Give the Robots Our Money
Now that we are firing up robot brains better at stock-picking than humans, thanks to another fairly loathsome cyber-eugenicist, one might ask:
If computers become better at picking investments, what is the purpose of reward incentives for investments? In other words, what is the justification for capitalism - privately held wealth invested in speculative enterprises -at all? Computers do not require the future promise of cash, women, and cars weighed against the risk of failure to make sound investments. They have no need for security in old age or the benefits of fine living. They are not swayed by toxic ideologies or tasteful fund marketing. They do not benefit from enhanced social status, partying with Paris Hilton, or getting their children into the better sort of pee-wee lacross team.
They do not desire. They are instructed to perform. They seek a set of numbers.
Let's say this works. By comparison, human investors would become hideously expensive, unreliable, unstable, over-tanned, and still apt to run off to the Caymans in their caviar-fueled yachts. The robots wouldn't overvalue outsourcing, layoffs, and inequality or undervalue externalities-like the continuation of life on earth- based on factless cultural ideologies of the market.
If the humans won't get in gear, give the robots our money.
posted by JAB at 10:17 AM
1 Comments:
Dr. X posts this during a break from writing his "Crowded Highway: The Cliche of American Blues Authenticity", to be published next autumn in the Journal of the American Musicological Society:
"I don't know much about this, of course, as my formal education training is in the field of qualitative musicology. But one or two points seem self-evident.
"First, computers beat humans at chess, too. But a human working with the assistance of a computer can easily beat a computer alone. The two intelligences are complementary - one data-driven and computational, the other based on pattern recognition and configural analysis.
"Second, it is apparent that the wide use of program trading has led to an arms race on Wall Street, in which the top firms race to bring in physicists who will build battle-bots to crush the other guy's battlebots. As Charlie Munger (Buffett's partner) has observed, we are increasingly seeing 'algorithms vs. algorithms.'
"But, as Benjamin Graham said long ago, in the short run the stock market is a voting machine, in the long run it is a weighing machine. These algorithms tend to be focused on the short term (1 day to 3 three months as the most), and as they contend are bound to introduce inefficiencies that longer-term traders could exploit.
"It is even possible that they could over-arbitrage short-term investment factors such as estimate revision, earnings surprise, and market reaction to earnings announcements, leading them to become negative, rather than positive metrics for performance over a 3-6 month forecast period. If that were true I would imagine one could find numerous opportunities in the market for volatility with astute purchases of straddles and strangles against the underlying sotck, or even, possibly, synthesize the desired derivatives exposure through judicious employment of convertible arbitrage.
"But maybe investors should think longer-term. It is probably not a coincidence that the most successful active managers over the past ten years have had a multi-year investment horizon. For the patient investor, anything that introduces noise to the markets is a chance to make money, so I imagine they are not too concerned about these developments.
"Perhaps we could invite a professional investor of some kind to say more about this."
Post a Comment
<< Home