At last Thursday's colloquium, Doug Shackelford presented a paper that uses extensive worldwide financial accounting data to look at how multinational entities (MNEs) are taxed around the world, compared to each other based on where they "reside" and invest, and compared to non-multinational public corporations.
The paper presented grist for the mills of both pro-worldwide and pro-territorial U.S. advocates. Aiding the latter, the relative tax burden on US MNEs has been rising relative to that on other countries' MNEs, and Japan, which currently has the highest such burdens, is switching to a territorial system. But aiding the former, the burdens on US companies are not greatly out of line with those for other OECD countries, and what's more US MNEs are shown as facing about the same worldwide tax burdens as purely domestic US companies. (This may reflect a draw in the battle between aggressive tax planning on the one hand and aggressive U.S. rules, e.g., limiting deductions, on the other.)
One clear caveat to keep in mind is that the data relates to companies that are acquired to make public income reports, as distinct from all businesses. So closely held and private firms are not in the data set. This might affect the significance of finding that purely domestic businesses and MNEs from the same country often face similar worldwide effective tax burdens. For example, it's long been assumed that MNEs are predominantly public corporations - although this is becoming less true over time - but the same certainly doesn't hold in the U.S. domestically. Private companies often are more aggressive and savvy in their tax planning, reflecting that managerial ownership significantly reduces agency costs.
Not surprisingly, discussants at the session agreed with the universal consensus (in my experience) among knowledgeable people that managers at publicly traded U.S. companies are far more interested in boosting reported earnings for accounting purposes than in actually saving tax. Hence a suggestion was made that U.S. companies' behavior might be far less affected, in some respects, by the actual U.S. rules for foreign source income than by the accounting rule permitting them to treat foreign earnings as "permanently reinvested" abroad, and thus as not potentially subject to ultimate U.S. tax on repatriation.
Absent that accounting benefit, the firms would have to book the hypothetical future U.S. repatriation taxes without any time value discount for the economic value of deferral. This might cause U.S. managers to start acting as if the U.S. had a purely worldwide system even though, in effect, it is actually partly territorial. By contrast, with the "permanently reinvested" accounting rule they may be inclined to act today as if we had a purely territorial system even though, in fact, it is partly worldwide.
Strange to think that mere accounting changes might matter behaviorally, to some taxpayers, more than changes in the income tax rules that actually apply.