This paper uses firm level data from the World Bank Enterprise surveys conducted in 2019 and from the COVID-19 follow-up surveys conducted in 2020 in eight European countries to investigate the link between firm characteristics before the pandemic and firm survival until 2020. For the first time the marginal effects of firm characteristics are computed by a new machine-learning estimator, Kernel Regularized Least Squares (KRLS), which makes no restrictive assumptions regarding the functional form of the empirical model used. A comparison with results from a standard parametric approach, Probit regression, reveals important differences.
Joachim Wagner (Thu,) studied this question.