First, let’s quantify the day yesterday. Yesterday’s nearly 40% jump in the VIX was the 11th biggest VIX daily spike in history. There are 6726 trading days in the VIX so the chance that any given day will be in the Top 11 is 11/6726 or 0.16%. As you can see from the table below 5 of the top 20 biggest daily spikes have happened in 2015 and 2016. There is no other 2 year period with 5 such days in the top 20 – not 2010-11, not 2007-08 and certainly not many before 2007. In fact 14 of the top 20 (70%) are since 2005 when Regulation NMS was passed by the SEC thus giving rise to the High Frequency Traders.

In addition, even though there have been spikes in volatility when the SPX is near an All-Time High, the SPX usually has not lost 50 points in a day this close to an All-Time High. The biggest such day prior to yesterday was -36 points on April 15th in 2013. The other days the SPX lost around -15 points.

Not only was the 50 point loss the biggest this close to an all-time high, but also the drop below the MA50 which is normally a big support level was also rather astounding. There are only 10 instances when the SPX opened above the MA50, dropped below the MA50 and proceeded to lose more than 30 points after that. The average drop through the MA50 is about 7 points with a standard deviation of 10 points. So dropping 36 points below the MA50 is a 3 standard deviation event (3 sigma event or 0.3% occurrence). As you can see most of these have happened during Bear Markets or during some large international crises like in 1997 or in June 2016 (Brexit). For this to happen out of the blue yesterday on no real news is rather astounding.

In fact, the only other day that bears resemblance to this VIX spike is 2/27/2007 which was triggered by a selloff caused by the government of China intervening in the Shanghai market to stop a 10% selloff there.

A Unicorn Day

Yesterday was a day that couldn’t be predicted with any of the usual methods of analysis – fundamental, technical or volatility analysis.

From a fundamental perspective stocks are fully valued, but GAAP earnings are on the way up. Also the US and overall global GDP is on an upswing in Q3. Moderately rising yields in Japan, Europe and the US normally trigger a rotation from expensive defensives to cheaper cyclical sectors. All of these are factors that dampen volatility and make the presence of high volatility days less likely.

From a technical perspective, there was no topping or reversal pattern. Big days like this are preceded by big hesitation, high volume days, some kind of a multi-day topping pattern as well. There are no such patterns to be seen. From support/resistance perspective, we had 3 pretty big support levels at 2160, 2150 and 2135 that should have held but didn’t.

From “trend following” perspective, indeed the Bollinger Bands are tightening and the near-term averages are bunched up, but the slope of these is positive so it is very unlikely for a big down day like this to transpire.

From a “mean reversion” perspective, you really can’t say that yesterday was a “mean reversion” move because the 50MA is the mean. The SPX had mean reverted by going nowhere for 2 months.

From a volatility perspective, volatility has been very calm, I would say too calm, but still this is a 3 sigma event based on the recent volatility profile and 2.5 sigma (2% occurrence) event on a more normal volatility profile. Again something that normally would not happen even if the VIX was in the 15-16 range the day before.

So what was the cause for the selloff yesterday?

I find it hard to believe that a speech by a FED president many people ignore triggered such a massive move. So all of a sudden the market doesn’t listen to Stanley Fischer and Janet Yellen, but it does listen to Eric Rosengren. Really?!? I also don’t really buy the argument that a 0.10% jump in the 10-Year Treasury yield triggered this selloff either. Yields are still at 1.67%, levels the market has traded below in only 143 out 13649 trading days in its history. That would be 99.9895% of the time people bought treasuries at higher yields. The story that you have a bond selloff on the horizon because Japan and Europe think that negative rates are not beneficial is not credible. Central banks may not like negative but they still very much like zero. So rates will go up 25-50% basis points. I am not sure why that would somehow trigger a larger escalation. So forgive me if I don’t think 1.67% 10-Year Treasury yield is a reason for a bond driven stock market selloff. I also don’t buy that North Korea nuclear test is the reason. North Korea has had other such tests, they didn’t trigger large market moves.

The most plausible explanation I have is that the market has traded in that 2160-2190 range for so long that any technical break of these levels will trigger stops, short covering and the algorithmic closing of trades tends to extend the move. What I have a hard time reconciling is that the move would be so extended. Essentially, we got a 1 week worth of selloff in 1 day. The only way I explain the 2.5% drop yesterday and the 5% drop after Brexit is high-frequency traders (HFT) algorithms. Once certain thresholds are broken, they hit kill switches and the liquidity evaporates from the market. For example, the market should’ve rallied at 2160 and 2150 yesterday but after 2150 broke, it was free fall from there.

This is a very disconcerting phenomenon, in my opinion, because it means that high frequency trading hasn’t made the markets more stable. They are more stable until they aren’t. There is a lot of liquidity up to a point and after that liquidity vanishes into thin air. I don’t really know what the fix for that it. So far it seems the SEC and the FED are really behind the ball in regulating the HFTs and making them responsible market participants. Until the FED and the SEC come up with a proper policy for the HFTs, these intraday flash-crashes are something that unfortunately market participants will have to continue to expect. We have had at least 3 of those crashes (Aug 2015, Brexit and yesterday) in a 1 year span – which is kind of amazing. I thought January’s controlled selloff proved that we have put the “flash crashes” behind us but apparently we have not.

Key Market Levels

This September was following the seasonal chart until Friday. Friday was supposed to be a down day and I expected that but it wasn’t supposed to be THAT big of a down day.

I don’t think the selloff that started last Friday will last very long or be very deep. I expect at most a 5-7% drawdown before the election which puts the SPX bottom in the 2035 to 2080 area. This is a typical pre-election drop and given the direction of earnings, with 2017 in mind I think this is an excellent opportunity to buy the dip. If rates are really rising, financials are a good buy here and if rates aren’t rising, what is different than before? Over the next couple of weeks and months, I will be looking to allocate the remaining 50% in the family 401(k) to stocks. If we get to the 2050-2100 area in the SPX, I think this is a really solid area to put money to work for the rest of 2016 and 2017.

SPX will meet serious support at the current 2130 level where the prior ATH was and after that at the 2100 level. The 2100 level where the higher low trendline of February and June is should be tough to overcome. And after that we have the 200MA at 2050 which is a big line in the sand.

Resistance areas on the topside is the 2165 where the 50MA is and 2175 level where the 10-40MAs are.