Factors in Corporate R&D Location: Where and Why?

Jerry Thursby

Emory University

Marie Thursby

Georgia Institute of Technology and NBER

Supported by the National Academies’ Government University Industry Research Roundtable (GUIRR) & the Kauffman Foundation

Objective:

The idea that the United States dominates cutting edge science and technology is increasingly challenged as the US share of patents and scientific awards declines and corporations report increasing reliance on offshore research and development (R&D)—either in their own new R&D facilities or collaboration with universities or other companies abroad. Indeed, recent discussions within GUIRR reveal that nearly all GUIRR industrial members are considering locating their next R&D facility or project outside the United States. Somewhat surprisingly, the reasons given were largely related—not to cost—but to the need for scientific expertise located elsewhere and the ease of accessing external expertise in other countries.

Our goal is to provide survey evidence that can serve as the basis for informed policy discussions on these issues. We will survey US and European based companies in order to identify key factors in R&D siting decisions and to develop a framework to evaluate the relative importance of these factors. Such data and analysis are critical for fruitful discussions of policy initiatives to enable the United States to attract and retain the advanced technology sectors that drive job creation, high productivity, and economic growth.

Project Description:

As increasing numbers of countries become technologically capable, the globalization of R&D capabilities is inevitable. The geographical location of industrial R&D investments, either virtual or "bricks and mortar," has a cascading and long-term effect on the economy of the region. Thus, it is not surprising that US policymakers view any loss in the attractiveness of the United States as the location for industrial R&D with some concern. The problem is that available evidence is anecdotal and while agencies such as the Bureau of Economic Analysis (BEA) and the National Science Foundation (NSF) provide aggregate data on R&D expenditure, there is no source of information related to the origins of R&D decisions. For informed policy discussions, it is important to understand the factors affecting these decisions and their relative importance.

The attached table lists nine types of factors that GUIRR members suggested are important in their R&D-siting decisions.

FACTORS	ISSUES
1. Support for Pre-existing Manufacturing Facilities	Technical support to solve manufacturing problems Product adaptation to local market
2. Proximity to New Markets (Product or Service)	Same types of support as in 1 Variety of factors in new facilities, including new markets, cheap labor, regulatory environment, etc.
3. Supply of Scientists or Technical Workers in Critical Areas	University expertise in relevant fields located in countries where R&D support is best (e.g. Nuclear engineering, metallurgy (China and Korea), thermodynamics, processing technologies) Cost of scientific expertise
4. Technology Transfer Landscape	Ownership of intellectual property Policies and financing for research collaboration Environment for spin-offs
5. Intellectual Property Protection	Patent enforcement post TRIPS Differential enforcement across countries
6. Legal & Regulatory Environment	Tort law Restrictions on research (e.g. stem cell) Regulatory hurdles (e.g. drugs)
7. Export Requirements	Ability to export tied to establishing local R&D facility Quid pro quo for access to local market (e.g. China) Quid pro quo for sale to government (e.g. aerospace)
8. Tax Policies	Tax holiday for new R&D facilities Differential taxes
9. Infrastructure	Financial Transportation Communication

These factors are not listed in any particular order, nor are they exhaustive. The proposed survey will focus first on identifying the importance of these and other factors and second on developing a weighting scheme and framework to guide policy discussions. The need for the latter should be apparent since all of these factors are likely to influence R&D decisions. Moreover, factors have very different policy implications. For example:

· R&D facilities to support manufacturing either to provide technical support for processes or product development may be desirable (contrary to notions from the popular press) particularly in the case of item 2, where both the manufacturing and R&D sites represent market expansion (as opposed to reallocation of production).

· Items 3 and 4 are perhaps the most surprising and troubling if, in fact, they are driving R&D decisions in many industries. As noted in item 3, several members gave examples of areas in the physical sciences where the preeminent expertise in now located outside the United States. Others, particularly representatives of electronics companies, cited increasing difficulties in dealing with US universities as a reason for increased collaboration with universities elsewhere and location of major R&D facilities to facilitate collaboration.

· Since the last round of the WTO, with the extensive revision of trade related intellectual property, and a phase in of patent protection in developing countries, it would not be surprising to find that an increased ability to protect intellectual property developed outside the United States is an important factor. Moreover, such a response would be the desired response to improvements in the multilateral policy environment.

· Items 5-8 represent other policies, most of which are examples of discriminatory policies that may be troublesome. For example, research restrictions on stem cell research will likely influence not only current, but also future R&D, as it may affect the future supply of scientists in this area.

It is anticipated that the factors relevant to basic research location may be different than those for applied research or development, and that the factoral weights may also differ depending on industry sector. Finally, the survey and subsequent analysis will need to consider current R&D decisions compared with historical outsourcing of R&D (as many companies located research facilities offshore in the 1960s and 70s). One of the important results of the study will be a better understanding of whether the movement towards greater R&D outsourcing in the last decade or two has been "incremental" or whether it is "accelerating" in intensity.

Survey Design and Analysis:

The survey design and analysis will be contracted to Emory University and the Georgia Institute of Technology, with Jerry Thursby and Marie Thursby, professors of economics and strategic management, respectively, as principal investigators. The project is envisioned as a 2 year project with the first year devoted primarily to survey design and administration, as well as preparation of an interim report of results. The second year will be devoted to statistical and policy analysis.

As part of the survey design process, individual CTO’s/CEO’s and key executives responsible for relocation decisions will be asked to preview the questions to ensure they are unambiguous to the intended survey respondents and can elicit a meaningful response (cognitive testing). The survey will be developed as a telephone instrument to achieve optimal response rate and maximize accuracy of the data. Current plans are each phone interview to last approximately 30 minutes and to be professionally coded for efficient data collection and analysis. Firm identity will be kept strictly confidential by the investigators.

Respondents will be recruited from the Industrial Research Institute (IRI), whose members comprise over 240 R&D performing companies in the U.S. and abroad. IRI’s Board has approved involvement in the project and Ross Armbrecht, IRI President, has held preliminary discussions with leadership of the European Industrial Research Management Association (EIRMA), with over 100 members, regarding participation by European companies. Direct involvement in this initiative by the leadership of IRI and EIRMA will help ensure high survey response rates, as will peer pressure from the GUIRR corporate membership, many of whom are also members of IRI.

Survey data will be augmented with data from the Census Survey of Industrial Research and Development in order to allow econometric analysis which controls for various firm characteristics (e.g., firm size, research “intensity”, etc.) in analysis of responses. A portion of these data are confidential and application to the BEA for access will be made. As we noted, any presentation of results will be done in such a manner as to protect confidentiality of the respondents.