Skip to main content
Log in

Efficient nonanthropocentric nature protection

  • Original Paper
  • Published:
Social Choice and Welfare Aims and scope Submit manuscript

Abstract

This paper analyzes nature protection by a social planner under different ‘utilitarian’ social welfare functions. For that purpose we construct an integrated model of the economy and the ecosystem with explicit consideration of nonhuman species and with competition between human and nonhuman species for land and prey biomass. We characterize and compare the efficient allocations when social welfare is anthropocentric (only consumers have positive welfare weights), when social welfare is biocentric (only nonhuman species have positive welfare weights) and when social welfare is nonanthropocentric (all species have positive welfare weights). Not surprisingly, biocentric social welfare calls for suspending all economic activities. It is more important, however, that both anthropocentrism and nonanthropocentrism make the case for nature protection through different channels, though. Our analysis suggests that one may dispense with the concept of nonanthropocentric social welfare provided that in the anthropocentric framework the consumers' intrinsic valuation of nature is properly accounted for.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Notes

  1. This line of reasoning already has significant impacts on the real world. In a recent amendment to the German constitution the intrinsic value of nature has been recognized. Almost 90% of the German population believe that the plants' and animals' right to live ought to be respected.

  2. If letters denote functions, subscripts attached to them indicate derivatives.

  3. 0 N is the N-dimensional vector consisting of N zeros.

  4. The conceptual origin of Eq. 2 is Hannon (1976) whose approach has been further elaborated by Tschirhart (2000). Our approach differs from theirs mainly in that these authors focused on the generation of net energy rather than net offspring.

  5. One might think about different specifications of the function V i in Eq. 3 depending on (future) insights of welfare biology. Note, however, that the main contribution of the present paper is conceptual rather than an inquiry into which are the most appropriate variables determining the nonhuman organisms' welfare.

  6. Ng (1995) defines sentients as beings that are capable of subjectively perceiving or feeling by means of the senses. Affective sentients are happy if feelings are positive and unhappy if feelings are negative. According to Ng (1995) positive feelings are those that feel good in themselves, including sensuous pleasures and spiritual delights, and negative feelings are those that feel bad in themselves, including sensuous pain and mental sufferings. Non-affective sentients perceive the world without feeling happy or unhappy.

  7. \(U_{{n_{i} }} > 0\) for all i is a stylized assumption, of course, since negative partial derivatives can be considered plausible for viruses and bacteria, e.g.

  8. Note that in contrast to Eqs. 5 and 6, in Eq. 7 the strict equality sign is indispensable; see Eichner and Pethig (2003).

  9. They also suffer from losing own biomass to their predators, but this is an intra-system interdependence.

  10. The issue of species extinction (n i =0) is in the realm of our analytical framework but beyond the scope of the present paper.

  11. Note that the notion of anthropocentric efficiency is equivalent to the conventional concept of Pareto efficiency.

  12. These conditions are analogous to the efficiency conditions governing the efficient production in disaggregated models of the economy.

  13. The specification Eq. (24a) is reminiscent of the concept of strong sustainability where substitutability between natural and man-made capital is denied.

  14. We don't aim at defending either specification Eq. (24), (24a), or (25), (25a) regarding its empirical relevance. The main reason for introducing the quite restrictive functions Equations (24), (24a), (25) and (25a) is clearly analytical tractability, and the main reason for employing alternative specifications of human welfare and species net offspring generation is to examine the robustness of our results.

  15. To explore the impact of changing welfare weights it suffices to change the parameter a while keeping a h constant. This allows us to further simplify the notation by setting a h ≡ 1.

  16. Similarly, Norton (1991) recommends the indicator ‘percentage of ground area with pervious surfaces’ as representing values of advocates of green space, of biological diversity, of water quality, and of anti-sprawl. One might consider the habitat variable \(\overline{r} - r_{y} \) from our model as a proxi for such an indicator.

  17. Total differentiating Eq. (62b) one gets (β W n rY r +zY z )dn+Wdβ+Y r ndrY z ndz. In Eq. (63) we have taken into account that (61a), (61b) and B r dr+B z dz=0 imply Y r ndrY z ndz=0.

References

  • Callicott JB (1989) In defence of the land ethic. State University of New York, Albany NY

    Google Scholar 

  • Dailey GC (1997) Introduction: what are ecosystem services? In: Dailey GC (ed) Nature's Services: Societal Dependence on Natural Ecosystems. Island, Washington, DC

    Google Scholar 

  • Eichner T, Pethig R (2003) A microfoundation of predator–prey dynamics. Discussion Paper 110-03, University of Siegen

  • Hannon B (1976) Marginal product pricing in the ecosystem. J Theor Biol 56: 253–267

    PubMed  CAS  Google Scholar 

  • Johannson-Stenman O (1998) The importance of ethics in environmental economics with a focus on existence values. Environ Resour Econ 11: 429–442

    Article  Google Scholar 

  • Ng Y-K (1995) Towards welfare biology: evolutionary economics of animal consciousness and suffering. Biol Philos 10: 255–285

    Article  Google Scholar 

  • Ng Y-K (1999) Utility, informed preference, or happiness: following Harsanyi's argument to its logical conclusion. Soc Choice Welf 16: 197–216

    Article  MATH  Google Scholar 

  • Ng Y-K (2003) From preferences to happiness: towards a more complete welfare economics. Soc Choice Welf 20: 307–350

    Article  MATH  Google Scholar 

  • Norton B (1991) Towards unity among environmentalists. Oxford University Press, New York

    Google Scholar 

  • Norton B, Minteer BA (2002), From environmental ethics to environmental public philosophy: ethicists and economists, 1973—future. In: Tietenberg T, Folmer H (eds) The international yearbook of environmental and resource economics 2002/2003. Edward Elgar, Cheltenham UK Northampton MA USA, pp 373–407

    Google Scholar 

  • Regan T (1983) The case for animal rights. University of Berkeley, Berkeley, CA

    Google Scholar 

  • Rolston III H (1988) Environmental ethics. Temple University Press, Philadelphia

    Google Scholar 

  • Routley R (1973) Is there a need for a new, an environmental ethic? Proceedings, 15th World Congress of Philosophy 1: 205–210

  • Spash O (2000) Multiple value expression in contingent valuation: economics and ethics. Environ Sci Technol 34: 1433–1438

    Article  CAS  Google Scholar 

  • Tschirhart J (2000) General equilibrium of an ecosystem. J Theor Biol 203:13–32

    Article  PubMed  CAS  Google Scholar 

Download references

Acknowledgements

Helpful comments from an anonymous editor of the journal Social Choice and Welfare are gratefully acknowledged. Remaining errors are the authors' sole responsibility

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Thomas Eichner.

Appendices

Appendix A

Proof of Proposition 1

The associated first-order conditions to the Lagrangean (11) are

$${\user1{\mathcal{L}}}_{y} = \lambda _{y} - \lambda _{Y} \leqslant 0,\quad y{\user1{\mathcal{L}}}_{y} = 0,$$
(50a)
$${\user1{\mathcal{L}}}_{{y_{h} }} = \delta _{h} a_{h} n_{h} U_{y} - \lambda _{y} n_{h} \leqslant 0,\quad y_{h} {\user1{\mathcal{L}}}_{{y_{h} }} = 0,$$
(50b)
$${\user1{\mathcal{L}}}_{{r_{y} }} = - \delta _{h} a_{h} n_{h} U_{r} + \lambda _{Y} Y_{r} - \lambda _{r} \leqslant 0,\quad r_{y} {\user1{\mathcal{L}}}_{{r_{y} }} = 0,$$
(50c)
$${\user1{\mathcal{L}}}_{{z_{{yj}} }} = \lambda _{Y} Y_{{z_{{yj}} }} - \lambda _{{z_{j} }} \leqslant 0,\quad z_{{yj}} {\user1{\mathcal{L}}}_{{z_{{yj}} }} = 0,$$
(50d)
$${\user1{\mathcal{L}}}_{{r_{i} }} = n_{i} \lambda _{{b_{i} }} B^{i}_{r} - \lambda _{r} n_{i} \leqslant 0,\quad r_{i} {\user1{\mathcal{L}}}_{{r_{i} }} = 0,$$
(50e)
$${\user1{\mathcal{L}}}_{{z_{{ij}} }} = n_{i} \lambda _{{b_{i} }} B^{i}_{{z_{{ij}} }} - n_{i} \lambda _{{z_{j} }} \leqslant 0,\quad z_{{ij}} {\user1{\mathcal{L}}}_{{z_{{ij}} }} = 0,$$
(50f)
$${\user1{\mathcal{L}}}_{{b_{i} }} = \delta _{s} a_{i} n_{i} V^{i}_{b} + \mu _{{n_{i} }} n_{i} - n_{i} \lambda _{{b_{i} }} = 0,$$
(50g)
$$\begin{array}{*{20}c} {{\mathop \mu \limits^. }_{{n_{i} }} = \rho \mu _{{n_{i} }} - {\user1{\mathcal{L}}}_{{n_{i} }} = \rho \mu _{{n_{i} }} - \mu _{{n_{i} }} b_{i} + \lambda _{r} r_{i} + {\sum\limits_j {\lambda _{{z_{j} }} z_{{ij}} - \delta _{h} a_{h} n_{h} U_{{n_{i} }} } }} \\ { - \delta _{s} a_{i} {\left( {V^{i} + n_{i} V^{i}_{n} } \right)},} \\ \end{array} $$
(50h)
$$ \begin{array}{*{20}l} {{{\user1{\mathcal{L}}}_{{\lambda _{{b_{i} }} }} \geqslant 0,\lambda _{{b_{i} }} {\user1{\mathcal{L}}}_{{\lambda _{{b_{i} }} }} = 0,} \hfill} & {{{\user1{\mathcal{L}}}_{{\lambda _{y} }} \geqslant 0,\lambda _{y} {\user1{\mathcal{L}}}_{{\lambda _{y} }} = 0,} \hfill} & {{{\user1{\mathcal{L}}}_{{\lambda _{Y} }} \geqslant 0,\lambda _{Y} {\user1{\mathcal{L}}}_{{\lambda _{Y} }} = 0,} \hfill} \\ {{{\user1{\mathcal{L}}}_{{\lambda _{r} }} \geqslant 0,\lambda _{r} {\user1{\mathcal{L}}}_{{\lambda _{r} }} = 0,} \hfill} & {{{\user1{\mathcal{L}}}_{{\lambda _{{zj}} }} \geqslant 0,\lambda _{{z_{j} }} {\user1{\mathcal{L}}}_{{\lambda _{{zj}} }} = 0.} \hfill} & {{} \hfill} \\ \end{array} $$
(50i)
  1. (1)

    Set δ h =0 and δ s =1 in Eqs. 50–50i and consider a maximum of 10 in which r i >0 for all i, z ij >0 for all i, j, ij, and (hence) z jj <0 for all i. In view of Eqs. 50e and 50f it is straightforward that either (a) λ r >0 and therefore \(\lambda _{{b_{i} }} > 0\) and \(\lambda _{{z_{j} }} > 0\) or (b) λ r =0 and therefore \(\lambda _{{b_{i} }} = \lambda _{{z_{j} }} = 0\). Consider first the case (b). Raise \(\overline{r} \;\,{\text{by}}\;\,{\text{d}}\,\overline{r} > 0\) and set \(n_{i} {\text{d}}r_{i} = {\text{d}}\overline{r} \) and \({\text{d}}b_{i} = B^{i}_{r} {\text{d}}r_{i} > 0\). The impact of db i >0 on social welfare is given by Eq. 50g where \(a_{i} n_{i} V^{i}_{b} > 0\) is the direct impact on instantaneous welfare and where \(\mu _{{n_{i} }}\) captures the shadow price of the population of species i or, more technically, the present value of the marginal future welfare induced by \({\text{d}}{\mathop n\limits^. }_{i} = n_{i} {\text{d}}b_{i} > 0\). Since the welfare index 9 is strictly increasing in n i (for all i) \(\mu _{{n_{i} }} > 0\) follows cogently. Yet \(\lambda _{{b_{i} }} = 0\) and Eq. 50g imply \(\mu _{{n_{i} }} = - a_{i} V^{i}_{b} < 0\) which is incompatible with a maximum of 10. This leads us to conclude that case (a) applies with \(\mu _{{n_{i} }} = \lambda _{{b_{i} }} - a_{i} V^{i}_{b} > 0\).

    Suppose now case (a) holds but y>0, contrary to the assertion in Proposition 1i. Then λ y =λ Y ≥0 follows via Eq. 50a. If λ y >0 then y h =0 from Eq. 50b and therefore \({\user1{\mathcal{L}}}_{{\lambda _{y} }} > 0\). However, \(\lambda _{y} {\user1{\mathcal{L}}}_{{\lambda _{y} }} = 0\) implies λ y =0. This contradiction establishes λ y =λ Y =0. Moreover, Eq. 50c yields λ Y Y r λ r =−λ r <0 and therefore r y =0 via \(r_{y} {\user1{\mathcal{L}}}_{{r_{y} }} = 0\). Since Y(0, z y ) = 0 by assumption, y=0 follows from r y =0. This contradiction proves that y=0 when 10 is maximized. y=0 and Eq. 6 yield y h =0 for all h. Finally, r y =0 and z yj =0 for all j is the unique efficient way to produce y=0.

    In the remaining proof of (1) we establish Eqs. 1215. Equation 12 follows from Eq. 50f. More precisely

    $$\frac{{n_{i} \lambda _{{b_{i} }} B^{i}_{{z_{{ij}} }} }} {{n_{i} \lambda _{{b_{i} }} B^{i}_{{z_{{ik}} }} }} = \frac{{\lambda _{{z_{j} }} }} {{\lambda _{{z_{k} }} }} = \frac{{n_{m} \lambda _{{b_{m} }} B^{m}_{{z_{{mj}} }} }} {{n_{m} \lambda _{{b_{m} }} B^{m}_{{z_{{mk}} }} }}\quad {\text{for}}\,i,j,k,m = 1, \ldots ,N.$$
    (51)

    Equation 13 is derived from Eqs. 50e and 50f:

    $$ \frac{{n_{i} \lambda _{{b_{i} }} B^{i}_{{z_{{ij}} }} }} {{n_{i} \lambda _{{b_{i} }} B^{i}_{{z_{{ik}} }} }} = \frac{{\lambda _{r} }} {{\lambda _{{z_{k} }} }} = \frac{{n_{m} \lambda _{{b_{m} }} B^{m}_{r} }} {{n_{m} \lambda _{{b_{m} }} B^{m}_{{z_{{mk}} }} }}\quad for\;i,j,j,m = 1, \ldots ,N. $$
    (52)

    Equation 14 follows from Eqs. 50e and 50g, and finally, using Eqs. 50c, 50e and 50g in Eq. 50h we obtain Eq. 15.

  2. (2)

    Set δ s =0 and δ h =1 in Eqs. 50a50i. Now we restrict our attention to an interior solution such that all inequalities Eqs. 50a50i hold as equalities. Equation 16 follows from Eqs. 50d and 50f. To establish Eq. 17 we combine Eqs. 50c50f. To obtain Eq. 18 we combine Eqs. 50e and 50g, and using Eqs. 50c, 50e and 50g in Eq. 50h we get Eq. 19.

  3. (3)

    Here we set δ h =δ s =1 and make use of the same manipulations as in the derivation of Eqs. 1619.

Appendix B: Comparative dynamics in the one-species model

Derivation of Eqs. 36a36c: Total differentiation of Eqs. 35a35c yields:

$$J \cdot {\left( {\begin{array}{*{20}l} {{{\text{d}}r} \hfill} \\ {{{\text{d}}z} \hfill} \\ {{{\text{d}}n} \hfill} \\ \end{array} } \right)} = {\left( {\begin{array}{*{20}l} {0 \hfill} \\ {{ - {\left( {{\text{d}}\beta W_{n} + {\text{d}}a{\left( {V + nV_{n} + \rho V_{b} } \right)}} \right)}B_{r} } \hfill} \\ {{ - {\left( {{\text{d}}\beta W_{n} + {\text{d}}a{\left( {V + nV_{n} + \rho V_{b} } \right)}} \right)}B_{z} } \hfill} \\ \end{array} } \right)}$$
(53)

where

$$ J: = {\left( {\begin{array}{*{20}c} {{B_{r} }} & {{B_{z} }} & {0} \\ {{nY_{{rr}} + \lambda B_{{rr}} + \phi _{1} Br}} & {{ - nY_{{rz}} + \lambda B_{{rz}} + \phi _{2} B_{r} }} & {{ - zY_{{rz}} + rY_{{rr}} + \phi _{3} B_{r} }} \\ {{ - nY_{{zr}} + \lambda B_{{zr}} + \phi _{1} B_{z} }} & {{nY_{{zz}} + \lambda B_{{zz}} + \phi _{2} B_{z} }} & {{ - zY_{{rz}} + zY_{{zz}} + \phi _{3} B_{z} }} \\ \end{array} } \right)} $$

and

$$ \begin{array}{*{20}c} {\lambda = {\left( {\beta W_{n} + a{\left( {V + nV_{n} + \rho V_{b} } \right)} - rY_{r} + zY_{z} } \right)}/\rho ,} \\ {\phi _{1} : = {\left( { - Y_{r} + nrY_{{rr}} - nzY_{{zr}} } \right)}/\rho _{1} ,} \\ {\phi _{2} : = {\left( {Y_{z} - nrY_{{zr}} + nzY_{{zz}} } \right)}/\rho _{1} ,} \\ {\phi _{3} = {\left( {\beta W_{{nn}} + a{\left( {2V_{n} + nV_{{nn}} } \right)} + r^{2} Y_{{rr}} + z^{2} Y_{{zz}} - 2rzY_{{rz}} } \right)}/\rho .} \\ \end{array} $$

Solving the equation system 53 by using Cramer's rule we obtain

$$ {\text{d}}r = \frac{{{\text{d}}\beta W_{n} + da{\left( {V + nV_{n} + \rho V_{b} } \right)}}} {{{\left| J \right|}}} \cdot B_{z} \cdot {\left[ {{\left( {zY_{{zz}} - zY_{{rz}} } \right)}B_{r} - {\left( {rY_{{rr}} - zY_{{zr}} } \right)}B_{z} } \right]}, $$
(54a)
$${\text{d}}z = \frac{{{\text{d}}\beta W_{n} + da{\left( {V + nV_{n} + \rho V_{b} } \right)}}}{{{\left| J \right|}}} \cdot B_{r} \cdot {\left[ {{\left( {zY_{{rz}} - zY_{{zz}} } \right)}B_{r} - {\left( {zY_{{zr}} - rY_{{rr}} } \right)}B_{z} } \right]},$$
(54b)
$$dn = \frac{{{\text{d}}\beta W_{n} + {\text{d}}a{\left( {V + nV_{n} + \rho V_{b} } \right)}}}{{{\left| J \right|}}} \cdot {\left[ {B^{2}_{r} {\left( {nY_{{zz}} + \lambda B_{{zz}} } \right)} + B^{2}_{z} {\left( {nY_{{rr}} + \lambda B_{{zr}} } \right)} - 2B_{r} B_{z} {\left( { - nY_{{rz}} + \lambda B_{{rz}} } \right)}} \right]}.$$
(54c)

Derivation of Eqs. 37a37d: While Eqs. 37a, 37c and 37d follow directly from total differentiation of Eq. 24, nr and nz, respectively, totally differentiating \(Y{\left( {\overline{r} - nr,nz} \right)}\) yields

$$ {\text{d}}y = nY_{z} {\text{d}}z - nY_{r} {\text{d}}r + {\left( {zY_{z} - Y_{r} } \right)}{\text{d}}n. $$
(55)

B r =Y r /λ and B z =−Y z /λ from Eqs. 34a and 34b, respectively, inserted into B r dr+B z d z =0 yields Y r drY z d z =0 such that Eq. 55 simplifies to Eq. 37b.

Proof of Proposition 2

Model A: The results in lines 1 and 5 of Table 1 follow directly from Eqs. 36a36c and 37a37d by taking into account that ∣J∣<0, B r >0, B z <0, Y rr <0, Y zz <0, B rz ≤0, Y rz ≥0 and λ>0.

Model B: In case of the Leontief net offspring function \(b = \min {\left[ {r,c{\left( {\overline{z} - z} \right)}} \right]} - \varepsilon \) efficiency requires \(r = c{\left( {\overline{z} - z} \right)}\) such that Eq. 33 turns into

$${\user1{\mathcal{H}}} = \beta W{\left( n \right)} + Y{\left[ {\overline{r} - nr,n{\left( {\overline{z} - \frac{r} {c}} \right)}} \right]} + anV{\left( {n,r - \varepsilon } \right)} + \mu n{\left( {r - \varepsilon } \right)}$$
(56)

and we obtain the first-order conditions

$$ {\user1{\mathcal{H}}}_{r} = - nY_{r} - n\frac{{Y_{z} }} {c} + anV_{b} + n\mu = 0, $$
(57a)
$${\mathop \mu \limits^. } = \rho \mu - \beta W_{n} - a{\left( {V + nV_{n} } \right)} + rY_{r} - {\left( {\overline{z} - \frac{r}{c}} \right)}Y_{z} - \mu {\left( {r - \varepsilon } \right)}.$$
(57b)

Then the long-run equilibrium is characterized by nb=0 which implies \(r = c{\left( {\overline{z} - z} \right)} = \varepsilon\) and by

$$ \beta W_{n} + a{\left( {V + nV_{n} + \rho V_{b} } \right)} - {\left( {r + \rho } \right)}Y_{r} + {\left( {z - \frac{\rho } {c}} \right)}Y_{z} = 0. $$
(58)

Since the long-run levels of r and z are constant Eq. (58) determines the long-run level of n. Implicit differentiation of Eq. (58) yields

$$ \frac{{{\text{d}}n}} {{{\text{d}}\beta }} = - \frac{{W_{n} }} {{\widetilde{J}}}\quad {\text{and}}\;\;\frac{{{\text{d}}n}} {{{\text{d}}a}} = - \frac{{V + nV_{n} + \rho V_{b} }} {{\widetilde{J}}} $$
(59)

where \( \widetilde{J}: = \beta W_{{nn}} + a{\left( {2V_{n} + nV_{{nn}} } \right)} + r{\left( {r + \rho } \right)}Y_{{rr}} + z{\left( {z - \frac{\rho } {c}} \right)}Y_{{zz}} - {\left( {2rz + \rho z - \frac{{\rho r}} {c}} \right)}Y_{{rz}} \) is assumed to be negative (stability of the equilibrium) implying that dn/dbeta;>0 and dn/da>0. Since dr/dβ=dr/da=dz/dβ=dz/da=0, the results in lines 2 and 6 are straightforward.

Model C: In case of \(u = \min {\left[ {\beta W{\left( n \right)},Y{\left( {\overline{r} - nr,nz} \right)}} \right]}\) the social planner's objective is to maximize

$${\user1{\mathcal{L}}} = \beta W{\left( n \right)} + anV{\left( {n,b} \right)} + \mu nb + \lambda {\left[ {\beta W{\left( n \right)} - Y{\left( {\overline{r} - nr,nz} \right)}} \right]} + \lambda _{b} n{\left[ {B{\left( {r,z} \right)} - b} \right]}$$
(60)

and the first-order conditions read

$$ {\user1{\mathcal{L}}}_{r} = \lambda _{b} nB_{r} + \lambda nY_{r} = 0, $$
(61a)
$$ {\user1{\mathcal{L}}}_{z} = \lambda _{b} nB_{z} - \lambda nY_{z} = 0, $$
(61b)
$$ {\user1{\mathcal{L}}}_{b} = anV_{b} + n\mu - n\lambda _{b} = 0, $$
(61c)
$${\user1{\mathcal{L}}}_{\lambda } = \beta W{\left( n \right)} - Y{\left( {\overline{r} - nr,nz} \right)} = 0,$$
(61d)
$${\mathop \mu \limits^. } = \rho \mu - \beta W_{n} - a{\left( {V + nV_{n} } \right)} + \lambda {\left( {zY_{z} - rY_{r} - \beta W_{n} } \right)} - \mu b.$$
(61e)

Observe that the efficient long-run allocation (r, z, n) is characterized by

$$n_{B} {\left( {r,z} \right)} = 0,$$
(62a)
$$\beta W{\left( n \right)} - Y{\left( {\overline{r} - nr,nz} \right)} = 0,$$
(62b)
$$B_{r} Y_{z} + B_{z} Y_{r} = 0.$$
(62c)

It is interesting to note that Eqs. 62a62c are independent of the parameter a. Total differentiation of 62a62c yieldsFootnote 17

$$ {\left( {\begin{array}{*{20}c} {{B_{r} }} & {{B_{z} }} & {0} \\ {0} & {0} & {{\beta W_{n} - rY_{r} + zY_{z} }} \\ {{\phi _{1} }} & {{\phi _{5} }} & {{\phi _{6} }} \\ \end{array} } \right)} \cdot {\left( {\begin{array}{*{20}c} {{{\text{d}}r}} \\ {{{\text{d}}z}} \\ {{{\text{d}}n}} \\ \end{array} } \right)} = {\left( {\begin{array}{*{20}c} {0} \\ {{ - {\text{d}}\beta W}} \\ {0} \\ \end{array} } \right)} $$
(63)

where

$$\begin{array}{*{20}c} {\phi _{1} : = B_{{rr}} Y_{z} - B_{r} Y_{{zr}} n + B_{{zr}} Y_{r} - B_{z} Y_{{rr}} n < 0,} \\{\phi _{5} : = B_{{rz}} Y_{z} + B_{r} Y_{{zz}} n + B_{{zz}} Y_{r} + B_{z} Y_{{zr}} n < 0,} \\{\phi _{6} : = B_{r} {\left( {zY_{{zz}} - rY_{{zr}} } \right)} + B_{z} {\left( {zY_{{rz}} - rY_{{rr}} } \right)} < 0.} \\\end{array}$$

Solving the equation system 63 we obtain

$$ \frac{{{\text{d}}r}} {{{\text{d}}\beta }} = \frac{{B_{z} W\phi _{6} }} {{{\left| {\overline{J} } \right|}}} < 0,\frac{{{\text{d}}z}} {{{\text{d}}\beta }} = - \frac{{B_{z} W\phi _{6} }} {{{\left| {\overline{J} } \right|}}} < 0,\frac{{{\text{d}}n}} {{{\text{d}}\beta }} = \frac{{W{\left( {B_{r} \phi _{5} - B_{z} \phi _{4} } \right)}}} {{{\left| {\overline{J} } \right|}}} > 0, $$
(64)

where \({\left| {\overline{J} } \right|}\) is assumed to be negative. Finally we differentiate u=βW(n)=y to get

$$du = Wd\beta + \beta W_{n} dn = dy$$
(65)

which establishes du/dβ>0 and dy/dβ>0.

Model D: Now suppose u and b are specified as \(b = \min {\left[ {r,c{\left( {\overline{z} ,z} \right)}} \right]} - \varepsilon\) and \(u = \min {\left[ {\beta W{\left( n \right)},Y{\left( {\overline{r} - nr,nz} \right)}} \right]}\). Then it is easy to see that the long-run allocation is determined by \(r = c{\left( {\overline{z} - z} \right)} = \varepsilon \) and

$$\beta W{\left( n \right)} - Y{\left( {\overline{r} - nr,nz} \right)} = 0.$$
(66)

Implicit differentiation of Eq. 66 yields

$$\frac{{dn}} {{d\beta }} = - \frac{W} {{\beta W_{n} + rY_{r} - zY_{z} }} > 0,$$
(67)

since βW n +rY r zY z is assumed to be negative. Summarizing, we have dr/dβ=0, and dz/dβ=0, d(rn)/dβ>0, d(zn)/dβ>0, and du/dβ=dy/dβ=W+β W n (dn/dβ)>0.

Appendix C: Comparative dynamics in the two-species model

First, observe that

$$B^{2}_{z} dz_{{21}} = 0.$$
(68)

Next, total differentiation of Eqs. 49a, 49c49e leads to

$$ \widetilde{J} \cdot {\left( {\begin{array}{*{20}c} {{{\text{d}}r}} \\ {{{\text{d}}z_{{11}} }} \\ {{{\text{d}}n_{1} }} \\ {{{\text{d}}n_{2} }} \\ \end{array} } \right)} = {\left( {\begin{array}{*{20}c} {0} \\ {0} \\ {{ - {\text{d}}\beta W_{{n_{1} }} B^{1}_{r} - {\text{d}}a_{1} {\left( {V^{1} + n_{1} V^{1}_{n} + \rho V^{1}_{b} } \right)}B^{1}_{r} + {\text{d}}\gamma {\left( {rB^{1}_{r} + \rho + z_{{11}} B^{1}_{z} } \right)}}} \\ {{ - {\text{d}}\beta W_{{n_{2} }} B^{1}_{r} - {\text{d}}a_{2} {\left( {V^{2} + n_{2} V^{2}_{n} + \rho V^{2}_{b} } \right)}B^{1}_{r} + {\text{d}}\gamma {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}}} \\ \end{array} } \right)} $$
(69)

where

$$ \widetilde{J}: = {\left( {\begin{array}{*{20}c} {{B^{1}_{r} }} & {{B^{1}_{z} }} & {0} & {0} \\ {0} & {{n_{1} }} & {{z_{{11}} }} & {{ - z_{{21}} }} \\ {{\phi _{7} B^{1}_{{rr}} - \gamma B^{1}_{r} }} & {{ - \gamma B^{1}_{z} - \gamma z_{{11}} B^{1}_{{zz}} }} & {{{\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]}B^{1}_{r} }} & {0} \\ {{\phi _{8} B^{1}_{{rr}} }} & {{\gamma B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}}} & {0} & {{{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}B^{1}_{r} }} \\ \end{array} } \right)} $$

and

$$ \phi _{7} : = \beta W_{{n_{1} }} + a_{1} {\left( {V^{1} + n_{1} V_{{n_{1} }} + \rho V^{1}_{b} } \right)} - \gamma r < 0\;{\text{and}}\;\phi _{8} : = \beta W_{{n2}} + a_{2} {\left( {V^{2} + n_{2} V^{2}_{n} + \rho V^{2}_{b} } \right)} > 0. $$

Then solving Eq. 69 by Cramer's rule gives

$$ \frac{{{\text{d}}r}} {{{\text{d}}a_{1} }} = - \frac{1} {{{\left| {\widetilde{J}} \right|}}}z_{{11}} {\left( {B^{1}_{r} } \right)}^{2} B^{1}_{z} {\left( {V^{1} + n_{1} V^{1}_{n} } \right)}{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}, $$
(70)
$$ \frac{{dz_{{11}} }} {{da_{1} }} = \frac{1} {{{\left| {\widetilde{J}} \right|}}}z_{{11}} {\left( {B^{1}_{r} } \right)}^{3} {\left( {V^{1} + n_{1} V^{1}_{n} } \right)}{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}, $$
(71)
$$ \frac{{{\text{d}}n_{1} }} {{{\text{d}}a_{1} }} = - \frac{{{\left( {V^{1} + n_{1} V^{1}_{n} } \right)}B^{1}_{r} }} {{{\left| {\widetilde{J}} \right|}}}{\left( {B^{1}_{r} } \right)}^{2} n_{1} {\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]} + z_{{21}} \gamma B^{1}_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)} - z_{{21}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} , $$
(72)
$$ \frac{{{\text{d}}n_{2} }} {{{\text{d}}a_{1} }} = \frac{{{\left( {V^{1} + n_{1} V^{1}_{n} } \right)}B^{1}_{r} }} {{{\left| {\widetilde{J}} \right|}}} - z_{{11}} \gamma B_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)} + z_{{11}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} , $$
(73)
$$ \frac{{{\text{d}}r}} {{{\text{d}}a_{2} }} = \frac{1} {{{\left| {\widetilde{J}} \right|}}}z_{{21}} {\left( {B^{1}_{r} } \right)}^{2} B^{1}_{z} {\left( {V^{2} + n_{2} V^{2}_{n} } \right)}{\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]}, $$
(74)
$$ \frac{{{\text{d}}z_{{11}} }} {{{\text{d}}a_{2} }} = - \frac{1} {{{\left| {\widetilde{J}} \right|}}}z_{{21}} {\left( {B^{1}_{r} } \right)}^{3} {\left( {V^{2} + n_{2} V^{2}_{n} } \right)}{\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]}, $$
(75)
$$ \frac{{{\text{d}}n_{1} }} {{{\text{d}}a_{2} }} = - \frac{{{\left( {V^{2} + n_{2} V^{2}_{n} } \right)}B^{1}_{r} }} {{{\left| {\widetilde{J}} \right|}}}{\left\lceil {\gamma z_{{11}} B^{1}_{{zz}} + z_{{21}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right\rceil }, $$
(76)
$$ \frac{{{\text{d}}n_{2} }} {{{\text{d}}a_{2} }} = - \frac{{{\left( {V^{2} + n_{2} V^{2}_{n} } \right)}B^{1}_{r} }} {{{\left| {\widetilde{J}} \right|}}}{\left\lceil {{\left( {B^{1}_{r} } \right)}^{2} n_{1} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)} + \gamma {\left( {z_{{11}} } \right)}^{2} B^{1}_{{zz}} + z_{{11}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right\rceil }, $$
(77)
$$ \frac{{{\text{d}}r}} {{{\text{d}}\beta }} = - \frac{{{\left( {B^{1}_{r} } \right)}^{2} B^{1}_{z} }} {{{\left| {\widetilde{J}} \right|}}}{\left[ {W_{{n_{1} }} z_{{11}} {\left( {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right)} - W_{{n_{2} }} z_{{21}} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)}} \right]}, $$
(78)
$$ \frac{{{\text{d}}z_{{11}} }} {{{\text{d}}\beta }} = \frac{{{\left( {B^{1}_{r} } \right)}^{3} }} {{{\left| {\widetilde{J}} \right|}}}{\left\lceil {W_{{n_{1} }} z_{{11}} {\left( {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right)} - W_{{n_{2} }} z_{{21}} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)}} \right\rceil }, $$
(79)
$$ \frac{{{\text{d}}n_{1} }} {{{\text{d}}\beta }} = \frac{{B^{1}_{r} }} {{{\left| {\widetilde{J}} \right|}}}{\left[ { - W_{{n_{1} }} {\left\{ {{\left( {B^{1}_{r} } \right)}^{2} n_{i} {\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]} + z_{{21}} \gamma B^{1}_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)} - z_{{21}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} } \right\}} - W_{{n_{2} }} {\left\{ {z_{{11}} \gamma B^{1}_{r} B^{1}_{{zz}} + z_{{21}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right\}}} \right]}, $$
(80)
$$\frac{{{\text{d}}n_{2} }}{{{\text{d}}\beta }} = - \frac{{B^{1}_{r} }}{{{\left| {\widetilde{J}} \right|}}}{\left[ {W_{{n1}} {\left\{ {z_{{11}} \gamma B^{1}_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho }{{B^{2}_{z} }}} \right)} - z_{{11}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} } \right\}} + W_{{n_{2} }} {\left\{ {{\left( {B^{1}_{r} } \right)}^{2} n_{1} {\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]} + {\left( {z_{{11}} } \right)}^{2} \gamma B^{1}_{r} B^{1}_{{zz}} + z_{{11}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right\}}} \right]},$$
(81)
$$\frac{{{\text{d}}r}}{{{\text{d}}\gamma }} = \frac{{B^{1}_{r} B^{1}_{z} }}{{{\left| {\widetilde{J}} \right|}}}{\left[ {{\left( {rB^{1}_{r} + \rho + z_{{11}} B^{1}_{z} } \right)}z_{{11}} {\left( {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right)} - {\left( {z_{{21}} + \frac{\rho }{{B^{2}_{z} }}} \right)}z_{{21}} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)}} \right]}$$
(82)
$$ \frac{{{\text{d}}z_{{11}} }} {{{\text{d}}\gamma }} = \frac{{{\left( {B^{1}_{r} } \right)}^{2} }} {{{\left| {\widetilde{J}} \right|}}}{\left[ { - {\left( {rB^{1}_{r} + \rho + z_{{11}} B^{1}_{z} } \right)}z_{{11}} {\left( {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right)} + {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}z_{{21}} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)}} \right]}, $$
(83)
$$ \frac{{{\text{d}}n_{1} }} {{{\text{d}}\gamma }} = \frac{1} {{{\left| {\widetilde{J}} \right|}}}{\left[ {{\left( {rB^{1}_{r} + \rho + z_{{11}} B^{1}_{z} } \right)}{\left\{ {n_{1} {\left( {B^{1}_{r} } \right)}^{2} {\left( {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right)} + z_{{21}} \gamma B^{1}_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)} - z_{{21}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} } \right\}} + {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}{\left( {z_{{11}} z_{{21}} \gamma B^{1}_{r} B^{1}_{{zz}} + z_{{21}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right)}} \right]}, $$
(84)
$$ \frac{{{\text{d}}n_{2} }} {{{\text{d}}\gamma }} = \frac{1} {{{\left| {\widetilde{J}} \right|}}}{\left[ {{\left( {rB^{1}_{r} + \rho + z_{{11}} B^{1}_{z} } \right)}{\left\{ {z_{{11}} \gamma B^{1}_{r} B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)} - z_{{11}} \phi _{8} B^{1}_{z} B^{1}_{{rr}} } \right\}} + {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}{\left\{ {n_{1} {\left( {B^{1}_{r} } \right)}^{2} {\left( {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right)} + {\left( {z_{{11}} } \right)}^{2} \gamma B^{1}_{r} B^{1}_{{zz}} + z_{{11}} \phi _{7} B^{1}_{z} B^{1}_{{rr}} } \right\}}} \right]}, $$
(85)

where

$$ {\left| {\widetilde{J}} \right|} = B^{1}_{r} {\left\{ {n_{1} {\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]}{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}{\left( {B^{1}_{r} } \right)}^{2} + {\left[ {z_{{11}} {\left( {\gamma B^{1}_{z} + \gamma z_{{11}} B^{1}_{{zz}} } \right)} + z_{{21}} \gamma B^{1}_{{zz}} {\left( {z_{{21}} + \frac{\rho } {{B^{2}_{z} }}} \right)}} \right]}{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}B^{1}_{r} } \right\}} - B^{1}_{z} {\left\{ { - z_{{11}} {\left( {\phi _{7} B^{1}_{{rr}} - \gamma B^{1}_{r} } \right)}{\left[ {\beta W_{{n_{2} n_{2} }} + a_{2} {\left( {2V^{2}_{n} + n_{2} V^{2}_{{nn}} } \right)}} \right]}B^{1}_{r} + z_{{21}} \phi _{8} B^{1}_{{rr}} {\left[ {\beta W_{{n_{1} n_{1} }} + a_{1} {\left( {2V^{1}_{n} + n_{1} V^{1}_{{nn}} } \right)}} \right]}B^{1}_{r} } \right\}}. $$
(86)

The results of Table 2 follow from Eqs. 7086 and the properties of the functions W, B 1, B 2, V 1 and V 2. In addition, we assume \(\beta W_{{n_{i} n_{i} }} + a_{i} {\left( {2V^{i}_{n} + ^{i}_{{nn}} } \right)} < 0\) for i=1,2 and ϕ7 B rr 1−γB r 1<0 which are sufficient for the negative definiteness of \(\widetilde{J}{\left( {{\left| {\widetilde{J}} \right|} > 0} \right)}\).

Rights and permissions

Reprints and permissions

About this article

Cite this article

Eichner, T., Pethig, R. Efficient nonanthropocentric nature protection. Soc Choice Welfare 26, 47–74 (2006). https://doi.org/10.1007/s00355-005-0029-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00355-005-0029-3

Keywords

Navigation