Syndicate content

On Economic Rationality, Bubbles, and Macroprudence

Biagio Bossone's picture

(Non-)rationality in economic decisions

As last year’s choice of the Nobel award for economic sciences well reflects, economists are deeply divided as to whether, and how, rationality should be modified as a basic assumption for modeling asset allocation and pricing decisions.

The three Nobel laureates for 2013 — Eugene Fama, Lars Peter Hansen, and Robert Shiller — epitomize the economics profession’s broad spectrum of positions currently existing on the subject: from Fama’s unflinching faith in the full rationality of economic action to Shiller’s recognition of the influence of non-rational and irrational factors upon human economic determinations, passing through Hansen’s acceptance of “distorted beliefs” as explanations of some otherwise inconsistent economic behaviors empirically observed.

The unresolved differences bear on the scientific status of contemporary macroeconomic analysis, especially since the crisis of 2007-09 has demonstrated the inadequacy of its underlying microfoundations. Particular attention has since been placed by economists on what they really know about asset bubbles, as these cannot be endogenized within purely rational choice models, and policymakers have re-considered whether bubbles can (or should) be managed in the public interest.

(Ir)rationality and bubbles

In fact, while it is unnecessary to abandon the rationality hypothesis to understand real-world economic phenomena, and financial crises in particular, combining human rational thinking with changing emotional states should feature prominently in the economics research agenda both to gain a deeper appreciation of real decision-making processes and to design policy tools that can re-orient individuals’ decisions, when necessary, toward superior public good objectives.

This is what I have tried to do with the general utility-based approach to asset allocation and pricing, which I have recently developed to study agents’ responses to shocks when expectations reflect the interaction of knowledge and changing market sentiment.

The approach rests on three building blocks:

  • All assets deliver utility. Assets of all types are considered as “vehicles” to future consumption, each characterized by its own peculiar “speed” (that is, the immediacy and the cost of converting it into consumption) and “power” (that is, its capacity to accumulate and store wealth over time at some risk). Greater speed would come at some power cost, and vice versa. Also, at each instant the agent would be faced with the likelihood of having to liquidate the asset to face a consumption shock: changes in likelihood affect differently the utility from asset with different speed and power load. The instantaneous utility of every asset is thus calculated through the agents’ time-horizon as the expected value of the discounted summation of stochastic (uncertain) consumption utility, to which the asset gives access, net of the (uncertain) consumption utility lost to the asset liquidation cost.
  • Variable cost of asset liquidation. Every asset is characterized by an optimal speed, defined as the shorter time-interval possible for the asset to be sold at the minimum liquidation (or transaction) cost possible. Asset optimal speeds are structural parameters determined by the economy’s level of institutional and technological development: all else equal, a more efficient and safe financial infrastructure allows asset liquidation to be effected more rapidly and at lower costs.[1] Some assets can be liquidated and converted into consumption immediately and at no cost, while others require longer time-intervals and involve positive costs. Having to liquidate an asset at a higher than optimal speed (owing, for instance, to immediate and unexpected consumption needs) results in higher liquidation costs or forces the agent to accept larger discounts on the asset sale price. Uncertainty affects asset utility also by influencing expected asset liquidation costs.
  • Rational expectations and emotions. (1) Expectations depend on the state of knowledge: better knowledge and information help the agents to form more precise expectations about future relevant variables, while lower quality knowledge and information cause their predictions to be less determinate. (2) Expectations depend also on emotions: defining optimism (pessimism) as the state of mind that induces the agents to expect superior (inferior) outcomes of future events than would otherwise be reasonable for them to expect exclusively on the basis of the given knowledge and information, optimistic (pessimistic) expectations derive from a “distortion” process. This process introduces a deviation between purely rational predictions and predictions that affected by emotional states of mind (e.g., “animal spirits”): optimism (pessimism) distorts the way agents process and interpret information.

All assets can then be directly comparable based on the utility they deliver, as differently affected by perceived uncertainty and emotional states. In equilibrium, asset allocations and prices must be such that, at each date, all agents extract from the last resource unit still left unallocated the same utility from all assets, which in turn must equal the marginal utility derived from optimal intertemporal consumption at the given date. The agent’s consumption and portfolio decisions can be simultaneously determined as inter- and intra-temporal solutions to an optimal programming problems: in deciding about which asset to convert into consumption, when, and how much, the agent evaluates the expected utility gains/losses associated with the different speed vs. power tradeoffs featured by the different assets under the possible future states of nature and emotions.

Bubbles as asset price “overshooting”

The approach can be used, inter alia, to perform an instructive dissection of the microeconomic dynamics of bubbles, and to draw a policy proposition.

Let’s take a new production technology being launched in the market, which promises to yield large increases of output productivity. The offer price of the new asset and its existing stock of shares are both initially very low, making the asset’s relative marginal utility (that is, standardized to the asset’s unit price) abnormally high relative to that of consumption and the other assets available.

Under complete knowledge and full information, and assuming a fully flexible supply of the asset stock, both its price and quantity would settle instantly at their rational-expectations equivalents. However, under optimistic expectations and a sluggish supply of the asset stock (due to short-term rigidities), the dynamics can give origin to speculative booms. Specifically, if positive news on the asset leads the agents to expect increasing trends of its price, an upbeat market lowers the asset optimal liquidation cost and raises its utility. Also, as optimism prevails, agents keep delaying the expected time of stock liquidation, once again raising its utility. Finally, optimism lowers the asset’s expected liquidation cost, once again raising its utility.

Meanwhile, as the asset stock supply adjusts only slowly to demand, the stock return grows as the asset’s yield and price both rise. All this contributes to keeping the relative marginal utility of the asset (and, hence, its demand) increasing. But as the stock supply responds slowly to increasing demand, equilibrium requires that its price overshoot its rational-expectations equivalent. Price overshooting feeds into optimistic expectations and causes the stock price to stay (or to rise farther up) above equilibrium even as its supply starts adjusting to demand. This can generate a sequence of temporary equilibria characterized by levels of the asset stock and price higher than their rational-expectations equivalents, thus forming a bubble.

Pricking bubbles: macroprudential policy options

As the crisis has challenged the consensus whereby monetary policy can ensure financial stability by targeting the price inflation of goods and services, alternative (macroprudential) policy tools have been designed to prevent financial imbalances.

My approach suggests a type of tool that could be activated when the asset price dynamics indicate the presence of speculative activity. Bubble growth could be pre-empted by increasing the asset’s liquidation cost, thus “throwing sand in the wheel” of its trading; this could be accomplished by an authority imposing an ad-hoc (adjustable) fee on the trading price of the asset. Set at an appropriate level, the fee would discourage excess demand for the asset, as it would lower its utility. By altering the asset’s (current and future) optimal trading cost, the fee would force the economy to settle at a lower equilibrium level of its stock.

To the extent that the tool would become entrenched in financial policymaking, and credible, it could stabilize market behaviors ex ante — much as a street speed control limit and fine would discourage drivers from exceeding the limit. Clearly, the fee would distort the given asset market price, but it would act as a corrective device to the distortive effect of optimistic expectations discussed above.

Compared with using monetary policy to prick bubbles, as strongly suggested by Roubini, the ad-hoc fee tool would be market- (or asset-) specific, thus averting the disruptive consequences on the rest of the economy from using coarser (non-selective) instruments like interest rates or credit restraints, as noted by Posen. The ad-hoc fee tool, also, could be calibrated to any level necessary to discourage excess trading of an asset, a flexibility feature that is unavailable to the monetary authorities (see Bernanke). The tool could be more specifically targeted than the central banks’ macroprudential instruments, which are intended to constrain the supply of funds from financial intermediaries.[2]

Finally, the use of a quasi-fiscal tool would call for reconsidering the question of whether macroprudential policy should overburden central banks, or should not better be assigned to an inter-institutional systemic-risk supervisory institution with broader (not just financial) instruments and a horizontal (not just vertical) responsibility extending across banks, securities firms, financial and nonfinancial markets, and geographies, as was alluded to by the New York’s Fed chairman William Dudley soon after the crisis.

*I wish to thank Luigi Passamonti and Richard Wood for their comments and suggestions. Obviously, I remain the only responsible for the content of the commentary.
[1] For an early theoretical and empirical study on the effects of financial infrastructure development, see my work with Sandeep Mahajan and Farah Zahir.
[2] For a comprehensive state-of-the-art evaluation of central banks’ macroprudential tools, see the report from the Committee on the Global Financial System.


Add new comment