Initially Optimistic?
The Influence of Investor Optimism on the Long-Run
Underperformance of Initial Public Offerings
Michelle de Jong
Student number: 10209093
Specialization: Finance
Supervisor: Mr. M.A. Dijkstra
Abstract
This thesis researches the effect of optimistic investors on the underperformance of initial public offerings (IPOs) in the United States in the first three years after the offering. This thesis adds to the existing literature, because not only the level of underperformance is measured, the effect of investor optimism on underperformance is estimated as well. Excessive optimism among investors is expected to drive IPO underperformance, because overoptimistic investors overvalue firms that go public and their expectations can’t be realized once more information about the IPO firm becomes available. Three different optimism measures, the momentum factor, the return on the S&P 500 and consumer confidence, are used as proxies for optimism among investors. A sample of 327 IPOs in the period 2006-2009 in the United States is collected to measure performance over the three year period after offering, over a wide variety of model specifications. Overall IPO underperformance is found, when benchmarked against the S&P 500. No significant effect of momentum on IPO performance is found, so that overoptimistic investors are not the main cause of the long-run underperformance of IPOs.
Table of Contents
1. Introduction ... 4
2. Literature Review ... 5
3. Data and Methodology ... 10
4. Results ... 15
5. Conclusion ... 20
References ... 21
Appendix ... 23
Statement of Originality
This document is written by Student Michelle de Jong who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
1. Introduction
An anomaly in the field of initial public offerings (IPOs) that has been documented by Ritter (1991), Aggarwal and Rivoli (1990) and King and Banderet (2014) is the long-run stock price underperformance of IPO firms compared to firms that did not recently have an IPO. Ritter (1991) finds that U.S. IPO’s in the period 1975-1984 significantly underperform comparable firms, matched by market capitalization and industry, in the first three years after offering. One explanation for this underperformance is that overoptimistic investors, who overvalue firms that go public, are active in the stock market and pay a price for IPO stocks that is too high (Ritter, 1991). Once more information about the IPO firms becomes available, these investors’ valuation of the IPO firms converges towards the mean valuation, which makes them want to sell the stocks again. The higher demand that follows is what causes stock prices to decline. King and Banderet (2014) find that crisis period (2008-2009) IPOs outperform comparable firms, by size and industry, by 26% in the first three years after the offering. IPOs that took place in the non-crisis years (2003-2007 and 2010) underperform comparable firms by 22%.
The research question is: Did the influence of optimistic investors on
underperformance of IPOs change during the most recent financial crisis? In this thesis the level of overoptimistic investors is captured by three measures of optimism: the momentum factor, the return on the S&P 500 and consumer confidence. This is to find out whether overoptimistic investors are the cause of IPO underperformance in the period 2006-2009 in the United States. The expectation is that the three optimism measures have a negative effect on three year IPO performance, measured in buy-and-hold returns (BHAR) and cumulative abnormal returns (CAR) .It is also expected that IPO underperformance is lower in the period of financial crisis, because the overoptimistic investors, who cause underperformance, are not active in that period (King, 2014).
Regressions are estimated for three different measures of optimism and a variety of control varibles, with different model specifications. The results do not show any significant effect of momentum on the three year performance of IPOs. On average momentum does have a negative effect on IPO stock performance, but this is not statistically significant, nor robust over different specifications of the model. It can be concluded that overoptimistic investors are not the main factor that causes the long-run underperformance of IPOs.
The rest of this thesis is structured as follows. Section two will first explain what an initial public offering is and it offers possible explanations for IPO underperformance,
including overoptimistic investors. Section three describes the regressions that are performed and the data that is used for these regressions. Section four presents and discusses the results of the regressions. Section five is a conclusion and summary of this thesis.
2. Literature Review
An initial public offering is the event when a company offers its stock to a public securities exchange for the first time, also referred to as a company going public. It is the first time that a company can see what its value is on the market and for current shareholders of the private firm it is the moment they can sell their shares to the public and make a profit on their investment (King, 2014). The main reason for a firm to go public is because it has a need for financing and it can’t get enough funding through private equity anymore. When the
economy has high growth expectations, which can be seen by, for example, growth forecasts by the World bank and the IMF, the demand for funding by companies tends to rise (Lowry, 2003). Another reason for firms to go public is to acquire funding for debt repayment.
Previously private firms can acquire capital with bank loans, venture capital or public equity, which means selling stock to the public. A company chooses an IPO over other types of funding when this option has the lowest cost of capital (Lowry, 2003). The demand for capital of a firm does not only rise when the economy is growing, but also when the firm identifies new investment opportunities, for example, an innovation in a certain industry (Buttimer, 2005).
The underperformance of IPOs in the long run is an anomaly in the IPO market that was first documented by Ritter (1991). Long-run underperformance of an IPO means that the stock earns lower returns than a specified benchmark1. Different time periods are used to specify the long-run, varying between one to five years after issuance (King, 2014). Ritter (1991) researches the long-run performance of IPOs in the United States using a sample of 1526 companies that went public in the period between 1975 and 1984. He examined whether, in the first three years after the offering, these companies significantly underperformed comparable firms that did not recently have their IPO. To compare
performance each IPO firm is matched to another firm by market capitalization and industry
1
Ritter (1991) uses the following benchmarks: the NASDAQ index, the Amex-NYSE index, firms that are already listed matched to each IPO by market capitalization and industry and an index of the smallest size decile of the NYSE.
(Ritter, 1991). After the first three years these IPO firms were public, they had an average buy-and-hold abnormal return (BHAR) of 34.47% while a set of comparable firms in the same period had an average BHAR of 61.86%. This shows that IPO firms underperformed comparable firms that were already public (Ritter, 1991).
IPO underperformance is supported by Aggarwal and Rivoli (1990) who studied the performance of US IPOs in the first year after the issuance of stock took place in the period 1977-1987. Their results show that investors who purchase an IPO stock at the moment of offering on average do not gain a positive return. Investors who purchase an IPO stock in the aftermarket, which means after the first trading day, and hold it for one year are found to underperform the market by 13.73% on average (Aggarwal, 1990).
King and Banderet (2014) identify two possible explanations for IPO underperformance. The first explanation is that overoptimistic investors in the market overvalue the stock at the time of the offering. This causes underperformance because the high expectations can’t be realized and the stock price declines. The second explanation concerns high discretionary current accruals for firms that are planning an IPO. This is associated with poor three year stock performance, because discretionary accruals cause reported earnings to be higher than actual earnings and this causes the initial stock price to be too high.
Overoptimistic Investors and High IPO volume
Ritter (1991), King and Banderet (2014) and Cornelli, Goldreich and Ljungqvist (2006) discuss the theory of overoptimistic investors who overvalue issuing firms. Overoptimistic investors will have a higher valuation of the IPO firm than less optimistic investors. The demand for a certain security comes from the group of investors who have the most optimistic expectations about that security (Miller, 1977). After the offering, more
information about the firm’s future performance becomes available and the valuation by the overoptimistic investors shifts towards the mean valuation. This causes the stock price to decline and results in underperformance of the stock (King, 2014).
Ljungqvist, Nanda and Singh (2006) add the difficulty in short selling stocks to the optimistic investor explanation for IPO underperformance. It may be difficult to sell the acquired stocks right after the offering, which means that investors who have more
pessimistic expectations will never buy the IPO stock. This means that the market price of an IPO stock tends to reflect more of the optimistic investor’s view on the stock, because the
only demand comes from the optimistic investors, so the stock is even more optimistically priced (Buttimer, 2005).
Rajan and Servaes (1997) conclude that brokerage house analysts are overoptimistic in their forecasts of earnings and growth of IPOs. Forecasts used in their paper come from the Institutional Brokers Estimate System2 (IBES). Their sample of IPOs in the period 1975-1987 was divided into four groups according to the height of growth forecasts. The group with the lowest growth forecasts on average outperform the NYSE/AMEX index by 35.6% and the group with the highest forecasts on average underperform this index by 62.8% (Rajan, 1997). The expectations of analysts reflect on the decisions of investors, because investors get their information from analysts who know more about the IPO firm’s characteristics and future prospects (Rajan, 1997).This implies that investors who base their expectations on those of brokerage house analysts are even more optimistic than those analysts.
King and Banderet (2014) suggest that when investors are optimistic, the number of IPOs rises. This is supported by Rajan and Servaes (1997) who state that because of
optimistic expectations, firms see opportunities to do an equity offering with a higher stock price than when this optimism is not present. This is referred to as a firm taking advantage of windows of opportunity. Firms choose to go public at a time when they are overvalued, because then they can get higher proceeds from an equity offering than in other times. Growth and earnings forecasts, by brokerage house analysts, are indicators for a firm that investors are optimistic about its value (Rajan, 1997). This window of opportunity results in a period of high IPO volume (Ritter, 1991). King and Banderet (2014) define the last three months of 2007 as a high volume period as each of those months still had at least eleven IPOs, so these months are associated with IPO underperformance. The years 2008 and 2009 are referred to as a low volume period, because there were never more than four IPOs per month. In the sample of Lowry, Officer and Schwert (2010), the number of IPOs dropped from 59 to 21 in September 2000, which is recognized as the end of a high volume period.
Loughran, Ritter and Rydqvist (1994) state that private firms are successful in timing their IPOs in periods when they are overvalued by optimistic investors. In their sample of IPOs from 25 countries in the period 1960-1991 they find that an increase in the level of the stock market is associated with an increase in IPO volume. They also find that an increase in
2The Institutional Brokers Estimate System is a system that collects the future earnings estimates made by more
than 900 brokerage house analysts for the majority of firms in the United States (Thomson Reuters).
the level of the stock market at the time of offering is accompanied by a decrease in the three year IPO stock return (Loughran, 1994).
This thesis uses three different measures of optimism, the momentum factor from the Carhart four factor model (Carhart, 1997), the return on the S&P 500 and consumer confidence, measured in the Consumer Confidence Index (CCI). These three factors are used to identify market sentiment.
The Carhart momentum factor measures the tendency of the stock market to rise further when it is already rising or the tendency to fall further when prices are already falling. It is a proxy for the level of optimism among investors, because investors can base their expectations of firm performance on the performance of the stock market, as not much information is available on newly issuing firms. Investors become optimistic when stock market performance is high.
𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = 12(𝑠𝑠𝑀𝑀𝑠𝑠𝑠𝑠𝑠𝑠 ℎ𝑖𝑖𝑖𝑖ℎ − 𝑠𝑠𝑀𝑀𝑠𝑠𝑠𝑠𝑠𝑠 𝑠𝑠𝑀𝑀𝑙𝑙) +12(𝑏𝑏𝑖𝑖𝑖𝑖 ℎ𝑖𝑖𝑖𝑖ℎ − 𝑏𝑏𝑖𝑖𝑖𝑖 𝑠𝑠𝑀𝑀𝑙𝑙)
Kenneth French calculates the momentum factor monthly by forming six value-weighted portfolios of stocks listed on the NYSE, AMEX and NASDAQ. Two portfolios are constructed based on size, measured in market equity, where the firms are split in two portfolios with the monthly median market equity. Three portfolios are constructed based on prior returns where firms are divided according to how high prior returns are (low, medium and high prior returns). The intersections of these size and prior return based portfolios form the six stock portfolios with which momentum is calculated. In the final monthly momentum calculation four portfolios are used, the two portfolios with high prior returns (small high and big high) and the two portfolios with low prior returns (small low and big low). The variable measures the stock markets’ tendency to move in a certain direction by taking the difference between firms with high prior returns and firms with low prior returns for both small and big companies and taking the average of these two differences.
The second measure of optimism used is the monthly return on the S&P 500 index. The S&P 500 index consists of 500 large-cap companies listed on the NYSE or NASDAQ, which are also the two stock exchanges used to study IPO performance in this thesis. The S&P 500 shows the level of the market in a certain month instead of the trend of the market, like the momentum factor does. It measures how well the market is performing each month,
which is part of what investors base their expectations on, so it could be a proxy for how optimistic investors are.
The third optimism measure used is the Consumer Confidence Index which is calculated monthly by The Conference Board in the United States. It indicates the optimism among consumers about their financial situation and about the economy. The level of optimism among consumers reflects on the spending or saving activity in the country. The expectation is that consumer confidence is an indicator of the level of optimism in the stock market. CCI is measured with a monthly survey among 5000 households in the U.S., which contains five questions about business conditions which can be answered positively or negatively. Of each question the positive responses are divided by the positive and the negative responses of that question, this gives an index value which is compared with base year 1985. The index values of the five questions are averaged and this gives the total monthly CCI.
In a period when the optimism measures are high, the volume of IPOs is high and the assumption is that more overly optimistic investors are active in the market, because investors become optimistic when the market is performing well. Looking at the theory about
overoptimistic investors this would mean that when the optimism measure increases, IPO performance decreases, which results in underperformance. Ritter (1991) shows a negative relation between IPO volume and IPO performance between 1975 and 1984. For eight of the ten years in his sample, three year holding period return increased in the subsequent year when the number of issues decreased and it decreased when the number of issues increased (Ritter, 1991). In other words, IPO stocks issued in low volume periods perform better than IPO stocks issued in high volume periods. For periods of negative optimism measures, and low IPO volume, it could even be the expectation that there is no underperformance at all. If no optimistic investors are active, all investors are rational so securities are expected to be priced correctly.
High Discretionary Current Accruals
Teoh, Welch and Wong (1998) discuss the theory that firms with higher current accruals at the time of offering have poorer stock performance in the first three years following the offering. A firm’s management can inflate reported earnings by making use of discretionary accrual adjustments. Total accruals is measured as all accounting adjustments made to cash flows (Teoh, 1997). Teoh, Welch and Wong (1998) document that the discretionary cash flow accruals of an IPO firm are high at the time of the offering relative to that of comparable
firms that are not issuing equity. Firm management uses discretionary accruals in such a way that reported earnings appear higher than they actually are and if investors base their
expectations on this, they could pay a stock price that is too high. It is not directly observable to outsiders how much of the total cash flow accruals is discretionary, so Teoh and Wong (1997) try to estimate it. With a regression model based on an industry benchmark, a normal part and an excess part of total cash flow accruals of a company is estimated. This estimated excess part is used as a proxy for the discretionary part of total accruals (Teoh, 1997). Teoh, Welch and Wong (1998) focus on current working capital accruals instead of total accruals at the time of offering as predictor of three year stock performance. In this paper a sample of 1649 U.S. IPOs was divided into four groups based on how high the estimated discretionary accruals were in the year of the IPO. They find that IPO firms in the group with the highest cash flow accruals had approximately a CAR that was 20 to 30% lower than for firms in the group with the lowest accruals and a BHAR that was 15 to 30% lower. This result is robust to different test specifications so they conclude that three year underperformance of IPOs is caused by discretionary current accruals (Teoh, 1998).
3. Data and Methodology
To find out whether optimism has a significant effect on the three year performance of IPOs and to check whether this effect changed during the most recent global financial crisis, the following regression is estimated:
𝑦𝑦𝑖𝑖 = 𝛽𝛽0+ 𝛽𝛽1𝑂𝑂𝑂𝑂𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖𝑠𝑠𝑀𝑀𝑖𝑖+ 𝛽𝛽2𝑂𝑂𝑂𝑂𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖𝑠𝑠𝑀𝑀^2𝑖𝑖 + 𝛽𝛽3𝐶𝐶𝐶𝐶𝑖𝑖𝑠𝑠𝑖𝑖𝑠𝑠𝑖𝑖 + 𝛽𝛽4𝑂𝑂𝑂𝑂𝑀𝑀𝑖𝑖𝑀𝑀𝑖𝑖𝑠𝑠𝑀𝑀 ∗ 𝐶𝐶𝐶𝐶𝑖𝑖𝑠𝑠𝑖𝑖𝑠𝑠𝑖𝑖+ 𝛽𝛽5𝑉𝑉𝐶𝐶𝑖𝑖
+ 𝛽𝛽6𝑅𝑅𝑂𝑂𝑅𝑅𝑖𝑖+ 𝛽𝛽7𝑠𝑠𝑀𝑀𝑖𝑖𝑀𝑀𝑙𝑙𝑀𝑀𝐶𝐶𝑠𝑠𝑂𝑂𝑖𝑖 + 𝛽𝛽8𝑠𝑠𝑀𝑀𝑖𝑖𝑅𝑅𝑠𝑠𝑠𝑠𝑀𝑀𝑀𝑀𝑠𝑠𝑖𝑖 + 𝛽𝛽9−15𝐼𝐼1−7,𝑖𝑖
+ 𝛽𝛽16−18𝑌𝑌2006−2008,𝑖𝑖 + 𝜀𝜀𝑖𝑖
Dependent variable yi is the performance of IPO stocks in the first three years after the
offering. IPO returns will be measured in two ways, cumulative abnormal return (CAR) and three year buy-and-hold return (BHAR). This three year period is in line with Ritter (1991).
𝐶𝐶𝑅𝑅𝑅𝑅𝑖𝑖 = �(𝐶𝐶𝑖𝑖𝑖𝑖 36 𝑖𝑖=1
− 𝐶𝐶𝑚𝑚𝑖𝑖)
𝐵𝐵𝐵𝐵𝑅𝑅𝑅𝑅𝑖𝑖 = ��(1 + 𝐶𝐶𝑖𝑖𝑖𝑖) 36 𝑖𝑖=1 � − ��(1 − 𝐶𝐶𝑚𝑚𝑖𝑖) 36 𝑖𝑖=1 �
𝐶𝐶𝑖𝑖𝑖𝑖 is defined as the monthly return of each firm and 𝐶𝐶𝑚𝑚𝑖𝑖 is the monthly return on the market
portfolio. In the CAR equation the abnormal return is calculated each month by taking the difference between a stock’s return and the return on the market portfolio. Then these
abnormal returns are summed for each firm for the 36 months after the offering. BHAR has a similar approach, except for that it assumes that returns gained from investing in the stock are reinvested in the next month. This is done by each month calculating the product of the returns. These abnormal return calculations are only based on changes in stock prices, dividend payments are not taken into account. An advantage of BHAR is that it more accurately measures the experience of investors than CAR, because in its calculation it accounts for gained returns that are reinvested and CAR doesn’t as it only sums all monthly returns (Lyon, 1999). Barber and Lyon (1997) document that CAR is a positively biased predictor of BHAR, because BHAR includes the effect of monthly compounding and CAR ignores this effect. This could lead to an incorrect conclusion that a stock earns abnormal returns if only CAR is used to measure stock performance. A disadvantage of BHAR is that there is a risk of cross-sectional dependence among firms in the sample, which means that the number of firms in the sample is not equal to the actual number of independent observations, and CAR controls for this better (Lyon, 1999). Lyon, Barber and Tsai (1999) argue that BHAR is a less biased measure for abnormal returns, but that since both measures have advantages and disadvantages it is best to use both.
Both CAR and BHAR require a benchmark to compare the performance of the firm’s stock with. The return on the S&P 500 is chosen as this benchmark. To calculate
performance, for each firm monthly price data is collected from The Center for Research in Security Prices (CRSP) for the three years following the IPO, using tickers to identify the companies. The optimism variable is measured by three proxies: momentum, the return on the S&P 500 and consumer confidence. The momentum factor measures the tendency of the stock market to keep moving in the same direction as it did in the previous month. It is retrieved from the Fama French database in WRDS. For each firm the momentum of the month in which the IPO took place is used. To find out whether stock performance is different for firms that went public in a month of positive momentum than in a month of negative momentum, the sample is also split in positive and negative momentum IPOs. 144
offerings occurred in months with negative momentum and 183 offerings took place in months with positive momentum (table 1).
Table 1: Number of IPOs per year
The fact that more companies go public when momentum is positive is consistent with the theory of windows of opportunity. The second optimism measure, the return on the S&P 500, is the level of the market in the month of the IPO. This variable is simply calculated as 𝐶𝐶𝑀𝑀𝑀𝑀𝑀𝑀𝐶𝐶𝑀𝑀 = 𝑃𝑃𝑃𝑃𝑖𝑖𝑃𝑃𝑃𝑃𝑡𝑡−𝑃𝑃𝑃𝑃𝑖𝑖𝑃𝑃𝑃𝑃𝑡𝑡−1
𝑃𝑃𝑃𝑃𝑖𝑖𝑃𝑃𝑃𝑃𝑡𝑡−1 , dividend payments are not included in the calculation. It measures
how well the market performs in each month and data can be found at the website of Yahoo Finance. The third measure is the Consumer Confidence Index (CCI). It measures the level of optimism among consumers in the United States and it is calculated monthly by the
Conference Board using a survey among 5000 U.S. households.
Crisis is a dummy variable which is 1 if the IPO took place in 2008 or 2009 and 0 if the IPO took place in 2006 or 2007. Following the method of King and Banderet (2014), the years 2008 and 2009 are defined as crisis period. Of the IPOs in the sample, 61 occurred in 2008-2009 and 266 occurred in 2006-2007 (table 1). The fact that more IPOs occur in the non-crisis period than in the crisis period is consistent with the theory that the crisis is a low volume period (King, 2014). Squared terms of the optimism variables are added because the relation between these optimism variables and the dependent variable may not be linear. In an additional regression the optimism variable is substituted by a dummy variable which takes on a value of 1 if the optimism measure is positive and 0 if the optimism measure is negative. VC is a dummy variable which identifies whether a firm was backed by a venture capitalist at the time of the offering. This variable is retrieved from the Thomson One database at the same time as the full set of IPOs is collected. VC could have a positive effect on IPO
performance, because venture capitalists are specialized in bringing companies to the market. Brav and Gompers (1997) find that venture-backed IPOs outperform IPOs that were not IPO Year Number of IPOs
in sample
Positive Momentum Negative Momentum
2006 127 48 79 2007 139 100 39 2008 27 11 16 2009 34 24 10 Total 327 183 144 12
brought to the market by a venture capitalist. A number of dummy’s is added to control for industry fixed effects and year fixed effects. All IPOs are divided into seven industries based on two-digit SIC codes. These industries are mining, manufacturing transport & public utilities, wholesale trade, retail trade, finance and services.
Market capitalization, assets and return on assets (ROA) are firm specific
characteristics that could have an influence on the long-run performance of IPO stocks. These variables are retrieved from the COMPUSTAT database. For the variables assets, market capitalization and ROA applies that the value is taken at the end of the year in which the firm went public, because the COMPUSTAT database does not report these fundamentals on a daily basis. ROA is an indicator of operational performance so a higher ROA could indicate better firm performance (King, 2014). ROA is calculated as net income of the full year in which the IPO took place divided by total assets at the end of that year. The variable assets is added to the regression, because companies with higher total assets could be less affected by market conditions during the global financial crisis and this could affect stock performance (King, 2014). Market capitalization is the total dollar value of a firms outstanding equity, also referred to as firm size. It is calculated as total shares outstanding multiplied by share price. Firm size is an important forecaster of financial performance (Hendricks, 2001). Aggarwal and Rivoli (1990) argue that IPO underperformance possibly occurs more in smaller firms than in larger firms. The reason for this is that small issues are less followed by investors and by investment banks which makes it more difficult to estimate intrinsic value. To control for this, they account for a size effect (Aggarwal, 1990). Assets and market capitalization have a skewed distribution. Using the logarithm of these variables gives a more normal distribution.
The sample of initial public offerings was retrieved from the Thomson One database. The initial sample consists of all companies in the United States that performed an IPO in the period 2006-2009 and that listed on the Nasdaq or on the New York Stock Exchange
(NYSE), formerly known as American Stock Exchange (AMEX). All closed-end funds and real estate investment trusts (REITs) are excluded from the sample, in line with the method of Loughran and Ritter (1995) and Lowry (2003). After these firms are excluded, a sample of 667 IPOs is left. The IPOs were then matched with the CRSP database, resulting in a sample of 432 IPOs. A part of the firms in this sample delisted before the 36th month or no stock price data is available for the full 36 trading months. Because the performance of these firms is not comparable with the performance of firms for which full data is available, they are left out of the sample. This might indicate survivorship bias, because only the firms that survived
the first three years after issuing are included. The final sample, including only firms that did not delist before the 36th trading month, then consists of 327 IPOs (table 1).
Table 2 reports descriptive statistics for the sample. Overall, IPO stocks underperform the market portfolio, both for CAR and BHAR.
Table 2: Descriptive statistics
Table 3 shows that CARs are higher for IPOs that took place in positive momentum periods. BHARs are higher for firms that went public in negative momentum periods.
Table 3: Mean of dependent variables CAR and BHAR Performance Measure Mean Momentum>0 Mean Momentum<0 Mean S&P 500>0 Mean S&P 500<0 Mean CCI>100 Mean CCI<100 CAR –0.52 –0.57 –0.49 –0.63 –0.50 –0.61 BHAR –0.12 –0.09 –0.12 –0.10 –0.13 –0.07
To know whether the difference between mean performance of positive momentum IPOs and negative momentum IPOs is significant, a two sample t-test is performed and the p-value is constructed for both CAR and BHAR. The two sample t-test for CAR gives a t-value of 0.156
Variable Mean Mom>0 Mean Mom<0 Mean Standard deviation Min Max N Cumulative Abnormal Return (in decimals)
–0.52 –0.57 –0.54 2.86 –20.41 3.72 327
BHAR (in decimals) –0.12 –0.09 –0.11 0.74 –1.90 3.89 327
Momentum (in decimals) 0.02 –0.02 0.00 0.04 –0.35 0.13 327 Return on S&P 500 (in decimals) 0.00 0.01 0.00 0.03 –0.09 0.09 327 CCI 93.45 98.04 95.47 19.92 40.80 111.90 327 Market Capitalization ($millions) 1277.5 993.5 1151.1 3510.0 0.43 40578.3 245 Assets ($millions) 1102.4 1416.1 1240.4 4520.0 1.86 55743.0 266
ROA (in decimals) –0.09 –0.11 –0.10 0.86 –7.66 0.58 266
and a p-value of 0.562, the difference is not significant. The difference for BHAR has a t-value of -0.383 and a p-t-value of 0.351 so this is also not significantly different from zero. Figure 3 and 4 in the appendix also don’t show a difference in abnormal returns between positive momentum IPOs and negative momentum IPOs. The difference between the mean CAR of the IPOs in months of negative and positive returns on the S&P 500 yields a t-value of 0.427 and a p-value of 0.665. For the mean BHAR the t-value is 0.085 and the p-value is 0.534. For the difference between high and low consumer confidence the mean CAR has a t-value of 0.358 and a p-t-value of 0.639. For BHAR the t-t-value is 0.087 and the p-t-value is 0.535. None of these differences are significantly different from zero.
4. Results
Table 4 presents the results. It shows a negative effect of the momentum factor on three year BHAR. The coefficients on momentum are not significant and also not robust over the reported specifications of the model. This negative effect is found in the quadratic variable of momentum as well. In the regressions with CAR as independent variable, momentum has a positive relation with IPO stock returns. Again, the result is not statistically significant. The quadratic momentum variable also shows this positive effect. Table 4 documents a negative relation between the dummy crisis and both BHAR and CAR, when regressed on momentum. This result rejects the expectation that IPOs which occurred in the most recent financial crisis perform better in the first three years after offering than IPOs that took place in the non-crisis period. This result is not statistically significant, but it is robust over different specifications of the model. Table 1 of the appendix shows the BHAR regressions where the optimism measure is replaced with a dummy indicating whether the measure is positive or negative. Table 2 of the appendix shows the same regressions, but with CAR as independent variable. A negative effect of the momentum factor dummy on IPO stock returns is found, so IPOs that took place in months of positive momentum perform worse in the three years after issuing than IPOs that that took place in months of negative momentum. This results are not significant, but they are robust over different model specifications.
To measure whether the effect of momentum on three year IPO performance changes in 2008 and 2009 relative to 2006 and 2007, the interaction variable between momentum and the crisis dummy is added. The coefficient on this interaction variable (𝛽𝛽4) is positive for the reported specifications of the model (table 4). This indicates that the effect of Momentum on
IPO performance is also positively affected by whether the IPO took place in the crisis or not. These coefficients are not significant so it is not possible to say that the effect of momentum in the crisis period differs from the effect in the non-crisis period. Market capitalization is found to have a significantly positive effect on BHAR and CAR. Table 4 shows a coefficient on the logarithm of market capitalization that is positive and statistically significant at the 1% level for three of the four model specifications. This is the expected effect, because
underperformance is expected to be less in larger firms (Aggarwal, 1990).
Table 4: Regressions with Momentum as optimism measure (standard errors in parentheses)
BHAR BHAR BHAR BHAR CAR CAR CAR CAR
Momentum –1.10 (2.43) –1.39 (1.70) 0.10 (1.07) 0.08 (7.60) 0.27 (5.38) –2.12 (3.97) Momentum2 –2.36 (8.55) –6.21 (6.75) 21.00 (26.71) 7.83 (21.35) Crisis 0.03 (0.14) –0.05 (0.11) –0.66 (0.43) –0.34 (0.40) Momentum* Crisis 1.43 (3.56) 5.52 (11.11) Log(Assets) 0.00 (0.06) –0.05 (0.20) Log(Market Capitalization) 0.11* (0.07) 0.12*** (0.04) 0.72*** (0.20) 0.61*** (0.12) Return on assets 0.02 (0.10) –0.30 (0.30) Venture Capital 0.05 (0.12) 0.49 (0.39) Constant –2.97*** (1.06) –3.06*** (0.97) –0.62 (0.53) –0.57 (0.52) –13.41*** (3.31) –12.81*** (3.08) –0.01 (1.95) 0.16 (1.95) Industry fixed effects
Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effects
No Yes Yes No No No Yes No
N 245 245 245 327 245 245 327 327
R2 0.07 0.08 0.03 0.02 0.23 0.23 0.12 0.11
Adjusted R2 0.01 0.03 0.00 0.00 0.18 0.18 0.09 0.08
*
, ** and *** denote statistical significance at the 10%, 5% and 1% level, respectively.
Table 5 reports a negative effect of optimism among investors, here measured with the return on the S&P500, on the full model regression with CAR as independent variable. A positive relation is found between optimistic investors and BHAR in the regression of the full model. These coefficients are not significantly different from zero and the results are not robust over different specifications of the model so the null hypothesis can’t be rejected. One significant result can be found in the regression of the complete model with CAR as
independent variable. The coefficient on the crisis dummy is significantly negative, which could indicate that IPOs that occurred in the crisis period performed worse than IPOs in the non-crisis period. This is not the expected effect, but it is not robust over different
specifications of the model. In the regressions with S&P 500 as independent variable and where firm level controls are included, the coefficient of market capitalization is positively significant for all specifications of the model. Table 1 of the appendix reports a positive effect of the S&P 500 dummy on three year BHAR which indicates that IPOs that took place in months in which the S&P 500 yielded positive returns performed better in the three years following the offering than IPOs that took place in other months. The coefficients on this dummy are not statistically significant and this effect is also not found in table 2 of the appendix where CAR is the independent variable.
Table 5: Regressions with return on the S&P 500 as optimism measure (standard errors in parentheses)
*
, ** and *** denote statistical significance at the 10%, 5% and 1% level, respectively.
BHAR BHAR BHAR CAR CAR CAR
S&P500 0.99 (2.28) 0.88 (1.87) –0.03 (1.57) –5.31 (7.09) –1.47 (5.86) 0.60 (5.82) S&P5002 –8.85 (43.13) –25.80 (43.30) 169.94 (134.44) 150.81 (135.48) Crisis 0.06 (0.15) –0.87* (0.47) S&P500*Crisis –3.81 (3.41) 4.61 (10.64) Constant –2.98*** (1.06) -2.96*** (1.08) –0.62 (0.53) –13.90*** (3.31) –14.45*** (3.39) –0.01 (1.97)
Firm level controls Yes Yes No Yes Yes No
Industry fixed effects Yes Yes Yes Yes Yes Yes
Year fixed effects No Yes Yes No Yes Yes
N 245 245 327 245 245 327
R2 0.07 0.08 0.03 0.24 0.24 0.12
Adjusted R2 0.01 0.02 0.00 0.19 0.19 0.09
In table 6 consumer confidence is used as measure for the level of optimistic investors. It shows that the effect of CCI on BHAR is positive in two regressions, and
negative when BHAR is regressed on only CCI. The effect of CCI on CAR is negative for all specifications of the model and in the regression where CCI and CCI^2 are included the negative effect of CCI is statistically significant at the 5% level. Table 1 of the appendix shows that the effect of the dummy CCI is positive for the complete model and is negative for the model where only the optimism dummy is included. Table 2 of the appendix does not report a robust effect of CCI as well. For the regressions in table 6 the coefficient on market capitalization is again significantly different from zero.
Table 6: Regressions with consumer confidence as optimism measure (standard errors in parentheses)
*, ** and *** denote statistical significance at the 10%, 5% and 1% level, respectively.
On average the momentum factor, the return on the S&P 500 and CCI have a negative relation with BHAR and CAR, in the reported model specifications. A possible explanation for the lack of significance in the results could be that the final sample of IPOs does not consist of all IPOs that took place between 2006 and 2009 in the United States. Out of the total number of 667 firms, 235 were dropped because price data was not available in the CRSP database. Another 105 firms were not included in the final sample, because they
BHAR BHAR BHAR CAR CAR CAR
Consumer Confidence Index 0.17 (0.13) 0.03 (0.04) –0.01 (0.01) –0.37 (0.40) –0.28** (0.13) –0.01 (0.01) Consumer Confidence Index2 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00** (0.00) Crisis 3.91 (3.90) –8.97 (12.22) Consumer Confidence Index*Crisis –0.04 (0.05) 0.09 (0.15) Constant –11.14* (6.43) -4.26** (1.84) –0.30 (0.60) 4.31 (20.17) –3.73 (5.73) –0.30 (0.60)
Firm level controls Yes Yes No Yes Yes No
Industry fixed effects Yes Yes Yes Yes Yes Yes
Year fixed effects No Yes Yes No Yes Yes
N 245 245 327 245 245 327
R2 0.08 0.08 0.03 0.24 0.25 0.03
Adjusted R2 0.02 0.02 0.00 0.19 0.20 0.00
delisted before the 36th trading month so they could not be compared on the basis of three year returns. A different explanation for the absence of a negative relation between
momentum and performance is that the used optimism measures do not capture the presence of overoptimistic investors in the IPO stock market. It could be that other indexes than the S&P 500, such as the NASDAQ composite index or the NYSE composite index explain optimism better, because these are the two exchanges used in this thesis. However, after having used three different measures of optimism it can be concluded that overoptimistic investors are not the main cause of three year underperformance of IPOs. These results are not as expected, since King and Banderet (2014) and Cornelli, Goldreich and Ljungqvist (2006) find that investor optimism has a significant effect on long-run IPO performance.
Figure 5 in the appendix shows the effects that the seven different industries have on IPO abnormal returns. It is found that IPOs in the mining and retail trade industry do not underperform on average. There is underperformance on average in the other industries and it is also documented that the underperformance in the finance industry is more than twice as high as in other industries.
When CAR is used as dependent variable it is found that standard errors are larger than when BHAR is the dependent variable. For example in table 4, with momentum as the optimism measure, the standard error of the momentum coefficient is 2.43 in the BHAR regression and it is 7.60 in the CAR regression. Figure 2 in the appendix also displays these large standard errors for CARs. This holds for all model specifications and all variables and this is consistent with the theory of Lyon, Barber and Tsai (1999) that BHAR is a less biased measure for abnormal returns in a sample of firms.
5. Conclusion
Ritter (1991) finds that IPOs on average underperform the benchmark in the first three years following the offering. A possible explanation for this underperformance is the presence of overoptimistic investors who overvalue the firm at the time of the IPO. This thesis studies the effect of overoptimistic investors on three year IPO stock performance, for a sample of 327 IPOs in the United States in the period 2006-2009, by capturing the level of optimism among investors in three measures of optimism: the Carhart momentum factor, the return on the S&P 500 and consumer confidence. These measures explain investor optimism, because investors can partially base their valuation of IPO firms on these factors, as not much information is public yet for newly issuing firms. IPO stock performance is measured in cumulative abnormal returns and in three year buy-and-hold returns. It is also studied what the effect of the most recent global financial crisis is on three year IPO performance and whether the effect of optimistic investors changed during this crisis. The expectation is that the presence of overoptimistic investors has a negative effect on three year IPO stock performance, because they overvalue a stock at the time of the offering and once more information about the firm becomes public their optimistic view converges towards the view of the more pessimistic investors. This causes investors to want to sell the stocks, the excess demand that follows is what causes the stock price decline and underperformance in the long-run.
None of these optimism measures have a significant relation with the two measures of abnormal returns. On average a negative effect of momentum on BHAR and CAR is found, but this effect is not robust over all specifications of the model. The effect that IPO
performance is driven by investor optimism is not found so the conclusion is that long-run IPO underperformance is not caused by optimistic investors. This is not as expected, as King and Banderet (2014) and Cornelli, Goldreich and Ljungqvist (2006) conclude that optimistic investors are the cause of long-run IPO underperformance. For the effect of the financial crisis on IPO performance no significance is found as well, so it can’t be concluded that three year IPO stock performance is different for IPOs that take place in the crisis period than in the non-crisis period.
References
Aggarwal, R. and Rivoli, P. (1990). ‘Fads in the Initial Public Offering Market’, Financial
Management, 19(4), pp. 45-57.
Barber, B.M. and Lyon, J.D. (1997). ‘Detecting Long-Run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics’, Journal of Financial
Economics, 43(3), 341-372.
Brav, A. and Gompers, P.A. (1997). ‘Myth or Reality? The Long-Run Performance of Initial Public Offerings: Evidence from Venture and Non-Venture Capital-Backed
Companies’, The Journal of Finance, 52(5), pp. 1791-1821.
Buttimer, R.J., Hyland, D.C. and Sanders, A.B. (2005). ‘REITs, IPO Waves and Long-Run Performance’, Real Estate Economics, 33(1), pp. 51-87.
Carhart, M.M. (1997). ‘On Persistence in Mutual Fund Performance’, The Journal of
Finance, 52(1), pp. 57-82.
Cornelli, F., Goldreich, D. and Ljungqvist, A. (2006). ‘Investor Sentiment and Pre-IPO Markets’, The Journal of Finance, 61(3), pp. 1187-1216.
Hendricks, K.B. and Singhal, V.R. (2001). ‘Firm Characteristics, Total Quality Management, and Financial Performance’, Journal of Operations Management, 19(3), pp. 269-285. King, E. and Banderet, L. (2014). ‘IPO Stock Performance and the Financial Crisis’
(Working Paper No. 2456220). Retrieved from Social Sciences Research Network website: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2456220
Ljungqvist, A., Nanda, V. and Singh, R. (2006). ‘Hot Markets, Investor Sentiment and IPO Pricing’, The Journal of Business, 79(4), pp. 1667-1702.
Loughran, T. and Ritter, J.R. (1995). ‘The New Issues Puzzle’, The Journal of Finance, 50(1), pp. 23-51.
Loughran, T., Ritter, J.R. and Rydqvist, K. (1994). ‘Initial Public Offerings: International Insights’, Pacific-Basin Finance Journal, 2(2-3), pp. 165-199.
Lowry, M. (2003). ‘Why Does IPO Volume Fluctuate So Much?’, Journal of Financial
Economics, 67(1), pp. 3-40.
Lowry, M., Officer, M.S. and Schwert, G.W. (2010). ‘The Variability of IPO Initial Returns’,
The Journal of Finance, 65(2), pp. 425-465.
Lyon, J.D., Barber, B.M. and Tsai, C.L. (1999). ‘Improved Methods for Tests of Long-Run Abnormal Stock Returns’, The Journal of Finance, 54(1), pp. 165-201.
Miller, E.M. (1977). ‘Risk, Uncertainty, and Divergence of Opinion’, The Journal of
Finance, 32(4), pp. 1151-1186.
Rajan, R. and Servaes, H. (1997). ‘Analyst Following of Initial Public Offerings’, The
Journal of Finance, 52(2), pp. 507-529.
Ritter, J.R. (1991). ‘The Long-Run Performance of Initial Public Offerings’, The Journal of
Finance, 46(1), pp. 3-27.
Teoh, S.H., Welch, I. and Wong, T.J. (1998). ‘Earnings Management and the Long-Run Market Performance of Initial Public Offerings’, The Journal of Finance, 53(6), pp. 1935-1974.
Teoh, S.H. and Wong, T.J. (1997). ‘Analysts’ Credulity About Reported Earnings and Overoptimism in New Equity Issues’ (Working Paper No. 8529). Retrieved from Social Sciences Research Network website:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=8529
Appendix
Table 1: regressions with BHAR and a dummy variable for the optimism measures (standard errors in parentheses)
*, ** and *** denote statistical significance at the 10%, 5% and 1% level, respectively.
Table 2: regressions with CAR and a dummy variable for the optimism measures (standard errors in parentheses)
*, ** and *** denote statistical significance at the 10%, 5% and 1% level, respectively.
Momentum Momentum S&P 500 S&P 500 CCI CCI Positive Optimism –0.13 (0.11) –0.09 (0.09) 0.08 (0.12) 0.02 (0.10) 0.17 (0.32) –0.05 (0.13) Optimism^2 –0.51 (8.69) –4.99 (43.77) 0.00 (0.00) Crisis 0.05 (0.14) 0.06 (0.15) –1.70 (1.28) Optimism*Crisis 1.25 (2.53) –3.52 (2.77) 0.02 (0.02) Constant –2.88*** (1.06) –0.57 (0.53) –3.09*** (1.07) –0.65 (0.54) –2.29* (1.33) –0.63 (0.53)
Firm level controls Yes No Yes No Yes No
Industry fixed effects Yes Yes Yes Yes Yes Yes
Year fixed effects No Yes No Yes No Yes
N 245 327 245 327 245 327
R2 0.07 0.03 0.07 0.03 0.07 0.03
Adjusted R2 0.01 0.00 0.01 0.00 0.01 0.00
Momentum Momentum S&P 500 S&P 500 CCI CCI Positive Optimism –0.09 (0.33) –0.36 (0.33) –0.25 (0.37) 0.12 (0.37) –1.16 (1.00) 0.28 (0.50) Optimism^2 –23.56 (27.21) 162.16 (136.54) 0.00 (0.00) Crisis –0.65 (0.43) –0.89* (0.47) 5.43 (4.00) Optimism*Crisis 6.33 (7.92) 1.53 (8.63) –0.08 (0.05) Constant –13.33*** (3.32) 0.26 (1.96) –13.53*** (3.35) –0.09 (1.98) –16.39*** (4.16) 0.02 (1.95)
Firm level controls Yes No Yes No Yes No
Industry fixed effects Yes Yes Yes Yes Yes Yes
Year fixed effects No Yes No Yes No Yes
N 245 327 245 327 245 327
R2 0.23 0.12 0.24 0.12 0.24 0.12
Adjusted R2 0.18 0.09 0.19 0.09 0.19 0.09
Figure 1: three year BHAR for issues in the period 2006-2009
Figure 2: 36 month CAR for issues in the period 2006-2009
-2 0 2 4 BH AR
01jan2006 01jan2007 01jan2008 01jan2009 01jan2010
IPO date -2 0 -1 5 -1 0 -5 0 5 C AR
01jan2006 01jan2007 01jan2008 01jan2009 01jan2010
IPO date
Figure 3: Distribution of three year BHAR over Momentum
Figure 4: Distribution of 36 month CAR over Momentum
-2
0
2
4
-.4 -.3 -.2 -.1 0 .1
Fitted values BHAR
-2 0 -1 5 -1 0 -5 0 5 -.4 -.3 -.2 -.1 0 .1
Fitted values CAR
Figure 5: CAR and BHAR in different industries -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 CAR BHAR 26
