Ecological Economics 70 (2011) 721–728

Contents lists available at ScienceDirect

Ecological Economics

j ourna l homepage: www.e lsev ie r.com/ locate /eco lecon

Analysis

Can domestication of wildlife lead to conservation? The economics of tiger farmingin China☆

Brant Abbott a, G. Cornelis van Kooten b,⁎a University of British Columbia, Department of Economics, 997-1873 East Mall, Vancouver, BC, Canada V6T 1Z1b University of Victoria, Department of Economics, P.O BOX 1700, STN CSC, Victoria, BC, Canada V8W 2Y2

☆ Brant Abbott and G. Cornelis van kooten, 2010. All rimay not be reproduced in whole or in part, by photocoppermission of the authors.⁎ Corresponding author. Tel.: +1 250 721 8539; fax:

E-mail addresses: abbottb@shaw.ca (B. Abbott), koo

0921-8009/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.ecolecon.2010.11.006

a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 June 2010Received in revised form 2 October 2010Accepted 10 November 2010Available online 22 December 2010

Keywords:Endangered speciesExtirpationWildlife farmingBioeconomics

Tigers are a threatened species that might soon disappear in the wild. Not only are tigers threatened bydeteriorating and declining habitat, but poachers continue to kill tigers for traditional medicine, decorationpieces and so on. Although international trade in tiger products has been banned since 1987 and domestictrade within China since 1993, tigers continue to be poached and Chinese entrepreneurs have establishedtiger farms in anticipation of their demise. While China desires to permit sale of tiger products from captive-bred tigers, this is opposed on the grounds that it likely encourages illegal killing. Instead, wildlifeconservationists lobby for more spending on anti-poaching and trade-ban enforcement. In this study, amathematical bioeconomic model is used to investigate the issue. Simulation results indicate that, unlessrange states are characterized by institutions (rule of law and low corruption) similar to those found in therichest countries, reliance on enforcement alone is insufficient to guarantee survival of wild tigers. Likewise,even though conservation payments could protect wild tigers, the inability to enforce contracts militatesagainst this. Our model indicates that wild tigers can be protected by permitting sale of products from tigerfarms, although this likely requires the granting of an exclusive license to sellers. Finally, it is possible totradeoff enforcement effort and sale of products from captive-bred animals, but such tradeoffs are worsenedby deteriorating tiger habitat.

ghts reserved. This manuscripty or other means, without the

+1 250 721 6214.ten@uvic.ca (G.C. van Kooten).

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Wild tiger (Panthera tigris) populations have declined from about100,000 in 1900 to perhaps as few as 5000 today. The Bali tiger becameextinct during the 1930s, the Caspian tiger during the 1970s, and theJavan tiger disappeared a decade later. Six species of tiger (Bengal,Indochinese, Amur, South China, Malaysian and Sumatran) remain,scattered throughout eastern Russia, Indochina, the Indian subcontinentand Southeast Asia (Table 1). Alongwith tiger poaching anddepletion ofprey, habitat degradation and destruction caused primarily by illegallogging contribute greatly to the demise of the tiger. Ninety-threepercent of the tiger's historic range has disappeared, while the areaknown to have been inhabited most recently by tigers has declined by41% over the past decade (Dinerstein et al., 2007).

International trade in tigers has been prohibited since 1975 whenthe species was listed under Appendix I of CITES, although the Amurtigerwas listed only in 1987 andChina imposed a domestic ban on trade

in tiger bones and medicine from tiger bone in 1993. Nonetheless,evidence indicates that illegal trade in wild tigers continues with tigerbone still used in some traditional medicines (Bhalla, 2006). WithinChina, the domestic ban coincidedwith the establishment of tiger farmsthat now house some 4000–5000 animals (Kirkpatrick & Emerton,2010; Gratwicke et al., 2008; CATT, 2007), with some evidenceindicating there are two tiger breeding facilities in Vietnam and one inThailand (Conrad, 2010).Wildlife groups are concerned that (seeminglyinevitable) sales of products from tiger farms will increase the demandfor tigers and facilitate marketing of poached animals.

The government of China has considered partially lifting itsdomestic ban on trade in tiger products to allow products fromcaptive breeding farms to be sold legally. The carcasses of tigers thathave died in captivity are currently frozen and stored as ownersspeculate that the domestic trade banwill be relaxed, although there isconcern that tiger farms are already a significant source of illegallytraded products that contain tiger bone (Nowell & Ling, 2007; EIA,2007). Opponents to the sale of captive tigers argue that anyweakening of the trade ban will legitimize consumption and increasethe demand for tiger parts; this, in turn,will increase poaching becausedetection of products from poached tigers would be more difficult(Gratwicke et al., 2008). Researchers have surveyed tiger populationsand the extent of their habitat, the availability of tiger products inChinese and international markets, the state of captive tiger breeding

Table 1Remaining tiger species, location and estimated populations, 2008.Source: http://www.savethetigerfund.org/AM/Template.cfm?Section=Community(as viewed 14 April 2009).

Tiger species Locationa Estimatedpopulation

Estimatedhabitat (km2)

Amur tiger Russia (Siberia), China 431–529 156,000Panthera tigris altaicaBengal tiger India, Nepal, Bangladesh,

Bhutan, Myanmar, China3500–4700 210,000c

Panthera tigris tigrisSouth China tiger China 20–30b 10,000c

Panthera tigris amoyensisIndochinese tiger Cambodia, Laos, Burma,

Thailand, Myanmar,Vietnam, China

750–1300 300,000c

Panthera tigris corbetti

Malayan tiger Malaysia(Malayan peninsula)

N500 45,000Panthera tigris jacksoniSumatran tiger Indonesia (Sumatra) 400–500 130,000c

Panthera tigris sumatrae

a Although four species of tiger were historically found in China, evidence suggeststhat numbers of any species would now be extremely small (Walston et al. 2010).

b Some sources indicate that the South China tiger may even be extinct (Walstonet al. 2010).

c Own calculations based on data from www.savethetigerfund.org. Sanderson et al.(Sanderson et al., 2006) provide information by region for effective potential habitat(land cover with low human influence), although this is potential and not currenthabitat. For Sumatran tiger, information is from Shepherd and Magnus (Shepherd &Magnus, 2004).

722 B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

in China, and confiscations of poached tigers, concluding that wildtigers will likely become extinct if the status quo is maintained(Gratwicke et al., 2008; Nowell & Ling, 2007; Dinerstein et al., 2007;Sanderson et al., 2006; Shepherd & Magnus, 2004; Bolze et al., 1998).

The most common recommendation for preventing extirpationof wild tigers is to increase enforcement of the trade bans, whileopposing tiger farming on the grounds that farmed output removesthe stigma of using tiger-based products and facilitates the launderingof illegal tiger parts. The so-called ‘stigma effect’ (Fischer, 2004)postulates that the demand for illegal wildlife products falls whentrade is banned. Proponents of tiger farming and trade in tiger parts,on the other hand, favor a supply-side approach to conservation,arguing that a captive breeding industry could meet all demandfor tiger products, thereby eliminating illegal killing of wild tigers andpreventing their extinction.

Damania and Bulte (Damania & Bulte, 2007; Bulte & Damania, 2005)assume imperfect competition to demonstrate theoretically thatmultiple equilibria are possible in a game between organized purveyorsof illegal wildlife products (criminal poaching gangs) and domesticwildlife farms. In their model, it is not possible to determineunambiguously whether products from captive-bred wildlife willincrease or decrease harvests of wild animals. If poachers and farmerscompete on the basis of quantity (Cournot competition), the solutionto the game leads to higher populations of wild tigers; but, if com-petition is on the basis of price (Bertrand competition), wild stocksare reduced. However, as Singh and Vives (Singh & Vives, 1984)demonstrate, the poacher and farmer are unlikely to compete on thebasis of price because they can both do better if they compete onthe basis of quantity when the goods they market are substitutes(which they must necessarily be). That is, the Cournot outcome (withhigher wild stocks) dominates the Bertrand outcome (reduced wildstock) if wild and farmed products are substitutes, and especially ifdemand is linear (as assumed by Damania and Bulte).

Damania and Bulte also argue that, since the costs of raising tigersare considerable (some $4000–$5000 annually for four years),farmers are unable to undercut suppliers of illegal wildlife products.However, they underestimate the costs of processing and marketingwild animals. As a consultant from Singapore notes: while it ischeaper to kill a tiger in the wild than raise one on a farm, the “wildtiger must be transported across numerous borders …. As an illicit

good, bribes and payoffs would be required; the rule of thumb is adoubling of price each time the cargo is handed off from one dealer toanother. China has imposed the death penalty for trafficking tigerparts, and this strong deterrent further raises the price” (Conrad,2010). Tiger farms need a retail price of $20,000 per tiger to breakeven, but current prices are higher (Conrad, 2010). But clearly theexistence of tiger farms despite a trade ban suggests that costs maynot be onerous, that there might be benefits specific to tiger farming(e.g., paid public viewing), that farmed products are somehowcircumventing the current trade ban, or some combination of thesefactors.

Similar arguments have been raised concerning the ivory tradeban. There is fear that CITES-sanctioned intermittent sales of stock-piled raw ivory from southern African states with large elephant herdspromote illegal killing of elephants. Given the extent and scope ofpoaching, van Kooten (van Kooten, 2008) found that the elephantcould go extinct in some African states despite a trade ban and highlevels of enforcement. The stigma effect appears to have had littleeffect in reducing the rate of decline in elephant populations inwest and central Africa, although Blanc et al. (Blanc et al., 2007) findelephants in east and southern Africa to be increasing by 4% annually.Van Kooten argued that the elephant is best protected by effectivelyprotecting its habitat through actual on-the-ground payments tied toelephant numbers. While conservation of tiger habitat is an importantpolicy in range states, as are captive breeding programs designed toensure survival of various tiger subspecies, economic incentives toprevent poaching and promote tiger protection have seemingly beenignored.

The present study contributes to the debate about tiger farmingby using a bioeconomic model of wild tiger population dynamics,trade and habitat to analyze the potential of heightened anti-poachingenforcement and/or liberalization of the captive tiger breedingindustry to prevent extirpation of wild tigers. A major conclusionis that anti-poaching and trade-ban enforcement must be increasedto seemingly unattainable rates if extirpation of wild tigers is tobe prevented, but that a captive breeding industry and/or effectivetransfer payments from rich countries to poor ones for protecting thehabitat could potentially prevent the extirpation of wild tigers.

The fate of the wild tiger population is modeled by a tigersurvivability function that is derived from economic principles. Thesurvivability function is a differential equation that maps the tigerpopulation, the rate of poacher detection, the output of tiger farms,the stigma effect, available habitat, and other relevant variables to therate of change in the wild tiger population. Using the survivabilityfunction, we determine for any combination of parameters whetherthe tiger population will reach a stable positive equilibrium or goextinct. The model makes no distinction between poachers andfarmers, except that the ability to sell farmed animals increases thesupply of tigers while also shifting out the demand function, which istaken to be downward sloping. We estimate the current levels of allof the parameters and then calculate how much each must change,ceteris paribus, to prevent wild tigers from becoming extinct.

2. Model of Tiger Trade

An economic model of the interplay between killing of wild tigersand culling of farmed animals is provided in Fig. 1 (van Kooten, 2008;Heltberg, 2001). When there is no ban on farmed tigers, equilibriumoccurs at point z, with the number ofwild plus farmed tigers harvestedequal to q⁎ and corresponding price of p⁎; q1 wild tigers are poachedand q⁎−q1=qlegal farm-produced tigers are killed. When there is aban on products from tiger farms, the demand curve shifts inwards asindicated — it is assumed for simplicity that the slope of the demandfunction remains constant while the intercept shifts from k to s toaccount for the stigma effect. With a trade ban and demand functionDStigma, the market equilibrium shifts from z to w, with price p⁎⁎ and

Fig. 1. Tiger market.

1 Based on our readings of newspaper accounts and agency reports, in the case oftigers, anti-poaching efforts are usually initiated and paid for by internationalconservation agencies; this makes it less likely that poachers will incur penaltiesmuch beyond confiscation and perhaps a token fine. Tigers appear to be killed mostlyby local peasants, although professional poachers are also involved. Professionalpoachers target all sorts of wild animals (as tiger numbers are low), and may beindistinguishable from non-professionals.

723B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

illegal quantity equal to q⁎⁎. In the diagram, q⁎⁎Nq1, but, if the stigmaeffect was large enough so that the post-ban demand function(DStigma) intersected the illegal supply curve to the left of v then itcould happen that q⁎⁎bq1. Further, the qualitative analysis remainsunchanged if the marginal cost of bringing wild tigers to market waslower than the marginal cost for farmed animals— say, the illegal andlegal supply functions were interchanged.

In Fig. 1, the illegal supply function is upward sloping, whileNowell and Ling (Nowell & Ling, 2007) indicate that there has beenlittle change in tiger bone prices from the early 1990s to 2006. If pricehas indeed remained constant since the Chinese trade ban came intoeffect in 1993, then either DStigma has shifted further to the left thanindicated in the figure (so it intersects the horizontal price line p⁎ atv rather than y) or the illegal supply curve is horizontal (S′illegal),indicating constant marginal costs of poaching and marketing tigers.The only time that the trade ban will stop all illegal harvests is if, inFig. 1, the illegal supply function is upward sloping and its interceptlies above p⁎ (perhaps due to successful enforcement). If not, tigerswill always be poached. The question is whether illegal harvests willstill cause an extirpation of wild tigers.

2.1. Bioeconomic Model of Tiger Exploitation

The forgoing model neglects the dynamics of tiger reproduction,habitat loss and so on. A bioeconomic analysis begins by supposingthat the population of wild tigers x is characterized by the followingsingle-species growth function with Allee effect:

g x tð Þð Þ = γx tð Þ x tð Þ−mx tð Þ + m

� �1− x tð Þ

K

� �; ð1Þ

where m is the minimum viable population, K is the populationcarrying capacity and γ is a growth constant (Boukal & Berec, 2002;Eiswerth & van Kooten, 2009). The rate at which tigers are harvestedis given by a square-root variant of the standard constant returns toscale Schaefer production function (Clark, 1990):

h x; τð Þ = θ xOτO; ð2Þ

where θ is a catchability parameter and τ is the fraction of time spentpoaching.

The expected payoff to a poacher from both legitimate activities(such as farming) and poaching is modeled as follows:

E u x; τð Þ½ � = 1−πð Þ 1−τð Þw + pθxOτOh i

+ π 1−τð Þw−η½ �; ð3Þ

where the parameter π is the probability of apprehension, p is themarket price of a tiger, and w is the wage rate in other employment.1

The first term of Eq. (3) is the payoff to a poacher who avoidsdetection, weighted by the probability of detection, while the secondterm is the payoff to a poacher who is caught. There are two parts tothe penalty a poacher faces when apprehended: The variable part ofthe penalty is confiscation of the animal parts, and the fixed part η,represents additional penalties. There is very little evidence availableas to what these non-confiscation penalties are in practice. In China,the law calls for the death penalty to be imposed for any trade in tigerparts, but the international agencies that fund poaching enforcementare less likely to view poachers as criminals than victims of poverty.Thus, we leave questions relating to punishment schedules to beaddressed when more information is available.

An expected utility maximizing poacher will choose τ so that theexpected marginal benefit from spending time poaching equals theopportunity cost of that time given by the wage rate:

Oθ xOτ–O 1–πð Þp = w ð4Þ

In the event that a poacher is apprehended, themarginal utility from timespent poaching is zero, and thus terms related to apprehension do notappear in Eq. (4). Solving Eq. (4) for τ and substituting the result intoEq. (2) gives:

h xð Þ = θ2 1−πð Þxp2w

: ð5Þ

We can also solve for the catchability parameter:

θ =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2hw

1−πð Þxp

s: ð6Þ

2.2. Tiger Survivability Function

Whether tigers are likely to survive in the wild can be determinedby the sign on the following differential equation:

dxdt

= x = g xð Þ–h xð Þ = γx x−mð Þx + m

1− xK

� �− 1−πð Þ θ

2xp2w

: ð7Þ

If it can be shown that growth of wild tigers exceeds their (illegal)harvests, so that xN0, then tiger numbers will increase over time,while they will decline if xb0. Once the relations in Eq. (7) areappropriately defined, this differential equation can be thought of asa tiger survivability function.

If demand is perfectly elastic so thatp isfixed, the survival function caneasily be parameterized. If demand is not perfectly elastic then price is afunction of output andwe replace pwith p(h). As discussed later, the casewhere the output from tiger farms affects the price of tigers is the one ofmost interest, because, if this is not so, tiger farming has no effect on wildtigers andpolicy-makersneed to thinkof strategies to savewild tigers thatare independent of decisions regarding the legitimacy of tiger farms.

The illegal supply of tigers is given by (1−π) h, or

S pð Þ = θ2 1−πð Þ2xp2w

ð8Þ

724 B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

Assume a linear derived demand function for wild tigers, D(p)=α−β p, with α, βN0. Setting S(p)=D(p) and solving for price gives:

p =2wα

θ2 1−πð Þ2x + 2wβ: ð9Þ

Now substitute Eq. (9) into Eq. (7) to obtain the followingexpression for the tiger survivability function:

x =γx x−mð Þx + m

1− xK

� �− α 1−πð Þθ2x

1−πð Þ2θ2x + 2wβ: ð10Þ

2.2.1. Anti-poaching EnforcementThe effect of anti-poaching enforcement is to halt or slow the

decline in tiger populations. Upon partial differentiating x in Eq. (7)with respect to π and then substituting for p, we get:

∂x∂π =

θ2xp2w

=αθ2x

2wβ + θ2 1−πð Þ2xN 0: ð11Þ

Both the numerator and the denominator in Eq. (11) are positive,so ∂x

∂π N 0, which implies that the effect of increased enforcement, ordetection (π), will have a positive effect on the rate of increase in thepopulation of wild tigers, ceteris paribus.

This says nothing, however, about the threshold level of detectionπ⁎ required to cause the tiger population to rise, or at least stopdeclining. Set Eq. (7) equal to zero, substitute for p from Eq. (9) andsolve for π⁎:

π� = 1− 12g xð Þ α �

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiα2−8wβg xð Þ2

θ2x

s24

35: ð12Þ

The second term on the right-hand-side of Eq. (12) is required tobe less than 1.0. The value of θ is obtained from Eq. (6) and isconstant. For baseline values of the parameters (see later discussion),π⁎=1−0.431=0.569. Given the current low levels of enforcement,it might be difficult to increase enforcement sufficiently to preventextirpation of wild tigers without implementing other policies.

2.2.2. Stigma EffectUpon partial differentiating x with respect to the shift parameter

on the demand function,

∂x∂α = − 1−πð Þθ2x

1−πð Þ2θ2x + 2wβb 0: ð13Þ

The sign is negative because both the numerator and denominatorare positive. Result Eq. (13) indicates that, if a stigma effect reduces α,it leads to an increase in the growth of the population of wild tigers.

2.2.3. Sale of Captive-bred TigersIf tigers are farmed, there will be some number Ω produced by the

farms. The supply of tigers will differ from Eq. (8) because legal saleswill need to be added to the illegal supply:

S pð Þ = θ2 1−πð Þ2xp2w

+ Ω: ð14Þ

Again set S(p)=D(p), solve for p and substitute this result intoEq. (7). The new tiger survivability function is:

x =γx x−mð Þx + m

1− xK

� �− α−Ωð Þ 1−πð Þθ2x

1−πð Þ2θ2x + 2wβ: ð15Þ

Partial differentiating xwith respect toΩ gives a result identical toEq. (13), except the sign differs:

∂x∂Ω =

1−πð Þθ2x1−πð Þ2θ2x + 2wβ

N 0: ð16Þ

The direct effect of permitting sales of farmed tigers is to increasethe population growth of wild tigers over what it would otherwisehave been. If wild tigers are declining due to poaching or loss ofhabitat, the sale of farmed tigers will reduce the rate of decline. In themodel, if enough farmed tigers are marketed, extirpation of wildtigers is prevented.

The overall impact on wild tigers from sales of farmed tigers isuncertain. Although poaching remains illegal, Chinese domestic tradeusing bred captive tigers is now taken to be legal so that the stigmaeffect no longer holds. If there is a stigma effect, the demand interceptα would decrease, perhaps to what it was prior to the trade ban. Tosee the overall effect, partial differentiate x in Eq. (15) with respect

to (α−Ω), which gives ∂x∂ α−Ωð Þ b0 if αNΩ and ∂x

∂ α−Ωð Þ N0 if ΩNα,ceteris paribus.

2.2.4. Conservation PaymentsIt is theoretically possible that poachers and consumers of tiger

products are compensated so as not to undertake these activities. It istheoretical only because, lacking many essential governance institu-tions (e.g., rule of law), it is likely impossible to enforce contracts withtiger poachers and consumers in much of Asia. Alternatively, ratherthan compensating poachers and consumers of tiger products, itmight be possible to protect wild tigers by compensating resource(habitat) owners and others who might be negatively impacted bytigers or have the incentive to poach or help poachers (perhaps onlyby not reporting their activities). Here we seek to answer the questionof how much compensation might be required to prevent extirpationof wild tigers. We do so in roundabout fashion by considering thecompensation needed to eliminate consumer plus producer surplus.

We can rewrite the respective demand and supply functions fortigers as:

p qD� �

=1β

α–qð Þ;α;β N 0; and p qS� �

=2w

1−πð Þ2θ2x q; ð17Þ

where q refers to sales (harvest) of poached tigers. The total consumerplus producer surplus (SS) in the tiger market is given by the areabetween the demand and supply functions:

SS = ∫q�

0

αβ− q

β− 2w

1−πð Þ2θ2x q� �

dq: ð18Þ

Let qbq* be the sale of tigers that would prevent their extirpation;that is, xN0, which is slightly below the level of sales (and relatedillegal harvests) that would cause x≤0. A payment of amount M isrequired to protect q�− q tigers. The size of the payment increases asthe number of sales of illegally harvested tigers falls from q* toward q.Therefore, we define the marginal surplus value, M, as the surpluscreated by the last q�− q wild tigers harvested (see Fig. 2):

M = ∫q

q�αβ− 1

β+

2w1−πð Þ2θ2x

� �q

� �dq

=αβ

q�− q

− 12β

+w

1−πð Þ2θ2x

� �q�− q 2

: ð19Þ

The number of tigers that could theoretically be saved per year ifside payments were possible is q�− q

= 1−πð Þ. The divisor (1−π)

accounts for the fact that tigers are no longer confiscated from

M

Supply

Demand

p (q )

q

M

Supply

Demand

p (q )

qq*q

Fig. 2. Non-market values transfer framework.

2 One of the journal reviewers felt it would be important to point out that this is a“very conservative estimate” of habitat, but we are unable to comment either way.

3 Available from U.S. State Department at http://www.state.gov/r/pa/ei/bgn/4130.htm (viewed September 25, 2010).

725B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

poachers because poaching is assumed to cease when conservationpayments are made.

From Eq. (19), we can solve for q as a function of M and thesolution will be quadratic with two real roots. One of the roots will beless than q* and the other greater than q*; we select the smaller valueof q*. We can simulate the effect of conservation (side) payments onthe tiger survivability function by writing the tiger growth functionas:

x = g xð Þ−h xð Þ + q� xð Þ− q x;Mð Þ1−π

: ð20Þ

Eq. (20) is another form of the tiger survivability function thatis similar to Eqs. (7), (10) and (15), but with an added term that ispositive as (q*− q)N0 and (1−π)N0. Mathematically,

∂x∂M =

−∂q∂M

1−πN 0; ð21Þ

because ∂q/∂Mb0 (as evident from Fig. 2). Thus, x will be greater forall levels of x, which implies that there is greater incentive to preservetigers. Using Eq. (21) and knowledge about the model parameters, wecan determine how large M must be to prevent extinction — that is,the size of the conservation payment required to make dx/dt positive.

2.2.5. Effect of Tiger HabitatThe analysis thus far has not directly addressed the issue of habitat

loss. The only way in which habitat loss and loss of prey can be takeninto account in the current model is through the carrying capacity, K.Partial differentiating Eq. (7) with respect to K gives:

∂x∂K =

γx2 x−mð Þx + mð ÞK2 N 0;∀x N m: ð22Þ

Not unexpectedly, an increase in K will lead to an unambiguousincrease in wild tigers, given all else held constant. We assume thatpoaching productivity is constant regardless of the size of the habitat.While search effort might increase with habitat size, policies toincrease habitat also come with greater efforts to protect tigers,reducing hunting success (Walston et al. 2010).

2.3. Tiger Survival Parameterization

From the theoretical model, we find that the rate of change in tigerpopulation increases as a result of increased habitat, conservationpayments, increased anti-poaching enforcement, and a trade ban thatoperates through the stigma effect. But this says nothing about thesurvivability of wild tigers. If wild tiger populations are declining, anincrease in these parameters might only reduce the rate of decline.Further, the effect of sales of captive-bred tigers is uncertain as itdepends on the size of the stigma effect parameter (α) and the sales of

farmed tigers (Ω). Therefore, we consider plausible parameter valuesto provide some notion of the potential impact of various policies onwild tiger populations. Our model cannot take into account subspeciesdetails regarding reproduction, habitat and minimum viable popula-tions, because,while some subspecies and regional demand informationis available, too much is lacking for such a detailed analysis.

In rainforests, tigers occur at densities of one to two tigers per100 km² because of low prey densities; in other regions with higherprey densities and/or smaller tiger subspecies, habitat can supportan average of as many as five adult tigers per 100 km2 (Save theTiger, 2009). Dietary requirements, prey densities and other factorsdetermine the size of habitat required to support tigers. In the lastcolumn of Table 1, data on currently available habitat are provided.There currently exist approximately 6000 wild tigers (=x) over arange of some 850,000 km2 (Table 1); this habitat might support17,000 tigers (at two tigers per 100 km2), so K=17,000.2

To maintain a minimum viable population of six breeding females(perhaps as few as 20 animals total), reserves need to be aminimumofabout 500 km2 for Bengal tigers, since prey are generally abundant intheir native habitat, to more than 2000 km² for the Amur (Siberian)tiger. Thus, habitat and prey availability are important factors affectingwild tiger populations. If we assume there are 76 tiger conservationlandscapes (Sanderson et al., 2006) and a minimum population ofapproximately 20 animals in each to ensure survival, then m≈2000(Reed et al., 2003). To determine the growth constant γ, we need anestimate of the growth rate of awild tiger population that is not subjectto poaching. One such estimate is provided by Sanderson et al.(Sanderson et al., 2006) who indicate that only 20% of newborn tigerswill have the opportunity to breed, annual reproduction averages0.61 per young female, and there are 2.5 females per male. Then thegrowth constant is γ=0.087 (=0.2×0.61×2.5/3.5).

Economic data are even more difficult to find. Nowell and Ling(Nowell & Ling, 2007) report that, in 1992 (before the Chinese ban),some 200 wild tigers were harvested with a total industry value of$US 12.4 million, or $62,000 per tiger. There is evidence that poachersworking in the forest only receive about $800 per tiger, but that thosewith highly sophisticated criminal gangs receive considerably more(CITES, 1999). We choose a price of p=$5000 per wild tiger, but alsoconsider scenarios with lower prices to provide some notion of thedirectional effect of prices.

Each tiger produces between 5 kg and 12 kg of dry bone (Ng &Nemora, 2007, p.8); thus, 200 wild tigers would yield 1000–2400 kgof bone. Over the period 1999–2005, an average of 60 kg of tiger bonewas seized annually, or 2.5 to 6.0 % of the 1992 illegal harvest. Weassume a base-line detection rate of π=0.04, approximately halfwaybetween 2.5 and 6%.

To determine the stigma effect, it is necessary to know somethingabout the demand function and sales of tigers before and after theChinese domestic trade ban. Given that 23 tigers were marketed inshops in Indonesia (Ng & Nemora, 2007) and there might be 500tigers in that country (Table 1), we get a poaching rate of 4.5%. If thisrate applied just before the trade ban when there were some 7000tigers, then 315 tigers would have been harvested and sold. After theban, when there were an assumed 6000 tigers, 270 would have beensold. Given evidence that prices had remained relatively stable fora long period (Nowell & Ling, 2007), and given a linear demandfunction, we could conclude that the trade ban shifted the demandfunction so that, on the horizontal axis, the intercept shifted to the leftby 50 — the extent of the stigma effect.

Finally, average per capita GDP in Vietnam reached about $US 1000in 2008,3 which was still below the average in most other Asian

Table 2Model parameters: summary.

Item Parameters and their valuesa

Population dynamics γ=0.087 K=17,000 m=2000Harvest levels h=q=315 (trade) h=q=270 (ban)Economic parameters w=$450 π=0.04 p=$5000

Demand parameters p=$5000 p=$1500

α αstigma α αstigma

More elastic β=0.01 365 320 330 285Baseline β=0.1 815 770 465 420Less elastic β=0.2 1315 1270 615 570

a The demand function is q=α−βp. Slope parameters (β) are assumed values, priceis assumed to be $5000/tiger unless otherwise indicated, while pre- and post-Chinesetrade ban illegal harvests (q) are assumed to be 315 and 270, respectively. The ‘nostigma’ and ‘stigma’ values are then derived as the demand intercepts. Thecorresponding intercepts in Fig. 1 are k=α/β and s=αstigma/β. The values of h (=q),w, π, p and the assumed starting or current population (x=6000) are used to derive thecatchability coefficient (θ) and, together with θ, used in Eqs. (10), (15), (23) and (10)again to derive Figs. 3 to 6, respectively, and in Eq. (12) to derive results in Table 3.

Table 3Critical detection rates to prevent extirpation of wild tigers.

Price=$5000 per tiger Price=$1500 per tiger

β=0.01 β=0.1 β=0.2 β=0.01 β=0.1 β=0.2

Habitat capacity=17,00090.1% 56.9% 49.3% 96.8% 76.3% 64.7%

Habitat capacity=40,00086.4% 37.5% 29.1% 95.7% 65.6% 47.9%

200

726 B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

countries. However, peasants are unlikely to earn wages equal to theper capita GDP of the country in which they reside. We simply assumethat those actually killing tigers for the organized gangs have anopportunity annual wage w=$450.

A summary of the parameter values employed in the model isprovided in Table 2.

3. Numerical Simulation Results

The tiger survivability function can be viewed graphically byplotting dx/dt on the vertical axis against x(t) on the horizontal axis.For any x(t) for which dx/dtN0, the tiger population is growing, whileit shrinks whenever dx/dtb0. If dx/dt b 0 for all x≤x(t) then the tigerwill become extinct. This is the case in Fig. 3 for parameters associatedwith four of the scenarios in Table 2. Our model supports theconclusions of tiger researchers as it predicts extinction of the wildtiger, even under a Chinese trade ban, and this result is robust to alarge number of scenarios. Now consider policies that might preventextirpation.

To prevent tigers from becoming extinct, policies need to reducethe prevalence of poaching. Tigers are fairly susceptible to modestincreases in mortality, and less likely to recover quickly afterpopulation declines (Chapron et al., 2008). Since the current tigerpopulation is above the hypothesized minimum viable population,policy can affect the parameters of the tiger survivability functionso that it becomes positive at x(t)=6000. Increased expenditureson enforcement will result in higher rates of poacher detection, π. Ifthe rate at which poachers are caught increases sufficiently, it will

-270

-225

-180

-135

-90

-45

0

0 2000 4000 6000 8000 10000

dx/d

t

Population of wild tigers

β=0.1, p=$5000 β=0.01, p=$5000 β=0.2, p=$5000 β=0.01, p=$1500

Fig. 3. Rate of change in wild tiger populations for various population levels under atrade ban, x0=6000, π=0.04, selected values of β, and prices of $5000 and $1500.

become far less profitable for poachers to harvest tigers, and poachingpressure will be reduced to a level at which the survivability functionis positive. The critical levels of detection π⁎ required to keep wildtiger populations from declining can be determined from Eq. (12),and are provided in Table 3. Clearly, levels of enforcement must bedauntingly high to prevent tigers from going extinct.

A second policy option is to permit farmed tiger products. Thepresence of captive-bred tigers would increase the total supply oftiger products, reducing the price of those products and makingpoaching a less profitable occupation. If the price is reduced enough,the tiger survivability function will become positive and extinctionwill be prevented. However, by introducing farmed tigers to themarket, the stigma effect associated with the prohibition of tigertrade will disappear, which will tend to increase poaching. Resultsfor selected scenarios are provided in Fig. 4; these indicate that thecritical level of annual sales causing the rate of change in tigerpopulations to become positive is between 160 and well over 400farmed tigers.

We are also interested in a combination of policy options that canprevent extirpation of wild tigers. For example, we may want to knowwhether a particular combination of enhanced enforcement and farmedproducts can prevent extinction. There are many possible minimumcombinations of the two policies that can prevent extirpation, and thesecombinations can be used to determine the ‘extinction prevention’policy frontier. Setting ˙x=0 in Eq. (15) and solving for Ω gives thefollowing relation for the policy frontier:

Ω = α− 1−πð Þ + 2wβ1−πð Þ2θ2x

� �g xð Þ: ð23Þ

Plotting various combinations of Ω and π gives the extinctionprevention policy frontier plotted in Fig. 5. Clearly, detection rates(expenditure on enforcement) must be increased significantly if onewishes to reduce reliance on sales of farmed tigers to protect wild

-200

-100

0

100

0 50 100 150 200 250 300 350 400

dx/dt

Sales of captive-bred tigers ( )

β=0.01, p=$5000 β=0.2, p= $5000

β=0.1, p=$1500 β=0.2, p=$1500

Fig. 4. Effect that sales of captive-bred tigers have on rate of change in wild tigerpopulation, x0=6000, π=0.04, selected values of β, and prices of $5000 and $1500.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 100 200 300 400 500 600 700

Det

ectio

n R

ate

(π)

p =1500

p =$5000

Sales of captive-bred tigers ( )

Fig. 5. Extinction prevention policy frontier, x0=6000, β=0.01.

727B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

stocks, with the initial tradeoff steeper when poachers receive lowerprices.

Finally, we consider a policy option that focuses on habitat to theexclusion of other policies, except the current level of enforcementremains. The effect of protecting or expanding habitat is to increasethe ecosystem's carrying capacity (K) — its ability to support tigers.In Fig. 6, we provide an indication of this effect. Notice that, ifthe population of wild tigers is approximately 6000, an increase incarrying capacity to perhaps 35,000 (from 17,000) might lead to apositive growth in wild stocks. This occurs in the absence of increasesin poaching detection rates and sales of farmed tigers. However, itmay be necessary to implement some of the other policies in order toincrease tiger populations in the short termwhile efforts to protect andincrease habitat and ecosystem carrying capacity are implemented(see Table 3).

4. Discussion

Given the complexity of tiger protection, our results suggest that,because habitat is being eroded, neither legitimizing trade in productsfrom captive bred tigers nor increased enforcement are likely able toprevent wild tigers from being extirpated. Rather, a cocktail of policieswill be needed to give wild tigers a chance of surviving. Clearly, ifgovernance institutions found in developed countries (rule of law,low levels of corruption, etc.) characterized range states, the tigerwould survive in the wild (Bulte et al., 1–3 December 2003). Thesekinds of institutions lead to rates of detection that exceed thoserequired to preserve wild tigers. Our results also indicate thatconservation payments from rich countries to poor range states canbe effective in protecting tigers, and that such payments need not beonerous. But again, lack of adequate institutions precludes writing

-250

-200

-150

-100

-50

0

50

100

150

2000 4000 6000 8000 10000

dx/d

t

Tiger population

increasing carrying capacity

base case

K=45,000

K=35,000

K=25,000

Fig. 6. Effect of increased habitat availability on the tiger survivability function, basecase parameters, and various carrying capacities.

enforceable contracts that protect habitat and prevent poaching oftigers.

In the absence of the required institutions or effective community-based natural resource management regimes that inhibit illegaltakings, our results indicate that the sale of tiger products from tigerfarms could reduce poaching sufficiently to enable wild tigers toreproduce faster than they are killed. In the absence of othermeasures(additional food or habitat), a combination of increased enforcementand legal sales offers the best chance of wild tiger survival. However,the loss of quality habitat makes it extremely difficult to design aneffective strategy for saving wild tigers.

Our simulation results assume that anti-poaching enforcementefforts and the demand for poached tigers will be unaffected if tigerfarming is legitimized, other than through a stigma effect that causesdemand to shift outwards if trade is permitted. These assumptions arecertain to generate debate, as there are several arguments againstthem. Most commentators have argued against legalization of tigerfarming on the following grounds (CITES, 2001; Kirkpatrick & Emerton,2010; Nowell & Ling, 2007):

1. It increases the demand for poached tigers because farmers willpurchase them to increase their captive stocks.

2. It increases poaching because it is much cheaper to poach a tigerthan to raise one in captivity; thus, producers of tiger productspurchase poached animals and sell them as if they were bred incaptivity. This is one means by which a legalized activity facilitatesan illegal one.

3. Because there will be a legitimate as well as illegal supply of tigers,it will be harder to recognize poached tigers and the effectivenessof anti-poaching efforts will be reduced.

These claims would certainly be warranted if tiger farming wasunregulated with many competitive firms. However, if tiger farmingis concentrated in a single or a very small number of well regulatedmonopolistic firms, these concerns may not materialize, as wedemonstrate using a related historical example.

For 200 years spanning almost the entire length of the colonialNorth American fur trade, the Hudson's Bay Company (HBC) held amonopoly over most of what is now Canada, east of the continentaldivide (Rocky Mountains). The HBC aggressively self-policed theregion they controlled to ensure that their full monopoly rights wereupheld (Gough, 2007). Consequently, the HBC was able to restrict theflow of beaver pelts out of North America. This had two importanteffects: first, the price of beaver pelts was much higher than it wouldotherwise be, leading to large profits for the HBC. Second, and perhapsmore important in the current context, the population of beaversremained viable because of the conservation effect of restricted trade.

The lands to the west of the North American continental dividewere not controlled by any form of monopolistic company until 1824.In fact, during the late 18th and early 19th centuries, many companiesfrom Great Britain, the United States, Spain and Russia competed forfurs along the northwest coast of North America. The prize of the furtrade in this region was the sea otter, and, as a result of unrestrictedcompetitive trade, the sea otter population plummeted rapidly fromperhaps 300,000 to less than 2000 (Reidman et al., 1990). Even today,the sea otter remains an endangered species.

Although there is a fundamental difference between granting amonopoly to harvest wild animals and one to sell farmed animals, thelesson about the incentive to defend property rights still applies. Thestructure of the tiger farming industry could have a tremendous effecton the fate of the tiger. To the extent that China is the primary marketfor tiger products and/or Chinese tiger farms can collude with thesmall number of tiger farms outside China, the granting of a monopolycharter could ensure that tiger products provide monopoly rents.In this case, the granting of a monopoly charter could lead to greateranti-poaching enforcement and reduce the demand for poached

728 B. Abbott, G.C. van Kooten / Ecological Economics 70 (2011) 721–728

products as the charter holder may take action to prevent poachersfrom dissipating those rents.

It would be in the interest of a monopoly cartel to take actionsagainst poachers to protect their profits. A highmonopolistic price canonly be maintained if the cartel prevents poaching or the sale ofpoached tiger products. The monopoly charter should give the cartelthe right to police poachers. Given the current detection rate is onlysome 3%, extant methods of anti-poaching enforcement are clearlyineffective, perhaps because government enforcement officers aresusceptible to corruption. Those involved in anti-poaching activitiesare not impacted financially by the success of poachers, except to theextent that they can be bribed. A cartel, on the other hand, has greatincentive to prevent poaching because poaching threatens their rents.

To help ensure that poachedwild tigers are not ‘laundered’ into thestock of captive bred tigers, an animal registration program similar tothat used for cattle in Europe and North America could be adopted. Inthe cattle sector, animals are registered with the government at birth,identified by ear tags, frequently branded, and so on. All captive bredtigers could similarly be registered at birth with the registrationsystem monitored for compliance not only by various governmentsbut also by a credible international organization such as CITES thatwould certify products. Only animals born in captivity to registeredparents could be culled to produce medicines and other goods,with monitoring again performed by an outside certifier. Such ascheme would address the three concerns raised by various wildlifeprotection groups. Monopolistic power, regulatory restrictions andmonitoring are all required to give wild tigers their best chanceof survival, although other policies (e.g., conservation payments)would need to address the problem of habitat loss and depletion ofthe tiger's prey.

Ethical and ideological considerations are important factors nottaken into account in the forgoing analysis. Clearly, one can object tothe slaughter of tigers, but it occurs despite our objections. One canalso object on ethical grounds to the sale of products from tiger farms,except that it is difficult to argue against tiger farms while acceptingthe production of beef, poultry, pig and other commodities from whatare best described as animal manufacturing facilities. Unfortunately,about all that we can conclude from our analysis is that, if wild tigersare to be preserved, we must adopt a pragmatic strategy that includesefforts to protect habitat, enforce bans on poaching and internationaltrade, and enable countries to develop and implement institutionsthat reduce opportunities for illegal activities of all kinds. But wemust also be prepared to adopt approaches that might be difficult toaccept from an ethical and ideological perspective, and that could wellinclude the sale of products from tiger farms (Rao, 2008 March 17).

Acknowledgements

The authors wish to acknowledge funding support from theCanada Research Chairs program, and comments and suggestionsprovided by Kirsten Conrad, Erwin Bulte and participants at theTRAFFIC Workshop on tigers held at Cambridge, UK, August 2008.Remaining errors are attributable solely to the authors.

References

Bhalla, N., 2006. Forest people must help conserve India's tigers. Available from Reuters,New Delhihttp://www.alertnet.org/thenews/newsdesk/SP308865.htm. accessedMarch 2008.

Blanc, J.J., Barnes, R.F.W., Craig, G.C., Dublin, H.T., Thouless, C.R., Douglas-Hamilton, I.,Hart, J.A., 2007. African elephant status report 2007: an update from the elephantdatabase. Occasional Paper Series of the IUCN Species Survival Commission, No. 33.Gland, Switzerland: IUCN/SSC African Elephant Specialist Group. vi+276 pp.

Bolze, D., Chetkiewicz, C., Mingjiang, Q., Krakower, D., 1998. . Available from Theavailability of tiger-based Traditional Chinese Medicine products and publicawareness about the threats to the tiger in New York City's Chinese communities:a pilot study. Working Paper No. 12. Wildlife Conservation Society, New York.http://wcs.org/media/file/mime-9967825651.pdf (accessed June 2008).

Boukal, D.S., Berec, L., 2002. Single-species models of the Allee effect: extinctionboundaries, sex ratios and mate encounters. Journal of Theoretical Biology 218,375–394.

Bulte, E.H., Damania, R., 2005. An economic assessment of wildlife farming andconservation. Conservation Biology 19, 1222–1233.

Bulte, E.H., van Kooten, G.C., Swanson, T., 1–3 December 2003. Economic incentives andwildlife conservation. Available from CITES, Geneva, Switzerland. http://www.cites.org/eng/prog/economics/CITES-draft6-final.pdf. accessed July 2008.

CATT, 2007. CATT Alert #49. Thousands ask China to keep Tiger Trade Ban. Availablefrom http://www.savethetigerfund.org/CATT. accessed June 2008.

Chapron, G., Miquelle, D.G., Lambert, A., Goodrich, J.M., Legendre, S., Clobert, J., 2008.The impact on tigers of poaching versus prey depletion. Journal of Applied Ecology45, 1667–1674.

CITES, 1999. Issues relating to species Tiger. Available from Technical Missions.Convention on International Trade in Endangered Species of Wild Fauna andFlora, Forty-second meeting of the Standing Committee in Lisbon, 1999. StandingCommittee 42, Doc. SC42.10.4. http://www.cites.org/eng/com/SC/42/42-10-4.pdf(accessed June 2008).

CITES, 2001. Tiger enforcement task force to target criminal networks. Available fromPress Release. http://www.cites.org/eng/news/press/2001/010330_tiger.shtml(accessed June 2008).

Clark, C.W., 1990. Mathematical Bioeconomics, 2nd ed. John Wiley & Sons, New York,NY.

Conrad, Kirsten. 2010. Personal Communication, April 28, 2010. Communication related to:http://online.wsj.com/article/SB10001424052748704454304575081042120355402.html?mod=WSJ_latestheadlines (accessed April 2010).

Damania, R., Bulte, E.H., 2007. The economics of wildlife farming and endangeredspecies conservation. Ecological Economics 62, 461–472.

Dinerstein, E., et al., 2007. The fate of wild tigers. Bioscience 57, 508–514.EIA, 2007. Availability of tiger bonewine atwild animal parks in China. BriefingDocument.

http://phttp://www.eia-international.org/files/reports164-1.pdf. Available from.Eiswerth, M.E., van Kooten, G.C., 2009. The ghost of extinction: preservation values and

minimum viable population in wildlife models. Ecological Economics 68 (7),2129–2136.

Fischer, C., 2004. The complex interactions of markets for endangered species products.Journal of Environmental Economics and Management 48, 926–953.

Gough, B., 2007. Fortunes a river: the collision of empires in Northwest America. HarbourPublishing, Madeira Park, BC.

Gratwicke, B., Mills, J., Dutton, A., Babriel, G., Long, B., Seidensticker, J., Wright, B., You,W., Zhang, L., 2008. Attitudes toward consumption and conservation of tigers inChina. PLoS ONE 3 (7), e2544. doi:10.1371/journal.pone.0002544.

Heltberg, R., 2001. Impact of the ivory trade ban on poaching incentives: a numericalexample. Ecological Economics 36, 189–195.

Kirkpatrick, R.C., Emerton, L., 2010. Killing tigers to save them: fallacies of the farmingargument. Conservation Biology 24, 655–659.

Ng, J., Nemora, 2007. Tiger trade revisited in Sumatra, Indonesia. TRAFFIC SoutheastAsia, Petaling Jaya, Selangor, Malaysia. 20 pp.

Nowell, K., Ling, X., 2007. Taming the tiger trade. China's markets for wild and captivetiger products since the 1993 domestic trade ban. 64pp. Available from TRAFFICEast Asia, Hong Kong. http://www.ifaw.org/ifaw/dimages/custom/2_Publications/Wildlife/MadeinChina_farmingTigersToExtiction.pdf. (accessed June 2008).

Rao, J., 2008 March 17. How to kill a tiger, the Indian express, Monday.Reed, D.H., O'Grady, J.J., Brook, B.W., Ballou, J.D., Frankham, R., 2003. Estimates of

minimum viable population sizes for vertebrates and factors influencing thoseestimates. Biological Conservation 113, 23–34.

Reidman, M.L., Estes, J.A., Reidman, M.L., Estes, J.A., 1990. The sea otter (Enhydra lutris):behavior, ecology, and natural history. Washington, D.C.: U.S. Fish and WildlifeService Biologist Report, 90(14).

Sanderson, E.W., et al., 2006. Setting priorities for the conservation and recovery of wildtigers: 2005–2015. Available from The Technical Assessment. Washington, DC:WCS, WWF, Smithsonian and NFWF-STF. http://www.savethetigerfund.org/AM/Template.cfm?Section=Full_Reports (accessed April 2009).

Save the Tiger, 2009. http://www.savethetigerfund.org/Content/NavigationMenu2/Community/Distribution/default.htm. (accessed 16 April 2009).

Shepherd, C.R., Magnus, N., 2004. Nowhere to hide: the trade in Sumatran tiger.TRAFFIC Southeast Asia, Petaling Jaya, Selangor, Malaysia. 98pp.

Singh, N., Vives, X., 1984. Price and quantity competition in differentiated duopoly. TheRand Journal of Economics 15, 546–554.

van Kooten, G.C., 2008. Protecting the African elephant: a dynamic spatial priceequilibrium model of ivory trade. Biological Conservation 141, 2012–2022.

Walston, J., Robinson, J.G., Bennett, E.L., Breitenmoser, U., da Fonseca, G.A.B., et al., 2010.Bringing the tiger back from the brink — the six percent solution. PLoS Biology 8.doi:10.1371/journal.pbio.1000485.