Large multinationals based in the United States - among them General Electric, Pfizer, Apple and Citigroup - have been hoarding record amounts of cash overseas, mainly because of the 35 percent tax they would have to pay if they brought it back to the United States.
The American Jobs Creation Act of 2004 offered a temporary tax holiday that allowed firms to repatriate cash at about a 5 percent tax rate, but there were strings attached. The repatriated money was to be used only on permissible activities like research and development, capital expenditures and pension funding. It was not to be used for shareholder dividends or share repurchases.
The purported goal of the legislation was to create jobs, not simply to enrich shareholders at the expense of federal tax revenue. In recent years, companies have lobbied for another tax holiday.
Tax policy experts are suspicious of tax holidays, and most experts question the effectiveness of attaching strings to such legislation. Because cash is fungible, companies might be expected to use the repatriated money for permitted domestic activities that they would have conducted anyway, freeing up other cash to be used for dividends and stock buybacks. If companies merely reshuffle the use of cash without changing behavior, then the tax holiday amounts to a windfall to shareholders, not an effective economic stimulus.
The 2004 tax holiday brought back $312 billion in extraordinary cash dividends from foreign subsidiaries. How much of that cash was used for permitted activities, and how much for impermissible dividends and stock buybacks? A 2011 paper by Dhammika Dharmapala, C. Fritz Foley and Kristen J. Forbes, published in The Journal of Finance, estimated that 60 cents to 92 cents of every repatriated dollar was spent on shareholder payouts in 2005. The paper is Exhibit A in the case against future tax holidays.
A new paper by Thomas J. Brennan of the Northwestern University School of Law challenges that study and finds that, for the 20 companies that repatriated the most cash, 78 cents of every dollar was spent on permissible uses, and just 22 cents on impermissible shareholder payouts. Extending the analysis to 341 companies outside the top 20, Mr. Brennan estimates that about 40 cents of every dollar was spent on impermissible shareholder payouts, still much lower than the earlier estimate.
Mr. Brennan attributes the different findings to a change in research methodology. The Dharmapala paper uses econometric techniques to estimate the marginal increase in repatriation caused by the legislation, focusing on a certain subgroup expected to benefit from the legislation รข€" namely companies that face a low tax rate abroad and have a subsidiary in a tax haven. The Brennan study also uses statistical estimates, but it constrains those estimates with actual publicly available data on what each individual company spent on various activities. Using the data on actual share buybacks and dividends, Mr. Brennan shows that the Dharmapala estimate is too high.
The Brennan study is important for its explanation of how companies responded to the 2004 tax holiday. But it hardly endorses tax holidays as an instrument of international tax policy. Indeed, Mr. Brennan notes that, in a separate paper, he has made the case against temporary tax holidays.
It is important to remember the big picture: there is no evidence that the 2004 tax holiday created any jobs.
Mr. Brennan does not attempt to provide such evidence. Nor does he argue that spending on domestic activities is necessarily more effective in creating economic stimulus than dividends or stock buybacks would have been. After all, a shareholder who receives a dividend from Company A might reinvest the cash in Company B, an action that seems just as likely to create a new job as would spending by Company A.
Moreover, the Brennan study still finds significant amounts of shareholder payouts from repatriated profits, ranging from 20 cents on the dollar for the largest companies to 40 cents for smaller ones. It is sad, and unfortunate, that our expectations of corporate managers have fallen so far that 20 to 40 percent noncompliance might be viewed as a success.
The fact that the managers only incompletely evaded the purpose of the legislation should not be read as an endorsement of tax holidays.
