Hedge accounting incentives for cash flow hedges of forecasted transactions

Abstract

US GAAP as well as IAS (IFRS) contain specific accounting regulations for hedging activities. Basically the hedge accounting rules ensure that an offsetting gain or loss from a hedging instrument affects earnings in the same period as the gain or loss from the hedged item. However, due to the way hedge accounting rules are set up, their application turns out to be an option rather than an obligation for firms. Recognizing this fact, the paper analyses corporate incentives for hedge accounting in the presence of a moral hazard problem. We consider a two-period LEN-type agency model with a risk averse agent and a risk neutral principal. The principal decides upon hedging and motivates effort through an incentive contract based on accounting income. We find that in such a setting the principal strictly prefers hedging as opposed to no hedging. Whether he prefers hedge accounting or not depends on how the firm's overall risk exposure is allocated over periods. If risk exposures differ largely over periods the principal prefers hedge accounting. Otherwise no hedge accounting is preferred.

Acknowledgements

Helpful comments from John Christensen, Alfred Luhmer, Ulf Schiller, Jens Robert Schöndube, Dirk Simons, Marco Trombetta, Alfred Wagenhofer and two anonymous referees as well as from participants at the Accounting Workshop 2004 in Vienna and the Workshop on Accounting and Economics in Frankfurt 2004 are gratefully acknowledged. An earlier version of this paper was titled ‘Hedge Accounting versus No Hedge Accounting for Cash Flow Hedges’.

Notes

1. SFAS 133 was published in 1998. It revised hedge accounting rules comprehensively superseding a number of previously issued standards. SFAS 138 and 149 contain some minor amendments with respect to very specific problems. The latest revisions of IAS 32 and 39 were published in December 2003. Since then some minor amendments have been integrated into IAS 39 throughout 2004 and 2005. Further Exposure Drafts are still subject to discussion.

2. See Trombley (2003, p. 46).

3. See Melumad et al. (1999, p. 279).

4. Such intention is documented, for example, in the IASB preface P-4 (13).

5. Trombley (2003, p. 33).

6. See Trombley (2003, p. 33).

7. Different risk exposures over periods possibly stem from different products or parts of products produced in both periods. Alternatively one might assume that production processes last two periods and that either early steps or final steps in the process are subject to higher risk.

8. Contrary to the principal we assume that the manager is not perfectly diversified. This is reasonable to assume as managers typically face compensation and future career risks that account for a large fraction of their overall risk exposure and are difficult to diversify.

9. See Demski (1998, p. 265).

10. See, for example, Spremann (1987, pp. 17–22), Holmström and Milgrom (1991, p. 29), Feltham and Xie (1994, pp. 431–433).

11. Note that this contract design incorporates a contract that compensates the agent based on overall income, π₁ + π₂, as the principal is free to choose s ₁ = s ₂ = s.

12. For example, Murphy (1999, pp. 2500–2502) states that virtually all large US corporations use bonus plans as part of management compensation and that earnings are the most heavily used performance measure in these plans.

13. Details are given in Appendix A.

14. The properties of the LEN model ensure that first-order conditions are sufficient to characterize the agent's effort choice. See, for example, Datar et al. (2000).

15. This is ensured by choosing the fixed payment S appropriately. The fixed payment, however, has no role in the rest of the analysis and we do not present explicit solutions in what follows.

16. See Appendix B.

17. In fact with y = 0.1 Δ(k) is a bell-shaped function, too. The maximum of this function, however, is reached for k < 0 and is therefore outside the feasible domain.

18. Note that the conditions developed in what follows are necessary but not sufficient for NHA to be preferred to HA. The point here is to convey the intuition of the results.

19. Whenever we get and therefore (5) is positive.

20. Within the analysis not much attention has been devoted to the role of rσ ². The right-hand side of (7′), however, is increasing in rσ ² and thus higher rσ ², other things equal work in favour of HA.

21. See, for example, Gebhardt et al. (2004) who analyse hedge accounting within the banking industry.