There are two factors driving computer automation fears: complete elimination of jobs and that a “human logic” will be taken away in decision making. To the extreme, there are fears that robots will engage in a dastardly plot to dominate human free will. And then there is a reality, which might possibly be seen through the lens of increasing Bloomberg terminal subscriptions.
Bloomberg terminal sales increase is a sign humans are not extinct on Wall Street… yet
No, computers might not be plotting to take over all human trading jobs. At least, not yet.
As if to punctuate the point, perhaps the most ubiquitous sign of a computer aiding the success of the human trader is the Terminal, emphasized with a capital “T” for its importance to a relatively independent and at times aggressive news organization.
We learned today that terminal subscriptions rose to 325,017 in October, up 1,036 on a year-over-year basis, Quartz first reported. Categorizing terminal sales as a “rough proxy for hiring and firing trends in the industry,” author John Detrixhe pointed to a correlation between Terminal sales and human bond trading.
Over the years, bank analysts, Bloomberg’s largest customers, have been discussing its demise, but they seem to power back nonetheless. They have since diversified their business line to the point Terminal sales are one major revenue point along with technology. And to the surprise of the naysayers, they have flourished, as the recent uptick in sales indicates.
“Trading desk reduction and subsequent personnel upgrades (hiring more people that are tech savvy in lieu of old-school-only traders) is behind us,” Kevin McPartland, head of research for market structure and technology at Greenwich Associates was quoted as saying. “The businesses for the most part have been made as efficient as they can be and need to be.”
Taking the concept further, the Terminal sales numbers actually make a bigger point. It is at this historical juncture we can see where computers are better than humans and vice-versa. From this perspective, rising terminal sales can be a barometer of how traders might succeed in the future.
Bond trading is an example of understanding paradoxes that don’t have historical precedent
In trading, there are thought processes that work well with computers and, paradoxically, thought processes that are uniquely human.
Consider, for instance, Hayman Capital’s Kyle Bass and his famous Japanese quantitative economic modeling. The shorting of Japanese government bonds (JGB) is an example of how bond trading requires a distinctly human approach.
Computers, at present, can only provide analysis based on past history. Because the thought process is mathematically driven at its core – with Boolean logic driving execution triggers – it can have difficulty with analysis that doesn’t have a previous history. Enter Bass and his JGB trade to illustrate the point.
Bass recognizes that sometimes irrationality plays into markets and too much quantitative stimulus tilts free market scales. Its just a matter of when. The thesis is yet to be proven out — the Fed was deft under Yellen at slowly reducing the market dependency, so a market adjustment might not occur. The power of free markets is not just a slogan, but an understanding of how the law of supply and demand works in nature. Like it or not, it is an eternal mechanism to establish value. At its core, Bass appears to recognize this.
The point to recognize is the human component that a computer likely couldn’t pick up, which speaks to the thesis behind why more Bloomberg terminals are a victory for human traders.
While some of Bass’s QE analysis is in the public domain and shared by other hedge fund managers for a computer to scrape up and analyze from the standpoint of a formula, what Bass and others are betting on a computer would have trouble understanding. Bass connecting the dots on the odd effects of what QE might do to markets when they mean revert is yet to be a proven theory… yet. There is no historical precedent, thus humans armed with terminals are likely to prevail in this type of trade environment.
It wasn’t a long time ago when yield curve traders would consider central bank buying corporate bonds and the concept of negative interest rates absurd, tin-foil hat material, but this isn’t something a computer would be likely to recognize. So how can a computer pick up on the if-then connection that undue or overly aggressive free market influence by non-economic players might lead to a significant mean reversion? Maybe they could pick up pieces of this analysis, but the full trade thesis would require humans to some degree. Maybe one day they will entirely understand. But at present, most computer-based trading analysis relies on mathematical formulas – and that is the paradox of their success and weakest point.