Take a guess: Who do you think pays more in corporate taxes, companies with CEOs who favor the Republican party, or those with CEOs who tend toward the Democrats?

If you guessed the latter, you’re wrong, according to a paper that will be presented Aug. 7 at the annual meeting of the American Accounting Association in Washington, D.C. The study, which was led by researchers at the University of Arizona, finds that companies helmed by executives who lean Republican each pay an average of $12 million more in annual taxes than companies headed by Democratic-leaning managers. With a mean pretax income of $541 million for companies in the sample, the difference results in an extra 2.22 percent in taxes paid.

Yet why would these companies, the ones run by leaders who support a political party that is currently refusing to consider higher taxes, be the very ones to pay more? Because they tend to be more conservative, the researchers say, and “[p]sychology research shows that conservative individuals tend to be more risk-averse, fear losses, value financial security, and try to avoid uncertainty.” As a result, the researchers Dane Christensen and Dan Dhaliwal found, conservative nature trumps political ideology.

To perform their study, Christensen and Dhaliwal looked at all executives listed in the ExecuComp database between 1992 and 2008, used personal political contributions data from the Federal Election Commission to determine those executives’ political persuasion, and examined two common measures of tax avoidance over the years: each company’s effective tax rate (GAAP ETR) and cash effective tax rate (Cash ETR). “We find that when the top executive team’s political preferences lean toward the Republican party, their firms have significantly higher GAAP ETRs and Cash ETRs,” the two summarize in their paper.

Christensen and Dhaliwal make a convincing case that it is the conservative nature of managers, rather than any number of alternate interpretations, that explain their findings. Many skeptics, for example, are likely to question whether CEOs’ campaign contributions might have to do more with trying to influence legislation or corporate interests than their true political leanings. But by averaging contributions over a number of years, they say, motive-driven financial donations don’t have an effect on the results. In addition, Christensen and Dhaliwal write, when looking at personal contributions, the size of the amount is relatively small, and most executives consistently donate to one party.

The study also accounts for a range of other variables, from age, industry and gender to whether or not the company is a family firm (which tend to be more conservative) or uses a large amount of stock options (which produces a tax benefit), but they continue to end up with similar results: Companies run by corporate leaders who support Republican candidates pay slightly more in taxes.

What I find most interesting about the study is not which political persuasion leads to paying more to the IRS. (Although it is interesting, if not surprising, to see that of the 6,818 executives they examined, 64 percent leaned Republican, 31 percent leaned Democrat, and another 4 percent were classified as “neutral.”) Instead, what intrigues me most is that the values and mindset of the CEO—a leader who is rarely an accountant, and spends little of his or her time bogged down in tax minutiae—could have such an influence on how much taxes are paid.

Of course, as Christensen and Dhaliwal point out, CEOs might influence how much to spend on outside tax advisers, how well to compensate the company’s top tax managers, or how much to try to reduce a company’s tax burden. More important, however, they hire and shape an executive team that is likely to share their approach to uncertainty and losses. And they set a tone at the top that, whether intentionally or not, either encourages or discourages risk-taking throughout their organizations.

Leaders may think their personalities and political preferences can be separated from how they run their companies. But if Christensen and Dhaliwal are right, maybe they can’t.

Update: Two additional researchers, Steven Boivie (University of Arizona) and Scott Graffin (University of Georgia), also contributed to study. 

More from On Leadership:

Photos | Inspiring quotes from sports

Is Chris Christie the political leader nearest the center?

How to completely destroy an employee’s work life

View Photo Gallery: The best and the worst leadership moments of recent U.S. presidents.

Like On Leadership? Follow us on Facebook and Twitter:

@post_lead | @jenamcgregor | @lily_cunningham