大型僵尸企业，小型受害者：基于人口层面数据的拥堵效应证据

Large Zombies, Small Victims:Evidence on Congestion Effects fromPopulation-Level Data Kyeongtae Lee 2026. 6 Large Zombies,Small Victims:Evidenceon Congestion Effects fromPopulation-Level Data Kyeongtae Lee∗ The views expressed herein are those of the authors,and do not necessarily reflect the officialviews of the Bank of Korea.When reporting or citing this paper,the authors’names shouldalways be explicitly stated. Contents I.Introduction····························································1 II.Data······································································3 III.Empirical results······················································13 IV.Welfare implications of zombie firm exit·····················23 V.Conclusion······························································25 Large Zombies,Small Victims:Evidence onCongestion Effects from Population-Level Data Most studies on zombie firms focus on listed firms—which account for only asmall fraction of the firm population—largely because of limited access to data onprivate firms.Using a novel dataset that covers virtually the entire universe of firmsin Korea,this study provides new evidence on the extent of zombification in the over-all economy,with particular emphasis on private firms,where systematic evidencehas been scarce. We show that the aggregate share of assets and financial liabilities held by zom-bie firms is driven by externally audited,and thus larger firms,while the economicpresence of private zombies remains limited and stable over time.Consistent withthe existing literature,we find economically meaningful congestion effects of zom-bie firms on healthier firms.Higher zombie prevalence is associated with lowerinvestment,employment,productivity,and profitability among non-zombie firms,and these effects persist over time.The adverse spillovers are substantially strongerfor smaller firms and for firms operating in non-tradable sectors.These effects areparticularly pronounced for investment and employment outcomes,with the hetero-geneity being especially strong among private firms.We further show that the exitof zombie firms can generate non-trivial gains in aggregate productivity and valueadded,while acknowledging that such reallocation may involve distributional trade-offs across firms,workers,and creditors. Overall,the findings suggest that the primary economic cost of zombificationlies not in its prevalence among small firms,but in the disproportionate burden thatzombie congestion imposes on smaller firms.This underscores the importance ofpolicies that facilitate corporate restructuring and timely firm exit,while limitingincentives for inefficient credit continuation. Keywords:Zombie firms;Misallocation;Congestion effects;Firm dynamics;Pro-ductivity;Firm exit;Private firms I.Introduction To the extent that small and medium-sized enterprises,which are not cap-tured in the data,account for a significant share of some sectors in the re-gion,vulnerabilities could be even higher. (IMF,Regional Economic Out-look:Asia and Pacific,May 2023) This observation is particularly salient in the study of zombie firms.Empirical evi-dence on zombification has been drawn predominantly from listed firms,while the vastmajority of small and medium-sized private firms—often accounting for most firms anda substantial share of employment—remain largely outside the scope of analysis.Thisraises concerns about external validity:it remains unclear whether conclusions drawnfrom listed firms extend to the broader economy,where private firms dominate in num-ber and play a central role in employment and value creation. Recent work has begun to incorporate private firms into the analysis(e.g.,Albu-querque and Iyer,2024),but available data sources remain subject to important selectionconcerns.Financial-statement datasets provided by credit bureaus tend to cover privatefirms that engage with external finance or maintain business relationships with largerfirms,implying that firms without borrowing needs or such linkages are systematicallyunderrepresented. A related Korean literature has examined the effects of zombie firms on productivity,pricing,accounting behavior,and restructuring,but the evidence remains limited in cov-erage and is often based on listed firms,externally audited firms,or specific sectors.1)Despite the policy emphasis on restructuring zombie firms in Korea,there is still limitedevidence on the prevalence of zombie firms among private firms,and it remains unclearto what extent the congestion effects documented in existing studies generalize to the broader firm population. This paper addresses these limitations by exploiting comprehensive administrativetax data from Korea’s National Tax Service,which cover nearly the entire universe offirms required to file corporate tax returns.Because reporting is mandatory,concernsrelated to sample selection and survivorship bias are substantially mitigated,allowingus to provide population-