Employer Losses and Deferred Compensation

David I. Walker is Professor of Law at the Boston University School of Law. This post is based on his recent paper.

Companies and their employees may choose whether to structure pay as cash or other currently includable and deductible compensation or as deferred compensation, including equity-based pay, which will be included in income and deducted in the future. It is well understood that taxes affect the attractiveness of deferred compensation relative to current compensation and that deferred compensation is relatively more attractive when employee tax rates are expected to be lower in the future than today, when employer tax rates are expected to be higher in the future, and when an employer can earn a greater after-tax rate of return on any compensation that is deferred.

In recent years, well over half of U.S. public companies have reported having a net operating loss (NOL) carryforward, reflecting deductions (including prior year NOLs) in excess of gross income. If significant, these NOLs reduce employer effective marginal tax rates (MTRs) and could make deferred compensation relatively more attractive by improving employer after-tax returns on deferred amounts and/or by reducing the value of current employer deductions relative to future employer deductions.

However, empirical studies that utilize variations in employer NOL positions in search of evidence that taxation affects the choice between current and deferred compensation have met with only limited success, which raises several questions. Are firms failing to consider joint tax minimization in the design of compensation programs? If so, is that failure a result of ignorance or laziness, or might it be rational? Alternatively, are researchers failing to detect tax sensitivity in compensation arrangements due to poor experimental design or the use of poor instruments?

The goal of this paper is to advance our understanding of the challenges that companies and researchers face in incorporating employer NOL positions into their compensation design analyses. The paper’s novel strategy is to bifurcate and unpack the two distinct ways in which NOLs may affect employer taxes and (generally) improve deferred compensation economics. First, NOLs may increase an employer’s after-tax rate of return on deferred sums. Second, NOLs often increase the value of employer deductions at deferred compensation payout relative to grant. I show that these two effects may work in tandem to boost the attractiveness of deferred compensation when an employer has a large NOL position, but that they may not. In some cases, employers deploy deferred funds in such a way that after-tax rates of return are less affected or unaffected by NOLs. I also show that while NOLs at grant tend to improve the economics of deferred compensation, this is not necessarily the case. In addition, even when grant date NOLs do improve the economics of deferred compensation, a larger NOL position at grant does not necessarily produce a larger driving force for deferral. In mathematical terms, the relationship is not always monotonic.

In short, while conventional wisdom holds that employer NOLs boost the economic attractiveness of deferred compensation, this paper both complicates and clarifies this picture. And this analysis leads to two main conclusions. First, it may not be irrational for firms to ignore grant date NOLs in making some deferred compensation decisions. Given the complexity of the relationship between NOLs and the economics of deferred compensation as well as unpredictable payoffs, it would not be surprising to find managers making decisions based on simple heuristics, such as statutory rates, rather than upon full-blown expected values incorporating the impact of NOLs.

Second, even if firms are sensitive to taxation in making compensation decisions, it is not clear that the research methods that have been deployed to test this sensitivity would detect it. Researchers often use NOL dummy variables, sometimes combined with a lack of profits, as proxies for low MTRs. These may (or may not) be good proxies for grant date MTR, but they are unlikely to be effective as proxies for the tax advantage of deferred compensation, which is not always monotonically related to grant date MTR. More sophisticated research techniques employ simulations of MTR, both at grant and payout, which, in theory, is an improvement. However, different simulations produce dramatically different estimated MTRs for the same firms in the same years. Moreover, recent work suggests that the input data used in some of these simulations is more error-riddled than previously suspected. It appears that we are a long way from being able to confidently test the relationship between employer tax rates and the use of deferred compensation.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.