Are the profits in the pharmaceutical industry too large, as is often claimed?

Are the profits in the pharmaceutical industry

too large, as is often claimed?

Mo Leo

*

14 July 2016

Faculty of Economics and Business

Bachelor Thesis

Supervisor: Dr. András Kiss

*

_{Student BSc Economics and Business at the University of Amsterdam, specialization Economics and Finance.}

Student number: 10526595.

Statement of Originality

This document is written by Student Mo Leo who declares to take full responsibility for the contents

of this document.

I declare that the text and the work presented in this document is original and that no sources other

than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the

work, not for the contents.

Table of Contents

1 INTRODUCTION ... 4

2 LITERATURE REVIEW ... 5

2.1

Capital Asset Pricing Model ... 5

2.1.1

Previously conducted studies on Jensen’s alpha ... 7

2.2

Determinants of risk in the pharmaceutical sector ... 7

2.3

Previously conducted studies on R&D in the pharmaceutical sector ... 8

3 HYPOTHESIS ... 9

4 METHODOLOGY ... 10

4.1

OLS regression ... 10

4.2

Historical revenue and net income ... 12

5 RESULTS ... 14

5.1

OLS regression ... 14

5.1.1

Examining the alphas ... 14

5.1.2

Examining the betas ... 16

5.2

Historical revenue and net income ... 17

6 CONCLUSIONS ... 20

7 REFERENCES ... 21

8 APPENDICES ... 24

Appendix 1 – Revenue and net income growth rates ... 24

Appendix 2 – OLS regression output ... 25

Abstract

The purpose of this paper is to examine whether profits in the pharmaceutical industry are

excessive relative to the risk pharmaceuticals experience. First, it is examined if stock

returns in this particular industry are higher than expected with respect to the risk involved

by performing an OLS-regression using the Capital Asset Pricing Model. Second, it is

investigated whether historical returns and net income are volatile and thereby impose risk

on pharmaceutical firms. With respect to the OLS-regression, the overall result is that it

cannot be concluded that historical stock returns are significantly higher than expected in

relation to the risks involved. With respect to the historical growth rates, it appears that for

return as well as for net income, both growth rates are volatile. But, given that the

pharmaceuticals’ profits remain substantial, the study does not show any support for the

claim that there are serious risks present.

1 Introduction

For more than 50 years, there has been a debate about the remarkably high profits of the

pharmaceutical industry (Scherer, 2004). It is known that between 1988 and 2009, the pharmaceutical

industry gained profits that were 3 to 37 times as large as the all-industry average profits (Spitz and

Wickham, 2012). According to Angell (2004), these remarkably high profits are the result of

excessively high prices. Many people find these excessively high prices a major issue since such high

prices pose a real problem for hospitals, as their budget is not always sufficiently large to pay for all

needed medications and treatments (Tax and Van der Hoeven, 2014).

Paul Ginsburg (2008) showed that between 1960 and 2006, total medical costs rose on average

with 9.9 percent per year in the U.S. According to Spitz and Wickham (2012), the U.S. accounts for

more than 45 percent of total sales in the pharmaceutical industry, hence the U.S. represents a large

share of the market. This increase in national spending occurred while GDP rose on average with 7.3

percent per year. When adjusting these percentages for inflation, it appears that the trend in health care

spending is more than two times as large as the GDP trend between 1960 and 2006 in the U.S.

(Ginsburg, 2008). This finding obviously raises the question whether this gap is sustainable and

whether it does not come at the expense of the patient. The fact is that health care spending is

increasing and pharmaceutical firms continue to enjoy substantial profits compared to other

comparable firms. According to Spitz and Wickham (2012) this creates welfare loss since

pharmaceuticals will innovate less due to lack of competition and because the profits in this particular

industry are substantial and sustained.

In order to fully investigate whether profits in the pharmaceutical industry are too large, it is

essential to clarify what is considered as too large. The main argument put forward by pharmaceuticals

to the question why their profits are much higher than the all-industry average is that pharmaceuticals

face a lot of risk due to high investment costs (Toll, 2004). Therefore, this paper will primarily focus

on the risk-return relationship. According to Ho, Xu and Yap (2004), R&D-intensive firms’ stocks

carry greater systematic risk than stocks of firms that conduct less R&D. Since the pharmaceutical

sector is by far the greatest research-intensive industry of all sectors that support their R&D with

private funding, risk is likely to play an important role in this sector (Scherer, 2004). In the

pharmaceutical industry, the main risks arise from scientific and regulatory uncertainties as both

uncertainties make the development of a new medicine a long process and will thereby result in

economic uncertainty. The longer the process of the scientific development takes, the greater the

chance that a competitor will come up with the same finding, and thereby strongly reduce the

opportunity for a return on the investments made. The issue with regulatory uncertainty is that it slows

the process of developing a new medicine even more since a firm is not allowed to implement product

marketing as long as there is no approval. The relevant time period causing risk is from the moment

the IND application is submitted until the NDA approval or abandonment of research is given.

Together, the scientific and regulatory time span account for the major part of the long duration of

medicine development (Dickson and Gagnon, 2004).

Apparently there are various risks involved in this industry. Though, the real question is whether

these risks are substantial enough and thereby justify the excessive returns that are gained in the

pharmaceutical industry, given that many firms reject the accusation of earning excessive returns by

stating that pharmaceutical firms experience many risks. For example, when it would often happen

that much research eventually yields nothing, it could be justified that when research does deliver it

has to cover the other costs that have been made. Therefore the research question of this paper is: Are

the profits in the pharmaceutical industry on average too large with respect to the risk the firms are

subjected to?

2 Literature Review

This paper will focus on the risk-return relationship of pharmaceutical companies. An economic

model that describes the relationship between risk and return is the Capital Asset Pricing Model,

CAPM. According to this model, riskier assets go together with a higher expected return to give

investors an incentive to invest in them (Chen, 2003). Therefore, this section will first discuss the

basic theory of this model. Subsequently it will focus on the determinants of risk in the pharmaceutical

sector. Lastly, an overview is given consisting of previous studies that have been done regarding risk

in the pharmaceutical industry.

2.1

Capital Asset Pricing Model

The Capital Asset Pricing Model is an economic model that can provide the required rate of return an

investment needs to yield, to be acceptable for investors (Bodie, Kane and Marcus, 2011). The basic

version of the CAPM makes use of the following simplifying assumptions:

1. There are many investors who are all price takers and can lend and borrow unlimited.

2. All investors have the same holding period.

4. There are no taxes or transaction costs included in the model.

5. All investors are rational mean-variance optimizers.

6. All investors have homogenous expectations and all investors are in complete agreement.

These assumptions are clearly restrictive, which makes it a less realistic description of the real world.

However, it can serve as a useful benchmark model and give some insights in the securities market.

Figure 1 shows the relationship between expected return and risk according to CAPM. The

horizontal axis represents portfolio risk, measured by the standard deviation of portfolio return. The

vertical axis represents portfolio expected return. As an assets’ individual risk can be diversified in a

portfolio containing risky assets, investors can minimize the risk of the expected return. The minimum

variance frontier shows combinations of the expected return and the risk of a portfolio, when having a

portfolio consisting only of risky assets. When risk-free borrowing and lending is included, new

possible combinations of expected return in relation to risk arise. The Capital Market Line, the straight

line that is tangent to the minimum variance frontier, shows this. This line is the best attainable capital

allocation line (Bodie, Kane and Marcus, 2011).

Figure 1. Fama and French, 2004

Since investors will hold the same risky portfolio with unlimited risk-free lending and borrowing, all

investors will hold the same risky portfolio T. The CAPM assumptions also imply that the market

portfolio M must lie on the minimum variance frontier. All these assumptions result in the familiar

Sharpe-Lintner CAPM equation (Fama and French, 2004):

!(#

$

) = #

'

+ *

$+

!(#

+

) − #

'

,

. = 1, … , 1.

In this equation, E(Ri) is the expected return of asset i, E(RM) is the expected market return, Rf is the

risk-free rate which is uncorrelated with the market return and *

$+

is the asset’s market beta which

to influence the required return (Bodie, Kane and Marcus, 2011). The market beta of asset i is the

covariance of the market return and the return of i, divided by the variance of the market return.

Therefore, beta is often referred to as the measure of the systematic risk involved (Gençay, Selçuk,

and Whitcher, 2003).

The stock’s alpha is defined as follows (Bauer, Koedijk and Otten, 2005):

3

$

= #

'

+ *

$+

!(#

+

) − #

'

− #

$

,

. = 1, … , 1.

As shown in the formula above, 3 is the difference between the expected return according to CAPM

and the actual return of asset i. If the actual return is higher than the expected return, 3 is positive

(Bodie, Kane and Marcus, 2011). Typically 3 is seen as a measure for out- or under-performance with

respect to market performance (Bauer, Koedijk and Otten, 2005).

Altogether, CAPM is based on unrealistic assumptions such as complete agreement along

investors and the possibility of unlimited borrowing and lending. However, many economic models

rely on unrealistic simplifications and provide nevertheless useful information (Fama and French,

2004).

2.1.1 Previously conducted studies on Jensen’s alpha

In a study performed by Kim, Matilla, and Gu (2002) it was investigated whether hotel real estate

investment trusts outperformed the market between 1993 and 1999, using the Capital Asset Pricing

Model. A regression was carried out using monthly returns, measured by the percentage change in

stock price plus dividend yield. It was found that five out of nine portfolios had no significant alpha

and hence didn’t outperform the market.

In a bachelor thesis, Van Ballegooijen (2015) investigated whether the S&P Global Luxury

Index outperformed the MSCI World Index during the financial crisis, using the Capital Asset Pricing

Model. An OLS-regression was carried out using stock price returns of 67 listed companies from 2007

until 2009. When using daily stock returns, no significant alphas were found. However, when using

monthly stock returns to reduce potential noise caused by daily stock returns, some alphas appeared to

be significant. But, nevertheless, using a monthly interval did not provide enough empirical evidence

to show outperformance.

Both studies focused on sectors with remarkable growth, and, for this reason, for both studies

it was expected that the particular sectors would outperform the market. The results show that this is

apparently not always the case.

2.2

Determinants of risk in the pharmaceutical sector

As mentioned in the Introduction, pharmaceuticals claim to be a high risk business due to high

investment costs. Pharmaceuticals generally devote significant amounts of funds into R&D. In 2011,

the industry spent around 10-15 percent of total revenues on R&D (OECD, 2015). This may indeed

entail risks when it is uncertain whether a significant investment will eventually be regained.

Furthermore, R&D activities do not only bring financial risks with it. R&D activities demand

time-consuming clinical trials as well that could worsen a company’s competitiveness (Giaccotto, Santerre,

and Vernon, 2005).

To investigate pharmaceutical supply chain risks, Jaberidoost, Nikfar, Abdollahiasl, and

Dinarvand (2013, p. 69) examined nine articles with regard to pharmaceutical supply chain risks

which they presented it in a systematic review. They stated that:

Most of reported risks were related to supply and supplier issues. Organization and strategy

issues, financial, logistic, political, market and regulatory issues were in the next level of

importance. […] It was shown that the majority of risks in pharmaceutical supply chain were

internal risks due to processes, people and functions mismanagement which could be managed

by suitable mitigation strategies.

Another possible risk for pharmaceutical companies comes from the threat of price regulation.

In several countries price regulation is already applied (Vernon, 2005). The U.S.’ pharmaceutical

industry is largely unregulated with respect to price (Giaccotto et al., 2005). To investigate what the

impact of political price regulation in the pharmaceutical industry would be, Giaccotto et al. (2005)

performed a multiple regression where drug prices’ growth rates between 1980 and 2001 were

restricted to the overall consumer price index growth rate in the U.S. Giaccotto et al. (2005) concluded

that if the U.S. would have implemented price regulation in the pharmaceutical industry, R&D

investment would have been 30 percent lower. They also estimated that this decline in R&D

investment would result in a decrease of at least one-third newly developed medicines in that time

period. This finding is particular interesting since one would expect that consumers would always

benefit from lower prices. However, when the pharmaceutical industry creates less medicines, then

this is, of course, also to the detriment of the consumer. Though it is questionable whether a less

restrictive price regulation would also result in significant drops in R&D spending and medicine

launches.

2.3

Previously conducted studies on R&D in the pharmaceutical sector

According to Dimasi (1991) the discovery and development of new drugs is a very long and costly

process. In order to estimate the risk that pharmaceutical companies face when developing new drugs,

a more recent study conducted by DiMasi, Grabowski, and Hansen was performed in which they

collected data from 1442 different compounds originating from top 50 firms in the pharmaceutical

industry. They found that the overall probability that a drug that enters clinical testing ultimately got

approved was 11.83 percent (DiMasi, Grabowski, and Hansen, 2016). Therefore, 88.17 percent of the

compounds does not pass clinical testing. However, although more than 88 percent of developed drugs

does not pass clinical testing, the development costs of these drugs should not be seen as a wasted

investment. In general, the knowledge acquired during the whole development process may have a

huge value that can be utilized for the development of other drugs. The outcome of 11.83 percent is

much lower than the probability of 20.5 percent the researchers estimated in their previous study in

2003. Dimasi et al. (2016) believe the failure rates could have increased due to several changes over

the years. Firstly, regulators have become more risk averse and therefore act more cautious. Another

reason could be that during the past few years the pharmaceutical industry focused more on areas

where the science is lagging behind and therefore failure risk increased. Last, they believe that the

growth in identified drug targets, many of which may be poorly validated, may have encouraged

pharmaceuticals to pursue the development of more compounds with a large uncertainty about their

success.

According to DiMasi, Hansen, and Grabowski (2003) the estimated drug development cost per

new drug between 1990 and 2001 amounts to be $802 million in 2000 dollars, which is around $1044

2013 dollars. This estimation also includes R&D costs of drugs that ultimately were not sold on the

market and the opportunity costs associated with R&D capital. They also concluded that the

development of a new compound takes on average 16 years before the chemical entity is sold on the

market. In a more recent study, the estimated average drug development cost per new drug between

2000 and 2010 amounts to be around $2558 million in 2013 dollars (DiMasi, Grabowski, and Hansen,

2016). As mentioned before, 88.17 percent of all compounds does not enter clinical testing. When

looking at these significant amounts of money needed for the development of a new medicine, it

becomes clear that a lot of money is invested while in fact only a small fraction of the developed

medicines will be sold on the market.

3 Hypothesis

As mentioned in the Introduction, the pharmaceutical industry enjoys profits that are substantial and

sustained. At the same time, as was discussed in Section 2.3, the R&D investments of the

pharmaceutical industry is very high and more than 88% of the drugs developed does not pass clinical

testing. However, this does not necessarily mean that the risks are truly a large threat, since the profits

are substantial and sustained. The hypothesis to be tested in this paper is twofold:

1. The profits in this industry are higher than expected with respect to the risk involved in this

industry.

4 Methodology

As mentioned in the Introduction, the pharmaceutical industry gained profits that were remarkably

high compared to the industry average profits. In addition, a lot of money is put into research, while

only a small fraction of the drugs developed ultimately will be sold on the market. Since

pharmaceuticals in general reject the accusation of earning excessive returns by stating they

experience severe risks, it is interesting to investigate if the risks are indeed large enough. This will be

done by two different approaches. Section 4.1 covers the OLS regression in which stock returns from

multiple pharmaceutical companies over a longer time period are examined to test whether historical

stock returns are significantly higher than expected with respect to the exposed risk. Section 4.2 covers

an analyses of the historical growth rates in both revenue and net income from multiple

pharmaceutical companies where, in particular, potential fluctuations can be interesting from a

risk-perspective.

4.1

OLS regression

In Section 2.1, the Capital Asset Pricing Model has been outlined. This model is particularly

appropriate for testing this paper’s first hypothesis as this model concerns the relationship between

risk and return.

To find out if the returns are indeed excessive, given their risk, daily stock prices from 56

listed U.S. pharmaceuticals from January 1994 until January 2014 will be used in an OLS regression.

These pharmaceuticals are selected based on the fact that all companies operate in the U.S. and that

their stock prices are available from 1994 until 2014. The firms differ considerably in size. This is not

surprising since the top ten pharmaceutical firms globally accounted for 46 percent of global sales in

2006 (IMS Health, 2007). The differences in size can be reflected in several features. Since firms’ net

income can be influenced by acquisitions or other external factors, it can give a distorted view.

Therefore, not the net income but the revenues of this dataset are considered. The revenues for the

year 2014 are displayed in Figure 2.

1

The percentages represent the companies’ revenue share with

respect to the total revenue of this dataset.

2

1

_{Retrieved from}

_{https://ycharts.com}

_;

_{http://www.nasdaq.com}

_;

_{http://finance.yahoo.com}

_.

2

_{For Pharmos, Fertil-A-Chron, Regenerx Biopharmaceuticals, Unigene Laboratories, Heron Therapeutics, Emisphere}

technologies, USA Equities, Protein Polymer Technologies and Protide Pharmaceuticals, 2014 revenue data is not available.

Therefore, these nine pharmaceuticals are excluded from the diagram.

25%$ 17%$ 14%$ 8%$ 7%$ 7%$ 7%$ 5%$ 3%$ 3%$ 2%$ 1%$ 1%$ JOHNSON$&$JOHNSON$ PFIZER$$$$$ MERCK$&$COMPANY$$ GILEAD$SCIENCES$$ ABBOTT$LABORATORIES$$ AMGEN$$ ELI$LILLY$ BRISTOL$MYERS$SQUIBB$ BIOGEN$ MYLAN$$$ PERRIGO$$ REGENERON$PHARMACEUTICALS$$ The$other$35$companies$

Figure 2. Dataset revenue share in 2014

The diagram shows that twelve pharmaceuticals in this dataset outperform the other 35 companies

significantly with respect to revenue. The revenues of these twelve outperforming pharmaceuticals

range from $2.8 billion to $74.3 billion in 2014. As a result, these companies are all included in the

S&P 500 index in 2014.

3

Seven of the outperforming pharmaceuticals displayed in Figure 2 belonged to the ten largest

U.S. pharmaceuticals in 2002 that are discussed in a paper written by Pattison and Warren (2002). In

their paper, these ten companies were criticized for their high earnings compared to other industries.

The other three pharmaceuticals criticized in 2002 by Pattison and Warren are not part of this papers’

dataset since they were either acquired by other pharmaceutical companies or merged.

The overlap between the companies in the paper from 2002 and the largest companies in

Figure 2 shows that pharmaceuticals consistently outperform the other companies with respect to the

markets’ revenue share.

Furthermore, it is remarkable that only one percent of total revenues is coming from 35 listed

U.S. companies. Among these 35 companies, large differences in revenue exist. Their revenues range

from $0.1 billion to $618.79 million. These noticeable differences in firm size makes it a varied

dataset.

The daily stock prices from these 56 pharmaceuticals are retrieved from Datastream. First, the

daily stock prices are converted into daily returns. To estimate the market rate, it is common to use

data from the S&P 500 index (Berk & DeMarzo, 2014). The S&P500 index is a stock portfolio

including data from the 500 largest companies of the U.S. Therefore S&P500 data will be used to

estimate the market return. The risk-free rate is retrieved from the Federal Reserve Economic Data,

where 10 year treasury bond rates were obtained.

4

_{In order to use the risk free rates in the OLS}

regression, the rates have to be converted into daily rates. This was done through the following

formula:

(1 + #

'

)

4 567

_{− 1}

The OLS regression will be performed in order to figure out whether alpha, as mentioned in Section

2.1, has been significantly positive between 1994 and 2014. If this is the case, it would imply that the

actual return of the specific company has been higher than expected. The estimates for each companies’

alpha will be found by using a simple OLS regression in its usual form:

8

$

= *

7

+ *

4

9

$

+ :

;

In this form, 8

$

represents the actual daily return of asset i subtracted with the daily risk free rate #

'

,

*

7

represents alpha, *

4

represents the asset’ market beta, the independent variable 9

$

represents the

excess market return which is obtained by subtracting the daily risk-free rate from the daily market

returns, and :

_;

is an error term. The regression is done on an annual basis. This means that for each

company i, 21 alphas and betas are found. After that, the output will tell if there are positive alphas

and whether they are significant. The significance of the alphas is based on the p-value. The p-value

can tell if the coefficient is significantly different from zero. The smaller the p-value, the more

significant the coefficient is.

In addition to the estimated alphas, it is interesting to examine the stability of firms’ betas over

the years. Stability or instability of the betas matters from a risk perspective since it reflects the

volatility of the involved systematic risk. This can be evaluated by looking at firms’ highest and

lowest beta between 1994 and 2014. Besides that, it might be interesting to investigate whether the

betas are similar across companies since similarity could imply that the whole pharmaceutical market

is equally sensitive to systematic risk. This can be tested by observing every firms’ beta for one

specific year. The OLS regression will be performed using STATA.

4.2

Historical revenue and net income

As mentioned in the Introduction, the main argument put forward by pharmaceuticals to the question

why their profits are much higher than the all-industry average, is that pharmaceuticals face a lot of

risk due to high investments costs. One would expect that high investments costs takes risk along with

it when it would cause fluctuations in a company’s earnings. When earnings would be completely

stable, high investment costs shouldn’t pose a real risk. To see whether earnings are stable or not,

historical revenue and net income data of various major pharmaceutical companies are collected and

displayed in a graph. Historical revenue and net income between 2000 and 2015 are collected for 10

different U.S. based pharmaceutical companies that are listed since at least 1986 and are still

operating. The data is retrieved from YCHARTS.

5

_{First, the historical revenues and net profits are}

converted into growth rates per year since growth rates provide a better insight when looking at

multiple firms. Subsequently, both the historical revenue growth rates and the net income growth rates

from the year 2001 until 2015 are displayed in two separate graphs. Displaying growth rates visually

should create a clearer view in terms of fluctuations.

5 Results

5.1

OLS regression

5.1.1 Examining the alphas

The OLS regression results can be found in Appendix 2.

6

Table 1 summarizes the results for the whole

time period with respect to the estimated alphas.

Table 1

Number of positive

alphas

Number of significant

alphas (

< = =%)

Number of significant

alphas (

< = ?@%)

1994

22

0

1

1995

51

10

11

1996

43

2

4

1997

37

0

0

1998

33

0

0

1999

42

5

8

2000

51

1

4

2001

42

0

1

2002

19

0

1

2003

46

2

5

2004

29

0

1

2005

25

0

0

2006

28

0

1

2007

42

6

11

2008

33

1

1

2009

40

0

2

2010

34

1

2

2011

30

1

1

2012

39

1

2

2013

49

8

15

2014

40

0

5

As seen from table 1, for most years the regression resulted in many positive alphas. However, very

few are significant. Of the nine pharmaceuticals with three or more significant alphas, five

pharmaceuticals are companies that outperform the others in this dataset with respect to revenue

earnings displayed in Figure 2.

7

Their significant alphas are summarized in Table 2.

Table 2

<

<

<

<

Bristol Myers Squibb

0.0013678

0.0041951

0.0017451

Eli Lilly

0.0018568

0.0044707

0.0010763

Biogen

0.0090528

0.0062844

0.0031595

0.0021079

Perrigo

0.0071328

0.0060674

0.0012977

Gilead Sciences

0.0049243

0.0039149

0.00238

As mentioned in Section 4.1, seven of the ten pharmaceutical companies criticized in a paper written

by Pattison and Warren (2002) are also considered in this study. Of these seven companies, Bristol

Meyer Squibb and Eli Lilly have three significant alphas as shown in Table 2. It is not clear what

meaning can be attached to this observation. First of all, it might be purely accidental that they are

significant, because there are quite a number of alphas in this study. Second, the significant alphas are

quite small. For Bristol Myers Squibb the average annual price return between 1994 and 2015 was

11.40 percent, and for Eli Lilly over the same period 13.53 percent.

8

_{The effect of the alphas found for}

these two companies, as displayed in Table 2, is therefore small as well.

Since very few alphas are significant, no conclusions can be drawn whether these 56 pharmaceutical

companies earned excessive returns according to their expected return estimated by CAPM. This is

contrary to what was expected. A possible explanation for this statistically insignificant result, could

be that only stock returns were taken into account. One could expect dividend distributions to temper

share value and therefore, by taking only share value in consideration a distorted picture of returns

could be the result. A more accurate picture could be obtained if dividend returns were also taken into

account. The study by Kim, Matilla, and Gu (2002) mentioned in Section 2.1.1. includes dividend

yields in their return measurement as well as stock price returns. Relatively, their result contains

considerably more significant alphas than the regression result in this paper. Therefore, it is worth

examining whether including dividend yields would result in more significant alphas.

Looking at daily stock prices may perhaps have contributed to mostly insignificant alphas since

the OLS method assumes normally distributed errors. Using monthly returns instead of daily returns

could reduce the noise coming from daily returns since monthly returns are at least approximately

normal distributed. But one should be cautious in concluding that it is better to use monthly returns.

According to Daves, Ehrhardt, and Kunkel (2000) their results show that daily returns provide a

smaller standard error of the estimated beta than monthly returns. The issue requires a more detailed

analysis than the scope of this paper permits.

7

_{Bristol Myers Squibb, Eli Lilly, Teligent, Celldex Therapeutics, Akorn, Biogen, Perrigo, Gilead Sciences, and Protide}

Pharmaceuticals have 3 or more significant alphas (3 = 10%).

5.1.2 Examining the betas

Besides alphas, betas are estimated for the used dataset between 1994 and 2014. The estimated betas

are presented in Appendix 2 as well. It appears that the number of significant betas is seven times as

large as the number of significant alphas, when 3 = 10%. To examine the stability of the betas over

the years, the largest and the smallest beta for the whole dataset is displayed in Figure 3.

9

Figure 3. Highest and lowest beta between 1994 - 2014

Figure 3 shows that the estimated betas were not stable between 1994 and 2014. The differences

between the highest and lowest beta are remarkably large for almost every pharmaceutical firm. To

find out whether significant betas would provide more stability, this selection is also performed using

only significant betas. Then, the differences between the highest and lowest beta are smaller, but still

not stable at all. A possible explanation for this instability is that several economic downturns have

taken place in this time period and therefore influenced the betas.

To see whether betas are similar across companies, the estimated betas for 2013 are displayed

in Figure 4. Nearly every company’s beta is located between 0 and 0.6. A similar beta across

companies should make sense since one would expect pharmaceutical firms to be equally sensitive to

systematic risk.

9

_{12 outliers have been removed from the data. These outliers were part of Abeona, Fertil-A-Chron, Peregrine}

Pharmaceuticals, Champions Oncology, USA Equities, and Protide Pharmaceuticals.

!2,5% !1,5% !0,5% 0,5% 1,5% 2,5% 3,5% 4,5% BRIS TOL%MY ERS%SQ UIBB% ABBO TT%LABS .% ELI%LILL Y%% JOHN SON% &%JO HNSO N% MERC K%&%CO MPAN Y%% PFIZE R%%%%% MYLA N%%% CYBRD I%% AVIRA GEN% THERA PEUT ICS% ALSE RES%P HARM. % POLY DEX%P HARM. %% ENZO %BIOCHE M%% ABEO NA%THE RAPE UTICS % TELIG ENT% AMG EN%% PHARMO S% CEL!S CI%% ENZO N%PHA RM.%% FERT IL!A! CHRO N% IMMU NOME DICS%% ACURA %PHA RM._XO%%MA%% REPL IGEN %% AMA G%PHA RM.%% CELLD EX%THE RAPE UTICS %% CYTRX% REGE NERX% BIOPH. % IMMU CELL% SOLIG ENIX% % CELGEN E%% UNIG ENE%L ABS.% HERO N%THE RAPE UTICS % AKORN %% JUNIP ER%PHA RM.%% PERE GRIN E%PHA RM.% CHAMP IONS %ONC OLOG Y%% BIO!TE CHN E%% EMIS PHERE %TECH. %% RESP IRERX% PHARM. % IMMU NOGE N%% USA%E QUITIE S%% REGE NERO N%PHA RM.%% IONIS %PHA RM.%% WOU ND%MA NAGE MENT %TECH. % ALKE RME S%% VERT EX%PHA RM.%% BIOGE N% BIOSP ECIFIC S%TEC H.% PERRIG O%% GILEA D%SCIE NCES %% PROTE IN%PO LYME R%TEC H.%% PDL%B IOPHA RMA %% PROTID E%PHA RM.% BIOTIME %% SCICL ONE%P HARM. % STEMC ELLS %% Highest%β% Lowest%β%

Figure 4. Betas in 2013

The dataset’s beta estimation is rather different from betas of other research-intensive

industries estimated by Damodaran (2016). According to EU R&D Scoreboard (2015), examples of

research-intensive industries are Aerospace-Defense, Automobiles, and Software. Their betas

amounted to be 1.33, 0.96, and 1.33, respectively. These betas are clearly larger than the betas

displayed in Figure 4. However, the estimated beta for the pharmaceutical sector is estimated at 1.02,

which is also considerably higher than this papers’ estimations for 2013. Possible explanations for this

difference could be that Damodaran (2016) used data originating from other years or that the dataset

was completely different from this papers’ dataset.

5.2

Historical revenue and net income

Graph 1 shows the movement of revenue growth rates between 2001 and 2015 for 10 different

pharmaceuticals. Graph 2 shows the movement of net income growth rates between 2001 and 2015.

The corresponding figures can be found in Appendix 1.

!1# !0,8# !0,6# !0,4# !0,2# 0# 0,2# 0,4# 0,6# BRIS TO L# MY ERS #S Q U IB B# ABBO TT #L ABS .# EL I#L IL LY ## JO HN SO N #& #JO HN SO N # ME RC K# & #C O MP AN Y# # PF IZ ER# ### # MY LA N ### CY BRD I## AV IRA G EN #T HE RA PE U TIC S# AL SE RE S# PHA RM. # PO LY D EX# PHA RM. ## EN ZO #B IO CHE M# # AB EO N A# THE RA PE U TIC S# TE LIG EN T# AMG EN ## PHA RMO S# CE L! SC I## EN ZO N #P HA RM. ## FE RT IL !A !C HRO N # IMMU N O ME D IC S# # AC U RA #P HA RM. ## XO MA ## RE PL IG EN ## AMA G #P HA RM. ## CE LL D EX# THE RA PE U TIC S# # CY TRX# RE G EN ERX# BIO PH. # IMMU CE LL # SO LIG EN IX# # CELGEN E## U N IG EN E# LA BS .# HE RO N #T HE RA PE U TIC S# AK O RN ## JU N IP ER# PHA RM. ## PE RE G RIN E# PHA RM. # CHA MP IO N S# O N CO LO G Y# # BIO !T EC HN E# # EMIS PHE RE #T EC H. ## RE SP IRE RX# PHA RM. # IMMU N O G EN ## U SA #E Q U IT IE S# # RE G EN ERO N #P HA RM. ## IO N IS #P HA RM. ## W O U N D #MA N AG EME N T# TE CH. # AL KE RME S# # VE RT EX# PHA RM. ## BIO G EN # BIO SP EC IF IC S# TE CH. # PE RRIG O ## G IL EA D #S CIE N CE S# # PRO TE IN #P O LY ME R# TE CH. ## PD L# BIO PHA RMA ## PRO TID E# PHA RM. # BIO TIME ## SC IC LO N E# PHA RM. # ST EMC EL LS ## 2013#

!750%& !550%& !350%& !150%& 50%& 250%& 450%& 650%& 850%&

2001& 2002& 2003& 2004& 2005& 2006& 2007& 2008& 2009& 2010& 2011& 2012& 2013& 2014& 2015&

ABBOTT&LABORATORIES&& BRISTOL!MYERS&SQUIBB&& LILLY&(ELI)&AND&CO.&& JOHNSON&&&JOHNSON&& MERCK&&&CO&INC&& PFIZER&INC& MYLAN&NV&& ENZO&BIOCHEM&INC& TELIGENT&& AMGEN&INC&

Graph 1. Revenue growth

Graph 2. Net income growth

Graph 1 shows multiple outliers, both negative and positive. In addition, every firms’ revenue

growth rate appears to move up and down. This seems to indicate that revenue is unstable for

companies that operate in this industry. Graph 2 also shows multiple outliers. These remarkable

!75%% !50%% !25%% 0%% 25%% 50%% 75%% 100%% 125%% 2001% 2002% 2003% 2004% 2005% 2006% 2007% 2008% 2009% 2010% 2011% 2012% 2013% 2014% 2015%

ABBOTT%LABORATORIES%% BRISTOL!MYERS%SQUIBB%% LILLY%(ELI)%AND%CO.%% JOHNSON%&%JOHNSON%% MERCK%&%CO%INC%%

!150%& !100%& !50%& 0%& 50%& 100%& 150%& 200%&

2001& 2002& 2003& 2004& 2005& 2006& 2007& 2008& 2009& 2010& 2011& 2012& 2013& 2014& 2015&

ABBOTT&LABORATORIES&& BRISTOL!MYERS&SQUIBB&& LILLY&(ELI)&AND&CO.&& JOHNSON&&&JOHNSON&& MERCK&&&CO&INC&& PFIZER&INC& MYLAN&NV&& ENZO&BIOCHEM&INC& AMGEN&INC&

growth rates are in almost each case the result of major acquisitions.

10

_{However, for the company with}

most outliers, Teligent, the remarkable rates are not caused by acquisitions. Extreme outliers result in

a less visible volatility net income growth rates. In order to obtain a proper outline of this industry’s

net income volatility, the greatest outliers are filtered out. These outliers are highlighted in Appendix

1. Graph 3 shows the same data used in Graph 2, only in this graph 4 outliers caused by acquisitions

and a penalty as well as the growth rates from Teligent are filtered out. Less extreme outliers, caused

by acquisitions and potential penalties, will still be captured in Graph 3.

Graph 3. Net income growth

This graph shows very clear that net income growth rates fluctuate for each pharmaceutical

company frequently. It is important to realize that a curve located below the x-axis does not

necessarily imply a negative net income. On the contrary, of all 9 companies only Enzo Biochem

experienced a negative net income between 2001 and 2015. For example, in 2007 Pfizer’s net income

decreased with more than 50 percent while its net income still amounted to be more than $8 billion

dollar.

11

This example clearly shows that a sharp decline in net income does not necessarily have a

detrimental impact. There is indeed a great volatility visible for all companies, although this seems to

be no compelling argument for setting excessive high prices as profits remain substantial when there is

a strong decline in net income growth.

When comparing Graph 1 with Graph 3, it appears that revenue growth rates are less volatile than

net income growth rates. Revenue growth rates move roughly between [-0.25% ; 25%] while net

10

_{AMGEN acquired Immunex Corporation for $16 billion in 2002 ; MYLAN acquired the generics arm of Merck KGaA in}

2007 for $6.6 billion ; LILLY (ELI) AND CO. received a $1.4 billion penalty in 2009; MERCK & CO. acquired

Schering-Plough for $41 billion in 2010 (retrieved from

http://www.reuters.com

).

income growth rates move roughly between [-50% ; 50%]. This difference in volatility may suggest

that sales are relatively stable and that a large share of net income volatility is the result of other

circumstances, for example, fluctuating investments. Relatively stable sales provide pharmaceutical

firms more certainty, which weakens the pharmaceuticals’ argument of facing substantial risk.

6 Conclusions

Previous studies showed that pharmaceutical firms in general gain profits that are many times as large

as the all-industry average profits. These high profits are the result of controversial high prices which

ultimately may come at the expense of the patient. In addition, high profits also create welfare loss

since pharmaceuticals will innovate less. Pharmaceutical firms justify the high prices by pointing to

the enormous development costs that the industry demands, and that therefore entail huge risks.

OLS regression

To determine if historical returns in this particular industry did exceed expected returns, this paper

examined if historical stock returns obtained from 56 pharmaceutical companies exceeded the

expected return according to the Capital Asset Pricing Model, CAPM. This economical model

estimates the expected return for an asset based on the assets’ exposed risk. By performing an OLS

regression, historical returns that actually exceeded expected return result in a significant positive

alpha. The performed OLS-regression results showed that between 1994 and 2014, most alphas were

positive but non-significant. Companies with the most significant alphas had very small significant

alphas that probably have little impact on return. It might even be purely accidental that these alphas

are significant since many alphas are estimated in this study. Since most alphas were non-significant,

no conclusions can be drawn based on this data. A possible explanation for this insignificant result is

that including stock returns without including dividend returns creates a distorted picture since

dividend distributions will temper share value and hence historical returns are not correctly reflected.

Therefore it is definitely worth to investigate in a follow-up study if including dividend returns will

result in a significant result. It is also worth to examine if using monthly returns instead of daily

returns could reduce potential noise coming from daily returns and hence result in more significant

alphas. As for the estimated betas, it appeared that the number of significant betas is seven times as

large as the number of significant alphas, when 3 = 10%. It was found that the estimated betas were

not stable between 1994 and 2013. However, the betas were quite similar for most companies in 2013.

Historical revenue and net income

To investigate whether pharmaceutical companies truly experience financial risks, this paper examined

if individual firms’ profits were sustained. Historical revenue and net income from 10 different listed

pharmaceuticals were collected and growth rates were examined. The growth rates showed that both

revenue growth rates and net income growth rates fluctuated frequently for all firms. However, the

volatility seems to be no compelling argument for setting excessive high prices as revenues and profits

in almost every case remained substantial when a strong decline in growth rates occurred. When

comparing the revenue and net income growth rates, it appeared that revenue growth rates moved less

volatile than net income growth rates. This may indicate that the pharmaceutical market is stable and

pharmaceutical firms experience less risk due to stable sales.

Altogether, pharmaceutical firms experience volatile profits. Since the pharmaceutical

industry demands high investment costs, it is no surprise that the net income growth rates are volatile.

Given that pharmaceuticals’ profits remain substantial, there is no reason to believe that there are

serious risks present. And therefore, no support was found for the claim that prices and hence profits

should be excessively high.

7 References

Angell, M. (2004). Excess in the pharmaceutical industry. Canadian Medical Association

Journal, 171(12), 1451-1453.

Ballegooijen, L. van. (2015). “The performance of the S&P Global Luxury Index compared to the

MSCI World Index, during the financial crisis” BA Thesis, University of Amsterdam. Accessed

at

http://dare.uva.nl/cgi/arno/show.cgi?fid=626648

.

Bauer, R., Koedijk, K., & Otten, R. (2005). International evidence on ethical mutual fund performance

and investment style. Journal of Banking & Finance, 29(7), 1751-1767.

Berk, J. B., & DeMarzo, P. M. (2007). Corporate finance. Pearson Education.

Bodie, Z., Kane, A., & Marcus, A. J. (2011). Investment and portfolio management. New York:

McGraw-Hill.

Chen, M. (2003). Risk and return: CAPM and CCAPM. Quarterly Review of Economics and Finance,

43(2), 369-393.

Damodaran, A. (2016). “Betas by Sector (US)” Accessed at

http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/Betas.html

.

Daves, P. R., Ehrhardt, M. C., & Kunkel, R. A. (2000). Estimating systematic risk: the choice of return

interval and estimation period. journal of Financial and Strategic Decisions, 13(1), 7-13.

Dickson, M., & Gagnon, J. P. (2004). Key factors in the rising cost of new drug discovery and

development. Nature Reviews Drug Discovery, 3(5), 417-429.

DiMasi, J. A., Hansen, R. W., Grabowski, H. G., & Lasagna, L. (1991). Cost of innovation in the

pharmaceutical industry. Journal of health economics,10(2), 107-142.

Dimasi, Hansen, & Grabowski. (2003). The price of innovation: New estimates of drug development

costs. Journal of Health Economics, 22(2), 151-185.

DiMasi, J. A., Grabowski, H. G., & Hansen, R. W. (2016). Innovation in the pharmaceutical industry:

New estimates of R&D costs. Journal of Health Economics, 47, 20-33.

EU R&D Scoreboard (2015). “The 2015 EU Industrial R&D Investment Scoreboard” Accessed at

http://iri.jrc.ec.europa.eu/scoreboard15.html

.

Fama, E., & French, K. (2004). The Capital Asset Pricing Model: Theory and Evidence. Journal of

Economic Perspectives, 18(3), 25-46.

Gençay, R., Selçuk, F., & Whitcher, B. (2003). Systematic risk and timescales. Quantitative

Finance, 3(2), 108-116.

Giaccotto, C., Santerre, R. E., & Vernon, J. A. (2005). Drug prices and research and development

investment behavior in the pharmaceutical industry. Journal of Law and Economics, 48(1),

195-214.

Ginsburg, P. B. (2008). High and rising health care costs: Demystifying US health care spending.

Princeton, NJ.

Ho, Y. K., Xu, Z., & Yap, C. M. (2004). R&D investment and systematic risk. Accounting &

Finance, 44(3), 393-418.

IMS Health (2007). “Intelligence 360: Global Pharmaceutical Perspectives 2006” Accessed at

http://www.amcp.org/WorkArea/DownloadAsset.aspx?id=13140

.

Jaberidoost, M., Nikfar, S., Abdollahiasl, A., & Dinarvand, R. (2013). Pharmaceutical supply chain

risks: a systematic review. DARU Journal of Pharmaceutical Sciences, 21(1), 69.

Kim, H., Mattila, A. S., & Gu, Z. (2002). Performance of hotel real estate investment trusts: a

comparative analysis of Jensen indexes. International Journal of Hospitality Management,

21(1), 85-97.

OECD. (2015). “Research and development in the pharmaceutical sector” Accessed at

http://www.oecd.org/health/health-systems/health-at-a-glance-19991312.htm.

Pattison, N., & Warren, L. (2002). drug industry profits: hefty pharmaceutical company margins dwarf

other industries. Washington (DC): Public Citizen Congress Watch; June 2003.

Scherer, F. (2004). The Pharmaceutical Industry — Prices and Progress. The New England Journal of

Medicine, 351(9), 927-932.

Spitz, J., & Wickham, M. (2012). Pharmaceutical High Profits: The Value of R&D, or Oligopolistic

Rents? American Journal of Economics and Sociology, 71(1), 1-36.

Tax, S. & Van der Hoeven, K. (2014). Nieuwe dure geneesmiddelen tegen kanker; een dilemma.

Nederlands tijdschrift voor geneeskunde, 158, A8034.

Toll, P. J. M. M. (2004). Ontwikkelingskosten van nieuwe geneesmiddelen in maatschappelijk

perspectief. Medisch-Farmaceutische Mededelingen, 42(7), 193-193.

Vernon, J. A. (2005). Examining the link between price regulation and pharmaceutical R&D

investment. Health economics, 14(1), 1-16.

8 Appendices

Appendix 1 – Revenue and net income growth rates

Revenue growth rates

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 ABBOTT%LABORATORIES% 0,185 0,086 0,113 25E205 0,135 0,006 0,153 0,139 0,042 0,143 0,105 0,026 20,452 20,073 0,008 BRISTOL2MYERS%SQUIBB% 0,066 20,067 0,153 20,072 20,009 20,067 0,08 0,065 20,087 0,036 0,09 20,171 20,07 20,031 0,043 LILLY%(ELI)%AND%CO.% 0,063 20,04 0,136 0,101 0,057 0,071 0,188 0,094 0,072 0,057 0,052 20,069 0,023 20,151 0,017 JOHNSON%&%JOHNSON% 0,133 0,1 0,153 0,131 0,067 0,056 0,146 0,043 20,028 20,005 0,055 0,034 0,061 0,042 20,057 MERCK%&%CO%INC% 0,182 0,085 20,566 0,02 20,04 0,028 0,069 20,014 0,15 0,674 0,048 20,016 20,07 20,041 20,065 PFIZER%INC 0,093 0,009 0,396 0,162 20,023 20,057 0,001 20,003 0,035 0,356 20,006 20,125 20,125 20,038 20,015 MYLAN%NV% 0,072 0,304 0,15 0,083 20,088 0,003 1,311 0,769 20,009 0,07 0,126 0,108 0,017 0,117 0,221 ENZO%BIOCHEM%INC 0,167 20,075 20,023 20,211 0,042 20,082 0,328 0,47 0,151 0,084 0,051 0,01 20,091 0,024 0,017 TELIGENT% 20,174 20,726 20,185 3E204 20,194 20,086 0,748 20,112 20,072 0,613 0,281 0,097 1,128 0,851 0,311 AMGEN%INC 0,106 0,375 0,513 0,263 0,178 0,148 0,035 0,016 20,024 0,025 0,038 0,108 0,082 0,074 0,08

Net income growth rates

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 ABBOTT%LABORATORIES% +0,444 0,802 +0,015 0,175 0,042 +0,491 1,1 0,354 0,177 +0,195 0,022 0,261 +0,568 +0,113 0,937 BRISTOL+MYERS%SQUIBB% 0,026 +0,558 0,453 +0,231 0,256 +0,472 0,366 1,424 1,022 +0,708 0,196 +0,472 0,308 +0,218 +0,219 LILLY%(ELI)%AND%CO.% +0,091 +0,026 +0,054 +0,293 0,094 0,345 0,109 +1,702 +3,089 0,171 +0,142 +0,06 0,139 +0,487 0,008 JOHNSON%&%JOHNSON% 0,144 0,005 0,263 0,137 0,23 0,098 +0,043 0,224 +0,053 0,086 +0,274 0,122 0,275 0,18 +0,056 MERCK%&%CO%INC% 0,067 +0,018 +0,045 +0,147 +0,206 +0,043 +0,261 1,384 0,652 +0,933 6,285 +0,017 +0,286 1,707 +0,627 PFIZER%INC 1,09 0,172 +0,572 1,905 +0,288 1,392 +0,579 0,118 +0,052 +0,044 0,212 0,456 0,51 +0,585 +0,924 MYLAN%NV% +0,759 6,009 0,046 0,229 +0,392 +0,094 0,177 +1,834 +2,283 0,484 0,555 0,194 +0,027 0,49 +0,088 ENZO%BIOCHEM%INC 0,028 0,016 +0,445 +2,621 +1,482 +1,522 +0,154 +0,197 1,212 +0,056 +0,417 2,03 +0,536 +0,453 +0,771 TELIGENT% +0,848 +6,046 +1,037 1,77 0,455 0,284 +0,801 4,578 1,657 +0,304 +0,122 0,306 +0,786 +7,251 0,27 AMGEN%INC +0,016 +2,243 +2,623 0,046 0,555 +0,197 0,073 0,28 0,136 0,005 +0,204 0,18 0,169 0,015 0,345

Appendix 2 – OLS regression output

1994

1995 1996 1997 1998 1999

Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value

BRISTOL MYERS SQUIBB BF 0.1395739 0.013 0.1424491 0.028 0.3510785 0 0.4703471 0 0.4142199 0 0.1204251 0.083 _cons 0.0000933 0.876 0.0013678 0.016* 0.0008897 0.209 0.0013977 0.292 0.0008892 0.436 -0.0001194 0.93 ABBOTT LABORATORIES BF 0.1543603 0.089 0.1692953 0.098 0.3677531 0 0.382542 0 0.341463 0 0.1707649 0.011 _cons 0.0005103 0.599 0.0007825 0.38 0.0006877 0.416 0.0003037 0.752 0.0011757 0.256 -0.001156 0.38 ELI LILLY BF 0.1766173 0.071 0.2578143 0.005 0.3227356 0.003 0.4861149 0 0.4126923 0 0.1875238 0.017 _cons 0.000524 0.617 0.0018568 0.021* 0.0010223 0.386 0.0017054 0.189 0.0005675 0.676 -0.001056 0.493 JOHNSON & JOHNSON BF 0.1981271 0.006 0.1331535 0.121 0.3785718 0 0.341698 0 0.2589943 0 0.2094965 0 _cons 0.0007282 0.347 0.0015526 0.039* 0.0004852 0.572 0.0004824 0.646 0.0006389 0.52 0.0002756 0.793 MERCK & COMPANY BF 0.1349906 0.098 0.0209413 0.817 0.3502926 0 0.3557999 0 0.3310274 0 0.2277375 0 _cons 0.0003521 0.687 0.0020721 0.009** 0.0006806 0.44 0.0005251 0.65 0.0009301 0.395 -0.000452 0.698 PFIZER BF 0.1644878 0.035 0.0804433 0.44 0.4651258 0 0.518586 0 0.4706187 0 0.2226136 0.003 _cons 0.0005154 0.536 0.001817 0.046* 0.0009661 0.295 0.0013711 0.243 0.0015457 0.27 -0.0009782 0.502 MYLAN BF 0.0980297 0.488 -0.1660727 0.237 0.5256934 0 0.4176197 0 0.3689931 0.005 0.0450444 0.617 _cons 0.0005697 0.707 0.0012369 0.313 -0.001324 0.323 0.0003852 0.826 0.001504 0.452 -0.000611 0.73 CYBRDI BF -1.095318 0.155 -0.9593066 0.406 -0.2023361 0.799 -0.1487448 0.715 -0.5715184 0.233 0.376346 0.265 _cons 0.0075854 0.358 0.0114656 0.255 0.009308 0.287 0.0028928 0.639 0.0064373 0.382 0.0044086 0.506 AVIRAGEN THERAPEUTICS BF 0.572897 0.016 0.2043072 0.331 0.3340881 0.124 0.4464726 0.01 -0.4271925 0.145 0.1057579 0.532 _cons 0.0038824 0.126 0.0014351 0.434 -0.0002682 0.911 -0.0036978 0.158 0.00197 0.662 0.0032255 0.332 ALSERES PHARM. BF -0.3108947 0.725 2.35042 0.024 0.5891033 0.162 0.0857788 0.752 0.1855165 0.791 -0.0899348 0.67 _cons -0.0037106 0.695 0.0141534 0.117 0.0020109 0.664 -0.0019435 0.635 0.0081404 0.449 0.0024518 0.554 POLYDEX PHARM. BF 0.6033102 0.092 -0.3264222 0.608 0.204395 0.596 -0.095217 0.693 0.5265295 0.121 -0.2843696 0.269 _cons -0.0031471 0.412 0.0028929 0.602 0.0031117 0.463 0.0018333 0.615 -0.0003687 0.944 0.0042746 0.398 ENZO BIOCHEM BF 0.5598137 0.006 -0.0662962 0.774 0.2173998 0.195 0.3486124 0.005 0.5351038 0.001 0.2208787 0.137 _cons -0.0009553 0.659 0.0027019 0.181 0.0001822 0.921 -0.0011817 0.524 -0.0012393 0.602 0.0064902 0.026* ABEONA THERAPEUTICS BF 0.3742281 0.451 0.2462212 0.733 0.1665532 0.663 -0.4669563 0.249 0.0384481 0.902 -0.5862789 0.01 _cons 0.0004931 0.926 0.0069335 0.271 0.0007412 0.86 0.0041081 0.502 -0.0033191 0.491 0.0024617 0.582 TELIGENT BF 0.4349455 0.01 0.4483046 0.05 0.294548 0.197 0.0001591 0.999 -0.082853 0.736 -0.0563169 0.743 _cons 0.001566 0.385 -0.0001392 0.944 -0.000199 0.937 -0.0016751 0.397 -0.000973 0.797 0.0013643 0.686

1994 1995 1996 1997 1998 1999

Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value

AMGEN BF 0.1348128 0.276 0.4811239 0.001 0.2671583 0.019 0.2983162 0 0.172473 0.08 0.1720118 0.088 _cons 0.000748 0.573 0.0022439 0.074** -0.0003587 0.774 -0.0005284 0.672 0.0024582 0.105 0.0034043 0.085** PHARMOS BF -0.1238374 0.724 -0.6139393 0.292 0.3932817 0.23 0.2286941 0.203 0.1603452 0.395 0.0059881 0.967 _cons -0.004698 0.212 0.0040501 0.426 0.0013918 0.699 0.0015175 0.576 -0.0001643 0.955 0.0020238 0.477 CEL-SCI BF -0.0599051 0.869 -0.3305598 0.525 -0.1580216 0.685 0.5675083 0.034 0.5065443 0.014 0.2320153 0.368 _cons -0.0036091 0.353 0.0022116 0.626 0.0027349 0.524 0.0024615 0.542 -0.0045782 0.147 0.0036078 0.476 ENZON PHARM. BF 0.0558223 0.858 0.9567178 0.017 0.7889159 0.008 0.1218743 0.583 0.7842937 0 0.1344813 0.216 _cons -0.0030727 0.36 0.0011738 0.736 0.0022435 0.488 0.0033964 0.312 0.0035718 0.26 0.0048605 0.023* FERTIL-A-CHRON BF 0.0001153 0.243 0.0000274 0.814 0.000067 0.256 0.0000155 0.668 0.0000684 0.125 -3.30E-08 0.999 _cons -0.0001877 0 -0.0001745 0 -0.0001709 0 -0.0001687 0 -0.0001406 0 -0.0001505 0 IMMUNOMEDICS BF 0.4796561 0.149 0.2146014 0.637 0.2334739 0.359 0.0047999 0.979 0.3511836 0.185 -0.1611009 0.491 _cons -0.0003233 0.928 0.0036475 0.359 0.0004055 0.885 -0.0008049 0.772 0.0020176 0.62 0.0067933 0.14 ACURA PHARM. BF 0.1988235 0.557 -0.1581096 0.702 0.2812941 0.309 0.1597736 0.445 0.0328992 0.906 -0.055608 0.799 _cons -0.0018195 0.616 0.0036066 0.318 0.00299 0.326 -0.0044227 0.163 0.0016107 0.708 0.0007094 0.868 XOMO BF 0.3039212 0.324 0.533369 0.283 0.7318369 0.018 0.5896022 0.001 0.123984 0.525 0.2386947 0.283 _cons -0.0015462 0.64 0.0025055 0.562 0.0025724 0.448 8.85E-06 0.997 -0.0012678 0.672 0.0022373 0.608 REPLIGEN BF 0.1033962 0.767 -0.3583292 0.547 1.310843 0.008 -0.0247871 0.927 -0.4181251 0.262 -0.0545328 0.772 _cons -0.0037362 0.318 0.0016744 0.747 0.0035555 0.512 0.0002982 0.942 0.0056314 0.326 0.0050454 0.173 AMAG PHARM. BF 0.1985777 0.159 0.3954957 0.012 0.2383702 0.082 0.139789 0.204 0.1016785 0.612 0.0614517 0.669 _cons 0.0005723 0.704 0.0020397 0.135 -0.0019909 0.186 -0.0022663 0.173 0.0000978 0.975 -0.0013696 0.627 CELLDEX THERAPEUTICS BF 0.0593024 0.846 0.2427682 0.522 0.458568 0.097 0.2961568 0.313 0.3137422 0.251 -0.3044997 0.19 _cons -0.0035533 0.278 0.0017984 0.586 -0.0015603 0.607 0.0031453 0.478 0.0003588 0.932 0.0039904 0.382 CYTRX BF 0.5091864 0.059 -0.9031199 0.078 -0.0441947 0.858 0.0135222 0.942 0.4194034 0.158 0.2371622 0.337 _cons -0.0050555 0.08** 0.0030483 0.495 -0.0002829 0.917 0.0002001 0.944 -0.0023569 0.606 0.002855 0.556 REGENERX BIOPHARM. BF -0.1457259 0.775 -0.6163725 0.443 -0.0496177 0.965 -0.0054851 0.993 0.8464101 0.286 1.864282 0.052 _cons -0.0067635 0.216 0.0007873 0.91 0.0161986 0.191 0.0082079 0.364 0.0118365 0.332 0.0309029 0.1** IMMUCELL BF 0.0180186 0.959 0.613383 0.134 -0.0040191 0.991 0.2111718 0.378 0.2454392 0.356 0.0858781 0.721 _cons -0.0008403 0.822 0.0039976 0.262 0.0015237 0.693 0.0012218 0.736 -0.0009505 0.816 0.0050572 0.285

1994 1995 1996 1997 1998 1999

Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value Coef. p-value

SOLIGENIX BF -2.091695 0.291 1.106234 0.633 -0.1947116 0.829 0.7720027 0.83 -0.1895583 0.473 0.3295763 0.121 _cons 0.0111327 0.6 0.0301901 0.136 0.0195327 0.05* 0.0654777 0.229 -0.0014878 0.714 0.0028404 0.495 CELGENE BF 0.5863249 0.002 0.1331902 0.62 0.9330636 0 0.1716571 0.293 0.5906558 0.007 0.181047 0.199 _cons -0.0004696 0.817 0.0037225 0.113 -0.0001626 0.95 -0.0007787 0.752 0.0028758 0.39 0.0064959 0.019* UNIGENE LABS. BF -0.437011 0.09 -0.1036538 0.834 0.5531812 0.085 0.3111926 0.114 0.3865053 0.131 0.0162659 0.941 _cons 0.0007733 0.779 -0.0000434 0.992 0.0029555 0.402 0.0013505 0.65 -0.0016809 0.669 -0.0006598 0.878 HERON THERAPEUTICS BF 0.2877475 0.242 -0.2470846 0.38 0.5304065 0.04 0.3550895 0.007 0.3175882 0.099 -0.1220058 0.481 _cons -3.39E-06 0.999 0.001709 0.486 0.0019824 0.484 -0.0008435 0.672 -0.0002525 0.932 -0.0003874 0.909 AKORN BF -0.3100757 0.167 0.447118 0.138 -0.1631061 0.511 0.6447395 0.001 0.3905072 0.047 0.0962992 0.364 _cons 0.0005463 0.82 -0.0006613 0.801 -0.0002338 0.932 0.0022067 0.46 0.0016649 0.581 0.0003221 0.877 JUNIPER PHARM. BF 0.4093132 0.055 -0.1899274 0.478 0.5469001 0.005 0.3242632 0.02 0.8127357 0.002 0.3461761 0.036 _cons 0.0000499 0.983 0.0028604 0.221 0.0021277 0.321 0.0005188 0.805 -0.0058119 0.141 0.0042928 0.183 PEREGRINE PHARM. BF 0.0001153 0.243 0.0000274 0.814 5.636051 0.29 0.0224455 0.896 -0.0790619 0.837 0.1431233 0.68 _cons -0.0001877 0 -0.0001745 0 0.0563447 0.336 -0.0021166 0.414 0.0013613 0.818 0.0022212 0.744 CHAMPIONS ONCOLOGY BF 1.255589 0.18 -1.349008 0.297 0.4025333 0.684 0.3074746 0.463 0.0514847 0.93 0.7870125 0.037 _cons 0.0100577 0.316 0.0116931 0.3 0.0167395 0.125 0.0053105 0.402 0.0034089 0.704 0.0067009 0.364 BIO-TECHNE BF 0.1714341 0.369 0.2358258 0.253 0.0230306 0.893 0.228628 0.037 -0.0733018 0.578 0.1501081 0.11 _cons 0.0003535 0.863 0.002597 0.149 0.0012571 0.504 0.0009098 0.582 0.0012 0.554 0.0038331 0.038* EMISPHERE TECH. BF 0.0455272 0.884 0.4719967 0.384 0.6399157 0.028 0.1999911 0.224 0.7145395 0 0.1638095 0.281 _cons -0.0036665 0.275 0.0058652 0.215 0.0048752 0.127 0.0002472 0.921 -0.0004987 0.872 0.0033676 0.26 RESPIRERX PHARM. BF 0.2459141 0.434 -0.0953315 0.854 0.7935279 0.08 0.4593767 0.011 0.5888859 0.15 -0.0025367 0.993 _cons -0.0033431 0.321 0.0045901 0.309 0.001869 0.707 -0.0035116 0.194 0.0007158 0.909 0.0038845 0.484 IMMUNOGEN BF 0.0451515 0.91 -0.3215015 0.534 0.8472857 0.035 0.2835059 0.223 0.7503123 0.015 0.1544325 0.427 _cons -0.0033722 0.431 0.0029012 0.52 0.00213 0.628 -0.002983 0.396 0.0052374 0.266 0.0049667 0.194 USA EQUITIES BF -0.0936572 0.785 -0.4099713 0.413 0.2038396 0.481 0.1221384 0.592 0.2398069 0.521 -0.0725912 0.778 _cons -0.0028007 0.446 0.0049323 0.259 0.0025741 0.418 -0.0018131 0.599 -0.0000663 0.991 0.0015368 0.761 REGENERON PHARM. BF 0.4638664 0.116 -0.1807521 0.633 0.8683043 0 -0.0460094 0.853 0.3540517 0.055 0.6274781 0 _cons -0.0050822 0.108 0.0068652 0.038* 0.0013573 0.581 -0.000711 0.849 -0.0001447 0.959 0.002481 0.375