Computers vs. the News: What’s Behind the Stock Market Chop?


Today marked another triple-digit move for the Dow Jones Industrial Average, which closed up 272 points. Of the 45 trading days over the last two months, 28 of them (including today) have seen triple-digit moves, meaning the Dow has gone up or down by 100 points (or more) 62% of the time since July 25. The average daily move for the Dow during that time has been 188 points, or 1.6%. Here’s a snapshot showing the performance of the Dow over the last two months:

Pretty choppy, right? I’m no stock market historian, but I’d imagine that you’d be pretty hard-pressed to find such a sustained period of volatility. Which brings up the question: what’s causing this? Obviously, there is a lot of uncertainty (and fear) in the market right now. From Europe’s sovereign debt problems, to America’s toxic political climate, to the sputtering global economy, there is a lot to be anxious about. Anxiety breeds indecision, which characterizes the bumpy market pretty well.

There’s also been a lot of speculation about how much high frequency traders are driving the choppiness. These guys are pretty much the only ones who have been consistently making money the last several weeks. High-frequency traders feed off of volatility. No matter which way the market moves, they can make money by taking advantage of tiny price differentials, trading in and out of stocks in fractions of a second. But feeding off of volatility is very different from causing it.

In a recent op-ed, Manoj Narang, founder of the HFT shop Tradeworx, insists that high-frequency trading is not driving volatility. Rather, it’s media exposure and access to markets that is really driving it. Narang writes:

From 2000-2006, the S&P 500 moved an average of 0.37% per day when the market was closed (that is, between the close of one day and the open of the next), whereas from 2007 onward, it moved 0.61% per day during this period, a 65% increase. Logically, it is impossible to blame high-frequency traders for this 65% rise in close-to-open volatility, because there is no trading when the market is closed.  This volatility reflects one thing and one thing only — markets react to news, and since 2007, there has been an abundance of news which has caused investors to panic.

This 65% increase in volatility is particularly revealing when it is juxtaposed against the comparable difference when the markets are open. From 2000-2006, the S&P 500 moved an average of 0.76% between the open and close of the same day,  compared with 0.85% since 2007. In other words, volatility during trading hours has increased only 12% during the exact same period when volatility during non-market hours has increased 65%, less than a fifth of what would be expected.

Still, regulators remain increasingly skeptical about the world of computer-driven trading. According to a Reuters report earlier this month, the SEC and FINRA have asked high-frequency firms to hand over details about their proprietary trading codes. It remains to be seen whether they hold any secrets to what’s behind all the market volatility.

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. rationalrevolution says:

    The issue isn’t high frequency trading, the issue is predictive modeling.

    Thumb up 0 Thumb down 0
  2. RGJ says:

    Not to be sarcastic, but isn’t this how a vibrant, competitive, well-informed free market capitalist system trades?

    A little jittery, but the world is a little jittery.

    Thumb up 2 Thumb down 0
  3. Statspotting says:

    Here is a stunning stat: did you know that ONLY TWO FIRMS contribute to ONE-FIFTH of ALL U.S. EQUITIES TRADING?

    Thumb up 2 Thumb down 0
    • JFW says:

      Wouldn’t it be less exacerbating to say that one firm contributes 10% of all equity trading, or however the number breaks down, in the US?

      Constantly talking in terms designed to stand out from everyone else is how the news gets so crazy to begin with.

      Now, if one firm is responsible for 10% of equity trading, I do not find that shocking at all. How about VISA which is responsible for greater than 35% of all credit clearing transactions? Capitalism breeds large companies.

      Thumb up 2 Thumb down 1
      • Statspotting says:

        you are right in so many ways. Remember what Steve Jobs said: “we are selling an iPad every 3 seconds” suddenly sounds great.

        but the idea behind this one third, one fifth etc is for the magnitude to really sink in. I think.

        Thumb up 1 Thumb down 0
      • steve says:

        Capitalsim and heavy regulations breed large companies.

        As widely recognized, large capital requirements breed large companies. When a new semiconductor factory costs a billion dollars small semiconductor manufacturers don’t exist.

        What is less well recognized is that the same effect (economies of scale) occurs in heavily regulated industries that don’t otherwise require large amounts of capital. For example, small food processing companies don’t exist in the US not because the factories are expensive. Butchering cows isn’t very capital intensive. Rather it is due to regulation since you need both a compliance department to keep up with the regulations and the ability to periodically absorb multi-million dollar fines when you inevitably screw up. Now I don’t really mind regulation of food processing (who wants unsafe food), but it should be recognized that this definitely encourages companies to grow larger.

        Thumb up 0 Thumb down 0
  4. FreakonomicsFan says:

    Predictive modeling could mean “painting the tape” to regulators. The % of daily trading volume handled by HF systems is staggering.

    Thumb up 1 Thumb down 0
  5. Adam says:

    This issue can’t be explained by HFT or news. It’s never that easy. My view is we are seeing the culmination of various factors converging together and it has directly resulted in increased, large swings in volatility. The speed at which news is available has increased close to the rate that HFT speed has increased. In 1999, HFT took a few seconds to execute, whereas today, we are in microsecond trade rate execution (electronic trading actually goes back to the 1987 crash…). News travels fast because of increased media outlets and social networks. Leveraged ETF’s don’t get the attention they deserve. Double and Tripled levered ETF’s short underlying securities and are advisable only for day trading because they reset every night. These instruments were a direct product of the 2007-2009 downturn where the average investor didn’t have many shorting opportunities outside of Options and many don’t understand options, so creating these ETF’s was a viable alternative, and it surely made some people a load of money. The problem with these is as they have become adopted, the underlying shorts help exacerbate down moves as news breaks on “The Euro Zone crisis looms…” or “Looks as though Greek default more likely”. Eliminating the up tick rule makes shorts drive down prices violently as well (removed in 2007). HFT algorithms have the technical indicators set, so as news breaks, positive or negative, prices move toward those set levels and once hit, HFT takes over on the downside or upside. Hedge Fund selling and buying contributes to this affect as well along with other sources.

    It’s just a different game and for those who decide to continue playing, it’s going to stay volatile for a while…

    Thumb up 2 Thumb down 0
  6. Joshua Northey says:

    Hey if you can get more churn and more total transactions the people running Wall Street from the inside can skim more money out of the economy for themselves.

    Thumb up 0 Thumb down 0
  7. mfw13 says:

    Look at the volume numbers.

    These days, volume is about 5 billion shares a day, give or take.

    Ten years ago, volume was about 1 billion shares a day, give or take.

    Fifteen years ago, volume was about 500 million shares a day, give or take.

    What’s changed? Technology. Fifteen years ago, short-duration trading for the most part didn’t exist, since you couldn’t buy or sell stocks online. Nowadays, probably 70-80% of all trades are done by computer programs and algorithms, with nary a human in sight. Most of these trades last only a few seconds or minutes. And because the length of these trades is so short, actual real-world news doesn’t have any impact on them, since they are being made by entities trying to profit from microscopic short-term moves in the market.

    Therefore you have a market which appears to move with no relationship to real world events, since the number of people making trades based on real world events (maybe 20-30% of volume) are being drowned out by computers making lightning quick short-term trades that are completely algorithmic in origin (maybe 70-80% of volume).

    Well-loved. Like or Dislike: Thumb up 10 Thumb down 0
  8. arno says:

    “there is no trading when the market is closed”

    S&P Future are trading, even when the index is closed.

    It may be that Manoj Narang doesn’t know that, but I would guess he is a clever guy.

    Maybe it’s just that as a founder of a HFT shop, he is just pushing his own agenda, by putting out shamelessly misleading Op-ed.

    Who knows ?

    Thumb up 2 Thumb down 0
    • Forstingale says:

      I guess HFT models just don’t work that well during the after/pre-market hours, when trading volume is sooo small.

      Thumb up 0 Thumb down 0
      • arno says:

        Saying trading volume are small is different from saying there is no trading occuring.

        I just had a quick look at bloomberg data, the E-mini S&P 500 futures volume after/pre-market hour on the 27th of september was 540k lots (around 31mrds USD) around 20% of the volume during mkt hours (2706k lots )

        That’s only a day , so may not be representative, but there is trading even when the US market is closed.

        And I didn’t even had a look at european/asian based etf trackers, etc…

        Thumb up 0 Thumb down 0