Certification and socially responsible production


In many markets, consumers are unable to distinguish between goods that are produced in a socially responsible way and goods that are not. In such situations, socially responsible production is not a profit-maximizing strategy, even if the premium that consumers are willing to pay exceeds the costs. Only firms that are genuinely motivated by social responsibility would then produce in this way, and there would be too little socially responsible production. Improved opportunities for voluntary certification could potentially reduce this problem by allowing firms to signal their type. We examine how the possibility of certification affects the share of socially responsible production. Our main result is that increased certification may reduce the share of socially responsible production by reducing prices in the market for uncertified products and thus crowd out socially responsible producers who do not certify. This provide a mechanism through which certification might have adverse effects on socially responsible production, even when the certification process is perfect and when there is perfect competition among the producers.

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

Fig. 1
Fig. 2


  1. 1.

    The assumption of rational expectations simplifies the analysis, but the main results of this paper hold as long as the consumers’ expectations about the fraction of socially responsible production is an increasing function of the actual fraction.

  2. 2.

    For \(c \ge \bar{z}\), \(D(e,\bar{z})=0\) for all \(e \ge 0\) , i.e. there is no demand for certified products. Equation (15) is then identical to (4) from Sect. 2.3; this means \(e(p;c) \equiv e(p)\).

  3. 3.

    Differentiating H(epc) twice with respect to e shows that (i) holds if \(D_{ee}\) is not too negative.

  4. 4.

    If c is sufficiently low, the e(pc)-curve starts above v, which means it may not cross the p(e)-curve and \(e=0\) may be the only solution.


  1. Amacher GS, Koskela E, Ollikainen M (2004) Environmental quality competition and eco-labeling. J Environ Econ Manag 47:284–306

    Article  Google Scholar 

  2. Baksi S, Bose P (2007) Credence goods, efficient labelling policies, and regulatory enforcement. Environ Resour Econ 37(2):411–430

    Article  Google Scholar 

  3. Baland J-M, Duprez C (2009) Are labels effective against child labor? J Public Econ 93:1125–1130

    Article  Google Scholar 

  4. Basu AK, Chau NH, Grote U (2006) Guaranteed manufactured without child labor: the economics of consumer boycotts, social labeling and trade sanctions. Rev Dev Econ 10(3):466–491

    Article  Google Scholar 

  5. Bjørner TB, Hansen LG, Russel CS (2004) Environmental labeling and consumers’ choice—an empirical analysis of the Nordic swan. J Environ Econ Manag 47:411–434

    Article  Google Scholar 

  6. Bonroy O, Constantatos C (2008) On the use of labels in credence goods markets. J Regul Econ 33:237–252

    Article  Google Scholar 

  7. Cason TN, Gangadharan L (2002) Environmental labeling and incomplete consumer information in laboratory markets. J Environ Econ Manag 43:113–134

    Article  Google Scholar 

  8. Costa S, Ibanez L, Loureiro ML, Marette S (2009) Quality promotion through eco-labeling: introduction to the special issue. J Agric Food Ind Organ 7(2):1–8

    Google Scholar 

  9. Darby MR, Karni E (1973) Free competition and the optimal amount of fraud. J Law Econ 16(1):67–88

    Article  Google Scholar 

  10. Dranove D, Jin GZ (2010) Quality disclosure and certification: theory and practice. J Econ Lit 48(4):935–963

    Article  Google Scholar 

  11. Dulleck U, Kerschbamer R (2006) On doctors, mechanics, and computer specialists: the economics of credence goods. J Econ Lit 44(1):5–42

    Article  Google Scholar 

  12. Dulleck U, Kerschbamer R, Sutter M (2011) The economics of credence goods: an experiment on the role of liability, verifiability, reputation, and competition. Am Econ Rev 101(2):525–555

    Article  Google Scholar 

  13. Feddersen TJ, Gillian TW (2001) Saints and markets: activists and the supply of credence goods. J Econ Manag Strategy 10(1):149–171

    Article  Google Scholar 

  14. Hamilton SF, Zilberman D (2006) Green markets, eco-certification, and equilibrium fraud. J Environ Econ Manag 52:627–644

    Article  Google Scholar 

  15. Mason CF (2009) Certification of socially responsible behavior: eco-labels and fair-trade coffee. J Agric Food Ind Organ 7(2):1–21

    Google Scholar 

  16. Rode J, Hogarth RM, Le Menestrel M (2008) Ethical differentiation and market behavior: an experimental approach. J Econ Behav Organ 66:265–280

    Article  Google Scholar 

  17. Teisl ML, Roe B, Hicks RL (2002) Can eco-labels tune a market? evidence from dolphin-safe labeling. J Environ Econo Manag 43:339–359

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Alexander W. Cappelen.


Appendix 1: Derivation of e(pc)

Define the right-hand side of (8) as H(epc), i.e.

$$\begin{aligned}{}[\epsilon G(p-m)-D(e;c) ]/[\epsilon G(p-m)-D(e,c)+(1-\epsilon ) G(p)] \equiv H(e;p,c). \end{aligned}$$

Equation (8) can then be written as

$$\begin{aligned} e=H(e;p,c). \end{aligned}$$

We focus on the case where there is a demand for certified products, i.e. where \(c<\bar{z}\) such that \(D(\cdot )>0\) for some \(e>0\).Footnote 2 As D(ec) is decreasing in e and c, it is straightforward to show that \(H(\cdot )\) is increasing in e, c, and p. Assume that (i) \(H_{ee}<0\) and (ii) \(H_e(0;p;c)=- D_e(0,c)<1\) for the range of c and p we study.This ensures a unique solution to (16).Footnote 3

Consider c-values such that \(H(0;p,c)\ge 0\) for some p, and let p be sufficiently high to make \(H(0;p,c)> 0\). When (i) holds, (ii) implies that \(H_e(e;p,c)<1\) for all \(e>0\). The \(H(\cdot )\)-curve then crosses the 45-degree line only once and from above. This implies that (15) gives a unique and stable solution with \(e>0\). Equation (15) determines e as an increasing function of p for a given c, written as e(pc). For each c, the e(pc)-curve starts at the lowest p-value that yields \(H(0;p,c)=0\). As \(D(\cdot )>0\), it follows from (15) that \(H(0;p,c)=0\) requires \(G(p-m)>0\), which in turn requires \(p>m\). Thus, all e(pc)-curves start above m on the p-axis.

Differentiating (16) with respect to p yields

$$\begin{aligned} de/dp=\frac{1-H_e(e;p,c)}{ H_p(e;p,c)}. \end{aligned}$$

As \(H_p(e;p,c)>0\) and \(H_e(e;p,c)<1\), it follows that \(de/dp>0\).

Differentiating (16) with respect to c yields

$$\begin{aligned} de/dc=\frac{1-H_e(e;p,c)}{ H_c(e;p,c)}. \end{aligned}$$

As \(H_e(e;p,c)<1\) and \(H_e(e;p,c)<1\), it follows that \(de/dc>0\).

Appendix 2: Proof of Proposition 2

As shown in Sect. 2.3, Eq. (7) determines the price p as an increasing function of e, starting in v on the p-axis. For \(c<\bar{z}\), the curve e(pc) starts above m on the p-axis. As p goes to infinity, e(pc) goes towards \(\epsilon \), while \(p(\epsilon )\) is a finite value. For c not too low, such that the e(pc)-curve starts below v, it crosses the curve p(e) at least once from below. At the crossing, the e-curve is steeper than the p-curve, i.e. \(1/e_p(p,c)>p(e)\), which means the solution is stable. Thus, there is at least one stable solution to (7) and (8) with \(e>0\).Footnote 4

Consider a stable solution with \(e>0\). Differentiating (7) and (8) with respect to c yields

$$\begin{aligned} de/dc=\frac{e_c(p,c)}{1-e_p(p,c)p'(e)}. \end{aligned}$$

As \(e_p(p,c)p'(e)<1\) in a stable solution, it follows that \(de/dc>0\) , i.e. e goes down when c goes down. As \(p'(e)>0\), p goes down when e goes down. With a lower c, the e(pc)-curve shifts to the left and crosses the p(e)-curve at a lower value of e. Thus, a lower c gives rise to a lower fraction of ethical producers in the uncertified market.

Appendix 3: Proof of Proposition 3

Without access to certification, total welfare is given by (9). By using \(e_0mx_0=\epsilon \int \limits _{0 }^{p_0-m } mdG(q)\), we can rewrite the welfare function as

$$\begin{aligned} W=v x+e_0\int \limits _{0 }^{x} [z(x)-m]dx-\left[ \epsilon \int \limits _{0 }^{p_0-m } qdG(q)+(1-\epsilon )\int \limits _{0 }^{p_0 } qdG(q) \right] . \end{aligned}$$

to show that \(x_0\) is lower than optimal, we evaluate \(W'(x)\) at \(x_0\). Differentiating W with respect to x, taking into account that p and e are functions of x, and evaluating at \(x_0\), yields

$$\begin{aligned} dW/dx_0=v+e_0z(x_0)-p_0 +(\lambda -e_0) \frac{1}{x_0} \int \limits _{0 }^{x_0 } [z(x)-m]dx, \end{aligned}$$

where \(\lambda = \epsilon g(p_0-m)/[\epsilon g(p_0-m)+(1-\epsilon )g(p_0)\). As e is increasing in p, \(\lambda - e_0>0\). From (5), \( v+e_0z(x_0)=p_0\) in the no-certification equilibrium. Inserting this into (21) gives \(dW/dx_0>0\), which means output is too low in the absence of certification.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Cappelen, A.W., Ognedal, T. Certification and socially responsible production. Econ Gov 18, 71–84 (2017). https://doi.org/10.1007/s10101-016-0184-2

Download citation


  • Social responsibility
  • Certification
  • Labeling
  • Ethical production

JEL Classification

  • Q5
  • M14
  • D82