Business%20R

1
Business RD Intensity in Canada and the United
StatesDoes Firm Size Matter?

• Presentation to
• The 2008 World Congress on National Accounts and
  Economic Performance Measures for Nations
• Arlington VA
• Daniel Boothby
• Bernice Lau
• Thitima Songsakul
• May 16, 2008

2
Outline

• Introduction
• Decomposition of Business Expenditures on RD
  (BERD) intensity Five factors to examine
• size class distribution of firms
• weight of firm size classes in value added
• RD intensity for RD performers in each size
  class
• RD incidence of firms performing RD in each
  size class and
• ratio of value added by the average performing
  firm to the average value added of all firms in
  each size class.
• Summary of findings
• Directions for further research

3
Introduction

• Canada lags other advanced economies in its
  innovation performance on a number of indicators,
  such as low BERD intensity.

BERD intensity across OECD economies, 2004,
(percentage of total GDP)
Industry-financed BERD intensity across OECD
economies, 2004, (percentage of value added in
industry)
Source OECD MSTI (2006/2) Table 34
Source OECD MSTI (2006/2) Table 24
4
Is smaller firm size the reason for lower RD
intensity?

• Smaller firm size in Canada is often mentioned as
  a possible reason for poor innovation performance
  and low intensity of business RD.
• Two possibilities
• Canada might have too many small businesses
  that is, small firms might be more prevalent in
  Canada than in the U.S.
• Small businesses in Canada may do less RD, while
  larger Canadian firms do as much RD as their
  U.S. counterparts.

5
Objective

• Our goals
• Assess whether smaller firms are indeed a
  significant source of low business RD intensity.
  
• Describe further measures that can be carried out
  using existing data and that will explain the
  relation between firm size and business RD
  performance in Canada.
• Suggest directions for future research.
• We use a formal decomposition to account for the
  size-related factors that affect aggregate BERD
  intensity.
• RD intensity, RD incidence and firm size
  distribution are factors in the decomposition.

6
Five factors to examine

• A. the size class distribution of firms ni /n

B. the average-firm value-added weight by size
class bvai /bva
C. the RD intensity for RD performers in each
size class, berdip
D. the percentage of firms performing RD in each
size class (i.e., RD incidence) nip/ni and
E. the ratio of value added by the average
performing firm to the average value added of all
firms in each size class bvaip/bvai.
7
A. The size class distribution of firms (ni /n)
Number of business enterprises in Canada and the
U.S. by employment size class, 2002

• We begin with the relative frequency of firms by
  size class (ni /n) where size is defined by the
  number of employees.
• U.S. and Canadian data on firm size for the
  entire business sector are available.
• There is little U.S.-Canada difference in the
  distribution of firms by size class.
• There is a much larger US.-Canada difference in
  the employment shares of various firm sizes.

Employment shares in Canada and the U.S. by
employment size class, 2002
Source Canada-Statistics Canada (2006a) Table 1a
and 5a, U.S.-U.S. SBA (2006) Note
Zero-employee firms cannot be removed from
Canadian LEAP data. For the U.S., there are 13.5
percent of enterprises with zero employment.
8
Why do firms in the U.S. account for a larger
share of business sector employment?

• The greater share of employment of U.S. firms
  with 500 employees is primarily due to the fact
  that these firms are bigger in the U.S. than in
  Canada.
• This has implications for value-added shares
  (another factor of the decomposition) suppose in
  each country value added per employee is the same
  for all firm sizes. Then the value-added share in
  each firm size would be the same as its
  employment share. Consequently, the largest firms
  would account for more of the value added in the
  U.S.

9
B. The average value added weight (bvai /bva)

• No direct measure so we will discuss evidence on
  value-added shares of establishments (not firms)
  in manufacturing since we do not have firms
  economy-wide.
• Small establishments in Canadian manufacturing
  account for a greater share of employment and
  value added than those in the U.S.
• We have seen employment-share comparisons that
  would lead us to think that this is true for
  firms economy-wide.
• While small firms dominate in the number of
  businesses in both countries, they do not
  dominate in measures of economic weight.

10
C. RD intensity for RD performers in each size
class (berdip)

• The biggest Canada-U.S. differences in RD
  intensity for performers are for large firms, not
  small firms.
• There is a pattern of overall higher intensity in
  all size classes in the U.S.
• We need to be carefulneither numerator nor
  denominator are the same for the U.S. and Canada.

11
D. RD incidence (ni p/ni)

• The smallest size class shows much lower RD
  incidence than larger enterprises in both
  countries.
• However, Canadian small firms are not lagging
  their U.S. counterparts in terms of RD
  incidence.
• In fact, these results show very similar
  incidence of RD in Canada and the U.S. in all
  size classes.

12
E. The ratio of value added by the average
performing firm to the average value added of all
firms (bvaip/bvai)

• There is little difference in the number of
  employees in performing and non-performing firms
  in the U.S. at most performing firms have 10
  more employees.
• Thus large productivity differences would be
  required for bvaip/bvai to be large. As a
  guesstimate, we think it is unlikely that this
  factor will be greater than 1.5 in any size class
  in Canada or the U.S.

13
Summary of findings firm size and Canada-U.S.
BERD intensity gap

• Small firms are not more prevalent in Canada.
• Greater prevalence of small firms cannot
  contribute to the gap.
• Value-added share of large firms is bigger in the
  U.S.
• This is likely to contribute to the gap, but it
  takes large changes in the value-added shares of
  size classes to have much of an impact on overall
  BERD intensity.
• BERD intensity tends to be lower in Canada than
  in the U.S., more so for large firms.
• Small firms have little to do with the reasons
  for the BERD intensity gap.

14
Summary of findings firm size and Canada-U.S.
BERD intensity gap (contd)

• There are only small differences between Canada
  and the U.S. in the percentage of firms in each
  size class that perform RD.
• Differences in RD incidence cannot contribute
  significantly to the gap.
• The ratio of value added by the average RD
  performer to the average firm is not likely to
  differ much among size classes nor between Canada
  and the U.S.
• Little impact of differences in this ratio on
  this gap.

15
Small firms are not responsible for low RD
intensity in Canada

• The RD share of all firm size classes except the
  largest (5000 employees) is much bigger in
  Canada than in the U.S.
• Lower intensity for very large firms in Canada
  relative to the U.S. may contribute to low BERD
  intensity.

3
16
Directions for further researchData development

• Linking the establishments reporting in the
  Census of Manufacturers and the companies
  reporting in the Research and Development in
  Canadian Industry (RDCI) Survey to the firms of
  LEAP or of the Business Register. This would
  allow direct computation of RD incidence,
  intensity, the relative value added of performers
  and all firms within a size class and the
  value-added share of a size class.
• Other projects would be to link LEAP and the RDCI
  to provide better information on RD incidence
  and to use LEAP to provide a finer division of
  large firms by size class.
• In the long term, a longitudinal, integrated data
  set on business strategies, RD activities,
  value-added and employment would be a valuable
  asset for researchers.

17
Directions for further researchBroad directions

• If very large firms are in fact multinational
  firms, a better understanding of how
  multinationals choose where to conduct RD
  activities would be useful.
• From the RDCI survey, about 15 of Canadian BERD
  is funded by foreign sources. Are these research
  mandates from multinational parents?
• If so, is this large part of Canadas already low
  BERD at risk from emerging economies?

18
Conclusion

• In our view, we have ruled out small firms as an
  important factor in the BERD intensity gap
  between Canada and the U.S. If firm size plays a
  role in the gap at all, it is differences in the
  presence of very large firms that may matter. Our
  suggestions for research directions are aimed at
  better understanding the role of these firms in
  BERD intensity.