A Sentimental Journey
How bullish are we feeling? If the American Association of Individual Investors’ weekly survey is anything to go by, not very. It has been carried out for decades and is much followed as a measure of sentiment. Retail investors are simply asked if they’re feeling bullish, bearish, or neither. And this week, fewer called themselves bullish than in any week since September 1992:
This is quite something. There were more bulls after 9/11, during the Covid lockdown, and in the throes of the Global Financial Crisis, than there are now. Most of us can intuit that sentiment is weak and shaken at present. The nastiness of pandemic-era life, and now the war in Ukraine and the sharp rise in inflation, will see to that. But lack of bullishness on this scale is startling.
A more complete measure of sentiment takes the spread between bulls and bears. On this basis, sentiment looks far more nuanced, as many investors are noncommittal when answering the questionnaire, and the proportion calling themselves bearish is still just under 50%:
I’ll admit to being very fond of the AAII as a market-timing sentiment gauge, as I used it in what turned out to be one of my most fortuitously timed columns ever. On March 9, 2009, the day the great post-GFC rally started, I greeted my readers with the words: “Perhaps the greatest reason for hope at present is that almost all hope seems to have been lost.” The generational low for the AAII bulls-minus-bears measure, plainly visible in the chart, was the center of my argument. As it made me look good, I tend to be thankful to the survey.
But how good a buying signal does the AAII survey send, really? The charts above show some obvious extreme moments. Looking just at the bulls, there were big dips in September 2002, and again in April 2005. Using the bulls-bears measure, the buying signals flashed in October 1990 (when Saddam Hussein had occupied Kuwait and there was great uncertainty over how he could be moved out) and March 2009, the nadir of the GFC. Here is how the S&P 500 performed in the five years (please disregard the calendar years on the x-axis) after each of those sentiment buying signals:
It looks as though my March 2009 call was lucky rather than anything else. The subsequent rally was on a different scale from the other three. But all three were good short- and medium-term times to buy, with the S&P up after one and three years. (If you bought at the previous low-bullishness landmark in 2005, you would have been caught up in the GFC, and lost money after five).
The sample size is far too small to make any conclusions. On balance it seems fair to suggest that the lack of bulls is indeed somewhat bullish, and needs to be listed on the positive side of the ledger for anyone considering whether to buy. If there are that many people unconvinced, good news should have more of a positive effect on share values as time goes by. And I suppose the buy signal this measure sent in September 1992, not regarded as any particularly big turning point for the markets or the world, was quite a strong “win.” But I wouldn’t go much further than that.
Pile into the market now and you might look as good as I did after that call in March 2009. But that will require some luck.
Why So Few Bulls?
A more interesting question is why so few are bullish. Covid is miserable but it’s nearing its end and most keen investors should know that that is positive. Ukraine might also dampen sentiment, of course. But the regular surveys conducted by the Yale School of Management under the economist Robert Shiller suggest that valuation is a key concern. You can find an archive of the surveys, conducted regularly since 1995, here.
One question aims to measure confidence in valuations:
Stock prices in the United States, when compared with measures of true fundamental value or sensible investment value, area: [Circle 1, 2, 3, or 4]
Too low
Too high
About right
Do not know
The following chart shows how the proportion answering “too low” or “about right” has moved as a proportion of all those who expressed an opinion. Both institutions and particularly individual investors are far less confident that the market is fairly valued than they used to be. Put differently, the proportion that think the stock market is too expensive is very high. These people cannot be happy about rising bond yields. So it’s fair to suggest that there would be appetite to exit stocks if the bond selloff resumes:
For what looks like a great contrarian indicator, try the Shiller “crash confidence” index. This asks respondents to put a percentage probability on “a catastrophic stock market crash in the U.S., like that of Oct. 28, 1929, or Oct. 19, 1987, in the next six months,” including one that starts outside the U.S. The index traces the proportion that think the chance of such a catastrophe is 10% or less. Higher readings denote greater confidence that there won’t be a crash:
Particularly among individuals, a lot of investors find an imminent crash easy to imagine at present. That doesn’t accord with the still-average levels of bearishness in the AAII survey. And it may not be bad news anyway, because on the limited evidence available, this would be a contrarian indicator. There was indeed one 1929-scale crash in the last two decades, at the end of 2008. But confidence remained quite high that a crash could be avoided as late as that summer. In early 2007, when with hindsight we can see that all the warning signals were already in place, crash confidence hit its highest ever.
It’s possible, even, that the much lower crash confidence since 2008 has been caused by that experience. The “availability heuristic” first promulgated by the great Israeli behavioral economists Amos Tversky and Daniel Kahneman finds that we all suffer from a tendency to use easily recalled events to estimate the probability of an event occurring. The days when prices crashed are, by a country mile, the days I most easily recall from my career covering the market, and I imagine that’s true for most of us. Once the crisis following the fall of Lehman Brothers gave us a crop of very powerful memories of what a crash was like, it’s no surprise that we overestimate the probability that it happens again.
Judging the probability of unlikely but damaging events (tail risk) has always been a problem for markets. As none of us has ever experienced nuclear Armageddon, it’s possible we’re under-estimating such a risk; as most of us have clear memories of a market crash, the chances are that this is a risk we’re overestimating. That in turn creates more opportunity for stocks to rise from here. But that leads us to a special case:
The Netflix Crash
The collapse in the share price of the dominant video-streaming group Netflix Inc. is now a phenomenon for the ages. Tuesday night’s news that its subscriber base declined very slightly in the first quarter was enough to spark a 33% selloff when the market reopened. Netflix was worth $306 billion as recently as November last year, and it was bigger than Walt Disney Co. Now, it’s worth $100.5 billion. Its value has crashed after both its last two earnings reports:
What most alarmed the markets was the news that Netflix’s subscriber base had declined by 200,000, quarter-on-quarter. This was its first decline in more than a decade, and came as a reminder that growth couldn’t be assumed to be perpetual. But the fall in the share price wasn’t so much about the underlying figures, and more about the valuation that the market was prepared to put on. As we’ve seen, confidence in valuations is low, and Netflix suffered from this.
The following chart shows the price/earnings before interest, tax, depreciation and amortization ratio for Netflix since its initial public offering, compared to the S&P. As it has gapped lower today, you might not be able to make out the final number; Netflix now trades at 15.5 times Ebitda while the market trades at 13.5 times. Substantially all the valuation premium that Netflix used to command because of the perception that it offered perpetual growth into the future has now been removed:
How can the market possibly have marked the value down so much so quickly? I suspect much depends on another concept from Kahneman and Tversky — loss aversion, or “prospect theory.” We don’t like losses, and we care about them far more than about gains of the same magnitudes. An outright loss will always grab our attention more than a declining gain.
Netflix’s subscriber figure was almost universally reported as a fall of 200,000, which sounds like a lot. Nobody expressed it as a percentage. The previous figure was 221,840,000, so the fall in the first quarter was 0.09%. That’s a rounding error. Given that Netflix had to close its Russian operation during the quarter, and that some falloff in subscriptions was surely inevitable after the bulge caused by the pandemic, it really doesn’t seem that bad. Had Netflix’s subscribers risen by 200,000, it’s a fair bet that this would (reasonably) have been described as “static.”
But that’s not how it went. There’s a lack of bulls, and rising concern that investors have left themselves too exposed to duration risk. When a company’s value is locked in the future, like that of Netflix, then rising interest rates or declines in projected growth will have an outsized effect on valuation. Combine that with the natural human aversion to anything that can be called a loss, even if it’s only 0.09%, and you have the screeching accident that Netflix has suddenly become.
Allez la France
No news can be very good news. As far as global markets, that’s certainly true of French politics, which has turned reassuringly boring in the run-up to Sunday’s second and conclusive round of the presidential election.
Emmanuel Macron has not turned himself into one of the great popular presidents, but his ideas about finance and about the European Union are well-established, very predictable, and much more palatable to international capital markets than those of his opponent, Marine Le Pen. At the beginning of this month, a surge in polling support for the hard-right populist Le Pen began to get markets worried. Now, taking the Predictit prediction market as a proxy for conventional wisdom, a Presidente Le Pen looks much less likely:
This is because the polling evidence steadfastly shows Macron reopening his gap over Le Pen, after a month in which he’d allowed it to narrow. The latest survey from the French polling group Elabe is typical, and shows Macron with a 9-percentage-point lead, up from only 4 percentage points on the eve of the first round:
This is having an effect on markets. The spread of French over German 10-year bond yields has narrowed by some 8 basis points in the last two weeks:
However, as the chart shows, that spread is still elevated. This is probably because investors are still applying the lessons from 2016, when unpopular technocratic options like Macron (EU membership and Hillary Clinton) fell to more popular alternatives (Brexit and Donald Trump). For a reminder, betting markets showed the chances of Britain voting to stay in the EU rising sharply in the last week before the referendum. On that occasion, the markets had it wrong, and large numbers of speculators who had piled in to bet on the outcome in the last few days of the campaign lost their shirts:
Now we come back to yet more flaws in human reasoning. Because the Brexit referendum happened, and it was a very memorable event, I am probably overstating the chance of a Le Pen victory. The polling evidence on the face of it suggests overwhelmingly that France will have another five years of President Macron. But Brexit did happen, and I’m surely not the only one who can’t help but overweight the chance of another populist surprise. If, as is very likely, Macron prevails, there will be gains to be made on Monday morning.
Survival Tips
A suggestion for devotees of British politics. At the turn of the year, the New Statesman offered a series of podcasts called Prime Ministerial. It’s an intelligent but accessible delve into what made the personalities of some of the biggest political figures in recent history, and it also gives rise to some fascinating analysis of how to judge effectiveness and leadership. Well worth listening, even if you aren’t a British political junkie.
Meanwhile, Stephen Bush, who as the New Statesman’s political editor fronted the podcast, now has a new perch as an associate editor at the Financial Times. His newsletter can be found here. And he confided on Twitter last week that he was in the final stages of sorting out his format: “We just need to finesse the last little bits that still have names like ‘that cool thing from John Authers’ email, I’ve always wanted to do something like that.’” So it looks like he has the right role model in mind.
Bloomberg News provided this article. For more articles like this please visit
bloomberg.com.
Read more articles by John Authers
