U.S.-based multinationals hold $2.1 trillion in foreign cash and insist that the only way they can feasibly bring that money back home is if Congress grants them a tax holiday—an idea that even President Obama now appears to support. But the argument they (and the President) are making for a holiday is based on a series of illusions. If policymakers fail to see through the artifice, they will find themselves disappointed once the smoke has cleared, and the Treasury will have needlessly lost substantial tax revenue. That bundle of overseas cash is more than these multinationals hold domestically, and equivalent to 18 percent of the national debt. The companies blame the 35 percent tax they’d have to pay to bring their foreign earnings back to the U.S., the so-called repatriation tax. And they’ve been lobbying for a temporary reduction in the rate. Their argument: If only Congress would grant them a partial tax holiday, they’d bring the money home, pay some tax rather than none, and hire American workers with the imported cash. Last week, President Obama said, “We have discussed the possibility of being able to bring in some of the dollars that are trapped outside of the country right now, and in a one-time transaction, potentially use that to pay for some infrastructure improvements. I think there is some openness to that [emphases added].” The President is saying several things here, all of which echo the multinationals’ argument, and all of which are illusory:

1. The dollars are “trapped outside of the country”.

The President invokes the red-blooded American nightmare of being trapped abroad. Multinationals themselves like to deploy the social-scientific term “lock out effect”. Almost everyone refers to what’s not happening as “repatriation.” Yet recent Senate hearings reveal that companies like Apple, Cisco, and Google actually hold at least 75 percent of their “unrepatriated” foreign earnings in the U.S. You read that right: the money is back in the U.S., and yet “unrepatriated.” As such, it entered the U.S. without “repatriation tax.” That’s because the word “repatriation” in this context doesn’t really mean “reinvestment in the U.S.” Accordingly, the “repatriation tax” is not really a tax on bringing foreign earnings home. Rather, it is a levy on shifting the legal ownership of foreign earnings from a firm’s foreign subsidiary to itself. “Repatriation” refers merely to this legalistic change of pockets. While the money remains in the subsidiary’s name, the parent, who controls the subsidiary, may generally invest the money in the U.S.—including in U.S. bank accounts and in the stocks and bonds of unrelated U.S. companies—without paying repatriation tax.

2. The tax holiday would be a “one-time transaction.”

In fact, it would be the second one-time-only repatriation tax holiday. The last was in 2005. However reasonable it was to suppose that the first one-time holiday would not be followed by a second, it seems significantly less reasonable to suppose that a second would not be followed by a third.

3. The holiday would raise revenue.

No doubt, a partial tax holiday would boost revenue in the year it is offered. That’s what happened in 2005, when the tax rate on repatriations was cut from 35 percent to 5.25 percent. The question is: at what cost to future revenues? The problem is worse than it might seem. Of course, firms would accelerate repatriation of legacy foreign earnings to take advantage of the holiday, and the government would thus forgo the full 35 percent tax on this pre-existing income at some future date. But firms are also likely to delay the repatriation of future foreign earnings while they await the next one-time holiday. When waiting is a viable option—when, for instance, firms can borrow at an interest rate that’s less than the return they’re getting on unrepatriated cash—a 35 percent rate that everyone knows will be periodically reduced to 5.25 percent is really a 5.25 percent rate.

4. The holiday is good way to pay for urgent infrastructure improvements.

With a repatriation holiday, the government has more to spend now, but has to “give it back” later in the sense of sacrificing the revenues that it would have received sometime in the future. That’s like government borrowing. Borrowing is the right way to handle urgent infrastructure needs whose neglect may have dire consequences. But is this implicit holiday loan a good way for the government to borrow? It certainly doesn’t look good on the back of an envelope. Consider, generously, only the effect on legacy earnings and assume the holiday rate is again 5.25 percent. For every $1 of such earnings “repatriated” at a 5.25 percent rate, the government gets 5.25 cents now and loses 35 cents at some future date. In other words, it “repays” about 7 times more than it “borrowed.” That’s a sizable multiple. If the government explicitly borrowed at its current long term rate of 3 percent, the amount that it would have to repay would only reach that multiple after 64 years. The question then becomes: absent anticipated holidays, would US multinationals repatriate their foreign earnings at some point in the next 64 years? It seems likely that they would. Our roads and bridges may well need urgent repair. Our economy may well need an infusion of cash. Our government may well need revenue. But the belief that a repatriation tax holiday is the answer to these needs is founded on a series of illusions.