FEATURED MEDIA
Financial markets are an environment of immense complexity and uncertainty. People make investment and business decisions based on their best guesses of future economic conditions, which compounds uncertainty.
Often the economic models that economists use to predict the future are not designed to cope with those unknown factors, according to Lars Peter Hansen, David Rockefeller Distinguished Service Professor in Economics and Statistics. Hansen opened his Becker Brown Bag lunchtime lecture on April 17 by placing emphasis on the need to think about the concept of uncertainty in broader terms than is typical in economic analyses.
Drawing from macroeconomics, asset pricing, and statistics, Hansen highlighted how the interplay between these fields helps illuminate a broader understanding of uncertainty. Using those insights, he showed how to assess economic uncertainty from the inside and outside of the economic models.
Hansen offered a brief overview of the history of probability theory, starting with Swiss mathematician Jacob Bernoulli’s fundamental 1713 theorem “Law of Large Numbers.” From historical precedents, Hansen identified three forms of uncertainty: risk, where outcomes are unknown but probabilities are known; ambiguity, where the underlying model determining the probabilities is unknown; and model misspecification, where one can’t be sure of the inevitable flaws in the model.
Hansen explained that Bernoulli’s proof that repeated experiments produce results that converge on reliable probabilities provided the underpinnings for the theory of rational expectations, outlined by John Muth in 1961 and University of Chicago’s Robert E. Lucas in 1972. The theory says that economic actors, such as investors, have long used histories of data to predict future outcomes and as well as expectations determined within specific economic models so that their collective expectations match those of the economic model.
This provides a tractable way to embed expectations and uncertainty into economic models and a coherent way for assessing alternative economic policies, Hansen explained. However, in some applications, too much emphasis has been placed on risk and not enough on uncertainty, according to Hansen.
Models are essential for analyzing today’s highly complex and intrinsically linked global economy. But, because such models tend to be highly stylized and simplified, they are therefore imperfect. Hansen argued that model misspecification—the challenge of figuring out how to use these flawed models—is the most important and challenging component to uncertainty. How we determine what forms of misspecification are most consequential for resource allocation and the behavior of financial markets is most important to analyze, according to Hansen.
Hansen, the 2013 Nobel laureate in economics, used various pieces of art to illustrate his points. Georges de La Tour’s The Cheat appeared on the cover of one of his books—but without one person as depicted in the original painting. Hansen explained how using probability models with a naïve view of their value, whether in a card game depicted in a painting or information in the real economic world, can lead to damaging outcomes.
Economists have been focusing on flawed models which are currently too simplistic to accurately capture investor behavior. Hansen hopes that economists can go beyond the rational expectations models and embrace the uncertainty that comes along with living in a vastly complex economic world. He sees significant value in pushing beyond the risk models to assess statistical complexity, however uncertain it may be.
Hansen used asset pricing to illustrate the price of economic uncertainty. In financial markets, risks can be mitigated by investing in a diverse portfolio. But in times of major macroeconomic shocks, all sectors of the economic are affected, leaving no paths to diversify against these shocks. Investors require compensation for the potential of exposure to such shocks.
The lens through which an existing economic model is viewed and the economic climate at the time can affect the intrinsic value of model. For instance, persistence in macroeconomic growth can be perceived differently depending on the economic climate at the time.
In good economic times, investors value persistence and want the boom to continue, while in bad economic times, investors hope for a lack of persistence. Conversely, what they fear changes as well. More generally, by assessing the statistical complexity of the economic environment, economic researchers can establish a range of behavioral anomalies that could potentially occur without being obviously counterfactual. This allows them to contemplate the compensations necessary to alleviate poor economic outcomes.
Turning to policy implications, Hansen called for more acknowledgment and acceptance of uncertainty. Policymakers tend to like economists who make clear and confident predictions, but that confidence is often not justified by concrete evidence. He pointed to climate change as an area where models forecasting the human impact on the climate are flawed and incomplete.
It’s better to acknowledge the potential gaps in human understanding and intelligently assess the range of probabilities of the potential adverse impacts. “The fact that our models aren’t perfect doesn’t mean we shouldn’t act to address climate change,” he concluded, “The potential for bad consequences can suffice for acting now instead of waiting until uncertainty is more fully revealed.”
— by Diana Petrova