It’s been a while since I’ve written about the market overall, mostly because I shouldn’t be advising anyone – please don’t take this as advice – and partly because I’ve been mystified by what’s going on – please see digression #1, above.
So what has been going on? Here’s the DJIA over the last six months:
It recently hit a local high, and then took a bit of a plunge, before recovering slightly on Friday.
I tend to see the market as relentlessly forward looking, projecting confidence in the future by inflating prices, and concern about the future by deflating prices. Sometimes stocks “get ahead of themselves,” and we’ll see a drop in the middle of a boom – that’s the nature of the beast.
But it’s important to note that the market, more and more, is controlled not only by investor expectations, but by computer algorithms as well. We can, for fun, split the investors into three groups, based on my scanty knowledge of investment technology these days, but extrapolating from what little I do know and how people use new technologies.
- Traditional investors, such as myself. We invest for the long term, and computers are used for little more than implementing how stocks are bought, held for a while, and sold, as well as trivial portfolio management. The computers are a necessary convenience, but not an active assistant in shaping investing tactics and strategies.
- Short-term Investors who use first generation computer algorithms to shape their strategies. These algorithms, which no doubt are being driven into obsolescence, are the result of investors and programmers teaming up to create algorithms that survey market conditions and data about the business world, ranging from numerical results to analysis of news articles, and, making mechanical predictions, execute trades in response. Successful systems include, if I’ve heard rightly, the conservative Mercer family, who made a fortune in this arena.
- Short-term Investors who use second generation computer algorithms, or ML (machine learning) systems. These systems learn the rhythms of the stock market using advanced ML, again using numerical results and semantic analysis of news articles. The trick is in the ML algorithms, which will typically see patterns missed by human analysts.
Now, I just wrote that down to point out that all three of these groups are in a situation new to them: the Covid-19 pandemic, including the governmental response, followed by the well-known economic downturn.
Now, to my naive mind, any sober adult who’s been following the situation should be well aware that we’re not yet at the terminus of the pandemic. Neither effective vaccines nor treatments have been identified, although they may be in the pipeline. But how are algorithms and ML systems to know and evaluate this information? That is, how can they project and trade in such a way as to leave investors, ummm, happy?
Yeah. The first generation algorithms are not going to do well, as they’re essentially static.
The ML systems will also not do well, because they don’t have a comparable situation to learn from.
For that matter, investors who do not take advantage of the two generations of computer support may also do poorly, because they don’t know how this is all going to work out, either.
I’ve been viewing the market results as driven by short-term traders, mostly via computer algorithms, who do not know how to properly evaluate the markets and the economy. It’s like a long sugar-rush, and I worry what will happen when some of these highly valued companies fail to deliver results promised in their stock prices. I suspect a lot of people will get hurt.
But even more, when these computers semantically analyse these news releases, what the hell are they doing with those from President Trump? The man lies compulsively, but I must confess that I do not know how ML systems deal with information that is almost certainly false – but comes from a source that can almost uniquely move the markets.
I’m left wondering if the markets have been pushed higher simply because the underlying agents simply don’t know how to properly manage information deriving from President Trump.
And that’s scaring me.
