Do IPOs perform better during economic crises?

Do IPOs perform better during economic crises?

Amsterdam Business School Name Wouter van Roessel Student number 105558373 Program Economics & Business Specialization Finance & Organization Number of ECTS 12 Supervisor Ilko Naaborg Target completion 31 / January / 2017 Abstract

This thesis researches if IPO performance is better during economic crisis. This research compares the one-year performance of IPOs between 2001-2006 with the one year IPO performance between 2008-2011, for companies listed on the NYSE. Also the returns will be calculated, excluding the first week after the IPO. This will be researched due to the fact that IPOs often have high first day returns. By excluding the first week the IPO underperformance could be even bigger. To calculate the abnormal returns, this research makes use of the Fama and French five-factor model. To test the difference in the performance a two-sample t-test is used. This paper finds underperformance of the whole period, as for the separate periods. Also for the whole period and the period between 2001-2006 the underperformance is worse if the first week is excluded. By doing a t-test there is a significant difference found between the crisis period and the period before the crisis. Therefore the IPO performance is better compared to the market, during economic crisis, than before the economic crisis. However the robustness check shows the problems with calculating long term stock performance. Keywords: IPO, underperformance, financial crisis

Statement of Originality This document is written by student, Wouter van Roessel, 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.

Table of Contents

1. Introduction ... 3

2. Literature review ... 4

3. Methodology and Data ... 6

3.1. Methodology ... 6

4. Analysis ... 12

4.1. Empirical Results ... 12

4.2 Robustness check ... 15

5. Conclusion and discussion ... 15

6. Appendix ... 19

1. Introduction IPO’s underperform compared to the market (Ritter & Welch, 2002). This is because optimistic investors think they found the new Facebook. Therefore the price opens high but if opinions of investors come closer to each other the price becomes more on the average, and lowers. Lerner (1994) shows that venture capitalist try to take firms to the capital markets, during market peaks. By this they try to get the highest price. However there has not been many research if IPOs perform better or worse during times of economic crisis. Therefore this paper investigates if the IPO performance is better during the economic crisis 2008-2011. To investigate this research will try to answer the following question: Is the one year IPO performance, compared to the market, better during the economic crisis of 2008-2011, than before? If IPOs perform better during the economic crisis, than a crisis period will be a relatively good time to invest in IPOs. This could lead to more investments in IPOs during times of a financial crisis. This research will make use of an OLS regression based on the Fama and French five-factor model. This is a relatively new model. Therefore there has not been done research on the performance of IPOs, using this model. This model will help me to better estimate the abnormal returns of IPOs, than in previous researches. Because there are problems measuring long-term stock return (Ritter & Welch, 2002). This new way of measuring abnormal returns could lead to different findings in the underperformance of IPOs. To see which companies went to the stock market, the Zephyr database is used. Then the CRSP daily stock prices of the IPO companies are used to calculate the daily returns. The different factors of the five-factor model are downloaded from Kenneth French’s website. These give the 2x3 daily factors. First the literature review will show theories on the IPO underperformance and the difference in the performance of IPOs during market peaks and non-market peaks. Then the methods used to calculate the IPO performance are discussed. Afterwards the data used is shown, and if there are problems with the data. The robustness check shows if the model gives the same

results during a different time period. Afterwards the results will be shown, and discussed. At last a conclusion is given, and suggestions for further research 2. Literature review The probability that a company goes public rises with age and size (Pagano, Panetta, & Zingales, 1998). They mostly do this to get access to alternative finance methods instead of banks (Pagano, Panetta, & Zingales, 1998). This is especially interesting for companies with high future investments. Also going public makes it easier to spot a takeover target. Therefore it’s easier for a shareholder to sell his assets (Zingales, 1995). Furthermore going public gives diversification opportunities for shareholders, and it raises the liquidity (Pagano, Panetta, & Zingales, 1998). The biggest downside of going public is that it is very costly. Underwriters get a percentage of the IPO price, accounting cost rise and stock exchange fees have to be paid (Pagano, Panetta, & Zingales, 1998). Also the investors are less informed about the stock than the issuers. This leads to a lemons problem, which leads to a decline in the average quality of the IPO. Therefore investors are not prepared to pay a high price for the IPO stock. This under-pricing leads to a big value loss for the issuer (Leland & Pyle, 1977). Despite the high first day returns, IPOs still have bad long-term performance (Ritter & Welch, 2000). Welch and Wong (1998) say that IPO firms try to make the firm look as good as possible on paper. When the firm goes public they can no longer keep up the “optimistic” accounting (Ritter and Welch, 2002). Ritter and Welch (2002) state that not only the choice of econometric model is of influence of the long run IPO performance, but also the choice of sample period. This is especially during the Internet bubble. IPOs had initial high returns, but the bubble collapsed and the returns vanished (Ritter and Welch 2002). Therefore the IPO underperformance was really high during this period. Fama (1998) says there is a problem with the calculation of long-term stock returns, because of market inefficiencies. Therefore it could be that the measured underperformance is a result of mis-measurement. Brav and Gompers (2000) use the Fama and French three-factor model. He already improves the

model by adding the factor “no equity issued in the last 5 years” instead of the HML factor. This research will try to improve estimations even more by using the Fama and French five-factor model (2015). Lerner (1994) found that IPO’s, which are backed by seasoned venture capitalist, are better in finding market peaks. Inexperienced venture capitalist may want to take there firm public before the market peak. This way they can sign to the market that the company is of high quality (Lerner, 1994). This would mean that the average IPO quality is better if the market is “cold”. However another explanation is that less experienced venture capitalist, can’t find an underwriter who’s willing to take their firm to the market during market peaks (Lerner 1994). This is in line with Schultz (2003). He looked at IPOs between 1973 and 1997. He calculates abnormal returns by subtracting the average market returns by the returns of the IPOs. He says that firms go public when, investors are opportunistic, and firms can get a high price for their stock. Because of this less firms try to go public during an economic crisis, as can be seen in the data of this research. Loughran and Ritter (1995) say that high growth of IPOs, make it easy to justify high opening price. Optimistic investors believe they found companies, which are going to grow very rapidly. Over the years the price goes more to the average of the optimistic investors and the rest of the market, therefore the price drops (Ritter & Welch, 2002). During a crisis investors are less optimistic and the opening price will be lower. Therefore the drop in price will be lower, which leads to a smaller underperformance. Lowry (2003) researched the three-year stock return of 1526 IPOs between 1975-1984. Listed on Amex-NYSE or NASDAQ. She finds that high IPO volume leads to low IPO returns. However during times of High IPO volume, companies with similar characteristics as IPOs, are also valued high by the market. Times of high IPO value are followed by low market returns. So during this period the IPO performance is not good, but also the market is not performing well. Therefore the IPOs abnormal performance is not worse during a period of high IPO volume. This would mean there is no difference between the performance of IPOs before and after the crisis.

IPOs try to go public when the market is at its peak (Lerner 1994). They try to do this, because investors are more optimistic and will pay a higher price for the new stock. Firms can try to signal their quality to the market by going public during a “cold” market (Lerner 1994). This leads to higher quality firms going public in “cold” period than during market peaks (Lerner 1994). This will lead to less underperformance during cold markets, which is in line with the findings of Schultz (2003). Ritter and Welch (2002) find that optimistic investors buy the IPO stock in hope to get high returns. Over time, the price goes to the average of the optimistic investors and the rest of the market. Because investors are less optimistic, the difference between the market and the opinion of the optimistic investors is smaller. This leads to a smaller drop in price, which leads to a lower underperformance. This would mean that IPOs perform better, compared to the market, during the crisis period than before the crisis period. Therefore the hypothesis is: The one year IPO performance, relative to the market, is better during the economic crisis, than before. 3. Methodology and Data 3.1. Methodology There are problems with calculating long run stock performance (Ritter, 2002). Because of this, it’s hard to determine how abnormal the IPO returns are (Ritter, 2002). But because this paper will look at a shorter time period, than previous researches, the estimations will be less biased. Also the one-year returns will be used because then, the crisis, and the non-crisis period can be easily distinguished. If the 3-year returns are measured, as used in most papers, there will be an overlapping time period. Time periods longer apart from each other could be used, but then external factors will probably influence the result more, than if the time periods closer to each other. This paper will make use of the Fama and French five-factor model (Fama and French, 2015). Where 𝑅_!" is the monthly return of the IPO portfolio. 𝑅_!" Is the risk free return. 𝛼 Is a constant. 𝑅_!" Is the market return. Small minus big (SMB) is the difference between the stock returns, with size above the median and below the median. High minus low (HML) is the difference between returns

of portfolios with a high book to market equity and low book to market equity. Robust minus weak (RMW) is the difference between firms with a high and low profitability. Conservative minus aggressive (CMA) is the difference in returns between high and low investment firms. 𝑅_!"− 𝑅_!" = 𝛼 + 𝐵 𝑅_!" − 𝑅_!" + 𝑏_!𝑆𝑀𝐵 + 𝑏_!𝐻𝑀𝐿 + 𝑏_!𝑅𝑀𝑊 + 𝑏_!𝐶𝑀𝐴 + 𝜀 The betas will be calculated by doing a regression in Stata on the data set of the market. The abnormal returns are the real stock prices subtracted by the estimated stock prices. By doing a two-sample t-test, the significance of the difference can be tested. Afterwards the same regression will be done, but excluding the first week returns. This way high first day returns are excluded. This could lead to even higher underperformance. A two-sample t-test will be used, to test if the difference is significant, compared to the whole first year returns. Fama and French (1993) found that except for the beta on the market return, also the size and the book to market ratio is significant in predicting stock returns. Because small companies are of a higher risk than big companies, their returns are higher on average. Therefore SMB will give a positive relation on the abnormal returns (Fama and French, 1993). Fama and French (1993) found that firms with low a low book to market ratio have a 0.40% lower monthly return that firms with a high book to market ratio. This is because high book to market firms tend to have persistently high earning, whereas low book to market ratio persistently have poor earnings. Novy-Marx (2013) found that profitability of a company leads to significant abnormal returns. Profitable firms have on average a low book to market ratio. This would lead to lower abnormal returns. But analysis, on the premium, for companies with a high book to market ratio, Show that this premium is not driven by companies with low profitability (Novy-Marx, 2013). Titman, Wei and Xie (2004) found a negative relation between the level of investment and the stock return. They found that high investment companies tend to overinvest and thereby lowering their stock returns. Another possibility is that companies try to increase their stock return by investing, when their stock return is low (Titman, Wei, & Xie, 2004). Both theories would give a good

explanation on why the investments are a significant predictor in stock returns. Because of these researches the different factors in the Fama and French five-factor model, would give causal effects on the excess return instead of just correlations. It’s uncertain if the factors have a positive or a negative effect on the regression, because that depends on the characteristics of the IPOs. Most IPOs are small growth firms the HML will probably be positive. Because small firms mostly have higher returns than big firms. In the research of Fama and French (2014) and (2015) the HML is a redundant factor. This is because the HML return is captured by the other factors in the five-factor model (Fama and French, 2014). Therefore there is a chance that it does not have a significant influence on the stock return. The RMW is negative if companies are in the lower portfolios of profit (Fama and French, 2014). IPOs sometimes have negative profits if they go public. Therefore RMW will probably be negative. Fama and French (2014) found that the CMA factor give a negative factor for small companies with high investments, but a positive coefficient if the firms is big and has high investments. Because IPOs are mostly small firms and need to invest a lot to grow, the CMA factor will probably be negative. It’s possible that there is an omitted variable bias, because a lot of factors have influence on the price of stock. Therefore I will use a Ramsey reset test, to test if there are omitted variables. 3.2. Data and descriptive statistics First the Zephyr database is used to check which companies did an IPO on the NYSE, between the years 2001-2006 and between 2008-2011. By using this time period, the dot-com bubble will be avoided. Because of the bubble collapse most of the tech companies had negative returns. Because most of the IPO’s during this period where tech companies, is the average IPO returns are strongly negative. Because it was a bubble, this period is not representative for the IPO performance in different time periods (Ritter, Welch, 2002). The crisis period used is between 2008 and 2011. This time period is used as the economic crisis because the crisis started in 2007. With the collapse in value of mortgage backed

securities. At this time there were only problems in America, but because I use the New York stock exchange this will be a good time as the starting point. I use 2011 as the end because most of the problems were over, and the economy was in recovery. For some stocks there were missing prices in the data, therefore daily returns couldn’t be calculated. For this reason these companies are deleted from the data set. Also the calculated returns had really big outliers. These are the result of penny stocks. Because some companies had stock prices below $5, therefore a small difference in price leads to a big difference in the return. Therefore returns bigger than 50%, or smaller than -50% are excluded. This is still big for daily returns, but because part of this research is during a crisis period, returns of this size could be possible. Some research exclude penny stock, this does improve the estimation of long-term stock return (Ritter and Welch, 2002). However some of the companies with extreme returns did not start as penny stock. But during the crisis their stock price dropped dramatically. But because they started as normal companies with higher stock prices, they won’t be excluded form the sample. Figure 1: IPO returns-risk free rate Source: Wharton Research Data Services (2017) -1 00 -5 0 0 50 10 0 15 0 re tu rn -rf

01jan2000 01jan2002 01jan2004 01jan2006 01jan2008 01jan2010 01jan2012 date

The SMB, HML, RMW and the CMA factors, from the Fama and French five-factor model, are downloaded from the Kenneth French’s database. These give the daily factors from 2001-2011. If from a certain date the factors are unknown, then the closest date with known factors will be used. French (2017) uses al stocks on the NYSE, AMEX and NASDAQ to calculate the factors. Small minus big (SMB) is calculated by subtracting the average return of small companies by the average return of big companies. (French, 2017) The small companies are all companies underneath the median, and the big companies are all companies above the market median. The size of the companies is based on their market value (Fama and French 2015). The high minus low (HML), are the return of firms with a high book to market value minus the returns of firms with a low book to market value (Fama and French 2015). The robust minus weak (RMW), are the returns of firms with high operating profits minus firms with low operating profits (Fama and French 2015). The operating profitability is not just the operating profitability, because Fama and French (2014) also subtract the interest expenses. The conservative minus aggressive (CMA), are the returns of firms with low investments minus the returns with high investments (Fama and French 2015). The HML, RMW and CMA are calculated in the same way. French makes 6 portfolios based on size and book to market ratio, 6 portfolios on size and operating profit and 6 portfolios on size and investment (French, 2017). The two breakpoints are on the 30th _{and the} 70th _{percentile (Fama and French 2015). The coefficients are then calculated by} subtracting the top two portfolios by the lowest two portfolios (French, 2017).

Table 1: Descpriptive statistics Return-rf (Std. dev) Min. Max. Mktrf (Std. dev.) Min. Max. SMB (Std. dev.) Min. Max. HML (Std. dev.) Min. Max. RMW (Std. dev.) Min. Max. CMA (Std. dev.) Min. Max. #obser-vations # of IPOs 2001-2006 -1.1355 0.027279 0.01134 0.021354 0.00907 0.000423 63,870 260 (2.18291) (0.87591) (0.48983) (0.32261) (0.4048) (0.2750) -41.8188 -5.03 -3.16 -4.15 -2.67 -5.92 35.80218 5.43 2.58 2.39 3.01 2.32 2008-2011 -0.03545 0.03626 0.014795 -0.0139 0.0218 0.01336 42,105 166 (2.90897) (1.51516) (0.65542) (0.75680) (0.4265) (0.3049) -49.9886 -8.95 -3.42 -4.22 -2.36 -1.67 43.61702 11.35 4.53 4.8 1.99 1.25 Total period -0.69845 0.030847 0.012713 0.007347 0.01413 0.005563 106,405 426 (2.55414) (1.17239) (0.56150) (0.53905) (0.4136) (0.2873) -49.9886 -8.95 -3.42 -4.22 -2.67 -5.92 43.61702 11.35 4.53 4.8 3.01 2.32 Source: Wharton Research Data Services (2017) Table 2: correlations whole period returnrf MktRF SMB HML RMW CMA returnrf 1 MktRF 0.3441* 1 SMB 0.1746* 0.3055* 1 HML 0.1104* 0.2749* 0.1035* 1 RMW -0.1601* -0.4569* -0.3156* -0.2791* 1 CMA -0.0138* 0.0106* 0.1111* 0.2023* -0.2682* 1 *significance with 5%

This table shows that all coefficients are significantly correlated over the whole period to each other; this leads to higher standard errors. Fama and French (2014) also describe this. For example the negative correlation between SMB and RMW means, that big firms invest less. This correlation is also the reason that in Fama and Frech(2014) and Fama and French (2015) the HML becomes redundant. 4. Analysis 4.1. Empirical Results Table 3: coefficients rm-rf (Std. dev.) SMB (Std. dev.) HML (Std. dev.) RMW (Std. dev.) CMA (Std. dev.) Constant (Std. Dev.) 2001-2006 0.634802* (0.013991) 0.439736* (0.02214) 0.0471945* (0.034936) 0.047195 (0.03299) -0.26684* (0.04741) -1.17139* (0.008419) 2008-2011 0.7445517* (0.021437) 0.245362* (0.03443) -0.06381*** (0.03594) -0.25185* (0.05607) -0.37764* (0.06250) -0.05657* (0.012919) total 0.693186* 0.370404* 0.109640* 0.05858** -0.25205* -0.72758* (0.01332) (0.02027) (0.023681) (0.02827) (0.03910) (0.007385) Significance: *p<0.01, **p<0.05, ***p<0.10 Source: Wharton Research Data Services (2017) The regression over the whole period gives only significant factors. This means all the factors should be included in the regression. Therefore the five-factor model explains the abnormal returns better than the three-factor model. Also there is a negative constant, which is also significant. This means that over the whole period there is underperformance of the IPO’s compared to the market. In the non-crisis period the, robust minus weak, is not significant. This is not in line with Fama and French (2015). Fama and French (2015) found that HML was a redundant factor in their research. This means that in their regression the influence of the robust minus weak is captured by the other factors. In the crisis period the HML factor is redundant, which is in line with Fama and French (2015). Also the constant is less negative than the constant before the

crisis period. Therefore there seems to be less underperformance of IPO’s during the crisis than before.

Table 4: Two sample t-test

# obs Underperformance Std. Dev. 2001-2006 63870 -1.171391 0.008419 2008-2011 42105 -0.0565663 0.012189 diff>0 Pr(T > t) 1.0000 To test if the underperformance of the two time periods differs significantly, a two-sample t-test is conducted. The test shows that the difference between the two time periods is bigger than 0, whit significance higher than 99%. Therefore H0 is rejected. So the one-year IPO performance is better during the crisis relative to the market, than before the crisis. On average the underperformance of the IPO’s is not really high. This could be due to the high first-day returns. Therefore a regression will be done excluding the first week returns. Table 4: t-test comparing 1 year and 1 year-first week returns Coefficients Underperformance (Std. Dev.) Two sample t-test

Period 1 year 1 year without first week Pr(T > t)

2001-2006 -1.171391 -1.175763 1.0000 (0.0084185) (0.0084523) >0 2008-2011 -0.0565663 -0.0526998 1.0000 (0.0129189) (0.0130371) <0 Total -0.7275785 -0.7299796 1.0000 (0.0073851) (0.7299796) >0 Source: Wharton Research Data Services (2017) This table shows the one-year IPO underperformance. The first column shows the coefficient of the whole year. The second column shows the returns of the first year, excluding the first week. For 2001-2006 as for the total period the

underperformance is bigger if the first week is excluded. The difference is small but that’s because the first week is just a small part of the period, so it doesn’t influence the 1-year return by a large amount. But they do significantly differ. In the period between 2008-2011 the underperformance is smaller, excluding the first week returns. This means that the first week returns are worse than the rest of the year. This is not in line with the theory, because all the previous researches found high first day returns. It’s possible that in this period only the first day returns are high, but for the rest of the first week, there are negative returns. Table 5: Ramsey reset test.

Period F-test p-value

2001-2006 10.24 0.000 2008-2011 47.53 0.000 Total 66.38 0.000 H0= Model had no omitted variable bias. Because the P-values are really low the H0 is rejected and there is an omitted variable. This was expected up-front because there are a lot of different factors, which influence stock price. To improve the regression there has to be done more research on the company characteristics that influence stock prices.

4.2 Robustness check To test the robustness of the model, the one-year IPO performance of the first two years after the crisis will be tested. Table 6: 2013-2014 regression coefficients. returnrf Coef. P>|t| MktRF 0.012239 0 SMB 0.007196 0 HML 0.004761 0.052 RMW -0.00187 0.428 CMA -0.0025 0.516 _cons 0.000176 0.771 Source: Wharton Research Data Services (2017) During this period HML, RMW and CMA are not significant with a significance of 5%. This means that there is no evidence that the factors predict the returns of the IPOs. Another notable coefficient is the constant. In this regression the constant is positive, which would mean that IPOs perform better than the market, which is the opposite of what theory predicts. However the constant is also not significant, so no conclusions can be drawn from this regression. Because the constant is not significant, it is not possible to compare the underperformance of the crisis period, with the 2013-2014 period. An F-test on RMW and CMA shows that they are jointly significant, see table 8. So the Fama and French five-factor model would give better estimations than the three-factor model. The insignificant coefficients during this period could mean, that the Fama and French five factor model, is not a good model in estimating stock performance. However the Ramsey reset test on the regression from 2001-2011 already shows that there is an omitted variable bias. Therefore it’s possible that during this period factors not included in this model have a bigger influence than in the period between 2001-2011.

A possible explanation for the insignificance of the CMA is that the risk free rate is 0% during the whole period. Therefore it’s cheap to borrow and companies will invest more. Because of that many companies are investing a lot. Therefore the difference between the highest investing companies and the lowest investing companies is small. This leads to a small CMA, which leads to an insignificant coefficient. The insignificant RMW could also be explained by the low interest rates. The interest expenses are subtracted from the operating profits. Therefore companies who invest a get lower operating profits. However because the risk free rate is 0%, the interest expenses are also low. This leads to a reduction in the spread between companies with high operating profits to companies with low operating profits. Therefore the RMW is smaller and the factor becomes insignificant. The HML is also redundant in Fama and French (2015) and during the 2008-2011 period of this research. Therefore is not notable that this is also the case during the 2013-2014 period. 2013-2014 is a short period, therefore deviations from the model are of bigger influence than if a longer period is measured. Also because the period is short, it’s hard to generalize the results. Fama and French (2015) find that the HML is a redundant factor. And this research finds that RMW is a redundant factor for the 2001-2006 period and that HML is a redundant factor for the 2008-2011 period. But because during most time periods researched the five-factor model gives significant coefficients, the model does predict returns reasonably well. The 2013-2014 time period regression shows again problems with calculating long-term returns. So on average the Fama and French five-factor model does give good estimations of the stock prices. However the 2013-2014 period shows you have to be careful drawing conclusions from the abnormal returns, using the Fama and French five-factor model. Because a lot of factors, which are not included in the model, influence stock prices. 5. Conclusion and discussion In this paper a Fama and French five factors regression is used, to calculate abnormal returns, of firms one year after their IPO, between the period of 2001-2006 and 2008-2011. This paper finds that IPOs underperform compared to the

market, for the whole period as for the separate periods. This paper also finds that the underperformance of IPOs is less during the crisis than before. This is line with the theory because investors are less optimistic, so IPOs get less overpriced. Also good companies want to signal their quality, by going to the public market. So that investors know the company is doing good despite the bad economic times. Because of the smaller underperformance of IPOs during the economic crisis, is this period a relatively good time to invest in IPOs. The 2013-2014 period shows insignificant result. It’s curious why it differs so much from the 2008-2011 period, despite the short time period between them. An explanation for the insignificant factors could be the risk free rate, which is set at 0%. This could lead to higher investments of many companies, which leads to a reduction in the spread between the high and low investing firms. The low interest rates could also explain the insignificance of the RMW. Because of the low interest rate, the interest payments drop, which leads to a rise in the operating profit. High investing firms probably have low operating profits. But because of the low interest rate the spread between the operating profits of high and low investing companies could be reduced. This shows again the problems with calculating long-term stock returns. Although the 2013-2014 is a short period, therefore it’s hard to draw conclusion for the whole model. However the Fama and French five-factor model should be tested on multiple time periods, to test if it’s not just coincidence that the 2001-2011 periods give significant results. Also the one-year performance excluding the first week is researched. This is to see if the underperformance would be even bigger, excluding the high first day returns. For the 2001-2006 period as for the whole period, the underperformance was indeed significantly bigger. Although it was just a small difference. However the underperformance of the 2008-2011 period got smaller by excluding the first week returns. This is not in line with the theory, because all previous researches found high first day returns. A possible explanation is that, the first day returns are high, but the returns of the rest of the first week are negative. The first limitation of this research is that there are problems with calculating abnormal returns. This paper uses the Fama and French five-factor

model. The Ramsey test shows that there is an omitted variable bias. This is not strange because a lot of factor influence stock prices. If there has been done more research on the factors that influence stock prices, a better model could be used. This new way of calculating stock returns could lead to different conclusions. Also only one crisis period is used. It’s possible that the underpricing is bigger during different crisis period. For example the dot.com bubble is excluded. This drop in the market was mainly due to over-priced IPOs. So it’s possible that IPOs perform worse during a different crisis. Furthermore the period, to which the crisis period is compared, is just after the dot-com bubble. Therefore it’s possible that investors are not optimistic about IPOs. Maybe the IPOs would perform even better during the crisis period, if they are compared to a different period. Only IPOs on the NYSE are researched. It’s possible that the performance of IPOs is different on other stock markets. Future studies could try to measure the IPO performance with different ways of measuring abnormal returns. Also they should check multiple crises to see if the rise is IPO performance is not just limited to this crisis.

6. Appendix Table 7: two sample t-test comparing 2001-2006 with 2008-2011 Table 8: F-test on CMA and RMW on the 2013-2014 period ( 1) CMA = 0 ( 2) RMW = 0 F( 2, 49936) = 0.84 Prob > F = 0.4308 Pr(T < t) = 0.0000 Pr(|T| > |t|) = 0.0000 Pr(T > t) = 1.0000 Ha: diff < 0 Ha: diff != 0 Ha: diff > 0 Ho: diff = 0 Satterthwaite's degrees of freedom = 65587.9 diff = mean(x) - mean(y) t = -1.6e+04 diff -1.114825 .0000712 -1.114964 -1.114685 combined 105,975 -.7284592 .0016761 .5456323 -.7317444 -.7251741 y 42,105 -.0565663 .000063 .0129189 -.0566897 -.0564429 x 63,870 -1.171391 .0000333 .0084185 -1.171456 -1.171326 Obs Mean Std. Err. Std. Dev. [95% Conf. Interval] Two-sample t test with unequal variances

References Brav, A., Geczy, C., & Gompers, P. (2000). Is the abnormal return following equity issuances anomalous?. Journal Of Financial Economics, 56(2), 209-249. Fama, E. & French, K. (1993). Common risk factors in the returns on stocks and bonds. Journal Of Financial Economics, 33(1), 19-32. Fama, E. (1998). Market Efficiency, Long-Term Returns, and Behavioral Finance. Journal Of Financial Economics, 49(3), 291-293. Fama, E. and French, K (2014). A five factor asset pricing model. Retrieved from; http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/working_papers.html Fama, E. and French, K. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), pp.1-22. French, K. (2017). Kenneth R. French - Description of Fama/French Factors. [online] Mba.tuck.dartmouth.edu. Available at: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/Data_Library/f-f_5_factors_2x3.html [Accessed 20 Jan. 2017]. Leland, H. & Pyle, D. (1977). Informational Asymmetries, Financial Structure, and Financial Intermediation. The Journal Of Finance, 32(2), 371. Lerner, J. (1994). Venture capitalists and the decision to go public. Journal Of Financial Economics, 35(3), 293-316. Loughran, T. & Ritter, J. (1995). The New Issues Puzzle. The Journal Of Finance, 50(1), 23-51. Lowry, M. (2003). Why does IPO volume fluctuate so much?. Journal Of Financial Economics, 67(1), 3-40. Novy-Marx, R. (2013). The other side of value: The gross profitability premium. Journal Of Financial Economics, 108(1), 1-28. Pagano, M., Panetta, F., & Zingales, L. (1998). Why Do Companies Go Public? An Empirical Analysis. The Journal Of Finance, 53(1), 27-64. Ritter, J. & Welch, I. (2002). A Review of IPO Activity, Pricing, and Allocations. The Journal Of Finance, 57(4), 1796, 1817-1821. Schultz, P. (2003). Pseudo Market Timing and the Long-Run Underperformance of IPOs. The Journal Of Finance, 58(2), 489-492. Titman, S., Wei, K., & Xie, F. (2004). Capital Investments and Stock Returns. Journal Of Financial And Quantitative Analysis, 39(4), 677-700.

Teoh, S., Welch, I., & Wong, T. (1998). Earnings Management and the Long-Run Market Performance of Initial Public Offerings. The Journal Of Finance, 53(6), 1964-1965.