This thesis investigates dirty surplus accounting practice in different countries and its implications for performance measurement and equity valuation. First, I examine the characteristics of dirty surplus accounting permitted by accounting regulation and reported by companies in four countries: France, Germany, the U. K and the U. S during the period 1993 to 2001. I find that dirty surplus flows are negative on average. I also find substantial cross-country variation in dirty surplus accounting both in accounting rules and in companies' reporting practice. Dirty surplus accounting seems more severe in France and Germany than in the U. K and the U. S. Second, I analyse the implications of dirty surplus accounting and crosscountry variation therein for accounting-based measures of abnormal performance. I find that the omission of dirty surplus flows creates inaccuracy in abnormal performance measurement for all classes of dirty surplus flows and across the four accounting regimes studied. Bias in abnormal performance measurement is largely caused by goodwill-related flows. Third, I explore the valuation implications of dirty surplus accounting. I demonstrate that the residual income valuation model (RIVM) and the abnormal earnings growth model (AEGM) should yield identical intrinsic value estimates provided there is consistency in projections of accounting numbers. Accordingly, omission of dirty surplus flows from these projections results in identical valuation error in both models. I then perform empirical tests of the relationship between valuation errors and dirty surplus flows, both in terms of bias and inaccuracy. Only in the case of the U. S. do I find some evidence of such relationship. For this country, I also find evidence of industry differences in the relationship between financial and non-financial companies. Finally, results suggest cross-country differences in the relationship between the U. S. and the other three countries, particularly in terms of inaccuracy.