Markets have been profoundly unexciting over the few weeks since the bank collapse headlines. In dollar terms stocks have advanced, bonds mostly moved sideways, and gold has rallied. Stocks had a nice little selloff this afternoon, but much of that was recovered in the final minutes of trading.
The near term outlook is, as is usually the case, dominated by factors pulling in opposite directions. Calling which ones will prevail is always fraught with peril, but that hasn’t stopped us from speculating yet. The technical picture for stocks says the path of least resistance over the next few weeks is lower, though that resistance is far from negligible due to predominantly bearish sentiment. Bonds and gold are even less clear, but seem likely to at least outperform stocks.
Let’s focus on stocks. One of the challenges in the study of charts is determining when stocks are overbought or oversold. This is complicated by having to first address on what time frame. It’s possible for a market to be overbought on a time frame of days or weeks and oversold on a time frame of months or years. Just to cite an extreme case, at this point in April of 2009, stocks had rallied very sharply from the March lows, and were overbought on a weekly time frame. Yet in the bigger picture, they were profoundly oversold, having been decimated over the course of the preceding months.
Most Financology readers aren’t day traders. With that in mind, let’s lean towards the longer term and take a look at a chart of stocks since the turn of the millennium. Further, we normally don’t allocate between just stocks and cash. Realistically, we may adjust our exposure to stocks versus traditional defensive assets like Treasuries and gold. So for this purpose, we’ll plot stocks in terms of a very conservative portfolio of Treasuries and gold, three quarters to one quarter. The Treasuries allocation is assumed to be half in one year bills and half in ten year bonds. As a practical matter, this would correspond closely to about 5% cash, 70% in the broad treasury fund GOVT, and 25% in a gold bullion fund such as IAU. Stocks themselves would be represented by the comprehensive global index fund VT.
Another twist here is that we’re detrending the result. We want to see whether and to what degree stocks may have “gotten ahead of themselves” or are “low” in relation to their broader trend. Strictly speaking, we need to compare current levels with both past and future levels to do this. This of course is fine if you’re only interested in history, but if you’re trying to assess the current positioning of the stock market, it’s of no help. You don’t have future levels to input into your formulas.
So instead what we’ve done here is apply a cubic array of exponential moving averages. The methodology is akin to that used by technicians to construct oscillators, but the use of three time constants allows us to go a step further and ensure that there is no persistent upside or downside bias. It’s not quite the “noncausal” ideal we might use if we had future data available, but it behaves similarly in that the net area under and over the curve tends to zero … it’s a mean reverting series.
From the following two-plus-decade chart we can see that historically buying when the plot is less than zero and selling when it has been greater than zero has been rewarding. The further below or above zero, the better. No forecasts are involved, just a mathematical evaluation of past data. The data is purely technical and based only on the returns of stocks, bonds and gold. As with Synthetic Systems, they are total return data, so that the real world assets are put on an equal footing … there is no need to separately consider dividend yields, bond yields, or the non-yielding nature of gold. There are no fundamentals considered either, so it is best regarded not as a substitute for SS or for one’s fundamental views or valuations, but as a technical supplement to these. Think of it as an attempt to assess whether and how much stocks are “high” or “low” in relation to their prevailing established trend.
Since periods spent above or below the zero line tend to last some time, buying or selling would be done not all at once, but measured out in a series of small purchases and sales, ideally on dips and rallies. If anything this indicator tends to lean to the early side, so there is ample time to respond.
The light dashed line is the same data before detrending.
What do we see? As of today stocks are a bit below the zero line, having partially recovered from the more deeply oversold state of last fall. On this indicator alone, some modest buying would be justified.
Of course fundamentals may differ. My primary concern, as I’ve discussed several times over the past year, is the inversion of the yield curve. To briefly summarize, yield curve inversion is a bearish omen. But not immediately so. The more imminent bearish flag is when short rates begin to fall, uninverting the curve. Ironically the anticipated Fed “pivot” often cited to support a rally is likely to require stocks selling off as a precondition. If this were to happen in the weeks ahead, the detrended stocks plot would turn lower, ultimately indicating better buying opportunities.
Yet that could be some time off. Based on what we know now, what’s the most likely scenario? The yield curve is positive through the first four months, but negative from there to six months. This says the bond market is pricing in lower short rates between four and six months out than from now to four months. That puts the start of declining short rates some time in August-September-October. The bond market is not assured to be correct, but is more likely so than Fed projections that there will be no rate cuts this year.
So are stocks a buy, sell or hold? As always, it depends on individual circumstances. A long term investor with no stocks at all could easily justify beginning to remedy that shortfall, while one with his entire portfolio in stocks could just as easily remedy the lack of diversification. The foregoing could serve as an aid in setting the pace.