On February 5, 2018, a series of market events triggered an "event acceleration" of an ETN called XIV calling for effective redemption of the security on February 21, 2018. Investors in XIV lost 90% of the value overnight. $2 billion of value was wiped out. ETFs and ETNs are an intrinsic part of modern markets, and the collapse of XIV was a warning shot. A collapse of more central ETFs would have catastrophic effects on the market. I look at the prospectus of NYSEARCA:SPY to understand the level of risk that exists in perhaps the most important ETF in the market.
XIV was an ETN, whilst SPY is an ETF. There are significant structural differences of these two instruments. SPY is an instrument that holds underlying assets that represents the value held in the ETF. XIV did not hold any assets. It held derivative instruments that tried to mimic a short trade of the VIX, an index that tracks market volatility. There are several excellent articles that describe how XIV failed, and I recommend Stuart Barton's summary. A mismatch between VIX futures risk and SPX options risk, along with the size of the trades, resulted in a level of risk that the markets were not prepared to backstop in trades. At that point, selling the inverse volatility trade created more volatility and a feedback cycle that broke the system. These events triggered an "event acceleration" in XIV's prospectus and redemption of the security at over 90% lower than its traded value.
SPY has no such event acceleration. Its liquidation is trigger by events that will almost definitely play out over a very long time frame of over 100 years. Furthermore, upon liquidation, the value of each unit of SPY is equal to the value of the underlying assets, minus minimal costs. The exact wording from the prospectus can be found below:
The Trust has a specified lifetime term. The Trust is scheduled to terminate on the first to occur of (A) January 22, 2118 or (B) the date 20 years after the death of the last survivor of eleven persons named in the Trust Agreement, the oldest of whom was born in 1990 and the youngest of whom was born in 1993. Upon termination, the Trust may be liquidated and pro rata Units of the assets of the Trust, net of certain fees and expenses, distributed to holders of Units.
The 11 persons named in the Trust Agreement are aged 27-30. Even with a lethal pandemic, the chances of event (B) are minuscule at 0.2% to the power of 11.
The prospectus lists a number of risk factors:
Six of the reasons listed in the prospectus are generic risks related to the nature of the underlying securities. If you invest in S&P 500 stocks, you will be exposed to these risks. In other words, by investing in SPY, you are exposed to market risk. Unsurprising, as this is what this instrument was designed to do - mimic an investment in the market.
Two reasons, however, are unique to SPY. Reasons (4) and (8) are additional risks not borne through an investment in stocks.
The prospectus details the risk associated with fluctuation of NAV, Unit Premiums and discounts as follows. NAV is the value of the assets underlying one unit of SPY. A unit premium or discount to NAV means that the market is trading SPY at a different price to the underlying assets.
The NAV of the Units will generally fluctuate with changes in the market value of the Trust’s securities holdings. The market prices of Units will generally fluctuate in accordance with changes in the Trust’s NAV and supply and demand of Units on the Exchange or any other exchange on which Units are traded. It cannot be predicted whether Units will trade below, at or above their NAV. Price differences may be due, in large part, to the fact that supply and demand forces at work in the secondary trading market for Units will be closely related to, but not identical to, the same forces influencing the prices of the securities of the Index trading individually or in the aggregate at any point in time. The market prices of Units may deviate significantly from the NAV of the Units during periods of market volatility. While the creation/redemption feature is designed to make it likely that Units normally will trade close to the Trust’s NAV, disruptions to creations and redemptions and/or market volatility may result in trading prices that differ significantly from the Trust’s NAV. If an investor purchases Units at a time when the market price is at a premium to the NAV of the Units or sells at a time when the market price is at a discount to the NAV of the Units, then the investor may sustain losses that are in addition to any losses caused by a decrease in NAV. (emphasis mine)
The S&P 500 is in a period of unprecedented volatility. The VIX has reached record heights.
The key question is whether that level of volatility is sufficient to create premiums of discounts to NAV. We indeed have seen such premiums and discounts recently. Whilst they are the largest spikes in well over a decade, they are by no means historical record deviations.
The premiums and discounts were higher in 2002 and in 1998, but this time something is different.
Below are three panels showing the Discount / Premium to NAV of SPY, the VIX and level of S&P 500.
The maximum premium or discount to NAV coincide with highs of the VIX. This is exactly what the prospectus lists as a risk factor. Note that in both the exceptional discount and premium we witnessed, the premium / discount quickly returned to zero.
In the previous discount spike in 2002, these charts looked very different.
The dip in the discount actually corresponds to a VIX low. The S&P 500 moved around 4%, a movement significantly lower than those that we recently experienced. I could not find a reason for the unusual fall in the NAV discount. Whilst July 2002 was a month of market turmoil, with the Enron scandals in the background and 7 S&P 500 market constituents consequentially exiting and entering the index, nothing that I could find happened on that day.
It was a similar story in 1998.
The fall in NAV discount happens irrespective of market volatility. Again, this seemed to be a localized event, not connected to market events.
This appears to be the first time that market volatility has had a significant impact on the Discount / Premium to NAV. And this time, the AUM of SPY is much larger than ever before. In fact, over 10X what it was during the previous NAV discount dip in 2002.
Whilst the risk appears to have played out on this occasion, it was localized to a couple of trading days. The investors that would have lost or gained from this event were limited to those that traded SPY on March 13 and 16. The creation/redemption features appear to have mitigated that risk and did not allow it to become a structural risk. If future volatility opens similar discounts or premiums to NAV, an arbitrage trade against the underlying S&P 500 assets becomes available. The trade itself, however, will push the premium / discount lower, as a pair trade against the S&P 500 will close the gap, not push it further apart.
The risk as defined in the prospectus:
Even if an order is processed through the continuous net settlement clearing process of NSCC, Portfolio Securities or Units, as applicable, may not be delivered on settlement date, due to liquidity or other constraints in the clearing process. Orders expected to settle outside of the continuous net settlement clearing process of NSCC are not covered by NSCC’s guarantee of completion of delivery.
Creation Units are blocks of 50,000 units of SPY stock. These blocks are how SPY is created, redeemed and trades in a cost-effective manner. This risk is a breakdown of trading functionality due to clearing mechanisms breaking down. The one specific risk listed is liquidity constraints. Note that this risk, compounded with the above discount / premium to NAV risk, could lead to a breakdown in the mechanisms that align the market price of SPY with its NAV. It is unclear what would happen in such a scenario. In such a scenario, $240 billion of traded assets would not trade at the value of the underlying assets and would not have a mechanism to revert to that value. Such a scenario could create a selloff of the security and an eventual breakdown of the ETF. It is unclear what could stop clearing in such a manner. In 2008, when the markets experienced a significant liquidity crunch, this did not occur.
The mechanisms designed to keep SPY trading at NAV have functioned well during recent market volatility. Whilst short-term, small arbitrage opportunities have opened up, we have seen nothing that suggests structural risk at this point. If clearing in SPY becomes significantly hindered, this could lead to a worst-case scenario. However, at this point, there is no visible tangible risk of that happening.
If you liked this article please follow me by clicking on the "follow" button by my picture at the top of this page.
This article was written by
Disclosure: I am/we are long SP500. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.