"Using a detailed data set of employee stock option grants, we compare observed stock-option-based pay plans to hypothetical cash-only or restricted-stock-based plans. We make a variety of assumptions regarding the possible benefits of options relative to cash or stock, and then use observed option grants to make inferences regarding firms' decisions to issue options to lower-level employees. If the favorable accounting treatment is the sole reason underlying firms' choices of options over cash-only compensation, then we estimate that the median firm in our data set incurs $0.64 in real costs in order to increase reported pre-tax income by $1. This figure is several times larger than the willingness-to-pay for earnings reported by Erickson, Hanlon, and Maydew (2002), who study firms that (allegedly) commit fraud in order to boost earnings. If, on the other hand, firms' option-granting decisions are driven by economic-profit maximization then observed stock option grants are most consistent with explanations involving attraction and retention of employees"--National Bureau of Economic Research web site.