1824 The University of Manchester Alliance Manchester Business School Methodologies: Survey analysis Directly ask managers whether...

Transcribed Image Text:

1824 The University of Manchester Alliance Manchester Business School Methodologies: Survey analysis Directly ask managers whether M&A creates value Strength: -Might yield insights unknown by the stock market p Z (Managers are very familiar with their own transactions) Weaknesses: -Might not be focused on economic value -Memories of past results can be hazy and biased -Typically very low participation rate insight. managers will think on their own purpose rather than the shareholders' clearly they are not best judge. If managers are saying that only the half of deal succeed-risky strategy The University of Manchester Alliance Manchester Business School Step 1: Identify the event A common procedure is to pick the date of the first announcement on Factiva, SEC Edgar, or SDC Platinum/Thomon One.com as the event date But some uncertainty about the event date is sometimes unavoidable (e.g. announcements made after stock market closure) procedurepick announcement event date Alliance Manchester Business School Step 2: Calculate normal stock returns We need a COUNTER FACTUAL and we need to compare the CF and Normal returns with the actual return to get abnormal/excess return To estimate the normal return Rit that one would expect to occur for firm j jt over the event day t, the researcher needs to use a benchmark model over an estimation window Should not include announcement As it is a benchmark. Unbiased estimate what happen to return of this company, so should be expected to carry on the return of the company if the announcement does not take place. The abnormal return rit is then equal to the actual return Rt minus the expected return jt Alliance Manchester Business School Step 2: Calculate normal stock returns Specify an estimation window: -Period over which you estimate the normal returns, using one of the three benchmark models specified below -Normally before event period, sometimes after -Generally, event period is excluded from estimation period. -For example, from trading day 240 to trading day 41 COUNTERFACTUAL Choose a benchmark model: - -Mean adjusted return method Part -Market adjusted return method rumours - -Market model method day generate normal Time profile return Alliance Manchester Business School Step 2: Calculate normal stock returns Mean adjusted return method Normal return = average daily stock return of firm over estimation period Past rumours -41 jt=jt &T period 240 Rjt = Announcement 200 day generate boom conditions B - Stock Market adjusted return method - Normal return = normal return 13% make 19 adjustment for market moveme return on the market index Rit=Rmt Alliance Manchester Business School Step 2: Calculate normal stock returns Allow for market movement & Risk. Market model method - Normal return is obtained by estimating the market model regression Rjt = ; + ; Rmt + Ejt = Constant over the estimation period 'measure of different risk 8=1 uss risky stock. move exactly with the market far greater - This regression produces estimates of a; and ;; we call these ; and The predicted return for a firm j for a day t in the event period is the return given by the market model on that day using these estimates, i.e. Over the announcement period. jt=j+ Rmt it allows of to What Normal announcement calculate returns would be at the given movement in the market & changy Most widely used method in risk. The University of Manchester Alliance Manchester Business School Step 3: Calculate and analyze abnormal stock returns Once we have calculated the normal return, we can easily obtain the abnormal return as follows: - rjt = Rjt jt Question: How should we interpret r.,? What does it represent? rit The performance of the market judgement of the merger The market is going to make an assessment of whether the dealing is good or bad and that is being captured by abnormal return. This is market immediate judgement. MANCHESTER 1824 The University of Manchester Alliance Manchester Business School . Step 3: Calculate and analyze abnormal stock returns For each day t in the event period, the residuals are averaged across firms to produce the average residual for that day (as individual firm estimates are noisy): ; rjt ART = N Finally, we sum the average residuals for each day over the entire event window, to obtain the average total effect of the event across all firms: 40 CAR = ARt Past rumours t=-40 The market overall assessment of this window Announcement f day generate normal MANCHESTER 1824 he University of Manchester nce Manchester Business School Step 3: Calculate and analyze abnormal stock returns Statistical tests answer the question whether abnormal returns are significantly different from zero (at a given confidence level). The null hypothesis is of the form: E(AR) = 0 for a given time t. Not affect shareholders' wealth That is, the expected abnormal return is zero. Or, over a number of consecutive days: E(CAR) = 0 The University of Manchester Alliance Manchester Business School Methodologies: Clinical research Analyze one transaction in great depth Strengths: -In-depth analysis of actual experience You concentrate only on one deal And you look all aspect of the deal All STAKEHOLDERS -Ideal for discovering new patterns and developing new hypotheses Get some insight which You can explore it in other deal Weaknesses: -Ill-suited to hypothesis testing -Results often not generalizable pattern In practical useless However one data point to observe all deal One data point cannot give you any result 1824 The University of Manchester Alliance Manchester Business School Methodologies: Accounting study Examine reported financial results (EPS, ROA,...) of acquirers, before and after acquisitions (often matched-sample approach) Strengths: COUNTERFACTUAL of benchmark, -Credibility (checked by auditors) -Based on measures that are often used by investors Profitability Price the deal. Return after the deal data things change measure Weaknesses: -Data before and after might be non-comparable -Backward-looking we need forward looking -Differences across companies and countries 1824 The University of Manchester Alliance Manchester Business School Methodologies: Event study Academic studies that empirically examine the stock price responses to financial decisions are referred to as event studies decisions stock price responses to financial In event studies, researchers collect the dates on which a sample of firms made similar announcements (e.g., initiating new dividends) announcement Next, they examine stock returns on the event's announcement date and the days immediately before and after the event, averaged across all firms in the sample how the stock price change after the announcement The fact that there is a sample of firms/events, allows one to perform a statistical analysis of the magnitude and determinants of stock price reactions Whether it is truthful, or random... Alliance Manchester Business School Event study Strengths: shareholder wealth impact -Direct measure of shareholder wealth impact -Forward-looking Weaknesses: - Relies on strong assumptions about stock markets (i.e. semi-strong form efficiency) share price efficiently reflect the -Vulnerable to confounding events situation You got other thing happen at the same time Firm engage in m&a One m&a in isolation is making another announcement about other thing. e.g. Issue in shares, occurring in the same time Other events which gonna create noise and confuse the original signal it is market judgement if it is good thing or bad thing. Alliance Manchester Business School Methodologies: Concluding remarks "Scientific approach" ie Formulating and testing a null hypothesis (H0) eg: "A merger announcement has no value impact" Key value: t-statistic of M&A performance metric. It is likely to be a random result Based on a large number of observation If absolute value of t-statistic > 2, results are significantly different from zero with 95% confidence Purely random result ( p-value <0.05) Accounting and event studies are most scientific", while surveys and case studies are mainly descriptive Since event study places the strongest emphasis on shareholder value and are most often used in M&A empirical research, we now discuss this method in more detail The University of Manchester Alliance Manchester Business School Methodologies: Concluding remarks "Scientific approach" ie Formulating and testing a null hypothesis (H) eg: "A merger announcement has no value impact" Key value: t-statistic of M&A performance metric. It is likely to be a random result Based on a large number of observation If absolute value of t-statistic > 2, results are significantly different from zero with 95% confidence Purely random result ( p-value <0.05) Accounting and event studies are most scientific, while surveys and case studies are mainly descriptive Since event study places the strongest emphasis on shareholder value and are most often used in M&A empirical research, we now discuss this method in more detail An event study generally consists of three steps: 1. Identify the event of interest and, in particular, the timing of the event -Sources: Thomson One.com, Factiva -Some uncertainty may be unavoidable 2. Calculate normal stock returns using a benchmark model Identify the return. Work out a benchmark. And then compare the benchmark with actual return to find an abnormal return 3. Calculate and analyze abnormal returns around the event date Counter factual What is difference between what you observe and what you expect? Alliance Manchester Business School Step 1: Identify the event In general, an event is an (observed) decision or announcement of the firm's management For example, stock splits, dividend initiations, stock repurchases, mergers and takeovers, earnings announcements, ... The event date is most often the announcement date. E.g. stock prices react when a merger is announced, rather than when the merger is completed The periods prior to or after the event may also be of interest to win a announce -To Fn} = rumours It is a window around the announcement days will capture hopefully not just the announcement, also any other information regarding specific deal. Q: Why do event studies typically consider days - 40 to 40 around the announcement date 0? What is the problem associated with considering stock returns over longer periods? Shareholders are every rights to response that rumour and they will do. 1824 The University of Manchester Alliance Manchester Business School Methodologies: Survey analysis Directly ask managers whether M&A creates value Strength: -Might yield insights unknown by the stock market p Z (Managers are very familiar with their own transactions) Weaknesses: -Might not be focused on economic value -Memories of past results can be hazy and biased -Typically very low participation rate insight. managers will think on their own purpose rather than the shareholders' clearly they are not best judge. If managers are saying that only the half of deal succeed-risky strategy The University of Manchester Alliance Manchester Business School Step 1: Identify the event A common procedure is to pick the date of the first announcement on Factiva, SEC Edgar, or SDC Platinum/Thomon One.com as the event date But some uncertainty about the event date is sometimes unavoidable (e.g. announcements made after stock market closure) procedurepick announcement event date Alliance Manchester Business School Step 2: Calculate normal stock returns We need a COUNTER FACTUAL and we need to compare the CF and Normal returns with the actual return to get abnormal/excess return To estimate the normal return Rit that one would expect to occur for firm j jt over the event day t, the researcher needs to use a benchmark model over an estimation window Should not include announcement As it is a benchmark. Unbiased estimate what happen to return of this company, so should be expected to carry on the return of the company if the announcement does not take place. The abnormal return rit is then equal to the actual return Rt minus the expected return jt Alliance Manchester Business School Step 2: Calculate normal stock returns Specify an estimation window: -Period over which you estimate the normal returns, using one of the three benchmark models specified below -Normally before event period, sometimes after -Generally, event period is excluded from estimation period. -For example, from trading day 240 to trading day 41 COUNTERFACTUAL Choose a benchmark model: - -Mean adjusted return method Part -Market adjusted return method rumours - -Market model method day generate normal Time profile return Alliance Manchester Business School Step 2: Calculate normal stock returns Mean adjusted return method Normal return = average daily stock return of firm over estimation period Past rumours -41 jt=jt &T period 240 Rjt = Announcement 200 day generate boom conditions B - Stock Market adjusted return method - Normal return = normal return 13% make 19 adjustment for market moveme return on the market index Rit=Rmt Alliance Manchester Business School Step 2: Calculate normal stock returns Allow for market movement & Risk. Market model method - Normal return is obtained by estimating the market model regression Rjt = ; + ; Rmt + Ejt = Constant over the estimation period 'measure of different risk 8=1 uss risky stock. move exactly with the market far greater - This regression produces estimates of a; and ;; we call these ; and The predicted return for a firm j for a day t in the event period is the return given by the market model on that day using these estimates, i.e. Over the announcement period. jt=j+ Rmt it allows of to What Normal announcement calculate returns would be at the given movement in the market & changy Most widely used method in risk. The University of Manchester Alliance Manchester Business School Step 3: Calculate and analyze abnormal stock returns Once we have calculated the normal return, we can easily obtain the abnormal return as follows: - rjt = Rjt jt Question: How should we interpret r.,? What does it represent? rit The performance of the market judgement of the merger The market is going to make an assessment of whether the dealing is good or bad and that is being captured by abnormal return. This is market immediate judgement. MANCHESTER 1824 The University of Manchester Alliance Manchester Business School . Step 3: Calculate and analyze abnormal stock returns For each day t in the event period, the residuals are averaged across firms to produce the average residual for that day (as individual firm estimates are noisy): ; rjt ART = N Finally, we sum the average residuals for each day over the entire event window, to obtain the average total effect of the event across all firms: 40 CAR = ARt Past rumours t=-40 The market overall assessment of this window Announcement f day generate normal MANCHESTER 1824 he University of Manchester nce Manchester Business School Step 3: Calculate and analyze abnormal stock returns Statistical tests answer the question whether abnormal returns are significantly different from zero (at a given confidence level). The null hypothesis is of the form: E(AR) = 0 for a given time t. Not affect shareholders' wealth That is, the expected abnormal return is zero. Or, over a number of consecutive days: E(CAR) = 0 The University of Manchester Alliance Manchester Business School Methodologies: Clinical research Analyze one transaction in great depth Strengths: -In-depth analysis of actual experience You concentrate only on one deal And you look all aspect of the deal All STAKEHOLDERS -Ideal for discovering new patterns and developing new hypotheses Get some insight which You can explore it in other deal Weaknesses: -Ill-suited to hypothesis testing -Results often not generalizable pattern In practical useless However one data point to observe all deal One data point cannot give you any result 1824 The University of Manchester Alliance Manchester Business School Methodologies: Accounting study Examine reported financial results (EPS, ROA,...) of acquirers, before and after acquisitions (often matched-sample approach) Strengths: COUNTERFACTUAL of benchmark, -Credibility (checked by auditors) -Based on measures that are often used by investors Profitability Price the deal. Return after the deal data things change measure Weaknesses: -Data before and after might be non-comparable -Backward-looking we need forward looking -Differences across companies and countries 1824 The University of Manchester Alliance Manchester Business School Methodologies: Event study Academic studies that empirically examine the stock price responses to financial decisions are referred to as event studies decisions stock price responses to financial In event studies, researchers collect the dates on which a sample of firms made similar announcements (e.g., initiating new dividends) announcement Next, they examine stock returns on the event's announcement date and the days immediately before and after the event, averaged across all firms in the sample how the stock price change after the announcement The fact that there is a sample of firms/events, allows one to perform a statistical analysis of the magnitude and determinants of stock price reactions Whether it is truthful, or random... Alliance Manchester Business School Event study Strengths: shareholder wealth impact -Direct measure of shareholder wealth impact -Forward-looking Weaknesses: - Relies on strong assumptions about stock markets (i.e. semi-strong form efficiency) share price efficiently reflect the -Vulnerable to confounding events situation You got other thing happen at the same time Firm engage in m&a One m&a in isolation is making another announcement about other thing. e.g. Issue in shares, occurring in the same time Other events which gonna create noise and confuse the original signal it is market judgement if it is good thing or bad thing. Alliance Manchester Business School Methodologies: Concluding remarks "Scientific approach" ie Formulating and testing a null hypothesis (H0) eg: "A merger announcement has no value impact" Key value: t-statistic of M&A performance metric. It is likely to be a random result Based on a large number of observation If absolute value of t-statistic > 2, results are significantly different from zero with 95% confidence Purely random result ( p-value <0.05) Accounting and event studies are most scientific", while surveys and case studies are mainly descriptive Since event study places the strongest emphasis on shareholder value and are most often used in M&A empirical research, we now discuss this method in more detail The University of Manchester Alliance Manchester Business School Methodologies: Concluding remarks "Scientific approach" ie Formulating and testing a null hypothesis (H) eg: "A merger announcement has no value impact" Key value: t-statistic of M&A performance metric. It is likely to be a random result Based on a large number of observation If absolute value of t-statistic > 2, results are significantly different from zero with 95% confidence Purely random result ( p-value <0.05) Accounting and event studies are most scientific, while surveys and case studies are mainly descriptive Since event study places the strongest emphasis on shareholder value and are most often used in M&A empirical research, we now discuss this method in more detail An event study generally consists of three steps: 1. Identify the event of interest and, in particular, the timing of the event -Sources: Thomson One.com, Factiva -Some uncertainty may be unavoidable 2. Calculate normal stock returns using a benchmark model Identify the return. Work out a benchmark. And then compare the benchmark with actual return to find an abnormal return 3. Calculate and analyze abnormal returns around the event date Counter factual What is difference between what you observe and what you expect? Alliance Manchester Business School Step 1: Identify the event In general, an event is an (observed) decision or announcement of the firm's management For example, stock splits, dividend initiations, stock repurchases, mergers and takeovers, earnings announcements, ... The event date is most often the announcement date. E.g. stock prices react when a merger is announced, rather than when the merger is completed The periods prior to or after the event may also be of interest to win a announce -To Fn} = rumours It is a window around the announcement days will capture hopefully not just the announcement, also any other information regarding specific deal. Q: Why do event studies typically consider days - 40 to 40 around the announcement date 0? What is the problem associated with considering stock returns over longer periods? Shareholders are every rights to response that rumour and they will do.