Looked at quarterly conference calls between managers of traded firms and analysts. They then measured how much time during the conversation was dedicated to talking about political issues and risks. The idea is that with firms who face relatively higher levels of risk and uncertainty due to political changes (say, a regulatory reform), either the analysts will ask or the managers will voluntarily dedicate more time to discussing the political environment. In other words, the ratio of political to non-political conversation in these conference calls is a proxy for the political risk faced by a specific firm.
Once managers who frequently alluded to political risks hang up the phone, do they behave differently than managers who were troubled less by political turmoil? The answer is yes, in more ways than one. First, firms that face more political risk tend to cut back on hiring and investments. This result was more pronounced when the specific political risks that were mentioned were government reform, health care, and the environment.
To be sure, there could be some cheap or manipulative talk going on. That is, managers may mention political risk in the conference calls merely as an excuse for steps they would have taken regardless of the political environment; or they may be angling for political favors. While Hassan and co-authors find indications that such dynamics are indeed in play in some instances, the overall wealth of findings strongly suggests that political risk and uncertainty do play a real and not insignificant role in firm behavior.
To this end, we adapt a simple pattern-based sequence-classification method developed in computational linguistics (Song and Wu, 2008; Manning et al., 2008) to distinguish between language associated with political versus non-political topics. For our baseline measure of overall exposure to political risk, we use a training library of political text (an undergraduate political science textbook and text from the political section of newspapers) and a training library of non-political text (an accounting textbook, text from non-political sections of newspapers, and transcripts of speeches on non-political topics) to identify two-word combinations (“bigrams”) that are frequently used in political texts.
For our topic-specific political risk measure, we similarly use training libraries of text concerned with eight political topics (e.g., “economic policy & budget,” “environment,” and “health care”), as well as the political and non-political training libraries mentioned above, to identify patterns of language frequently used when discussing a particular political topic.