Big Tech Has A Lot Farther To Fall – Seeking Alpha

Where there's smoke...

Getty Images/Getty Images News

A year ago, I argued that tech stocks would likely underperform the broader market over the subsequent seven years and that leadership would likely pass to more cyclical sectors, most notably energy, and I reiterated that in October on Seeking Alpha. Particularly with the price action year-to-date, the energy sector has outperformed the Nasdaq by about 30%, and this is likely only the beginning. In this article, I am going to take things a step further by arguing that the Nasdaq 100 will likely be sharply negative over both the short-term (the next two to three years) and the long-term (the next seven years).

First, I will briefly summarize and update the long-term sectoral thesis, and then place this thesis in the context of my long-term negative outlook on the wider market. I will then augment this with an analysis of the tech sector's current valuation and finally write about what shorter-term cyclical indicators are signaling.

There are a few key indicators of when a long-term sectoral rotation is about to commence.

First, a high degree of sectoral dispersion in long-term performance.

Chart A. Sectoral dispersion leads to sectoral rotation (Own calculations from Fama-French data)

This chart is based on the Fama-French 12-sector large-cap data going back to 1926. The shaded areas represent the respective performances of these various sectors relative to an average of all twelve. The green Business Equipment sector (BusEq) is our proxy for big tech, although it should be noted that in many of the previous peaks, other sectors also participated. For example, in the dot.com boom of the 1990s, Telecommunications outperformed much of the market. In the late 1920s, although a lack of data makes it difficult to be sure, Utilities were strong. And in the last few years, the Durable goods sector (roughly equivalent to the more modern Consumer Discretionary index (XLY)) has exploded almost entirely under the leadership of the auto sector (most notably Tesla (TSLA). At any rate, the Business Equipment sector is composed of "computers, software, and electronic equipment". It is the most consistent proxy for the tech sector.

The black line is a measure of the sectoral dispersion. This has declined modestly over the last year but remains elevated. These divergences almost always involve either the tech or energy sectors and frequently both. In fact, if we strip out every other sector other than Business Equipment, Energy, and Health (which often seems to play the role of caretaker in the transition from one sectoral regime to another), we can observe this relationship between these key sectors and the broader sectoral divergence.

Chart B. Sectoral divergences involve the usual sectoral suspects (Own calculations from Fama-French data)

The red line in this chart is the same thing as the black line in Chart A.

Peaks in the standard deviations tend to precede reversals in the sectoral hierarchy, especially when tech and energy have been moving in opposite directions, as they have in recent years.

Compare these patterns to those in the following chart.

Chart C. Sector rotations and sectoral hierarchies (Own calculations from Fama-French data)

In this chart, the red line measures the correlation between the sectoral hierarchy of the previous seven years with that of the subsequent seven years. The right-hand axis is inverted. Where the red line is high on the chart, the correlation is strongly negative, suggesting that hierarchies were almost entirely inverted in the following seven years.

Notice the similarities between the sectoral dispersions and the sectoral rotations in charts B and C.

This is highlighted in the following chart.

Chart D. Sectoral dispersions versus sectoral rotations (Own calculations from Fama-French data)

As I argued in my previous articles, the energy sector appears to be the most critical one. Not only is it involved in nearly every major rotation, either as king (such as in the 2000s) or vassal (as in the 2010s), but it seems to have a key relationship with the performance of the overall market.

Chart E. Sectors have relationships with overall market performance (Own calculations from Fama-French data)

Notice, for example, in Chart E that, typically, when the energy sector is strong on a relative basis, the broader market is underperforming. The light red line represents an equal-weighted index of the twelve sectors, essentially a proxy for the S&P Composite.

This is illustrated in the following table. The relative performance of the Energy sector is negatively correlated with the absolute performance of the FF12 Index.

Table 1. Relative performance in energy stocks is negatively correlated with stock market's absolute performance (Own calculations from Fama-French data)

Notice also in Chart E that the Business Equipment sector's peak performances tend to occur when the broader index is strong, and especially at the conclusion of bull markets. The table shows that the relative performance of the Business Equipment sector is most strongly correlated with the overall performance of the market.

Table 2. Relative performance in tech stocks is positively correlated with stock market index (Own calculations from Fama-French)

Although this is a separate issue from the one at hand, the relative performances of the Business Equipment sector and the Nondurable sector (this appears to be composed almost entirely of food producers and would be closely aligned with the defensive Consumer Staples sector) have always been inversely correlated.

Chart F. Relative performances in staples and tech are strongly inversely correlated (Own calculations from Fama-French data)

There is, finally, one more signal that a sector rotation is about to begin.

Chart G. Oil shocks tend to occur at the beginning of sectoral rotations (Own calculations from Fama-French data; St Louis Fed)

Here again we have historical sectoral rotations, but this time set beside year-on-year changes in the price of crude oil. Not every spike in oil is followed by a sector rotation, but every sector rotation has begun with a spike in oil, at least for as far back as we have data.

The reason for this appears to be that transitions from one energy regime to another (bullish to bearish, or bearish to bullish, or bearish even to flatish) has apparently always coincided with a sudden spike in oil prices. Thus, the spikes in oil prices in 2008 and 2010 were the conclusions of the oil boom of the 2000s, and the oil spike of 2021 is the end of a heretofore unrelenting bear market. Incidentally, although I think the historical evidence points to a relative outperformance of the energy sector in the 2020s, as I argued in my recent piece on the Energy sector, I believe we are transitioning not from a secular bear market to a secular bull market in energy but from a secular bear market to a secular flat market. You can find that piece here.

This is my sectoral transition thesis in a nutshell. The update shows that some of these pressures have perhaps eased slightly in recent months, but we are effectively at the same place we were last year. Dispersions are still high. Tech sector performance has softened but not changed the order of the hierarchy, and oil prices are still at 'shock' levels.

Keep in mind, renewed strength in the energy sector will likely imply declining absolute performance in the wider market and relative weakness in the tech sector. This is also suggested by the tech sector's history of peaking right at the conclusion of secular bull markets in the broader market.

So, if the S&P 500 were to remain flat for the remainder of the decade, we might expect tech stocks to be negative.

But, how bad are things likely to get?

My target for the S&P 500 in the Year 2029 (that is, seven years from now) is 3000. This is based on extrapolations of the history of growth and valuations in the S&P Composite index since 1871, which I described in "The Death of Irrational Exuberance".

Chart H. Historically implied price and earnings levels for 2020s (John Overstreet)

This is, in essence, a 1930s-style scenario. As difficult as it is to imagine now with inflation at 7.5%, it appears to me that we are still in a low-inflation/low-yield regime and that bouts of high cyclical inflation are likely to be followed by deflationary, or at least disinflationary, shocks.

As I argued in the irrational exuberance series, history suggests that bouts of high earnings growth in low-inflation/low-yield regimes tend to be followed by growth shocks. That will be the first blow to markets. What follows is powerful cyclical swings in growth and inflation. The second blow will be when one of those swings catches fire, raising growth, inflation, and yields but lowering PE multiples.

Now, if the S&P 500 is going to be as low as 3000 in 2029, that would be 33% lower than current levels. If we use a relatively modest estimate of the degree to which tech stocks will underperform the broader indices, say 25% (the relative rate of Business Equipment from 1929-1936), that would put the Nasdaq 100 index 50% below recent levels, somewhere in the neighborhood of 7000-8000 points.

If that seems unduly bearish, recall that the S&P 500 was effectively flat during the 2000s while the Nasdaq 100 ranged between -50% and -80% from 2000 to 2010.

This is not to say that the tech sector will not be profitable over the coming decade.

One thing I read in comments sections of articles about tech stocks is that the difference between 2000 and today is that tech stocks are now able to produce cash, but I am not sure that is a bullish sign.

I looked at long-term growth in dividends for each of the Fama-French 12 sectors, the closest historical proxy I can find for earnings growth, and their respective relationships with their dividend yields (the closest proxy I could find for valuations) and price performance (and implicitly total returns).

What I found was that the Business Equipment sector was the only sector to have a consistently negative correlation between 7.5-year dividend growth and price performance. The dividend yield also had a strong negative correlation with that dividend growth (a close third place among the twelve), suggesting that high valuations such as we have now are followed by high rates of earnings growth. The yield also had a middling correlation with price performance and a strong correlation with total returns.

In sum, high valuations in tech stocks had the unique combination among the twelve sectors of resulting in high growth and low returns. Note that the dividend yield is not as low as it was in 2000 in the Business Equipment sector but that the Durable goods sector has nearly reached that level.

Chart I. Dividend yields for tech stocks are approaching record lows (Own calculations from Fama-French)

The tech sector may be able to produce cash over the next decade, but this will likely be more than offset by a collapse in stock prices.

Over the short-term (say over the next two to three years), I see three indicators pointing to negative returns. Two of them are applicable to markets in general and one is applicable to the Nasdaq 100 specifically.

The first has to do with the interplay between growth and valuations described in the "The Death of Irrational Exuberance" series and suggests that the S&P 500 could go as low as 2500 points over the next two years.

The second has to do with the spike in oil prices that we described above. I showed that spikes in oil prices typically occur during transitions in energy price regimes and sectoral hierarchies. But, these spikes also tend to occur before general market crashes. As I wrote in previous treatments of sectoral rotations, it almost appears as if stock market crashes are part of the ecology of long-term rotations.

Chart J. Stock markets tend to crash around the time that rotations begin (Own calculations from Fama-French and Shiller data)

In this chart, I illustrate the relationship between rotations and stock market crashes. The latter tend to occur early on in the process of the former, and energy seems to be part of what links them.

Below is the relationship between sectoral divergences and market crashes.

Chart K. Stock market crashes seem to help resolve sectoral dispersions (Own calculations from Fama-French data and Shiller data)

From this angle, it appears that market crashes reduce the 'pressure' of sectoral divergence. By the time the crash has run its course, the divergence is reduced to much more normal levels. One might imagine, then, that those sectors that were most in favor during the boom might be the most severely punished during the crash.

In any case, the point is that markets are likely to see negative returns over the short-term, and it is unlikely that tech stocks will be immune from this.

The third reason I am bearish on tech stocks and the Nasdaq 100 in particular over the short-term is because of a set of short-term algorithms I have been working on over the last two weeks. Each of these sets is now signaling 'sell'.

Having backtested these algorithms against roughly 150 data sets spanning stocks, commodities, and bonds up to 150 years back, I feel rather confident that they are robust. (If there is sufficient interest in the comment section, I would like to write about them in greater detail in the future).

The following chart, perhaps somewhat unorthodox in its presentation, shows the log value of the average monthly alpha produced by one of these algorithms in the Business Equipment sector. Since these are log values, a '0' is equivalent to an alpha of '1'. It is not easy to think in terms of log values. At these levels, however, the percentage increases are roughly equivalent to the log values. That is, a log value of 0.01 is about 1% per month.

Chart L. Seeking alpha in tech stocks (Own calculations from Fama-French data)

There are a number of things I can say about this data.

First, alpha here is defined as being relative to the Business Equipment index itself rather than a benchmark index like the S&P 500. That is, implementing this formula on the Business Equipment index has only been demonstrated to beat the Business Equipment index. Second, over the last 90 years or so, if this algorithm were blindly followed, it would produce positive alpha over the entire period. The average alpha of the entire period is just over 1.4% per year over the last 90 years. Third, that alpha is generated primarily during bear markets. Fourth, it appears that inflationary bear markets like the 1970s permit less alpha than deflationary or lowflationary regimes as in the 1930s and 2000s, respectively. Fifth, it can be finnicky at tops and bottoms. That is, it can blink on and off from month to month at transitions, so it does not allow one to simply "call a top". Therefore, next week or next month, this signal could change. For now, however, this is where they stand.

It suggests that investors should shift to less risky equities (almost anything other than tech at this point, but staples (XLP) might hold up well in light of Chart F) or even commodities or bonds. I have been shorting the Nasdaq 100 through the QID ETF since November and because of the size of the downside risk, I added SOXS last week based on my reading of these tea leaves. I also balanced this with positions in XLRE and CMDY, even though I am bearish on commodities.

It appears to me that nearly everything, long term and short term, is stacked against the tech sector. That does not mean that tech companies will no longer continue to produce great products for customers or cash for investors, but that does not necessarily translate into price performance or total returns. I think another 40% down is likely, probably with even sharper cuts along the way. Eventually, the downward pressure being felt in tech stocks will spill over into other sectors and asset classes, but for now those places are less risky than tech.

Excerpt from:

Big Tech Has A Lot Farther To Fall - Seeking Alpha