Johannes' website

The auditor and dividend payout in small private firms: Evidence from a natural experiment

We use register data on private Norwegian firms and random variation induced by the introduction of a policy allowing small private firms to forgo the use of an auditor in submitting their financial statements in order to causally estimate the informational effect of the auditor on dividend payout. Identification is obtained by a regression discontinuity around the arbitrary cut-offs for the policy, while propensity score matching is used to create a balanced synthetic control. In addition we use a difference-in-difference design, making full use of the panel of data. We consistently find that the absence of an auditor led to a significant increase in dividend payout from small firms.