Here's is how I implement Probabilistic Tunneling posted in 2023:
Suppose I had $100 invested in UPRO before Trump's tariff announcement. Anticipating that the event would trigger market turmoil, I wanted to close the leveraged position and hold cash (SHV) until the volatility subsided.
In a tax-free world, this would be straightforward: sell the position and wait out the storm with no further risk. However, let’s assume the UPRO is held in a taxable account, and closing the position would trigger a capital gain of $90. If it's long-term gain, I’d owe 20% in federal tax,13.3% in state tax (CA), and potentially plus 3.8% NIIT. If short-term gain, the marginal federal tax rate can be as high as 37%.
Regardless of whether my bet is right, I would surely pay a significant tax bill. Instead of closing the position in the taxable account, I can open an opposite position in Roth IRA accounts - In practice, I used YANG and TMF to bet the market crash.
If the market does drop as expected, I’ve effectively transferred the loss in the taxable account to the tax-free account. If the market rises instead, the money flows in the opposite direction, but my overall portfolio remains market neural. As long as my betting odds are better than 50/50, this strategy can gradually move money in some desirable direction over time.
If the assets in the taxable account aren't highly appreciated, the strategy becomes even more favorable. Whenever the taxable account incurs a significant loss, I can simply close the position, realize the loss, and use it to reduce my overall tax burden.
Very interesting idea, thx for sharing. To be precise, I thought it’s better to use expected value>0 in place of probability > 50%, right? You can bet with < 50% but still expect to make money if the reward/loss ratio is high enough. Also it’s a little weird to me: if I have a strong conviction on a bet — high expected value, why not just bet in the same direction in both accounts? If I only have a weak opinion on the bet but still positive ev, why not just bet in smaller bet size? If I am completely market neutral, then there is no guarantee which direction the money would flow between the two accounts, plus the cost incurred. It seems to me the problem here is that the market neutral conception seems to be binary. I.e you can’t be half market neutral, half not. Once you give up on the market neutral objective, you fall into the expected value system .
ReplyDeleteExpected Value only makes senses when the number of bets N is very large. Here, my base case is buy and hold for 99% of time. Only in the rare events such as COVID-19 and Liberation Day, I may decide to stay out of the market for a brief period. It requires me to sell highly appreciated position and buys it back later. Instead of selling, I can implement such bets by opening an opposite position in a tax-advantaged account.
DeleteThere is no real way to calculate EV for such bets because these are non-repeatable events and any estimate based on historical data is probably irrelevant.