Accounting Information as Political Currency

It is well known that firms contribute money to politicians. It is also widely held that such money, in the form of campaign contributions and lobbying expenditures, is used to buy access to and/or favors from politicians. Firms and politicians establish relationships with one another and the value to firms of such relationships likely increases over time. When a politician with a well-established relationship to a firm faces a tough election prospect, it is in the firm’s interest to secure that politician’s future. One obvious way to do so is to make further monetary contributions directly to the politician’s campaign. A priori, direct monetary contributions are not the only channel through which firms can deliver benefits to candidates during political campaigns. In a recent working paper, Sugata Roychowdhury of MIT and I investigate whether political contributions can take a non-cash form, specifically (accounting) information. In other words, we investigate whether (accounting) information can be used as political currency?

Our setting is the US congressional election of 2004, where outsourcing of US jobs was a campaign issue. Firms engaged in outsourcing activities had incentives to ensure that political candidates they were affiliated with did not suffer from negative media due to the outsourcing. These incentives were likely to be strongest when the candidates were in competitive races. We test whether outsourcing firms understated profits in the period leading up to the 2004 election, in circumstances where the firms’ affiliated candidates were in competitive races. Understating profits can help deflect attention away from the firms’ outsourcing activities, and thus spare the candidates considerable embarrassment. We find that outsourcing firms donating to congressional candidates in closely watched races managed their earnings downwards in the two quarters immediately preceding the 2004 election. We find no evidence of downward earnings management among outsourcing corporations donating to congressional candidates not in closely watched races.

Ceteris paribus, if donors’ downward earnings management is effective in deflecting attention away from outsourcing, thus sparing candidates from negative media, we expect such candidates to do better in the election (than the average candidate). In regression tests that control for likely determinants of election outcomes, we find vote shares for candidates are increasing in the extent of their corporate donors’ downward earnings management. Overall, our findings are consistent with firms managing accounting information in circumstances where this is likely to benefit allied politicians. The evidence is consistent the hypothesis that accounting information can be used as political currency.

You can read the entire paper here and an interview with me over the paper here.