. In most industries, as supply increases, prices and profits are squeezed, thus limiting expansion in the sector. This is not necessarily true of the financial sector. Once under way, a credit expansion will tend to boost output and push up asset values through leveraged acquisitions, thus promoting further credit expansion (Crockett 2001). Upswings in asset prices have been associated with high rates of growth in the volume of credit. There is also evidence linking credit growth to banking crises and periods of financial stress (Bell and Pain 2000; Eichengreen and Areta 2000). As the BIS argued in its 2001 Annual Report,
Financial factors have long played a role in shaping business cycles. However, as domestic financial systems and international capital flows have been liberalized, this role has grown. Developments in credit and asset markets are having a more profound effect on the dynamics of the typical business cycle than was the case a few decades ago, and have also contributed to the increased frequency of banking system crises. (123) Hence, as a consequence of financial liberalization, the links between the workings of the financial system and the health of the economy have become tighter. This connection is heightened by increased levels of business and household debt exposure in recent years and by the entry of pension funds and small investors to equity and property markets. While the broad link between credit growth and asset inflation seems clear, the exact dynamics of the relationship are still poorly understood. Borio and Lowe (2002) have argued that because of the limited nature of the existing research we are still unable to answer questions such as when credit growth should be considered "excessive," what the cumulative effects of credit expansions might be, or how credit booms might interact with other financial imbalances.
On the demand side, borrowers have been eager to increase their gearing. An important factor behind this has been price. The achievement of low inflation in many countries in the 1980s and 1990s has seen a very substantial reduction in interest rates and the cost of borrowing. Hence, while debt-income ratios among firms and households have risen steeply across many countries in recent years, debt-servicing ratios and interest-income ratios have remained relatively stable. Lower real interest rates have made debt more affordable and encouraged higher borrowing. Money illusion associated with lower nominal interest rates has also tended to play a role in promoting higher gearing, as have the wealth effects born of rising asset prices.
Low inflation or monetary stability is also implicated in asset inflation in another way. Conventional wisdom holds that sharp fluctuations in inflation could destabilize the financial system, for example, by increasing the cost of debt if inflation suddenly falls. Similarly, high inflation tends to encourage debt-based asset acquisitions and other forms of speculative behavior. Hence, monetary stability and financial stability have tended to be seen as complementary. This does not, however, mean that monetary stability and financial instability are mutually exclusive. Three of the biggest asset bubbles in the twentieth century-America's in the 1920s and 1990s and Japan's in the late 1980s-occurred in a low inflation context. In Japan in the late 1980s, CPI inflation remained at close to zero while equity prices almost tripled and commercial property in Tokyo more than tripled. More generally, inflation in most developed countries has...