Offsetting Versus Mitigation Activities to Reduce \(\hbox {CO}_{2}\) Emissions: A Theoretical and Empirical Analysis for the U.S. and Germany


This paper studies the voluntary provision of public goods that is partially driven by a desire to offset for individual polluting activities. We first extend existing theory and show that offsets allow a reduction in effective environmental pollution levels while not necessarily extending the consumption of a polluting good. We further discuss the impact of an increased environmental preference on purchases of offsets and mitigation activities. Several theoretical results are then econometrically tested using a novel dataset on activities to reduce \(\hbox {CO}_{2}\) emissions for the case of vehicle purchases in the U.S. and Germany. We show that environmental preference triggers the stated use of \(\hbox {CO}_{2}\) offsetting and mitigation channels in both countries. However, we find strong country differences for the stated purchase of \(\hbox {CO}_{2}\) offsets. While such activities are mainly triggered by a high general awareness of the climate change problem in the U.S., the perception that road travel is responsible for \(\hbox {CO}_{2}\) emissions to a large extent is more important for driver’s license holders in Germany.

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


  1. 1.

    For example, the difference may stem from a limited general acceptance of offsets in society.

  2. 2.

    Corresponding notation will be used for the second derivatives (e.g. \(U_{xx} =\partial ^{2}U/\partial x^{2}\) and \(U_{xz} =\partial ^{2}U/\partial x\partial z)\).

  3. 3.

    Note that the normality assumption on \(U(z,x,-r)\) does not imply normality of \(V(z,x)=U(z,x,-x)\) in (zx). The individual desires to spend the additional income on both increased consumption of both x and \(-r\). As they are linked through \(x=r\), however, not both can increase such that the effect of additional income on both x and \(-r\) is ambiguous.

  4. 4.

    Different from Kotchen (2009), our model does not allow to easily define increased environmental concerns as a larger demand for \(-r\) as this demand may interact with the demand for the polluting. When no offsets are available, for example, an individual may consume less of the polluting good not for environmental concerns, but just because of a lesser preference for the private good.

  5. 5.

    While the interviews were conducted during the first phase of the global financial crisis, the survey time period is clearly before the peak crisis time, which began with the insolvency of Lehman Brothers in September 2008. Therefore, it is very unlikely that the responses were influenced by this specific time period. It should be mentioned that the survey time period is also clearly before the introduction of the U.S. Car Allowance Rebate System (“cash for clunkers”) in 2009, and a similar car-scrap bonus in in Germany in the same year. Both these programs affected the purchase of new vehicles and thus would possibly have influenced the responses to our questions.

  6. 6.

    While this complementary paper also analyzes the determinants of the purchase of \(\hbox {CO}_{2}\) offsets, it focuses on the effect of prior knowledge of this mechanism, but does not consider the relevance of the feeling of responsibility for \(\hbox {CO}_{2}\) emissions and is additionally not based on a theoretical analysis. In particular, the former paper does not examine mitigation activities when purchasing a new vehicle.

  7. 7.

    Response options in the underlying question were “very likely”, “somewhat likely”, “somewhat unlikely”, and “very unlikely”, which implies an ordinal structure of the variable. However, we do not focus on the corresponding ordinal dependent variable and thus ordinal probit models since we include this variable in a multivariate binary probit model as discussed below. Furthermore, Ziegler et al. (2012) show that the estimation results for the main explanatory variables are very similar by including ordinal or binary dependent variables. This finding is confirmed in this new analysis (the estimation results are not reported, but available upon request).

  8. 8.

    Response options were “not at all”, “purchase of a smaller vehicle”, “purchase of a vehicle with a smaller engine with less hp”, “purchase of a vehicle with different fuel or alternative drive systems”, “purchase of a vehicle with lower fuel consumption”, “purchase of a vehicle with low emissions”, “I am going to give up my vehicle entirely in future”, and “other”.

  9. 9.

    Examples for alternative fuels in vehicles are gas (e.g. natural gas, liquid petroleum gas), hydrogen, or biofuel, while examples for alternative driving systems or propulsion technologies are electric or hybrid. It can be argued that the purchase of vehicles with alternative drive systems as well as the purchase of less-emitting vehicles are not optimal indicators for the mitigation channel. Unfortunately, however, data about the intensity of using a vehicle are not available. Nevertheless, we think that our indicators are appropriate alternative indicators and especially assume that the rebound effect of these purchases on the use of the vehicles is less than 100 % so that finally they actually lead to a reduction of \(\hbox {CO}_{2}\) emissions.

  10. 10.

    In this respect, we used 50 random draws in the GHK simulator. Furthermore, we consider the robust estimations of the standard deviation of the parameter estimates (White 1982). The corresponding simulated maximum likelihood estimations (in the same way as all further estimations and also the descriptive statistics as discussed above) were conducted with STATA.

  11. 11.

    Response options were “the topic is extremely important to me”, “the topic is important to me”, “the topic is only of little importance to me”, and “the topic is of no importance to me”.

  12. 12.

    Response options were “thoroughly convinced”, “largely convinced”, “rather unconvinced”, and “not convinced at all”.

  13. 13.

    Response options were “a very large proportion”, “a large proportion”, “a small proportion”, and “no role”.

  14. 14.

    Unfortunately, data on individual income were not available.

  15. 15.

    The estimated correlation coefficients in the error terms of the underlying latent variables are not reported for brevity, but available upon request. They reveal significantly positive correlations between all three dependent variables in the U.S., which underpins the importance of applying multivariate instead of univariate probit models. In Germany, only the correlation coefficient between the two mitigation variables is significantly different from zero. In contrast to the result in the U.S., this correlation coefficient is negative.

  16. 16.

    This result can be shown when “only purchase of less-emitting vehicle” instead of “no purchase of \(\hbox {CO}_{2}\) offsets and no purchase of less-emitting vehicle” or “only purchase of vehicle with alternative drive systems” instead of “no purchase of \(\hbox {CO}_{2}\) offsets and no purchase of vehicle with alternative drive systems” are used as base categories in the multinomial logit models. The corresponding estimation results are not reported for brevity, but available upon request.

  17. 17.

    The corresponding estimation results are not reported for brevity, but available upon request.

  18. 18.

    Therefore, they are not reported for brevity, but available upon request.

  19. 19.

    These estimation results are not reported due to brevity, but available upon request.


  1. Andreoni J (1989) Giving with impure altruism: applications to charity and Ricardian equivalence. J Polit Econ 97:1447–1458

    Article  Google Scholar 

  2. Andreoni J (1990) Impure altruism and donations to public goods: a theory of warm-glow giving. Econ J 100:464–477

    Article  Google Scholar 

  3. Andreoni J, Levinson A (2001) The simple analytics of the environmental Kuznets curve. J. Public Econ. 80:269–286

    Article  Google Scholar 

  4. Akter S, Brouwer R, Brander L, Van Beukering P (2009) Respondent uncertainty in a contingent market for carbon offsets. Ecol Econ 68:1858–1863

    Article  Google Scholar 

  5. Blasch J, Farsi M (2014) Context effects and heterogeneity in voluntary carbon offsetting–a choice experiment in Switzerland. J Environ Econ Policy 3(1):1–24

    Article  Google Scholar 

  6. Börsch-Supan A, Hajivassiliou VA (1993) Smooth unbiased multivariate probability simulators for maximum likelihood estimation of limited dependent variable models. J. Econ. 58:347–368

    Article  Google Scholar 

  7. Brouwer R, Brander L, Van Beukering P (2008) “A convenient truth”: air travel passengers’ willingness to pay to offset their \({\rm CO}_{2}\) emissions. Clim Change 90:299–313

    Article  Google Scholar 

  8. Business for Social Responsibility (2007) Getting carbon offsets right: a business brief on engaging offset providers. San Francisco

  9. Clean Air - Cool Planet (2006) A consumer’s guide to retail carbon offset providers. Portland

  10. Clean Air - Cool Planet and Forum for the Future (2008) Getting to zero: defining corporate carbon neutrality

  11. Conte MN, Kotchen MJ (2010) Explaining the price of voluntary carbon offsets. Clim Change Econ 1(2):93–111

    Article  Google Scholar 

  12. Geweke J, Keane M, Runkle D (1994) Alternative computational approaches to inference in the multinomial model. Rev Econ Stat 76:609–632

    Article  Google Scholar 

  13. Greene WH (2008) Econometric analysis. Prentice Hall, New Jersey

    Google Scholar 

  14. Gossling S, Haglund L, Kallgren H, Revahl M, Hultman J (2009) Swedish air travellers and voluntary carbon offsets: towards the cocreation of environmental value? Curr Issues Tour 12(1):1–19

    Article  Google Scholar 

  15. Hajivassiliou VA (1994) A simulation estimation analysis of the external debt crisis of developing countries. J Appl Econ 9:109–131

    Article  Google Scholar 

  16. Keane M (1994) A computationally practical simulation estimator for panel data. Econometrica 62:95–116

    Article  Google Scholar 

  17. Kotchen MJ (2009) Voluntary provision of public goods for bads: a theory of environmental offsets. Econ. J. 119:883–899

    Article  Google Scholar 

  18. Lu J-L, Shon ZY (2012) Exploring airline passengers’ willingness to pay for carbon offsets. Transp Res Part D Transp Environ 17(2):124–128

    Article  Google Scholar 

  19. MacKerron GJ, Egerton C, Gaskell C, Parpia A, Mourato S (2009) Willingness to pay for carbon offset certification and co-benefits among (high-)flying young adults in the UK. Energy Policy 37:1372–1381

    Article  Google Scholar 

  20. Mair J (2011) Exploring air travellers’ voluntary carbon-offsetting behaviour. J Sustain Tour 19(2):215–230

    Article  Google Scholar 

  21. White H (1982) Maximum likelihood estimation of misspecified models. Econometrica 50:1–25

    Article  Google Scholar 

  22. WWF Deutschland (2008) \({\rm CO}_{2}\) -Kompensation: Augenwischerei oder wirksamer Klimaschutz? Leitfaden des WWF für die wirksame Anwendung eines sinnvollen Instruments. Positionspapier, Frankfurt am Main

  23. Ziegler A, Schwarzkopf J, Hoffmann VH (2012) Stated versus revealed knowledge: determinants of offsetting \({\rm CO}_{2}\) emissions from fuel consumption in vehicle use. Energy Policy 40:422–431

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Andreas Lange.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lange, A., Ziegler, A. Offsetting Versus Mitigation Activities to Reduce \(\hbox {CO}_{2}\) Emissions: A Theoretical and Empirical Analysis for the U.S. and Germany. Environ Resource Econ 66, 113–133 (2017).

Download citation


  • Public good
  • Voluntary provision
  • Climate change
  • \(\hbox {CO}_{2}\) offsetting
  • Vehicle purchase
  • Discrete choice models

JEL Classification

  • H41
  • Q54