It is said that decommissioning costs on the order of $300 million. I don't see information readily available on how much utilities have in these funds, but going by anecdotal information, utilities believe that it will cost significantly less than what they've saved. The fund is thus a means for return of shareholder value. The incentive for the utility is to decommission for as low cost as possible, and whatever is left goes to their bottom line. The fund inherently has to be an overshoot, because the point is to prove existence of sufficient funds. Because they plan to get the money back, the operators don't mind packing away too much anyway. I've also heard it said that there are many plants out there that don't look very profitable, but they actually are because the earnings are being delayed by stashing it in the fund.
I don't understand the 20% myself. Our corporate tax rate is 35% for the highest bracket (above $13 M per year, which is tiny) and 15% for the lowest. As we often hear, however, the effective tax rate is much lower, at around 13.4%. I would imagine that the quoted tax rate on the decommissioning fund would act in place of the corporate tax rate, but I can't say for sure. Even assuming I've interpreted that right, whether the utility could have obtained a lower tax rate by absorbing the money into their general fund would be an open question.
Even so, that would be small potatoes compared to the fact that the fund has requirements placed on it. Since the goal is to satisfy the NRC that they have sufficient funds, it's sure to be in relatively safe and low-yield investments. That's not good for shareholders, and is really the opposite of a subsidy, although it may be a reality of decommissioning needs. In fact, the loss of return due to the requirements on the fund may form the argument for setting a special tax rate in the first place.
The main lesson I learned from industrial economics is that corporations basically bleed money. Deferring gains until later almost always incurs an economic penalty, since the rate of return you need to justify investments internally to a corporation is much higher than what we normally think of from personal finance. I think a part of the reason is that tax it has to incur to get from the corporate balance sheet to the investor's account. Worse still, we don't even bother correcting depreciation for inflation. That means the return from an investment must be vastly higher than a single calculation with single rate would suggest. Nuclear plants, as very long term investments, have extremely unfavorable depreciation schemes, and faster depreciation is sometimes noted as a subsidy to various renewables. In the big picture, there are many ways taxes disincentive nuclear investment, although the decommissioning fund may or may not be among these.