Accounting Information as Political Currency
Executive Summary — The study of accounting and the political process has long been viewed through the political cost hypothesis, the basic premise of which is that firms manage earnings in order to extract first-order benefits (or avoid first-order costs) from regulators. This paper develops and tests a distinct, yet likely, complementary hypothesis: Firms manage reported earnings in order to supply first-order benefits to regulators. Focusing on Democratic and Republican candidates in congressional races in 2004, Ramanna and Roychowdhury test whether the management of accounting information is in some circumstances akin to a political contribution from firms to politicians: in other words, whether accounting information can be used as political currency. The authors predict and find that identified corporate donors to candidates in closely watched races in 2004 managed information related to outsourcing, a hot-button issue in those races. Key concepts include:
- While corporate donors in general do not exhibit evidence of downward earnings management, corporate donors to candidates in closely watched congressional races exhibit significant evidence of downward earnings management in the second and third quarters of 2004.
- The evidence of downward earnings management is stronger for firms likely to have greater outsourcing activities.
- These findings are consistent with firms managing accounting information in circumstances where this is likely to benefit allied politicians, thus supporting the idea of a "political currency" hypothesis.
We test whether accounting can be used as political currency. Our setting is the US congressional election of 2004, where outsourcing of US jobs was a campaign issue. We find that the largest corporate donors to principal candidates in closely watched congressional races manage earnings downwards in the two quarters immediately preceding the 2004 election. We find no evidence of such downwards earnings management among corporate donors to candidates in all other congressional races. Election outcomes for candidates are also systematically associated with the extent of donors' downwards earnings management in closely watched races, but not all other races. The findings are consistent with firms managing accounting information in circumstances where this is likely to benefit allied politicians.