Lars Peter Hansen Prize Lecture: Uncertainty Outside and Inside Economic Models

---

397

Uncertainty Outside and 

Inside Economic Models

*

Prize Lecture, December 8, 2013

by Lars Peter Hansen

University of Chicago, Chicago, IL, USA.

We must infer what the future situation would have been without our 

interference, and what change will be wrought in it by our action. 

Fortunately or unfortunately, none of these processes is infallible, or 

indeed ever accurate and complete. Knight (1921: 201–202)

1 iNTRoduCTioN

Asset pricing theory has long recognized that financial markets compensate in-

vestors who are exposed to some components of uncertainty. This is where mac-

roeconomics comes into play. The economy-wide shocks, the primary concern 

of macroeconomists, by their nature are not diversifiable. Exposures to these 

shocks cannot be averaged out with exposures to other shocks. Thus, returns on 

assets that depend on these macroeconomic shocks reflect “risk” premia and are 

a linchpin connecting macroeconomic uncertainty to financial markets. A risk 

premium reflects both the price of risk and the degree of exposure to risk. I will 

be particularly interested in how the exposures to macroeconomic impulses are 

priced by decentralized security markets.

*

 I thank Manuel Arellano, Amy Boonstra, Philip Barrett, Xiaohong Chen, John 

Cochrane, Maryam Farboodi, Eric Ghysels, Itzhak Gilboa, Massimo Marinacci, Nan 

Li, Monika Piazzesi, Eric Renault, Scott Richard, Larry Samuelson, Enrique Sentana, 

José Scheinkman, Martin Schneider, Stephen Stigler, Harald Uhlig, Amir Yaron an 

anonymous referee and especially Jaroslav Borovička, James Heckman, Thomas Sargent 

and Grace Tsiang for helpful comments.

6490_Book.indb   397

11/4/14   2:30 PM

---

398 

The Nobel Prizes

How do we model the dynamic evolution of the macroeconomy? Follow-

ing the tradition initiated by Slutsky (1927, 1937) and Frisch (1933), I believe 

it is best captured by stochastic processes with restrictions; exogenous shocks 

repeatedly perturb a dynamic equilibrium through the model’s endogenous 

transmission mechanisms. Bachelier (1900), one of the developers of Brown-

ian motion, recognized the value of modeling financial prices as responses to 

shocks.

1

 It took economists fifty years to discover and appreciate his insights. 

(It was Savage who alerted Samuelson to this important line of research in the 

early 1950s.) Prior to that, scholars such as Yule (1927), Slutsky (1927, 1937) 

and Frisch (1933) had explored how linear models with shocks and propaga-

tion mechanisms provide attractive ways of explaining approximate cyclical 

behavior in macro time series. Similarities in the mathematical underpinnings 

of these two perspectives opened the door to connecting macroeconomics and 

finance.

Using random processes in our models allows economists to capture the 

variability of time series data, but it also poses challenges to model builders. As 

model builders, we must understand the uncertainty from two different perspec-

tives. Consider first that of the econometrician, standing outside an economic 

model, who must assess its congruence with reality, inclusive of its random per-

turbations. An econometrician’s role is to choose among different parameters 

that together describe a family of possible models to best mimic measured real 

world time series and to test the implications of these models. I refer to this as 

outside uncertainty. Second, agents inside our model, be it consumers, entrepre-

neurs, or policy makers, must also confront uncertainty as they make decisions. 

I refer to this as inside uncertainty, as it pertains to the decision-makers within 

the model. What do these agents know? From what information can they learn? 

With how much confidence do they forecast the future? The modeler’s choice 

regarding insiders’ perspectives on an uncertain future can have significant con-

sequences for each model’s equilibrium outcomes.

Stochastic equilibrium models predict risk prices, the market compensa-

tions that investors receive for being exposed to macroeconomic shocks. A chal-

lenge for econometric analyses is to ascertain if their predictions are consistent 

with data. These models reveal asset pricing implications via stochastic discount 

factors. The discount factors are stochastic to allow for exposures to alternative 

1 

See Davis and Etheridge (2006) for a translation and commentary and Dimson and 

Mussavian (2000) for an historical discussion of the link between Bachelier’s contribu-

tion and subsequent research on efficient markets.

6490_Book.indb   398

11/4/14   2:30 PM

---

Uncertainty Outside and Inside Economic Models 399

macroeconomic random outcomes to be discounted differently. Moreover, the 

compounding of stochastic discount factors shows how market compensations 

change with the investment horizon. Stochastic discount factors thus provide 

a convenient vehicle for depicting the empirical implications of the alternative 

models. I will initially describe the methods and outcomes from an econometri-

cian outside the model.

Stochastic discount factors are defined with respect to a probability distri-

bution relevant to investors inside the model. Lucas (1972) and others imposed 

rational expectations as an equilibrium concept, making the probability distri-

bution relevant to investors inside the model coincide with the probability dis-

tribution implied by the solution to the model. It is an elegant response for how 

to model agents inside the model, but its application to the study of asset pric-

ing models has resulted in empirical puzzles as revealed by formal economet-

ric methods that I will describe. These and other asset pricing anomalies have 

motivated scholars to speculate about investor beliefs and how they respond 

to or cope with uncertainty. In particular, the anomalies lead me and others to 

explore specific alternatives to the rational expectations hypothesis.

In this essay I will consider alternatives motivated in part by a decision the-

ory that allows for distinctions between three alternative sources of uncertainty: 

i) risk conditioned on a model, ii) ambiguity about which is the correct model 

among a family of alternatives, and iii) potential misspecification of a model or 

a family of possible models. These issues are pertinent to outside econometri-

cians, but they also may be relevant to inside investors. I will elaborate on how 

the distinctions between uncertainty components open the door to the inves-

tigation of market compensations with components other than more narrowly 

defined risk prices. Motivated by empirical evidence, I am particularly inter-

ested in uncertainty pricing components that fluctuate over time.

Why is it fruitful to consider model misspecification? In economics and as 

in other disciplines, models are intended to be revealing simplifications, and 

thus deliberately are not exact characterizations of reality; it is therefore spe-

cious to criticize economic models merely for being wrong. The important criti-

cisms are whether our models are wrong in having missed something essential 

to the questions under consideration. Part of a meaningful quantitative analysis 

is to look at models and try to figure out their deficiencies and the ways in which 

they can be improved. A more subtle challenge for statistical methods is to ex-

plore systematically potential modeling errors in order to assess the quality of 

the model predictions. This kind of uncertainty about the adequacy of a model 

or model family is not only relevant for econometricians outside the model but 

potentially also for agents inside the models.

6490_Book.indb   399

11/4/14   2:30 PM

---