SSRN Id288123
SSRN Id288123
SSRN Id288123
Abstract
In this paper, we use daily closing data on CBOE options of 30 stocks during February
through July of 1999 to investigate whether options open interest contains information that
can be used for trading purposes. Individual stock price at option maturity is first predicted
based on the distribution of options open interest. Several stock only and stock plus options
directional trading strategies are then considered after comparing the predicted stock price at
maturity and the actual stock price at the trade initiation date. In our sample, these trading
strategies generate better returns compared to the S&P 500, the buy and hold strategy
involving the sample stocks and the Merton et al (1978) style covered call strategy. Our
empirical evidence thus indicates that non-price measures of activity in the derivatives
market such as the open interest contain information about the future level of the underlying
asset. This lends support to prior works (such as Copeland and Galai (1983), and Easley et al
(1998)) that suggest that the derivatives cannot be considered redundant in a market with
information-related frictions. One implication is that the distribution of non-price
derivatives market activity may be helpful for other purposes where the physical instead of
the risk-neutral distribution of the underlying asset is needed. These include beta estimation,
volatility forecasting and volatility trading.
Keywords: information content, options, open interest, trading strategy.
JEL Classification: G13, G14.
Corresponding Address: Mo Chaudhury, Faculty of Management, McGill University, 1001 Sherbrooke Street
West, Montreal, Quebec, Canada H3A 1G5. Tel: (514) 398-5927, Fax: (514) 398-3876, Email:
[email protected].
* Rafiqul Bhuyan is from the School of Business, The University College of the Cariboo, 900 McGill Road,
Box 3010, Kamloops, BC, V2C 5N3, Canada. Tel: (250) 828-5055, Fax: (250) 371-5675, Email:
[email protected]. Mo Chaudhury is with the Faculty of Management, McGill University, 1001, Sherbrooke
Street West, Montreal, Canada H3A 1G5, Tel: (514) 398-5927, Fax: (514) 398-3876, Email:
[email protected]. This paper is related to Bhuyans doctoral dissertation work in progress at the
Department of Economics, Concordia University, Montreal. However, the research and the opinions
expressed in this paper are the sole responsibility of the authors.
DO PRICES AND TRADING ACTIVITY in security markets provide information about future
price movements? If so, can this information be used to generate trading gains? While these
questions have long generated significant interest among researchers and practitioners alike, the
existence and growth of derivatives such as options have added new and interesting issues to this
arena. The derivatives constitute an additional means for the informed traders to trade on their
information and others to discover that information. Not only the derivatives may lead the
underlying assets in impounding information, they may in fact provide information that simply
cannot be inferred from the markets in underlying assets. Further, there are measures of trading
activity, e.g., open interest, that are unique to the derivatives markets and as such provide a novel
way to examine the informational role of financial markets.
In this paper, we examine the role of options market open interest in conveying information
about the future movement of the underlying asset. More specifically, we use the open interest of
short term equity options to predict the stock price at maturity and show empirically that trading
strategies based on this predictor yield better returns than the buy-and-hold and passive covered call
strategies. Thus, according to our study, (unique measure of) activity in the equity options market
seems to contain information about future stock price that can be exploited for trading purposes. In
principle, this information can also be helpful in other applications where the physical or true
distribution of the underlying asset is needed.
Financial economists have long been interested in the process of price formation when
informed traders, uninformed liquidity (or noise) traders and market makers interact in the asset
market.1 With the introduction of options market, informed traders as well as liquidity traders have
an additional means to meet their trading needs. In fact, informed traders may find the options
market more lucrative than the stock market due to lower transaction costs, less capital outlays,
higher leverage, limited loss potential and lesser trading restrictions (e.g., no up tick rule for
shorting).2 If informed traders do choose to trade in the options market, not only the option prices
1
In an asymmetric information environment, informed traders may profit at the cost of noise or liquidity traders
loss (Copeland and Galai (1983)). Continued trading of the informed investors can, however, serve as signals to the
other (uninformed) market participants who can learn the underlying information in a Bayesian fashion and trade
accordingly (Glosten and Milgrom (1985), Easley and OHara (1987), Kyle (1985)). This possibility of multiple
rounds of trade arises if the first trade of the informed traders does not instantly reveal the new information.
Although the implications for the price paths, volume changes and trading strategies are different, information-
motivated trading may also be driven by differential information (He and Wang (1995)) or differential interpretation
of the same information (Copeland(1976)) by informed traders.
2
Black (1975) first hinted to the attractiveness of the options market to informed traders.
3
Under asymmetric information, information may flow from the option market to the underlying asset market. This
has important implications for the underlying asset and its options. Grossman (1988) suggests that under asymmetric
information traded derivatives are not the same as their synthetic counterparts due to their differential information
content. While Detemple and Selden (1991) argue that information asymmetry may alter the hedging opportunities
and as such impact the underlying asset price, Back (1993) examines the impact on option prices. Biais and Hillion
(1994) show that the price volatility of the underlying asset may be affected by information asymmetry. Brennan
and Cao (1996), on the other hand, show that in a noisy rational expectations equilibrium, option trades may not be
information based.
4
Using the Black-Scholes model, Manaster and Rendleman (1982) jointly imply the stock price and volatility from
the observed daily closing option prices and find support for the incremental information content of options based on
returns from ex post and ex ante trading strategies. Using the Berkley transaction data on options, Bhattacharya
(1987) also finds incremental information content of options although the trading benefits seem marginal. Vijh
(1988) questions the ex post results based on daily closing prices due to the bias arising from the nonsynchroneity of
closing stock and option trades and the bid-ask bounce. Anthony (1988) uses volumes of options bracketing the
daily closing price and find that for 64% of the sample stocks the options volume led the stock volume in a Granger
Causality sense. However, about 48% cases are statistically significant in both univariate and multivariate causality
tests. Stephan and Whaley (1990), on the other hand, find stocks to lead their options both in terms of intra-day price
change and trading activity. Chan, Chung, and Johnson (1993) argue that Stephan and Whaleys results are due to
different price discreteness rules in the stock and option markets. Examining how option prices move with option
trades, Vijh (1990) concludes that option trades are not information-driven while Srinivas (1993) attributes Vijhs
results to a sample selection bias. Sheikh and Ronn (1994) attributes unique patterns of returns in the options market
to information-based trading there. John, Koticha and Subrahmanyam (1993) and John, Koticha, Narayanan and
Subrahmanyam (2000) show that the impact of options trading depends on the margin and liquidity constraints faced
by the informed and uninformed or liquidity traders. Mayhew, Sarin and Shastri (1995) find that a reduced equity
options writing margin increases the bid-ask spread of the optioned stocks and that the uninformed traders are more
liquidity constrained than the informed traders. This evidence goes against the prevalence of information traders in
the options market.
5
The general equilibrium analysis of Leisen and Judd (2001) provides insights as to how open interests are
determined along with the option prices for various strikes in an incomplete markets setting. In their paper, agents
have heterogeneous risk preferences but homogeneous probability beliefs about the underlying assets. As such,
derivatives market activity (the distribution of open interests) is not informative about the future of the risky asset.
6
Converting the implied risk-neutral distribution into a physical distribution generally requires preference
specification.
Consider an equity instrument, or stock, for which there is a set of call and put options
maturing at T, the current time being T0. Let the price of the stock at time t be St. Let {Xi, i
=1,2, .,K} be the set of strike prices for call options and {Xl, l=1,2, ,L} be the set of strike prices
for put options. The payoff at maturity to the buyer of a call option with a strike price Xi is ic =
Max[0, ST - Xi]. For a strike price Xl, the put option buyers payoff at maturity is lp = Max[0, Xl-
7
We thus visualize the financial markets environment to be characterized by possible combinations of the
following: asymmetric information, differential private information or differential interpretation of same public
information, learning by the uninformed traders, hedging or insurance needs, and learning by the uninformed
traders.
8
Detemple and Murthy (1994) shows that in an economy with heterogeneous but rationally updated beliefs, while
the structure of intertemporal general equilibrium asset prices remain the same as under homogeneous beliefs, the
price levels and their dynamics are indeed affected by heterogeneity of beliefs. Detemple and Selden (1991)
considers a one-period general equilibrium model with a risky stock and two classes of investors with diverse beliefs
about the risk of the stock payoff. They show that when option is introduced, it is traded (bought by high risk
perception investor and sold by low risk perception investor) in a noisy rational expectations equilibrium and ends
up changing (increasing) the stock price as the investor-specific demand for the stock changes. A further
informational effect on the stock price is generated as the investors condition their beliefs on the option price as
well.
9
For example, Back (1993) finds that the perceived conditional density of the terminal stock price may be bimodal
under asymmetric information. Back considers the impact of information asymmetry in a market microstructure
model of Kyle (1985) to include a call option. With one risk-neutral informed trader, some uninformed liquidity
traders and market makers, options trading creates a stochastic volatility framework leading to an imperfect
correlation between the stock and the option. Not only Black-Scholes type option pricing no longer prevails, option
orders in fact convey different information than stock orders. For example, a call option buy order increases the
probability of option expiring in-the-money disproportionately more than a stock buy order
10
An extensive empirical literature exists on the effect of equity options introduction on the underlying stocks.
According to Conrad (1989), Detemple and Jorion (1990), Kim and Young (1991), earlier equity options
introductions led to an increase in the price of the optioned stocks. Such benefits are, however, not visible for later
option listings and may in fact have reversed into negative impacts (Sorescu (2000)). A number of studies (e.g.,
Conrad (1989), Skinner (1989), Damodaran and Lim (1991)) find option listings to stabilize the variance but having
no impact on the beta. The variance effect in the US market is by no means uncontested. In the Canadian market,
Elfakhani and Chaudhury (1995) find a variance and beta stabilization effect of the earlier option listings; the
K
S tC X i qit (1)
i =1
Cit
where qit = K
.
i =1
C
it
The weight, qit, attached to a given strike price, Xi, of a call option, is the net open interest of
that strike price relative to the aggregate net open interest of all call options at time t. Since these
weights are by definition between 0.0 and 1.0 and sum to 1.0, they can be construed as probabilities.
The price predictor COP can thus be viewed as an expected terminal stock price where the
expectation is with respect to the time t discrete distribution of the open interests for all call options
variance effect however reverses following the later option listings. While Conrad (1989) and Kim and Young
(1991) find no impact of the US listing of put options on the optioned stocks, Elfakhani and Chaudhury (1995) and
Chaudhury and Elfakhani (1997) report Canadian evidence that supports a risk reduction effect associated with put
option listings.
L
S tP X l qlt (2)
l =1
ltP
where qlt = L
l =1
P
lt
The weight, qlt, attached to a given strike, Xl, of a put option, is the net open interest of that
strike price relative to the aggregate net open interest of all put options at time t.
Since financial markets may not be complete and information-related imperfections may be
prevalent, a put option and a call option with similar terms may not be the mirror image of each
other. Hence, open interests of both put options and call options may convey information about the
terminal stock price. Accordingly, we derive a third predictor, CWOP, combining the open interests
of both call and put options:
K L
Cit X i + P
lt Xl
S
t
Z i =1
K
l =1
L
(3)
i =1
C
it +
l =1
P
lt
Assume that the positive private information of a group of informed investors reflect a more
favorable distribution of the stock price at T leading to a higher (than the current price St) expected
equilibrium stock price, E(ST).11 Since their information is only probabilistic, one likely speculative
strategy is to establish buy positions in out-of-the-money options of various strikes.12 The more
11
For simplicity, we assume risk neutral investors and a zero risk-free rate.
12
Throughout this paper, we assume that there are competitive risk-neutral market makers in both stock and options
markets.
13
For most optioned stocks, margin buying already represents some leverage. As such, in-the-money call options are
not likely attractive to the optimistic informed investors who choose to trade in the options market.
A. Data
The daily closing call and put options data (price and open interest by strike) used in this
study was collected from the online delayed quote reporting system of Dreyfus Brokerage Services
(DBS). These quotes are derived from the Market Data Report of the Chicago Board of Options
Exchange (CBOE).14 Our sample period spans the six consecutive option months of February,
March, April, May, June, and July of 1999. We define an option month as the period between the
two consecutive option expiration dates. For example, the February option month extends from the
first day of trading after the options expiration date in January to the last day of trading before the
options expiration date in February.15
Our sample consists of thirty popularly held companies chosen from the NASDAQ and the
New York Stock Exchange (NYSE). A list of these companies and the sectors that they represent
are provided in Table I. These companies were selected to represent major market indexes, a cross
section of important sectors and active options trading on the CBOE. The distribution by index is
as follows: DJIA (11), S&P 500 (7), NASDAQ 100 (7), NASDAQ Composite (3) and DJTA (2).
The sample firms represent seven broad sectors: technology (11), services (10), consumer products
(3), basic materials (2), conglomerates (2), energy (1) and healthcare (1).
14
We cross-checked the DBS data against the CBOE data available online and found no major discrepancies.
15
The options expiration date in any calendar month is usually the third Friday of the calendar month. In case, the
third Friday is a holiday, we use the last trading day before expiration as the last day of the option month.
B. Methodology
Our active investor (learner) may choose one of the following four types of basic strategies
with respect to a given stock: passive buy-and-hold (stock only), Merton et al type covered call
(stock plus option) strategies, open interest based active strategy using only stock, and open interest
based limited risk active strategies using stock plus options. For each type of strategy, equally
weighted portfolio returns are calculated averaging the returns from positions with respect to the
individual stocks in the sample.
For all strategies, positions are established using the closing prices of a trade initiation day (t)
within an option month and the positions are liquidated using the closing prices on the last day of
trading (T) during the option month. Thus, options that are permissible for trading have less than
thirty calendar days to expiration. Within an option month, we allow trades to be initiated on four
trading days: the second Monday (2M), the second Friday (2F), the Friday before the expiration
Friday (LF) and the last Monday, i.e., the Monday of the expiration week (EXM).16 Thus, the
corresponding holding periods in terms of calendar days are about 20 days (2M), 14 days (2F), 7
days (LF) and 4 days (EXM).
We chose to study short (less than a month) holding periods since information nowadays
circulate quite rapidly due to the explosive growth of internet usage specially among the investing
population. As such, learning has become easier and faster and information based trading is
expected to impound the information into prices without significant delays. Also, active traders tend
to have a short horizon. The 2M trade initiation day is chosen to allow about a week of option
trading during the option month following the last expiration, so it allows a week of learning or
16
As is customary, if a Monday is a holiday, trading is initiated using the closing prices of the next available trading
day. Similarly, if Friday is a holiday, positions are liquidated using the closing prices of the immediately preceding
trading day available.
In this section, we shall delineate the specifics of the trading strategies with respect to a given stock
that our hypothetical learner investor may follow. These strategies are respectively the passive buy-
and-hold (stock only) strategy, Merton et al type covered call (stock plus option) strategies, open
interest based active strategy using only stock, and open interest based limited risk active strategies
using stock plus options.
For a given stock, the buy-and-hold strategy simply involves buying shares at the closing
price of the trade initiation day. The number of shares bought is determined by dividing the total
dollar amount to be invested ($10,000) by the closing stock price on the trade initiation day and
rounding to the closest integer if necessary. It is assumed that investors are not allowed to borrow
securities on margin, and the risk free interest rate is assumed to be zero as the interest up to three
weeks is inconsequential. We do not consider brokerage fees for stock trading since many discount
brokers charge as low as ten dollars or less.
The expected dollar returns for the buy-and-hold strategy for a stock is:
BH =[
( S T S t ) N s ] I
(1)
N S = 10,000 / S t
where, ST is the closing stock price at the option maturity date, St is the closing stock price on the
trade initiation day, N S is the number of shares bought, and I is the net outlay. Here,
An equally weighted portfolio of all thirty stocks is then formed. We call this portfolio the nave
investor or NI portfolio. The holding period return on this portfolio is simply the average return on
the thirty stocks from the specific trade initiation day to the end of the option month.
A covered call strategy differs from a buy-and-hold strategy in that the dollar loss on the
stock is reduced by the option premium received and the dollar gains on the stock are capped by the
strike price (if the written call is out-of-the-money when the position is initiated). Thus, ex ante,
compared to the buy-and-hold strategy, the covered call strategy has lower return potential and less
[ ( ) { }
CC = (S T S t ) N S + (C it N C ) 100 TC (max 0, ( S T X i ) N C 100) ] (3)
where, Cit, is the closing price of the call option with strike Xi on the trade initiation day and TC is
the total transaction cost for the call option transactions, and NS/100 = NC is the number of call
option contracts written and is equal to $10,000/St rounded to the nearest integer. Since option
trades incur a fixed ordering cost plus a per contract transaction cost, TC can be significant
percentage wise for orders of small value.
The initial investment required on a fully covered position, ICC, is given by:
I CC = N S S t (C it N C ) * 100 (4)
Three different equally weighted portfolios are formed using the covered call positions of
the thirty individual stocks. These portfolios are named OMP (out-of-the-money calls written), AMP
(at-the-money or nearest-to-the-money calls written) and IMP (in-the-money calls written). The
portfolio return is just the average return on the thirty individual stockscovered call positions.
At the close of a trade initiation day, t, our active investor estimates the expected terminal
stock price, StZ, based on the distribution of open interests of call and put options of various strike
prices but all maturing at the same date, T. If the predictor StZ is greater (lower) than the
contemporary security price St, the investor considers this as a buy (sell) signal and goes long (short)
on the stock.
The expected dollar returns and the percentage returns for the active long stock strategy are
as in equations 1 and 2. The expected dollar and percentage returns for the active short stock strategy
are given by the following equations:
AS =[
S ti S Ti ] N S I (6)
We are assuming that whether the stock is bought or sold short, the investor has to deposit
the equivalent of trade value with the broker. In our case, this amount is $10,000, i.e., the assumed
investment sum. Thus, our active stock investment strategy, like the buy and hold strategy, assumes
no margin buying of the stock and 100% margin requirement for short selling.
When we present the empirical results, we shall mostly rely on the average percentage return
on the active strategy with respect to the thirty individual stocks in our sample. This means that the
average percentage return will reflect the returns from active long positions in some stocks and active
The open interest based active stock strategy can be quite risky for individual stocks. We,
therefore, consider limited risk active strategies involving stocks and options. As in the active stock
strategy, on each trade initiation day t, our investor compares the actual closing stock price St and
the open interest based price predictor StZ. Here, however, the investor considers the magnitude as
well as the direction of predicted price movement. Suppose X is the next available strike price in the
direction of the predicted price movement. It is considered a major price movement if StZ either goes
past X or at least is closer to X than it is to the current stock price St. Otherwise, it is categorized as
a minor price movement. By and large, the cases identified as major (minor) price movements in our
sample represented a predicted change of more (less) than 5% (2%) from the current stock price.
According to the direction and the magnitude of the predicted stock price movement, we
have the following four cases:
ASP1: The price predictor signals a minor upward movement in the stock.
ASP2: The price predictor signals a major upward movement in the stock.
ASP3: The price predictor signals a minor downward movement in the stock.
ASP4: The price predictor signals a major downward movement in the stock.
We shall now discuss the four cases above and the active trading strategies our investor
will follow under these cases.
D.1 ASP1: The Price Predictor Signals a Minor Upward Movement in the Stock
If the price predictor indicates a minor upside for a stock, our investor buys the stock at the
closing market price on the trade initiation day and writes a call option. The strike price of the
written call option is chosen based on the upside potential indicated by the price predictor and the
availability of strike prices. If, for example, the current stock price is $55, the predicted stock price at
[ ( ) { } ]
ASP1 = (S T S t ) N S + C xi N C TC 100 max 0, ( S T X i ) N C 100 I ASP1
(8)
whereX i S tZ
Here the symbol means closest to StZ in the direction of the predicted price movement.
The initial investment required for the strategy ASP1, is given by:
In the equations above, NS=NC x100, is the number of shares that can be purchased with
$10,000 rounded to the nearest 100. This is because in order to write one fully covered call option
contract, the investor needs to have 100 shares. For the sake of convenience, we allow the investor
to buy 200 shares and write 2 call contracts if the number of shares that could be purchased with
$10,000 is 180. Therefore, the actual net investment could be higher or lower than $10,000. The
percentage return, PROIASP1, is calculated on the net investments I ASP1 . The options transaction
cost, TC , is the transaction cost of writing NC option contracts. Lastly, C Xi is the time t closing price
of the call option with strike Xi.
If the stock does move up but does so substantially (not expected), the return from the
strategy ASP1 will be positive but less than the active stock strategy AS (long). However, if the stock
stays roughly the same, moves up but not beyond the chosen strike for the written call, or in fact
drops a little, then the strategy ASP1 may still provide a positive return that is greater than the return
from the strategy AS (long). If the stock moves down by more than the net proceeds from the written
call, the return from ASP1 will be negative but it will not be as bad as the return from the strategy
D.2 ASP2: The Price Predictor Signals a Major Upward Movement in the Stock
If the price predictor indicates a major upside for a stock, as in strategy ASP1 our investor
pursues a limited risk active covered call strategy. The investor buys the stock at the closing market
price on the trade initiation day and sells a deep-out-of-the-money (significantly higher strike) call
option; the strike is chosen based on the upside potential indicated by the price predictor and the
availability of strike prices. If, for example, the predicted stock price at maturity is $64, signaling a
major upside from the current stock price of $55, our investor writes a covered call at strike around
65. If, on the other hand, the predicted stock price at maturity were $61 or even $59, still signaling a
major upside according to our classification, the chosen strike would have been $60.
The dollar return, the investment and the percentage return for the strategy ASP2 are then as
in equations (8), (9) and (10). This, however, will only make sense if the premium received for the
deep-out-of-the-money calls written exceeds the options transaction cost. Since the options in our
where Cyi is the time t closing price of a call option with strike Xyi St . Since the premium for the
deep-out-of-the-money (Xyyi StZ >> St) call option net of the transaction cost (when jointly ordered
with the purchase of at-the-money call option) will be of relatively small magnitude, we continue to
include the writing of the deep-out-of-the-money call option. The dollar return, the initial
investment and the percentage return on ASP2 are then estimated as:
{( ) } ( )
(S T S t ) N S + max S T X Yi N C ,0 100 (CYi N C ) 100 + TC
ASP 2 =
+ ( (C yyi N C ) 100 TC) {(
max S T X yyi )
N C ,0}100
(12)
X yi S t , X yyi S t Z
The purchase of at-the-money call options to partially replace the purchase of shares makes
the strategy ASP2 more bullish (more leveraged) than when deep-out-of-the-money option premium
exceeds the options transaction cost. However, the number of such cases is rather small and has no
material impact on the overall (portfolio) returns of the strategy ASP2. Hence, for ASP2, we shall
pretend from hereon that the option premium received is large enough to handily offset the
transaction cost.
If the stock does move up as expected, the return from the strategy ASP2 will be positive
and greater than both the active stock strategy AS (long) and the active covered call strategy ASP1
D.3 ASP3: The Price Predictor Signals a Minor Downward Movement in the Stock
If the price predictor indicates a minor downside for a stock, our investor writes a covered
call at a strike near the closing stock price on the trade initiation day. If the stock goes down to the
somewhat lower predicted level but not below, the loss on the long stock will be outweighed by
premium received for the at-the-money call written assuming that the call premium is greater than
the expected drop in the stock. In case the stock stays flat or moves up, there is no loss on the long
stock as the investor either keeps the stock or gives it away at the strike close to the purchase price.
Meantime, the investor bags the call premium. The worst case for this strategy is when the stock
drops significantly below the predicted level. In that case, the strategy ASP3 will lose money
although the call premium cushions the loss to a degree.
If the investor instead went for short selling, the dollar gains would be more if the stock
drops as predicted. However, not only the required investment is larger (due to the premium
received under ASP3) relative to the strategy ASP3, the short selling strategy is destined to lose
money if the stock moves up instead of moving down. The same is true for naked put buying. In
other words, short selling and naked put buying are more volatile bearish alternatives than the
strategy ASP3.
The dollar returns, the initial investment and the percentage return for the strategy ASP4 are
given by the following equations:
In the above equations, PZi is the price of the put option with strike X Zi and CVi is the
price of the call contract with strike X Vi . The transaction costs are TC for NC call option contracts
D.4 ASP4: The Price Predictor Signals a Major Downward Movement in the Stock
If the price predictor indicates a major downside for a stock, our investor writes a covered call
at a strike close to the predicted stock price at maturity and buys a put option with the strike near or
above the closing stock price on the trade initiation day. If the stock goes down to the predicted
level or below, the loss from the long stock will be covered by the put options payoff, the written
call be worthless, and the net premium received (price of the in-the-money call written minus the
price of the at-the-money put option bought) can be considered as profit. The net premium plus the
put payoff would likely exceed any loss from the exercise of written call when the stock does not
drop all the way to the predicted low price. In case the stock moves up, profit from the stock would
offset further losses on the written call. The worst case for this strategy if the stock remains flat and
the net premium is less than the intrinsic value of the written in-the-money call.
If the investor instead went for short selling, the dollar gains would be more if the stock
does actually drop significantly as predicted. However, not only the required investment is larger
(when the net premium is positive), the loss is also greater if the stock moves up instead of moving
down. The same is true for naked put buying. In other words, short selling and naked put buying are
more volatile bearish alternatives than the strategy ASP4.
( ) {( ) }
(S T S t ) N S ( PZi N P ) 100 + TP + (max X Zi S T N P ,0 100)
ASP 4 =
( X Vi S T ) N C ,0 100) + (CVi N C ) 100 TC )
(max
(18)
X Zi = put strike closest toS t , and X Vi = call strike closest to S tZ .
In the above equations, PZi is the price of the put option with strike X Zi and CVi is the price
of the call contract with strike X Vi . The transaction costs are TC for NC call option contracts written
and TC for NP put option contracts bought, and NC = NP = NS/100 where NS is $10,000/St rounded
to the nearest 100.
In this section, we first discuss (Table III) how the active strategy performs for a given option month
and trade initiation day and compare this with the nave investors buy and hold strategy. Second, the
prediction accuracy of the open interest based price predictor is analyzed (Table IV). Third, the
comparative performance results for the various option months and trade initiation days are
reported for the various passive and active strategies (Table V). These strategies are the nave
investors buy and hold strategy (NI), the S&P 500, the Merton et al (1978) style passive covered call
strategy (OMP: out-of-the-money call written, AMP: at-the-money call written and IMP: in-the-
money call written), the open interest based stock only limited risk active strategy (AS) and the stock
plus options limited risk active strategy (ASP, ASP1: minor up predicted, ASP2: major up predicted,
ASP3: minor down predicted, and ASP4: major down predicted). For the active strategies, we only
report the performance results for the CN (Consider News) situation where the stocks with major
To obtain a sense of how the active positions are established and aggregated to arrive at
portfolio results, Table III provides an example of detailed stock by stock prediction, strategy and
performance for the stock only active strategy (AS). It shows for the February option month of 1999
the comparative performances of the buy and hold strategy (NI or Nave Investor Portfolio), the
stock only active strategy considering news (CN or OPP Portfolio) and the stock only active strategy
ignoring news (IN Portfolio). The active strategies here were initiated on 2F or the second Friday of
the February option month (the third last Friday counting backward from the option expiration
Friday in February 1999). That is, the options had about 14 days to expiration as of the trade
initiation day.
The first column of Table III indicates the symbols of the sample stocks. The second
column displays the closing prices, ST, at the option maturity date. The third column shows the open
interest based stock price prediction, StZ, where t is the second Friday of the February 1999 option
month. The fourth column indicates the closing stock price, St, on our trade initiation day t. The
fifth column shows the direction of movement of the stock signaled or predicted by the predictor. If
St> StZ, it is identified as Down. If, on the other hand, If St< StZ, it is identified as "Up". The sixth
column shows the stock only active trading strategy (BUY for long and SS for short selling) for our
hypothetical learner investor. The seventh column refers to the number of shares the investor buys
or sells short with $10,000. The last three columns display the PROI for 14 days holding period for
the NI, CN (OPP) and the IN portfolios.
The results from the table indicate that our hypothetical investor could have earned a return
of 0.86% for the two-week holding period in the February 1999 option month following the stock
only active strategy and excluding the one stock (MO) that had a major news event during the
holding period. If this wisdom or foresight is taken away, the return to the stock only active strategy
would have been 0.39%. While the impact of the news event is important, it however does not
B. How Accurate is the (Up or Down) Prediction of the Open Interest Based Predictor?
In Table III, we see that while the stock only active strategy did have a mean return
advantage during the two-week period of the February 1999 option month, the up or down
predictions for the sample individual stocks were not always right. If the return advantage arises due
to accurate prediction in only a minority of stocks and only in a few months, then the strategy may
not be reliable for replication in general. We now, therefore, look at the prediction (up or down)
accuracy of the open interest based predictor in the six option months in our sample considering all
four trade initiation days within an option month. Since there are thirty stocks in our sample, we
have a total of 120 cases of prediction in each of the six option months.
In the three option months of February (59%), April (70%) and May (65%), a majority of
the sample stocks actually lost their value during the holding periods. On the other hand, in the
option months of March (53%), June (77%) and July (63%), a majority of the sample stocks actually
marched higher during the holding periods. Thus, loosely speaking we might refer to February, April
and May as the down option months in our sample and to March, June and July as the "up
option months.
In comparison, our option based predictor calls for a down option month in all option
months except June. Thus there seems to be a downward bias in our open interest based prediction.
This is somewhat expected as our prediction method is solely based on the stock-specific open
interest of options and does not take into account any market wide factor, nor does it predict up or
Lastly, Table V presents the comparative performance results for the various passive and
active strategies in the six option months considering the four alternative trade initiation days (2F,
2M, LF, EXM) within each option month.
Considering all 120 cases of prediction, a passive investor investing in the S&P 500 would
have earned 1.53% return on average. If the passive investor followed the equally-weighted buy and
hold strategy, the average return would have been 1.00%. Considering that the average holding
period is about 11 to 12 days, these returns translate to about 49% annualized return for the S&P
500 and about 32% annualized return for the equally weighted portfolio of the thirty blue chip
sample stocks. Given that 1999 was a stellar year for stocks, these returns appear realistic.
Now consider the aggregate performance of the open interest based stock only active
strategy. Following this strategy, our hypothetical learner investor could expect to earn 9.05% return
on average. This translates to an annualized return of about 2600%. By any means, the return
17
The standard deviation of the S&P 500 holding period return for a given option month/ trade initiation day
combination cannot be calculated as it is a single number.
In this paper, we use daily closing data on CBOE options of 30 stocks during February through July
of 1999 to investigate whether options open interest contains information that can be used for
trading purposes. Individual stock price at option maturity is first predicted based on the distribution
of options open interest. Several stock only and stock plus options directional trading strategies are
then considered after comparing the predicted stock price at maturity and the actual stock price at
the trade initiation date.
We find the prediction of stock price movement based on the distribution of options open
interest to have reasonably good accuracy. The prediction accuracy is not due to accuracy only in
specific option month or trade initiation day. In our sample, the open interest based active trading
strategies generate better returns compared to the passive benchmarks. The stock only active
strategy yields significantly higher return than the S&P 500 and the nave investors buy and hold
strategy involving the sample stocks. Since our hypothetical learner investor faces the risk of
incorrect information or inaccurate learning, the investor might prefer limited risk speculative
strategies involving stock plus options. In this context, our benchmark is Merton et al (1978) style
covered call strategy. Here also, we find that the open interest based active strategy provides
significantly higher return than the passive covered call strategies.
Not only the risk of the active strategies (naked and limited risk) seems close to the risk of
the benchmarks, the magnitude of the return advantage seems too high to be nullified by any risk
disadvantage there may be due to the use of a rough proxy for risk in this paper. We, therefore,
conclude that the equity options open interest contains valuable information that is attractive for
trading purposes.
18
Alternatively, one may combine the distribution of open interests, the distribution of option volume, historical
time series based distribution, the risk neutral distribution and/or the implied volatility surface to minimize forecast
error of the predicted price or higher order moments of the assets true future distribution.
[ 2] Back, Kerry, 1993, Asymmetric information and options, Review of Financial Studies 6, 435-
472.
[ 3] Back, Kerry, Henry Cao, and Gregory Willard, 2000, Imperfect competition among
informed traders, Journal of Finance 55, 2117-2155.
[ 4] Bhattacharya, Mihir, 1987, Price changes of related securities: The case of call options and
stocks, Journal of Financial and Quantitative Analysis 22, 1-15.
[ 5] Biais, Bruno, and Pierre Hillion, 1994, Insider and liquidity trading in stock and options
markets, Review of Financial Studies 7, 743-780.
[ 6] Black, Fisher, 1975, Fact and fantasy in use of options, Financial Analysts Journal 31, 36-41,
61-72.
[ 7] Brennan, Michael J., and H. Henry Cao, 1996, Information, trade, and derivative securities,
Review of Financial Studies 9, 163-208.
[ 8] Chaudhury, Mo, and Said Elfakhani, 1997, Listing of put options: Is there any volatility
effect?, Review of Financial Economics 6, 57-75.
[ 9] Cherian, Joseph A., and Anne Fremault Vila, 1997, Information trading, volatility, and
liquidity in options markets, The Research Foundation of the Institute of Chartered Financial Analysts:
AIMR Publication Series.
[10] Cherian, Joseph A., 1998, Discretionary volatility trading in option markets, Working Paper,
Boston University.
[11] Cherian, Joseph A., and William Y. Weng, 1998, An empirical analysis of directional and
volatility trading in options markets, Working Paper, Boston University.
[12] Conrad, Jennifer, 1989, The price effects of option introduction, Journal of Finance 44, 487-
498.
[13] Copeland, Thomas E., 1976, A model of asset trading under the assumption of sequential
information arrival, Journal of Finance 31, 1149-1168.
[14] Copeland,Thomas E., and Dan Galai, (1983), Information effects and the bid-ask spread,
Journal of Finance 38, 1457-1469.
[16] Detemple, Jerome, and Phillipe Jorion, 1990, Option listing and stock returns, Journal of
Banking and Finance 14, 781-801.
[17] Detemple, Jerome, and Larry Selden, 1991, A general equilibrium analysis of option and
stock market interactions, International Economic Review 32, 279-302.
[18] Detemple, Jerome and Shashidhar Murthy, 1994, Intertemporal asset pricing with
heterogeneous beliefs , Journal of Economic Theory 62, 294-320.
[19] Easley, David, and Maureen OHara, 1987, Prices, trade size, and information in security
markets, Journal of Financial Economics 19, 69-90.
[20] Easley, David, Maureen OHara, and Srinivas P.S., 1998, Option volume and stock prices:
Evidence on where informed traders trade, Journal of Finance 53, 431-465.
[21] Elfakhani, Said, and Mo Chaudhury, 1995, The volatility effect of option listing: Some
Canadian evidence, Quarterly Review of Economics and Finance 35, 97-116.
[22] Glosten, Lawrence, and Paul Milgrom, 1985, Bid, ask, transaction prices in a specialist
market with heterogeneously informed traders, Journal of Financial Economics 13, 71-100.
[23] Grossman, Sanford, 1988, An analysis of the implications for stock and futures price
volatility of program trading and dynamic hedging strategies, Journal of Business 61, 275-298.
[24] He, Hua, and Jiang Wang, 1995, Differential information and dynamic behavior of stock
trading volume, Review of Financial Studies 8, 919-972.
[25] John, Kose, Apoorva Koticha, and Marti Subrahmanyam, 1993, The micro-structure of
options markets: Informed trading, liquidity, volatility, and efficiency, Working Paper, New
York University.
[26] John, Kose, Apoorva Koticha, Ranga Narayanan, and Marti Subrahmanyam, 2000, Margin
rules, informed trading in derivatives and price dynamics, Working Paper, New York
University.
[27] Kim, Wi Saeng., and Colin M. Young, 1991, The effect of traded option introduction on
shareholder wealth, Journal of Financial Research 14, 141-151.
[28] Kyle, Albert, 1985, Continuous auctions and insider trading, Econometrica 53, 1315-1335.
[29] Leisen, Dietmar and Kenneth Judd, 2001, Equilibrium open interest, Working Paper, Faculty
of Management, McGill University.
[31] Merton, Robert C., Myron Scholes, and Mathew Gladstein, 1978, The returns and risk of
alternative call option portfolio investment strategies, Journal of Business 51, 183-242.
[32] Mayhew, Stewart, Atulya Sarin, and Kuldeep Shastri, 1995, The allocation of informed
trading across related markets: An analysis of the impact of changes in equity-option margin
requirements, Journal of Finance 50, 1635-1653.
[33] Sheikh, Aamir M., and Ehud I. Ronn, 1994, A characterization of the daily and intraday
behavior of returns on options, Journal of Finance 49, 557-580.
[34] Skinner, Douglas, 1989, Options markets and stock return volatility, Journal of Financial
Economics 23, 62-78.
[35] Sorescu, Sorin, 2000, The effect of options on stock prices: 1973 to 1995, Journal of Finance
55, 487-514.
[36] Stephan, Jens A., and Robert E. Whaley, 1990, Intraday price change and trading volume
relations in the stock and stock options markets, Journal of Finance 44, 115-134.
[37] Vijh, Anand M., 1988, Potential biases from using only trade prices of related securities on
different exchanges, Journal of Finance 43, 1049-1055.
[38] Vijh, Anand M, 1990, Liquidity of the CBOE equity options, Journal of Finance 45, 1157-1179.
*A major news event took place for the stock after the trade initiation day.
If St < StZ , then the prediction or signal is an upward move (UP) for the stock from t to T and the hypothetical active
investor goes long on the stock (BUY). If St > StZ , then the prediction is a down move (DOWN) for the stock from t to
T and the hypothetical active investor goes short on the stock (SS). For each stock, the number of shares (NS) bought or
shorted is $10,000/ St. We assume 100% margin deposit for both stock purchase and short selling. Stock transaction
costs are assumed to be zero. The percentage return on investment (PROI) is calculated as 100x dollar return ()/
Investment (I). For the buy and hold strategy and the stock only active strategy, I is $10,000. For the buy and hold
strategy and the long stock active strategy, = ST St per share. For the short stock active strategy, = St ST per share.
Portfolio type NI (Nave Investor) indicates a strategy of buying on trade initiation day and holding till next option
expiration day with respect to each of the thirty sample stocks. In the CN and IN portfolios, on the other hand, the
hypothetical investor goes long (short) on the stock if the open interest based price predictor signals an UP (DOWN)
market for the stock from the trade initiation day to the next option expiration day. The difference between CN
(Consider News) and IN (Ignore News) is that IN includes active strategy positions in all stocks even if some of the
stocks had major news events after trade initiation day that led to major swings in the stock price. CN, on the other
hand, excludes the news event stocks and as such assumes that our directional trader had expected such news although
the direction of the stock impact was not known.
June July
Actual Up (% of Total) 92(77%) 76(63%)
Actual Down (%of Total) 28(23%) 44(37%)
---------------------------------------------------------------------------------------------------------------------------------------------------
Predicted Up (% of Total) 78(65%) 54(45%)
Predicted Down (% of Total) 42(35%) 66(55%)
Correct Up (% of Predicted Up) 60(77%) 47(87%)
Correct Down (% of Predicted Down) 10(24%) 37(56%)
_____________________________________________________________________________
Portfolio/Strategy Portfolio/Strategy
Option Trade AS NI S&P ASP Merton et al (1978) Covered Call:
Month Initiation (CN) 500 (CN) OMP AMP IMP
Day, t
FEBRUARY 2F 8.03 -8.06 -0.02 9.46 -1.95 0.64 2.07
17.18 18.31 12.34 13.59 9.73 4.47
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Aggregate:
Average PROI 9.05 1.00 1.53 11.20 1.96 2.43 2.14
Average St. Dev. 22.64 25.07 19.05 22.48 14.34 7.74
.
If St < StZ, then the prediction or signal is an upward move (UP) for the stock from t to T and the hypothetical active
investor goes long on the stock (BUY). If St > StZ, then the prediction is a down move (DOWN) for the stock from t to
T and the hypothetical active investor goes short on the stock (SS). For each stock, the number of shares (NS) bought or
shorted is $10,000/ St. We assume 100% margin deposit for both stock purchase and short selling. Stock transaction
costs are assumed to be zero. The percentage return on investment (PROI) is calculated as 100x dollar return ()/
Investment (I). For the buy and hold strategy and the stock only active strategy, I is $10,000. For the buy and hold
strategy and the long stock active strategy, = ST St per share. For the short stock active strategy, = St ST per share.
The stock plus options limited risk active strategy (ASP) is based on the direction (Upward or Downward) as well as the
magnitude of stock price movement as predicted by the open interest based predictor StZ. The magnitude is considered a
major move if StZ either goes past or at least is closer to the next available strike in the direction of the price move than it
is to the current stock price St. By and large, this meant a change of more than 5% in our sample. Minor moves mostly
meant a change of less than 2% in our sample. If a flat to minor upward movement is predicted, the investor pursues
ASP1: writes covered call with strike close to the predicted price. If a major upward movement is predicted, the investor
pursues ASP2: writes covered call with strike close to the predicted price. However, if the call premium is not large
enough, the strategy is instead to buy shares and at-the-money call options and write calls with strike close to the
predicted price. If a minor downward movement is predicted, the investor pursues ASP3: writes a covered call with strike
close to the current stock price. If the price predictor indicates a major downside for a stock, the investor pursues ASP4:
writes a covered call at a strike close to the predicted stock price and buys a put option with the strike near or above the
closing stock price on the trade initiation day. The initial investment for the ASP strategies applied to a stock varies
somewhat from $10,000. This is because the number of shares was rounded to the nearest 100 as a CBOE equity option
contract is for 100 shares.