【正文】 s milliondollar status and find that, relative to other firms, there has been an increase in the sensitivity of CEO wealth to shareholder wealth from 1993 to 1996 for milliondollar firms after controlling for other factors that affect CEO incentives. In sum, our results suggest that pensation mittees have taken the regulatory environment into account, and that these regulations have had a real economic impact on pensation and on the overall pay for performance relation. The reaction of firms with a pay for performance emphasis in their pensation policy to the currently changing economic conditions will generate interesting observations on the durability of this phenomenon. 。s effect, this finding is especially pronounced for milliondollar firms. Robustness tests show that the increased pay for performance relation is not driven by surging stock option grants, the definition of milliondollar firms, or CEO ownership. To further understand the effect of 162(m) and the disclosure regulation on our large sample of firms, we also examine the sensitivity of three measures of CEO39。P 500, Midcap, and Smallcap index firms. Compensation data are classified into the seven pensation categories identified in the Summary Compensation Tables of the proxy statements including salary, bonus, other annual pensation (including perquisites and amounts for reimbursed for payment of taxes), restricted stock awards, options or stocks appreciation rights (SARs), and longterm incentive plan payouts (LTIP), and “all other pensation.” Total pensation consists of all seven ponents reported in the proxy statement. The Disclosure Rules require firms to provide dollar values in the Summary Compensation Table for each of the ponents, except options and SARs, and ExecuComp reports the dollar values provided in the proxy statements. For options and SARs, firms are only required to disclose the number granted to their top executives in the Summary Compensation Table, however additional information relating to the stock option grants is available elsewhere in the proxy statement and ExecuComp separately putes the present value of the options granted to executives using a modified BlackScholes method. We retrieve accounting data from Compustat and stock returns from the CRSP files. For a smaller subset of our firms, we also collect information on the performance measures explicitly used by pensation mittees to evaluate the CEO from the pensation mittee reports in the proxy statements. For several of our tests discussed in the sections below, we pare pensation before 162(m) to pensation after 162(m). For the pre162(m) period, which we identify with the post1993 dummy variable in our tests, we only have 1991, 1992, and 1993 pensation data that are truly parable to the post162(m) data. This data limitation reduces the power of our tests. 5. Measures of changes in CEO wealth for changes in shareholder wealth Thus far, we have focused on the sensitivity of annual pensation flows (., salary, bonus, restricted stock, stock options, longterm incentive plan payouts, and other pensation) to firm performance. We have taken this approach because pensation critics have concentrated on these annual flows and because this is the flow that was apparently targeted by the regulations. Jensen and Murphy (1990), Hall and Liebman (1998), Murphy (1999), among others, show that CEO equity and option ownership in the firm accounts for the bulk of the sensitivity of CEO wealth to firm performance. While our prior results indicate that pensation structure and the pensation to performance sensitivity of milliondollar firms have been affected by the regulations, a crucial question to determine the economic impact of the regulations is whether these effects translate into a higher sensitivity of CEO wealth to shareholder wealth. 6. Ownership or shareholder targeting We perform various tests to examine whether our results are sensitive to measurement method or sample selection. Thus far we have focused on the subset of milliondollar firms because they are more likely to be affected by the regulations. We now examine whether the milliondollar effect is due to the fact that milliondollar firms are large firms with mon governance characteristics that make them more sensitive to shareholder pressures relating to pensation. Core et al. (1999), among others, document that governance structure affects CEO pensation. We explore the characteristics of milliondollar firms using an untabulated logistic model and confirm that milliondollar firms tend to be large firms with low CEO ownership. Firms with low CEO ownership are firms where the CEO has less control or where the board is more likely to face investor pressures to increase the alignment between CEO and shareholder incentives. Next, we estimate the models from Table 6 and add interaction variables for (i) firms with low CEO ownership (a dummy equal to one if the CEO39。s performance. As part of the Omnibus Budget Reconciliation Act of 1993, Congress enacted Section 162(m) of the Internal Revenue Code (1995), a provision that disallows deductions for nonperformance related pensation over one million dollars for the CEO and the other executives whose pensation must be reported in the proxy statement. The applicable employee remuneration does not include remuneration payable on a mission basis, pensation that is performancebased, and pensation under a binding written contract in effect on February 13, 1993. For the purpose of a deduction, pensation is classified as performancebased only if (i) the performance goals are determined by a pensation mittee prised solely of two or more outside directors, (ii) the performance goals under which the remuneration is to be paid are disclosed to the shareholders and approved by