Huiyu Li, Economist, Federal Reserve Bank of San Francisco

Huiyu Li


Macroeconomic Research

Growth, Firm dynamics, Computational methods (at)


Working Papers
Hide this section
Leverage and Productivity

Job Market Paper | July 2016

abstract (+)
Financial frictions can reduce aggregate productivity, in particular when firms with high productivity cannot borrow against their profits. This paper investigates the quantitative importance of this form of borrowing constraint using a large panel of firms in Japan. The firms are young and unlisted, precisely the firms for which credit frictions are expected to be the most severe. In this data, I find that firm leverage (asset-to-equity ratio) and firm output-to-capital ratios rise with firm productivity, both over time in a firm and across firms of the same age and cohort. I use these facts in indirect inference to estimate a standard general equilibrium model where financial frictions arise from the limited pledgeability of profits and capital. In this model more financially constrained firms have higher output-to-capital ratios. The model matches the two facts the best when firms can pledge half of their one-year-ahead profits and one-fifth of their assets. Compared to the common assumption that firms can pledge only assets, aggregate productivity loss due to financing frictions is one-third smaller when profits are also pledgeable to the degree seen in Japan.
Missing Growth from Creative Destruction

2017-04 | With Aghion, Bergeaud, Boppart, and Klenow | January 2017

abstract (+)
Statistical agencies typically impute inflation for disappearing products from the inflation rate for surviving products. As some products disappear precisely because they are displaced by better products, inflation may be lower at these points than for surviving products. As a result, creative destruction may result in overstated inflation and understated growth. We use a simple model to relate this “missing growth” to the frequency and size of various kinds of innovations. Using U.S. Census data, we then apply two ways of assessing the magnitude of missing growth for all private nonfarm businesses for 1983–2013. The first approach exploits information on the market share of surviving plants. The second approach applies indirect inference to firm-level data. We find: (i) missing growth from imputation is substantial—0.5 percentage points per year when using the first approach, 1 percentage point per year using the second method; and (ii) almost all of the missing growth is due to creative destruction (as opposed to new varieties).
supplement (+)
wp2017-04_appendix.pdf – Supplemental Appendix
wp2017-04_slides.pdf – Slides
Entry Costs Rise with Development

Stanford Manuscript | With Bollard and Klenow | June 2016

abstract (+)
Across cohorts of firms and plants within the U.S., Indonesia, India and China, we find that average discounted profits rise systematically with average labor productivity at the time of entry. The number of entrants, in contrast, is weakly connected to average labor productivity but closely tied to aggregate employment. In many models of firm dynamics, growth, and trade, these facts imply that the cost of creating a new business is increasing with average productivity given a zero profit condition for entrants. Entry costs could rise as development proceeds because entry is laborintensive and/or because it is more expensive to set up firms using more skilled workers and more sophisticated technology.
The Asymptotic Distribution of Estimators with Overlapping Simulation Draws

Manuscript | With Armstrong, Gallant, and Hong | August 2015

abstract (+)
We study the asymptotic distribution of simulation estimators, where the same set of draws are used for all observations under general conditions that do not require the function used in the simulation to be smooth. We consider two cases: estimators that solve a system of equations involving simulated moments and estimators that maximize a simulated likelihood. Many simulation estimators used in empirical work involve both overlapping simulation draws and non-differentiable moment functions. Developing sampling theorems under these two conditions provides an important complement to the existing results in the literature on the asymptotics of simulation estimators.
Published Articles (Refereed Journals and Volumes)
Hide this section
Numerical Policy Error Bounds for η -Concave Stochastic Dynamic Programming with Non-interior Solutions

Computational Economics 46, (Issue 2), August 2015, 171-187

abstract (+)
This paper derives explicit error bounds for numerical policies of η-concave stochastic dynamic programming problems, without assuming the optimal policy is interior. We demonstrate the usefulness of our error bound by using it to pinpoint the states at which the borrowing constraint binds in a widely used income fluctuation problem with standard calibrations and a firm production problem with financial constraints.
Solving the Income Fluctuation Problem with Unbounded Rewards

Journal of Economic Dynamics and Control 45, August 2014, 353-365 | With Stachurski

abstract (+)
This paper studies the income fluctuation problem without imposing bounds on utility, assets, income or consumption. We prove that the Coleman operator is a contraction mapping over the natural class of candidate consumption policies when endowed with a metric that evaluates consumption differences in terms of marginal utility. We show that this metric is complete, and that the fixed point of the operator coincides with the unique optimal policy. As a consequence, even in this unbounded setting, policy function iteration always converges to the optimal policy at a geometric rate.
Generalized Look-Ahead Methods for Computing Stationary Densities

Mathematics of Operations Research Publication 37 (3), August 2008, 489-500 | With Braun and Stachurski

abstract (+)
The look-ahead estimator is used to compute densities associated with Markov processes via simulation. We study a framework that extends the look-ahead estimator to a broader range of applications. We provide a general asymptotic theory for the estimator, where both L_1 consistency and L_2 asymptotic normality are established. The L_2 asymptotic normality implies root-n convergence rates for L_2 deviation.
FRBSF Publications
Hide this section