What’s up with SPY?
by HunterAstute market observers may have noticed some significant stock market volatility this year (/sarcasm). An interesting Bloomberg story caught my eye a few weeks ago regarding arguably the biggest and most liquid equity instrument on planet earth – State Street’s behemoth SPDR S&P 500 ETF, $SPY. I probably could have stopped with all the preamble because you probably would have known what I was talking about by the ticker alone, but I digress. While it’s no longer the largest ETF by assets (the crown technically belongs to Vanguard’s $VOO now, although they are basically neck and neck), it’s still the most liquid ETF in U.S. equity markets, trading well over 40M shares a day even in the quietest of trading days. Given all the volatility we’ve been experiencing in domestic equity markets this year.
Anyway, the Bloomberg article accurately points out that amid the unprecedented buying frenzy on April 10th following the pause in the Trump administration’s global tariff rollout, the single most liquid and widely used vehicle to access U.S. stock market beta experienced a significant dislocation from its NAV. Personally, I was flabbergasted that $SPY hit a 90bps premium to NAV, even on such a crazy day in the markets. In my day job I traffic in $SPY regularly and, given the size and liquidity of the thing, I am quite used to single basis point premiums/discounts. Even with all the volatility I couldn’t have imagined $SPY getting that out of whack. Which raised the question in my mind about how dramatic this day’s 90bps premium was in the long history of $SPY. Using NAV data from State Street and daily closing prices, I took a look at the historical premium/discount to close for some historical context:
90bps premium was the highest closing price to NAV premium in ~16.5yrs (the State Street NAV data goes back to 2003). Next highest was 11/18/2008 at +0.98% – which, interestingly enough, was right around a particularly gnarly drawdown in mid/late November of 2008. Less than 3 months later the absolute bottom of GFC would be in on March 9, 2009. 10/13/2008 & 10/22/2008 were also notably high at 1.00% and 1.01%, respectively. The highest premium in my data was 7/1/2004 which was 1.05%. The ‘08s make sense as we were nearing the bottom of the GFC, but 2004 seems odd to me and I can’t find any obvious rationale for it. I’m mentally chalking that up to ETF (and I suppose equity markets in general) being pretty different from the current markets given all the changes we’ve experienced with market infrastructure over the last two decades.
On the flipside, I show 11 trading days of <-90bps discount. The max discount was -1.61% on 10/10/2008 with a total of 9 of those 11 days occurring at some point in the fall of 2008 (mostly October and November). The only outlier was again an odd 8/18/2004 day where $SPY closed at a -1.31% discount.
Zooming out a bit, across the full 5,386 trading days in the dataset the average premium/discount was -.0024% (median +0.0020%), so roughly zero as one might expect. The standard deviation across everything was 0.11%, which would make the +/- 90bps observations I highlighted something like an +8 standard deviation event. For what it’s worth, the big premiums/discounts don’t seem to be *necessarily* tied to big market movements in the S&P itself. Across the combined 17 trading days I highlighted above with >90bps of absolute premiums/discounts from NAV, there’s a smattering of returns. Sure, you have 4/9/2025 where the S&P was +10.5% on the day. But there are also days with pretty mild returns like -.01%, -.16%, -.60%, etc. See the chart below for a graphical representation of the data:
I think my takeaway here is that while these big swings are obviously possible for $SPY, it really takes a high volatility environment to push it out to these outer limits. Not sure that this is new information for most practitioners: duh, crazy environments lead to crazy premiums/discounts on even the most liquid vehicles. But I think it does put 4/10/2025 in some appropriate historical context.
