Will high computing power substitute the certainty-equivalence assumption?

Question

Bloom in a recent JEP paper considers that "the increase in computing power has made it possible to include uncertainty shocks directly in a wide range of models, allowing economists to abandon assumptions built on “certainty equivalence,” which refers to the amount of money that would be required as compensation for risk." (2nd paragraph of page 154, third point).

My understanding of Bloom's point is the idea that with computing power we can treat and exploit the heterogeneity of data. We can use high-frequency and/or large-scale data, in combination with computing power, to identify the role of uncertainty shock on economic outcomes.

My guess is that Bloom's point could be an implicit criticism of the expected utility theory given the importance of the certainty equivalent concept, and the related concept of risk premium, in this framework. Is this guess/interpretation right?

Answer 1

This is perhaps a good opportunity to point out that the "certainty equivalence" concept means one thing in microeconomics/choice under uncertainty theory, while it means something different in macroeconomics.

Microeconomics/choice under uncertainty
The Certainty Equivalent of a lottery/gamble, is the amount of wealth which, if given with certainty, provides the same utility as the lottery/gamble. See for example Jehle & Reny's "Advanced Microeconomic Theory" (2011, 3d ed.), p.113.

Macroeconomics
"Certainty Equivalence" is the situation in a stochastic model, where optimal decision rules prove to be identical to those that would have been derived in a deterministic framework (see for example Lungqvist & Sargent's "Recursive Macroeconomic Theory" 2004, 2nd ed, p. 113-115). Informally, this is sometimes described as "agents behave as if the stochastic processes are not stochastic", or "the decision rule is not affected by stochastic variability".

It is a funny coincidence that both references treat the subject in the same page number...

The concepts are evidently linked by the "as if there was no uncertainty" angle, but they are different in essential aspects: the micro-concept is a "buy-out" of uncertainty, providing a certain alternative towards which the agent that faces the gamble is indifferent in terms of utility, while the macro-concept emerges as a property of the solution (and only under restrictive model structures, or linear approximations thereof).

Answer 2

Expected utility theory is based on valuing expectations. There is more to that than just uncertainty.

First, a risky investment is different from an uncertain investment. Risk is usually defined as future events occurring only with a measurable probability, while uncertainty is called upon when the likelihood of future events is indefinite or incalculable, but only estimated in itself (it's like a risk on risk). What I'm saying here is actually agreed upon by the author. Note that the first definition of expected utility models did not include any other considerations that risk. Part of the "risk premium" is... a risk premium, rather than an uncertainty premium, and shouldn't be modified by a better understanding of uncertainty.

Another part of the risk premium is the fact that it is usually set outside the micro-economic model, usually at the moment of raising capital. No due diligence can give an investor an "inside" view of the model, and so there will always be a cost of borrowing money, which is to be taken into account in a risk premium for any project that requires money over time.

What the article discusses seems more to me as a motivation to decrease the impact of uncertainty, and so the risk premium, rather than change the way we look at it. Being able to translate uncertainty into risk is great, but won't make risk itself disappear. For some projects so, it is likely to increase the risk premium as risks taken are better quantified, as uncertainty premium is meant to cover hidden/ unknown risk premium only on average.