Testing the Porter hypothesis: the effects of environmental investments on efficiency in Swedish industry


The main objective of this paper is to test the Porter hypothesis by assessing static and dynamic effects of environmental policy on productivity. According to the hypothesis, stringent environmental regulations have dynamic effects on firm performance, and these effects eventually generate profits that offset the adaptation costs. We extend previous analyses by using unique data on environmental protection investments in the Swedish manufacturing industry as a proxy for environmental stringency. These data enable us to separate environmental protection investments into pollution prevention and pollution control. This distinction is crucial since the hypothesis claims that it is investments in prevention that have positive dynamic effects on firm performance. To test the hypothesis, a stochastic production frontier model is estimated where firm inefficiency is a function of investments in environmental protection. In general, we find no support for the Porter hypothesis within the time frame of our study, indicating that environmental regulations lead to efficiency losses. This result is even stronger in the harshly regulated pulp and paper industry.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2


  1. 1.

    For example, Byung and Sickles (2004) studied 17 OECD countries from 1980 to 1990 and found that Sweden showed a relatively high productivity growth on average due to technological development and efficiency improvement. The reason for the productivity growth was that Sweden reduced CO2 emissions at the same time as its GDP increased (p. 580). However, their study does not really say anything about whether the efforts of reducing CO2 actually contributed to the GDP growth.

  2. 2.

    For instance, the Confederation of Swedish Enterprise claims that such a policy will cause dramatic changes in industry structure, and fears that Swedish industry will suffer considerable costs and lose competitiveness (Resvik and Furbeck 2005).

  3. 3.

    Porter and van der Linde (1995) use the term pollution control, which is synonymous to pollution treatment.

  4. 4.

    Investments in pollution control is defined as capital expenditures for methods, technologies, processes, or equipment designed to collect and remove pollution after its creation, prevent the spread of and measure the level of the pollution, and treat and dispose of pollutants. Investments in pollution prevention is defined as capital expenditures for new, or adaptation of, existing methods, technologies, processes, equipment designed to prevent or reduce the amount of pollution created at the source (Eurostat 2001).

  5. 5.

    Even if a positive relationship is found, improved technical efficiency, resulting in increased competitiveness and profits, does not necessarily fully offset the initial costs of adapting to regulations.

  6. 6.

    For a general introduction to this literature, see, e.g., Coelli et al. (2005), Kumbhakar and Lovell (2000) and Grosskopf (1993).

  7. 7.

    When “well-designed” environmental regulations result in complete cost neutralization, it is referred to as “a strong Porter effect”.

  8. 8.

    The advantage of stochastic production frontier models, compared to deterministic ones, is that the impact on output of random shocks due to variation in labor and machinery performance or to measurement errors can be separated from the contribution of variation in technical efficiency.

  9. 9.

    Battese and Coelli (1993) provide the likelihood function and its partial derivatives.

  10. 10.

    A likelihood ratio test supports the choice of a translog over a Cobb-Douglas specification.

  11. 11.

    According to Shadbegian and Gray (2006) other distributions can be used, e.g., truncated normal. In our case the truncated normal model was rejected as it gives rather unstable results depending on the modeling of heteroscedasticity in u. When estimating a time-invariant efficiency model without modeling heteroscedasticity, half-normal and truncated normal distributions give similar results judging from Kernel distributions of the estimated inefficiencies (u it).

  12. 12.

    Among the manufacturing industries, the pulp and paper, chemical, and basic metal industries are the largest investors in environmental protection, amounting to 30, 7, and 4 % respectively of total investments in environmental protection in 2004. The wood products and rubber and plastic industries are included since they had enough observations on environmental protection investments.

  13. 13.

    Implicit deflation factors have been derived from national account series of sector-specific value added in current and fixed prices. The national accounts use different deflation factors when transforming current prices to fixed prices (i.e., the production values are deflated by producer price index and the input values are deflated by the factor price index).

  14. 14.

    Since the write-offs reflect the annual loss of economic value in fixed assets, we have not deflated the net capital stock explicitly.

  15. 15.

    The capacity utilization rate is an average for the sector at the two-digit level (on the three-digit level for NACE20 and NACE21). Insufficient demand is the main reason for not utilizing full capacity (Swedish National Institute of Economic Research 2010).

  16. 16.

    See Olsson and Eberhardson (2003) for an evaluation of the environmental protection expenditures from a data quality and collection perspective, and for an English version of the questionnaire.

  17. 17.

    The response rate is typically lower for firms with fewer than 250 employees. The high response rate from 2001 and onwards reduces the risk for sample selection problems, and is due to the fact that the survey became better known among the responding firms and that some of the questions became compulsory.

  18. 18.

    About 40 percent of the environmental protection investments, in monetary terms, are made to control or prevent air pollution. Table 5 in Appendix exemplifies these investments for the industries in the analysis.

  19. 19.

    Since 1991, Sweden has had a CO2 tax that is also considered dynamically efficient.

  20. 20.

    Useful energy can be steam, hot water, or electricity produced in a boiler and used in production processes or heating of factory buildings. After deducting administrative costs (1 % of the revenues), the revenues are refunded to the same sources that paid the charge, but in proportion to output of useful energy.

  21. 21.

    The use of two lags reduces the number of observations from 1,294 to 524.

  22. 22.

    As mentioned in Sect. 4, the estimated parameters have been multiplied by −1 and, hence, their signs are to be interpreted in terms of effects on efficiency.

  23. 23.

    Table 6 in Appendix shows the results for a stochastic frontier model that does not specify heteroscedasticity in inefficiency. The results indicate technological progress as the trend parameter is positive and significant. The results also show significant differences between sectors. The parameter that captures time-varying inefficiency (η) is positive, meaning that inefficiency decreases over time. However, η is not statistically significant.

  24. 24.

    In general, the goodness of fit of the Models 1–3 does not improve by separating environmental investments (judged from the Akaike and Bayesian information criteria).

  25. 25.

    We have decomposed the model further with respect to lags and environmental domains, see Appendix Table 8, Models 4 and 5. The results do however not alter the conclusions made.

  26. 26.

    Table 7 in Appendix provides the results for a stochastic frontier model that does not specify heteroscedasticity in inefficiency. The results indicate technological progress as the trend parameter is positive and significant. The results also show significant differences between sub-sectors. The parameter that captures time-varying inefficiency (η) is negative and significant, meaning that inefficiency increases over time. Compared to the results for the pooled sample in Table 3, the trend parameter is larger suggesting that technological progress has been relatively high in the pulp and paper industry. This may explain the negative trend in efficiency.

  27. 27.

    We have decomposed the model further with respect to lags and also environmental domains, see Appendix Table 9, Models 4 and 5. The results do however not alter the conclusions made.


  1. Battese GE and TJ Coelli (1993) A stochastic frontier production incorporating a model for technical inefficiency effects, working papers in econometrics and applied statistics, no 69, Department of Econometrics, University of New England, Armidale, Australia

  2. Battese GE, Coelli TJ (1995) A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empir Econ 20(2):325–332

    Article  Google Scholar 

  3. Brännlund R (2008) Productivity and environmental regulations, Umeå Economic Studies, no. 728

  4. Brännlund R, Lundgren T (2009) Environmental policy without costs? A review of the Porter hypothesis. Int Rev Environ Resour Econ 3(2):75–117

    Article  Google Scholar 

  5. Brännlund R, Lundgren T (2010) Environmental policy and profitability: evidence from swedish industry. Environ Econ Policy Stud 12(1/2):59–78

    Article  Google Scholar 

  6. Byung MJ, Sickles RC (2004) The role of environmental factors in growth accounting. J Appl Econom 19(5):567–591

    Article  Google Scholar 

  7. Coelli TJ, Rao DSP, O’Donnell CJ, Battese GE (2005) An introduction to efficiency and productivity analysis. Springer Science + Business Media, LLC, New York

    Google Scholar 

  8. Eurostat (2001) Definitions and guidelines for measurement and reporting of company environmental protection expenditure, Eurostat Task Force, “Environmental Protection Expenditure—Industry Data Collection, Doc. ENV/01/3.6A”, European Commission

  9. Government Bill 2008/09:162, En sammanhållen klimat- och energipolitik (in Swedish)

  10. Grosskopf S (1993) Efficiency and productivity. In: Fried HO, Lovell CAK, Schmidt SS (eds) The measurement of productive efficiency—techniques and applications. Oxford University Press, New York

    Google Scholar 

  11. Kumbhakar SC, Lovell CAK (2000) Stochastic frontier analysis. Cambridge University Press, Cambridge

    Google Scholar 

  12. Lanoie P, Patry M, Lajeunesse R (2008) Environmental regulation and productivity: testing the Porter hypothesis. J Prod Anal 30(2):121–128

    Article  Google Scholar 

  13. Managi S, Opaluch JJ, Jin D, Grigalunas TA (2005) Environmental regulations and technological change in the offshore oil and gas industry. Land Econ 81(2):303–319

    Google Scholar 

  14. Olsson N, Eberhardson M (2003) “Environmental protection expenditure in Swedish industry—evaluation of new variables, data quality and estimation methods”, Statistics Sweden and Eurostat

  15. Porter ME (1991) America’s green strategy. Sci Am 264:168

    Article  Google Scholar 

  16. Porter ME, van der Linde C (1995) Toward a new conception of the environment-competitiveness relationship. J Econ Perspect 9(4):97–118

    Article  Google Scholar 

  17. Resvik B, Furbeck A (2005) Varför ska vi låta klimatmålet sänka vår industri i onödan? Dagens Ind 2005-12-20 (in Swedish)

  18. Schmidt P (2011) One-step and two-step estimations in sfa models. J Prod Anal. doi:10.1007/s11123-011-0228-0

    Google Scholar 

  19. Shadbegian RJ, Gray WB (2006) Assessing multi-dimensional performance: environmental and economic outcomes. J Prod Anal 26(3):213–234

    Article  Google Scholar 

  20. Statistics Sweden (2001) “Miljöskyddskostnader i industrin 1999 och 2000” (Environmental Protection Expenditure in Industry 1999 and 2000), MI 23 SM 0101, Statistics Sweden

  21. Statistics Sweden (2002) “Miljöskyddskostnader i industrin 2001” (Environmental Protection Expenditure in Industry 2001), MI 23 SM 0201, Statistics Sweden

  22. Statistics Sweden (2003) “Miljöskyddskostnader i industrin 2002” (Environmental Protection Expenditure in Industry 2002), MI 23 SM 0301, Statistics Sweden

  23. Statistics Sweden (2004) “Miljöskyddskostnader i industrin 2003” (Environmental Protection Expenditure in Industry 2003), MI 23 SM 0401, Statistics Sweden

  24. Statistics Sweden (2005) “Miljöskyddskostnader i industrin 2004” (Environmental Protection Expenditure in Industry 2004), MI 23 SM 0501, Statistics Sweden

  25. Statistics Sweden (2009) Personal communication Nov, 2009

  26. Swedish National Institute of Economic Research (2010) Personal communication with Roger Knudsen

Download references


The authors would like to thank Lennart Hjalmarsson, Thomas Forsfält, Maria Vredin Johansson, Göran Östblom, the editor, and two anonymous referees for valuable comments. Financial support from The Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS) is gratefully acknowledged.

Author information



Corresponding author

Correspondence to Eva Samakovlis.



See Tables 5, 6, 7, 8 and 9.

Table 5 Investments in pollution prevention and control for the category air
Table 6 Parameters of the estimated production function, and the mean efficiency scores; pooled data sample for the included industries (z-values in parentheses)
Table 7 The parameters of the estimated production functions, and the mean efficiency scores in the pulp and paper industry (z-values in parentheses)
Table 8 Test of the Porter hypothesis on the manufacturing industries; time-varying inefficiency model. Parameters multiplied by −1 presented
Table 9 Test of the Porter hypothesis on the pulp and paper industry; time-varying inefficiency model. Parameters multiplied by −1 presented.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Broberg, T., Marklund, PO., Samakovlis, E. et al. Testing the Porter hypothesis: the effects of environmental investments on efficiency in Swedish industry. J Prod Anal 40, 43–56 (2013). https://doi.org/10.1007/s11123-012-0335-6

Download citation


  • Environmental protection investments
  • Industry efficiency
  • Porter hypothesis
  • Stochastic production frontier

JEL Classification

  • C61
  • D21
  • D24
  • Q50
  • Q58