My research asks a single question in changing settings: when formal finance does not reach people, what explains it, and what can institutions do about it?
The theoretical root is information asymmetry. In capital markets it appears as idiosyncratic volatility and liquidity premia; in policy finance as market failure and moral hazard; in small-business lending as thin credit files; in community finance as the case for relationship banking. The empirical signature is heterogeneity — across regions, sectors, income groups, and countries, the same policy rarely produces the same result.
Credit Guarantee Systems
Public credit guarantee schemes are the main instrument by which states extend credit to firms that banks will not serve alone. My work asks how much guarantee supply such a scheme can sustain and where its fiscal limit lies.
The recurring finding is that the binding constraint is the subrogation rate, not the size of the endowment. Work on Korean regional guarantee foundations (2014, 2015) derived the endowment and default-payout conditions required to meet latent demand; the same framework applied during the COVID-19 expansion (2021) identified a threshold range beyond which existing fiscal capacity cannot absorb losses.
A second, counter-intuitive result concerns operating principles. Guarantee accidents track past guarantee volume more closely than the business cycle, which implies that steady and adequate supply outperforms counter-cyclical flexibility — the opposite of standard policy instinct.
Current work extends this to international comparison, including commissioned research on optimal leverage for the Korea Credit Guarantee Fund and the Korea Technology Finance Corporation (2024).
Financial Inclusion and Community Finance
This strand shifts the unit of analysis from borrowers to the institutions that lend to them. Using a twelve-year panel of Korean municipalities (2010–2021), I constructed a sub-national financial inclusion index and linked it to the management performance of community credit cooperatives.
Two findings stand out. First, inclusion and soundness are not in conflict: institutions with stronger profitability, stability, and asset scale also serve their communities more widely. Second, and more consequential for policy, the capital region and the rest of the country behave in opposite directions. Branch closure improves efficiency in the capital region but damages both inclusion and efficiency elsewhere; expanded real estate collateral lending raises profitability only in the capital region.
The policy implication is that uniform national restructuring rules deepen financial exclusion outside metropolitan areas.
Credit Assessment and Alternative Data
If exclusion follows from the absence of assessable information, can new information sources reverse it?
The answer turns out to be conditional. Unstructured non-financial information adds little beyond commercial bureau scores (2016). But structured external data improves prediction meaningfully — commercial-district grades combined with bureau scores outperform bureau scores alone, with the largest gains in the low-score range (2017). Later work extends the information base to personality traits and financial behavior (2019) and applies gradient-boosted trees with Shapley value decomposition to preserve interpretability alongside predictive accuracy (2022).
The general lesson is that what matters is not the volume of data but its design.
Cooperative Finance and Development
My current work places Korean community finance within the global lineage of cooperative banking rather than treating it as a national success story.
The Raiffeisen model (1849) was transplanted widely and failed in most places, including documented cases in Ireland and Bengal. Korean village credit cooperatives, founded in 1963 and later merged with the Saemaul Undong, are among the rare cases where the model took root. Framed as a Korean case study this is area studies; framed as "why did the transplant fail in Ireland and Bengal but work in Korea?" it becomes institutional economic history, with comparison cases already established in the literature.
A related concern is the absence of a standard English-language reference on Korean community finance, which produces recurring factual errors in international scholarship — the founding date and the relationship to the Saemaul Undong are misreported with some regularity.
Development Finance and International Cooperation
Recent work extends the inclusion agenda beyond Korea: the growth effects of remittances in Southeast Asia, which are positive but diminish non-linearly with income level (2023), and the transfer of Korean commercial-district information systems as a form of data-infrastructure ODA (2022).
The PSPS Alumni Panel is the current centerpiece. Global Findex, the IMF Financial Access Survey, and OECD DAC CRS all measure whether institutions exist; none measures whether they function. Working through PSPS alumni who serve as front-line practitioners in sixteen countries, the panel is designed to track the gap between law and implementation on a semi-annual basis over ten years, producing evidence on announced-but-unimplemented policy, actual collection practices, and the final incidence of loss under risk-sharing schemes.