Junkan Li is an economist and interdisciplinary scientist whose research addresses contemporary environmental challenges such as natural disasters and climate change—phenomena that are inherently dynamic, uncertain, and spatially heterogeneous. To evaluate the consequences of these challenges and the policies designed to address them, he leverages both statistical and computational methodologies to develop quantitative frameworks that identify and quantify impact mechanisms and the endogenous adaptation behaviors of agents within economic systems. By drawing on his expertise in quantitative modeling, he aims to build a distinctive interdisciplinary research profile that bridges economics, policy analysis, and environmental science.
Working Papers
1. Voting with Their Feet: Insights from Storm Impacts on Internal Migration in the U.S. Latest Version
with Ian Sue Wing, Linda Nozick, Jamie Kruse, Meghan Millea, Dahui Liu, and Jiahang He
Hurricanes are among the most costly natural hazards in the United States, yet existing evidence on storm-induced migration largely relies on net population changes or event-specific analyses. This paper provides a comprehensive national assessment of hurricane-induced migration by examining bilateral county-to-county migration flows following tropical cyclones. We link Internal Revenue Service migration data to wind exposure from 137 storm events affecting 22 Eastern and Gulf Coast states between 1997 and 2018. Using a gravity-based migration framework within a causal inference design, we estimate how hurricanes affect movements both out of and into impacted counties, allowing for dynamic and severity-dependent responses. Our results reveal simultaneous increases in migration flows from directly impacted counties to non-impacted counties and from non-impacted counties to impacted counties, rejecting the hypothesis that destination preferences are invariant to storm exposure. Extending standard one-dimensional identification to an origin–destination setting, we distinguish hazard effects operating through origins, destinations, and their interaction. Mechanism analysis using a difference-in-difference-in-difference framework suggests that storm-exposed households exhibit safety-seeking behavior, while non-exposed households respond more strongly to post-storm economic opportunities in impacted areas. Aggregating estimated responses, we find that hurricanes generated approximately 1.7 to 2.6 million moves nationally from 2000 to 2018, with gross flows far exceeding net population changes. These findings highlight substantial population turnover following hurricanes and underscore the importance of flow-based approaches for understanding climate-induced migration, regional resilience, and inequality.
2. Heterogeneity in Migration Decisions Following Tropical Cyclones in U.S. Southeast (accepted at Population & Environment)
with Qiyuan Huang, Linda Nozick, Dahui Liu
This paper investigates how tropical cyclones influence both out-migration from affected areas and the attractiveness of storm-impacted destinations in the southeastern United States, with attention to racial, ethnic, and socioeconomic heterogeneity. Using microdata from the American Community Survey linked with physically modeled tropical cyclone wind-speed estimates, we find that storm exposure has meaningful effects on migration behavior and exhibits substantial heterogeneity that may reshape demographic distributions across affected regions. Storm exposure increases out-migration among low-income families while reducing it among high-income families. Examining racial and ethnic heterogeneity indicates that the income effect is concentrated among White families, whereas the same income pattern does not emerge for Hispanic or African American families. In particular, African American families exhibit limited mobility across income levels, suggesting that non-economic barriers constrain relocation decisions. At the same time, storm-exposed destinations become significantly less attractive to potential migrants, particularly those with lower incomes. Notably, areas with favorable pre-storm economic conditions, such as high socioeconomic status and rapid growth, lose their appeal even further, especially among minority populations, highlighting unequal exposure to the secondary effects of disasters through the erosion of local opportunities. Overall, migration emerges as an uneven adaptive response shaped by intersecting racial, ethnic, and economic factors, emphasizing the need to integrate equity considerations into post-disaster recovery and adaptation planning.
3. Quantify the Aggregate Effects of the Indirect Cost of Financial Distress, Latest Version
This paper studies the interactions between the firm's financial conditions and the demand for the goods it produces, focusing on the physical capital manufacturers in Canada. The interactions are constructed through the capital’s maintenance and repair (M&R), which is also supplied by capital manufacturers and is necessary to keep the used capital operative. When the possible bankruptcy of a financially distressed manufacturer makes the existence of the M&R uncertain in the future, rational buyers devalue the capital and reduce their demand. This downward shift in the demand curve generates an indirect cost of financial distress, lowering revenues and profits and making the manufacturers further distressed. To quantify the indirect cost, I construct a capital market model with a secondary M&R market, using a dynamic game framework, and incorporate it into a stochastic dynamic partial equilibrium model with heterogeneous capital manufacturers making decisions on debt financing and default. The results indicate the real effects of the risky M&R are phenomenal: in the long run, the aggregate investment, consumption production, and hours worked all decrease by 10%, and the default rate increases by 15%. Extra experiments on M&R technology are conducted, ensuring the significance of M&R both in the long run and over the business cycles.
4. Heterogeneous Unemployment Risk over the Business Cycle, Latest Version
I construct a DSGE model to study the dynamics of income and consumption over the business cycle when households face heterogeneous unemployment risks and heterogeneous channels of insurance. To match the real-world pattern, I model the unemployment risks and the availability of insurance channels as functions of households' idiosyncratic labor productivity. The model can generate a realistic Gini coefficient and produce non-monotone distributional effects, which are consistent with the data but haven't been studied in previous literature. In terms of wealth loss, wealthy households suffer most due to the drop in the price of financial assets during the crisis, while in terms of consumption drops, the middle class suffers more than the upper class and the lower class. The mechanism is that the middle class, who cannot access the complete financial market, lacks tools of self-insurance, but when getting unemployed, only a small proportion of the income will be compensated by government-supported unemployment insurance.
5. VC Investment Network in China: Formation and Influence on Future Performance (Draft available upon request)
with Zhi Li
This paper studies the investment network formed by Venture Capital (VC hereafter) joint investment in the domestic market in China. Using VC investment data in China from 1999 to 2013, we investigate the formation of domestic VC investment networks, factors that explain the individual network attributes, and the impact of investment networks on VC investment performance. The empirical results show that the previous cooperation experience and similarity in backgrounds lead to VC joint investment; the individual experience of VC has the most significant explanatory power for network attributes. In terms of investment performance, the role of the VC network is to bring more investment projects in the future rather than to increase the probability of investment projects IPO, and this result is consistent with the "structural hole" theory. In addition, we find that from 1999 to 2013, the advantages of foreign VCs have diminished. Once focusing on the later periods of our sample, the foreign VCs showed no superior performance in improving their VC investment projects after controlling the network characteristics.
6. Re-valuating the VC investment in China: Value Seeking vs Value Creating (Draft available upon request)
with Zhi Li
VC-backed startups feature high growth potential in China. We apply matching methods and Heckman's sample selection model to check whether it is because of effective screening (selection hypothesis) or effective monitoring (value-added hypothesis). The empirical results show evidence for the co-existence of those two effects. Quantitatively, our "counterfactual inference" shows that the value created by VCs only accounts for 58%-71% of the overall premium. In contrast, the remaining 29%-42% of the value comes from the VC selection effect from the entrepreneurial enterprise itself. When including more VC characteristics, we find the market share is significantly negatively related to startups' IPO probability, which is contrary to previous studies. We argue that this result implies the investment strategy of VCs in China: to raise profits, VCs choose to increase the number of projects instead of improving their monitoring services.
Journal Articles
Computing Multi-region Competitive Prices for Hurricane-Related Insurance, with Dahui Liu, Linda Nozick, Meghan Millea, Jamie Kruse, and Rachel Davidson, International Journal of Disaster Risk Reduction, 2025, 120: 105383. (Link)
Insurability and government-funded mitigation: safer but costlier, with Dahui Liu, Linda Nozick, Meghan Millea, Jamie Kruse, Rachel Davidson, Joseph Trainor, and Caroline Williams, The Geneva Papers on Risk and Insurance-Issues and Practice, 2024: 1-16. (Link)
Unequal Economic Consequences of Coastal Hazards: Hurricane Impacts on North Carolina, with Dahui Liu, Ian Sue Wing, Brian Blanton, Jamie Kruse, Linda Nozick, and Meghan Millea, Environmental Research Letters, 2024, 19(10), DOI 10.1088/1748-9326/ad6d81. (Link)
Non-Linear Effects in Equity Value: Analysis Based on Continuation Value and Conversion Options, with Bing Deng and Yiming Wang, Journal of Financial Research (金融研究 ), 2015. (Working Paper Version)
Other Publication
The Impact of Marketing Capabilities on Image of Exhibition: Based on the Perceptions of Demand-Side, with Wang Jingna, Ye Chen, Jinlin Zhao, and Jing Han, ICTCHS 2014 (2014.6.29)
Work in Progress
Dynamics Computational General Equilibrium Model with Recursive Investment, with Ian Sue Wing
The Economic Consequences of Hurricane Disasters: A General Equilibrium Analysis, with Dahui Liu, Ian Sue Wing, Linda Nozick
Optimal Land-Use Policy under Natural Disaster Risks, with Qiyuan Huang, Linda Nozick
Housing, Migration, and Disaster Shocks: A General Equilibrium Framework, with Ian Sue Wing, Linda Nozick, Jamie Kruse, Meghan Millea, Dahui Liu
Understanding the Heterogeneous Consequences of Hurricanes: A DSGE Framework, with Ian Sue Wing, Linda Nozick, Jamie Kruse, Meghan Millea, Dahui Liu
Captive Finance in the Fore-Market, Maintenance, and Repair in the After-Market and Credit Shocks