Those that follow my personal account on Twitter and StockTwits will be familiar with my weekly S&P 500 #ChartStorm in which I pick out 10 charts on the S&P 500 to tweet. Typically I'll pick a couple of themes and hammer them home with the charts, but sometimes it's just a selection of charts that will add to your perspective and help inform your own view - whether its bearish, bullish, or something else!
The purpose of this note is to add some extra context beyond the 140 characters of Twitter. It's worth noting that the aim of the #ChartStorm isn't necessarily to arrive at a certain view but to highlight charts and themes worth paying attention to.
So here's the second S&P 500 #ChartStorm write-up
1. This year's market vs the historical average: This chart surprised me when I put it together. I decided to use 2 axes because the average (black line) by virtue of being an average moves in a relatively narrow range vs the swings and gyrations of each year. The pattern is clear and YTD 2016 appears to be tracking very closely to its historical seasonal pattern. This means we're probably going to have to go through a rough patch before that nice looking year-end rally comes into play.
Bottom line: 2016 is playing out remarkably similar to the historical seasonal pattern, this is bad news short-term.
2. An average year in the life of the S&P500: The below graph shows the average seasonal pattern across the year (by business day, from 1990-2015) for the S&P500 and the CBOE option implied volatility index or VIX. This is a chart that's worth keeping on hand as a reference guide to seasonality for equities and equity risk across the year. The key takeaway as above is that we're heading into the rough part of the year right now, so it warrants a cautious bias.
Bottom line: The seasonal pattern for equity risk and equities is for more volatility and weakness at this time of the year.
3. Seasonality by the month: This chart shows the seasonal pattern by month and also includes the "% positive returns" (i.e. in September the historical record is that only 45% time are returns positive). It's worth noting the % positive because averages can and do deceive - there are always exceptions to the rule when it comes to seasonality in the stockmarket, and if you think about the 45% stat - look at the inverse, i.e. the market only fell 55% of the time - it's not far from 50/50. So while seasonality is worth factoring in, and can help form a more balanced and broader view, it's not the be all and end all.
Bottom line: September is the worst month historically, but historically September saw negative returns only 55% of the time.
4. VIX seasonality: It should be reasonably unsurprising to learn that the VIX displays a similar pattern of seasonality to the S&P500. What may be surprising or concerning is that 2016 has tracked the historical trend quite closely (with the exception of the January panic and Brexit vote). The reason it would be concerning is that usually the VIX trends up around this time of the year, and Friday's spike may be just the beginning of a more volatile patch. So definitely something worth noting for VIX traders.
Bottom line: The VIX also displays seasonality, and it tends to go higher this time of the year.
5. VIX options volume: Schaeffers research highlights the merits of monitoring volume in options trading on the VIX and how it often functions well as a market timing tool. The rule of thumb is that it spikes in a sell-off, and when it hits low levels that's usually the time to get cautious. Friday's trading makes you think when you look at this chart.
Bottom line: VIX options volume is a useful timing tool, it recently signaled a bearish/cautious outlook.
6. Conditional seasonality: For all the talk of seasonality it's worth noting the conditional seasonality work in the table below by BAML. What they do is look at the seasonal effect when you run a condition on whether the market is above or below its 12 month moving average and whether the moving average is sloping up or down - i.e. really they're asking if the market was in an uptrend or a downtrend headed into September. The result is when the market is in a strong up trend (price > 12mma and 12mma +ve sloping) the average September performance is what I would say mediocre vs terrible. So given this conditions has been satisfied for 2016 the bright side is that September could end up just being mediocre.
Bottom line: Based on conditional seasonality analysis September 2016 may not be as bad as unconditional seasonal analysis suggests.
7. Sell side consensus indicator: The below chart tracks sell side analyst consensus expectations, and the rule of thumb is to interpret it in a contrarian fashion i.e. extreme bullish expectations is actually a bearish signal, and extreme pessimism is a case of "so bad it's good". The latest observation was the most negative since around the time of the European debt crisis. So another bright side is that this indicator points to upside potential for equities. Of course the counter argument is simply that for once the consensus on the sell side are right and we should be worried. I'll leave it to you to decide that one.
Bottom line: Sell side consensus expectations have become extremely pessimistic which is a positive/bullish contrarian signal for equities.
8. High dividend paying stock valuations: The chart below from LPL Research shows that high-dividend paying stocks have become extremely expensive. This is a logical result of the secular bull market in bonds over the past couple of decades and more lately a reflection of central bank policy and the "search for yield". This is a good example of a potential distortion effect of ultra-easy monetary policy, and is a key risk area. Any signs of higher bond yields or tighter monetary policy will put this sector at risk and there would most likely be spillover effects to the rest of the market. While it's unlikely that bond yields will head structurally rapidly higher at this point, bond yields are prone to spikes e.g. during the 2013 Fed taper tantrum. All it might take is an aggressive Fed statement or the wrong move by the BOJ or ECB to trigger another bond market tantrum.
Bottom line: Low bond yields have pushed valuations to extremes for high dividend paying stocks, this is a risk for the broader market and for yield chasers in particular.
9. Falling number of stocks: This one falls in the interesting category. If you're having trouble finding good stocks to buy one reason may be that the number of stocks trading has fallen to the lowest level since 1984. Those using a straight-line projection in the mid-90's would have been wildly wrong to predict a higher number of stocks trading in the future. I think this reflects a number of trends e.g. lower IPO volume, more private equity, and more M&A. Those 3 factors are generally consistent with a favourable supply environment for equities (thinking about supply of equities vs demand). So perhaps this has been a good thing for the market...
Bottom line: The number of stocks in the CRSP database is the lowest since 1984.
10. Election poll spreads vs the market: This one is a bit of "fun with charts" comparing the spread in the presidential election polls to the stockmarket. There is potentially some albeit weak correlation - when the spread is more in favor of Clinton the market is usually headed higher, and vis versa for Trump. Whether this actually reflects market sentiment about the potential policy outlook is up for debate. It's also plausible that a falling market is associated with more fear and negativity, the likes of which would favour an anti-establishment candidate (as we learned in Europe). Of course, as we learned with Brexit, correlations between markets and politics can be nonexistent for a while and then dranmatically existent in an instant, so political risk is still worth thinking about.
Bottom line: There is a weak correlation between the polls and the market, but political risk tends to exert itself dramatically rather than gradually.
So where does all this leave us?
There's probably 2 main themes that came out of this week's charts:
1. Seasonality
The seasonality aspect was examined from a couple of angles, first on a business day basis and then by month. Whatever way you look at it the worst part of the year is the September-October period. This period is characterised by a generally higher VIX and weakness in the stockmarket. There are exceptions to the rule, particularly when you look at conditional analysis, which would suggest that this year may not be that bad. That said, the overall takeaway on seasonality is that while it's just one factor it warrants a cautious bias over the next few weeks.
2. Risks to keep in mind
Aside from negative seasonality a couple of other risks were highlighted. One is the extreme valuations seen in the high-dividend paying stocks, clearly this one is verymuch reliant on low bond yields and ultra-easy monetary policy. The other one is political risk; aside from the myriad political risks offshore, in the US the presidential election poses risks to sentiment and the policy outlook. Brexit showed us how dramatically political risk can show up. Another short-term risk is the lull in options trading on the VIX, which is a warning sign for a correction.
Summary
This week's ChartStorm highlighted a number of risks. Besides the well documented seasonality which is at its worst around this time of the year, other risks include stretched valuations in yield sensistive assets, political risk, and bearish short-term timing indicators. Overall a cautious bias is warranted on equities in the short-term. Of course if it all works to plan, the seasonality identified above could offer the patient or opportunistic investor a nice end to the year.
See also: Weekly S&P500 #ChartStorm - 4 Sep 2016