Robot Analysts Outwit Humans on Investment Picks, Study Shows (bloomberg.com) 52
They beat us at chess and trivia, supplant jobs by the thousands, and are about to be let loose on highways and roads as chauffeurs and couriers. Now, fresh signs of robot supremacy are emerging on Wall Street in the form of machine stock analysts that make more profitable investment choices than humans. From a report: At least, that's the upshot of one of the first studies of the subject, whose preliminary results were released in January. Buy recommendations peddled by robo-analysts, which supposedly mimic what traditional equity research departments do but faster and at lower costs, outperform their flesh-and-blood counterparts over the long run, according to Indiana University professors.
"Using this type of technology to make investment recommendations or to conduct investment analyses is going to become increasingly important," Kenneth Merkley, an associate professor of accounting and one of the authors, said by phone. Whether getting stock calls right is a critical mission of human analysts is debatable. Wall Street research departments serve a variety of functions, among them connecting investors with company executives and gathering earnings and other corporate data. While their buy, sell and hold recommendations still garner attention and can move stocks, the number of clients premising investment decisions off them is probably limited. The study looked at a small and still largely experimental branch of fintech, firms founded on the premise that digital technology does a better job than humans in making equity recommendations. While all analysts use computers, a handful of start-ups has been seeing if programs can handle every aspect of the stock-picking process.
"Using this type of technology to make investment recommendations or to conduct investment analyses is going to become increasingly important," Kenneth Merkley, an associate professor of accounting and one of the authors, said by phone. Whether getting stock calls right is a critical mission of human analysts is debatable. Wall Street research departments serve a variety of functions, among them connecting investors with company executives and gathering earnings and other corporate data. While their buy, sell and hold recommendations still garner attention and can move stocks, the number of clients premising investment decisions off them is probably limited. The study looked at a small and still largely experimental branch of fintech, firms founded on the premise that digital technology does a better job than humans in making equity recommendations. While all analysts use computers, a handful of start-ups has been seeing if programs can handle every aspect of the stock-picking process.
Duh (Score:5, Insightful)
Robot gamblers perform slightly better than human gamblers. Human cheaters running the show continue to laugh.
Re: (Score:2)
And just like that a couple thousand stock analysts are looking for a new career.
Re: (Score:2)
Comment removed (Score:3)
Who needs GAs (Score:4, Insightful)
About 10 years ago an experiment was done with schoolkids who picked better stocks than the so called professionals.
All it proves is that stock market investing is about as scientific and reliable as betting on the horses.
Re:Who needs GAs (Score:5, Funny)
That experiment has been carried out numerous times using chimps, darts thrown at a page of the Wall Street Journal, and even chickens who picked stocks by crapping on them. The professional investment advisors always lose.
Re: (Score:3)
That's what gives statistical studi
Re: (Score:2)
Re: (Score:2)
Investment advisors aren't being paid to make money for you, they're being paid to make money for their employer. Why did common investors take a bath in the Dot Bomb and the Mortgage Meltdown, while their employers came out smelling like roses? Up until the last minute they were cramming their small investor funds with Enron stock, pulling it out of the house accounts hand over fist before it tanked completely.
Re: (Score:2)
Common investors took a bath because they don't know how to incorporate risk management into their investing. They follow the trends they hear others talking about or that they see on the news without any regard for the risk involved. Your statement about advisors is pretty cynical, or could be said of any profession by the same token I suppose, but good advisors help their clients be disciplined with saving money, be prudent with investing their money, be sane when it seems like the walls are caving in, an
Re: (Score:2)
A reference that points to multiple studies: https://www.wesmoss.com/news/a... [wesmoss.com]
Re: (Score:1)
The main goal of many stock analyst is to get clients to do some action that gives themselves a cut and/or makes them look good. Beating the market or other analysts is only a secondary goal. Sometimes the client is trying to hedge against certain risks, which may supersede concerns about the beating the market or an average analyst.
Do they beat index funds? (Score:2)
Stock market analysts and actively managed mutual funds on average do not beat index funds. There is no mention in the article about whether these robo-analysts beat index funds.
I suspect not
Re: (Score:2)
Re: (Score:3)
"In short, you pay a little in opportunity cost of better returns and gain some comfort in avoiding downside risk."
They are no good at that either. If you want to lower your risk, and probably returns, then add a short term bond index fund to your portfolio mix.
Re: (Score:2)
In reality, picking stocks by chance leads to better results than pretty much any of the fund managers. The only exceptions are people like Warren Buffet, who will move the entire market simply by sneezing into a Kleenex rather than a Puffs.
Re: (Score:2)
https://www.fool.com/investing... [fool.com]
Re: (Score:2)
I'm not so sure about that. As a counterpoint, American Funds has 17 of their 18 equity funds outperforming the index over their lifetime and they've been doing this since 1934. Sure you can pick periods of time where they lagged the indexes, but as a poster above mentioned, active funds tend to earn their stripes in volatile markets where there is actually an edge to picking high quality stocks and not owning "all of the junk" along with it. One of the reasons so many people lost money in the dot-com bust
Re: (Score:3)
It's all about risk management. Depending on your goals, there's a certain balance of investments you take - between bonds, money market, stock equity
Re: (Score:1)
Re: (Score:2)
There's a big difference between actively managed mutual funds / ETF's and hedge funds, so don't confuse the two. Hedge funds typically have a long list of requirements as far as the amount deposited, when you can get access to it, etc. and they usually employ pretty complicated options strategies or very concentrated holdings to invest based on a specific goal, whereas typical actively managed funds are looking to gain performance similar to a benchmark while achieving a targeted risk tolerance. E.g., they
Re: (Score:2)
This. I suspect that the success of the robot is not so much that it's good at picking stocks, but that humans on average are bad at it.
Re: (Score:1)
Re: (Score:2)
Actively managed funds as a whole cannot beat the market index. It's mathematically impossible. In fact, managed funds as a whole underperform the market net of fees.
No, it is not impossible, mathematically or otherwise. Some managed funds do beat their comparative market-index, maybe not consistently, but possibly well enough to make them worth owning over the long-term.
If it were possible to prove mathematically that index funds cannot be beaten, then all funds would be index funds.
What is a robot? (Score:2)
So I have this strange definition of a robot that includes moving parts and having an effect on the physical world. This seems more like a pattern matching application/AI application... not a robot.
Don't get me started on how we are applying the term AI these days... most of the use of AI today seems to be a pattern matching application. Let me show you a million pictures of a dog, then the computer can pick out dogs on further pictures... ask it to find a greyhound and it would be lost (unless you includ
Re: (Score:2)
We use the term 'robotics' at work to describe what was known as 'automation'. It gets more funding, and it actually improves a great deal of stuff, so win.
But I can see where the humans in stock trading will decry this as exceedingly dangerous. The 'analyitics', those who rely on data, will likewise see this as a threat, and discredit it. Paww...
Can't wait to see this bear fruit, and make the stock market the logical equivalent of a poker site, since those are pretty much just bots taking your money, or a
My robot definition (Score:2)
Two characteristics:
1. A machine that is designed to do more than one task
2. Includes sensors which are used to help it determine what to do next
I agree, it's a misnomer when used to describe programming.
Re: (Score:2)
I for one appreciate the use of the term "Robot".
Because it's not AI, a term that marketing assholes are making bank on.
It is also a pleasant departure from the low hanging fruit this website usually gets it's headlines from.
Re: (Score:2)
Would you say an ATM is a robot but online credit card processing is not?
Securities used to be pieces of paper that would be physically moved from one place to another when traded.
I am not really arguing that modern trading algorithms are robots, just that whether a task requires moving parts in the physical world is sometimes just a function of the state of technology for accomplishing that task.
Micro trades (Score:2)
When the stock market devolved into day-traders making fractions of a cent per share on trades, and locating high power computers near the market to make trades faster, you knew someone was going to have an AI driving to gain a few more milliseconds. The research department will be relegated to what it has been for the past 15 years. . . a front for insider trading.
Re: (Score:2)
a front for insider trading.
Pretty much the only investment advisors who ever beat darts thrown at the financial pages to pick stocks.
Mostly because they buy low and sell high (Score:2)
Over the decades, I've learned that most people attach emotion to their purchases, and find it hard to sell, and overagonize about purchases.
Everyone always tries to sell at the exact bottom and the exact peak, but you're better off buying on trade news (e.g. public in the industry, not publicized) and selling on PR campaign (e.g. publicized after people knew it). You sell into the top of the wave and buy in when you approach the bottom.
Also, a wise investor over the past decade would have invested in the h
Re: (Score:2)
Re: (Score:2)
RDS was one of my first investments, bought on bad news (bottom) but after board change, sold on good news.
But I wouldn't invest in any of them right now. "Proven" reserves are not what any of them say they are, and it's priced to presume nothing goes wrong ...
Say it isn't so! (Score:2)
Robots saw patterns in data better than humans did?! How can it be! ;-)
So do monkeys (Score:2)
Excellent (Score:2)
Now we can get rid of those overpaid charlatans who come begging to the U.S. taxpayer every time they screw up.
Enough with the socialism. Let them fail. Or in this case, let the robots take their place at a much lower cost.
Based on what track record (Score:1)
It should be considered that this statement applies to a time-frame of the last 10+ years. Until i see robo analysts being capable to navigate through multiple financial or economic crisis i wouldn't entrust my money to an algorithm parsing news and trends. The whole aspect of diversification, risk management, long investment strategy and business valuation is being left out it seems.
Re: (Score:2)
Thank you for adding this to the conversation. A lot of this passive vs. active debate tends to focus on the performance since the Great Recession and leaves out how decimated index funds were during the dot-com bubble when technology's sector weighting had become so large by capitalization that it took down the S&P 500 by 47% while some conservative actively managed funds were down only 7%. It's nice making money on the way up, but keeping more on the way down can be even more important. Remember that
The market has negative feedback for chartism. (Score:5, Interesting)
The financial markets are a feedback mechanism that penalizes any algorithmic system for trading, once it becomes widespread.
Buying drives prices up, selling drives them down. When a sufficiently large amount of the trading is done the same way, based on the same (or closely similar) decision rules driven by the same inputs, this effect more than counters any benefit from the rules being good in the absense of the feedback.
(This is also why low latency in high-speed trading is so important: First to trade gets just the benefit of the rule set, by getting in before the crowd betting the same way drives the market into a loss for themselves.)
Right now most of the trading is still done by humans, sometimes aided by computers running canned investment rulesets. So human psychology and common-wisdom expertise create some commonality in decisions and receive some mob-trading penalty.
If the "Robot Analysts" mad the same recommendations as the human analysts, and the trading done on their advice had the same latency, the results would be about the same and there would be no news. To the extent that they differ, the robots drive a smaller amount of the market and receive a smaller penalty.
Of course that's assuming the study even used actual trading. If it used "what would have happened if" analysis the robots get NO penalty.
So there's nothing surprising here. And there's nothing indicating, to me, that the robots are, in effect, anything more than yet another new "chartist rule" that only works, so far, because it's not yet being played with enough money for the market to break it.
Re: (Score:2)
Thriving in a market is done by executing better and staying one step ahead. Not by having an idea so brilliant nobody else can figure it out.
Guh (Score:1)
If you think professional flesh-and-bone analyst can out-trade the enslaved-AI in Goldman Sachs basement, you are gravely mistaken. All traders and investors are pure gamblers and the poor bastard in sachs basement is the one to out-trade them all. The game was rigged from the start - stonks always go up.
Would help if market was logical! (Score:2)
Re: (Score:2)
Because nobody wants to get rich slowly! (-- Warren Buffet)
Your broker isn't there to pick stocks (Score:2)
Well sure... (Score:2)
What did the research paper find? (Score:3)
The research paper [ssrn.com] found that compared to human advisors, robo-advisors (1) produce a more balanced distribution of buy, hold, and sell recommendations, (2) revise their reports more frequently and consider more data, (3) "exhibit weaker short-window return reactions, suggesting that investors do not trade on their signals", and (4) portfolios formed based on the buy recommendations of Robo-Analysts appear to outperform those of human analysts, suggesting that their buy calls are more profitable.
Of course, it's really the last point that is what is touted by this thread's summary: Robo-advisors' buy recommendations earn annualized returns of 6.4-6.9% compared to 1.2-1.7% for human advisors. Of course, 1.7% is horrible when the market indices perform better than that, indicating that it's not that the robo-advisors are so great but that the human advisors are so bad, i.e., investing in a no-load index fund would have far outperformed the human advisors.
This isn't my field, so I'm not certain how to assess the quality of the SSRN site hosting this research paper. However, the paper itself appears to not have undergone peer review. Had it been subjected to peer review, I imagine that comparisons to index fund performance would have been included in the study.
Re: (Score:1)
You don't need to outrun the bear, just your buddy (Score:2)