The Case for Not Taking the Models too Seriously
“A change in perspective is worth 80 IQ points.” At least according to Alan Kay, the man who pioneered windowing graphic user interface. More plainly, he’s the reason you can point and click your computer into submission without knowing the first thing about programming. I suppose that I should be quoting the late Jim Simons here, the investor who pioneered computer-based trading and introduced the quants to the world in the 1980s. Simons certainly had a spare 80 or so IQ points at his disposal, but a generation later we’re all quants now. And that may be a problem.
On Friday, the jobs numbers were better than expected, and the figures on unemployment were worse. None of which is going to make this week’s Fed meeting easier despite the fact that it is almost certain that nothing much will happen. The conflicting picture can be explained – but not necessarily reconciled - by the fact that the models draw from different pools of data. Payroll data sets are much larger so you might think it paints the more accurate picture, but that doesn’t exactly stand to reason.
The numbers going into the models are probably accurate as can be expected. Barring some “deep state” conspiracy nonsense, there is no reason for government to lie about that sort of thing; it’s the way the data are processed. Understand that the larger the data set, the easier it is to draw the wrong conclusion because every time you average, or multiply or even add something you lose information about the situation on the ground. The more complicated the model, the easier this fact is to obscure.
Let’s take the average net worth of a 14-year old white girl. Or her parents, because those Taylor Swift tickets aren’t cheap. Now let’s fill an NBA arena with the little knee-biters and whomever had to drive them. The average net-worth of the parents will likely be considerably higher than that of the surrounding zip codes – that paints a reliable picture. When the lady of the hour enters, the average net worth of everyone in the arena goes up by about a million dollars a head – including the unemployable middle schooler. At least on paper. In reality, the only person who is likely to be a million dollars better off by Taylor Swift entering the arena is Taylor Swift.
Maybe the analogy is overly simple, but the reliance on overly complex models causes similar the blind spots. A more pressing example on the effects of an outsized outliers is the S&P500 and the Nasdaq. You could make a fair argument that we’ve got four Taylor Swifts in the class, which will warp the average plenty, but the cap-weighted S&P500 grades on a curve. In a normal situation you’d expect roughly 250 of the listed firms to underperform the index average – right now 350 firms are underperforming – about 20% higher. Throw the outliers out, or even don’t weigh the thing for a market cap, and the stock market looks... just fine, but not all that remarkable.
Now, the stock market and the economy are two different creatures, and they are not nearly as corelated as people think that they are. This insight is about neither, really. It’s about the dangers of hiding behind bad math.
Economist, investment bankers and other spreadsheet jockeys, dealing as they do with infinitely quantifiable money, has caused them to place inordinate faith in the computable data that fit easily onto models that produce accurate “narrow context” output. Which is all you need to sort out P&L, cash flow et al. Never, I repeat, never hire a novelist as a CFO. On the other hand, the World Bank announced last week that they’d tweaked the math and found an extra $7 trillion in the global GDP... so...
Adam Smith wrote one of the few economics books that has aged well. Or, it’s concepts and theories have aged well in that they’ve been repeatedly proved correct over the centuries. He filled 1,700 pages of economic theory without a single graph. Just words... so many words. Somewhere in that Scottish enlightenment thinking that is so precise as to be almost garble, he mentions the “invisible hand” of the market. And to understand that aspect of the market you’d do well to pull in someone – a novelist (only one), or an architect, or a computer graphics designer – anyone to offer a different perspective.
Why? Because while the market’s invisible hand is guided by enlightened self-interest, this doesn’t mean what economists think it does. If the people who make up the market were that calculating, we wouldn’t have asset bubbles, and that girl who is “so crazy it’s hot” wouldn’t be a thing. She is not in anyone’s enlightened self-interest, let alone hers. What’s really moving the invisible hand is an unholy concoction of hype, fashion, ego and risk-aversion counter-balancing the very human fear of missing out. To plot that on a global economic any model would need to map the human heart, superimpose a sort of nonbinding logical algorithm, throw in a parental baggage multiplier, and then make the model reflexive to the entire population of the planet. How do you do that?
Evidently, you don’t. Just because the analysts got things so wrong last year doesn’t mean they’ll get it right this year. The market isn’t dealing with a distorting anomaly, but a suite of them – market lockdown, a snap back, two rounds of massive stimulus triggering the worst inflation in a generation, followed by steepest rate hikes in the same – and no one should think it’s all worked its way out of the system.
In the past 11 Fed hike cycles, a recession has started about two years out. This round started in March of 2022 – which would have put us in March of this year. That’s on average – and you know what Taylor Swift, Nvidia or that crazy gal can do to an average.
So how do you sort things out without the math? Again, you don’t. Not well at any rate. Just understand that you won’t sort it out entirely with the financial models either. You need a different perspective. Alan Kay, the graphics programmer mentioned a thousand words back, was right: It will give you another 80 IQ points. Jim Simons, the father of the quants, was a different perspective in his day. A generation later, that revolutionary quantitative thinking dominates to the exclusion of all else. And it shows: economist and analysts have been consistently wrong for close to three years now, and are still operating in blind spots. The good news is that the innovation in the markets and investing is in the very blind-spot this blinkered thinking produces. The bad news is that so are all the bat-guano failures.
Sadly, in the words of John Maynard Keynes “worldly wisdom teaches that it is often better for the reputation to fail conventionally than to succeed unconventionally.”
Hell, he’s probably right.
Comments