The Market Response to Stock Option Re-pricing: An ... - CiteSeerX

0 downloads 0 Views 72KB Size Report
We investigate the market response to the firm stock option re-pricing event. By ... Incentive realignment theory (Saly 1994) suggests that underwater options are.
The Market Response to Stock Option Re-pricing: An Empirical Study Salil K. Sarkar Department of Finance and Real Estate College of Business Administration The University of Texas at Arlington Box 19449 Arlington, Texas 76019 E-mail: [email protected] Phone: 817 272 3836 Fax: 817 272 2252

Yen-Ling Chang Department of Finance and Real Estate College of Business Administration The University of Texas at Arlington Box 19449 Arlington, Texas 76019 E-mail: [email protected] Phone: 817 272 3083 Fax: 817 272 2252

1

The Market Response to Stock Option Re-pricing: An Empirical Study Abstract We investigate the market response to the firm stock option re-pricing event. By employing event study, we observe a simultaneous, significant down-U shaped curve and up-U shaped curve during the event window for cumulative abnormal trading volume and cumulative abnormal return, respectively. We hypothesize a highly diverse market opinion on the timing of re-pricing. Our result shows that market is aware of the deep drop in the stock price but not quite sure, if and when, the firm will re-price the executive stock option. We examine the relationship among volume, return and bid-ask spread and find little evidence of any dynamic relationship. However the contemporaneous relationship among the three is significant and strong.

2

The Market Response to Stock Option Re-pricing: An Empirical Study 1. Introduction Stock option re-pricing is one of the mechanisms for firms to restore the value of stock options given to managers if the stock price has dropped severely. However this issue is controversial since several competing but not mutually exclusive theories have suggested. Incentive realignment theory (Saly 1994) suggests that underwater options are no longer able to provide enough incentive for mangers to act in the best interest of the firm, therefore firms need to reset the strike price lower than the original exercise price. For those high-technology, young growth firms, option grant is a way to motivate the management team, once the options are deep out of money, it loses the purpose. Carter and Lynch (2001) propose the theory of retaining key employees and suggest option re-pricing can help the retention of key employees such as CEO. Managerial entrenchment theory explains that option re-pricing is the result of not being able to fire poor-performing mangers, and hence, managerial rent extracted from shareholders. Saly (1994) analyzes the relationship between stock option re-pricing and down market. She found that the two parties (employer and employee) would recognize the damage that down market causes to the stock option and renegotiate through the re-pricing mechanism. Later, Carter and Lynch (2001), Chen (2004) and Carter and Lynch (2004) analyze the determinants of options re-pricing and compare the re-pricing firms to non-re-pricing firms to investigate management turnover after re-pricing. Chidambaran and Prabhala (2003) even relate the re-pricing to internal governance mechanism. We find that few articles focus on the reaction of market (shareholders) to the re-pricing event. Callaghan, Saly and Subramaniam (2004, CSS hereafter) investigate the stock option re-pricing events for American market and document a U-shaped cumulative abnormal return (CAR) 50 days prior to the re-pricing event. Grein, Hand and Klassen (2005) use Canadian firms to demonstrate a re-pricing phenomenon. Their findings suggest that markets respond positively to the re-pricing announcements since they assist in retaining key employees. However, CSS also document that firms will re-price before favorable news and after unfavorable news, suggesting that the re-pricing event in different timing contains different information content. To investigate even further, we are not just looking at the

3

CAR around the event but also examine the changes in market depth and width. This might give us an even better and clear idea of how market makers and investors actually respond to the re-pricing event. Our analysis focuses on trading activities during the re-pricing event date. We look at the bid-ask spread, abnormal trading volume and abnormal return to examine the reaction of market makers and investors to the event. According to the market microstructure literature, market makers are trading with different types of investors. They will carefully examine the size of transaction in order to discern whether their counterparty is informed trader or not. By widening the bid-ask spread, they are able to transfer the loss to the liquidity traders. Admati and Pfleiderer (1988) proposed a stock pricing model consisting of informed traders, discretionary liquidity traders and nondiscretionary liquidity traders. Their idea suggests that informed traders would like to trade when they could be covered by those liquidity traders. In empirical studies, several papers document that during the regular corporate announcements such as earnings announcement, the bid-ask spread is larger and trading volume is thin before the announcement. Those suggest that market makers do suspect informed trading and most of the investors realize they have to share a part of the loss with market makers, therefore they choose not to trade in the market at that time. On the other hand, Kim and Verricchia (1994) derive a model suggesting that trading volume might not inversely relate to information asymmetry. Chae(2005) separates corporate announcements to two types of announcements: scheduled and unscheduled events. His findings suggest that abnormal trading volume is higher for unscheduled events and lower for scheduled events. In our paper, we find that abnormal trading volume is higher 10 days prior to re-pricing event and back to normal level afterwards. Furthermore, the bid-ask spread is slightly wider during the re-pricing event window. Those empirical findings raise the issue whether market makers suspect informed traders but the discretionary and non-discretionary liquidity traders are not able to identify the information content beforehand since the re-pricing event is not a regular and expected corporate announcement. The other result we notice is that trading volume is decreasing along with the narrow range of bid-ask spread after re-pricing events. The explanation we give is that during the downturn of stock price, especially for those high-tech firms, investors start to wonder when the firms are going to re-price the stock options since they want to retain key employees. Chidambaran and Prabhala(2003) document that firms with negative historical return are more likely to re-price their employee stock options.

4

Garfinkel and Sokobin (2006) analyze the post-earnings announcement trading activity and propose that unexplained (abnormal) trading volume could be an indicator of opinion divergence. Hence, investors speculate that firms are going to re-price the underwater options. However, there is no consensus about the timing of the re-pricing announcement. It is possible that higher level of divergence of opinion results in higher trading volume while, market makers suspecting informed traders might camouflage during this period, choose to widen the bid-ask spread. 2. Data and methodology 2.1 Data description We obtain the re-pricing events for each firm from executive compensation database. Because some of the re-pricing events are recorded repeatedly over time, we cross check with proxy statements for those re-pricing events. In total, we collect 311 re-pricing events for 224 firms from 1992 to 2005, excluding events where the exercise price is higher than stock price, spin-off, re-pricing date is not recorded, the time for the second consecutive re-pricing event is less than 6 months apart from the first re-pricing event for a firm.1 Since the goal of this paper is to examine investors’ response to the re-pricing events, we eliminate the re-pricing events if firms’ earnings announcement occurs 5 days prior to and after the re-pricing event. After the elimination, our sample size reduces to 66 events. In order to investigate the extent of information asymmetry due to stock re-price, we extract and calculate daily return, daily bid-ask spread and trading volume from CRSP for a re-pricing event within a time window of 505 days. We obtain accounting data such as total assets, sales, market value and other financial ratio for each firm from Compustat. Following Carter and Lynch (2001), we also match the re-pricing firms with a group of non-repricing firms where the executive stock options are out of money, based on the 4-digit SIC, firm size, and past performance (one year or two year annual returns). Carter and Lynch (2001) found that most of the re-pricers are computer related firms. Our sample firms also confirm this fact. One-third of the firms are technology related. Table 1 shows the statistical description of firm characteristics. Again, it suggests re-pricing firms are more likely to be small firms with low profitability and poor performance for the prior two years.

1

We also examine the firms which only have one repricing event. The result is similar to the sample.of multiple repricing event within a firm.

5

2.2 Descriptive statistics: Previous literature document that if market makers suspect potential informed traders in the market, they will increase the spread in order to reduce the loss from transaction. Therefore, in our paper, we use bid-ask spread as the indicator of information asymmetry among informed and uninformed traders. We examine the market response to the re-pricing event. We hypothesize that once the event is disclosed, the market worries less about the information asymmetry and market makers will lower the bid-ask spread. Other than spread, financial literature also treats abnormal trading volume as an indicator of asymmetric information. Therefore, we investigate the trading activities during the re-pricing announcement. This follows from our second hypothesis that trading volume will decrease if investors have homogeneous expectation based on the announcement. Chae (2005) investigates the market response to expected and unexpected corporate announcements. They find that trading volume decreases on average before the earnings announcements. However, it is not necessary for trading volume to decrease when corporate announcements are not expected. Furthermore, we postulate that investors are expecting the firm to announce but they do not know the timing. Before the actual announcement, market opinion is diverse and result in increased trading volume. Our finding supports that investors expect firms to re-price the stock options granted to employees but are not clear about the timing. 2.3 Bid-Ask spread, abnormal return, abnormal trading volumes: 2.3.1 Bid-Ask spread Bid-Ask spread is defined as follows

Spread = (ask − bid ) /[(ask + bid ) / 2] We collect the bid and ask price for each event for 30 day prior to and after the re-pricing date. The spread during this period enables us to look at the fluctuations of spread. Figure 1 shows the average pattern spans for 61 days centered on re-pricing event. We see a slight increase of spread after re-pricing date from day 0 to day 3. 2.3.2 Abnormal return: We employ the market model event study to compute the abnormal return for each

6

event. Ri ,t = α + β * Rm ,t + ε i ,t ARi ,t = Ri ,t − E ( Ri ,t ) CARi ,t =

t

∑ AR

t = −31

CAARt =

i ,t

1 n ∑ CARi,t n i =1

where Ri ,t : actual return for individual firm at time t Rm ,t : market return ARi ,t : abnormal return for ith firm at time t CARi ,t : cumulative abnormal return for ith firm at time t CAARt : cumulative average abnormal return at time t

We adopt market model and select a time period of 180 days as estimation window, ending on the 31st day prior to event date2. The event window is from 30 days prior to event date to 30 days after. Table 2 indicates that the decline of stock return is not statistically significant however the stock return increases by 2.5% significantly on the first day after re-pricing announcement. Figure 2 shows a dramatic u-shaped curve that the cumulative abnormal return is turning to positive 15 days after event announcement, suggesting that market responds positively to the events. 2.3.3 Trading volume market model: On the other hand, we analyze the trading volume prior and after the event using market volume index as benchmark. Harris (1986) finds trading volume to be positive skewed around event window. Ajinkya and Jain (1989) document that inclusion of market trading volume and autocorrelation adjustments are advantageous to the power of test. It’s appropriate for us to employ the following model and adopt EGLS test statistic to solve for the autocorrelation problem. 2

We choose 180 days as estimation window since in our sample we combine two re-pricing events if their time gap is less than 6 months.

7

Voli ,t = α + β * Volm ,t + ε i ,t AVi ,t = Voli ,t − E (Voli ,t ) CAVi ,t =

t

∑ AV

t = −30

CAAVt =

i ,t

1 n ∑ AVi,t n i =1

Where Voli ,t : relative trading volume for ith firm at period t; Vol=log(trading volume/share outstanding) AVi ,t : abnormal relative trading volume for ith firm at period t CAVi ,t : cumulative abnormal relative trading volume for ith firm at period t CAAVt : cumulative average abnormal relative trading volume at period t Volm ,t : the market volume index at time period.

The result is shown in figure and table. The positively increasing cumulative abnormal volume indicates active trading activities before and five days after the event date. 2.4 The relationship among spread, return and volume 2.4.1 The dynamic relation among spread, return and volume To know whether the trading volumes carry any information content and sequentially affect spread and return, we adopt Vector autoregressive model to examine the dynamic causal relationship among those three variables. We are testing whether the abnormal trading volume and higher spread before re-pricing date is triggered by investors’ expectation of firms re-pricing decision. To analyze the casual relation between these three variables, we employ the Vector autoregressive (VAR) model. AVt = α1,t + ∑ β1, p AVt − p + ∑ γ 1, p ARt − p + ∑ η1, p Spread t − p +ε 1

ARt = α 2,t + ∑ β 2, p ARt − p + ∑ γ 2, p AVt − p + ∑η 2, p Spread t − p + ε 2

Spread t = α 3 + ∑ β 3, p Spread t − p + ∑ γ 3, p AVt − p + ∑η 3, p ARt − p + ε 3

8

where t = 1…..30. First of all, we check for unit root of each variable, the results (not displayed) show that they are all stationary time series. Next we run the VAR model and check for the stability of the model and Granger causality using Granger test. Using the AIC and SBC criteria, we choose the lags of 3 in each equation. This result is consistent with Ajinkya and Jain (1989) that trading volume is positive first order serially auto-correlated. The Granger causality test results in table 4 indicate that there is a negative dynamic relationship between spread and abnormal return. We also conclude that the abnormal trading volume is Granger non-causal to abnormal trading return and spread. To check for any existence of co-integration, we employ Johansan’s procedure to check whether there is any co-integration relation among abnormal return, abnormal trading volume and spread. The trace test and lambda test statistics show no indication of co-integration. The non-causal relation between trading volume and return might indicate that stock return does not incorporate the extra information from abnormal trading activity or it might be the case that the increasing trading volume does not contain any information content. 2.4.2 Contemporaneous relationship among spreads, return and trading volumes AVt = α 1,t + β1 AVt −1 + β 2 ARt + β 3 Spread t + β11 Dum + ε 1 ARt = α 2,t + β 4 ARt −1 + β 5 Spread + β 6 AVt + β10 Dum + ε 2 Spread t = α 3 + β 7 Spread t −1 + β 8 AVt + β 9 AR + β12 Dum + ε 3

The variables are defined in the earlier VAR model; here we include a dummy variable for post-event days. So Dum is 1 if it is post-event day, and 0 otherwise. Because of the existence of endogeniety in our model, we employ instrumental variable method. The results in table 5 show that there is negative contemporaneous relation between abnormal volume and return. We neither find any contemporaneous relation between volume and spread nor between return and spread. The regression result suggests that spread is the exogenous variable that positively affects abnormal return and volume. The positive relation between spread and abnormal volume discloses that trading volume increases even as market makers widen the spread, suggesting that opinion divergence is outweighing the informed trading.

9

3. Conclusion Our findings on the market reaction to firm re-pricing event suggest there is no dynamic but contemporaneous relationship between abnormal trading volume and abnormal return, The two variables endogenously interact with each other negatively, suggesting that market opinion diverge when the executive stock options are underwater. Consistent with CSS, we also find that re-pricing firm’s stock return turn positive after re-pricing. Furthermore, the trading volume is back to normal level afterwards, suggesting markets gradually incorporate the re-pricing information. Our finding on the positive relation between trading volume and spread still needs further investigation. Therefore the future work should focus on decomposition of spread and volume during the event window and examine the extent of information asymmetry.

10

Table 1: descriptive statistics of re-pricers and non-repricers Descriptive statistics are shown for 65 re-pricers from the period of 1992 to 2005. We compute sales, assets and market values, reported in millions. Return on Asset (ROA) is pretax/assets; profit margin is pretax income/sales; EPS is primary earnings per share, excluding extraordinary items; debt to assets is total liabilities/total assets; and market-to-book is market value/net book value. We also compute annual and two-year cumulative return ending at re-pricing date.

Re-pricers mean

median

One year return ending at re-pricing date

-0.43

-0.49

Two year return ending at re-pricing date

-0.29

-0.41

Asset

516.8707

344.946

Sales

585.9347

317.127

EPS

-0.117

0.45

ROA

0.006787

0.059678

Profit margin

-0.02254

0.050227

market value

779.4023

374.646

market to book equity

10.87513

2.061737

debt ratio

0.44876

0.400887

numbers of firms

65

11

Figure 1: average bid-ask spreads

figure 1: spread plot

0.03

0.025

average spread

0.02

0.015

spread

0.01

0.005

24

26

28

30

24

26

28

30

20

22

20

22

16

18

16

18

12

14

12

14

8

10

10

4

6

0

2

-4

-2

-8

-6

-1 2

-1 0

-1 4

-1 6

-1 8

-2 0

-2 2

-2 4

-2 6

-2 8

-3 0

0

time

Figure 2: Cumulative average abnormal return

cumulative average abnormal return

0.08 0.06 0.04 0.02

-0.04 -0.06 -0.08 -0.1 -0.12 day

12

8

4

6

2

0

-4

-2

-6

-8

-1 0

-1 2

-1 4

-1 6

-1 8

-2 0

-2 2

-2 4

-2 8

-2 6

-3 0

caar

0 -0.02

CAAR

Figure 3: Cumulative average abnormal trading volumes plot of cumultive average abnoral trading volme

500.00%

400.00%

300.00%

CAAV

200.00% caav 100.00%

-100.00%

-200.00% day

13

30

28

26

24

22

20

18

16

14

12

8

10

6

4

2

0

-2

-4

-6

-8

-1 0

-1 2

-1 4

-1 6

-1 8

-2 0

-2 2

-2 4

-2 6

-2 8

-3 0

0.00%

Table 2: Abnormal return around the re-pricing date for re-pricing firms Here we use different model to estimate the abnormal return: market model, mean-adjusted model, the results are quite similar. The table below shows the GELS estimates and within parenthesis is p-value. *, ** and *** indicate significance level of 10%, 5% and 1% respectively. abnormal return Equally-weighted index Days

market modelmean-adjusted market model

(-30, -2) -6.05%***

(-1,0)

( 0,1)

(0, 10)

(11,30)

Valued-weighted mean-adjusted

-8.93%

-7.98%***

-8.93%***

(0.037)

(0.0086)

(0.013)

(0.0086)

-0.99%

-0.34%

-0.72%

-0.34%

(0.1961)

(0.4179)

(0.291)

(0.4179)

2.51%***

3.11%***

2.59%***

3.11%***

(0.0035)

(0.0009)

(0.003)

(0.0009)

6.49%***

7.93%***

6.95%***

7.93%***

0.0002

(0.0001)

(0.0001)

(0.0001)

6.31%***

8.95%***

7.6%***

8.95%***

(0.008)

(0.001)

(0.002)

(0.001)

14

Table 3: Abnormal trading volume around re-pricing date for re-pricing firms Here we use different model to estimate the abnormal trading volumes: market model, mean-adjusted model, the results are quite similar. The table below shows the GELS estimates and within parenthesis is p-value. *, ** and *** indicate significance level of 10%, 5% and 1% respectively. Abnormal trading volume Equally-weighted index window

Valued-weighted

market model

mean-adjusted

market model

mean-adjusted

(-30, -2)

404.47%**

446.01%**

176.16%

446.01%**

P-value

(0.096)

(0.088)

0.169

(0.088)

50.51%***

52.12%***

29.4%**

52.12%***

(0.049)

(0.047)

(0.094)

(0.047)

35.24%**

36.9%**

13.17%

36.9%**

(0.069)

(0.062)

(0.158)

(0.062)

29.12%

49.86%

-87.56%

49.86%

(0.426)

(0.377)

(0.295)

(0.377)

-272.57%

-227.63%

-453.57%

-227.63%

(0.11)

(0.158)

(0.018)

(0.158)

(-1,0)

( 0,1)

(0, 10)

(11,30)

15

Table 4 Test of causal relationship among stock return, trading volume and bid-ask spreads AVt  α1  A11(L) AR  = α  +A (L)  t   x  21 Spread     t  αe  A31(L)

 A22(L) A23(L)  A32(L) A33(L) 

A12(L) A13(L)

AVt−1  u1t  AR     t−1  + u2t  , Spread    t −1 u3t 

where

3

Ai , j ( L ) = ∑ a ij ( s ) Ls −1 , for i , j = 1, 2,3 s =1

Granger causality test Hypothesis Equation for volume Equation for

.C . H o : A12 ( L) = 0 Return G  → Volumes H o : A13 ( L) = 0 G .C . → Volumes Spread  G .C . H o : A21 ( L) = 0 Volume   → return

H o : A23 ( L) = 0 Return Equation for Spread

Comment

→ return Spread  G .C .

H o : A31 ( L) = 0 Volume  → Spread H o : A32 ( L) = 0 G .C . → Spread Return  G .C .

16

χ 2 Statistics

df

0. 62

3

3.78

3

1.61

3

6.78*

3

2.15

3

5.24

3

Table 5: contemporaneous relation among abnormal return, trading volume and spread AVt = α1,t + β1 AVt −1 + β 2 ARt + β 3 Spread t + β11 Dum + ε 1 ARt = α 2,t + β 4 ARt −1 + β 5 Spread + β 6 AVt + β10 Dum + ε 2 Spread t = α 3 + β 7 Spread t −1 + β 8 AVt + β 9 AR + β12 Dum + ε 3

Panel A:

Dependent variable: Abnormal trading volume

β1

Coefficient

-11.5264**

0.140535

-87.872***

493.4583**

1.584647**

t-statistic

-1.82

0.28

-2.02

1.8

1.98

Panel B: Coefficient

-0.25077***

t-statistic

-2.83

t-statistic

β5

β4

-0.16759 10.91914*** -0.51

β6

β 10

-0.0303**

0.025126***

2.81

-1.86

2.7

β8

β9

Dependent variable: Spread Intercept

Coefficient

β11

Dependent variable: Abnormal return Intercept

Panel C:

β2

β3

Intercept

β7

0.044346*** -0.77725 2.26

β12

-0.0021

0.924227

-0.01239

-0.25

1.6

-1.63

-0.99

17

References Ajinkya Bipin and Prem C. Jain, 1989, The behavior of daily stock market trading volume, Journal of Accounting and Economics, Vol. 11, 331-359. Admati, A.R. and P. Pfleiderer, 1988, A theory of intraday patterns: Volume and price variability, Review of Financial Studies, Vol. 1, 3-40. Brockman, Paul and Dennis Y. Chung, 2001, Managerial timing and corporate liquidity: evidence from actual share repurchases, Journal of Financial Economics, Vol.61, 417-448. Callaghan, Sandra, P. Jane Saly and Chandra Subramanian, 2004, The timing of option repricing, Journal of Finance, Vol. 59, No. 4, 1651-1675. Carter, Mary Ellen and Luann J. Lynch, 2001, An examination of executive stock option repricing, Journal of Financial Economics, Vol. 61, 207-225. Carter, Mary Ellen and Luann J. Lynch, 2004, The effect of stock option repricing on employee turnover, Journal of Accounting and Economics, Vol.37, 91-112. Chae. J., 2005, Trading volume, information asymmetry, and timing information, Journal of Finance, Vol. 60, No. 1, 413-442. Chidambaran, N. K. and Nagpurnanand R. Prabhala, 2003, Executive stock option repricing, internal governance mechanisms, and management turnover, Journal of Financial Economics, Vol. 69, 153-189. Chen, Mark A., 2004, Executive option repricing, incentives and retention, Journal of Finance, Vol. 59, No.3, 1167-1199. Coller, Maribeth and Teri Lombardi Yohn, 1997, Management Forecasts and information asymmetry: An examination of Bid-Ask spreads, Journal of Accounting Research, Vol. 35, No. 2, 181-191. Garfinkel Jon A. and Jonathan Sokobin, 2006, Volume, Opinion divergence and Returns: A study of post-earnings announcement drift, Journal of Accounting Research, Vol. 44, 18

No. 1, 85-112. Grein, Barbara M, John R. M. Hand and Kenneth J. Klassen, 2005, Stock price reactions to the repricing of employee stock options, Contemporary Accounting Research, Vol. 22, No. 4, 791-828. Harris, L., 1986, Cross-sectional tests of the mixture of distrubution hypothesis, Journal of Financial and Quantitative Analysis, Vol. 21, 39-46. Howe, John and Ji-Chai Lin, 1992, Dividend policy and the bid-ask spread: an empirical analysis, Journal of Financial Research, Vol. 1, 1-10. Kim, Oliver and Robert E. Verrecchia, 1994, Market liquidity and volume around earnings announcements, Journal of Accounting and Economics, Vol. 17, 41-67. Kanagaretnam, Kiridaran, Gerald Lobo and Dennis Whalen, 2005, Relationship between analyst forecast properties and equity bid-ask spreads and depth around quarterly earnings announcements, Journal of Business Finance and Accounting, Vol. 32, 1773-1799. Lee, Bong-Soo and Oliver M. Rui, 2002, The dynamic relationship between stock returns and trading volume: Domestic and cross-country evidence, Journal of Banking and Finance, Vol. 26, 51-78. Saly, P. Jane, 1994, Repricing Executive stock options in a down market, Journal of Accounting and Economics, Vol. 18, 325-356. Venkatesh, P. C. and R. Chiang, 1986, Information asymmetry and the dealer’s bid-ask spread: A case study of earnings and dividend announcements, Journal of Finance, Vol. 41, No. 5, 1089-1102.

19