Can Tian 田粲

Research

Macroeconomics, networks in macroeconomics, firm heterogeneity and dynamics, innovation and growth, search and matching, business cycles, information frictions.

Working papers

Innovation and Endogenous Knowledge Network Dynamics

with Jie Cai · Working paper, July 2024

PDF Earlier version (Feb 2023) Abstract & BibTeX

Abstract

We study innovation in evolving knowledge networks and its growth implications. Empirically, a new link in the knowledge network is approximated as the first citation between a pair of technology categories. We find that these new links likely connect high-quality innovations and large firms on both ends. After forming a new link, responsible firms’ innovation and real performance improve on both ends with positive spillovers to other firms. We build a tractable model to rationalize these findings and analytically characterize the stationary equilibrium. In the model, the knowledge network evolves endogenously as each firm searches for others as knowledge input suppliers and attracts others as users of its own knowledge. In equilibrium, faster growth is associated with denser and better knowledge networks. Endogenous link formation amplifies the effect of innovation incentives on growth. The calibrated model suggests that the majority of growth in the US has network origins.

Intermediaries in Decentralized Markets: Evidence from Used-Car Transactions

with Fei Li, Charles Murry and Yiyi Zhou · NBER Working Paper 34136 · revision requested, Journal of Political Economy

NBER Abstract & BibTeX

Abstract

We develop and estimate a spatial search-and-bargaining model to study the role of intermediaries and spatial frictions using data from the used-car market. We find that dealers earn price premiums by leveraging three key advantages: selective acquisition of higher-quality cars, superior matching efficiency, and greater bargaining power. Counterfactual simulations reveal that selective intermediation and spatial segmentation significantly affect market efficiency and consumer welfare. While dealers extract more surplus per transaction than sellers, policies reducing dealers' advantages increase search frictions and lower overall welfare. Counterintuitively, reducing spatial frictions harms consumers by shifting trade to less efficient private-seller channels.

Publications

Inventory Management in Decentralized Markets

with Fei Li, Charles Murry and Yiyi Zhou · International Economic Review, vol. 65(1), pp. 431-470, 2024

Journal version Abstract & BibTeX

Abstract

We present a directed search model of intermediaries' dynamic inventory and revenue management. Search frictions hinder instantaneous replenishment, prompting intermediaries to utilize dynamic inventory‐based pricing and ordering strategies. In equilibrium, when inventory is high, an intermediary posts a lower retail price to speed up sales and depresses wholesale price to slow down purchases. We characterize the evolution dynamics of inventory holdings and their steady‐state distribution, and extend the model to multiunit wholesale orders, product differentiation, and heterogeneous intermediaries. Using data from used‐car dealers, we quantitatively evaluate the welfare consequence of used‐car dealers' inventory management practice.

Learning and Firm Dynamics in a Stochastic Equilibrium

Journal of Economic Theory, vol. 203, art. 105486, 2022

Journal version Abstract & BibTeX

Abstract

New firms have unknown profitability types, but they learn in the market. With imperfect information, the learning mechanism matters at the firm level and in the aggregate. This paper studies the effect of learning on the distribution of firms, aggregate dynamics, and welfare when firm growth and the entry and exit dynamics are belief-driven. Aggregate uncertainty affects heterogeneous firm beliefs about own types. Faster learning raises firms' continuation values, lowers the selection threshold belief, and thickens the belief distribution's right tail. Aggregate uncertainty lowers firm value by reducing firms' learning speed. Transitory aggregate shocks have persistent impacts on the economy by shifting existing firms' belief updating paths, affecting firm decisions along both intensive and extensive margins. Faster learning improves social welfare.

Input-Output Linkages in Pigouvian Industrial Fluctuations

Journal of Monetary Economics, vol. 117, pp. 1078-1095, 2021

Preprint (PDF) Journal version Abstract & BibTeX

Abstract

Input-output linkages facilitate expectations-driven fluctuations. Signals about yet-to-be-realized aggregate (macro) and idiosyncratic (micro) fundamentals in the future affect current equilibrium outcome through chains of input needs when inputs require time to build. Depending on their importance on these chains, firms respond differently to common signals. The network structure, together with input time to build, determines the magnitude of signal-induced aggregate fluctuations. Macro signals need large input shares on average to produce sizable aggregate volatility, whereas micro signals require cross-sectional variations of input importance. Changing input time to build and altering the network structure have similar effects on equilibrium outcomes.

Firm-Level Entry and Exit Dynamics over the Business Cycles

European Economic Review, vol. 102, pp. 298-326, 2018

Journal version Earlier version (SSRN) Abstract & BibTeX

Abstract

This paper presents a set of facts on the cyclicality of firm births and deaths in the U.S. during the period 1979–2013. Asymmetry in the cyclicality between firm birth and firm death is observed: aggregate firm birth is generally significantly procyclical, while the (counter-)cyclicality in firm death is ambiguous and sensitive to the choice of cyclical indicator. Such a pattern exists within the majority of sectors as well and is mainly driven by small firms. Young firms are the main contributors to the findings on firm death activity. Further examination along the time-dimension discovers a reason for the contemporaneous cyclicality (or lack thereof) which can be documented as a new fact: firm births positively lead the usual business cycle measures and firm deaths positively lag them. Moreover, cyclicality in business openings and closings differs from establishment level to firm level due to the internal adjustment of continuing firms.

Riskiness, Endogenous Productivity Dispersion and Business Cycles

Journal of Economic Dynamics and Control, vol. 57, pp. 227-249, 2015

Journal version Slides Abstract & BibTeX

Abstract

In the data, cross-sectional productivity dispersion is countercyclical at both the plant level and the firm level, see e.g. Bloom (2009). I incorporate a firm's choice of risk level into a model of firm dynamics with real business cycle features to explain this empirical finding both qualitatively and quantitatively. In the model, in every period, each firm chooses the investment amount and the risk level associated with a production project every period. All projects available to each firm have the same expected flow return, determined by the aggregate and idiosyncratic shocks to the firm's productivity, and differ from one another only in their risk. The endogenous option of exiting the market and the limited funding for new investment jointly play an important role in motivating firms' risk-taking behavior. The model predicts that, in each period, relatively small firms are more likely to take risk and hence exhibit a higher exit rate, and that the cross-sectional productivity dispersion, measured as the standard deviation of the realized individual component of productivity, is larger in recessions.

Directed Search and Job Rotation

with Fei Li · Journal of Economic Theory, vol. 148(3), pp. 1268-1281, 2013

Journal version Abstract & BibTeX

Abstract

We consider the impact of job rotation in a directed search model in which firm sizes are endogenously determined and match quality is initially unknown. A large firm benefits from the opportunity to rotate workers so as to partially overcome the loss of mismatch. As a result, in the unique symmetric equilibrium, large firms have higher labor productivity and lower separation rates. In contrast to the standard directed search model with multi-vacancy firms, this model can generate a positive correlation between firm size and wage without introducing any ex ante productivity differences or imposing any non-concave production function assumptions.

Work in progress