Environmental Regulation and Innovation: A Panel Data Study

1
Environmental Regulation and Innovation A Panel
Data Study

• Adam B. Jaffe and Karen Palmer

Presented by Eliza Finley May 1, 2008
2
Porter Hypothesis

• Stricter environmental regulations provide
  incentives for firms to
• Develop new and less costly ways of reducing
  pollution
• Develop methods of production that eliminate
  emissions while cutting costs
• Stricter regulations in one country
• increase in innovation
  
• country becomes a net exporter of newly
  developed environmental technologies

3
Variations of Porter Hypothesis

• Narrow Version
• Outcomes are regulated but not processes
• Certain types of environmental regulation
  stimulate innovation
• Weak Version
• Regulations place constraints on firms profit
  maximizing behavior which will induce them to
  innovate in order to meet the constraint at a
  lower cost.
• Regulation will stimulate certain kinds of
  innovation
• Strong Version
• New regulation will induce firms to find new
  products/processes that comply with regulations
  while increasing profits (Paid lunch!)

4
Wheres the evidence?

• Evidence of Porter Hypothesis is largely
  Anecdotal
• Phase-out of CFCs led DuPont to develop less
  harmful substitute
• Clean Air Act regulations led to development of
  new paints and coatings with lower VOC content
• Regulations on water emissions in Sweden led to
  innovations in their paper production process
• Not much empirical assessment of impact of
  regulation on innovative activity

5
Previous Literature

• Oates et. al (1993)
• pollution tax rate increases adoption of more
  efficient abatement technologies
• Schmalensee (1994)
• RD devoted to environmental compliance may come
  at expense of more profitable research efforts
• Biglaiser and Horowitz (1995)
• Increase in emission tax rate will decrease
  aggregate research when inefficient firms adopt
  newly developed efficient technologies from more
  efficient firms
• Lanjouw and Mody (1993)
• Increases in environmental compliance costs lead
  to increases in patenting of new environmental
  technologies

6
The Model

• Purpose To analyze the relationship between
  stringency of environmental regulation and
  innovative activity in manufacturing firms
• Two Different Measures of Innovative Effort
• Industry-wide expenditures on RD
• Total number of successful patent applications
• Regulation stringency measured by environmental
  regulatory compliance expenditures

7
RD as Dependent Variable

• Measuring industry level RD is tough
• No data showing how real cost of scientists and
  research equipment vary across industries
• Result? This very crude equation
• log (RD)it ß1log(value added)it
  ß2log(government RD)it ß3log(PACE)i,t-1
• aiR µtR ?it
• Variables
• i denotes industries
• t denotes years
• Value Added
• Precludes spurious correlation between RD and
  pollution control expenditures due to variation
  in industry size
• Size-scaling variable
• Government RD
• One of the few measurable external drivers of RD
  at the industry level
• lagged PACE
• Pollution control expenditures

8
Patents as Dependent Variable

• Measuring patent data as a proxy for innovative
  output
• log(patents)?1log(value added)it ?
  2log(foreign patents)it
  
• ? 3log(PACE)i,t-1 aiP µtP ?itP
• Variables
• Value Added, PACE, Year and Industry Dummies
• same reason as in RD equation
• Foreign Patents
• Successful U.S applications in year t by foreign
  corporations
• Controls for variations in patent incentives
• incorporates idea of whether U.S. regulations
  causes U.S. firms to become more innovative
  relative to foreign competitors

9
Data Sources

• Environmental Compliance Cost Data
• PACE survey
• covers manufacturing firms
• collects information on capital costs of
  complying with environmental regulations
• RD Data
• Survey by Census for the National Science
  Foundation (NSF)
• tabulated at level of two- or three- digit SICs
• Government RD measured by employee-weighted
  fraction of firms in industry that report
  receiving government research funding
• Patent Data
• Industry panel of U.S. patents by year of
  application
• problems with classifying patents by industry
• solution Use totals for U.S. corporations

10
Results RD Regression

• Two different forms of lagged pace variable
• single year lagged value (LPACEL1)
• moving average of previous five years (LPACE5)
• Two models
• Pooled vs. Industry Fixed Effects
• Coefficient on each lagged PACE variable
• significant and negative in pooled model
• significant and positive in fixed-effects
• as (industry dummies) are positively correlated
  with RD but negatively correlated with PACE
  expenditures
• high-tech industries are less pollution-expenditur
  e intensive than low-tech industries
• Therefore, must control for industry-specific
  effects to see true effects of PACE on RD
• within-industry elasticity of RD with respect to
  lagged PACE expenditures is around 0.15

11
More Results from RD Regression

• Positive relationship between growth in RD and
  growth in pollution control expenditures
• Radio and TV receiving equipment, other
  transportation, and other instruments have
  highest rates of growth in both PACE and RD
• Ferrous metals, nonferrous metals, textiles, and
  lumber all have lowest RD growth and lowest PACE
  growth

12
Results Patents Regression

• Foreign patenting and domestic value added have
  significant positive effects
• None of coefficients on lagged PACE variables is
  statistically significant
• Regulatory compliance costs have no detectable
  impact on patenting activity

13
Conclusion

• No statistically significant relationship between
  patenting and regulatory compliance expenditures
• Significant positive relationship between
  regulatory compliance expenditures and RD
  activity by industry
• consistent with weak version of Porter
  hypothesis, that environmental regulation will
  stimulate certain types of innovation
• Insights are limited
• These are not strictly outcome-oriented
  environmental regulations
• No conclusion about narrow Porter hypothesis
  where regulations of this type would stimulate
  innovation
• Results do suggest that weak Porter hypothesis
  may be true
• Environmental regulation will stimulate certain
  types of innovation
• But not when using patenting to measure
  innovation

14
Conclusion

• Disincentives to RD caused by stricter
  regulations may be overcome by high returns to
  new pollution-control technology
• We do not know why RD increases
• RD may increase due to stricter regulation
  diverts RD away from regular RD efforts in
  order to comply with regulation (weak porter
  hypothesis)
• But RD may also increase because firms have been
  shocked into thinking up new and creative ways
  to produce goods that comply with regulation and
  reduce costs (strong Porter hypothesis)
• We do not know if this increased RD was
  productive
• There can be no major policy implications If
  increased RD only facilitates compliance but not
  actual productivity

15
Future Research

• Instead of PACE, use
• regulatory impact analyses of proposed
  environmental regulation
• number of pages in federal registry devoted to
  environmental regulations (for each industry)
• Focus on specific regulations and see the
  industrys response

16
Discussion Questions

• What are some problems with using patents as a
  measure for innovation?
• The RD results show that the PACE variable is
  positively and significantly correlated with RD,
  but this varies across industries. Comment.
• Is this study useful? Is it too broad?
• Is there a better measurement for regulation
  stringency?
• What are some problems with measuring innovation
  using RD?
• Jaffe and Palmer use both RD and patents that
  are unrelated to environmental innovation. What
  are some consequences of this approach?
• Whats the point?

17
Hmm