The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography...

22
Michelle Gittelman Department of Business and Management Rutgers Business School Newark-New Brunswick, New Jersey The Geography of Markets for Technology: Evidence from Bio- Pharmaceuticals

Transcript of The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography...

Page 1: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Michelle Gittelman Department of Business and Management

Rutgers Business School Newark-New Brunswick, New Jersey

The Geography of Markets for Technology: Evidence from Bio-

Pharmaceuticals

Page 2: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Approval

New

Drugs

Discovery Clinical development

Preclinical development

Marketing

3-6 Years 6-7 Years ½ - 2 years

5,000-10,000 compounds

250 compounds

5 compounds

1 drug

The pharmaceutical industry value chain Old Organizational Paradigm

Page 3: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

New

Drugs

Discovery Clinical development Approval

Preclinical development

Marketing

The pharmaceutical industry value chain New Organizational Paradigm

Source: Ed Holmes, UCSD, Presentation to Institute of Medicine 2005

Page 4: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Geography and licensing markets

Biotech represents not only a technological and organizational shift, but also a shift in the geographic locus of R&D

– Biotech clusters are often outside the geographic footprint of traditional pharma

Do licensing market operate effectively across space – or does co-location matter for licensing between small biotech and large pharma? (Alcacer, Cantwell, Gittelman, 2010)

Interviews with licensing execs Analysis of licensing deals What factors influence the choice of local vs. distant partnerships? (Gittelman, 2011)

Serendipitous encounters in local spaces vs. scientific communities that span geographic distance

Page 5: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

5

Does co-location matter for licensing?

We know that proximity facilitates markets for upstream inputs in biotechnology (scientists, universities, venture capital). . .

We conceptualize licensing as downstream product markets for biotechnology firms and hypothesize that they may also be subject to similar co-location pressures.

Problems in contracting for technology (Gambardella, 2008, Gans, Hsu, Stern 2008):

Search costs

Information asymmetries

Opportunism, monitoring and coordination costs

IP uncertainty

Technological uncertainty

Knowledge transfer

If co-location reduces

these frictions, we

should see proximity

increasing the

likelihood of licensing

Page 6: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

6

Interviews suggest that proximity is unimportant in licensing markets

1. Big pharma searches globally for in-licensing opportunities.

– Licensing a formal organizational activity

– “Our experts know where the frontier is. . .we follow up on things that pique their interest.”

– “We are border blind as to where an asset comes from. There could be great scientists anywhere”

2. Patent protection lower geographic bias – “We don’t care if it’s coming from the US or Brazil – a good compound

can come from anywhere - as long as the drug is patented in the US or Europe, where we intend to sell it”

Page 7: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

3. Biotechs around the world sell globally to big pharma. – “If you’re a small firm in India - all it takes is one BD guy sitting

in Manhattan.”

4. Post-license collaboration frequently discouraged by pharma

Patents give detailed roadmap to technological and legal landscape

Extensive due diligence before the license mitigates need for post-license collaboration

Big pharma seeks to take control of the property – “we know how to do it [compound development], they think they do -- but they don’t”

Virtual research teams can conduct modularized projects with little need for frequent interaction or coordination

Interviews suggest that proximity is unimportant in licensing markets

Page 8: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Does localization matter to licensing?

Page 9: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

9

Data

• 36,646 drug compounds (Pharmaprojects) – Late 1980s-2006, including failed drugs – Ceased (80%), Currently Active (17%), Currently Launched (3%)

• All licenses • Type of firm/organizations developing and licensing the drug

– Global Pharma – Biotech firms – Small and mid-sized firms – Universities and tech transfer

Co-location - Any geographic overlap between two firms at the time of license at the city level

Page 10: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

10

Selection model

Colocation Model, Conditional Logit: Pr(Licenseijk) = f(Co-location, city leveljk)

i – compound j – biotech firm developing compound i k – one of 25 global pharma

Selection model, was drug ever licensed: Pr(Ever Licensedi) = f(Compound Age, development status, co-location with

any big pharma, fixed effects for therapy codes and each of the 25 Big pharma firms)

Page 11: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

11

All compounds

Licensed to firm j

Ever licensed

Co-location, firms i and j 0.17** (5.3)

Co-locate, any big pharma

0.13** (14.9)

Age 0.18** (120.8)

Ceased -0.03** (3.7)

Fully Launched -0.33** (14.7)

Fixed effects, Therapy codes Y

Obs 152975

Chi Sq 28.09

ρ -0.25

Robust Z statistics in parentheses, SE clustered on originating firm.

Page 12: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Does location matter?

“Pfizer's Research Technology Center (RTC) is located in Cambridge, MA, one of the richest scientific environments in the world, surrounded by more than 300 biotechnology companies and thousands of scientific innovators in world-leading research hospitals and academic institutions. A key element of our success is the growing, dynamic partnerships with these innovators to address the needs of patients worldwide.” Pfizer website

Page 13: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Geography of knowledge flows When do firms partner locally for alliance partners, and when do they seek partners in distant locations? Conceptualize local and distant spaces as distinct opportunity sets for acquiring external technology: Distant locations – Scientific networks are designed to span geographic distance. Distant partnerships reflect search of scientists for others working on similar problems: strategic, intentional search based on specific in-house R&D expertise. Local spaces – Opportunities for serendipitous encounters between individuals/firms who otherwise might not know of eachother. Firms prior broadscope (general) knowledge enable them to exploit local opportunities.

Page 14: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Geography of knowledge flows The more a pharma firm has specific

experience in a disease market, the more likely it will partner with a distant firm.

Geographically distant partnerships will be more similar with respect to in-house R&D than local partnerships

Geographically proximate partnerships correlated with general knowledge and technological variety

Page 15: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Data

1. Alliances – 1038 alliances involving US biotechnology firms and 14 of the

largest pharmaceutical companies, 1993-2008 (Recombinant Capital)

2. Distance between alliance partners – Minimum distance (in miles) between a biotechnology firm and

the closest R&D lab of the pharma firm at the time of alliance

3. Drug portfolios of firms (Pharmaprojects) – Measure in-house R&D prior to alliance

– Drugs by disease and by discovery platform (biotech vs. other)

Page 16: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

0

.00

05

.00

1.0

015

Den

sity

0 250 500 750 1000 1250 1500 1750 2000 2250 2500Mindist

Dashed red line = distances betw. allied firms only; solid blue = distances all firms

Distances between alliance partners and non-alliance partners

Page 17: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Co-location - Alliance partners are within 100 miles of eachother Firm Disease Experience % of firm’s drug portfolio in same

disease as listed on alliance

Firm Biotech experience % of firms’ drugs developed through biotech techniques

Number of prior alliances by firm Need for collaboration - “Co-development” or “Collaboration” on description of

alliance Stage of project: Research, Development, Preclinical, PhaseI, etc.

Logit regressions of co-location

Page 18: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Figure 3. Correlations between alliances and in-house drugs

Alliances < 100 miles and > 100 miles

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Wye

th

Eli Lilly

Nov

artis

Glaxo

Smith

Kline

Abbot

t

Bristo

l-Mye

rs S

quibb

Pfizer

Astra

Zenec

a

Hof

fman

n-La

Roc

he

Sanof

i-Ave

ntis

Mer

ck &

Co

Scher

ing-

Ploug

h

Bayer

John

son

& Joh

nson

Firm

Co

rrela

tio

n c

oeff

icie

nt

Alliance partners less than 100 miles Alliance partners greater than 100 miles

Page 19: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Logit models of co-location (all include fixed effects for pharma lab locations &

random effects for pharma firms)

Co-location=1 if firms are <100 miles Coeff SE

Disease experience pharma firm -3.0*** 1.44

Biotech experience pharma firm 9.23*** 3.25

Disease experience biotech firm 0.60** 0.31

Biotech experience biotech firm 0.04*** 0.31

Alliance experience, Biotech firm -0.06 0.07

Alliance experience, Pharma firm -0.09 0.12

Early Stage -0.43* 0.2

Collaboration 0.32* 0.20

Controls for Prior Alliances, disease specified

N 709

Wald Chi Square 103.4

Log Likelihood -346.7

Page 20: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

The geography of licensing

Despite global search, co-location increases the probability of licensing

Clusters that mix biotech firms & large pharma important nodes in markets for technology

Distant and local partnerships leverage different in-house knowledge and generate different complementaries with respect to in-house R&D

– Distant partnerships deepen technological expertise in specific disease markets

– Local partnerships associated with general knowledge and technological variety – speaks to the importance of proximity for serendipitous learning and exploration

Page 21: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

Thank you!

Page 22: The Geography of Markets for Technology: Evidence from Bio ... 1.3. Gittelman.pdf · The Geography of Markets for Technology: Evidence from Bio-Pharmaceuticals . Approval New ...

22

Number of

drugs

developed,

1990s - 2008

Outlicensed

drugs

Internationally

licensed*

Companies in

California and Mass 4479 23% 12%

Global Pharma

Firms (Top 14) 11846 12% 10%

Rest of World 18981 17% 12%

Total 35306 16% 11%

Bio-regions most likely to out-license but - most drug compounds are never licensed

*Internationally licensed drugs may also be domestically licensed

Source: Gittelman, Alcacer and Cantwell, 2008.