See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/225466494
Analysis of Manufacturing Costs in Pharmaceutical Companies
Article in Journal of Pharmaceutical Innovation · March 2008
DOI: 10.1007/s12247-008-9024-4
CITATIONS
READS
73
67,881
5 authors, including:
Prabir K Basu
37 PUBLICATIONS 624 CITATIONS
Girish Joglekar
Purdue University
35 PUBLICATIONS 471 CITATIONS
SEE PROFILE
SEE PROFILE
Pradeep Suresh
PricewaterhouseCoopers
16 PUBLICATIONS 407 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
C-SOPS NSF ERC View project
LIU/FDA View project
All content following this page was uploaded by Girish Joglekar on 03 August 2015.
The user has requested enhancement of the downloaded file.
J Pharm Innov (2008) 3:30–40
DOI 10.1007/s12247-008-9024-4
PERSPECTIVE
Analysis of Manufacturing Costs
in Pharmaceutical Companies
Prabir Basu & Girish Joglekar & Saket Rai &
Pradeep Suresh & John Vernon
Published online: 4 March 2008
# International Society for Pharmaceutical Engineering 2008
Abstract In the pharmaceutical industry, costs attributed
to manufacturing are a major part of a company’s total
expenses. In this paper, trends in various expense and
income categories of pharmaceutical companies have been
analyzed with particular emphasis on manufacturing costs
to gain an insight into their relationships and how they may
differ among types of pharmaceutical companies such as
brand name, generics, and biotechs. The study includes data
published in the annual reports of leading pharmaceutical
companies from 1994 to 2005. Twenty-two pharmaceutical
companies were selected based on the annual revenues. The
set was further divided into three groups: brand names,
generics, and biotechs. The analysis shows that, between
1994 and 2005, manufacturing costs (as a percentage of
total sales) are different for the three groups of companies
listed above. Additionally, each group of companies differs
in how savings are leveraged strategically. The data on
brand-name pharmaceutical companies also indicate that
there is a strong correlation between the reduction of the
cost of goods sold (COGS) and the increase in R&D
expenditure. This suggests the validity of Vernon’s theory
that for brand-name companies, a reduction in COGS will
P. Basu (*) : G. Joglekar : S. Rai
Pharmaceutical Technology and Education Center,
Purdue University,
W. Lafayette, IN 47907, USA
e-mail: prabir1960@purdue.edu
P. Suresh
School of Chemical Engineering, Purdue University,
W. Lafayette, IN 47907, USA
J. Vernon
Department of Finance, University of Connecticut,
Storrs, CT, USA
likely have a positive impact on investments in R&D,
presumably resulting in much needed innovations and
future health benefits for the society.
Keywords COGS . Pharmaceutical manufacturing
Background
Prescription drugs often provide effective alternatives to
expensive medical procedures and hospital stays. Consequently, spending on prescription drugs as a percentage of
the total national health care spending is increasing. In
1999, prescription drugs accounted for 8.2% of the total
national health spending; that share was 11% in 2003 and is
expected to reach 14% by 2010 [1]. Although this is still a
relatively small proportion as a percentage of the total
national healthcare spending, it is one of the fastest growing
components of healthcare spending, increasing at double
digit rates from 1995 to 2003 [2].
The cost of bringing a new drug to the market place has
also been steadily increasing, with recent estimates projecting a required investment of over $2 billion to progress
from a laboratory idea to successful commercialization [3].
Pharmaceutical companies are spending more money on
R&D [4], while the productivity of their R&D investment,
computed as the number of drugs introduced to the market
place per year, is declining. Manufacturing costs are a
substantial part of their total cost structure [5]. According to
some estimates, these costs can be as high as 27–30% of
sales for manufacturers of brand-name pharmaceuticals
[1, 3], more than double the share of costs for research and
development [1]. No such estimates are currently available
in the literature for generic pharmaceuticals or drugs
manufactured by biotech companies.
J Pharm Innov (2008) 3:30–40
The manufacturing cost of pharmaceuticals, commonly
known as cost of goods sold (COGS), was approximately
$90 billion in 2001 for the top 16 drug companies [1]. In
2008, it is estimated that (in absolute terms) the total COGS
for all pharmaceutical products could be as much as $200
billion (estimated using 27% of estimated global pharmaceutical sales of $735 billion in 2008 [6]). In comparison,
the total spending in absolute terms on R&D by the
Pharmaceutical Research and Manufacturers of America
companies was $55 billion in 2007 [4].
31
generic, and biotech. The financial information for top
companies in each group, as disclosed in the annual reports
published by publicly traded companies, was used in the
analysis. The main objective can be divided into the
following sub-objectives:
–
–
Factors Affecting COGS of Pharmaceutical Companies
The pharmaceutical industry is strictly regulated due to its
direct impact on consumer health and well being. To ensure
that pharmaceutical products are safe and efficacious, the
Food and Drug Administration (FDA) periodically inspects
the facilities and procedures of all manufacturing operations
in the USA and those overseas operations that sell their
products in the USA. Consequently, these facilities and
procedures must be registered with the FDA and must
comply with the “(current) good manufacturing practices”
(cGMP) established by the FDA. The high COGS of
pharmaceutical products is the consequence, in part, of the
methods by which excellence in delivered product is
achieved. Other factors contributing to the high COGS
and the potential for savings in COGS have been
summarized in a previous report [3].
To offset the effects of rising costs of commercialization,
shorter effective exclusivity periods, and diminishing
returns on R&D investment, manufacturing costs may be
a source of savings for the pharmaceutical industry [3].
Reducing manufacturing costs without sacrificing quality
could be a way to effect social good in an environment
where more and more investment is required to find new
therapies for unmet medical needs along with a need to
control or slow down the rate of price increases of
prescription drugs. In fact, there are important linkages
between the efficiency of pharmaceutical manufacturing,
drug prices, and public health in the USA [7]. It can be
predicted that reductions in manufacturing costs will lead to
gains in consumer surplus (the difference between consumers’ willingness to pay and what consumers actually
pay; the standard economic measure of social welfare)
worth trillions of dollars.
Objectives
The main objective of this study was to assess and analyze
manufacturing costs of pharmaceutical products across
important industry categories. The companies comprising
this sector were divided into three groups: brand name,
–
Gain a better understanding of the trends in various
expense and income categories within companies of
each group.
Study the trends of the overall expenses and revenues.
Identify the differences in the trends, if any, of various
expense and income categories for each group of
companies. If there are differences, identify the causes.
Explore possible correlations between various expense
and income categories and propose the cause of the
relationships.
Methodology
The data used in this work were extracted from the annual
reports of the pharmaceutical companies studied. The
selected companies, based on annual revenues for the year
2005, are shown in Tables 1, 2, and 3. Brand-name
pharmaceutical companies are the original developers of
the drugs and have annual revenues of at least $10 billion.
Generic pharmaceutical companies manufacture off-patent
drugs, with typical annual revenues of less than $5 billion.
In 2008, more than two thirds of all prescriptions written in
the USA are expected to be for generics, and the generics
sales is expected to grow to more than $70 billion [6].
Biotech pharmaceutical companies are the original developers of drugs made through biosynthetic processes. Brandname companies account for the greatest share of the
market, though the weighted coverage of all companies is
well over 50% of the total market.
Table 1 Sample US brand-name pharmaceutical data set
Company
% Total market share
Pfizer Inc.
Johnson & Johnson Inc.
GlaxoSmithKline PLC
Sanofi-Aventis
Novartis
Astra Zeneca PLC
Abbott Laboratories
Merck & Co. Inc.
Bristol-Myers Squibb
Wyeth
Eli Lilly & Co.
Schering-Plough Corp.
8.5
8.4
6.3
5.4
5.4
4.0
3.7
3.7
3.2
3.1
2.4
1.6
55.6
32
J Pharm Innov (2008) 3:30–40
Table 2 Sample US generic pharmaceutical data set
Company
% Generics market share
Teva
Ivax* (2004)
Watson
Mylan
Barr
Alpharma
Par
9.5
3.3
3.0
2.3
1.9
1.0
0.8
21.9
Financial Information
The financial information about the companies was
extracted from the Wharton Research Data Services
(WRDS) Compustat [8] database. The data for the
following revenue and expense categories were extracted:
(a) Sales (SALES)—aggregate sales of a company’s
complete product offering.
(b) Cost of goods sold (COGS)—also referred to as
materials and production cost. Cost of goods sold is
an aggregate figure that includes all costs incurred in
producing the goods including write-offs from plant,
property and equipment, raw materials, inventory, etc.
(c) Research and development expense (R&D)—research
and development expenses that are separate from the
cost of goods sold.
(d) Selling, general, and administrative expense (XSGA)—
these expenses provide for sales, marketing, as well as
general expenses incurred by the product pipeline.
XSGA provided by the WRDS database includes
salaries, rent, and research and development (R&D)
cost. In this study, R&D and general expense were
treated as separate categories.
(e) General expense = XSGA − R&D.
(f) Taxes—taxes paid by the company.
(g) Depreciation—the steady loss in the value of capital
goods over a specified time period.
(h) Operating income (after depreciation and taxes)—
profits after depreciation and taxes; these are company
earnings from core operations after deducting the cost
of goods sold, and selling and general operating
expenses.
Operating income ¼ Sales COGS XSGA
Depreciation Taxes
For the categories given above, data were extracted for a
span of 12 years, from 1994 to 2005.
Data Collection and Analysis
To compare data in various categories across the three
company groups, the extracted data were normalized with
respect to the annual sales for the corresponding year
(represented as COGS%, general expense, etc.). The
normalized data were used for the detailed analysis.
Assumptions and Limitations
The authors have assumed that the data reported by the
companies in the financial statements are based on the same
interpretation of various categories. For example, COGS
should be truly inclusive of all costs pertaining to drug
manufacturing only. Although all companies follow generally accepted accounting practices, it is difficult to enforce
and monitor uniformity in their interpretations. One of the
limitations of the information obtained from financial
statements pertains to the lack of itemized information by
product, that is, we cannot associate a certain cost to a
certain product or process. From financial statements alone,
we cannot determine a firm’s unique or competitive
advantage in manufacturing capabilities. Also, some companies are, in reality, large conglomerates reporting all
business units, including non-pharmaceutical operations,
under one single filing.
For our analysis, we have chosen a sample space of
companies. Although these companies account for a major
share of the market, a significant number of companies
have not been included in the study.
The following procedures were applied to study the
aggregate data for each group of companies, namely, brand
names, generics, and biotechs:
–
–
Polynomial or linear trend lines were fitted to the data
set and general trends were observed.
The significance of change in each category was
established by using a t test, which assesses whether
the means of two groups are statistically different from
each other. Relationships were proposed among the
Table 3 Sample US biotech pharmaceutical data set
Company
% Total share
Amgen Inc.
Genentech Inc.
Genzyme Corp.
Biogen Idec Inc.
24
13
5
5
46
J Pharm Innov (2008) 3:30–40
33
Brand Name (with merck)
Generics
Biotech
Brandname Avg. (26%)
Generics Avg. (52%)
Biotech Avg. (14%)
Reinhardts Avg. (27%)
COGS%
70%
COGS %
60%
50%
40%
30%
20%
10%
0%
1994
1996
1998
2000
2002
2004
2006
Year
Fig. 1 COGS% for different types of companies
–
–
factors that showed significant changes and trends. The
significance level was chosen to be 0.05 (alpha value).
The period under consideration was divided into two
sub-periods of 6 years each. The arithmetic averages of
various expenses and incomes were calculated for each
sub-period. A t test was then done to find the
significance of the change from one sub-period to
another.
The relationships were quantified by computing the
correlations among different factors.
For the factors that showed significant changes, the
corresponding averages were compared with Reinhardt’s
average from financial data collected for eight brandname companies in 1998 [1]. To provide an economic
perspective on the pharmaceutical industry, Reinhardt
cites a Deutsche Banc Alex Brown research report and a
Banc of America Securities LLC report that shows
breakdowns of the disposition of the sales revenue
earned by the largest research-based pharmaceutical
manufacturers (defined as brand-name companies in this
paper) in 1988. Reinhardt also urges caution in
interpreting the data and provides some of the reasons
for that. Similar caution must be exercised in interpreting the financial data reported in this study.
was analyzed. The variations over the years for different
groups of companies were plotted, and the corresponding
averages were compared with Reinhardt’s average. Figure 1
shows the trends in the COGS% data for brand name,
generics, and biotechs, along with their corresponding
arithmetic averages. The average COGS% for brand-name
companies is nearly equal to the average estimated by
Reinhardt. However, the average COGS as a percentage of
sales, is almost half the arithmetic average for COGS as a
percentage of sales for generics and roughly double that of
biotechs. The higher value of COGS as a percentage of
sales for generics is possibly a reflection of lower
expenditure on R&D and sales, and marketing for the
generic industry. However, it is also quite revealing to
realize that the COGS% is significantly lower for biotechs.
The COGS as a percentage of total sales (COGS%) for
brand names appears to have declined during the years
2000 to 2005. On the other hand, the COGS% for generics
increased until the year 1996, and since then, there appears
to be a gradual decline. The COGS% values of generics
show a gradual reduction over the last 8 years. Biotechs
show more fluctuations than either generics or brand-name
companies.
As a special case, the time series data for ScheringPlough are shown in Fig. 2. In the case of Schering-Plough,
COGS% was significantly lower than the industry average
until about 2001. However, it increased significantly over
the years 2002–2004, growing more than the industry
average and then appears to have dropped back to the level
of the industry average. Correspondingly, the operating
income, which was slightly above the overall industry
average, dropped sharply over the years 2002–2004 and
then showed an increasing trend in 2005. The possible
explanation for this is that Schering-Plough entered a
consent decree with the FDA in 2002, agreeing to
Schering-Plough
Brand-Name
COGS%
Key Observations
35%
OIADP&T%
30%
R&D%
25%
20%
Trends Analysis COGS%, R&D%, Operating Income,
and General
15%
10%
5%
COGS
The trend of COGS, R&D, operating income, and general
expense as a percentage of total sales for the different types
of pharmaceutical companies over a 12-year time period
0%
1994
-5%
1996
1998
2000
2002
-10%
-15%
Year
Fig. 2 Schering Plough Corporation time series plot
2004
2006
34
J Pharm Innov (2008) 3:30–40
40%
35%
30%
R&D%
25%
20%
15%
10%
5%
0%
1994
1996
1998
2000
2002
2004
2006
Year
Fig. 3 R&D for different groups of companies (without 2002 data for
biotechs)
revalidate the manufacturing processes at several sites in
the USA and Puerto Rico, resulting in significant increase
in COGS% since 2002. In addition, it discontinued certain
older profitable products and outsourced other products.
The fact that, until 2002, the COGS% at Schering-Plough
was significantly lower than the industry average for brandname companies could be reflecting a lower investment in
plant, equipment, and cGMP systems, which could have
resulted in the findings of cGMP deficiencies.
R&D
The trends in R&D expenditures as a percentage of sales
(R&D%) are shown in Fig. 3. The data point for 2002 was
removed for biotechs, which is considered as an outlier.
The R&D% for brand-name companies appears to be
gradually increasing over the past few years. The average
value is very close to Reinhardt’s average. The trend for the
generics is substantially flat with peaks in 2000 and 2004.
The average value is slightly greater than half of the
average estimated by Reinhardt and the brand-name
industry average. This is expected since most generics are
not expected to devote resources to discovering new drugs.
As a result, their R&D% expenditures are largely directed
to developing marketable formulations of a large number of
drugs that are either out of patent or may be soon. The trend
for biotechs shows a record high R&D expenditure in the
year 1996 when it reached nearly 32%. Since 1996, the
R&D% has been fluctuating with an overall decreasing
trend. For biotechs, the average R&D expenditure as a
percentage of sales is twice Reinhardt’s average for brandname pharmaceuticals. Biotechs heavily invest in R&D to
discover new therapies; their R&D success rates are lower,
and they have a smaller portfolio of marketed products.
Operating Income
The trend in operating income as a percentage of sales is
shown in Fig. 4. The average operating income for brandname companies is around 19%. There is a gradual increase
in its value until the year 2003. After that, the trend
indicates a gradual decrease. However, the slope is very
small, and no definitive conclusions can be drawn. The
yearly average for generics is around 12%, which is lower
than that of brand-name and biotechs. The operating
income data for generics show a significant dip in 1997,
and after that, the trend indicates that it has been gradually
increasing. However, there is more fluctuation in the data
for generics as compared to that for brand-name companies.
For biotechs, the total number of companies in the data set
is small. However, there appears to be an increasing trend
of operating income in recent years.
General Expense
The trend in general expenses as a percentage of sales is
shown in Fig. 5. The trend for brand-name companies is
flat, and the average value is slightly less than the average
reported by Reinhardt in 2001. The trend for generics
shows more fluctuations compared to the brand-name
companies, with the lowest value being 12% in the year
2004. The average value of general expenses for generics is
nearly half of Reinhardt’s average, which confirms that the
generics spend less money in marketing and sales than
brand names or biotechs. The average value of general
expense for biotechs is slightly higher that that of generics
but lower than that of brand-name companies. Biotech
drugs are unique, have less competition from similar (metoo) drugs in the market, and treat specific unmet needs.
Therefore, the biotechs may not require as much marketing
and sales effort as do brand-name companies. The trend for
Brand name
Generics
Biotech
Brand name Avg. (19%)
Generics Avg. (12%)
Biotech Avg. (22%)
0.4
Operating Income/Sales
Brand name
Generics
Biotech (without year 2002)
Brand name Avg. (13%)
Generics Avg. (8%)
Biotech Avg. (26 %)
Reinhardt's Avg.( 13%)
R&D%
0.3
0.2
0.1
0
1994
1996
1998
2000
2002
2004
Year
Fig. 4 Operating income for different types of companies
2006
J Pharm Innov (2008) 3:30–40
35
Fig. 5 General expense for
different types of
companies
General Expense
Brand-name
Generics
Biotech
Brand-name Avg. (31%)
Generic Avg. (18%)
Biotech Avg. (21%)
Reinhardt's Avg. (35%)
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
1994
1996
1998
2000
2002
2004
2006
Year
biotechs, however, shows more fluctuations than brandnames as well as generics.
Relationship Between COGS, R&D, Operating Income,
and General Expense
The data were used to study the relationships between
COGS, R&D, operating income and general expense. Only
the following strong correlations were found from the data:
COGS and R&D for brand name, and COGS and operating
income for generics. The results are discussed in this section.
Brand-name Companies
The link between manufacturing costs, price of drugs, and
profits, etc. has recently attracted much attention [1, 7]. In
spite of the recent work by Vernon [7], the question is often
raised whether savings resulting from improvements in
manufacturing efficiencies would have any impact on
pharmaceutical prices or investment in R&D for discovery
of new drugs, or instead, would they merely increase the
profits of these pharmaceutical companies. This question
may in fact have been the main obstacle to public funding
of research that would result in reduction of COGS for
pharmaceutical companies. Although an in-depth answer to
such questions is beyond the scope of this paper, we
explore here the historical quantitative link between
manufacturing costs, profits, R&D, and general expenses.
Vernon [7] discusses two scenarios in which the various
expenses and incomes are related to each other. In scenario
1, Vernon has proposed that, by lowering the COGS
through improvement in manufacturing efficiency of
pharmaceutical companies, the market price for pharmaceutical products will decrease with all the other costs kept
constant. This will lead to gains in consumer surplus [7]. In
scenario 2, the market price is not lowered, and the
decrease in COGS, as a result of improved manufacturing
efficiency, will positively impact R&D spending of pharmaceutical companies, which will also eventually benefit
the consumers. This implies that if COGS can be reduced,
the pharmaceutical companies will invest those savings into
discovery of new therapies for unmet medical needs. In
fact, Vernon’s economic model calculates that the overall
gain in consumer surplus is higher in scenario 2 than in
scenario 1. Vernon also suggests that the actual effect of
improving manufacturing efficiency possibly exists between the above two scenarios.
Trend lines (linear or polynomial), fitted to the raw data
set using the standard function of Microsoft Excel, show
that there are significant changes in COGS%, operating
income, general expenses, and R&D% (Appendix 2
Fig. 13) for brand-name companies. The COGS% data
exhibit a decreasing trend, while the R&D% data exhibit
an increasing trend over the time period used in the study
(1994 to 2005). The operating income and general
expenses, however, show rising and falling trends,
respectively, with small gradients. Other expenses, such
as taxes and depreciation, show a flat profile.
The t test performed on the data, shown in Appendix 2
Table 4, supports the observation that, between the two subperiods of 1994 to 1999 and 2000 to 2006, there is a
significant decrease in the average COGS% and general
expenses, a significant increase in the average R&D% and
operating income, and no significant change in the average
values of taxes and depreciation.
The correlation between COGS% and R&D% was found
to be −0.93, showing a strong negative correlation between
the two. Figure 6 shows the trends in COGS% and R&D%
over the years. The fitted equations along with the data
points are also provided along with the R2 values, which
36
J Pharm Innov (2008) 3:30–40
% of sales
Biotechs
18%
y = -0.6038x + 0.2852
16%
R2 = 0.8645
R&D
14%
12%
10%
8%
6%
19%
21%
23%
25%
27%
29%
31%
COGS
Fig. 6 Relationship between COGS and R&D—brand names
indicate the goodness of the fit. A decreasing trend for
COGS% and increasing trend for R&D% can be observed.
Thus, the strong correlation along with the trends supports
Vernon’s scenario 2 that, during the recent past, for brandname companies, savings in COGS has positively impacted
investments in R&D, resulting in much-needed innovations
and future societal health benefits.
Trend lines for brand name companies
Fitted line
COGS%
R&D%
y ¼ 0:0013x þ 0:0071x þ 0:276
y ¼ 0:0004x2 þ 0:0004x þ 0:1067
2
R2
0.75
0.81
Appendix 2 Fig. 15 shows the variation in different
expenses over the studied period for biotechs.. The 2002
data for general expenses and R&D% showed unusually
high deviations due to the data from AMGEN and hence
were considered as outliers and removed for further
analysis (Fig. 10). Trend lines were fitted to the reduced
data. There is large fluctuation in the data, and the trend
lines do not fit well with the data points. A t test was done
to find the significance of changes in the various factors, as
shown in Appendix 2 Table 6. Only depreciation increases
significantly over the years. Thus, the analysis suggests that
for biotechs, all the expense categories such as COGS%
and R&D% have remained almost constant, other than
depreciation which has increased for the period studied.
General Observations
Aside from the key observations given above, the following
are some general observations based on analysis of the data:
–
Generic Companies
Fitted line
COGS%
R&D%
y ¼ 0:0005x þ 0:0038x þ 0:5616
y ¼ 0:0009x2 þ 0:0002x þ 0:0871
2
R2
0.63
0.59
–
–
% of sales
25%
y = -1.0304x + 0.6597
20%
Operating Income
Similar trend lines were also fitted to the data set over the
studied period (1994–2005) for generics. A decreasing
trend was observed for COGS% and general expenses,
while a rising trend was observed in operating income. The
trend for R&D% shows almost a flat profile. The
observations in trends were also supported by t test as
shown in Appendix 2 Table 5, where significant changes
were only observed in COGS%, general expenses, and
operating income. During the period under study, for
generics, the COGS% and general expenses decreased,
while operating income increased.
The correlation between COGS% and operating income
was found to be −0.92, showing a strong negative
correlation between the two observed in Fig. 7. From the
figure, it can be observed that the COGS% showed a
decreasing trend, and operating income showed an increasing trend over the years. Therefore, for generics, it appears
that the reductions in COGS% may be contributing to
increase in operating income.
Trend lines for generics
The total sales of the brand names were about 90%
of the total sales of all the companies in this study;
thus, the performance of the brand names has a
significant impact on the total market. This is in line
with the IMS global pharmaceutical market forecast
[6]. Therefore, an economic model, such as the one by
Vernon for the brand names, will be a good predictor of
the overall pharmaceutical market behavior.
For generics, the manufacturing expense is nearly 50%
of the total sales revenues as compared to about 27%
for that for brand names. It was surprising to note that
the COGS% for biotechs was the lowest.
Among all the three categories of pharmaceutical
companies, for biotechs, the expenditure on R&D%
as a percentage of sales is the highest. The biotechs
appear to invest the most in R&D with an average
R2 = 0.8528
15%
10%
5%
0%
40%
45%
50%
55%
60%
65%
COGS
Fig. 7 Relation between COGS and operating income for generics
J Pharm Innov (2008) 3:30–40
–
COGS%
Merck and Co.
General Expenses
Operating Income
70%
R&D%
60%
% of sales
–
R&D% of 27. As expected, R&D% for generics is the
smallest, nearly one third that of biotechs.
Operating incomes for brand names and biotechs are
nearly equal and higher than those for generics.
The general expense for brand names is nearly equal to
the industry average reported by Reinhardt in 2001 and
is much higher than those for generics and biotechs.
37
50%
40%
30%
20%
10%
0%
1992
1994
1996
Conclusions
The following conclusions can be drawn from the study:
–
–
–
2000
2002
2004
2006
Year
Fig. 8 COGS%, general expenses, operating income, and R&D% for
Merck and Co.
from 1994 to 2002 but dropped sharply to 15% in 2003 and
stayed close to that value for the next 2 years. Also, the
value of 15% in the past 3 years is significantly different
from the industry average. General expenses, operating
income, and R&D% showed an increase after 2002.
Although no explanation has yet been provided for the
sharp changes in the trends for Merck, the following factors
may have contributed these deviations. In 2001, the
divestiture of Medco Health Solutions (MHS) from Merck
and Co. was completed. The COGS% data for MHS is
shown in Fig. 9. As MHS is primarily focused on
distribution of drugs, the most important part of its
expenses is COGS (more than 90%). Therefore, before the
divestiture Merck’s COGS% was significantly higher than
the industry average and after the divestiture, it reduced
significantly. However, that does not explain why the COGS
% was reduced below the industry average after 2002.
In 2004, Merck withdrew Vioxx from the market due to
the revelation of potentially life-threatening side effects. It is
also possible that, because of Vioxx withdrawal, the company
may have made certain decisions to safeguard against large
litigation costs. We expect that, over an extended period, the
various expenses will stabilize to industrial averages.
Figure 10 shows the R&D% and general expense for
biotechs. There is a sudden rise in the R&D% and dip in
general expenses in the year 2002 due to high R&D%
120%
Medco +Merck
100%
Medco
Merck
80%
COGS%
–
As expected, it appears that a significant difference in
how the three different categories of pharmaceutical
companies, viz. brand names, generics, and biotechs,
spend their revenues.
For brand names, COGS per unit sales showed a significant
decrease, while R&D per unit sales displayed a significant
increase during 1994 to 2005. Moreover, COGS% was
found to be negatively correlated with R&D% across time,
suggesting that the cost savings associated with a reduced
COGS may have been budgeted for increased expenditure
on research and development for the discovery of new
therapies, as predicted by Vernon [7]. This work did not
consider alternative hypotheses for increased R&D expenditures during the period studied.
For generics, COGS% showed a significant decrease,
while operating income showed a significant increase
over the years. COGS% was found to be strongly
negatively correlated to operating income. Generics are
not involved in the discovery of new therapies, so their
R&D expenditure is likely more geared toward developing formulations of already approved drugs whose
patents are about to expire. Thus, for generics, the
savings from COGS is probably not invested in
research and appears to result in increased profits.
For biotechs, most of the factors studied do not show a
significant change over the years except for depreciation.
Fluctuations were found in the data compared to brand
names and generics, but the data failed to show any trend.
Interestingly, depreciation is significant for these companies, which might be an effect of the high capital
investment required to manufacture biotech products.
1998
60%
40%
Appendix
20%
Appendix 1: Anomalies in the Data
0%
1992
The data for various categories for Merck & Co. are shown
in Fig. 8. The COGS% increased steadily from 36% to 60%
1994
1996
1998
2000
Year
Fig. 9 COGS% trend
2002
2004
2006
38
J Pharm Innov (2008) 3:30–40
Biotech
IVAX Corporation
COGS%
R&D%
60%
Operating Income
General Expenses
100%
R&D
40%
80%
30%
60%
20%
10%
0%
-10%1992
1994
1996
1998
2000
2002
2004
2006
% of sales
% of Sales
50%
Year
40%
20%
0%
1994
-20%
Fig. 10 R&D and general expense per unit sales for biotechs
1996
1998
2000
2002
2004
-40%
-60%
Year
Par Pharmaceuticals
COGS%
Operating Income
100%
R&D
80%
60%
% of sales
expenses from Amgen Inc. We have not uncovered any
explanation for this variation. This data point was considered as an outlier and excluded from the calculation of the
overall R&D% and general expenses for biotechs.
As shown by the time series data in Fig. 11, Bristol
Meyers Squibb also experienced a similar setback in 2001,
resulting in higher manufacturing costs in that year. This
anomaly was apparently due to accelerated depreciation,
asset impairment, and restructuring expenses, as explained
in the 2004 annual report.
40%
20%
0%
The 1996–1997 Effect
1994
-20%
1998
2000
2002
2004
2006
-40%
Year
COGS%
Mylan Laboratories Inc
Operating Income
R&D
60%
50%
% of sales
An anomaly was observed for some of the generics during
the years 1996 and 1997 (Fig. 12). In this period, these
companies had sharp increase in COGS% and a corresponding decrease in operating income.
The plots show a dramatic change in COGS% and
earnings in 1996–1997. Some companies have provided
explanations in their annual reports. For example, IVAX
Corporation has provided the following explanation:.
“During 1996, certain national drug wholesalers instituted
programs designed to provide cost savings to independent
retail pharmacies on their purchases of certain Generic
pharmaceutical products. Pursuant to the programs, independent retail pharmacies generally agreed to purchase their
1996
40%
30%
20%
10%
0%
1994
1996
1998
2000
2002
2004
2006
Year
Fig. 12 1996–1997 effect for three generics, IVAX Corporation, Par
Pharmaceuticals, and Mylan Laboratories
Bristol-Meyers Squibb
COGS%
General Expenses
50%
Operating Income
R&D%
% of sales
40%
30%
20%
10%
0%
1994
1996
1998
2000
2002
Year
Fig. 11 Bristol Meyers Squibb time series plot
2004
2006
requirements of Generic pharmaceutical products from one
wholesaler and permitted the wholesaler to select the
product suppliers. Each wholesaler encouraged Generic
drug suppliers to participate in its program by offering to
purchase the wholesaler’s requirements of particular products from a single supplier. The programs encouraged
Generic drug suppliers to aggressively bid to be the
exclusive supplier of products under the programs. The
existence of the programs also resulted in reduced prices to
non-wholesaler customers. As a result, the Generic drug
industry experienced a significant reduction in the prices
charged by suppliers for many of its products during the
15%
10%
5%
0%
1994
1999
2004
Year
Fig. 13 Trend analysis for various expenses and incomes for brandname companies
COGS
Generics companies
General Expenses
Operating Income
R&D
DEPREC
60%
TXT
% of sales
50%
40%
30%
20%
10%
0%
1994
1996
1998
2000
2002
2004
2006
Year
Fig. 14 Trend analysis for various expenses and incomes for generics
6.6%
6.3%
−0.3%
0.20516
Taxes
General expense
31.9%
30.7%
−1.2%
0.18072
17.9%
19.4%
1.5%
0.034339
20%
11.2%
14.3%
3.1%
0.00381
25%
28.1%
24.2%
−3.8%
0.03573
% of sales
30%
Average (I, 1994–1999)
Average (II, 2000–2005)
Change (II−I)
Probability (p)
35%
Operating income
40%
R&D %
COGS%
Operating Income
R&D%
General Expenses
Taxes
Deprec%
Poly. (General Expenses)
Poly. (COGS%)
Linear (Operating Income)
Poly. (R & D %)
Brand Name companies
COGS%
Appendix 2: Trend Analysis
Table 4 The t test for the variation of the different costs as a percentage of sales between two sub-periods (1994–1999 and 2000–2005) for brand names
second and third quarters of 1996. Other wholesalers have
commenced or are expected to implement similar programs,
and such programs may be expanded to other product lines
and customer groups. Also during 1996, the Company
experienced increased competition some of its more
important domestic Generic pharmaceutical products as a
result of product approvals obtained by competitors.”
Although not explicitly stated, this explanation may be
applicable to all generic drug manufacturers.
For IVAX, the increase in COGS is also marked by a
restructuring program, which may explain the gradual
decrease in COGS over the following years. Although the
above summaries appear to be reflective of what has
occurred, a compelling explanation backed by numerical
analysis is yet to be seen.
4.0%
4.9%
0.8%
0.15001
39
Depreciation %
J Pharm Innov (2008) 3:30–40
40
J Pharm Innov (2008) 3:30–40
Table 5 The t test for the variation of the different costs as a percentage of sales between two sub-periods (1994–1999 and 2000–2005) for
generics
Average (I, 1994–1999)
Average (II, 2000–2005)
Change (II−I)
Probability (p)
COGS%
R&D%
Operating income
General expense
Taxes
Depreciation %
55%
49%
−5%
0.013304
7%
8%
2%
0.215482
9%
15%
6%
0.0078548
20%
17%
−4%
0.011321
5%
6%
1%
0.3284274
4%
4%
0%
0.56438
% of sales
Biotech companies
40%
35%
30%
25%
20%
15%
10%
5%
0%
1994
1996
1998
2000
COGS%
General Expenses
Taxes
Depre%
Operating Income
R&D%
Poly. (R & D %)
Linear (General Expenses)
Poly. (Operating Income)
Linear (COGS%)
Linear (Taxes)
Linear (Depre%)
2002
2004
2006
Year
Fig. 15 Trend analysis for various expenses and incomes for biotechs
Table 6 The t test for the variation of the different costs as a percentage of sales between two sub-periods (1994–1999 and 2000–2005) for
biotechs
Average (I, 1994–1999)
Average (II, 2000–2005)
Change (II−I)
Probability (p)
COGS%
R&D%
Operating income
General expense
Taxes
Depreciation %
13%
14%
1%
0.716411
28%
24%
−3%
0.097504
23%
21%
−2%
0.0718329
21%
26%
5%
0.3020194
9%
9%
0%
0.734358
6%
9%
3%
0.005236
References
1. Reinhardt UE. Perspectives on the pharmaceutical industry. Health
Aff. 2001;20(5):1363–70.
2. Kaiser Family Foundation Report, September 2007.
3. Suresh P, Basu PK. Improving pharmaceutical product development and manufacturing: impact on cost of drug development and
cost of goods sold of pharmaceuticals. Pharmaceutical Technology
& Education Center, Purdue University, February 2006.
View publication stats
4. PhRMA Report. What goes into the cost of a prescription drug? 2006.
5. Abboud L, Hensley S. Factory shift: new prescription for drug
makers. The Wall Street Journal 2003;September 3.
6. IMS. Global pharmaceutical market forecast. 2008.
7. Vernon JA, Keener HW, Trujillo AJ. Pharmaceutical manufacturing
efficiency, drug prices, and public health: examining the casual
links. Drug Inf J. 2007;41:229–39.
8. Wharton Research Data Services (WRDS). Compustat. http://wrds.
wharton.upenn.edu/.