Researchers such as Miller and Scholes (1982), Smith and Watts (1982), and Scholes and Wolfson (1986) have assessed analytically the impact of taxes on different forms of compensation. The role that taxes play in compensation plan choice, however, has not been investigated empirically. In order for accountants to accurately measure and report the impact of firms' transactions, it is desirable that they have as much information as possible regarding the factors that influence the transactions. Lewellen et al. (1987) concluded that an understanding of firms' investment and financing decisions may require knowledge of their executive compensation plans. By investigating the relative importance of incentive and tax considerations in the design of executive pay packages, our understanding of why various compensation techniques are employed should be increased.
The study described in the following pages tests for an unexpected increase in the current compensation of chief executive officers (CEOs) pursuant to the 1986 Tax Reform Act (TRA '86). A methodology developed by McGahran (1988) is used to test for unexpected increases in current compensation. A regression model is constructed to predict the change in CEOs' current compensation for the years 1986, 1987, and 1988. Actual changes in current compensation for each year then are compared with expected changes (from the prediction model) to identify unexpected changes in the CEOs' current compensation. The objective is to test for changes in current compensation that are not explained by macroeconomic and firm-specific factors included in the model. Detection of significant unexpected increases in current compensation is consistent with a tax explanation for the change. Failure to detect significant unexpected increases is consistent with the view that the tax considerations do not have a great enough impact to overcome the incentive factors that favor deferred compensation.
The study also has tax policy implications. To the extent that TRA '86 increased the tax benefits of current compensation methods relative to deferred compensation methods, it can be argued that TRA '86 had undesirable tax policy consequences. This is especially significant in light of successive tax acts since 1982 eroding the ability to defer compensation although this may have been mitigated to a degree by the $1 million dollar limitation on deductibility of current compensation.
The remainder of the study is organized as follows. Section 2 reviews the previous literature and develops a theoretical base from which hypotheses are derived. Section 3 describes the research methods. Section 4 analyzes the test results. Section 5 contains the study's conclusions and limitations.
TAX EXPLANATION OF COMPENSATION PLAN CHOICE
The tax explanation posits that the tax effects of different executive compensation techniques explain or influence their use. Researchers have developed various analytical models that explain why a particular compensation technique is preferable to another, given the tax laws at the time. They do not always agree, however, as to when, how, and to what extent taxes play a role in compensation plan choice.
In applying the tools of financial economics to various compensation techniques, Miller and Scholes (1982) attempted "...to distinguish schemes that seem intended mainly to share tax benefits from those, of greater interest to economists, that appear designed as incentives to make the real pie bigger (p. 179)." They reasoned that tax-disadvantaged schemes must have compensating non-tax benefits such as incentive or incentive-signaling benefits to justify their use. Based on the results of their analysis, however, they drew the general conclusion that, under the tax rules in effect at that time, deferred compensation techniques were tax-preferred to current compensation techniques in most situations. They even speculated that labor economists would be disappointed in their not having found more deferred compensation arrangements for which tax-savings motives could be ruled out. …