Market Volatility - Part Four

A view from the Member's Gallery inside the NYSE

Image via Wikipedia

In part 1, part 2 and part 3 of our volatility series, we showed that depending on how you measure volatility, it is not really clear that the markets are truly more volatile now than they have been in the past. As Joe Mecane, CAO for U.S. cash markets for NYSE Euronext mentioned at SIFMA 2011, “"the perception of volatility -- and perception is what matters -- is high.  But it's not necessarily unprecedented."

We could not agree more (at least partially because he was referring to analysis presented in these pages by yours truly, partially because your humble blogger reports into his organization, but more importantly, because he is correct). Our data show that intraday trading ranges are up since 2007, but much as we would like to be able to attribute this to one or two factors, we cannot.

But, What Causes Volatility?

We will cover our conclusions below, but there is little doubt that a lack of liquidity is a proximate cause of volatility. Those that love HFT, for example, point out that many HFT’s post liquidity and that should help dampen liquidity. Others correctly point out that all HFT is not liquidity provision and that some HFT may exacerbate imbalances. Others would like to blame various parts of Reg NMS, and yet others believe that ETFs are a factor in market volatility.

There have been some recent academic attempts to try and find measures that will forecast when there is risk of an event such as a flash crash. Some academic papers have focused on fragmentation, others on a measure known as VPIN (Volume Probability of Informed Trading). We simply note that when there are less posted shares available for the taking, while at the same time, liquidity demanding orders rise, the risk of large price moves increases.

By way of example, we found that shares posted on the NYSE on the most volatile days have tended to show measurably less depth than when the market is not quite as “animated”. Interestingly, we found that, especially near the inside, liquidity is less impacted, and that the more liquid the stock is, the less it is impacted as well.

Conclusions

Our review of volatility over the years highlights how much the markets have changed. Significant modifications to market structure, such as decimalization and Reg NMS have impacted how markets trade. Our basic findings show:

  • Day-to-day volatility is arguably similar now when compared to years in the past. The 1930s were easily more volatile than the post-2000, at least when measured by likelihood of a large price change.
  • Historical volatility has risen somewhat since the 2000s began.
  • Intraday volatility has increased since the start of Reg NMS, but end of day trading is probably the least impacted.
  • Without further study it is difficult to attribute any single factor to volatility changes. The most obvious cause of higher volatility is very simply higher risk to the financial system. Coincident to Reg NMS, we have seen the housing bubble burst, the collapse of FNMA, Freddie MAC, AIG, Lehman Brothers and Bear Stearns, the risk of a U.S. default, U.S. debt downgraded, and the continuing European debt crisis. These have all led to real risk of complete financial meltdown and seizing of the global economy. This has nothing to do with equity market structure, ETFs or High Frequency Trading.
  • We are not saying that these items are not factors in volatility, but the macroeconomic picture gives 2007-2011 a similar look to 1929 through the early 1930s, when volatility was last this high, and when orders were filled pretty much by carrier pigeon. (Anybody who wishes to try and determine if the more efficient equity market of the past few years has helped lead to added risk-taking elsewhere, is welcome to do the analysis.)
  • During quieter periods, volatility remains stubbornly high on an intraday basis. This may be due to fragmentation and off exchange trading pulling liquidity from the lit exchanges. High volatility is caused by a lack of liquidity and as more and more trading goes dark, risk for out sized moves increases (as shown by Prof. Daniel Weaver’s paper which you download here).

There is definitely more work to be done. For example, a study of market depth during different volatility regimes may help us to understand better why volatility changes and if HFT is stepping away when they are needed most.

We look forward to comments and suggestions from our readers.

 

Enhanced by Zemanta