Commentary: Bringing a human touch to investment data analytics

As someone who grew up a blue-collar kid in New Jersey, I have very special memories of Bruce Springsteen. But most people don't remember that his first two albums were considered commercial failures. Many critics used that sales data to predict he wouldn't succeed. But in 1974, while The Boss was struggling, music critic Jon Landau wrote something very prescient after watching him perform: "I saw rock and roll future and its name is Bruce Springsteen. And on a night when I needed to feel young, he made me feel like I was hearing music for the very first time." What did Mr. Landau see that others had missed?

Predictions are important in the hedge fund business, too, and we often come up with them by accumulating a wealth of data. In the early 1980s, investors who found data others did not were the winners. However, in the past 10 years we've seen such an explosion of data that the edge for humans finding this hidden data is no longer enough. Computers are probabilistically forecasting what the data will be before it is announced.

The active management industry is filled with media headlines on the arms race of data and artificial intelligence. Hundreds of millions of dollars are pouring into the search for the holy grail of predictions through data. If we're not careful, the end of this race will look a lot like the gold rush, with a few winners and many more losers. Indeed, many large, successful quantitative funds have been doing this for decades and already have scraped much of the available alpha chicken off the bone. Meanwhile, we're seeing plenty of competition outside the hedge fund industry coming from algorithm builders who are providing data. This increased competition to immediately connect asset prices to real-time data will continue to lead to shorter and sharper investment cycles, as evidenced in the chart below.

Thanks to faster computing speeds and a low entry barrier, individual traders with home computers can compete with professionals, further eliminating the ability of short-term active managers to use data for arbitrage opportunities. Simply put, computers discount new data points faster than humans. In the future, there will be far fewer market participants than there are today; the survivors will have adapted to this efficient world.

If data alone no longer gives us an edge in investing, what "knowledge" can help us to predict market outcomes? Despite increased competition from computers, the human brain remains the answer. It still has the advantage when it comes to pattern recognition, intuition and connecting the dots. Computers are good at telling us what is happening today, but not at what people will do in the future. The power is shifting from accessing data to filtering and interpreting data to make a decision. This is what Jon Landau was able to do. The data said Bruce Springsteen would not be successful, but Mr. Landau believed he was witnessing something great and used his experience as a music critic — and from attending Bruce's concert — to make a prediction. This led him to forecast a regime shift. Like Steve Jobs with the iPod, Mr. Landau's intuition was that people were eventually going to see what he saw.

Making the transition away from math and into the art of converting "knowledge" to data has become more important. For years, many have tried to incorporate the art of positioning and sentiment analysis into decision-making. Now, those who will win against computers are those who use their brains most effectively to turn knowledge into a model that constantly reassesses the probabilities and makes predictions. It's a move away from short-term investing alpha to trading and allocation alpha, where patience and folding are more important. This means we will periodically have the sharper short-term cycles we mentioned earlier, in the form of increased volatility.

Computers, quantitative analysis and artificial intelligence will continue to force out those incapable of adapting. The survivors will be the ones that 1), do not race against the machines and are able to incorporate technology into their processes and businesses; and 2), are able to incorporate knowledge into their models of decision-making while AI tries to catch up to the power of the brain.

The world appears poised to remain in a lower-for-longer environment with low growth and inflation. We're also likely to see more violent market pendulum swings between reflation and deflation, causing shorter and sharper investment cycles that will feel more like emerging market swings.

One unexpected pendulum shift began in February 2016 when the most recent lows were set in risk assets. Commodity deflation and central banks without ammunition were the fears then. But today, it's hard to find fears. Stock market prices around the world are significantly higher with the MSCI World index up 35% from February 2016 lows. And there is still hope for growth built around expectations of U.S. tax cuts, increased infrastructure spending and deregulation.

Despite these positive trends, we're also seeing the central banks being less accommodative with the Federal Reserve and People's Bank of China tightening. Commodities are weakening, with some indexes near or at 52-week lows as of May 31. Momentum has tech on one side, and autos, retailers and energy on the other. Meanwhile, value has been down five months in a row — also as of May 31 — for the first time since June 2008, and utilities have outperformed financials by 9% so far this year within the Standard & Poor's 500 index.

However, the behavioral side is most troublesome. The best measure to catch these regime shifts is using the Investor's Intelligence sentiment index, which recently hit its lowest level since 2015. The lack of position cleansing we have noticed since the U.S. presidential election in 2016 is particularly concerning. From August 2015 to November 2016, there were at least 10 corrections of at least 5%. Without cleansings, volatility measures have collapsed.

I traded markets in the early 1990s and learned not to expect higher volatility just because it is low. But the longer we go without a rise, the more violent it will be when we finally do see one.

The chart below measures sentiment, showing the factor price to book or value. Since the low in 2009, when lower for longer became the norm, whenever value has slipped it has paid to be patient. Because of low rates, the price-earnings ratio of the market has proven to be a bad gauge of whether returns are about to suffer. However, when viewed from a market neutral-perspective, when the expensive stocks are outperforming the cheap stocks, it has at times recently been a dot preceding a cleansing.

The psychology of "value no longer matters" fits in with the chart below, which shows the relative strength of the Nasdaq 100 vs. the S&P 500. This suggests the risk reward from these levels argues for patience and sitting out a few hands, waiting for a better entry.

Given the returns year-to-date and current sentiment levels, I think it's prudent to be careful and let the enthusiasm from the good results early in the year burn off. Regime changes should not be thought of as directional calls. It seems clear that what the markets are offering an active manager is very different from last February, and that means we should be prepared for possible turbulence ahead.

Jon Landau ignored the sales data of Bruce Springsteen's first two albums, and instead made his prediction based on his experience. Bruce's next album was the iconic, "Born to Run," which propelled him to international stardom. The rest, as they say, is history. Similarly, in an investing world filled with too much data, predicting success is about filtering, behavioral inputs, connecting the dots and constantly reassessing the odds.

Jordi Visser is president and chief investment officer of Weiss Multi-Strategy Advisers LLC, New York. This article represents the views of the author. It was submitted and edited under Pensions & Investments guidelines, but is not a product of P&I's editorial team.