Corey Hoffstein had a thought provoking Tweet thread that started with "leveraged portfolios are subject to variance drain." The conversation went on to compare the effect of variance drain in leveraged portfolios, which is what the ReturnStacked ETFs do, versus variance drain in more common, long only portfolios.
Variance drain is a fancy term for the potential underperformance that could be caused by volatility working against a portfolio's result. Corey goes on to "prove" his point in the thread. When I see things like this, I will ask myself whether this can actually help what I do or is it overly academic which creates the possibility of being too clever by half.
The theory of leveraging up is probably correct but as we've looked at before, actually pulling it off in a fund is not simply a given. I've questioned the ReturnStacked funds as well as ETFs from Simplify, they are leveraged but the benefit, for now, is hard to find. Here's another example with the Simplify US Equity PLUS GBTC ETF (SPBC). It is 100% S&P 500 plus 10% in the GrayScale Bitcoin ETF (GBTC).
Portfolio 1 is straight SPBC, Portfolio 2 is a build it yourself version that does leverage up and Portfolio 3 is an unleveraged version of build it yourself. The comparison does a couple of different things. The lag of SPBC is larger than the expense ratio of the fund. Its best year was no where near as good as the other two portfolios, its worst year was worse than the other two, it had the biggest max drawdown and it's Sharpe Ratio is quite a bit lower.
The leverage as SPBC is using it, makes it worse. Portfolio 2 sort of supports Corey's point with the build it yourself leverage helping returns. But SPBC disproves his point along the lines of Karl Popper. All the positive results can't prove something, but it only takes one negative result to disprove it.
I don't feel the need to apply Popper literally, more like back to SPBC itself where something isn't quite right with how they implement the strategy. Maybe there is a friction with how the fund is rebalanced. That would be the easy thing to point to but I might be wrong.
Back to the original question of an idea being overly academic. We can each decide for ourselves whether Corey's point should influence us in any way. AQR seems to do a good job incorporating leverage and there are others like maybe WisdomTree, but it is difficult to do and I would not jump in too enthusiastically with investment dollars. I am all in on studying these and wanting to learn more but my actual exposure is quite limited.
The information, analysis and opinions expressed herein reflect our judgment and opinions as of the date of writing and are subject to change at any time without notice. They are not intended to constitute legal, tax, securities or investment advice or a recommended course of action in any given situation.
5 comments:
Like you mentioned, seems to be related to rebalancing. If you set the rebalancing to monthly the large discrepancy shrinks drastically.
I believe RSST rebalances daily FYI; perhaps this would limit variance drain?
I didn't realize RSST rebalanced daily, wow.
To be fair I think this is more due to the way people target the momentum factor and trend following. Resolve has a white paper where they discuss it's better to rebalance more frequently due to the signal to get more exposure to the trend signal.
I'm assuming that includes their exposure to SPY. Assuming their inflows/outflows cause them to sell either part of their TF program or their futures exposure to SPY to get back to 100%/100% as part of the creation/redemption process.
Corey also likes to reduce rebalance timing as he has discussed at length by more frequent smaller rebalances.
I'm not a big fan of that approach, seems to be less ergodic. I use quite a few individual stocks and some of them have really run over the years. At some point they do need to be trimmed but not too frequently IMO.
Hi Roger,
I agree with your point when rebalancing asset classes--why some research suggest rebalancing bands over time periods (assuming typical volatility of equity products). But for momentum/trend following products, they wouldn't cut the winners, they'd let those ride. Might be wrong but i thought MTUM just changed to quarterly rebalancing instead of semi-annual. I think initially everyone thought transaction costs would erode the momentum premium, but seems in practice transaction costs have not been the hurdle academics thought it would be.
Seems more frequent rebalancing increases the momentum exposure substantially. See M* chart showing momentum exposure in May and Nov (when it rebalanced historically):
https://pbs.twimg.com/media/FnunLmkWQAIvS0n?format=jpg&name=large
Post a Comment