The New Challenges of Price Discovery by Frank Holmes
September 5, 2014
Investing in the Age of High-Frequency Trading, Falling Volumes and Widening Bid-Ask Spreads
By Frank Holmes
CEO and Chief Investment Officer
U.S. Global Investors
As investment managers, one of our most important fiduciary responsibilities is buying and selling stocks for the best possible price and execution. We do this by using the statistical strategies I’ve previously covered, from monitoring short- and long-term cycles; implementing probability models such as standard deviation, mean reversion and oscillators; and identifying the relative valuation of stock with the portfolio manager’s cube.
If only it were that simple.
In the past few years, price discovery—or the act of finding the “right” price for a security—has become much more challenging because of falling stock volume and widening bid-ask spreads. These challenges are directly attributable to the infiltration of high-frequency traders into the market, not to mention the expansion of dark pools and non-exchange trading.
Simply put, when stock volume is high and transactions increase, the bid-ask spread narrows. Brokers and dealers accordingly price shares to move, and investors have a pretty good estimation of what they’re going to spend on a security.
But when there are fewer transactions and volume is down, the bid-ask spread widens. Price discovery, then, becomes difficult because stock valuation has a broader range in which to move. I discussed this last week using the intraday performances of the TSX Venture Exchange and Market Vectors Junior Gold Miners ETF (GDXJ) as examples: in the afternoon, after volume and activity tend to decrease, spreads widen.
Think of this in terms of real estate. If volume is up and homes are selling rapidly in Neighborhood A, both buyers and sellers have a good idea of what a fair price is, based on the dollar amount of square footage of nearby homes sold within a certain timeframe. Price discovery, therefore, is reasonable.
But if homes in Neighborhood B languish on the market for lengthy periods of time, relative price comparisons begins to dissolve. Who knows what the homes should go for? Closing deals becomes tough because, in such a scenario, a buyer’s bid might come in way under what the seller is willing to accept. As a result, the price of homes, even those in adjacent lots, can fluctuate wildly.
Volume Drops, Volatility Rises—But Opportunity Remains
To see these concepts in action, look at the chart below. The S&P/TSX Venture Composite Index, which lists about 500 Canadian micro-cap venture companies, has seen a drop in volume of more than 60 percent since mid-2011. This has widened the bid-ask spreads of individual equities in the index—not the index itself—complicating price discovery.
Despite the challenge, we try to take advantage of the volatility that other investors might flee from. Decisions to buy or sell a company are first fundamentally driven, and then trading is based on statistical analysis of fund flows, volatility over different time periods and relative performance to the gold indices we strive to beat. For the Gold and Precious Metals Fund (USERX), it’s the FTSE Gold Mines Index; for the World Precious Minerals Fund (UNWPX), the NYSE Arca Gold Miners Index.
Our style resembles that of the Navy SEALS, in that we prefer to be nimble, surgical and tactical. During the bear market that ran from mid-2011 until February 2014, we nibbled rather than munched on inexpensive companies that were lagging in relative performance over one day, one month and one quarter. And when these companies showed a surge in price and volume, we often trimmed our holdings rather than sold outright. This incremental “nibbling” strategy is a little like investment reconnaissance, enabling us to test our conviction in a company before taking a weightier position.
Another disruptive factor to price discovery has been the proliferation of exchange-traded funds (ETFs). Accounting for more than 30 percent of trading volume in the markets, some ETFs are influencing the markets they track and impacting their underlying holdings. A study by Goldman Sachs confirmed that ETF trades influence stock prices. The study looked at which individual stocks move more with the dynamics of the ETF than on their own fundamentals and found that those stocks most affected by ETF activity are in the Russell 2000, probably because of their lower levels of liquidity, lower volume and cheap prices.
We’ve witnessed this same phenomenon with some junior gold stocks in the GDXJ. A gold stock can have a significant price move based not on changes to its fundamentals or a corporate event but rather shifts in sentiment toward gold that is compounded by fund flows. The inclusion or exclusion of a stock in the underlying index can result in a flurry of disruptive trading unrelated to changes in the company’s fundamentals.
Just as one man’s trash is another man’s treasure, one man’s fear of volatility is another man’s opportunity. Part of successful active management is not getting discouraged, learning to adapt to a changing climate and coming to terms with the market’s often erratic behavior.
But the erratic behavior has only ramped up in recent years.
HFT: Trading at the Speed of Greed
As I said earlier, price discovery has become much more difficult in recent years because of growing high-frequency trading (HFT), dark pools and non-exchange trading—all of which have changed, perhaps irreversibly so, the formation of capital in the investment industry.
HFT is a controversial practice whereby automated computers using sophisticated algorithms transact orders at lightning-fast speeds. In a process called latency arbitrage, high-speed traders are able to gain access to crucial order information and other market data milliseconds before “normal” or “slow” traders. They manage to do this through a number of means, including placing their computers as close as possible to stock exchanges and using best-of-the-best fiber optic cables.
After acquiring the information, such traders can get in front of other buyers’ purchases and, almost instantaneously, turn around and scalp the shares within less than a blink of an eye. Often gains are less than a penny per share, but because they trade so frequently and rapidly, it’s easy to make fast money.
This new form of legalized front-running became the talk of Wall Street after the March 2014 publication of financial writer Michael Lewis’s critical book on the matter, Flash Boys: A Wall Street Revolt. In one passage, Lewis deftly recounts the infamous Flash Crash that occurred at 2:34 on May 6, 2010:
For no obvious reason, the market fell six hundred points in a few minutes. A few minutes later, like a drunk trying to pretend he hadn’t just knocked over the fishbowl and killed the pet goldfish, it bounced right back up to where it was before. If you weren’t watching closely you could have missed the entire event… Shares of Procter & Gamble, for instance, traded as low as a penny and as high as $100,000. Twenty thousand different trades happened at stock prices more than 60 percent removed from the prices of those stocks just moments before.
A spread of $99,999.99. If that doesn’t give a trader pause, I’m not sure what will.
The chart below shows just how dramatically HFT has heightened intraday volatility in the SPDR S&P 500 ETF, arguably the most popular of its kind in the U.S. Up until 2007, daily price changes had a relatively steady heartbeat. But in 2007, when HFT as we know it today emerged, the average intraday volatility more than doubled. In August 2011, the peak volatility climbed to one that was 10 times higher than in 2006.
Lewis’s book has created a much-needed awareness of what HFT has brought to the market: disruption, unsettlement and a loss of trust and transparency. Like thieves in the night, high-speed traders can swoop in to a market that you created and take advantage of it.
Michael Matousek, head trader at U.S. Global Investors, has experienced this unpredictability firsthand. On numerous occasions he has put in a buy order, based on up-to-the-second liquidity information, but received only a fractional amount.
As for liquidity, Michael says that HFT might increase it, “but when an order is ‘sniffed,’ [high-frequency traders] cancel. So the perceived liquidity is gone within fractions of a second.”
Fellow U.S. Global trader Mike Ellingsen notes how HFT has also affected depth of market, or the measure of the liquidity of open and, I should add, transparent buy and sell orders.
True depth in the equities market has become hard to gauge,” he says. “Trust is key in this entire conversation.”
This trust, however, has been tarnished in a system dominated by HFT, which currently accounts for more than 70 percent of all market activity in the U.S., according to research conducted by European Central Bank economist Peter Hoffmann.
Trading in the Dark
So what do you do if you’re a large institutional trader who has a million shares to move but doesn’t want to be preyed upon by high-speed front-runners? The solution for many is to use not a conventional exchange, where market information is publically shared, but a private exchange. In such exchanges, known as “dark pools,” transactions are conducted secretly and anonymously. There is no trading floor, no orders visible to the public and no transparency.
Dark pools aren’t anything new; they’ve been around since at least the 1980s, mostly to reduce market impact and lower transaction costs. But an increasing number of large investors are using these exclusive pipelines to (allegedly) hide and protect their transactions from high-speed traders. In recent years, non-exchange trading has surged, accounting for close to half of all stock trades today.
The problem, as you might guess, is that stock volume in the U.S. is being usurped from the trading floors and drying up faster than the Aral Sea. Again, when volume drops, bid-ask spreads widen, which complicates price discovery.
According to TabbFORUM, whose Equities LiquidityMatrix™ consolidates monthly exchange data, industry volume dropped 1 percent