Where is the missing value? Evidence from the game industry IPOs underpricing in Poland

In the first part of the study, the average IPO first-day returns rates and their statistical significance were computed for the game and no-game companies (Table 3). The results show that the underpricing phenomenon occurred in both game and no-game sectors in 2016–2020. The analysis of the average values and the median of the IPO discount shows that the level of underpricing was higher in the game sector compared to companies from other industries (55.0% vs. 48.5% for means and 32.7% vs. 13.6% for medians). In case of debutants in the game sector, 25 out of 32 of them (78.1%) made their debut above the IPO price. In the no-game sector, the proportions were similar. Underpricing occurred in 76.1% of cases (51 out of 67 companies).

The assumption that underpricing is higher in the game sector in comparison with other groups of companies was also verified by examining the difference in rates of return on debuts between game and non-game companies and its statistical significance (H₀: u₁–u₂ ≤ 0).

Comparing the phenomenon of IPO underpricing in both game and no-game sectors, we can state that greater underpricing characterized private offers than public ones. The same applies to the IPOs of companies aiming at NewConnect as compared to larger companies to be listed on the WSE's prime market. This difference may result from the fact that private placements are addressed to a limited group of investors. Here, investors who were not invited to take part in an IPO but would like to become shareholders of given companies are forced to buy shares during the stock exchange listing. We can also assume that investors who took part in a private placement are aware of their privileged position. They do not accept premiums that are lower than in case of the public offers. A majority of them are professional investors, known to and chosen by the entity offering shares. Private issues are also smaller and addressed to non-impulse investors who have more experience in stock exchange investments.

Private issues prevail on NewConnect because of their much lower values compared to the IPOs in the WSE main market. In the period 2016–2020, the average value of the IPO preceding the debut on NewConnect was € 0.96 million (median € 0.46 million), compared to the WSE, where the average value of the IPO was € 138.0 million (median € 17.5 million). Another reason why underpricing on NewConnect was higher compared to the WSE's main market is the structure of investors in both markets. Individual investors who dominate NewConnect

In the first half of 2020, individual investors generated 22% of the turnover on the Main Market and 93% on the NewConnect market (WSE 2020)

The dynamics of the occurrence of the phenomenon of IPO underpricing in individual years also deserve attention (Table 4). In 2016–2017, four out of seven debutants from the game industry made their debut above the IPO price. In the following years, however, underpricing became a common phenomenon. In 2019, as in 2018, in the group of game companies IPOs, there was only one case where the share price at the end of the first day of quoting was lower than during the IPO, and in 2020, among 10 companies that debuted on the stock exchange, underpricing can be observed in eight cases. The scale of underpricing also increased in subsequent years, with the increased number of debuts that appeared along with the growing WIG Games index. In 2018, the average underpricing was 20.9% (with a median of 23.3%), in 2019 it increased to 73.6% (median 39.1%), and in 2020 it was 93.0% (median 84.9%).

Comparing the underpricing in the game and non-game sectors in individual years, we can see greater underpricing in the no-game sector in 2017–2018, which also coincides with a much larger number of debuts from companies outside the game sector. The increased importance of the game sector, as evidenced by the establishment of the stock exchange index for this sector in 2019 and its rally thereafter, resulted in increased interest in this industry among investors, and consequently affected the number of game IPOs. Along with usually relatively small amounts of money raised privately to selected investors before the stock exchange debut, we can see an increase in the underpricing scale in the group of game companies in relation to underpricing for debutants from other industries.

The model for testing the potential determinants of IPO underpricing level is based on univariate linear regressions measuring the relationship between a dependent variable that stands for IPO underpricing and one independent variable (chosen from a set of potential explanatory variables), which were selected based on the hypotheses formulated in the paper and literature. The independent variables concern underpricing in the following groups of companies debuting in Warsaw: (i) all companies (all IPOs), (ii) game companies (game IPOs), and (iii) no-game companies (no-game IPOs). As Table 4 shows, a method of conducting an offer (public or private) is an especially important determinant of the IPO underpricing among game and no-game companies; and the regressions were also done for following sub-groups of companies: (iv) game companies with private offer IPOs, (v) game companies with public offer IPOs, (vi) no-game companies with private offer IPOs, and (vii) no-game companies with public offer IPOs. Given the dataset of dependent variables y_i and the explanatory variables x_i the models we applied were simple linear regressions, as indicated in Eq. (1). (1) yj=α+βxi+εi {y_j} = \alpha + {\beta _{xi}} + {\varepsilon _i} where ɛ_i is the random component of the regression and x_i represents for each model one of the following variables regarding the following potential IPO underpricing determinants: (i) belonging to a game or no-game group (zero-one variable), (ii) private or public offering (zero-one variable), (iii) debut on the NewConnect market or on the WSE main market (zero-one variable), (iv) year of IPO, (v) behavior of indices (variables related to changes in WIG, WIG_GAMES, and NCINDEX indices for four periods: change in the index price on the debut day [D1], the closing price on the debut day compared to the previous day close [D–1], change in the closing price on the debut day compared to the close from the week before [W–1], and change in the closing price on the debut day compared to the previous month's price [M–1]), (vi) the size (capitalization) of the company, and (vii) the variables relating to the size of the offer and the value of shares introduced to trading (offer value, value of instruments introduced, and value of free float). We used ordinary least squares (OLS) to estimate the parameters α (a constant term) and β (the coefficient term).

The regression results are set out in Tables 5 and 6.

A total of 97 out of 99 IPOs were analyzed in the econometric study. Two companies were excluded: genXone and Spyrosoft, which represented outliers. GenXone had a 523.2% return rate on the debut (the company doing COVID-19 tests), whereas in the case of Spyrosoft, the figure was 620.0% (following underpricing resulting from a specific IPO carried out at non-market prices and offered only to employees for the amount of PLN 100,000).

The regressions results for all companies show that underpricing can vary among different industries. The significance of the game variable implies higher potential underpricing compared to no-game companies. The variables for private offers and NewConnect debuts are highly significant at 1%. These results support the observations discussed in Section 4.1, showing that the way the offer is done has a huge impact on the underpricing level. NewConnect market, where offers are smaller and in majority addressed to a limited number of investors, favors IPO underpricing as well. The third most significant variable is the year of the issue. Along with an increase in market sentiment, accompanied by an increase in indices and a higher number of IPOs, higher rates of return on debuts were observed. The role of these four factors: (i) belonging to the game sector, (ii) the way the issue was conducted, (iii) the selection of the quotation market, and (iv) the choice of the IPO period, was decisive in terms of the influence of the analyzed factors on the level of underpricing. Other factors, such as the size of the issuer, the value of IPO, free float, or the stock market situation on the day and in the period immediately preceding the debut, turned out to be insignificant.

The analysis of the underpricing phenomenon in the game sector showed that there are only two statistically significant regressors (P < 0.05): (i) the way of conducting the offer (private placement or public offer) and (ii) the time of debuting. In case of the no-game, only the following factors were, again, discovered to be statistically significant: (i) whether the offer was private or public, and (ii) on which market the stock exchange debut took place.

Interesting conclusions emerge when analyzing the factors influencing the scale of underpricing depending on the method of conducting an IPO offer. An in-depth analysis of the issues of private IPOs of the game companies shows that the lower the free float the greater the underpricing. Other factors were not statistically significant. This is because of the potentially lower supply of shares during the debut, which is below the demand from investors who were not invited to take part in an IPO, but would like to become company shareholders. With public issues of the game companies, there are more determinants of underpricing, with the most important factor being the stock market situation on the day and in the period preceding the stock exchange debut.

Underpricing that occurs in no-game companies where private placements were conducted, the size of the offer, the size of the company, and the related number of admitted shares are statistically significant. With public offerings, the statistically significant regressors are related to the market quotation and the year of issue. Interestingly, the level of underpricing with public issues is negatively correlated with the behavior of stock exchange indices on the debut day, which may suggest a shift in the demand from shares of other companies to shares of debutants on that day.

The next stage of the analysis was to examine the price and volume behavior after the debuts in both groups of companies. The analysis of the rate of return and volume after the debut was carried out for two periods: 1 month (1M) and 1 year from the debut (1Y), and then it was compared with the IPO rate of return and the volume at the first day of listings. The purpose of this study was to search for the causes of underpricing, which are related to the activity of investors on the first trading day compared to later periods.

As far as the comparison of returns between D1, M1, and Y1 is concerned, we assume that if the stock exchange price fell in a short period (1M) after the debut, it could mean that the causes of underpricing should be sought, among others, in behavioral factors. The behavioral effect is related to the desire to have shares in the investment portfolio that could not be purchased in the IPO (because of the private nature of the issue or in case of the significant reduction of subscriptions for shares in public offerings). In such situations, IPO prices were inflated, first as a result of IPO investors’ lack of willingness to sell shares in the absence of satisfactory premiums, and second by the high demand from investors who wanted to have these shares in their portfolios. The assumption that there is a relationship between underpricing and the price development in the short term after the debut (1M) arises from the fact that, as a rule, little time passes from an IPO to a debut and a month later, and there is usually no change in the fundamentals of the company during this time that would justify a change in capitalization. Thus, behavioral factors remain the main factors responsible for changes in the company's capitalization in the short term, in contrast to the longer period (1Y) in which fundamental factors play a greater role. This is because of investors’ greater knowledge of financial results or implementing the strategy described in the offer documents.

Table 7 shows the average rates of return in 1 month and 1 year from the IPO for companies from the game sector and other companies with their statistical significance, while in Figure 3 we compare the average rates of return in the periods D1, M1, and Y1.

Taking into consideration the risk factor, there is a huge difference in standard deviations between the average rates of return in the periods D1, M1, and Y1 in two markets operated by the WSE (NewConnect vs. the Main Market) and under two offerings regimes (private vs. public placement) in both groups of companies, i.e. from the game and non-game sectors. The problem of risk related to IPOs, including the relations between the rates of returns and accompanying risk for IPOs carried out on different markets and under various legal regimes in Poland, will be the subject of our further research.

Figure 3

In the group of all game companies, the average rate of return 1 month after the debut slightly increased, resulting in an underpricing rise from 55.0% to 57.5%. A similar slight increase can be observed in the group of game companies that conducted private offers that dominate on NewConnect (from 58.9% to 64.3%). With public offerings, which had much less underpricing than private offers (25.7% vs. 87.7%), average prices 1 month after the debut increased by 7.6%. In the longer term, except for shares sold in offerings preceding the Main Market debut, prices were also rising as measured by the behavior of their mean. A similar pattern of behavior characterized no-game companies, with the difference that the IPO underpricing was smaller in their case. A greater difference is visible with public offerings (and IPOs on the Main Market, where such offers dominate). However, it is difficult to arrive at a definite conclusion as regards these groups of companies, because of the lack of statistical significance of the average rates of return, which is mainly the result of the large diversification of price behavior in these groups of companies.

The average volumes of debutants in trading were next examined for the first day, 1 month, and 1 year after debuts (Table 8). The trading volumes were higher on the first trading day than after a month and a year for the game and no-game debutants regardless of whether they were private or public issues, as well as whether they were debuts on NewConnect or the Main Market of the WSE. The behavior of investors taking part in the IPO, who sell shares at the debuts, is often associated with their privileged position in relation to those investors who could not take part in the IPOs because of their private nature or a large reduction during subscription in the process of public offerings. It confirms the disposition effect, i.e., relatively quick profit-taking [Shefrin and Statman, 1985], as a natural consequence of the utility function from the prospect theory [Kahnemann and Tversky, 1979].