Yesterday at the colloquium, we discussed Fadi Shaheen's above-titled paper, which is available here.
As background to the paper, the "new view" of dividend repatriations demonstrates that, if the U.S. repatriation tax for multinationals' earnings through foreign subsidiaries remains in place indefinitely at a fixed rate, and if repatriation at some point, at this fixed rate, is inevitable, then there is no lockout so far as shareholder-level economic incentives are concerned.
Thus, suppose for simplicity that there is only a U.S. tax, with no source-based foreign tax. Let's call the U.S. repatriation tax rate t, the per-period after-tax rate of return r (let's also make things simpler by assuming that you get the same after-source-tax return everywhere), and the amount of foreign source income that is at issue X. If the U.S. parent repatriates X immediately, it ends up with X(1 - t), and then at the end of the period it has X(1 - t)(1 + r). By contrast, if it repatriates at the end of the period, it has X(1 + r) to repatriate, and doing so leaves it with X(1 + r)(1 - t). Obviously (and it's a good thing too), we don't need advanced math skills to conclude that X(1 - t)(1 + r) = X(1 + r)(1 - t).
While you can kill the strict equivalence with different after-source tax rates of return as between home and abroad, all that shows is that it's desirable to keep X where it earns more, rather than less - not that the repatriation tax is discouraging repatriation. What makes the equivalence work is the fact that deferring the repatriation tax takes today's prospective liability and both discounts it at r (since it's better to pay a fixed sum later rather than now) and causes it to grow at r (since the amount to be repatriated keeps growing).
While this is an important conceptual marker that improves one understanding of the relevant incentives, the bottom-line conclusion - no lockout - is not true, because the underlying assumptions are not true. In particular, keeping they money abroad has option value, since you can simply wait for the repatriation tax rate to drop. This is especially significant when we don't just have an expectation of random walk changes in the future history of repatriation tax rates, but rather there is a very real chance that it will go down by reason of a tax holiday, corporate tax rate cut, or the enactment of another tax holiday like that in 2004.
Hence there's significant lock-out, contrary to the new view model, because only the fools among corporate CEOs and CFO's would believe that the repatriation tax rate is fixed. And whether or not these individuals are as brilliant as their paychecks are fat, they are certainly smart enough to realize that. (Further point: It's not clear that a taxable repatriation is inevitable even if we are comfortable with viewing all wealth as ultimately consumed. If capital markets work well enough, the FSI could effectively be repatriated, put in shareholders' pockets, and consumed by them without there every being a taxable repatriation.)
In the paper, howevcr, Shaheen explores the possibility of lockout if the new view is actually correct, rather than merely being an important explanatory tool. (But just one last comment on the new view - saying it's "false" in practice is no more a criticism than saying the Coase Theorem doesn't hold because there are transaction costs. That's actually, at least arguably, the point - to show us where the relevant bodies are buried.)
The paper discusses an alternative, and indeed complementary rather than contradictory explanation for lockout, which is that publicly traded firms tend to have corporate governance problems, which cause the managers to care about current or near-term financial accounting income, at the expense of actually maximizing present value. Accordingly, they seek the accounting status of "permanently reinvested earnings" (PRE) which they have persuaded their accountants will never come home. Once PRE status is attained, the deferred repatriation tax, rather than being charged against earnings as a current expense (without regard to the fact that it hasn't been paid yet), is completely ignored - valued at zero, since supposedly it is irrelevant once the managers swear on a tall enough stack of bibles that they will never repatriate particular funds.
This entry is already rather long, so I will post it now and resume in a follow-up.