Résultat de la recherche
Essays on the relationship between stock prices and accounting variables
Starica and Marton (2019) soutiennent qu’une méthode d’estimation non linéaire basée sur des algorithmes de Machine Learning simples et flexibles peut résoudre plusieurs problèmes de modélisation économétrique afin d’évaluer la relation entre les cours des actions et les bénéfices comptables. Dans cette thèse, nous utilisons leurs avancées méthodologiques pour aborder trois problèmes de comptabilité financière. Premièrement, nous utilisons la technique d’estimation qu’ils proposent pour mesurer l’association que les cours des actions ont avec les bénéfices comptables et les flux de trésorerie, respectivement. Deuxièmement, nous suggérons une explication selon laquelle l’association entre les cours des actions et les bénéfices comptables semble avoir diminué au cours des deux dernières décennies. Dans cette thèse, nous utiliserons le terme anglais “accruals” pour indiquer la di_érence entre les bénéfices comptables (traduits comme “accounting earnings”) et les flux de trésorerie (traduits comme “operating cash flows”). Nos résultats suggèrent qu’un changement de corrélation entre accruals et flux de trésorerie est la raison la plus probable du phénomène mentionné ci-dessus, et, par conséquent, les flux de trésorerie sont devenus au moins aussi importants que les bénéfices comptables au cours de ces vingt dernières années. Troisièmement, nous cherchons à savoir si le degré d’association entre les cours des actions et les bénéfices comptables dépend de la qualité de la comptabilité d’exercice. Nous observons qu’en moyenne, les bénéfices comptables obtenus à partir d’une comptabilité d’exercice de meilleure qualité sont plus pertinents pour l’évaluation, parmi les entreprises dont les cours des actions sont comparables, ansi qu’une incertitude opérationnelle et des “timing-errors” similaires. ABSTRACT : Starica and Marton (2019) argue that a nonlinear estimation method based on simple, flexible Machine Learning algorithms can solve several econometric modeling issues in order to assess the relationship between stock prices and accounting earnings. In this thesis, we use their methodological advancements to address three problems in financial accounting. First, we use the estimation technique they propose to measure the association that stock prices have with accounting earnings and operating cash flows. Second, we suggest an explanation behind the stylized fact that the association between stock prices and accounting earnings has decreased in the last two decades. We argue that a change in the correlation between accruals and cash flows is the most likely reason for this phenomenon, and, consequently, our results suggest that cash flows have become at least as relevant as earnings in these last twenty years. Third, we investigate whether the degree of association between stock prices and earnings depends on the quality of accruals. Our findings suggest that, on average, higher-quality earnings are more pertinent to valuation among firms that have comparable prices, operating uncertainty, and size of timing errors.
The ebbing of accrual accounting
This paper investigates the evolution of accruals ability to improve the information content of earnings over cash flows, as reflected in market prices. We documented a pronounced decline in the price association of earnings relative to cash flows. We find that, after year 2000, cash flows have informed prices at least as much as earnings. We explain the decline in earnings price association through the decrease of the timing role of accruals related to the growth in the frequency and the magnitude of non-timing accruals (Bushman et al. 2016). Our results imply that this evolution of accruals has severely limited the information content of earnings and, consequently, significantly weakened the pertinence of accrual accounting to prices.
New insights into equity valuation using multiples
The thesis focuses on the equity valuation using multiples. Based on the notion of stochastic dominance, a relative valuation framework for comparing accuracy is developed and employed in the thesis.
In the first paper, we investigate the feasibility of a relative valuation framework for tracking developments of value relevance of earnings and book values across time.
The second paper focuses on the performance of EV/EBITDA, a multiple that gained in popularity with practitioners during the last decade, relative to that of the traditional multiple P/E.
The last paper explores the possibility of applying non-linear methods to improve the precision of multiple valuation.