Recent posts have described how capital gains tax deferral allows investors to reduce their effective tax rate asymptotically to zero, how this distorts investment decisions, how market-to-market cannot fix the problem, and how bond taxation partially deals with the problem. Before the eagerly anticipated revelation how the tax code could fix this distortion to come in a future post, let me describe another method by which this deviation can be further exploited.
You may say, fine, I can massively reduce the taxes owed on the appreciation of investments by just holding on to them long enough. But I also want to hold some investments which pay current returns, such as dividends or interest, which I am forced to recognize as soon as I receive them. Those returns I cannot defer.
But thanks to your friendly neighborhood financial engineer, you can and here is how:
The engineer once again sets up a trust in a hospitable jurisdiction. Every dollar the trust receives in investments (minus a small management fee and it really can be quite small, because running the trust will require very little effort or skill) it immediately converts to SPYders which mimic the return of the S&P 500. In return the investors receive a new security issued by the trust. In the tradition of charming names for new securities, let's call them RATs and think of a retronym later.
The RATs have the following properties:
- Collectively, the RATs are backed by the trust's SPYders. Hence, they will in total have the same return as the same quantity of SPYders. As the trust already holds these securities, it is perfectly hedged, has no risk, and exposes the investors to no risk beyond that of holding SPYders.
- However, some time towards the end of the tax year, the engineer is legally committed to generate a random non-negative number less than 1,000. At that point, all of the return and value of the SPYders goes into just those lucky RATs with serial numbers whose last three digits match the random number. All the other, unlucky RATs become worthless.
Now, what would a RAT trade for on an arm's length basis before the randomization? The expected value of a RAT with a par value of $1,000 is also $1,000, that is the value of a one-in-a-thousand chance to receive $1,000,000 worth of SPYders. But beyond the expected value, there is also the enormous risk of RATs. While some are gamblers and would actually be willing to pay more than $1,000 for a RAT, most investors are at least somewhat risk-averse and hence would only pay less than $1,000 for a RAT. As the trust needs $1,000 to pay for the SPYders backing each RAT, one might think the scheme would be doomed to fail.
But there is one type of investor which will always be willing to pay $1,000 for a RAT: those clever enough to buy full matched sets of RATs evenly distributed in the last three digits of their serial numbers. Let's call such sets of RATs, RAT PACKs. $1 million of RAT PACKs will always be worth exactly $1 million of SPYders, because the huge return on the rare lucky RATs will be exactly offset by the total losses of the many unlucky RATs. So a rational investor will, on an intrinsic basis, be indifferent between owning RAT PACKs and SPYders.
But what happens after the randomization near the close of the tax year? All the unlucky RATs will be worthless and the lucky RATs will appreciate enormously. Now the clever investor, having at the beginning of the year bought let's say 1,000 RATs at the then-fair-market value of $1,000 each, sells all the unlucky RATs for their fair market value—$0. Then the investor records a investment tax loss of $999,000, or 999 times a loss of $1,000 each. It is true that the lucky RAT has appreciated from $1,000 to $1,100,000 (if the S&P 500 gained 10% over the year), representing an unrealized gain of $1,099,000, but that gain can avoid even eventual taxation by the magic of deferral.
Now the investor can take the $999,000 of tax losses to offset all the gains including current returns such as dividends or interest in the rest of the portfolio. And next year, repeat ad infinitum. All the investor needs to do is buy enough RAT PACKs at the beginning of the year to offset a generous estimate of all otherwise-taxable, but not immediately needed, investment returns from the rest of the portfolio. If the investor over-estimated the other gains, this is not a significant problem; the investor can just keep some unlucky RATs for the next year or even sell part of the lucky RATs, to reach the goal of exactly zero taxable gains.
Now it might be argued that this is too easy and the tax law couldn't possibly allow it. For example, the equivalence between RAT PACKs and SPYders is so blatant that the tax authorities might deem it as without economic substance and hence disregard these transactions for tax purposes.
That is probably true. But that is only so because the scheme was for purposes of exposition set forth in the cleanest, simplest terms. That is easily avoided:
- Instead of overt randomization, instead tie the RATs values to some obscure economic derivatives. Those can readily be designed in such a way that they arguably represent some genuine economic investment choice, but in fact still be essentially random.
- Instead of fully negating the value of the unlucky RATs, just ensure that a sufficient number lose sufficient value. That will generate all the losses you need.
- Instead of buying perfect RAT PACKs, buy slightly unbalanced ones; by the law of large numbers, you will still effectively, if not legally, get the same return.
- Instead of putting your complete RAT PACKs into a single entity, spread them around unevenly among a variety of entities you ultimately own, and each entity will observe very real gains and losses due to its RATs, even if they in the end balance out for you.
These are just some obvious ideas. Doubtless, the clever people expert at this sort of obfuscation could think of many more.
If you still cannot believe that this is legal, remember that a less-extreme form of it is very common. Instead of owning SPYders, own an equivalent portfolio of all the stocks that make up the S&P 500. At the end of the year, sell all your losers to realize the tax losses. In order to stay reflective of the S&P 500, either immediately reinvest the proceeds into something economically nearly identical, but legally distinct, or wait 30 days before repurchasing the same losers, thereby avoiding the wash sale rule, albeit at a small, temporary risk of deviation from the S&P 500 return.
In principle, this universally accepted form of tax planning is no different from the RAT scheme. And even if the completely overt RAT scheme may not be accepted, one suspects that something more obfuscated, but very nearly as efficient, could not be prevented.
