|Shares Out. (in M):||32||P/E||0||0|
|Market Cap (in $M):||999||P/FCF||0||0|
|Net Debt (in $M):||0||EBIT||0||0|
Long Proshares Ultrashort S&P 500 (SDS)
I expect this idea to get a relatively low rating because it is not a value-based, buy and hold recommendation, but rather a security that I believe could be a timely addition to a portfolio of value ideas based on the current macro and technical setup.
The VIC tends to respond coldly to these ideas. The two most recent ones received low ratings and significant pushback. Miser’s IAU recommendation on 6/2/18 was rated 3.7 and slapped with a “topic masquerading as an idea” tagline. Defy_Augury’s QQQ short recommendation on November 4, 2018 received a 2.7 rating and a very pointed post highlighting the inappropriateness of posting macro ideas on a value investors’ forum (see post 2 entitled “Value Boyz Playing Macro”).
Nonetheless, I have found these recommendations to be very useful and often timely additions to the portfolios I manage. They diversify my portfolios, adding alpha by altering an overly homogeneous return profile and dampening volatility. Value ideas can go through long periods of underperformance that challenge the patience of even those most committed to the strategy. Further, timely macro ideas tend to be more liquid and, depending on the idea, can be implemented in greater size in a portfolio because their risks tend to be less idiosyncratic.
I’ve noticed that when the rare macro idea appears on the VIC, the timing tends to be pretty good. Perhaps this is because our kind is so instinctively opposed to these things that we only dare to post them when we have very high conviction. It’s notable that Miser’s IAU idea is up 16.5% since it was posted, versus the SPY up only 7.9% over the same period. Defy_Augury’s QQQ short would have captured an 8.7% downdraft through the end of 2018 if held for the two months since it was posted.
What is SDS?
SDS is an exchange traded fund (ETF) sponsored by the Proshares fund family. It seeks to deliver a return on a daily basis that is twice the inverse of the percentage return generated by the S&P 500 index. For example, if the S&P 500 is up 1.5% on a given day, we should expect the value of SDS to fall by 3.0%. If the S&P 500 is down by 1.5% on a given day, the SDS should be up by 3.0%. SDS achieves this by using index derivatives. The tracking error over long periods can be quite significant because of the imperfect matching of the derivatives, the 90 basis point expense ratio, and the daily reset, which will be discussed in more detail below.
SDS has an NAV of approximately $1 billion. It is fairly liquid, trading approximately 6 million shares, or $180 million dollar value, per day.
The Daily Reset
SDS is designed to deliver the 2x inverse return on a percentage basis each day. It is therefore not a static portfolio like most index ETFs or closed end funds. This can result in performance that diverges greatly over time from what would be expected from directly shorting twice the nominal dollar value of the S&P 500 index. As discussed below, the holder of SDS should expect to be worse off than shorting twice the index in volatile but non-trending markets, and somewhat better off in trending markets.
To illustrate this, let’s look at the mechanics behind the daily reset using an assumed starting NAV of $30 per SDS share, first in a volatile, but non-trending market, then in a trending market.
Effect of Daily Reset in a volatile, but non-trending market
Suppose on Day 1 the S&P 500 falls 2.5%. SDS has been set up with an equity value of $30 per share backing a $60 synthetic short position in the S&P 500. So at the end of Day 1, the value of the short position has fallen 2.5%, or $1.50 per share (equal to $60 x .025). SDS’s value has moved up to $31.50, which is up 5%, as intended.
However, its synthetic short position in the S&P 500 is now only worth $58.50 per SDS share. This is inadequate to generate a double inverse percentage return on Day 2 because the short position will have to be double the SDS value of $31.50 to achieve this. So SDS’s managers have to synthetically short $4.50 per SDS share more of the S&P 500 before Day 2 begins to set up for the 2x inverse result. After the additional short is put on, Day 2 begins with SDS having an equity value of $31.50 per share and carries a synthetic short position of $63 per share.
Now suppose on Day 2, the S&P 500 rises exactly back to where it started on Day 1. This requires a 2.564% increase on Day 2. SDS’s synthetic short position is now worth $64.615 per SDS share and SDS’s equity value declines by $1.615 from the $31.50 Day 2 starting value to be worth $29.885 at the end of Day 2. Thus, over the two days the S&P 500 has not moved at all, but whipsaw has caused an 11.5 cent erosion, equal to 38 basis points on the starting value, in SDS’s NAV. If this non-trending movement continues, the effect of SDS having to continually “buy high and sell low” to maintain its targeted levered exposure will dissipate its NAV considerable relative to a static short position.
Effect of Daily Reset in a trending market
Suppose Day 1 occurs exactly as above, with the S&P 500 declining 2.5%, the equity value of SDS rising 5% from $30 to $31.50, and SDS’s managers increasing the S&P short position increased from its $58.50 Day 1 closing value to $63 per share before Day 2 begins.
However, in this scenario the S&P 500 falls another 2.5% on Day 2. SDS’s synthetic short position is now worth $61.425 per SDS share and SDS’s equity value increases by $1.575 from the $31.50 Day 2 starting value to be worth $33.075 at the end of Day 2. Thus, over the two days the S&P 500 has fallen by 4.9375%, but SDS’s equity value has risen by 10.25% over the two day period. A static 2x short would have increased the security’s equity by $2.9625 or 9.875% but the releveraging increased it by $3.075 per share, or 38 basis points more off the initial NAV. Here, maintaining SDS’s leverage on a daily basis tends to boost NAV relative to a static short position. This math works in SDS’s favor relative to a static 2x static short position in when the S&P 500 is trending upward as well, in that the nominal short position is reduced with each up day in the S&P 500 and has less impact on SDS’s NAV as the S&P 500 trends upward than the static short would.
Why am I recommending this particular security?
I’m recommending using SDS to hedge out some market exposure for several reasons:
The fact that it is an ETF means that it can be used in more traditional accounts. This security can be purchased, held, and valued like any stock in a typical brokerage account. Other market hedges like index futures, swaps or other derivatives typically require more sophisticated account arrangements and often higher account values to deploy.
The security is liquid, is adequately capitalized, and trades with a tight bid-ask spread. There are many other ETFs with similar features, but most are undercapitalized and illiquid. SDS has a billion dollar equity capitalization, trades $180mm in value per day, and usually has a penny bid-ask spread.
The expense ratio is not onerous. The expense ratio is 90 basis points, which isn’t wonderful, but since it’s double levered, it requires half the value held to offset each unit of risk. In other words, you’re paying 45 basis points in expense per unit of risk offset.
The daily reset feature is actually a benefit over most market outcomes. As mentioned above, the daily leverage reset feature tends to dissipate NAV in volatile, non-trending markets and supportive of NAV in trending markets. The reality is that markets have a greater tendency to trend than not. It is interesting to compare the performance of the 2x inverse SDS security with that of the 1x inverse SH security over the last 5 years. Between July 31, 2014 and July 31, 2019, the S&P 500 is up 71%, the SH 1x inverse is down 43.3% and SDS is down 70.6%. Both are down less than a static short would have been, and SDS offered double the hedge at initiation and would have cost less than a 1x static short over the period and somewhat less than twice the cost of the 1x reset SH did over the period. Note that I am not expecting this position to be held over a multiyear timeframe, but it is instructive to see what the impact would be if it were held should an attractive exit point not appear or somehow be missed over this period.
The value dissipation of the double leveraged SDS in non-trending markets is much more manageable than that of the triple short ETFs. If we look back at the relatively volatile but non-trending period between March 31, 2015 and March 31, 2016 when the S&P 500’s price moved up and down quite a bit but ended very close to where it started, we can get an idea of the potential cost of this hedge in an environment highly unfavorable to the daily reset feature:
Period Cost per Unit
Security Return Leverage of Risk Offset
SPY +1.70% 1.0x -1.70
SH -4.90% 1.0x -4.90
SDS -11.16% 2.0x -5.58
SPXU -19.45% 3.0x -6.48
As the table above demonstrates, none of these ETF solutions is particularly good relative to an outright short of SPY, but SDS is only marginally more costly (at 5.58 percentage points versus 6.48 percentage points) per unit of risk offset than the 1x Inverse SH ETF. Note, however, that this outcome would only have unfolded if the hedge was not taken off when the opportunities presented themselves. There were two separate multi-week drawdowns of over 12% during this period.
I think many VIC members would agree that after 10 years of extraordinary central bank policies, the distortions that have accumulated in the system are well beyond historical norms. Over $15 trillion of the world’s debt (25% of all bonds outstanding) now trades at negative yields - levels that my undergraduate economics textbooks from long ago stated were impossible.