Master of Finance

Master of Finance

Master Thesis Finance

EBM866B20

Academic year 2017-2018

Influence of IPO prices and other factors on nominal share prices over

the long-term. Comparison between civil and common law.

Abstract

This paper investigates whether the offer price in initial public offerings (IPOs) could serve as an anchor for future nominal share prices of companies. In regard to this topic, previous literature either analysed one country only or one common group of countries. This paper makes a distinction by pooling countries and firms into two samples to see any potential differences between them, namely, civil law sample and common law sample. We analyse 1745 firms from 15 countries from 2000 to 2017. Results show, that in both samples the offer price has a significant positive influence on the firm’s share prices over the period and could be considered as a natural anchor for them.

Keywords: IPO; anchoring; share price behavior; civil law; common law

Student number: s2489570 Name: Simon Zakharov

Table of contents

1. Introduction...3

2. Literature review...5

2.1. Nominal share prices...5

2.2. Initial Public Offerings and performance over time...7

2.3. Country-specific differences...8

3. Methodology and hypotheses...9

4. Data...11

5. Results...17

6. Conclusion...20

References...22

1. Introduction

Many researchers from the financial field and the field of economics have found that the level of nominal share prices of many companies is stable over time. Stock prices, themselves, are freely determined by supply and demand of the markets, but there is an evidence that there exists a certain range in which these prices are kept over the long-term.

Dyl and Elliott (2006) described this phenomenon as the nominal share price puzzle. Weld et al. (2009) found that the average nominal share price on the New York Stock Exchange

(NYSE) and the American Stock Exchange (AMEX) has been stable and approximately equal to $25 for about 70 years since the Great Depression. Dyl and Elliott (2006) and Weld et al. (2009) suggest that this constant level of share prices over such a long period does not “just happen” and there exists a reason why share prices behave this way. They suggest that managers conduct certain actions such as stock splits, paying out dividends, doing share repurchases or following certain norms to affect the prices and keep them constant. Share prices, themselves, serve as an indicator about the health of a company and, also, can impact company's reputation. Usually, investors and managers consider a higher share price of a company as a good sign, however, managers continue to conduct certain actions to control the prices. This is what makes it the price puzzle.

This paper considers another possible reason and explanation of the nominal share price puzzle. We investigate whether the offer price in an initial public offering (IPO) has any effect or influence on the future level of nominal share prices of companies in countries that operate under civil and common law systems. One potential mechanism behind this relationship is the anchoring theory. It was introduced by Tversky and Kahneman (1974) and defined as a type of cognitive bias, in which humans have a tendency to put more attention and preference to the information that they recognize first. They conducted an experiment in which two groups of respondents were asked to calculate a product of several numbers and were given only 5 seconds to complete the task. The first group was asked to compute a product of 8×7×6×5×4×3×2×1, while the second group of participants was asked to calculate 1×2×3×4×5×6×7×8. The result was that the first group assigned a higher number, than the second group, on average. In our case the offer price in an IPO is the first price for a company that is offered to investors, hence, this is the first price recognized by investors and managers. That is why, the offer price in an IPO can serve as a natural anchor for future share prices of companies, because managers would keep the prices close to the offer price in an IPO. This could be another way of explaining the nominal share price puzzle.

common law sample. It is obvious that countries can have differences in cultures, development of institutions, religions, models of corporate governance and other national-based factors. That is why, we separately analyze two samples of countries and firms to identify the differences between them. Additionally, this research focuses on the data starting in year 2000, when there was an IPO boom, especially in Europe. This is what is different from the previous papers, that analyzed a timespan in previous years. There have been many events in the past 10 years, at least, which could change the nature of exchanges and firms’ performance, e.g. financial regulation, the United States Securities and Exchange Commission’s (SEC) tighter control over markets, the financial crisis. In other words, the results could differ from many previous papers. Finally, we consider whether a cross listing of a company in other markets, other than the main one, has an effect on its price over time.

The central research question of the paper is: “Does the offer price of a firm in an IPO serve as an anchor for its nominal stock price over the long-term?”

In order to answer this question, we follow the method of Bae et al. (2016), who conducted a similar research, but for a common group of countries. The dependent variable in the model is a firm’s share price in year t and the main explanatory variable is a firm’s offer price in an IPO. The timespan of our analysis is from 2000 to 2017. We collect company’s offer prices in IPOs, IPO dates and yearly share price observations each 1st _{January. In order}

to be included into the analysis, each firm must have at least 5 yearly observations. After collecting the necessary data, we conduct cross-sectional regressions year by year. Due to the fact that we have 17 years in our timespan, we, hence, conduct 17 cross-sectional regressions and then we average the coefficients, t-statistics, R-squared values among them and report the percentage of the coefficient estimates that are positively or negatively significant. In order to identify the differences among our samples, we introduce a law system dummy. In addition, we include the market value variable into our model. According to Dyl and Elliot (2006) and Weld et al. (2009), the firm size is an important determinant of a nominal share prices and should be taken into account. Also, as noted above, cross listing dummy is included into our model to identify differences between cross listed firms and non-cross listed ones.

Results of the analysis show that in both samples, the offer price in an IPO is positively significant and that other factors such as the firm’s size and being a cross listed also have an effect on the future share price level. Among all variables, the offer price alone is the most important factor in explaining the stock prices. We can conclude that the offer price could be considered as a natural anchor.

2. Literature review

Given the fact that there is not much prior research that directly deals with the offer prices in IPOs and their influence on nominal share prices at the same time, we discuss share price behavior related and IPO related studies, that have a link to this paper, separately. 2.1. Nominal share prices

Anchoring theory is one of the main theories that can explain a constant level of nominal share prices over time. Tversky and Kahneman (1974) argue that humans tend to assign more attention, preference and value to the information that they recognize first. In our case, the offer price in an IPO serves as a potential anchor for the future share prices of companies, because this is the price for a company that investors and managers see in the first place. In prior literature, however, an anchor is not a certain number or nominal price, but rather a certain range within which the nominal share prices are controlled on purpose.

Dyl and Elliott (2006) analyze the stock price behavior of 1,019 firms between 1976 and 2001 in the United States. They argue that some public firms choose a certain price trading range for the stocks and keep the prices within the range to reflect the investors’ desires. Their main finding was, that when a company’s share price exceeds this certain established range, it predicts a stock split, eventually, and that when a split happens, it usually attracts more individual shareholders, that have more preference for low-priced stock in comparison to the institutional investors. In other words, the share price dynamics do not happen “as it is”, but are controlled and managers control share price levels to increase the value of a firm. So, stock splits conducted by management is one of the reasons of a stable level of nominal stock prices over time and could be considered as one of the factors explaining the price puzzle.

Also, De Ridder and Burnie (2016) conduct an analysis for both the nominal and real prices of the public Swedish firms for a time period from 1900 until 2011 and show that there is a price decline over time. The main idea of their results is similar to the findings of Dyl and Elliott (2006). Namely, firms’ managers undertake managerial actions such as stock splits, reverse stock splits or dividend payouts to control the share price level in order to attract new investors and cater current investors.

could also be an exhibition of anchoring theory. Additionally, Dyl and Elliott (2006) and Weld et al. (2009) argue that the firm size is an important determinant of share prices, hence, it could be considered as a control variable in our paper. The papers mentioned above suggest that managerial actions are the main tools to control the levels of nominal share prices within a certain range, that, in turn, serves as an anchor for the stock prices.

Following the findings of Weld et al. (2009), who suggest that nominal stock price anchoring is exclusively the United States phenomenon, Bae et al. (2016) suggest that the nominal stock prices tend to converge to the offer prices in IPOs, eventually, and that the offer price serves as a potential anchor. They continue analysis of Weld et al. (2009) by analyzing a pool of companies worldwide. They suggest that the IPO price, among all alternatives, has the most significant effect on nominal stock prices of public firms. In their analysis, they reject the findings of Weld et al. (2009), who claim that among their sample of the countries, only in the U.S. there is a nominal price fixation over the long-term. The authors analyze 38 countries comprising 21,285 firms. They collect stock prices of the firms between 1983 and 2010 and hypothesize that companies’ IPO price is the main anchor for the share prices and that the prices would eventually converge or would be close to it. Then they divide the stock price movements into three categories in terms of the strength of their movement and report that most of the firms’ share prices do not leave their price group. They pool all countries together in their analysis and find that price is, basically, anchored worldwide. Hence, the result was a clear influence of the offer price.

Another theoretical explanation of a stable nominal stock price behavior can be explained by dividend payment theories. The study by Baker and Wurgler (2003) sheds light on the theory that is called catering theory of dividends. They suggest that managers’ decision to pay out dividends to investors is, actually, driven by investors’ demand. Managers “cater” investors by paying dividends, when investors put a high stock price premium on payers, compared to non-payers. According to the authors, the main feature of this theory is that managers are able to satisfy investors by catering payouts to them. Therefore, managerial actions might be the main force that keeps the stock prices constant over time and within a certain domain.

2.2. Initial Public Offerings and performance over time

It is known that IPOs are underpriced, on average, and it has been a topic for a debate over years in the literature. It means, that the price of a newly listed firm trading on a secondary market is higher than its offer price in an IPO. According to Bansal and Khanna (2012), underpricing is an anomaly, due to the fact that it contradicts the efficient market hypothesis. However, Ritter (1991) suggests that this underpricing is eliminated in the short-run and that firms underperform in the long-run, on average. Ritter (1991) and Chambers and Dimson (2009) argue that there are three common periods we can consider regarding IPOs and their performance. The first one is called the initial return period. It comprises the offer price of a company as well as its first trading price on a secondary market. In other words, this period is one day long. The second period is called the short-term, which is less than one year long and, very commonly, takes only several days up to a month upon an offering. The third period that exists is called the long-term period and it takes more than one year of time, on average.

In regard to the initial and short-term dynamics, Barry and Jennings (1993) focus entirely on the first trading day for an IPO. They do not analyze the first trading day, but divide it in two stages. The first one is an opening price return and the second one is intraday trading itself. After collecting a sample consisting of 229 IPOs, their main finding is that the underpricing is, in general, corrected during the first day of trade, but not for all firms. Underpricing is adjusted by the price-setting process that is established by an opening price. It means that there might not be any return after the intraday dynamics and that only the most benefited customers (subscribers) are better-off. Additionally, Lowry et al. (2010) focus on the volatility aspect of the initial return period by analyzing 11,734 firms between years 1965 and 2005. The authors find that throughout the whole period initial returns demonstrate high volatility and conclude that underpricing is an issue, because of the underwriters. Underwriters are not able to adequately value a company, usually, due to information asymmetry or inability to predict investors’ aggregate demand properly. Lastly, relating to the short-term, another noteworthy finding was recognized by Boulton et al. (2010). They argue that underpricing as well as initial and short-term returns are higher in those countries, where Anglo-Saxon corporate governance model is established. In those countries and firms, the dispersed ownership of firms prevails, which means that firms are owned by many investors. Hence, there are many investors and more dispersed ownership, rather than several major inside owners with high concentration of ownership. They analyze 4,462 firms in 29 countries and find significant results that higher underpricing is present more often in Anglo-Saxon corporate governance firms and that it is eliminated faster in those countries.

at the first day of an IPO, would get a 83 cent return after 3 years, which means that those firms underperformed. He states that the main reason why the companies underperform over the long-run is not exactly explicit, but claims that a possible explanation might be that firms try to go public at the peak of industry-specific fads. Furthermore, Yi (2001) argues that firms underperform just after three years after going public. He analyzes a sample of public firms consisting of 1,032 companies between years 1987 and 1991. His finding is that the firms underperform the market as well as a control group of firms after 3 years upon going public. The finding is in line with Ritter (1991) and supports the idea of investors’ over-optimism, who believe in a successful future of newly listed companies. Lastly, Chambers and Dimson (2009) claim that despite the fact that nowadays there is a big improvement in terms of IPO regulation and underwriters’ reputation comparing to the past, the firms still often underperform over the long-run. They analyze a sample of 3,661 IPOs between years 1935 and 1972. The authors find different results. With buy-and-hold returns underperformance is present, but insignificant, while with cumulative abnormal returns the underperformance vanishes at all.

In general, usual positive and dynamic returns over the first day and short-term periods could be explained by investors’ optimism towards newly listed companies, which corrects IPOs’ underpricing quickly and drives the prices up. And this over-optimism can also be a problem why many companies often underperform within just several years after going public, due to strong relying on high positive earnings in the future. It is clear from the findings that share prices go above the offer prices, in most of the cases, in the short-term period and that companies underperform in the long run, meaning that the share prices start to go down together with the returns. It is expected that the stock prices of firms would eventually converge or be close to the offer prices of firms in IPOs, supporting our idea of anchoring and converging to the offer prices.

2.3. Country-specific differences

It is appropriate to mention that countries differ in certain characteristics. These could be culture, religion, mentality, development of institutions and markets. These differences could also explain particular nominal share price levels across countries. According to the culture theory of Hofstede (1980), nations are comparable between each other through five dimensions1 _{and, according to the theory, e.g. some nations might have different risk}

preferences, which, in turn, can explain variation in share price levels among countries. Importantly, countries operate under different law systems, where the biggest of them are civil law system and common law system. According to Pejovic (2001), two systems have a different approach to the legal process. Common law system takes its roots in England and its main feature is that precedents (previous court judgments) are not just taken into account, but

1 _{Hofstede’s culture dimensions are: individualism - collectivism; uncertainty avoidance; masculinity -}

courts mostly tend to base their judgments on precedents unless there is a significant reason to change precedent. On the contrary, civil law system originates from continental Europe and its main characteristic is that systems are based on codifications of the laws and precedents are not binding. Due to the fact that these systems are different in their nature, there could be differences in terms of regulation between the countries that operate under these systems. Essentially, the countries that operate under civil law system all have the German corporate governance model applied in business, whereas countries where common law prevails use the Anglo-Saxon corporate governance model.

When there is a comparison between two systems, one of the most important elements to compare are shareholders’ protection and dividend payout policy. Based on findings of Ferris et al. (2007), investors in countries that operate under common law system face more dividend payments, than ones who own a company in a civil law country. They analyze two samples between 1995 and 2005 and find that the dividend payment frequency is higher in common law countries. Additionally, La Porta et al. (1998) analyze 45 countries and find that in common law countries there is the strongest form of investor protection in comparison to civil law countries with weak legal protection of shareholders. They find that higher concentration of ownership is negatively related with legal protection of investors. Farinha and López-de-Foronda (2005) support these findings and argue that in common law countries, there is a lower concentration of ownership and there is better protection of rights for minority shareholders than in civil law countries. Their result supports the idea that civil law and common law regimes have a common parallel between Anglo-Saxon and German corporate governance models where there are the same properties regarding the ownership structure and dividend payout policy within the firms. Based on these findings we can expect that the final results would differ among our samples.

3. Methodology and hypotheses

In order to conduct our analysis, obtain the results and answer our research question, we follow the methodology of Bae et al. (2016), who conducted a similar type of research.

to Bae et al. (2016), we do not analyze one common group of countries and firms, but assign countries and firms to two different samples, which are the civil law and the common law samples, to identify any potential differences between them. To determine the differences among two samples, we create a law dummy and interact it with our main explanatory and control variables. Based on prior literature findings, we expect that the offer price of firms in IPOs as well as the market capitalization of companies would have a positive and significant relationship with the nominal share prices. However, due to country-specific distinctions across countries, we expect to see different results among two constructed samples. Finally, we obtain the information whether each firm is listed on any other exchange besides its main one. Hence, we introduce a dummy variable for cross listed firms and communicate it with the main explanatory variable of the offer price in an IPO to see the differences.

Over half of the countries from our civil law sample use euro as their main national currency, however, some countries have their own national currency. That is why we convert all currencies into the United States dollar (USD) and conduct our analysis using the USD as the main currency for both samples. All offer prices in IPOs and market values of firms are converted into USD as well.

To test the relationship, we use the following regression model:

𝑃_𝑖,𝑡 = 𝛽₁+ 𝛽₂𝐼𝑃𝑂_𝑖 + 𝛽₃𝑆𝑖𝑧𝑒_𝑖,𝑡+ 𝛽₄𝐶𝑟𝑜𝑠𝑠 ∗ 𝐼𝑃𝑂_𝑖 ( 1 )

where 𝑃𝑖,𝑡 is the nominal stock price of firm i on the 1st of January in year t, 𝐼𝑃𝑂𝑖 is the offer

price in an IPO of firm i, 𝑆𝑖𝑧𝑒𝑖,𝑡 is related to the market capitalization of firm i at time t and

𝐶𝑟𝑜𝑠𝑠 is a dummy variable for cross listed companies. The timespan of our analysis is from 2000 to 2017. In order to be included into the analysis, each country must include at least 10 firms and each firm must have a certain offer price, certain IPO date and at least 5 yearly observations of share prices and market capitalization.

To test the statistical relationship between variables, we follow the procedure that was conducted by Bae et al (2016). We run cross-sectional regressions for each year of the sample period, hence, there are 17 regressions in total. The dependent variable is a share price of a firm in USD on the 1st _{of January in year t. The independent variables are the offer price of a}

company in an IPO and the market value of equity on the 1st _{of January in year t, both in USD.}

non-stationarity of the variables. Additionally, to reduce the influence from outliers, we winsorize our continuous variables at 5% and 95% percentiles and, additionally, all of our coefficients and t-statistics are based on White-Hinkley heteroscedasticity corrected errors. Firstly, we regress our stock prices only on offer prices. Then we substitute market value of a firm instead of offer prices to check the relationship. Finally, we include all variables into the regression model. Overall, 17 regressions implement 17 coefficients for each independent variable, t-statistics, R-squared and adjusted R-squared values for each sample. We average these measures across 17 regressions to get a single, general result for the civil and the common law samples. Finally, we report the percentage of obtained coefficient estimates that are positively or negatively significant at 5% level.

Based on our research question and prior literature discussion, the main hypotheses of this paper can be defined as follows:

H1: There is a positive relationship between the nominal stock price of a public firm and its offer price in the long-run in a country operating under the civil law system

H2: There is a positive relationship between the nominal stock price of a public firm and its offer price in the long-run in a country operating under the common law system.

4. Data

As discussed in the previous section, to apply our regression model and test our hypotheses, we collect the data in regard to three variables. The first variable for which we obtain the data is a company’s offer price in an IPO. Offer price in an IPO is the price per share offered to investors by a firm when it becomes public. Zephyr and Orbis databases contain firm specific and IPO specific information in a global context. The databases are used to identify the available companies through International Securities Identification Number (ISIN) codes and also used to collect the offer prices in IPOs in the United States dollars and in the local currencies together with IPO dates for these companies. Additionally, Zephyr database is used to retrieve the country codes for each firm and to obtain the information whether a firm is listed on other exchanges, besides its main one. The second variable of our interest is a firm’s nominal share price. To collect the prices, we use Thomson Reuters DataStream database. This database contains different time-series financial and economic data. We obtain the share prices of identified companies in the United States dollars and the local currencies between 2000 and 2017 each January 1st_{. In order to be included into our analysis, each country must comprise at}

upon their IPO date. As mentioned in the previous section, the period of our analysis is between 2000 and 2017. This period is chosen, due to the fact that there is a serious lack of IPO related data prior year 2000. According to Ritter (2003) and Burhop and Chambers (2016), IPOs were a common practice in the United States for a long period, whereas in Europe IPOs started to appear more often only in the end of 1990s, when there was a boom in the informational technologies and Internet development. Finally, the last variable in our analysis is the market capitalization of a firm. According to DataStream definition, it is a firm’s share price multiplied by its number of shares outstanding. We obtain the market capitalization in the same manner as we obtain a firm’s share price. In other words, market capitalization of each firm is collected in USD on the 1st _{of January, yearly, and 5 consecutive observations per firm are required.}

Given the fact that this paper analyzes two distinctive samples, we assign countries and firms to a certain group, based on the law system that a country operates under. The first sample, which is related to the civil law includes the countries located in continental Europe. All countries, that are located in continental Europe, with the exception of the United Kingdom and Ireland, operate under the civil law system and apply the German corporate governance model in business. The reason for analyzing European countries is that many of them use Euro as a common currency and that most of their markets are strong and liquid. The civil law sample, as a result, comprises 13 countries, which are Austria, Belgium, Denmark, France, Germany, Greece, Italy, Norway, Poland, Spain, Sweden, Switzerland and the Netherlands. Lithuania, Latvia, Czech Republic, Estonia, Romania and Bulgaria were excluded from our analysis, either due to unavailability of 10 firms per country or due to missing observations of some variables. Additionally, all the countries from this sample have not had any currency regime changes throughout the period of our analysis. Overall, the civil law group comprises 900 public companies.

hypothesize that share prices would be anchored around the firms’ offer prices, thus, the share prices are expected to converge with the offer prices, eventually. To check this, we present the offer-to-share price ratio for each country in column (6). This ratio fluctuates from the minimum of 0,68 in the Netherlands to the maximum of 2,06 in Greece. However, it is interesting to notice, that the average of this ratio across countries is equal to 1,02, which is almost equal to 1. On the level of the sample, it indicates that the average share price is close to the average offer price in IPOs and this supports the idea of Bae et al. (2016) about the price convergence. Next, columns (7), (8) and (9) present the average offer prices, average share prices and offer-to-price ratios per country in the local currencies. The ratios vary from the minimum of 0,49 in Spain to the maximum of 2,71 in Greece, but for most of the countries, ratios are close to 1.

The second sample of our analysis is the common law sample. This group includes two countries, which are the United States and the United Kingdom. These countries both operate under the common law system and apply the Anglo-Saxon corporate governance model in

United States Dollar Local Currency (1) (2) (3) (4) (5) (6) (7) (8) (9)

Country # of

firms IPOs Period

Offer price average Share price average Offer/Share price ratio Offer price average Share price average Offer/Share price ratio Austria 16 2000-2011 24,89 30,13 0,83 21,14 20,28 1,04 Belgium 26 2000-2008 18,35 18,25 1,01 14,39 14,88 0,97 Denmark 11 2000-2011 50,40 29,69 1,70 135,97 105,21 1,29 France 196 2000-2012 15,02 15,52 0,97 12,48 14,78 0,84 Germany 158 2000-2012 23,96 18,30 1,31 20,74 20,11 1,03 Greece 41 2000-2009 9,15 4,44 2,06 9,16 3,38 2,71 Italy 70 2000-2012 12,41 9,28 1,34 12,58 8,05 1,56 Norway 30 2000-2012 5,29 4,36 1,21 34,17 37,55 0,91 Poland 230 2000-2012 4,12 4,49 0,92 12,39 18,58 0,67 Spain 24 2000-2011 16,42 15,53 1,06 13,04 26,65 0,49 Sweden 51 2000-2012 4,06 4,65 0,88 30,38 46,85 0,65 Switzerland 32 2000-2012 77,65 96,77 0,80 105,42 88,54 1,19 The Netherlands 15 2000-2011 13,27 19,43 0,68 13,24 30,67 0,43 Total 900 21,15 20,83 1,02

Table 1. Summary of data of civil law countries

business. The reason why these countries are included into the analysis is due to the fact that they are one of the biggest and few countries operating under common law, that they have one of the most developed and liquid markets and exchanges in the world and due to availability of IPO related data. We collect the data in the same manner as for the civil law sample. As a result, 845 firms were collected for the analysis and a total of 1,745 firms were collected for both samples.

Table 2 presents a summary for the common law sample countries. Again, column (1) presents the name of a country that is included into our sample and column (2) shows a number of firms that we collected for each country. Column (3) presents the periods for each country during which IPOs were conducted. In the United States, the end of IPOs period is 2013, because several companies had an IPO date on the 1st _{of January of 2013. Next, columns (4),}

(5) and (6) present the average offer price in an IPO in USD, the average share price in USD and the offer-to-share price ratio, respectively, while columns (7), (8) and (9) present the same measures, but in the local currency. The United States dollar is the domestic currency in the US, that is why the local currency analysis is required only for the UK.

We notice, that in the UK the offer-to-price ratio is close to 1, while in the US it is 0,51, which means that the prices in the US did not converge to the offer prices closely. One potential explanation for this could be the financial crisis of 2007-2008, when the number of IPOs as well as the offer prices were low. Hence, the average offer-to-price ratio of the sample is only 0,55 overall, which is lower in comparison to the civil law sample.

United States Dollar Local Currency (1) (2) (3) (4) (5) (6) (7) (8) (9)

Country # of

firms IPOs Period

Offer price average Share price average Offer/Share price ratio Offer price average Share price average Offer/Share price ratio United Kingdom 269 2000-2012 1,90 1,62 1,17 1,26 79,41 0,02 United States 576 2000-2013 13,39 26,14 0,51 Total 845 7,64 13,88 0,55

Table 2. Summary of data of common law countries

Next, we present the graphical interpretation of the development of the nominal share prices and the offer prices of civil law and common law firms over time. Figure 1 presents the dynamics of the average nominal stock prices and the average offer prices of the civil law sample countries over 17 years. Vertical axis in the graph presents the prices in the United

States dollars, whereas horizontal axis represents the years of our analysis. Each year we calculate the average nominal share price as well as the average offer price in an IPO for the whole sample. The average offer price line starts in year 2000 and the average share price line starts in year 2001, because we collect the share prices each January 1st _{and the earliest year,}

when we can obtain the prices is 2001. The average offer price line ends in year 2012, because, as noted earlier, we require at least 5 sequent share price observations per firm, hence, 2012 is the limit for an IPO date. The graph comprises 900 firms from 13 countries, that operate under civil law. It is noteworthy, that both lines are different from each other in terms of the prices, but they move in a similar way over the years. We can also notice a drop in the offer prices starting in year 2010. A possible explanation for this could be that European firms were recovering from the financial crisis of 2007-2008 and that countries had stagnation in terms of economic development. We can also notice that despite the financial crisis appearance during the period, the average share price for the whole sample is very similar for years 2001 and 2017

Figure 1. Development of average share prices and average offer prices of the civil law sample over time

Figure illustrates the trends of the average nominal stock prices and the average offer prices in IPOs for the period from 2000 to 2017 for 900 firms that exercise the civil law system. There are 13 countries, which are Austria, Belgium, Denmark, France, Germany, Greece, Italy, Norway, Poland, Spain, Sweden, Switzerland and the Netherlands. The local currencies were normalized by converting them into the United States dollar(USD). The vertical axis shows the prices in USD, while the horizontal axis represents the years of the period.

0 5 10 15 20 25 30 35 40 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 $US

given the price fluctuations throughout the period, which supports the idea of Bae et al. (2016) that the nominal share price puzzle is not just the United States phenomenon, but can be recognized worldwide.

Furthermore, Figure 2 illustrates the same graph as the previous one, but for the common law sample countries. The graph is constructed in the same way as the graph for the

civil law sample countries. We can notice that both lines illustrate behavior that is different from the civil law sample. The average share price line and the average offer price line are both close to each other and look quite constant over the period. The average offer price of the sample firms is less volatile than the average offer price of the civil law sample firms, while the average share price is constant, which can also indicate, that the nominal share price puzzle exists. Additionally, Appendix 3 and Appendix 4 exhibit the scatterplots for both samples. Each scatterplot illustrates a relationship between the average offer and share price for the whole sample. Both plots show that there exists a linear relationship between the average offer price and the average share price, however, this relationship is negative in the common law sample. A reason for this could be that the average offer price in the common law firms is higher, during most of the years, than the average share price, while it is vice versa for the civil law firms. This can be seen in Figure 1 and Figure 2.

Figure 2. Development of average share prices and average offer prices of the common law sample over time

Panel illustrates the dynamics of the average nominal stock prices and the average offer prices for the period from 2000 to 2017 for 845 firms that exercise the common law system. There are 2 countries, which are the United Kingdom and the United States of America. The local currencies were normalized by converting them into the United States dollar(USD). The vertical axis shows the average stock price and the average offer price in USD, while the horizontal axis represents the years of the period.

1 0 5 10 15 20 25 30 35 40 45 50 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 $US

5. Results

This section presents the results of our statistical analysis and includes a discussion of them. Table 3 presents a summary of our regression results for both samples. As was mentioned in methodology section, 17 regressions were conducted for each variable resulting in 17,721 observations each time. We introduced the law dummy to see the differences among two samples and report the results for both samples based on the interaction of the dummy with the variables. Column (1) from the table shows the civil law sample results of the regression, where the only independent variable is the offer price in an IPO. The average coefficient among 17 regressions is equal to 0.353. The average t-statistic is 5.103 and the proportion of the coefficient estimates that are positively and significantly significant at 5% level is 0.88, meaning that almost every coefficient across the regressions was positively significant. We can conclude that the offer price in an IPO has an impact on the future nominal share prices of the firms that belong to the countries that operate under civil law. Column (2) presents the results of the same regression, but for the common law sample. The average coefficient is higher, than for the civil law sample and equals to 0.743 with the average t-statistic that is 5.904. The proportion of the coefficient estimates that are positively significant is higher, than in the civil law is equal to 0.94. Again, we conclude that in the common law sample countries, the offer price in an IPO has an impact on the future level of share prices. R2 _{and adjusted R}2 _{for the}

whole regression sample are equal to 0.274 and 0.271, respectively, meaning that, approximately, a third of variation in share prices can be explained by the offer price only.

Furthermore, column (3) presents the civil law sample results of the regression, where the independent variable of the market value of a firm is substituted instead of the offer price. The average coefficient across the regressions is quite low and is equal to 0.001, whereas the average t-statistic is equal 3.321. The proportion of positively significant coefficients at 5% level is equal to 0.76, meaning that the firm size also has an impact on the share prices. Column (4) presents the results of the same regression as in column (3), but for the common law sample countries. The average coefficient is a little higher than for the civil law sample and is equal to 0.004 and the average t-statistic is 3.895. The proportion of the positively significant coefficient estimates at 5% level is the same as for the civil law sample and equals to 0.76. R2 _{and adjusted}

R2 _{fall in comparison to the previous regressions, where the offer price was the only}

independent variable, and are equal to 0.066 and 0.065, respectively. We can conclude that, in general, the market value of equity does have partial influence on the share prices of firms as was confirmed by Dyl and Elliott (2006), Weld et al. (2009) and Bae et al. (2016).

Civil law Common law Civil law Common law Civil law Common law Variable (1) (2) (3) (4) (5) (6) Offer price 0.353 (5.103) [0.88 ; 0] 0.743 (5.904) [0.94 ; 0] 0.690 (7.576) [0,94 ; 0] 1.073 (8.072) [0.88 ; 0]

Market value of equity

0.001 (3.321) [0.76 ; 0] 0.004 (3.895) [0.76 ; 0] 0.005 (7.322) [0,94 ; 0] 0.006 (5.446) [0.82 ; 0] Cross listing 0.077 (1.603) [0.53 ; 0] 0.016 (1.813) [0.59 ; 0] Regressions 17 17 17 Observations 17721 17721 17721 R2 _{0.274 0.066 0.428} Adjusted R2 0.271 0.065 0.421 Table 3

Summary of regressions of nominal stock prices on offer prices and other variables in the civil law and common law samples

now equal to 0.94. The average coefficient of the market value variable equals to 0.005 with the average t-statistic of 7.322. The proportion of the positively significant coefficients at 5% level equals to 0.94. We notice that all the coefficients, t-statistics and proportions are higher for the civil law sample, when we consider the full model instead of separate variables. Additionally, the average coefficient of interaction of the cross listing variable with the offer price in an IPO is 0.077, the average t-statistic is 1.603 and the proportion of positively significant coefficients equals to 0.53. Usually, in order to be listed on more exchanges other than the main one, a firm faces more requirements and must establish good reputation, high-quality management and healthy financial base. That is why, the coefficient is positive and the effect is stronger for cross listed companies. Finally, column (6) presents the results of the full model for the common law sample countries. The average coefficient of the offer price estimate is equal to 1.073. The average t-statistic equals to 8.072 and the proportion of positively significant coefficient estimates at 5% level is 0.88. The average coefficient of the market value variable estimate is 0.006 and the average t-statistic equals to 5.446. The proportion of positively significant coefficient estimates is 0.82. Again, we get higher estimates of the coefficients, t-statistics and the proportions of positively significant estimates, when we run the full model including all variables into the analysis, with the exception of the proportion of positively significant estimates of the offer price that decreases from 0.94 to 0.88. In regard to the cross listing variable interaction with the offer price, we get the average coefficient of 0.016, which is a little smaller than for the civil law sample, the average t-statistic of 1.813 and the proportion of positively significant estimates is 0.59, which is approximately equal to the civil law case. We conclude, that being cross listed across several exchanges reveals stronger effects than for non cross listed firms. We notice that R2 _{and adjusted R}2 _{both increase, when}

we run regressions with the full model and are equal to 0.428 and 0.421, respectively.

6. Conclusion

In this paper we investigate whether there is any relationship between the nominal share prices of public companies and their offer prices in IPOs, which were predicted to act as natural anchors for the nominal share prices. Additionally, we test whether the firm size and firm’s presence on other exchanges other than the domestic one would have an effect on the nominal share price as well. We create and consider two distinctive samples of firms and countries, namely, civil law and common law samples and try to investigate whether our findings would support our main hypotheses and be in line with the previous literature findings. After statistically analyzing 1,745 firms from 15 countries, we obtain the results that show that the offer prices could serve as real anchors for the future stock prices of public companies.

Firstly, Table 1 and Table 2 indicate, through the offer-to-share ratios, that the average share price is close to the average offer price in an IPO in most of the countries. Secondly, Figure 1 and Figure 2, that illustrate the dynamics of the nominal stock prices and the offer prices over time, show that the stock prices and the offer prices either move in a similar manner or that they converge, periodically. Statistical analysis proves that the offer prices do show significance and could act as natural anchors for the future stock prices in both samples of countries, supporting the findings of Bae et al. (2016) and the anchoring theory of Tversky and Kahneman (1974). In a given 17-year period of our analysis, which could be considered as the long-term, the offer price still serves as an important factor explaining share price behavior. When a cross-sectional regression for year 2017 was conducted, it was clear that the offer price still has an effect on firms after such a long period from their IPO date. Important to mention, that in line with Dyl and Elliott (2006) and Weld et al. (2009), we empirically investigate that the size of a firm is, also, an important factor explaining the future stock price in both samples, but not that important as the offer price. Lastly, the cross-listing variable, which was interacted with the offer price shows that cross listed companies exhibit stronger effects, than the ones that are listed only on their main exchange.

References

Bae, K., Kang, J., Rhee, G., 2016. Nominal Stock Price Anchors: A Global Phenomenon? Working paper.

Baker, M., Wurgler, J., 2004. A catering theory of dividends, Journal of Finance, 59, 1125-1165

Bansal, R., Khanna, A., 2012. Determinants of IPOs Initial Return: Extreme Analysis of Indian Market, Journal of Financial Risk Management, Vol.1, No.4, 68-74

Barry, C.B., Jennings, R.H., 1993. The Opening Price Performance of Initial Public Offerings of Common Stock, Financial Management, 22, No. 1, 54-63

Boulton, T.J., Smart, S.B., Zutter, C.J, 2010. IPO Underpricing and International Corporate Governance, Journal of International Business Studies, 41, No. 2, 206-222

Burhop, C., Chambers, D., 2016. Initial Public Offerings: A Historical Overview, Research Foundation Publications, Financial Market History: Reflections on the Past for Investors Today, 132–146.

Chambers, D., Dimson, E., 2009. IPO Underpricing over the Very Long Run, Journal of Finance, 64, No.3, 1407-1443

De Ridder, A., Burnie, D.A., 2016. Managerial actions and nominal stock price levels, The European Journal of Finance, 22:14, 1435-1456

Dyl, E.A., Elliott, W.B., 2006. The Share Price Puzzle, Journal of Business, 79, No. 4, 2045-2066

Farinha, J., López-de-Foronda, Ó., 2005. The relation between dividends and insider ownership in different legal systems: International evidence, The European Journal of Finance, 15, No. 1, 169-189

Hofstede, G., 1980. Culture’s Consequences: International Differences in Work–Related Values, Sage Publication, Beverly Hills, CA.

Jong-Hwan Yi, J.H., 2001. Pre-offering earnings and the long-run performance of IPOs, International Review of Financial Analysis, 10, 53-67

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1998. Law and Finance, Journal of Political Economy, 106, No. 6, 1113-1155

Lowry, M., Officer, M.S., Schwert, G.W., 2010. The Variability of IPO Initial Returns, Journal of finance, 65, No. 2, 425-465

Pejovic, C., 2001. Civil law and common law: two different paths leading to the same goal, Victoria University of Wellington law review, 32, No. 3, 817-841

Ritter, J., 1991. The Long-Run Performance of Initial Public Offerings, Journal of Finance, 46, No. 1, 3-27.

Ritter, J., 2003. Differences between European and American IPO Markets, European Financial Management, 9, No.4, 421-434.

Tversky, A., Kahneman, D.,1974. Judgment under Uncertainty: Heuristics and Biases, Science, New Series, 185, No. 4157, 1124-1131

Appendix

Appendix 1. Countries from the samples and corresponding Zephyr country codes

Appendix 2. National currency names and codes

Austria AT Belgium BE Denmark DK France FR Germany DE Greece GR Italy IT Norway NO Poland PL Spain ES Sweden SE Switzerland CH The Netherlands NL United Kingdom GB United States US

Country Currency name _{Currency code}

Austria Euro _EUR

Belgium Euro EUR

Denmark Danish krone _DKK

France Euro EUR

Germany Euro EUR

Greece Euro EUR

Italy Euro EUR

Norway Norwegian krone _NOK

Poland Polish zloty _PLN

Spain Euro EUR

Sweden Euro EUR

Switzerland Swiss franc _CHF

The Netherlands Euro EUR

United Kingdom Pound sterling _GBP

Appendix 3. Relationship between the average offer price and the average share price per year in the civil law countries.

Appendix 4. Relationship between the average offer price and the average share price per year in the common law countries.