Tristan Nighswander (University of Oregon) presented Orphanides and Williams (2007), which considers the role of learning and misperceptions at the agent level in a standard NK-type framework. The abstract to the paper is below, while slides from the presentation can be found here.

Abstract:  We examine the performance and robustness properties of monetary policy rules in an estimated macroeconomic model in which the economy undergoes structural change and where private agents and the central bank possess imperfect knowledge about the true structure of the economy. Policymakers follow an interest rate rule aiming to maintain price stability and to minimize fluctuations of unemployment around its natural rate but are uncertain about the economy’s natural rates of interest and unemployment and how private agents form expectations. In particular, we consider two models of expectations formation: rational expectations (RE) and learning. We show that in this environment the ability to stabilize the real side of the economy is significantly reduced relative to an economy under RE with perfect knowledge. Furthermore, policies that would be optimal under perfect knowledge can perform very poorly if knowledge is imperfect. Efficient policies that take account of private learning and misperceptions of natural rates call for greater policy inertia, a more aggressive response to inflation, and a smaller response to the perceived unemployment gap than would be optimal if everyone had perfect knowledge of the economy. We show that such policies are quite robust to potential misspecification of private sector learning and the magnitude of variation in natural rates.

The full text article can be found here.

Matt Wilson (University of Oregon) presented new work from the third chapter of his dissertation. The equity premium puzzle is well known to students of modern macro economics. If households value financial instruments based on their future expected discounted present value, as mainstream finance and economics asserts, then the spread between risky and risk free assets should reflect only the increased risk. The premium implied by standard calibrations of DSGE models is much too low. Much analysis has focused on our assumption for the particular value of the CRRA parameter. Economists often assume this to be around 1. To generate premiums commensurate with real world date, this parameter must be significantly higher.

Wilson shows that by considering disaggregated household data on consumption, he can generate empirically realistic equity premia with a CRRA in the range of 2-2.5. Consumption of individual households is much more volatile than aggregate measures, and this volatility allows confidence intervals for parameter estimates that can more closely accommodate standard calibrations.

Slides from the presentation are here.

George Evans (University of Oregon) presented joint work with Seppo Honkapohja and Kaushik Mitra estimating the size of the fiscal multiplier in a New Keynesian model modified to accommodate boundedly rational agents. Specifically, agents in the model behave as econometricians and are endowed with the reduced form MSV solution to the model. However, the coefficients of this reduced form VAR representation must be estimated by the agents, who update their estimates using recursive least squares. The abstract is below.

Abstract: We examine the output multiplier for temporary increases in government spending in the New Keynesian model when agents form expectations for future inflation and output using adaptive learning. Our model also imposes inflation and consumption lower bounds, which can be relevant when expectations are pessimistic. When the economy is near the targeted steady state the output multiplier is front-loaded under learning but comparable in size to RE. When initial expectations are pessimistic, the economy can sink into a deflation trap. Multipliers under learning will then be very large if the fiscal stimulus is large enough to avoid the trap.

Brian Dombeck (University of Oregon) presented Zhang (2014), an unpublished manuscript exploring the stability properties of rational expectations equilibrium in the well-known expectationally driven business cycle model of Jaimovich and Rebelo (2009). The abstract is below, and a link to the paper is here.

Abstract:  This paper investigates the stability properties of Jaimovich and Rebelo (2009)’s news-driven business cycle model. In doing so, this paper adopts expectational stability (“E-stability”) as a natural criterion for rationality: plausible equilibria should arise from an adaptive learning formulation where agents form forecasts from a correctly specified model whose parameters are updated in real time. In examining the model’s stability properties, I find that when agents do not observe current state  variables when forming expectations, the rational expectations equilibrium (REE) is not learnable for some parameter values capable of generating news-driven recessions. However, stronger informational assumptions regarding the timing of expectations and outcomes can facilitate E-stability.

Shankha Chakraborty (University of Oregon) presented Jones and Kim (2014), in which the authors exploit the tractability of Pareto distributions in creating and analyzing a model of creative destruction which predicts rising inequality, the likes of which we have observed over the past few decades in the United States.

Abstract:  Top income inequality rose sharply in the United States over the last 35 years
but increased only slightly in economies like France and Japan. Why? This paper
explores a model in which heterogeneous entrepreneurs, broadly interpreted, exert
effort to generate exponential growth in their incomes. On its own, this force leads
to rising inequality. Creative destruction by outside innovators restrains this expansion and induces top incomes to obey a Pareto distribution. The development
of the world wide web, a reduction in top tax rates, and a decline in misallocation
are examples of changes that raise the growth rate of entrepreneurial incomes and
therefore increase Pareto inequality. In contrast, policies that stimulate creative
destruction reduce top inequality. Examples include research subsidies or a decline
in the extent to which incumbent firms can block new innovation. Differences
in these considerations across countries and over time, perhaps associated with
globalization, may explain the varied patterns of top income inequality that we see
in the data.

A link to the full paper can be found here.

Jeremy Piger (University of Oregon) presented Auerbach and Gorodnichenko (2012), in which the authors try to estimate a state-dependent fiscal multiplier using a simple SVAR approach. Previous estimates of the fiscal multiplier have ranged from about 0.5 to 1, but these studies have looked at an average response of output to spending shocks, implicitly treating expansion and recession regimes as identical.

To allow for regime switching the authors create a smooth switching function which depends on lags of output growth to create a linear combination of coefficients in expansion and recession regimes. They then construct impulse response functions to a government spending shock and report peak responses as the multiplier. During expansions, estimates of the fiscal multiplier range from 0.5 to 1. During recessions, their estimates imply a multiplier between 1 and 1.5, suggesting fiscal policy can be very effective at diminishing the magnitude and possible length of recessionary episodes.

A link to the paper can be found here.

Erin Hunt (University of Oregon) presented recent work on the normative implications of a government sponsored savings program such as that administered by the Social Security Administration in the USA. Forcing agents to save can be justified if agents have a propensity to under-save relative to the first best solution. Caliendo and Findley (2012) present a non-stochastic, continuous time overlapping generations model in which a continuum of agents have preferences for saving over a distribution of savings rates. The results of their calibration suggest social security schemes can be welfare improving, with the gains from those predisposed to under-save outweighing the losses from those who would have already saved enough.

The slides can be found here, and the paper itself can be found here.

Bruce McGough (Univeristy of Oregon) presented new work on alternative expectations mechanisms agents may use to solve infinite dimensional dynamic programming problems. Under certain fairly weak observability and controllability conditions, agents can learn to optimally solve their Bellman system in the limit. It turns out this solution is equivalent to the solution obtained by classic rational agents solving their Bellman system.