Replication package for: Beyond the Pandemic Reversal: An Anatomy of Imbalance-Correctors in Non-Financial Corporate Credit

Replication package for the paper "Beyond the Pandemic Reversal: An Anatomy of Imbalance-Correctors in Non-Financial Corporate Credit" by Alejandro Jara (Central Bank of Chile).

Abstract: Contractions in credit to the non-financial corporate (NFC) sector after 2022 have been markedly uneven across economies. This paper classifies the cross-section into boom-reversers—where the post-2022 decline mainly unwinds a pandemic-era surge—and imbalance-correctors—where it also works off an older build-up accumulated during 2010–2018. Combining three BIS data products—Total Credit Statistics, Locational Banking Statistics and International Debt Securities Statistics—I split aggregate NFC credit into four channels (domestic bank, domestic non-bank, external loans and external bonds) and identify 11 imbalance-correctors (26%), including Chile, Canada and several financial centres. Three findings emerge. First, the pre-pandemic overhang is associated with deeper contractions: each additional percentage point of prior accumulation is associated with 0.24 pp of additional contraction depth (p = 0.002, OLS cross-section), and a permutation test corroborates this association (exact p = 0.052, 50,000 draws). Second, the correction is domestic-bank-led, partially offset by substitution into non-bank and bond-based channels: aggregate NFC credit does not mean-revert because bank and non-bank components move in offsetting directions; bank credit converges (half-life 21 quarters) and non-bank credit faster still (18 quarters), with imbalance-correctors converging 2.7× faster. Third, only 2 of 11 corporate imbalance-correctors also qualify on the household side (18% overlap), confirming a sector-specific mechanism. Capital-account openness on the outflow side is the strongest structural correlate, and its interaction with local-currency debt share is significant (p < 0.05), consistent with the spare-tire hypothesis.