1 11 !" ##$111 !"#$! % # # # % / ## % / # / # ! # % # ( / * # $ % # 01 # / ,’ # % # 01 /% * / %* / 2/* % ## % # ! # / / % # / / %&’(") * # * / 3*+ " -4 3 )() 3*+ " -4 3 2(%) "* #-4 1+ -* 5 - "2 65 7* ***0(+) /0 .!"()1$!,(. In a world characterised by uncertainty, a divergence of interests, a diversity of actors and implicit contracts, the concept of earnings management becomes increasingly important for companies. Earnings management reflects the idea that all measurable things can be rigorously managed. To reach the best compromise between shareholders and stakeholders, it is commonly accepted that some managerial discretion is necessary for earnings publication policy. The company directors seem to reconcile interests of various partners before determining the published results. This behaviour is called "earnings management" and depends on the relationships within the company. Schipper (1989) defined earnings management as "a deliberate intervention in the process of financial information presentation in order to capture personal gain". Earnings management corresponds to a set of discretionary actions (Teoh, Welch and Wong, 1998). To assess accounting manipulation, different models have been developed to calculate discretionary accruals, which are considered as a measure of the managerial discretion; these include the models developed by DeAngelo (1986), Healy (1985), Jones (1991), Dechow, Sloan and Sweeney (1995) and Teoh, Welh and Wong (1998). The discretionary accruals correspond to adjustments that can move from revenue accounting to accrual accounting. Thus, they correspond to the difference between net income and cash flows. However, these models imperfectly assess managerial discretion. They do not identify opportunist and altruistic behaviours of managers. An investigation into earnings management motivations enables corporate behaviour to be determined through the published accounts and, therefore, allows investors to make efficient decisions. There are various motivations of earning management: capital markets pressures such as compliance with analysts' forecasts (Dechow, Richardson and Tuna, 2003) or an increase in the issue price of shares (Teoch, Welch and Rao, 1998), a reduction in the default probability of debt (Dechow and Skinner, 2000), an increase in managers’ revenues (Holthausen, Laker and Sloan, 1995) and a decrease in political (Bartov, Gul and Tsui, 2000) and financing costs (DeAngelo, 1986). The relationship between corporate governance and earnings management has been emphasised by Carcello and Neal (2000) and Lobo and Zhou (2001). However, one of the most important motivations in earnings management is the desire to influence the financial market’s perception of the firm risk. Indeed, an increase in level of risk could be associated with high earnings management. In