Last week, Harvard Law professor John Coates IV posted a paper on the Harvard Corporate Governance and Financial Regulation blog that attempts to pin down the cost to shareholders of corporate political activity. The provocation for the study is the Citizens United case, in which the Supreme Court, as Coates writes, "relaxed the ability of corporations to spend money on elections." In doing so, the court, he writes, "rejected a shareholder-protection rationale for restrictions on spending, in part on the ground that shareholders are generally capable of defending their own interests through 'corporate democracy.' "

Now this is, frankly, a relatively arcane paper in a field most folks would rather avoid. But the subject -- corporate political activities, particularly in campaign funding -- is pretty inflammatory, particularly after Citizens United. And perhaps more importantly, Coates admits in this paper to the limits of governance research that is refreshing if mildly disconcerting.

Coates' paper concludes that between 1998 and 2004 companies that demonstrated "observable" political activity -- lobbying, participation in political action committees, campaign donations -- were "negatively correlated" with shareholder-friendly governance. And shareholder-friendly governance, in turn, is "strongly correlated" with firm value. In plain terms, Coates argues that corporations that engage in active politicking tend to be less friendly to shareholders and worth less in the market.

Coates marches through a dense thicket of factors to reach his conclusions. He tries to get some sense of "corporate political activities" (he calls it CPA) by measuring what he can: the amount of money spent on obvious political activities such as lobbying or campaign donations. Then he tries to generate a kind of score for these companies based on "shareholder friendliness," which he derives from some 24 corporate governance provisions tracked by RiskMetrics Group Inc. Coates also tosses in some CEO compensation material and concludes that higher pay tends to correlate with corporate political activity. He then tentatively suggests that political activity represents a way "for managers with weak shareholder rights to extract private value for themselves, in a manner that is hard to observe or control."

All this is pretty murky, and the latter conclusion seems to be a leap, though one of the real values of Coates' papers is his sweeping review of the literature on corporate political activities. Coates himself recognizes the limitations of many of these studies: "All empirical studies of CPA are challenged by the fact that only certain kinds of CPA are required to be disclosed, even by public companies. ... Any research claiming to have assessed the aggregate amount of political activity in businesses or corporations should be viewed skeptically." And Coates even hedges his own conclusions:

"The foregoing empirical analysis does not definitively prove that political activity harms shareholders. ... But, given the complexity of the relationship between corporate governance and social welfare, and the difficulty of finding clean empirical tests of competing hypotheses on how governance structures work, such proof is never likely to be forthcoming, and policy must be built on a combination of theory and suggestive evidence. In this respect, research on the effects of CPA -- and the findings reported above -- are no less reliable than the broader body of corporate governance research on the relationship between governance and values."

What are we to make of this? That's a pretty sweeping statement about the empirical frailties of basic governance research. The problem is that in the larger world of governance practice, such reservations as Coates offers tend to disappear, flattened beneath a doctrinal hammer. And, as Coates says, policy must be built. Despite these caveats, Coates himself constructs a kind of chain of correlations, each of which manages to suggest the other: high pay, hostility to shareholders (or to governance doctrine), political activities, value. He seems to be attempting to include corporate political activities as a symptom of weak shareholder democracy and thus something, like the poison pill or high CEO pay that should be eliminated or restrained for the welfare of shareholders.

In fact, Coates may well be right; in many cases excessive political activity in a corporate setting may not only distort democratic politics but also suggest that there are deeper corporate problems (psychologically, you can easily imagine companies with growth problems turning to Washington for help). But, of course, each situation in the real world is unique; and "policy" requires a very broad brush of correlations. So, if you want a policy, you have to use the ammo you can get. - Robert Teitelman