Intermittency in firm bankruptcies from register data
By analysing the clustering of bankruptcy activity from administrative register data in the Netherlands over time, the intermittency of this process is investigated. Using methodology from survival analysis a statistical characterization of bursts is obtained: short active periods alternating with quieter periods. Using Statistics Netherlands microdata on Dutch business bankruptcies declared on court hearing-days from 2009 to 2024, combined with input-output tables from the Dutch national accounts, it is studied whether this clustering can be explained by differences between sectors, years, and inter-sectoral exposure. Sector differences explain most of the heterogeneity in the data, while year effects matter to a lesser extent and the inter-sectoral exposure variables constructed from the input-output tables further improve model fit.