Internet Search Queries Predict Stock Market Volatility

Can Internet Search Queries Help to Predict Stock Market Volatility? by Thomas Dimpfl and Stephan Jank, link.

There is a very interesting research on using publicly available Google Trends data to predict realized stock market volatility. The researches use a number of standard models of realized volatility and add search query index results to come up with better forecasts. The most interesting part that I found was that the authors claim significant gain (over standard models) in prediction accuracy during high volatility period in the autumn of 2008. Unfortunately the analysis does not extend to the most recent period of high volatility.

Abstract:

This paper studies the dynamics of stock market volatility and retail investor attention measured by internet search queries. We find a strong co-movement of stock market indices' realized volatility and the search queries for their names. Furthermore, Granger causality is bi-directional: high searches follow high volatility, and high volatility follows high searches. Using the latter feedback effect to predict volatility we find that search queries contain additional information about market volatility. They help to improve volatility forecasts in-sample and out-of-sample as well as for different forecasting horizons. Search queries are particularly useful to predict volatility in high-volatility phases.

I have been working on a similar approach, also using Google Trends and Google Correlate to find drivers for volatility. While intuition was telling me that keywords like "bearish", "negative", "crash", "risk", "drawdown", or just "volatility" would have high predictive power - they did not, it was mostly benign searches like "stock futures" and various references to stock indexes that were most correlated to volatility. I guess people search more when market is crashing, not so much when there is a bull market. I am very interested in advancing research on this topic, and hope to share some of my own findings soon.

A similar topic is analysed in Web search queries can predict stock market volumes, link by a team of researches using proprietary company-specific search query dataset from Yahoo!. They observe one-way causality, that search queries help predict trading volume, but not the reverse.

Abstract:

We live in a computerized and networked society where many of our actions leave a digital trace and affect other people’s actions. This has lead to the emergence of a new data-driven research field: mathematical methods of computer science, statistical physics and sociometry provide insights on a wide range of disciplines ranging from social science to human mobility. A recent important discovery is that query volumes (i.e., the number of requests submitted by users to search engines on the www) can be used to track and, in some cases, to anticipate the dynamics of social phenomena. Successful exemples include unemployment levels, car and home sales, and epidemics spreading. Few recent works applied this approach to stock prices and market sentiment. However, it remains unclear if trends in financial markets can be anticipated by the collective wisdom of on-line users on the web. Here we show that trading volumes of stocks traded in NASDAQ-100 are correlated with the volumes of queries related to the same stocks. In particular, query volumes anticipate in many cases peaks of trading by one day or more. Our analysis is carried out on a unique dataset of queries, submitted to an important web search engine, which enable us to investigate also the user behavior. We show that the query volume dynamics emerges from the collective but seemingly uncoordinated activity of many users. These findings contribute to the debate on the identification of early warnings of financial systemic risk, based on the activity of users of the www.

Expiration Analysis and VIX Forecasts

In the month since last expiration volatility indexes moved in a broad range, but finishing relatively unchanged. VIX index trades as low 28.24 and as high as 45.45, while VSTOXX traded from 34.94 to 50.44. The forecasts that I made last expiration - for VIX to close lower and VSTOXX to finish higher - we half correct: VIX expired at 33.15 (vs 35.70 in futures) but VSTOXX closed on Wednesday at 39.89 (vs 40.70 in the futures). All the forecasts are stored separately on Forecasts Tracker page. I will be updating PL chart in the next day or two.

My forecast for November VIX to close at 36.10 vs 34.05 in the futures market (Thursday close)
My forecast for November VSTOXX to close at 42.76 vs 40.50 in the futures market (Thursday close)

Volatility Derivatives Around The World

The recent document about possible introduction of VXSLV Silver Volatility Futures and Options prompted me to revisit history of listed volatility products. At this time there are only three actively traded volatility products in the world: VIX in the US, VSTOXX in Europe, and RTSVX (usd-denominated!) in Russia.

VSTOXX futures volume has been increasing, however options volume is experiencing some growing pains. Eurex is planning for different ways to boost options volume.

RTSVX volatility futures were listed in the late May 2011, and had a strong start. Unfortunately it seems that trading activity has been falling since the latest economic crisis, probably because some shorts were blown out of their positions when the index rose from 26 to 71 in August (including a heart-attack inducing intraday high of 112 on 8/9, although that was probably just a bad print) No options are planned at this point.

There are few products that are not actively traded, and at this point are set for delisting:
OIV - Nymex Crude Oil volatility derivatives listed on CME. Both futures and options are listed, but there is absolutely no trading activity.
GIV - Comex Gold volatility derivatives, also on CME, also no trading at all.
GVZ - Gold volatility futures are options introduced in April of this year on CBOE, despite a promising début have not traded at all in the last few months.

There are also a few products in the works:
CME has been planning to list two CBOT-based indexes - SIV on soybeans and CIV on corn in the first quarter of 2011, but obviously these have been postponed because of the failure of their other volatility products.
Osaka Securities Exchange announced few weeks ago concrete plans to introduce volatility futures on VNKY Nikkei Volatility Index in the Q1 2012.
Australian Securities Exchange released product specifications and solicited comments from financial community regarding volatility futures and options on ASX 200 VIX Index. According to a private conversation with exchange representative they are looking for committed market makers in the product, in order to insure trading interest and liquidity.
India VIX - NSE started real-time dissemination of their volatility index in July 2010. According to some regulation there was year-long wait period before derivatives could be launched, and since NSE did not announce anything in July 2011 I assume there are no concrete plans for the product.
Hang-Seng Indexes Company announced some technical changes to calculation of VHSI - HSI Volatility Index. I am speculating that there is usually no reason to change index methodology unless the exchange is planning to do something, and it may indicate the first step toward Hong-Kong volatility derivatives.

For a complete list of volatility products and indexes visit Volatility Indexes and ETFs page.

VXSLV / CBOE Is Planning Silver Volatility Derivatives

According to this document filed with SEC CBOE is quietly planning to launch options (and I assume futures) on VXSLV - volatility index based on SLV options. There is no mention of VXSLV derivatives on the CBOE website, or anywhere else that I could find. Hope springs eternal, and CBOE is once again trying to repeat the success of VIX despite failures of other volatility products like RVX, VXN, VXD, GVZ (and OIV and GIV on CME) and others.

How To Manipulate VIX Settlement Price

VIX expiration day often coincides with particularly heavy trading activity in underlying SPX options. VIX settlement value, or VRO rarely matches either the Tuesday close or Wednesday open prices on the "cash" index, prompting pundits to blame VIX settlment for being manipulated. A popular theory is that VIX settlement value is being pushed up or down with huge SPX trades, referred to as "carpet-bombing". Some say that the manipulative trades are concentrated around high-vega strikes, others concerned specifically about puts. In this post I explain why large trades are not likely an explanation for VIX manipulation, and instead how VIX settlement value can be artificially increased for less than one hundred dollars, how VSTOXX futures and options are not subject to such manipulation, and propose a simple modification that makes VIX manipulation too expensive to be profitable.

VIX settlement value is determined by a Special Opening Quotation, based on the opening trades of SPX options instead of quotes. It is true that VIX settlement value can be made higher or lower by placing a big order that would result in a trade. The quantity necessary to move ATM SPX options is significant, and such trade would have to either be maintained and hedged until expiration, or exited immediately possibly with a large slippage. For this reason I don't think that it would make sense for a trader to attempt to manipulate VIX this way - it is very costly, and possibly very risky if the market moves against the trader, and I believe that heavy trading on the open of VIX expiration does not signify VIX manipulation, but rather legitimate SPX trading activity.

However a different form of VIX manipulation is possible. The VIX calculation formula is a weighted sum of option prices, with weight proportional to 1/K^2, where K is the strike. When K is getting smaller, 1/K^2 is getting bigger. While in theory such growth in a weighting term is mitigated with declining put prices (as K is getting smaller puts get cheaper) in practice it seems possible to "blow up" the VIX by placing orders - nickel bids that are most certainly would get executed - at very low strikes. It is a rather small investment - a few cheap options with limited risk since all options are bought - no short positions.

How this would work in practice:
1. On Tuesday before VIX expiration before the close, a trader purchases a significant amount of VIX calls, ATM or slightly OTM, that have the most potential gain from an unanticipated VIX increase.
2. On Wednesday before the open, a trader places 0.05 1-lot bids on low strikes SPX puts for the next month's expiration (the expiration that determines VIX settlement)

To illustrate the idea I downloaded SPX data from the September 2011 VIX expiration available from the CBOE website here. VIX settled at 33.72, with 550 being the lowest strike traded. As I mentioned above, by construction VIX is very sensitive to the low-strike puts, and if a 500 strike had traded at 0.05 VIX would have settled at about 33.73. If 400 and 500 strikes had traded at 0.05 VIX would have settled at about 33.86; adding a 300 strike trade would push VIX to 34.06; adding a 200 strike trade would push VIX to 34.50; adding a 100 strike trade would push VIX to 36.23. To summarize, for a total cost of 5*$5 = $25 a trader can artificially inflate VIX value by 2.51 points.

500	+0.01
400	+0.14
300	+0.34
200	+0.78
100	+2.51

This methodology applies to any VIX expiration. Using settlement data from August 2011 expiration I estimate potential effects of price manipulations as above.

500	+0.15
400	+0.26
300	+0.47
200	+0.92
100	+2.69

Since the CBOE provides what they call "likely VIX series" we can know which SPX strikes were available at the time of expiration, and calculate exactly the effect on VIX index for every expiration for which data is available. The greatest effect on VIX comes from the lowest strike, so for practical implementation the strategy would depend on which SPX strikes are listed.

Economic significance of such manipulation depends on VIX options on the last trading day, and opening price of SPX. Historically the overnight VIX move from Tuesday close to Wednesday morning settlement has been rather volatile (Russell Rhoads did a study of this in his book), but I believe in most cases the trade would have a very large upside with limited downside.

VSTOXX - a pan-European volatility index based on EURO STOXX 50 index is not subject to such manipulation. The contract is settled into an average of index values during a half hour period on the last trading day (here)

The Final Settlement Price is established by Eurex on the Final Settlement Day, based on the average of the index values of the underlying on the Last Trading Day between 11:30 and 12:00 CET.

This certainly makes sense given that VSTOXX expires in the PM. However now that SPXPM options are picking up some volume on C2 I think it would make perfect sense to calculate VIX values based on a similar averaging procedure as VSTOXX that makes manipulation very expensive and practically impossible, or the CBOE to offer a different settlement procedure that is more resistant to manipulation.