OPTIONS LISTING, MARKET LIQUIDITY AND STOCK BEHAVIOUR: SOME CANADIAN EVIDENCE

Journal of Btcsincss Finance @Accounting, 20(5), September 1993, 0306 686 X

OPTIONS LISTING, MARKET LIQUIDITY AND STOCK BEHAVIOUR: SOME CANADIAN EVIDENCE

TREVOR w. CHAMBERLAIN, c. SHERMAN CHEUNG AND CLARENCE C.Y. KWAN*

INTRODUCTION

Stock option trading activity in the United States expanded quickly following the advent of organized markets offering standardized contracts in the mid 1970s. According to popular wisdom of the time,' the growth of options markets would have important consequences for the liquidity, price behaviour and trading volume of the underlying stocks. Subsequent studies have generally indicated that price volatility and trading volume effects do, indeed, occur around the time of options listing. However, in most instances, the evidence is based on short sampling intervals or volatility measures that only consider the systematic risk component. As for the association between options listing and stock liquidity, no direct tests have been reported.'

The organization of options markets in Canada and other countries soon followed the lead of the United States. However, unlike the US experience, there is no reported evidence on the impact of introducing options on the underlying stock market. This is in spite of the fact that in recent years options have been frequently accused of contributing to the volatility of these markets.

In an effort to inform the discussion of the effects of option listing and to establish whether the effects observed in the United States also persist in Canada, the present study was undertaken. The latter issue is of particular interest because previous studies involving other aspects of the Canadian options market, such as Halpern and Turnbull (1985) and Chamberlain, Cheung and Kwan (1989), suggest that the Canadian experience does not always mirror that of the United States4 This may be because the Canadian markets are shallower than their US counterparts and also because the participation rate in the options market, relative to the stock market, is much lower among Canadian investors.

Unlike previous studies, the present one examines both immediate and gradual changes in stock return volatility and trading volume following the listing of options. A further concern, not considered in earlier work, is the impact of listing on the underlying stock's liquidity, as measured by the bid-ask spread.

The authors are from the Faculty of Business, McMaster University, Hamilton, Ontario, Canada. Financial support for this research was provided by the Social Sciences and Humanities Research Council of Canada. The authors wish to thank the anonymous referee for helpful comments on earlier versions of this paper. (Paper received June 1991, accepted July 1991)

0 Basil Blackwell Ltd. 1993, 108 Cowley Road, Oxford OX4 IJF, UK and 238 Main Street, Cambridge, MA 02142, USA.

688 CHAMBERLAIN, CHEUNC AND KWAN

The possibility of cross-sectional variations in volume and liquidity at the time of listing is also examined. The study begins with a brief review of the US evidence, followed by a description of the data. It then goes on to describe the analysis together with the results obtained. The study concludes with a summary of its main findings and their implications.

PREVIOUS STUDIES

The effects of options listing on the underlying stocks were recently examined by Skinner (1989). Utilizing daily data for varying periods around the listing date, Skinner found, among other things, that volatility, as measured by the variance of returns, fell after options listing whereas beta remained unchanged. The change in variance he ascribed to a possible increase in the liquidity of the market for the stock. Skinner’s results generally confirm those obtained in a number of studies done in the early years of organized options markets. In the first of these (Nathan, 1974), the standard deviation of weekly returns of the sixteen stocks originally listed on the CBOE was found to have declined after the listing date, even after taking account of market effects. This decline occurred in spite of the fact that, initially, options were, for most investors, unfamiliar instruments and thus did not attract a broad base of interest.

Two studies by the Chicago Board Options Exchange (CBOE 1975 and 1976) confirmed Nathan’s results; that is, in both, a statistically significant decline in price volatility occurred following the options listing date. Klemkosky and Maness (1980) also examined volatility, as measured by monthly return variance, but found no consistent relationship between the pre- and post-listing estimates.

The primary concern of Klemkosky and Maness, however, was the possibility of changes in beta. While no particular pattern of change was observed, they did find that positive excess risk-adjusted returns disappeared following options listing. This they interpreted as indicating that options enhance the efficiency of the stock market.

Evidence that beta remains unchanged is also reported by Trennepohl and Dukes (1979) and Whiteside, Dukes and Dunne (1983), who used daily data to calculate returns. However, Whiteside et al. go on to point out that when their results are evaluated on a year by year basis, there is a tendency for post- listing betas to become progressively smaller relative to their pre-listing counterparts.

Whiteside, Dukes and Dunne also found that the number of shares traded was unaffected by the listing of options. In contrast, Hayes and Tennenbaum (1979), employing monthly data, concluded that option listing is associated with increases in trading volume. Their results are confirmed by Skinner (1989), who, as indicated above, utilized daily data and a much larger sample than was available to earlier researchers. However, like his predecessors, Skinner

0 Basil Blackwell Ltd. 1993

CANADIAN EVIDENCE ON STOCKS AND OPTIONS LISTING 689

confined his attention to average return and volume behaviour before and after options listing. No consideration was given to the possibility of a gradual change in return volatility or to cross-sectional effects at the time of listing.

THE DATA

Stock option trading in Canada was initiated by the Montreal Stock Exchange in 1975 and commenced shortly thereafter on the Toronto and Vancouver Stock Exchanges. A common clearing corporation - Trans-Canada Options Inc. - was subsequently formed to issue and guarantee option contracts. However, listing continued to be done on one of the three stock exchanges, with occasional changes in listing from one exchange to another.

The sample utilized in the present study comprises thirty-seven companies. As such, it includes all, but of the companies whose options were listed on any one of the three Canadian exchanges between November 1979 and January 1987. Though a number of options were trading prior to November 1979, their precise listing dates could not be established. No companies whose option listings occurred subsequent to January 1987 were considered because of the potentially confounding effects of the stock market ‘crash’ in October of that year. The companies included in the sample, together with the initial listing dates of their options, are reported in Table 1.

For each stock several daily price series (closing prices, highs, lows and final bid and ask prices) together with daily trading volume on the Toronto Stock Exchange (TSE) and monthly shares outstanding were obtained for a period beginning four months before and extending four months after each option listing date.6 The ex-dividend dates falling during the period together with the corresponding dividends were also collected. Daily returns for each stock, as measured by by (P, + - P, + D, + ,)/Pt, were then calculated. In addition, TSE 300 total return index data were obtained for each period and used to calculate daily market returns. This was done in order to allow tests for changes in volatility that abstracted from possible market effects.’

CHANGES IN RETURN VOLATILITY

T o test, first of all, whether there is a change in return volatility at the time of options listing, the before and after standard deviations of daily returns were calculated for each stock. Of course, a change in volatility could occur for reasons other than the listing of options. While other factors peculiar to individual stocks generally could not be identified, possible changes in daily market volatility were accommodated by dividing each stock’s standard deviation of returns by the standard deviation of market returns for the same period. In addition, in order to examine the possibility that any change in


690 CHAMBERLAIN, CHEUNG AND KWAN

Table 1

Optioned Stocks in Sample

Firm .%ling Date Firm Listing Date

Alberta Energy Apr. 21, 1981 Lac Minerals Mar. 21, 1984 Bank of Nova Scotia Dec. 11, 1979 Magna International ‘A’ Aug. 18, 1986 B.C. Forest Products May 21, 1986 Mite1 Jan. 11, 1982 B.C. Telephone Feb. 22, 1984 Numac Oil & Gas Jan. 15, 1985 BP Canada Jan. 05, 1981 PWA Jan. 05, 1987 Campbell Red Lake Minesa Jan. 05, 1981 Pagurian ‘A’ Aug. 18, 1986 Canada Development Corp.a Nov. 11, 1985 Placer Development Apr. 21, 1981 Canadian Imperial Bank Power Financial Aug. 18, 1986

of Commerce Mar. 24, 1980 Rogers ‘B’ Aug. 18, 1986 Canadian Occidental Dec. 03, 1984 Royal Trustco Sept. 02, 1986 CP Investmentsa Dec. 11, 1979 Sears Canada May 20, 1986 Canadian Tire ‘A’ Apr. 04, 1984 Spar Aerospace Dec. 03, 1984

Oct. 01, 1985 Carling O’Keefea Nov. 30, 1983 Teck ‘B’ Cominco Apr. 21, 1981 Texaco Canadaa Jan. 11, 1982 Echo Bay Mines Dec. 18, 1984 T D Bank Nov. 15, 1979 Falconbridge Mines May 20, 1986 Transalta Utilities Dec. 15, 1983 Hudson Bay Oil & Gas” May 09, 1980 Trilon Financial ‘A’ Jan. 08, 1985 International Corona Aug. 18, 1986 Walker-Consumersa Jan. 05, 1981 Inter-City Gas Jan. 05, 1981 Woodwards Jan. 13, 1986

Note: a Bid-ask spread data not available.

volatility does not occur immediately, the period following the listing date was divided into two subperiods of approximately equal length and the standard deviation of returns of each considered. Summary statistics for the before and after periods and subperiods, both with and without the market adjustment, are presented in Panel A of Table 2. While casual inspection suggests a slight decline in volatility immediately after the listing of the option, in no instance does a Wilcoxon signed rank test, using the pre-listing period as the benchmark, allow rejection of the null hypothesis that the distributions are the same.8

The use of closing prices to measure stock volatility is consistent with the practice of most previous authors. However, as Grammatikos and Saunders (1986) point out, the classical approach to the estimation of volatility does not consider other readily available information which might improve estimator efficiency. In order to take advantage of all ‘readily available’ daily price data Parkinson (1980) has developed estimators based not only on daily opening and closing prices, but also on daily highs and lows. One such estimator, which assumes that prices follow a Brownian motion with zero drift over both trading and nontrading periods, is

(InH, - 1nLf)*/(4nln2), I = 1

where H, and L, are the high and low prices observed on day t and n is the


Tab

le 2

Vol

atili

ty o

f S

amp

le o

f T

hirt

y-se

ven

Op

tio

ned

Sto

cks

Bef

ore

and

Aft

er O

pti

on

s L

isti

ng"

0

m 8. -

Pane

l A

: E

stim

ated

Sta

ndar

d D

evia

tion

of D

aily

Ret

urns

Stat

istic

B

efre

A

fkr

Afe

r I

Afe

r 2

&fore

Afte

r A

fer

I A

fter

2 $

Mea

n 0.

0204

7 0.

0200

8 0.

0196

7 0.

0201

0 3.

0083

2.

9266

2.

8941

3.

2379

5 P

3

Sta

ndar

d de

viat

ion

0.00

783

0.00

645

0.00

731

0.00

68 7

1.06

31

1.14

24

1.00

05

1.69

65 'L

W

ilcox

on te

st

-0.1

13

-0.5

81

0.20

4 -0

.656

-0

.596

0.

581

2

D n

'p <

Una

djus

ted

Mar

/xt-&

juste

&

P

-

Pane

l B

: D

aily

Hig

h-lo

w S

tatis

tic'

Stat

istic

Bc

forc

A

jer

Afe

r I

Afle

r 2

BeJa

re Afier

After

I

$2 U

nadj

tlrte

d M

drhf

-Odj

ustc

dd

z A

fter

2 0

Mea

n 0.

0002

90

0.00

0269

0.

0002

69

0.00

027

1 6.

0784

5.

8237

5.

5709

7.

5830

S

tand

ard

devi

atio

n 0.

0002

23

0.000 1

63

0.00

0197

0.

000 1

66

3.63

70

4.15

58

3.22

48

7.46

40 3

W

ilcox

on t

est

-0.4

60

-0.8

68

-0.0

30

-0.6

71

-0.3

39

1.23

0 8

Pane

l C

: E

stim

ated

Res

idua

l V

aria

nce

and

Bet

ae

Stat

istic

St

jiore

A

fn

A* I

Afte

r 2

Befor

e Af

tn.

Afe

r I

Aft

er2 '

Mea

n 0.

0003

58

0.00

0357

0.

0003

44

0.00

0344

1.

I987

1.

0369

1.

0497

1.

1143

2

Sta

ndar

d de

viat

ion

0.00

0288

0.

0002

13

0.00

0290

0.

0002

03

0.71

97

0.53

34

0.72

83

0.75

14 0

W

ilcox

on te

st

0.41

5 -0

.581

0.

060

- 1.

380

-0.3

55

-0.5

66

5

a A

fter

1 a

nd A

fter

2 r

efer

to th

e fir

st a

nd s

econ

d su

bper

iods

aft

er t

he l

istin

g of t

he o

ptio

n. T

he W

ilcox

on t

ests

use

the

pre-

listin

g pe

riod

as

the

benc

hmar

k. 3

0

E Re

sidua

l Va

rianc

e Be

ta

ii %

Not

es:

6 T

hat

is,

for

each

firm

the

stan

dard

dev

iatio

n of

dai

ly r

etur

ns d

ivid

ed b

y an

est

imat

e of

the

con

tem

pora

neou

s sta

ndar

d de

viat

ion

of d

aily

mar

ket

retu

rns.

C

alcu

late

d us

ing

the

Park

inso

n (1

980)

est

imat

or;

that

is,

"

(lnH

, -In

L,)

2/(4

nln2

),

1-

1

whe

re H

, and

L, a

re t

he h

igh

and

low

pri

ces

on d

ay 1

and

II

is th

e nu

mbe

r of

obs

erva

tions

. T

hat is,

for

each

fir

m t

he P

arki

nson

est

imat

e di

vide

d by

an

estim

ate

of t

he c

onte

mpo

rane

ous

vari

ance

OC

daily

mar

ket

retu

rns.

B

eta

is e

stim

ated

usi

ng th

e Sc

holc

s-W

iHja

ms

(197

7) e

stim

ator

.


number of observations. This estimator has been shown by Parkinson to have an estimation variance 5.2 times lower than that of the classical estimator based on closing prices.

The results of using Parkinson’s estimator to measure volatility are presented in Panel B of Table 2. Possible changes in market volatility were accommodated by dividing the price range statistic for each stock by the standard deviation of daily market returns. However, again, for both the unadjusted and market- adjusted volatility measures and all before and after periods and subperiods considered, the Wilcoxon test statistics indicate that the hypothesis that the before and after stock volatility distributions are the same cannot be rejected at any reasonable level of significance.

A further test of the possible impact of options listing on stock volatility involved using market model regressions to obtain before and after residual variances for each stock. Market effects were thus removed as part of the estimation procedure. The results reported in Panel C of Table 2 parallel those obtained for the standard deviation of daily returns. That is, while there appears to be some decline in the residual variance following options listing, the null hypothesis that the before and after distributions are the same cannot be rejected.

Panel C also reports beta coefficients before and after options listing. For thinly traded stocks, ordinary least squares estimates of beta are likely to be biased when daily data are used. Compared to most stocks comprising the TSE 300, optioned stocks in the Canadian markets are relatively heavily traded. However, in order to allow for the possibility of bias, beta is estimated using the well-known Scholes-Williams (1977) estimator. The resulting values appear, on average, to decline slightly following the listing of the options. Nonetheless, as above, the null hypothesis of before and after equality cannot be rejected.

CHANGES IN TRADING VOLUME

Schwert (1987) reports that growth in trading volume is positively related to an increase in stock return volatility. Any decrease in volatility following options listing should therefore be associated with a fall in trading volume. Skinner suggests that this might occur because of a diversion of trading away from the market for the stock to the market for its option. Alternatively, it may be that options enhance the liquidity of the market for the underlying stock and thereby generate additional interest and trading.

The impact of option listing on trading volume was examined by comparing the average daily numbers of shares traded before and after the listing date. In order to control for possible changes in market volume, the daily volume for each stock was divided by the contemporaneous daily volume for the market. Panel A of Table 3 reports the results of these calculations for the same intervals


CANADIAN EVIDENCE O N STOCKS AND OPTIONS LISTING

Table 3

Changes in Trading Volume (Number of Shares) of Sample of Thirtyseven Optioned Stocks

693

Panel A: Trading Volume Before and After Options Listinga

Statistic Before Afn Afn 1 Aftn 2 Before Afln A@ 1 A f n 2

Mean 68867 71745 69097 75090 0.01170 0.01183 0.01183 0.01188 Standard deviation 82934 57658 55763 70512 0.00967 0.00703 0.00766 0.00795 Wilcoxon test 0.913 0.445 1.592 1.033 0.641 0.370

Panel B: Cross-sectional OLS Regressions of the After-to-before Trading Volume Ratio on

Unadjusted Markd-adjusted

the Ratio of Trading Volume to Shares Outstanding Before Options Listing

Unadjusted Tradinp Volume

Market-adjusted Tradinc Valunu

~~ ~

Slope coefficient - 18.243 -54.948 t-statistic -0.251 - 1.022 R* 0.002 0.029

Notes. a After 1 and After 2 refer to the first and second subperiods after the listing of the option. The

Wilcoxon tests use the pre-listing period as the benchmark. That is, daily firm volume divided by contemporaneous daily market volume.

as are presented in Table 2. Here there appears to be a gradual increase in trading volume on both an unadjusted and market-adjusted basis. However, interpreting the Wilcoxon test statistics at conventional levels of significance does not allow rejection of the null hypothesis that the before and after trading volume distributions are the same. This, of course, does not imply that the above conjectures concerning the volume effect of options listing are incorrect. It may be that any diversion of trading activity from the stock to the option is offset by greater interest in the stock as a result of its increased liquidity.

The possibility of cross-sectional variations in the response of trading volume to options listing was also examined. Specifically, we considered whether more actively traded stocks differ from less actively traded stocks in their reaction to listing. To this end the ratio of trading volume after listing to trading volume before listing was regressed on the before listing ratio of trading volume to shares outstanding. This was done using measures of trading volume that were unadjusted as well as measures that were adjusted for possible changes in market volume. In the case of the latter each stock’s volume was divided by the market volume for the same period.

The results of these regressions are presented in Panel B of Table 3. While the slope coefficient is negative in both cases, in neither is it statistically significant. Hence, the null hypothesis that any change in trading volume is unrelated to a stock’s relative liquidity before listing cannot be rejected.


694 CHAMBERLAIN. CHEUNG AND KWAN

CHANGES I N THE BID-ASK SPREAD

In order to assess the liquidity effect of options listing directly, the average daily before and after bid-ask spreads were also e ~ a m i n e d . ~ The bid-ask spread as a measure of liquidity has been proposed by a number of authors beginning with Demsetz (1968), who characterized the spread as the price traders must pay to obtain a market maker’s marketability service.” Options may contribute to lower spreads through a positive impact on the trading volume of the stock and the economies of scale in market making that greater volume permits. In addition, if the return volatility is reduced, the risk to the market maker of holding inventory will be lower. Hence, a smaller risk premium will be required. Further, if options listing attracts informed traders away from the stock, l1 market makers will be less likely to be trading with investors better informed than themselves and thus be prepared to accept lower bid-ask spreads.

Bid-ask spreads tend to vary with the price of the stock. Therefore, in order to compare the spreads before and after options listing the daily spread for each stock was divided by the closing price of the stock on that day. In addition, possible changes in market risk were taken into account by dividing the normalized bid-ask spread by the standard deviation of daily market returns. The results of these comparisons, reported in Table 4, suggest that the bid-ask spread declines following options listing. Nonetheless, the null hypothesis that the before and after spread distributions are the same cannot be rejected.

The evidence thus far indicates that, on average, option listing has no significant impact on the return volatility, trading volume or bid-ask spread of the underlying stocks. However, cross-sectional variations may exist. That is, options listing may affect some stocks, but not others or the effects may vary in magnitude across stocks.

To investigate the possibility of cross-sectional effects the ratio of the

Table 4

Bid-ask Spread” for Sample of Thirty Optioned Stocks Before and After Options Listingb

Statistic Unadjusied Markct-adjustedc

Before After After I Afier 2 Before Afier Afier I A f t e r 2

Mean 0.01014 0.01007 0.01020 0.00991 1.6516 1.5201 1.5594 1.6790 Standard deviation 0.00293 0.00334 0.00372 0.00324 0.5599 0.6050 0.5838 0.8109 Wilcoxon test 0.010 -0.134 0.010 -1.224 -0.812 0.545

Notes: a Defined as (ask-bid)/close.

After 1 and After 2 refer to the first and second subperiods after the listing of the option. The Wilcoxon tests use the pre-listing period as the benchmark. That is, the (ask-bid)/close ratio divided by the contemporaneous standard deviation of daily market returns.


CANADIAN EVIDENCE O N STOCKS AND OPTIONS LISTING

Table 5

Cross-sectional OLS Regressions of the Ratio of the Bid-ask Spreada After Options Listing to the Bid-ask Spread Before Options Listing on the

Ratio of After-to-before Trading Volume and the Ratio of After-to-before Return Volatilityb

695

Independent Unadjusted Market-adjusted VariablesC Bid-ask Spread Bid-ask Spread

Volume Volatility R2 Volume Volatility R2 (i,iii) -0.1388 0.2018 0.440 -0.1919 -0.2867 0.349

(-4.025) (2.014) (-3.456) (-1.777)

(-2.654) (1.334) ( - 1.143) ( - 1.581)

(-4.288) (2.154) ( - 3.25 1) ( - 1.680)

(-2.895) (1.415) ( - 0.970) ( - 1 .689)

(-3.720) (0.898) (-3.182) (3.239)

(-3.276) (1.791) (- 1.281) (3.454)

(ii,iii) -0.1433 0.1543 0.289 -0.1034 -0.3066 0.105

(i,iv) -0.1465 0.2339 0.450 -0.1815 -0.2980 0.342

(ii,iv) -0.1524 0.1745 0.294 -0.0853 -0.3484 0.115

(i,.) -0.1372 0.1139 0.374 -0.1603 0.5611 0.476

(ii,v) -0.1691 0.2345 0.323 -0.0988 0.6759 0.321

Notes: a Defined as (bid-ask)/close.

I-statistics given in parentheses. Various measures of volume and volatility were combined in the regressions. The volume

measures considered were (i) daily trading volume (number of shares) and (ii) daily trading volume divided by contemporaneous daily market volume. The volatility measures considered were (iii) the square root of the residual variance of daily returns, (iv) the standard deviation of daily returns and (v) the square root of the market-adjusted variance of daily returns.

That is, the (bid-ask)/close ratio divided by the contemporaneous standard deviation of daily market returns.

normalized bid-ask spread before options listing to the normalized spread after listing was regressed on the ratio of before to after trading volume and the ratio of before to after return volatility. The regressions involved the thirty firms for which bid and ask data were available. Various measures of trading volume and return volatility were considered, including measures intended to take account of changes in market conditions.

Table 5 presents the results of these regressions. The relationship between the before-to-after bid-ask spread and the before-to-after trading volume is negative - that is, a smaller spread is associated with more volume - for all variable measurement combinations considered. Indeed, in eight of twelve cases, the relationship is significant at the one percent level. Thus, it seems that, cross-sectionally at least, lower bid-ask spreads are associated with greater trading volume around the time of options listing.



The relationship between the before-to-after bid-ask spread and the before- to-after volatility is less clear. Though eight out of twelve variable combinations produced positive signs - that is, smaller spreads associated with lower volatilities - in only two instances are the relationships significant at five percent. At the same time, it is of interest to note that these are the only instances in which both the bid-ask spread and the volatility measure are adjusted for possible market effects.

SUMMARY

The listing of options on Canadian stock exchanges appears to have had little impact on the price behaviour, trading volume or liquidity of the underlying stocks. While a casual interpretation of the evidence suggests that return volatility declines, trading volume increases and liquidity - as measured by the bid-ask spread - is enhanced, the results are not statistically significant. As such, they contrast with those recently reported by Skinner for optioned stocks in the United States. In addition, tests to determine whether the impact of listing on volume depends on the pre-listing liquidity of a stock failed to detect any statistically significant relationship. At the same time, the present study does find cross-sectional evidence of an inverse and statistically significant relationship between the before-to-after bid-ask spread and the before-to-after volume of trading. This suggests that in so far as listing does have an impact on trading volume, the bid-ask spread for the stock will be affected as well.

NOTES

1 See, for example, ‘Call of the Wild’, i h r o n s (May 3, 1976). 2 Among financial economists the question of whether the introduction of options affects the

underlying stock market is expressed in terms of their impact on the allocational erficiency of financial markets. If the market is complete, options are redundant assets and would not be introduced. When the stock market is incomplete, the introduction of options will provide payoff patterns not previously available and, in doing so, enhance market efficiency and affect price hehaviour and trading patterns. While a number of papers, beginning with Ross (1976). have focused on the ability of options to complete financial markets, little attention has been given to modelling the effects of non-redundant options. Among those papers that do (Dybvig and Ingersoll, 1982; Detemple and Selden, 1986; and Detemple and Jorion, 1988), Detemple and Selden find that there is an unambiguous change (a decrease) in stock return volatility when options are introduced.

3 A recent example of this concern is the effort by the Tokyo Stock Exchange to limit trading in Nikkei warrants. See ‘Tokyo Stock Exchange Pushes to Halt Further US Sales of Nikkei Warrants,’ Wall StrcefJoumal (April 26, 1990), pp. c 1 , c11.

4 Halpern and Turnbull (1985) examined Canadian call option prices for violations of boundary conditions and found that, unlike the results of similar US studies, the Canadian market was inefficient even after taking account of transactions costs. Chamberlain, Cheung and Kwan (1989), in a study of the index options and futures expiration date behaviour of TSE 300 stocks, were unable to reject the hypothesis that trading volume was unaffected by the expirations.

0 Basil Blackwell Ltd. I993

CANADIAN EVIDENCE ON STOCKS AND OPTIONS LISTING 697

In this their results differed from those of US studies, though, like the latter, Chamberlain et al. did find evidence of higher returns and return volatilities on expiration dates.

5 The one exception, Aquitaine Canada, was excluded because stock return data could not be obtained for the requisite four months on either side of the listing date.

6 The trading volume tests described below were also conducted using aggregated data for all markets on which each stock was listed. Depending on the stock, these included other Canadian markets, US markets and, in a few instances, overseas markets as well. However, the results obtained are not materially different from those obtained and presented for the Toronto Stock Exchange. This is as one would expect inasmuch as the Toronto market accounted for a substantial fraction of the trading in all of the stocks included in this study.

7 The stock price and volume data were obtained from Reuters Canada, the Globe and Mail newspaper and the TSE Monthly Review. The market return data were collected from the Toronfo Stock Exchange Daily Record, while the information on dividends was obtained from the Financial Post Dividend Record and the Financial Posf newspaper.

8 Numerous other comparisons were made, involving before and after subperiods of various lengths. In all instances the results are consistent with those reported; that is, the null hypothesis of identical distributions is never rejected.

9 The possibility of and scope for such a test is discussed by Skinner (1989). 10 Demsetz’s work and the subsequent literature is summarized in Schwartz (1988), Chapter 12. 1 1 This is because options on common stock are similar to highly levered positions in the stock,

which will appeal most to informed investors. See Skinner (1989), p. 75.

REFERENCES

Chamberlain, T.W., C.S. Cheung and C.C.Y. Kwan (1989), ‘Expiration Day Effects of Index Futures and Options: Some Canadian Evidence’, Financial Analysts Journal, Vol. 45 (September-October 1989), pp. 67-71.

Chicago Board Options Exchange (1975), Analysis of Volume and Price P a f f m s zn Sfocks undnlytng CBOE Optionsfrom December 30, 1974 to April 30, 1975 (Chicago, July 1975).

(1976), Analysis of Volume and Price Pafterns in Stocks Underlying CBOE Optionsfrom December 31, 1975 to January 16, 1976 (Chicago, February 1976).

Demsetz, H. (1968), ‘The Cost of Transacting’, Quarfer~Journal @Economics, Vol. 82 (February 1968), pp. 33-53.

Detemple, J . and P. Jorion (1989), ‘Option Listing and Stock Returns’ (Columbia University, First Boston Working Paper No. 89-13, 1989). - and L. Selden (1986), ‘A General Equilibrium Analysis of Option and Stock Market

Interactions’ (Columbia University, Mimeo, 1986). Dybvig, P. and J. IngersoU (1982), ‘Mean-variance Theory in Complete Markets’, Journal OfBusincss,

Vol. 55 (April 1982). pp. 233-251. Grammatikos, T . and A. Saunders (1986), ‘Futures Price Variability: A Test of Maturity and

Volume Effects’, Journal of Business, Vol. 59 (April 1986), pp. 319-330. Halpern, P J . and S.M. Turnbull (1985), ‘Empirical Tests of Boundary Conditions for Toronto

Stock Exchange Options’, Journal of Finance, Vol. 40 (‘June 1985), pp. 481-500. Hayes, S.L., 111 and M.E. Tennenbaum (1979), ‘The Impact of Listed Options on the Underlying

Shares’, Financial Managemmf, Vol. 8 (Winter 1979), pp. 72-76. Klemkosky, R.C. and T.S. Maness (1980), ‘The Impact of Options on the Underlying Securities’,

Journal of Porlfolio Mamgemmt, Vol. 6 (Winter 19801, pp. 12-18. Parkinson, M. (1980), ‘The Extreme Value Method for Estimating the Variance of the Rate of

Return’, Journal of Business, Vol. 53 (January 1980), pp. 61-65. Robert R. Nathan Associates (1974), Reuim @Initial Trading Expm’mce at the Chicago Board Options

Exchange (Chicago, December 1974). Ross, S. (1976), ‘Options and Efficiency’, Quarterly Journal ofEconomics, Vol. 90 (February 1976),

pp. 75-89. Scholes, M. and J. Williams (1977), ‘Estimating Betas from Nonsynchronous Data’, Journal of

Financial Economics, Vol. 5 (December 1977), pp. 309-327.



Schwartz, R.A. (1988), Equity Markets (New York: Harper & Row, 1988). Schwert, G. W. (1989), ‘Why Does Stock Market Volatility Change Over Time?’Journal ofFinance,

Vol. 44 (December 1989), pp. 1115-1153. Skinner, D.J. (1989), ‘Options Markets and Stock Return Volatility,’Jouml ofFinancia1 Economics,

V O ~ . 23 (1989), pp. 61-78. Trennepohl, G.L. and W.P. Dukes (1979), ‘CBOE Options and Stock Volatility’, Review OfBurineSs

and Economic Research, Vol. 14 (Spring 1979), pp. 49-60. Whiteside, M.M., W.P. Dukes and P.M. Dunne (1983), ‘Short Term Impact of Option Trading

on Underlying Securities’,Journal ofFinanciaf Research, Vol. 6 (Winter 1983), pp. 313-321.


OPTIONS LISTING, MARKET LIQUIDITY AND STOCK BEHAVIOUR: SOME CANADIAN EVIDENCE

Documents