Short-term performance in the Canadian marijuana sector

(1)

Short-term IPO Performance in the Canadian Marijuana Sector

Author:

M. K. Gavrilovic

Student number:

10757392

Thesis supervisor: Dr. J. J. G. Lemmen

Finish date:

31 January 2018

UNIVERSITY OF AMSTERDAM

BSc Economics & Business Economics

Bachelor Specialisation Finance and Organisation

Thesis Finance

(2)

STATEMENT OF ORIGINALITY

This document is written by Mike Klaas Gavrilovic who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are 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.

(3)

ABSTRACT

This thesis examines the short-term IPO performance of thirty-nine Canadian marijuana companies, between 1999 and 2017. For the short-term performance, an average underpricing-level of ninety-seven percent is found. The Canadian marijuana companies are more underpriced than the ten percent

underpricing found for comparable non-marijuana listed Canadian companies. From the regressions it appears that the natural logarithm of the variable offer price and the variable days till listing are most significant. The effect of twenty days trading volatility after listing day also has a significant effect on underpricing in two of the regressions, at a ten percent significance level.

Keywords: Underpricing, short-term, marijuana sector, Canada JEL Classification: G1, G32, G38

(4)

LIST OF TABLES

Table 1 Hypotheses 11

Table 2 Descriptive statistics 13

(6)

LIST OF FIGURES

Figure 1 Cross-sectional underpricing averages in Canada . 5

Figure 2 After market volatility on underpricing 6

Figure 3 Annual number of IPOs 12

(7)

CHAPTER 1: INTRODUCTION

1.1 Relevance

One of the first biggest steps in a company’s life cycle is going public, also known as an initial public offering or “IPO”. One of the main reasons to go public is enabling a company to raise capital on the stock market to finance expansion, instead of raising capital at financial institutions. This is a cheaper way to raise capital for the given company, because this kind of equity offer lacks the commitment to pay interest. In contrary, debt requires periodically payments as a price of lending capital and to partially cover the risk of a company going default. Other side effects of going public are; the possibility of attracting better performing managers, meeting requirements for disclosure and

operating in the name of a larger group of shareholders. The downside of IPOs that is discussed in this thesis, is the theoretical term “underpricing”, this means that the IPO offer prices are underpriced relative to the first day closing prices of the stocks. This leads to an inefficient capital raise for the firms that set up the IPO, because they could have changed the offer price closer to the first-day closing price. If this is a known fact supported by empirical theory then why do firms continue doing this (Draho, 1971)?

The market of focus in this thesis is the relatively new Canadian marijuana sector and the companies listed on the Canadian exchanges, which relate in any way to producing, selling or providing any goods for the marijuana sector. This sector is currently very volatile, full of uncertainty and in a state of bullish investor sentiment1, which causes investors to be willing to pay relatively very high prices for the stocks traded on the market, leading to all-time high valuations and monthly returns of 300 percent in capital gains are no exceptions. The marijuana companies listed in Canadian exchanges sum up to a total market capitalization of more than 20 billion Canadian dollars. One of the main reasons for the all-time high valuations is the fact that marijuana will be legalized for recreational use in Canada, by the first of July (Alexander & Skerritt, 2017).

All of the companies are in a state of expansion, because all of them want to be able to fulfil the highly expected consumers demand. While most of these companies are still making losses, they’re valued at hundreds of millions of Canadian dollars. The expectations for these companies are undeniably high, but this wasn’t always the case. These companies are known penny stocks, stocks that are offered at an offer price below five dollars, when they initially offered their shares to the public, and many of them are still trading below this level. Some expectations for the coming year say that a few share prices of these stocks will hit triple-digits, but that has yet to happen.

(8)

1.2 Contribution

Multiple researches have been done on this topic to observe and explain underpricing in IPOs. This led to multiple regression models, with some using different sets of variables than others. The reason behind the fact that there are multiple regression models, which are using different determinants, could be explained by the fact that the researches focus on different domains or different time periods. Research on IPO underpricing in the Canadian market has been done by several researchers like Jog and Riding (1987) but was mainly focused on the Canadian market as a whole and concentrated mostly on the time span from 1970-1999. This thesis focuses on the time span from 1999 to 2017 and isn’t equal to any existing literature, in terms of sector and domain focus.

1.3 Research question

The main question of this research states as follows:

“Is underpricing of IPOs of the current Canadian exchange venture listed companies, which produce or sell marijuana or make consumer devices for the purpose of smoking marijuana, higher than the average Canadian IPO underpricing? Can this be explained by existing literature?”

To answer this question a set of ten hypotheses has been set up, which are stated in chapter 3.

1.4 Main results

This research concludes that the natural log of the offer price, which is price for which the IPOs are offered, and the variable days till listing, which is the difference in days between announcement day and listing day, are the main explanatory variables of underpricing. Underpricing in the marijuana sector is ninety-seven percent, against roughly ten percent underpricing of other Canadian IPOs, which is stated in a research by Jog and Riding (1987). However, the high standard deviation in the sample leads to an insignificant result, when the two means are compared.

1.5 Outline

The thesis’ setup is as follows: chapter 2 focuses on the theoretical explanation of the theory used in this research and includes empirical findings of existing literature. Chapter 3 focuses on the

methodology of how the data has been retrieved and includes the stated hypotheses following from the regression models. Followed up by chapter 4, that presents the results and the data analysis. In this chapter the results are discussed, and the variables are explained in more depth. The last chapter, number 5, includes a general conclusion that answers the research question, followed up by several limitations to this research.

(9)

CHAPTER 2: THEORETICAL AND EMPIRICAL EVIDENCE

In this chapter theoretical and empirical findings on IPOs and underpricing are reviewed to get familiar with the concepts lying underneath the regression model.

2.1 Initial Public Offerings

An initial public offering is the first time that a private company offers stocks to the public. Through this offer the company raises capital to finance any future expansion plans, in a way that may be less costly than taking on new debts. The capital will be acquired through investors, who get a rightful share of the company in exchange. There are multiple reasons for a company to go public, but there are also some disadvantages to it. Not every IPO succeeds and the process of going public brings direct expenses with it. An underwriter has to be hired, which calculates the offer size, offer price and the best period to go public (Draho, 1971).

2.1.1 Reasons to go public

In the decision to go public the advantages should outweigh the disadvantages of going public. The disadvantages mainly consist of the fact that an IPO brings along direct cost, which might weight heavier than the cost of taking on debt in short-term. Thus, in some cases it is possible that one of the main benefits of going public, which is raising cheap capital, might be overshadowed by the direct cost (Draho, 1971). Except the advantages of going public don’t only consist of raising capital on the stock market, but also involves a few other things. One of these benefits is an increase in reputation, meaning that the company will get more known and might benefit from new deals given their

enhanced reputation. Another benefit is the fact that stock prices are a solid performance measure of a company, from the public’s perspective, which includes fundamental values as well as investor sentiment. It also allows the board to reward the management with stock options, which will cause incentives for further alignment between the management’s focus and shareholder value (Brealey, Marcus & Myers, 2001).

2.1.2 Process and involved parties

The process of going public is complicated and costly, without a guarantee of success. In the process multiple parties are involved namely; the issuer, the underwriter and the investors. The issuer is the firm that decides to go public and offers its shares to the investors on the stock market. The process is complicated; the issuer needs to hire an underwriter to smoothen the process and to collect all the necessary information before going public. The issuer and the underwriter work closely together to set up a prospectus. The prospectus will contain all the key factors for the IPO, including the offer size and the offer price. The offer size is how much shares will be issued, and the offer price will state at

(10)

what price these shares will be issued. The calculation of these variables will be based on peer group data analysis and the investor sentiment at the moment to adapt to the investors willingness to pay a certain price. The last one may be affected during crisis periods, because investors will be more cautious to spend their capital, which could lead to a lower price that they’re willing to pay. The IPOs also have to measure up to the specific country’s regulations and the regulations that the chosen venture exchange maintains, before they are actually issued (Jenkinson, & Ljunqvist, 2001).

2.2 Underpricing

When the prospectus is finished, and all the regulations are met, it is time for the issuer to offer its shares to the public. They are traded at a venture exchange for the offer price, which is filed in the prospectus. Numerous researches among others by Jog and Riding (1987) and Ritter (1984) have shown that the first day closing price of a share is substantially higher than the offer price of the offered shares. This means that the shares are being traded at a discount. This phenomenon is better known as underpricing. Underpricing is defined in formula by subtracting the offer price from the stocks first day closing price and divided by the offer price.

2.2.1 Theoretical evidence

The fact that multiple researchers have done event studies on this particular subject brings along the fact that numerous theories exist, which try to explain underpricing. The most entrenched theory among all is the model of asymmetric information. This theory states that when more parties are involved in a certain situation of any kind and one of the parties has more information about that particular situation than there is asymmetric information, because one party is better informed (Akerlof, 1970). According to Ritter (1984) IPOs are underpriced because of asymmetric information and can be explained from the fact that investors have less information than the issuer. The issuer knows more about the company’s and the management’s performance than the investors know. The investors want to be compensated for the possible risk of having less information, leading to a discount in share price Ritter. Another theory from the asymmetric information perspective is the theory that there exist two different types of investors. The first type is the ‘superior investors’, who possess more information than the company itself and more information than the second type of investors. The second type of investors is called the ‘uninformed investors’, who have no information available about the quality of equity offers. The so-called superior investors will only invest in profitable equity offers and won’t be willing to invest in non-profitable equity offers. The uninformed investors on the other hand have no idea which equity offers are a good investment, and which are a bad investment. Thus, to compensate for the losses that the uninformed investors incur, a discount on the IPO has to be offered to guarantee that they will invest (Rock, 1986).

(11)

2.2.2 Empirical evidence

Literature on IPOs in Canada shows evidence for underpricing. In the table underneath, extracted from an academic research done by Jog and Riding (1987), underpricing for the first ten days after the IPOs were issued, was calculated. To get the data the researchers searched the libraries of two major underwriters, because there is no centralized database. They used a total sample of 160 companies from the period January 1971 to December 1983, but had to exclude several observations due to lacking data. The table doesn’t only show that the Canadian IPOs were underpriced around ten percent on average, but the data also seemed to be significant (Jog, & Riding, 1987, pp. 8-13).

Figure 1: Cross-sectional underpricing averages in Canada. Source: Jog, V., & A. Riding (1987).

2.2.3 Determinants

Further Canadian evidence shows the effect of after market volatility on underpricing. In a research conducted by Jog and Wang (2002) evidence was found for average underpricing of Canadian IPOs and the significant effect of aftermarket volatility on underpricing. They researched the effects of twenty trading days, sixty trading days and hundred twenty trading days after market volatility of the IPOs on underpricing. They concluded that all the variables have a significant effect on underpricing, but the variable that is most significant is the twenty trading days’ volatility. The table below shows the results of the effect of after market volatility.

(12)

Figure 2: After market volatility on underpricing. Source: Jog, V., & Wang, L. (2002).

Research by Leone, Rock and Willenborg (2007) shows the effects of age on underpricing. They concluded that a company’s age is a significant explanatory variable for underpricing for a five percent significance level, when the natural logarithm is taken of this variable. The explanation for the significance is the fact that when a company goes public long after it’s founded, the investors are more familiar with the company and its performance. Investors want to be compensated for the fact that they aren’t familiar with a specific company, thus a lower age leads to more underpricing. Another IPO underpricing research from Theoh, Welch and Wong (1998) also shows evidence for underpricing. They examined the impact of market capitalization on underpricing. They concluded that smaller firms are more underpriced on average. The reason behind this comes from the fact that firms with a smaller market capitalization show more uncertainty about future performance, than firms with a large market capitalization.

(13)

CHAPTER 3: METHODOLOGY AND HYPOTHESES

This chapter reviews the method that was used to get the results that are displayed in chapter four and chapter five. Besides that, the hypotheses that are tested are stated in this chapter in combination with the regression model.

3.1 Method

The companies used for this research were selected by searching the official Canadian government’s website, the CSE-listing list and the TSX-listing list. The companies that were chosen to be included have to be involved in either the cultivation or retail of marijuana. Because this would lead to an insufficient amount of observations other companies were also included if their main objective was modifying marijuana using biotechnology, producing marijuana related products like cannabis-oil or produce technological consumer devices for the purpose of using marijuana products. On the official website of the Canadian government there is a list of Canadian companies who are licensed for marijuana cultivation and retail2. However, this list is not up to date, thus other companies had to be searched on the venture exchanges lists. At this point the total number of observations summed up to 85 different companies.

For each company a set of variables is needed to eventually run the regression. Data that was searched included: announcement day of the IPOs, offer prices, offer sizes, first day of trading, first day of trading closing share prices, the companies’ founding year, the companies’ licenses and the venture exchange where the companies were listed. With this dataset all the needed variables could be generated. The first three variables were extracted from the official financial filing site for Canadian based companies, called SEDAR. This site seemed to be lacking data about numerous companies’ prospectus, thus the Bloomberg terminals at Erasmus University were accessed to complete this list data as much as possible. Companies for which the prospectus isn’t available on these databases are excluded from the dataset.

The first trading day and first day closing price were retrieved from Datastream, by using the

unadjusted share price. The unadjusted share price was used, because this is the official closing price on that day, which has to be used to calculate underpricing at that time. The adjusted price includes stock splits that are done later on, which can lead to enormous differences between offer price and first day closing price. Datastream’s database was lacking a few number of observations, thus the

Bloomberg terminals were used to complete this list. Companies for which the share prices are not available are excluded from the dataset. The last variables were retrieved from the official CSE, TSX

2

(14)

https://www.canada.ca/en/health-canada/services/drugs-health-products/medical-use-and specific companies’ websites. After all the unusable observations were excluded a sample of 50 companies was left.

After all the necessary data was retrieved the regressions were run by using the ordinary least squares method in SPSS. First of all, a small regression was run, which includes the variables offer price, offer size, proceeds, days till listing, age and volatility. The variable offer size was debatable in terms of if the natural log should be used, thus a test for normality was performed. The reason behind this is that OLS only works for normally distributed variables. After the small model was tested with the normal value for offer size and one regression with the natural logarithm of offer size, the other variables were added to the regression for the final regressions. All the null-hypotheses except number one are tested in a one-sided test.

3.2 Variable explanation and hypotheses

For some of the variables the natural logarithm is taken, instead of the actual input as variable. The reasoning behind this is to make variables more normally distributed and to reduce the spread of the residuals. This linearizes the variables, which is important in this regression, because the ordinary least squares method is used to run the regression and estimate the coefficients.

The dependent variable in the model is underpricing, which is described in chapter 2. This is the variable that will be explained by all the used independent variables, which are selected based on research by Jenkinson and Ljungqvist (2001) and other papers that are cited in the specific variable explanation, if applicable. Underpricing in this market is expected to have a higher average compared to the IPO underpricing of other Canadian companies. Underpricing in formula:

The first independent variable is the offer price. Based on existing literature the offer price is negatively correlated to underpricing, which means that a higher offer price will cause less underpricing. The reasoning behind this is that a lower offer price is seen as signal for more

uncertainty for the future IPO’s performance, thus more underpricing is needed for investors to buy the stock (Ritter & Welch, 2002). The natural logarithm of the variable is used, because theory suggests this for dollar prices. The variable is defined as follows:

(15)

The second variable is offer size, which is the total amount of shares that is offered in the IPO. The offer size is expected to have a positive effect on underpricing according to existing literature. Higher levels of offer size will lead to more dilution and bigger dispersed number of shareholders, which leads to inefficiency, thus leading to an increase in underpricing.

The next explanatory variable is proceeds, which is defined as the total offer size times the offer price. This variable is expected to have a negative correlation with the variable underpricing, because a higher level of proceeds is seen as a sign of a well-established firm. In a research conducted by Ritter (1991), evidence is shown for the significance of the proceeds on underpricing. In formula:

The fourth independent variable is the amount of days till listing. This variable is defined by calculating the number of days between the announcement day and the first day of trading. Some observations seemed to be outliers and have been removed from the dataset, every observation counting more than 365 days till listing were excluded. This is the last part where data was excluded and reduced the useable number of observations to thirty-nine. The variable is expected to have a positive relation to underpricing, because when a company takes longer to go public it could indicate issues with the process from the investor’s perspective, which brings uncertainty with it.

The next variable is age, which is the difference between the listing date and the date that the company was founded. For several companies the exact date was missing so the founding year was used instead. The variable age has an expected negative relation to underpricing according to the literature. The natural logarithm was used and a one was added to each observation, because some observations equaled zero (Leone, Rock & Willenborg, 2007). This prevents the observations equaling zero to come up as missing in Stata and SPSS.

The last non-dummy variable is the aftermarket volatility. Based on the research from Jog and Wang (2002), the twenty trading days after market volatility is used, because this variable showed to be most significant. The effect of underpricing is expected to be positively correlated with underpricing,

(16)

compensated for the additional risk of uncertainty, thus resulting in higher levels of underpricing. The marijuana market is a relatively new market, without definite numbers about the total worth of the market. The expectation is that this leads to more uncertainty and volatility, which results in more underpricing, than the average market. The variable in formula:

The first two dummy variables are retail and cultivation. The dummy retail equals one if the specific company is licensed for the retail of marijuana and zero if otherwise. The dummy cultivation equals one if the company is licensed for the cultivation of marijuana and zero if otherwise. Both dummies are zero if the company has none of these licenses, for example for a company that only produces consumer devices to smoke marijuana. Both these dummies are expected to be negatively correlated with underpricing, because the licenses grant the companies advantages compared to the companies without licenses. Cultivation licensed companies can grow their own marijuana at low costs in contrast to companies that aren’t licensed and have to buy their marijuana from the cultivation companies. The retail licensed companies have an advantage compared to their peers, because they can reach a broader number of buyers compared to the firms that can only sell their goods to the limited number of

companies.

The other dummy variable is CSE, which is a one in the regression if the company is listed on the CSE venture exchange and a zero if the firms are listed on the TSX. All the firms are either listed on the CSE or the TSX, thus no other venture exchange variables are included. Underpricing is expected to increase when a company is listed on the CSE, because on average the firms with a smaller market capitalization are listed on the CSE. Based on the research of Theoh, Welch & Wang (1998) this should have a significant positive effect on underpricing.

All the variables stated above lead to the following final regression:

(17)

To answer the research question multiple hypotheses have to be stated. The expected effects of all the variables are stated above, but are summarized in the table below. means IPO underpricing of the Canadian listed marijuana firms and means IPO underpricing in Canada, based on literature.

Table 1: Hypotheses

1. H0: IPO underpricing in the Canadian marijuana sector is the same as the Canadian average based on existing literature.

H1: IPO underpricing in the Canadian marijuana sector is greater than the average Canadian IPO underpricing based on existing literature. 2. H0: 0 H1: 3. H0: 0 H1: 4. H0: H1: 5. H0: H1: 6. H0: H1: 7. H0: H1: 8. H0: H1: 9. H0: H1: 10. H0: H1:

(18)

CHAPTER 4: RESULTS ANALYSIS

This chapter firstly reviews the descriptive results from all the used variables to run the regressions. After that the regressions that were run are analyzed.

4.1 Descriptive statistics

The descriptive statistics table is separated in a panel for real values and a panel for the variable values after the natural logarithm was taken. Even though some variables are only used in natural log in the actual regressions, it is easier to explain them in their unimpaired values. The first variable is the dependent variable, underpricing. The numbers in this row have to be multiplied by one-hundred to get the underpricing in term of percentages. The mean of this variable states the average level of IPO underpricing of the sample, which equals roughly 97 percent. When this is compared to the average levels of underpricing of other Canadian samples from existing literature (10 percent) it is seen that the underpricing of this sample is seriously larger. The big difference could be due to the effect of outliers, but these weren’t excluded due to the already small sample. This can be backed up by the fact that the median is only forty percent, thus the most observations lie around this number. Underpricing changed over time, thus next to the fact that this sector is more volatile this can also be a reason for the big difference. The research that is used for comparison focuses on the time range from 1971 to 1983 and this research focuses on the time range 1999 to 2017, which is displayed in the chart below. The vast majority of observations are issued from 2008 to 2017. However, the time span was stretched to 1999 to get a better amount of observations. Another noticeable event in the chart is that 2009 is the only year where no IPOs took place, starting from 2003. The reason for this could in lie in the fact of the lasting crisis at that moment. The reason that the IPOs started increasing from 2003 is due to the fact of the increasing implementation of medical marijuana in the country.

Figure 3: Annual number of IPOs.

Using U1 from the research conducted by Jog & Riding (1987), stated in chapter 2, the average IPO underpricing is 9.33 percent. To check for the validity of hypotheses one, the two means of Um and

(19)

Uc are compared by using a t-test. The t-value equals (0.972-0.093)/1.630 = 0.497. The outcome shows no sign of significance. The reason of the lacking significance is the fact that the standard deviation is very high. Potential effects causing this are the low number of observations or a big variance of the residuals.

Table 2: Descriptive statistics

Another chart, about the proceeds, shows a typical characteristic for these stocks, namely low

proceeds. In the chart below the average amount of proceeds per year of the sample is shown. Most of these stocks are penny stocks, which are stocks with an offer price below five dollars. As is shown above, fifty percent of the observations have an offer price below five dollars, but there are a few exceptions. The chart doesn’t contain enough observations to conclude anything about the trend of the proceeds, but the peak in the last two years could be due to the effect of the upcoming legalization.

Variable N Minimum Maximum Sum Mean Standard

Deviation Median Underpricing 39 -0.800 7 37.903 0.972 1.630 0.400 Real values - - - - Offer price 39 0.050 12 41.253 1.058 2.799 0.200 Offer size 39 1,000,000 32,900,000 214,002,780 5,487,250.770 5,837,661.179 410,000 Proceeds 39 200,000 100700000 284945420 7306292.82 20980225.73 600,000 Ex post 20 days volatility 39 0 0.475 2.683 0.069 0.095 0.040

Days till listing 39 10 611 4823 123.670 108.858 112

Age upon announcement 39 0 25 114 2.920 4.726 2 Cultivation 39 0 1 25 0.64 0.486 1 Retail 39 0 1 24 0.62 0.493 1 CSE listed 39 0 1 17 0.44 0.502 0 Natural logs - - - - LN_PRO 39 12.210 18.430 536.180 13.748 1.659 13.305 LN_OP 39 -3 2.48 -54.210 -1.390 1.335 -1.609 LN_OS 39 13.820 17.310 590.400 15.138 0.865 15.227 LN_AGE 39 0.000 3.260 38.920 0.998 0.782 1.099 Valid N 39 - - - -

(20)

Figure 4: Annual average proceeds of sample, in millions

4.2 Regressions

The regressions, which are all included in the appendix, are summarized in the table on the following page.

4.2.1. Regressions 1 & 2

Regression one and two are the small regressions and only make use of the general variables that most of the existing literature uses. The difference between the first two regressions is that regression one has the additional variable ‘offer price’. To decide between the natural logarithm and the real values for offer size a test for normality was performed. The results conclude that the variable is normally distributed without taking the natural log. The literature seems to be divided about this, thus the real value is used. For both the regressions the variable LN_OP acts according to the literature. The effect of offer price on underpricing in this sample is negative, because the coefficient is negative. The variable is significant in both the regressions, but grows in terms of significance when the variable ‘Offer size’ is excluded. Offer size might be a biased estimator that raises the standard errors of the other variables.

(21)

Table 3: Regression results

The table shows all the used variables in every regression. LN_OP is the natural log of the offer price. Offer size is the IPOs offer size. LN_PRO is the natural log of the proceeds. Ex post 20 days volatility is the share price’s averaged standard deviation for the first twenty trading days. DTL is the difference between announcement day and listing day in days. LN_AGE is the natural log of the companies’ age at listing day. CSE is a dummy variable for CSE listed companies. Retail and cultivation are dummy variables for the specific stated licenses. Coefficients with *,** and *** are respectively significant at a ten percent, five percent and one percent significance level.

Variable _{Regression 1} _{Regression 2} _{Regression 3} _{Regression 4}

Constant -8.573 (7.957) -5.587 (4.388) -8.985 (4.824) -9.549 (8.388) LN_OP -1.248* (0.619) -1.021*** (0.359) -1.442*** (0.427) -1.482** (0.647)

Offer size -3.681E-8

(0.000) - - -7.235E-9 (0.000) LN_PRO 0.513 (0.561) 0.303 (0.310) 0.507 (0.328) 0.547 (0.581)

Ex post 20 days volatility 6.987 (4.472) 7.208 (4.391) 8.195* (4.359) 8.130* (4.502) DTL 0.005* (0.002) 0.005** (0.002) 0.005** (0.002) 0.005* (0.002) LN_AGE -0.110 (0.332) -0.145 (0.319) -0.130 (0.316) -0.123 (0.333) CSE - - -0.814 (0.510) -0.801 (0.542) Retail - - -0.435 (0.723) -0.416 (0.770) Cultivation - - 0.802 (0.801) 0.794 (0.819) Adjusted R square 0.252 0.270 0.297 0.273

The variable LN_PRO, the natural log of the proceeds, shows unexpected behavior in both regressions. It was expected to be negatively correlated with underpricing, but the coefficients are positive and aren’t significant in either of the regressions. The variable ‘ex post 20 days volatility’ does act according to theory. The variable seems to be positively correlated with underpricing, which means that the more volatile the stock is in the first twenty business days the more underpricing is expected. This is in compliance with the literature stated in chapter two. However, the variable doesn’t show any signs of significance in these regressions in contrary to the expectations. The reason that the variable is insignificant is due to the large standard error, but the underlying reason for this has yet to be discovered. The next variable is DTL, the difference between announcement day and listing day in days, and follows the literature. In this sample the variable effects underpricing in a positive way,

(22)

which means that the bigger the DTL was the more the IPO was underpriced. The same applies for this variable as for LN_OP, the natural log of the offer price, in terms of significance. The exclusion of offer price seems to increase both these variables significance. The last variable in these regressions is the variable LN_AGE, which is the natural log of the companies’ age in years at listing day. The variable’s relation to underpricing is consent with the existing literature, because an increase in a company’s age upon announcement leads to a decrease in underpricing in this regression. However, the variable shows no signs of significance. One of the reasons for this is, because the mean of the variable age is even less than three years and the median equals two. This means that fifty percent of the observations had an age upon announcement of less than three years. Existing literature explains that the longer a firm exists, the more well-known the firm is to the investors, which leads to lower levels of underpricing. In this sample the variable might show the same behavior, but most of the observations seem to be young companies at the time of the announcement. Thus, the lack of

significance could be due to the fact that most of the companies were quite young upon announcement and the difference in terms of reputation for a one-year old company isn’t very different from the reputation of a firm that’s three years old. A third small regression was performed, which included the natural logarithm of offer size, but this seemed to decrease the significance of the other variables and even led to the exclusion of LN_PRO. Excluding the variable offer size from the first regression increased the adjusted R square of the new regression. This means that regression two explains the variance in the model better than regression one.

4.2.2. Regressions 3 & 4

The last two regressions included the dummy variables retail, cultivation and CSE, but regression four also includes offer size to see if the variable might be significant in this regression or changes the significance of other variables. The offer price still shows the same expected outcome and is significant in both regressions. The variable LN_PRO isn’t significant and still shows inversed outcomes in relation to the theory. The variable ‘ex post 20 days volatility’ is significant in both regressions in contrast to the first two regressions where it was insignificant. The standard errors remained approximately the same, but the mean increased which led to the shift in significance. The variable DTL remains significant in both regressions and didn’t change in any way. The variable LN_AGE remains insignificant due to the effect described in the paragraph above and possible other causes. All the added dummy variables aren’t significant, but cause an increase in the adjusted R square, which filters out the effect of an increasing R square due to a simple increase in variables, thus the variables do have some explanatory value.

(23)

the exchange’s name recognition for listing relatively smaller firms3

. However, if you look at the variables offer price and proceeds in the descriptive table it can be concluded, from their medians, that most of the companies in the sample are small cap firms. However, only forty-four percent of the observations are listed on the CSE. According to the descriptive statistics table this number should at least equal the fiftieth percentile of the observations, thus this could be a reason for the insignificance. Retail, the dummy for if a company is licensed for the retail of marijuana, is negatively correlated to underpricing in the regression and thus decreases the level of underpricing when a company has a license for the retail of marijuana and is in compliance with the hypotheses. However, the effect is not significant in either of the regressions.

This also relates to the variable cultivation, however the sign of this variable in inversed. The prediction states that IPOs of companies with a cultivation license are less underpriced. One explanation for the insignificance of these variables lies in the fact that some of companies were granted a license after the IPO. Thus, the effect of decreasing uncertainty emerged later on for some companies. Adding the variable ‘offer size’ has the same effect on the big regressions as it has on the small regressions, namely the increase of the other variables standard error, which leads to less significance of the variables DTL and LN_OP. The adjusted R square increased when the dummy variables were added in regression 3 and this R square is designed to filter out the effect of adding new variables to increase the R square. However, none of the added variables were significant and the increase in R square is just over two percent compared to regression two, thus it is hard to assume that regression three is significantly better explaining the underpricing than regression two is.

(24)

http://thecse.com/en/trading/market-activity/cse-CHAPTER 5: CONCLUSION

In this last chapter of the thesis the final conclusion is summarized and the research question is answered. The conclusion relates to the stated hypotheses in chapter three. The second part of this chapter reviews the limitations of this research and includes recommendations for further research.

5.1 Conclusion

This thesis examined the short-term IPO performance of Canadian-listed marijuana companies. The research question was as follows:

“Is underpricing of IPOs of the current Canadian exchange venture listed companies, which produce or sell marijuana or make consumer devices for the purpose of smoking marijuana, higher than the average Canadian IPO underpricing? Can this be explained by existing literature?”

First of all, the average underpricing for the marijuana companies is calculated, which is compared to the average Canadian IPO underpricing, that was found in existing literature. The average

underpricing in the Canadian marijuana sector equals roughly ninety-seven percent, in comparison to the averaged ten percent underpricing, which is stated in a research from Jog and Riding (1987). The underpricing in the marijuana sector seems to be significantly higher, but the high standard deviation of underpricing in the regression accounts for an insignificant result. Thus, the first null-hypotheses, underpricing in the marijuana sector is equal to Canada’s average level of IPO underpricing, can’t be rejected at a ten percent significance level.

The second part of the thesis is about the explanatory variables for underpricing in the Canadian marijuana sector. The regression that is stated in chapter 3 is regression model four, however, the variable offer size, which is the total amount of offered shares in the IPO, seems to increase the standard errors of the other variables, thus it is left out. This makes regression three the main regression to explain underpricing, which is also the regression with the highest adjusted R square.

In none of the regressions are the variables ‘offer size’, LN_AGE, which is the natural log of the companies’ age in years at announcement day, LN_PRO, which is the natural log of the proceeds, CSE, which is a dummy variable for CSE-listed companies, retail, which is a dummy variable for if the company is licensed for the retail of marijuana and cultivation, which is the dummy variable for cultivation licensed companies, significant. This leads to the fact that null-hypothesis three, which tests for the significance of ‘offer size’, null-hypothesis four, which tests for the significance of

(25)

for the significance a retail license, null-hypothesis nine, which tests for the significance of cultivation licenses, and null-hypothesis ten, which tests for the significance of the exchange venture where the companies are listed, can’t be rejected.

The insignificance of age is due to the fact that the average and median of this variable are both under three years, and the effect of reputation significance for a one year old company compared to a three year old company won’t be very different. Retail and CSE did show the expected effect on

underpricing, except the effects are not significant. The effects of proceeds and cultivation are inversed compared to the literature, and insignificant.

The most important explanatory variables for the underpricing of IPOs of Canadian maiijuana companies are LN_OP, which is the natural log of the offer price, and DTL, which is the amount of days between announcement and listing day. In regression three they even showed significant values at a significance level of respectively one percent and five percent. The variable ‘ex post 20 day volatility’, acts according to literature, but seemed to show no sign of significance in the first two regressions. However, this changed in regression three and regression four and is significant at a ten percent significance level. This leads to the following conclusion, using regression three; the null-hypothesis number two, which tests for the significance of offer price, can be rejected at a one percent significance level, the null-hypotheses number five, which tests for the significance of DTL, can be rejected at a five percent significance level and the null-hypotheses seven, which tests the significance of ‘ex post volatility’, can be rejected at a ten percent level.

5.2 Limitations

Several problems were encountered during the process of writing this thesis, and this led to some limitations of this research. Most of the limitations were due to fact of insufficient time, because of the late change of subjects. The biggest limitation of this research is the absence of a peer group for underpricing comparison. Underpricing of this research is compared to the sample of another research, but this research was conducted in another time span and focused on different sized companies. The peer group should have existed of non-marijuana companies that are equal to the sample in all other factors. Another limitation was the limited amount of available data, which led to the loss half the initial observations. This also led to the fact that some variables were ruled out in advance, like ‘market capitalization on day one’ and ‘underwriter rank’, because most of this data is lacking in the consulted databases.

5.3 Recommendations

(26)

years, when the data is more widely available and more complete. This way the number of valid observations will increase, which will lead to a decrease of the standard errors of the estimators. This thesis only focused on the short-term IPO performance, thus an additional research on the IPO’s long-term performance should be done. Another option is to compare the IPO performance with the current performance of the companies, as the total marijuana market increased considerably last year. A last recommendation is about the CSE variable, the dummy variable that equaled one if the company is listed on the CSE. A research could be done that includes all the IPOs in Canada, and compares the CSE-listed companies IPO performance versus the TSX-listed companies IPO performance.

(27)

REFERENCES

Akerlof, G. (1970). The Market for "Lemons": Quality Uncertainty and the Market Mechanism. The

Quarterly Journal of Economics, 84(3), 488-500.

Alexander, D & Skerritt, J. (2017, December 28). Canada's medical marijuana producers are pushing to take over the world. The Financial Post. Retrieved from http://business.financialpost.com Beatty, R., & Ritter, J. R. (1986). Investment Banking, Reputation, and the Underpricing of Initial

Public Offerings. Journal of Financial Economics, 15(1-2), 213-232.

Brealey, R. A., Marcus, A. J., & Myers, S. C. (2001). Fundamentals of Corporate Finance. Retrieved from http://web.mit.edu/voila/www/trashbin/ForFun/corporatefinance.pdf

Clarkson, P. M. & J. Merkley (1994), Ex Ante Uncertainty and the Underpricing of Initial Public Offerings: Further Canadian Evidence, Canadian Journal of Administrative Sciences, 2(1), 54-67.

Daily, C. M., Certo, S. T., Dalton, D. R., & Roengpitya, R. (2003). IPO Underpricing: A

MetaAnalysis and Research Synthesis. Entrepreneurship: Theory & Practice, 27(3), 271-295. Draho, J. (1971) The IPO Decision: Why and how Companies Go Public. Retrieved from

https://books.google.nl/books?hl=nl&lr=&id=Uil5fLFVl3IC&oi=fnd&pg=PR6&dq=The+IPO

+Decision:+Why+and+how+Companies+Go+Public+&ots=1-J-3U_ZK5&sig=0ovcqEv8Nudkx0DBVIZDrTAz080#v=onepage&q=The%20IPO%20Decision %3A%20Why%20and%20how%20Companies%20Go%20Public&f=false

Eckbo, B. E. (2007). Handbook of Empirical Corporate Finance SET. Retrieved from

https://books.google.nl/books?hl=nl&lr=&id=l9CTEAZ0IP4C&oi=fnd&pg=PA375&dq=unde rpricing+hot+or+cold+market+venture+capital+proceeds+first+20+days&ots=mhhBWbM7Al &sig=ls3vFI4u2LD60105urfOtBbWflE#v=onepage&q&f=false

Habib, M.A., & Ljungqvist, A.P. (2001). Underpricing and Entrepreneurial Wealth Losses in IPOs: Theory and Evidence. Review of the Financial Studies, 14(2), 433-458.

Ibbotson, R., & Jaffe, J. (1975). "Hot Issue" Markets. The Journal of Finance, 30(4), 1027-1042. Jog, V., & Riding, A. (1987). Underpricing in Canadian IPOs. Financial Analysts Journal, 43(6), 48

55.

Jog, V., & Wang, L. (2002). Aftermarket Volatility and Underpricing of Canadian Initial Public Offerings, Canadian Journal of Administrative Sciences, volume 19(3), 231-248.

Jenkinson, T., & Ljungqvist, A. (2001) Going Public: The Theory and Evidence on how Companies

Raise Equity Finance. Retrieved from

https://books.google.nl/books?hl=nl&lr=&id=l_xdnKxeYIkC&oi=fnd&pg=PA3&dq=Going+ Public:+The+Theory+and+Evidence+on+how+Companies+Raise+Equity+Finance&ots=9N6 QsSdpWA&sig=PJ14pAwBE9IMNs2FEpVCeYwjl5g#v=onepage&q=Going%20Public%3A

(28)

%20The%20Theory%20and%20Evidence%20on%20how%20Companies%20Raise%20Equit y%20Finance&f=false

Kim, M., & Ritter, J. (1999). Valuing IPOs. Journal of Financial Economics, 53(3), 409-437. Leone, A., Rock, S. & Willenborg, M. (2007). Disclosure of Intended Use of Proceeds and

Underpricing in Initial Public Offerings. Journal of Accounting Research, 45(1), 111-153. Ljungqvist, A., Nanda, V., & Singh, R. (2006). Hot Markets, Investor Sentiment, and IPO Pricing. The

Journal of Business, 79(4), 1667-1702.

Loughran, T., & Ritter, J. (2004). Why Has IPO Underpricing Changed over Time? Financial

Management, 33(3), 5-37.

Ritter, J. R. (1984). Signaling and the Valuation of Unseasoned New Issues: A Comment. The Journal

of Finance, 39(4), 1231-1237.

Ritter, J. R. (1991), The Long-Run Performance of Initial Public Offerings, The Journal of Finance, 46(1), 3-27.

Ritter, J. R., & Welch, I. (2002). A Review of IPO Activity, Pricing, and Allocations. The Journal

of Finance, 57(4), 1795-1828.

Rock, K. (1986). Why new Issues Are Underpriced. Journal of Financial Economics, 15(1-2), 187-212.

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), 1935-1974.

Databases:

Bloomberg Terminals, consulted at Erasmus Rotterdam Datastream, consulted at the University of Amsterdam Sedar, the official Canadian filing site

(29)

APPENDIX INVESTOR SENTIMENT

The Y-axis represents the value of interest in a specific time period (100 being the highest value), the X-axis rapresents the date.

(30)

(31)

(32)

(33)

(34)

(35)