Today leveraged ETFs and inverse ETFs can be found for all sorts of markets such as stocks, bonds, and commodities. Before 2010 there were few leveraged ETFs in existence. For example, the 3X leveraged ETF for the S&P 500 with the ticker UPRO didn’t start until 2009. Fortunately, we can make some reasonable estimates as to how UPRO would have done during 2008, the dot-com crash, and the flash crash of 1987. This post will discuss a simplified method of analyzing how a leveraged ETF may have performed prior to its inception date.
Getting the Data
Yahoo provides some great free data for most mutual funds and ETFs. From yahoo we can easily grab the data for Vanguard’s VFINX (S&P 500 Mutual Fund) and ProShares UPRO (3X Leveraged S&P 500 ETF). Now you might think since UPRO is a 3X ETF we can just apply a 3X factor to the daily return of VFINX. However, with just a little bit more effort we can get a much more accurate synthetic UPRO prior to its start date.
Yahoo provides both the daily closing price and the adjusted daily closing price for UPRO and VFINX. The adjusted price simply include the affect of reinvesting dividends. Below we can see UPRO using the adjusted closing price compared to synthetic versions of UPRO created using the closing price or the adjusted closing price of VFINX with a 3X daily factor. As you can see none of them exactly match.
There is no free lunch, and using leverage is going to have some cost. Therefore, if we want to better estimate a synthetic UPRO prior to its inception, we need to add some kind of cost factor. In this particular case, if we subtract a 2.7% annualized fee we get very close to the real thing. It isn’t perfect, and there are certainly better and more detailed ways to estimate it. However, this simplified method is pretty decent. Using the adjusted prices for VFINX with a 2.7% annualized drag results in a synthetic UPRO that is very close to the actual UPRO.
Unfortunately, no synthetic is going to be perfect. This assumes the hypothetical UPRO could obtain the same leverage at a similar cost prior to 2009. For this and many other reasons we need to keep in mind that this is simply an estimate of a hypothetical situation. However, in the absence of certainty estimates will have to suffice.
Synthetic UPRO since 1980
Now that we have a formula for a synthetic UPRO that makes some sense, let’s look at the results of a hypothetical UPRO from 1980 to mid 2022.
The synthetic UPRO position may have generated a much larger return over the roughly four decade time period. However, without other diversification or drawdown mitigation virtually no one would have been able to emotionally stick with it. Don’t let the log scale fool you. In 1987 a user would have lost 61% in a single day! Then from roughly 2000 to 2019 they would have made $0 and at one point been down 97%! So don’t just go buy leveraged ETFs!
Leveraged ETFs offer a great deal of possibility, but they are extremely dangerous. They should only be bought and sold by professionals utilizing some well thought out methods. In some subsequent posts we will use this synthetic UPRO to discuss various topics and better explain Tines Capital’s use of leveraged ETFs. If you want to read about it, be sure to subscribe to the blog.