Vater, Tim
(2022).
Three Essays on Earnings Forecasts and Their Relation to Bankruptcy Risk and Earnings Management.
PhD thesis, Universität zu Köln.
Abstract
This thesis examines model-based earnings forecasts and their relation to firm-specific bankruptcy risk and earnings management. It consists of the following three essays. The first essay is based on the working paper “Analyzing Expected Accounting Losses in the Context of Bankruptcy Prediction” co-authored by Simon Wolf. We analyze the role of expected accounting losses in public firm bankruptcy prediction. To form expectations, we estimate firm-specific expected profitability distributions following Chang et al. (2021). We find that (i) an expected loss dummy, (ii) the expected probability to realize future losses, and (iii) the expected probability of future losses depleting current cash balances (i.e., illiquidity proxy) significantly improve forecast accuracy. Moreover, we provide evidence for a size effect in the relation between expected accounting losses and bankruptcy risk. The second essay is based on the working paper “Using the Expected Profitability Distribution to Predict SME Bankruptcy”. I examine how information imbedded in the expected profitability distribution, particularly in the area covering expected losses, improves SME bankruptcy prediction. Using a sample of German SMEs, I estimate firm-specific expected profitability distributions and compute three expected loss measures. I find that (i) the expected probability to realize losses and (ii) the expected probability of future losses consuming current cash balances (i.e., illiquidity proxy) significantly improve SME bankruptcy prediction accuracy, whereas (iii) the expected probability of future losses exceeding book equity (i.e., overindebtedness proxy) does not. The third essay is based on the working paper “The Relation Between Earnings Management and Model-Based Earnings Forecast Accuracy” co-authored by Luca Brunke. We analyze the effect of earnings management on model-based earnings forecast accuracy. We provide evidence that firms with a higher level of earnings management exhibit larger forecast errors, i.e., forecast accuracy is lower. Further, we show that accounting for the level of earnings management in earnings forecast models increases forecast accuracy, that then translates to more reliable ICC estimates. In conclusion, this thesis shows that information about expected profitability should be considered in public firm and SME bankruptcy prediction models, and that firm’s earnings management should be taken into account when forming these profitability expectations.
Downloads per month over past year
Export
Actions (login required)
 |
View Item |
