Latest Research Publications

A common problem in econometrics, statistics, and machine learning is to estimate and make inference on functions that satisfy shape restrictions. For example, distribution functions are nondecreasing and range between zero and one, height growth charts are nondecreasing in age, and production functions are nondecreasing and quasi-concave in input quantities. We propose a method to enforce these restrictions ex post on generic unconstrained point and interval estimates of the target function by applying functional operators. The interval estimates could be either frequentist confidence bands or Bayesian credible regions. If an operator has reshaping, invariance, order-preserving, and distance-reducing properties, the shape-enforced point estimates are closer to the target function than the original point estimates and the shape-enforced interval estimates have greater coverage and shorter length than the original interval estimates. We show that these properties hold for six different operators that cover commonly used shape restrictions in practice: range, convexity, monotonicity, monotone convexity, quasi-convexity, and monotone quasi-convexity, with the latter two restrictions being of paramount importance. The main attractive property of the post-processing approach is that it works in conjunction with any generic initial point or interval estimate, obtained using any of parametric, semi-parametric or nonparametric learning methods, including recent methods that are able to exploit either smoothness, sparsity, or other forms of structured parsimony of target functions. The post-processed point and interval estimates automatically inherit and provably improve these properties in finite samples, while also enforcing qualitative shape restrictions brought by scientific reasoning. We illustrate the results with two empirical applications to the estimation of a height growth chart for infants in India and a production function for chemical firms in China.

We study the transmission of financial news and opinions through social interactions among retail investors in the United States. We identify a series of plausibly exogenous shocks, which cause “treated investors” to trade abnormally. We then trace the “contagion” of abnormal trading activity from the treated investors to their neighbors and their neighbors’ neighbors. Coupled with methodology drawn from epidemiology, our setting allows us to estimate the rate of communication and how it varies with the characteristics of the underlying investor population.

We evaluate the connection between corporate characteristics and the reaction of stock returns to COVID-19 cases using data on more than 6,700 firms across 61 economies. The pandemic-induced drop in stock returns was milder among firms with stronger pre-2020 finances (more cash and undrawn credit, less total and short-term debt, and larger profits), less exposure to COVID-19 through global supply chains and customer locations, more corporate social responsibility activities, and less entrenched executives. Furthermore, the stock returns of firms controlled by families (especially through direct holdings and with non-family managers), large corporations, and governments performed better, and those with greater ownership by hedge funds and other asset management companies performed worse. Stock markets positively price small amounts of managerial ownership but negatively price high levels of managerial ownership during the pandemic.

The telegraph was introduced to China in the late 19th century, a time when China also saw the rise of modern banks. Based on this historical context, this paper documents the importance of information technology in banking development. We construct a data set on the distributions of telegraph stations and banks across 287 prefectures between 1881 and 1936. The results show that the telegraph significantly expanded banks’ branch networks in terms of both number and geographic scope. The effect of the telegraph remains robust when we instrument it using proximity to the early military telegraph trunk.

An agent may privately learn which aspects of his job are more important by shirking on some of them, and use that information to shirk more effectively in the future. In a model of long-term employment relationship, we characterize the optimal relational contract in the presence of such learning-by-shirking and highlight how the performance measurement system can be managed to sharpen incentives. Two related policies are studied: intermittent replacement of existing measures, and adoption of new ones. In spite of the learning-by-shirking effect, the optimal contract is stationary, and may involve stochastic replacement/adoption policies that dilute the agent’s information rents from learning how to game the system.

Organizations frequently rely on peer performance ratings to capture employees’ unique and difficult to observe contributions at work. Though useful, peers exhibit meaningful variance in the accuracy and informational utility they offer about ratees. In this research, we develop and test theory which suggests that raters’ social network positions explains this variance in systematic ways. Drawing from information processing theory, we posit that members who occupy core (peripheral) positions in the network have greater (less) access to firsthand and secondhand performance information about ratees, which is in turn associated with more (less) accurate performance ratings. To overcome difficulties in obtaining a “true” performance score in interdependent field settings, we employ an external criterion comparison method to benchmark our arguments, such that larger validity coefficients between established predictors of performance (i.e., a ratee’s general mental ability [GMA] and conscientiousness) and peer performance ratings should reflect more (less) accurate ratings for core (peripheral) members. In Study 1, we use an organization-wide network in a technology startup company to examine the validity coefficient of a ratee’s GMA on performance as rated by central versus peripheral members. In Study 2, we attempt to replicate and extend Study 1’s conclusions in team networks using ratee conscientiousness as a benchmark indicator. Findings from both studies generally support the hypotheses that core network members provide distinct, and presumably more accurate, peer performance ratings than peripheral network members.

We propose and estimate a quantitative model of exchange rates in which participants in the foreign exchange market are intermediaries subject to value-at-risk (VaR) constraints. Higher volatility translates into tighter VaR constraints, and intermediaries require higher returns to hold foreign assets. Therefore, the foreign currency is expected to appreciate. The model quantitatively resolves the Backus–Smith puzzle, the forward premium puzzle, and the exchange rate volatility puzzle and explains deviations from the covered interest rate parity. Moreover, the model implies both contemporaneous and predictive relations between proxies of leverage constraint tightness and exchange rates. These implications are supported in the data.

In this paper we investigate the impact of external monitoring from the government on state-owned enterprise performance, using the variation in monitoring strength arising from a nationwide policy change and firms’ geographic location in China. We utilise a structural approach to estimate input prices and productivity separately at the firm level using commonly available production data. We show that enhanced external monitoring, as a key component of corporate governance, can substantially reduce managerial expropriation in procurement (proxied by input prices) and shirking in production management (proxied by productivity). The results suggest that government monitoring can be an effective policy instrument to improve state-owned enterprise performance.

In social networks, social foci are physical or virtual entities around which social individuals organize joint activities, for example, places and products (physical form) or opinions and services (virtual form). Forecasting which social foci will diffuse to more social individuals is important for managerial functions such as marketing and public management operations. In terms of diffusive social adoptions, prior studies on user adoptive behavior in social networks have focused on single-item adoption in homogeneous networks. We advance this body of research by modeling scenarios with multi-item adoption and learning the relative propagation of social foci in concurrent social diffusions for online social networking platforms. In particular, we distinguish two types of social nodes in our two-mode social network model: social foci and social actors. Based on social network theories, we identify and operationalize factors that drive social adoption within the two-mode social network. We also capture the interdependencies between social actors and social foci using a bilateral recursive process—specifically, a mutual reinforcement process that converges to an analytical form. Thus, we develop a gradient learning method based on a mutual reinforcement process that targets the optimal parameter configuration for pairwise ranking of social diffusions. Further, we demonstrate analytical properties of the proposed method such as guaranteed convergence and the convergence rate. In the evaluation, we benchmark the proposed method against prevalent methods, and we demonstrate its superior performance using three real-world data sets that cover the adoption of both physical and virtual entities in online social networking platforms.