Bubbles are situations in which asset prices persistently deviate from their fundamental values that is, the prices warranted by the true earning potential of firms. Until recently, economists believed that asset prices as a rule reflected fundamental values, and that bubbles were highly implausible. This view relied on an argument of backward induction: the worth of any asset in the period before the final payoff datewould simply be equal to the discounted value of its face value in the period it expired that is, terminal value. Extending this argument backward, the value of an asset at any period would be equal to the discounted present value of its future stream of revenue. Otherwise, it was argued, an arbitrage opportunity would arise and the price would be pushed back to its fundamental value by the exploitation of this opportunity by informed traders.
Both the rise of rational expectations theory in economics and the idea that assets prices (such as shares and exchange rate) always incorporate all available public information (efficient market hypothesis in itsmost widely accepted semistrong form) went hand in hand with the backward induction argument, underpinning the view that the best way to understand asset prices changeswas by focusing on information about fundamental values.
With the rise of behavioral finance theory since the mid-1980s, however, the view that asset prices equal discounted present value of future streams of revenue has come under critical scrutiny (Shiller 2003). In fact, with the growing skepticismabout the efficient market hypothesis, much of the theoretical literature on asset pricing has been in a ‘‘vibrant flux’’ (Hirshleifer 2001). Given this change of paradigmin finance theory, it is hardly surprising that conceptual understanding of bubbles has been changing as well. In fact, in many of the new models with risk-averse traders that have asymmetric expectations it has become next to impossible to define what ‘‘fundamental value’’ is (Allen et al. 1993).
The older, standard view argued that all major asset price variations, including famous historical episodes of bubbles such as the Dutch Tulip Mania (1634 37), the Mississippi Bubble (1719 20), and the South Sea Bubble (1720), were not really ‘‘bubbles’’ in the sense that they were caused by changes in fundamental values (Garber 1990). Although it was always recognized that less than fully rational behavioral traders exist in financial markets, defenders of the standard view have held that informed arbitrageurs would undo anymispricing caused by them even when they did not cancel out each other’s effect (Fama 1970;Malkiel 2003). In this setting, bubbles could arise only in assets that had an infinite horizon (and thus lack a terminal value) because in this instance the backward induction argument did not apply. This case has largely defined the limits of initial theoretical interest in bubbles, going back to the economist Paul Samuelson (1958), and is the foundation ofmore recent work that now goes under the name of ‘‘rational’’ bubbles (also see the comprehensive survey by Santos and Woodford 1997).
With the rise of behavioral theory, however, the effectiveness of arbitrage became a central issue of contention. Behavioral theorists have produced numerous examples where arbitrage is severely limited in its ability to prevent persistent deviations of asset prices from their fundamental values, bolstering the view that major asset price variations are caused by mispricing as opposed to changes in fundamental values (De Long et al. 1990a, 1990b; Shleifer and Vishny 1997; Allen and Gorton 1993).
Contrary to what is often argued, bubbles are not necessarily the result of irrational trading behavior. A body of more recent work suggests that the main cause of bubbles is the uncertainty about higher-order beliefs, that is, rational agents not knowing what other rational agents will do, where irrationality is not a necessary ingredient (Allen et al. 2006). For instance, in Abreu and Brunnermeier (2003) fully rational traders know that the bubble will eventually burst, but in the meantime generate profits by riding it. They cannot individually bring down the price for lack of sufficient funds, but collectively can if they act in tandem. They all know that a common arbitrage opportunity exists, yet they are uncertain when other traders will begin to act on it.Thus each trader has to determine the right time to exit the market without knowing the exit strategies of other traders, and the varied opinions about the right time to exit cause the bubble to persist. In this setting, some news events can have a disproportionately strong effect beyond what their intrinsic information valuewouldwarrant by having the effect of causing traders to synchronize their exit strategies, and thus lead to a precipitous fall in price.
The idea that a bubble bursts only when rational traders’ subjective expectations begin to coalesce provides an interesting connection to some of the currency crisis models with self-fulfilling expectations. But the linkages between the newbody ofwork on bubbles by behavioral theorists and international finance so far remain few, despite the rich potential for fruitful connections between the two literatures. Such linkages are made in many economic models (e.g., Allen and Gale 1999; Morris and Shin 1999; Shin 2005), where it is argued that financial liberalization has led to asset price bubbles in numerous countries around the world, and the banking crises that usually accompany currency crises are seen to result from the bursting of the asset price bubble. See also asymmetric information; banking crisis; carry trade; currency crisis; financial crisis; financial liberalization; interest parity conditions; lender of last resort; speculation