TradingEdge for Sep 3 - More Thrusts in Sectors, Still Tough for Commodities
The goal of the Weekly Wrap is to summarize our recent research. Some of it includes premium content (underlined links), but we're highlighting the key focus of the research for all. Sometimes there is a lot to digest, with this summary meant to highlight the highest conviction ideas we discussed. Tags will show any symbols and time frames related to the research.
STOCKS | ||
In May, economic surprises in the U.S. turned negative for the first time in almost a year. They rebounded briefly, then surprises turned even more negative as economists didn't lower their forecasts enough.
Worldwide, however, economic reports had still been coming in better than economists' forecasts...until last week. For the first time in over a year, Citi's Global Economic Surprise Index turned negative.
Even though the Citi Global Economic Surprise Index just turned negative, the one focused on the U.S. is nearly -50. If we filter the table above to only include those signals when U.S. surprises were lagging the world, then the ratio of the Russell 3000 to the MSCI World Index tended to see losses.
The ratio between the two indexes, the U.S. versus the world, showed a negative bias at some point between 1-2 months later every time. Granted, there isn't much history, so the sample size is tiny.
THE CHALLENGE OF AN AUGUST HIGH
September and October are well-known as being the most challenging for traders, as stocks have tended to see their worst returns and highest volatility. It's relatively unusual for the S&P 500 to close at a record high in August. It would be even more unusual if it did it in September.
Stocks have rolled right over every possible negative in 2021, and seasonality is a tertiary factor anyway. So maybe we need to take the following table with a big grain of salt. It shows returns in the S&P 500 after it closes the month of August at an all-time high, and it's a sea of red.
Even though by definition, momentum was strong, the S&P had an incredibly hard time holding onto its gains over the next month. Most remarkably, the Risk/Reward Table shows that only two signals saw more reward than risk during the next month.
SEPTEMBER AND OCTOBER TURBULENCE
Will anything change with stocks' momentum between now and the end of the year? Jay looked at a particular seasonal tendency, which is kicking into gear now.
The chart below displays the Annual Seasonal Trend for the S&P 500 Index. The tendency for weakness in the months directly ahead is fairly obvious.
The table below breaks average yearly historical performance for the Dow Jones Industrial Average into two-month segments each year from 1900 through 2020.
To understand the results in this table, note that the first line looks at the performance achieved by holding the Dow ONLY during December and January every year. As you can see, this two-month period showed a gain 66% of the time with an average gain of +2.2%.
There have undeniably been several devastating market declines during this period even though from year to year, the results are roughly a coin-flip.
Most notably, since 1900, the Dow held ONLY during September and October has declined -73.7%. Still, since reaching a low in 2008, this period has seen the Dow gain +26.3% so it has not worked as well as a warning in recent years.
SOME FUNDAMENTAL WORRIES
Jay discussed our Macro Index Model, which is designed to track the ebb and flow of an economic expansion. To differentiate "temporary slowdowns" from real problems, we look for SIGNIFICANT macro deterioration. Our Macro Index combines 11 diverse economic indicators to determine the state of the U.S. economy.
- New Home Sales
- Housing Starts
- Building Permits
- Initial Claims
- Continued Claims
- Heavy Truck Sales
- 10 year - 3-month Treasury yield curve
- S&P 500 vs. its 10-month moving average
- ISM manufacturing PMI
- Margin debt
- Year-over-year headline inflation
At the end of August, the Macro Index Model dropped from .636 to 0.455.
Historically, S&P 500 Index returns in the year ahead following a drop below 0.46 have been lukewarm at best and certainly well below average. However, one mitigating factor is that this is only the 4th time that it's dropped this low while the S&P 500 was still trading above its 10-month moving average. Each of the other three saw the S&P trade even higher 9 months later.
BUT BREADTH CONTINUES TO RECOVER
After weeks, even months, of sketchy internal participation, Dean notes that the NASDAQ Composite Index ratio adjusted McClellan oscillator registered an oversold to overbought reversal signal at the close of trading on Thursday. The reversal signal occurs when the oscillator cycles from below -50 to above 50 in ten trading days or fewer.
The results look pretty good when the NASDAQ Composite is above its 200-day moving average, as is the case now.
Forward returns tend to be better when the McClellan Oscillator reversal signal occurs after a correction phase. However, breadth thrust signals under most market conditions are rarely outright bearish.
STOCKS AND SECTORS - FINANCIALS
Financial stocks have had a big thrust in recent weeks. Within 30 days, the sector cycled from having fewer than 20% of its members above their medium-term averages to having more than 95% of them above.
Whenever medium-term participation in Financials cycles from a low to a very high level in a relatively short number of days, the sector has tended to keep going. There have been 42 such signals over the past 70 years, and 38 of them led to a positive one-year total return (with one of the losses being a minuscule -0.3%).
Returns were just as good (or better) when Financials were at or near new highs at the time of similar thrusts. The most notable aspect of future returns was typically how little risk was involved.
It's interesting to note that the two losing signals started losing almost immediately, with very little upside. The five winning signals started winning almost right away, with minimal downside. It's tough to read that much into such a tiny sample size, but it would suggest that if Financials can keep buyers' interest in the week(s) ahead, then it bodes well for the longer term as well.
STOCKS AND SECTORS - KOREA
Dean showed that the percentage of Korea Stock Exchange KOSPI Index members trading below their respective lower Bollinger Band registered a mean reversion buy signal at the close of trading on 8/27/21.
With South Korea joining Taiwan, we now have two countries in the Asia Pacific region on a Bollinger band signal.
The table below shows forward returns in the Kospi after these volatility band signals.
The 1-2 week results look weak. That's not a surprise, as mean reversion signals can be accompanied by choppy conditions when a market is in the process of bottoming. Results on an intermediate to long-term basis look solid and the 3-month period has registered 15 consecutive winning trades.
STOCKS AND SECTORS - REAL ESTATE
Jay noted that the real estate sector has just done something fairly rare. Ticker IYR (iShares U.S. Real Estate ETF ) just closed higher for the 10th consecutive month.
The chart below displays the Annual Seasonal Trend for ticker IYR. While the short-term seasonal trend is still slightly favorable for a few more weeks, there is, um, slight potential for trouble in the months ahead.
The chart below displays the cumulative % +(-) for the S&P 500 Real Estate sector since 1953 if held ONLY from the close on Trading Day of Year #179 through the close on Trading Day of the Year #208.
Note that real estate DOES NOT decline every year during this time period; the problem is that when this sector is bad, it tends to be very bad.
COMMODITIES | ||
Jay highlighted the trouble that soybeans have had in September.
The chart below displays the Annual Seasonal Trend for soybeans in the top clip and the month-by-month Average Return in the bottom clip.
On a solely seasonal basis, the period just ahead is a perilous one for soybeans as September is historically the 2nd worst month of the year.
From the close on the last trading day of August through the close on the 3rd trading day of October, soybeans showed a gain 22 times (36% of the time) and there were a series of significantly large declines between 2008 and 2014. However, the last 4 years in a row have seen beans advance during this purportedly "Unfavorable" period, so nothing is 100% certain.
CHALLENGING TIME FOR CORN
Dean pointed out that Jay detailed the seasonal summer headwinds for corn in several notes in the last few months. The commodity peaked in May and has now declined by a little more than 30% as of 8/31/21. If I apply a 63-day or 3-month rate of change to corn futures, we see that corn was down over 20% during the seasonally weak June to August timeframe.
With seasonality about to turn more favorable, should we be thinking about bottom fishing in corn futures?
Let's conduct a study to assess the forward return outlook for corn futures when the 63-day rate of change is down 20% or more at the end of August. i.e., the current setup. The results look weak except for the 2016 instance.
COPPER CHALLENGED, TOO
Jay noted that while subjective, fundamental factors point to a rally in copper, several quantifiable factors may influence copper in the opposite direction.
The chart below displays the Annual Seasonal Trend for copper. We're on the onset of a seasonally unfavorable period.
Also, the 50-day average of copper's Optimism Index dropped below 65 at the end of June and has been declining mostly since then. Historically, copper's 6- and 12-month results have been negative after signals like this.
The "perception" seems to be that copper is headed higher as a key industrial metal. And if inflation and/or supply/demand issues play out a certain way, it may well be.
But if copper is going to break down, several objective measures suggest that that could happen in the months directly ahead.