Heifer-in-Trust: a model for sustainable livestock development?

K.A. Afifi-Affat

The author's address is: Chinook, 36 Windward Lane, Holcombe, Dawlish EX7 0JQ, UK.
Note: The work reported in this paper was completed while the author was at the School of Development Studies, University of East Anglia, Norwich NR4 7TJ, UK.

�G�NISSE EN PROPRI�T� FIDUCIAIRE�: UN MOD�LE POUR LE D�VELOPPEMENT DURABLE DE L'�LEVAGE?

L'objectif des plans �G�nisse en propri�t� fiduciaire� (GPF) est en principe de prot�ger et d'accro�tre les recettes des exploitants � faible revenu en leur permettant de se procurer des femelles relativement co�teuses � travers un accord de pr�t �en nature�. En effet, la possession d'un animal en mesure d'am�liorer l'�tat nutritionnel et de fournir d'int�ressants sous-produits comme le fumier, renforce avant tout la dotation en ressources de la personne concern�e. Ensuite, la production et la vente de produits laitiers et de viande, et l'utilisation de la traction animale (dans le cas des bovins) peuvent constituer une source additionnelle de revenus. Enfin, dans les p�riodes difficiles, par exemple en cas de s�cheresse, la vente des animaux en surplus peut permettre d'�viter la famine ou l'endettement. La viabilit� des plans GPF est �valu�e ici du point de vue du projet et des participants. Elle est mesur�e � l'aide d'un mod�le de calcul dans lequel les param�tres de reproduction et de mortalit� animales peuvent �tre divers. Les moyens d'am�liorer la durabilit� des plans GPF y sont �tudi�s. Le mod�le montre qu'il est fort peu probable que les plans GPF standard soient viables � la fois du point de vue du projet et des participants. Si les plans pr�voient le remboursement d'une seule g�nisse, le capital initial que repr�sentent les 100 g�nisses de base sera rapidement �puis�. En revanche, si le remboursement pr�vu consiste en deux g�nisses, de nombreux participants ne seront alors pas en mesure de rembourser leur emprunt, tandis que d'autres n'auront pas de g�nisse substitutive au moment o� la g�nisse de base mourra. Pour que les plans GPF profitent v�ritablement aux ruraux pauvres, ils doivent: i) mettre en œuvre un plan d'assurance; ou ii) pr�voir un remboursement en esp�ces qui permette aux exploitants de r�aliser � plein la valeur des veaux m�les; et iii) am�liorer de fa�on sensible le taux de v�lage et r�duire la mortalit�. Si le plan ne pr�voit pas d'assurance, il sera alors n�cessaire de rembourser plus d'une g�nisse pour chaque g�nisse de base re�ue.

NOVILLAS EN DEP�SITO: �MODELO PARA UN DESARROLLO PECUARIO SOSTENIBLE?

En teor�a, los planes de novillas en dep�sito buscan tanto proteger como aumentar los ingresos, proporcionando a los agricultores de bajos ingresos acceso a hembras de ganado de valor relativamente elevado mediante un acuerdo de pr�stamo �en especie�. En primer lugar, la dotaci�n de una persona mejora con la posesi�n de un animal que pueda contribuir a una mejor nutrici�n y proporcionar subproductos valiosos, como el esti�rcol. En segundo lugar, los ingresos pueden aumentar por medio de la obtenci�n y la venta de productos l�cteos y de carne y la utilizaci�n de la fuerza de tracci�n animal (en el caso de los vacunos). Por �ltimo, durante los per�odos dif�ciles, por ejemplo de sequ�a, los animales excedentes se pueden vender para evitar la inanici�n o la deuda. La sostenibilidad de los planes de �novillas en dep�sito� se eval�a por proyectos y por participantes. La sostenibilidad se valora utilizando un modelo de hoja de c�lculo en el que pueden variar los par�metros de la reproducci�n y la mortalidad del ganado. Se examinan las maneras de mejorar la sostenibilidad de dichos planes. El modelo muestra que los planes normales de novillas en dep�sito son poco sostenibles, tanto para los proyectos como para los participantes. Si en los planes se exige la devoluci�n de una novilla solamente, se agotar� con rapidez el capital inicial, representado por 100 novillas de partida. Si los planes exigen la devoluci�n de dos novillas, habr� un n�mero considerable de participantes que no podr�n devolver el pr�stamo, y otros no tendr�n ninguna novilla de sustituci�n cuando desaparezca la de partida. Para que los planes de novillas en dep�sito puedan beneficiar eficazmente a la poblaci�n rural pobre, ser� necesario aplicar un plan de seguro, o bien permitir la devoluci�n en dinero, con lo cual los agricultores podr�n aprovechar plenamente el valor de los becerros, e introducir mejoras considerables en la tasa de natalidad y reducir la mortalidad. Para que un plan funcione sin ning�n seguro es necesario devolver m�s de una novilla por cada una de partida recibida.

INTRODUCTION

One approach to identifying the poor is based on level of annual income and, with such a view, the reduction of poverty primarily involves increasing average income levels. Development schemes based on this understanding of poverty aim to improve people's incomes, and many have focused on the provision of credit. Unfortunately the very poor are often not in a position to take risks or use credit effectively (Johnson and Rogaly, 1997). This view of poverty was challenged by Sen (1981) who suggested that the key cause of chronic poverty, for example in a famine, was "entitlement failure". This entitlement failure can happen either because of a fall in a person's endowment, e.g. alienation of land, or because of an unfavourable shift in a person's exchange entitlement, e.g. loss of employment, fall in wages, rise in food prices (Dreze and Sen, 1989). The poor are therefore defined as those who are most vulnerable to entitlement failure. This alternative view of the "vulnerable poor" has been modified by others such as Swift (1989) and de Waal (1990) to consider people's coping strategies and their assets and claims in addition to their entitlements. Development schemes can therefore aim to reduce poverty by reducing the poor's vulnerability to entitlement failure, by improving their coping strategies and by increasing their access to assets and claims. In effect, the aim is income protection rather than income promotion (Dreze and Sen, 1989).

Heifer-in-Trust (HIT) schemes theoretically seek both to protect and to increase income by providing low-income farmers with access to relatively high-value female livestock through a loan "in kind" agreement. First, a person's endowment is improved by owning an animal that can contribute to improved nutrition and provide valuable by-products such as manure. Second, income can be increased through the production and sale of dairy products and meat and the use of animal draught (in the case of cattle). Finally, during difficult times such as drought, surplus animals can be sold to avoid starvation or debt.
HIT schemes generally distribute female dairy animals (i.e. foundation heifers) which farmers are loaned on condition that they repay to the project the first heifer calf born to each animal. Some schemes may also require the repayment of a second calf. The foundation animal remains the property of the HIT project until the credit is repaid. The animals returned to the project are "passed on" to additional farmers who must also repay the loan in the same manner. In theory, this receiving and passing on can continue indefinitely (de Wolff, 1996) and, in this sense, HIT is simply a type of revolving credit scheme.
The organization with the most experience in HIT schemes is Heifer Project International (HPI), which was founded in 1944 by Dan West as an interfaith, non-profit, Christian-inspired organization. HPI refers to the revolving credit element of HIT schemes as "passing on the gift". Pelant (1992) examined HPI's experience with HIT and acknowledged that it takes a long time for farmers to repay the loans. He suggests, for example, that 80 heifer calves may be repaid from 100 foundation heifers within seven years of their distribution. Despite this, Pelant (1992, p. 10) claimed that the "pass-ons continuously produce new generations of animals" and that HIT schemes stretch the value of donor funds by two to five times.
The greatest number of HPI schemes have been in the United Republic of Tanzania where 6 700 low-income farming families have received their "first improved dairy animal through an in-kind credit scheme" (Kinsey, 1996, p. 76). Based on schemes which require farmers to repay two calves for each heifer received, Kinsey (1994) claimed that on average, by the eighth year of a project, the number of project farmers doubles through pass-ons, but he also acknowledged that there is a wide variation in this figure from project to project. Despite the fact that the sustainability of benefits for farmers is not known because HPI does not monitor farmers once they have repaid their loan, Kinsey (1994) suggested that offspring free of debt far outnumber the foundation heifers plus the pass-on heifers.
Similar HIT schemes are run by other organizations in the United Republic of Tanzania. Houterman et al. (1993) report on the Kagera Livestock Development Programme (KaLiDeP) which requires the repayment of only one heifer calf. They estimate that only 63 percent of heifers are repaid owing to the high mortality among foundation heifers and that it takes 11 years for heifer numbers to recover to the original number distributed. However, after 14 years there are twice as many heifers. From this analysis, Houterman et al. (1993) concluded that the KaLiDeP HIT scheme has a limited capacity to generate bred heifers for new dairy farmers and is by no means a perpetual system. De Wolff (1996) concurred that KaLiDeP, and more generally HIT schemes, are unsustainable. Rutamu (1996) and Rutamu et al. (1994) tell a similar story based on the experience of the Tanga Smallholder Dairy Development Programme (TSDDP): after seven years only 20 percent of farmers had repaid their loans.
A number of suggestions of ways to improve the success of HIT schemes have been made. Kinsey (1994) suggested that, in certain circumstances, a bull calf may be exchanged for a heifer calf to repay the loan, and de Wolff (1996) suggested repayment should be possible in the form of cash, a bull calf or a heifer calf. Another way to improve the sustainability of HIT schemes is to insure foundation heifers against mortality (Houterman et al., 1993). To reduce corruption and defaulting by farmers, de Wolff (1996) suggests that a strong functional group should exist at the village level to exert peer pressure on members. At the participant level, Kinsey (1994) identified the need for additional credit so that farmers can purchase veterinary services and drugs. Rutamu et al. (1994) pointed to the need for good farm management supported by a dairy infrastructure their can supply feed, veterinary services and quality heifers, and market surplus milk. Rutamu (1996) also suggested the need for an effective extension service.
For any credit scheme to be financially sustainable it must be able to cover the cost of the funds; it must cover its operating costs and non-performing loans; and, finally, it must ensure its funds grow at the same rate as inflation (Havers, 1996). A credit scheme would normally cover these costs through charging fees and interest on its loans. Traditionally HIT schemes have not charged fees or interest on animal loans. While animals are not affected by inflation, the other costs noted above must be met if the scheme is not to be described simply as "grant giving". However, before one even considers the overall financial performance of HIT schemes, it is essential to determine whether such schemes are sustainable at the project and participant levels.
This article assesses the conditions under which HIT schemes are sustainable at both project and participant levels. Through the use of a spreadsheet model in which livestock reproductive and mortality parameters can be varied, a number of scenarios are evaluated against specific sustainability criteria. The article first describes the structure of the basic model as well as several possible modifications. The results of the use of the models are then presented and discussed and, finally, the implications for the sustainability of HIT schemes are considered.

MODELLING HEIFER-IN-TRUST SCHEMES

Initial Model

The Initial Model is designed to reflect the manner in which many HIT schemes are currently implemented. The majority of HIT schemes require the repayment of the first heifer calf born to the foundation heifer. Once the heifer calf has been repaid, any further calves belong to the farmer. However, some HIT schemes also require the repayment of a second heifer calf. Hence, the model is split into three stages: the first includes foundation heifers which have not yet repaid any calves; the second includes foundation heifers which have repaid one calf but must repay a second; and the third includes heifers which are free of debt.

All foundation heifers are initially placed in stage 1 of the model. Once the foundation heifer has repaid the first heifer calf, it is moved to stage 2 if a second calf is to be repaid. When the second calf has been repaid, the foundation heifer is moved to stage 3. If the repayment of only one calf is required the foundation heifer is moved to the third stage of the model directly, skipping stage 2. The structure of the Initial Model is illustrated in Figure 1.

1
Structure of the Initial Model
Structure du mod�le initial
Estructura del modelo inicial

In the tropics, temperate and tropical types of cattle have an average age at first calving of three and a half to four years (Chamberlain, 1989). Better management and better feeding can reduce the age at first calving to 24 months in an intensive system (Matthewman, 1993). In HIT schemes heifers are normally pregnant when distributed. The model assumes that all foundation heifers are three years old and heavily pregnant when they are distributed. Thus, the foundation heifers will give birth at the very start of their fourth year of life and, in the model, all calves are born at the beginning of the year. This simplifies the calculations, as one year after the foundation heifers are distributed all newborn calves will have reached exactly one year of age and the foundation heifers will be aged exactly four years. Heifer calves that are to be repaid must be returned to the project at one year of age.
The model is based on average data from three HIT schemes in the United Republic of Tanzania: Tanga Smallholders' Dairy Development Programme (TSDDP) (Rutamu et al., 1994); KaLiDeP (Houterman et al., 1993 and de Wolff, 1996); and Mbulu Livestock Development Programme (de Wolff, 1996). The two major control variables in the model are calving rate and mortality. The model assumes that the initial calving rate of foundation heifers will be 90 percent, and that this will drop to 57 percent in all subsequent years. While foundation heifers do not have their first calving until they are aged three years in the model, it is likely that some heifer calves born to the foundation heifers will have their first calving as early as 24 months of age. In the Initial Model calf mortality is assumed to be 11.4 percent during the first 12 months of life, while adult mortality is set at 6.7 percent. Cows can live a maximum of 11 years, after which they are culled. If mortality rates are high, most cows will have died before the age of culling.
To facilitate interpretation of results, the model is always run with 100 participants receiving one foundation heifer each. As the Initial Model is only concerned with the number of heifer calves born, 50 percent of the calves (i.e. all bull calves) are discarded from the model in the year in which they are born. Hence, the number of calves surviving their first year of life (and thus available for loan repayment) is equal to the number of calves born divided by two minus any mortalities.
The sustainability of HIT schemes can be viewed at two levels. A HIT scheme will only be considered sustainable at the project level if all loans are repaid. In the model this means that by the time all foundation heifers have died, the project must have received 100 heifer calves in repayment. In this sense, project sustainability is equated to the maintenance of the initial capital stock. A scheme will be considered sustainable at the participant level if all participants who have repaid their loan have also raised an additional heifer calf with which to replace the foundation heifer when it dies or is culled. The model is run for eight years at the end of which the foundation heifers have reached the close of their eleventh and final year of life. At the end of eight years there will be x number of participants who have repaid the required number of heifers. These participants will have a total of y cows free from debt. If y � x �1, sustainability has been achieved at the participant level. This figure will be referred to as the "average number of cows per participant free of debt".

Alternative models

Three alternative models are used to investigate the impacts of modifications to the implementation of HIT schemes. With the Insurance Model each foundation heifer will be replaced through an insurance scheme if it dies before giving birth to a heifer calf which survives the first year of its life and can be used to repay the loan. In effect, the insurance scheme eliminates mortality among foundation heifers (until they have repaid a heifer calf). In addition, if by the time the foundation heifer is culled at the end of year eight of the scheme a heifer calf has not been repaid, the insurance will pay to the HIT scheme the value of a heifer calf (i.e. clearing the participant's debt to the scheme).

Unlike the Initial and Insurance Models, Bull Calf Model 1 incorporates bull calves by assuming that each bull calf is worth as much as a heifer calf. Participants can therefore sell a bull calf and repay the loan in cash. Bull Calf Model 2 is a compromise between the Initial Model and Bull Calf Model 1 in that it assumes that a bull calf has half the value of a heifer calf. Hence participants would need to sell two bull calves to repay the loan.

RESULTS

Results for the Initial Model are given in Figure 2. During the life of the scheme only 79 heifer calves are repaid. At the end of year eight there are 123 cows free from debt and the average number of cows per participant free of debt is 1.6 (see Table). Therefore, the scheme depicted in the Initial Model would be considered unsustainable at the project level but sustainable at the participant level. However, it should be noted that over 20 percent of participants do not have a cow free from debt at the end of eight years.

2
Initial Model: repayment status when one heifer calf is repaid per foundation heifer received
Mod�le initial: situation au regard des remboursements, lorsqu'un seul veau femelle doit �tre rembours� pour chaque g�nisse de base re�ue
Modelo inicial: situaci�n cuando se devuelve una novilla por cada una de partida recibida

The number of heifer calves repaid is relatively insensitive to changes in calving rates or mortality (Figure 3). A 40 percent change in either of these variables does not result in a change in the number of heifer calves repaid of more than 12 percent. Even a 40 percent increase in the calving rate does not achieve sustainability at the project level. In the Initial Model the number of cows free from debt is most affected by the calving rate (excluding the year foundation heifers are distributed) (Figure 4). A 30 percent decrease in calving rate (to 0.4 calves per cow per year) results in the scheme becoming unsustainable at the level of the participant, as the average number of cows per participant free of debt drops below one. Sustainability at the participant level is less sensitive to changes in calf and adult mortality and the initial calving rate of foundation heifers.

3
Initial Model: sensitivity analysis for heifers repaid to project when one heifer calf is repaid per foundation heifer received
Mod�le initial: analyse de sensibilit� relative aux g�nisses rembours�es au projet, lorsqu'un seul veau femelle doit �tre rembours� pour chaque g�nisse de base re�ue
Modelo inicial: an�lisis de sensibilidad para las novillas devueltas al proyecto cuando se devuelve una novilla por cada una de partida recibida

4
Initial Model: sensitivity analysis for cows free from debt when one heifer calf is repaid per foundation heifer
Mod�le initial: analyse de sensibilit� relative aux vaches non grev�es de dettes, lorsqu'un seul veau femelle est rembours� pour chaque g�nisse de base re�ue
Modelo inicial: an�lisis de sensibilidad para las vacas libres de deuda cuando se devuelve una novilla por cada una de partida recibida

In marked contrast to results with the repayment of one heifer calf per foundation heifer, if repayment of two heifer calves is required, well over 100 heifers are repaid to the scheme. However, in this scheme the average number of cows per participant free of debt falls to 0.9, and 47 percent if participants are not able to repay their loan (see Table). Thus, by increasing the number of heifer calves to be repaid, the scheme becomes sustainable at the project level but unsustainable at the participant level. At the project level this result is relatively insensitive to changes in mortality or calving rate, as a 40 percent decrease in the calving rate (excluding the year foundation heifers are distributed) to 0.34 calves per cow per year still results in over 100 heifer calves being repaid. However, to achieve sustainability at the participant level, only a 1 percent increase in the calving rate is needed. Alternatively, sustainability could be achieved at the participant level by an increase of 5 percent in the initial calving rate of foundation heifers, or by a decrease of 5 percent in calf or adult mortality.

Sustainability of the different heifer models
Viabilit� des diff�rents mod�les
Sostenibilidad de los distintos modelos de novillas

By insuring the foundation heifer, project sustainability is ensured because exactly 100 heifers are repaid to the scheme and all participants are able to repay their loan.The scheme also remains sustainable at the participant level as the average number of cows per participant free of debt is greater than one. The insurance premium that each participant would have to pay before receiving a foundation heifer can be calculated as follows. For all years of insurance, any foundation heifers that die before repaying one heifer calf will be replaced. Also, at the end of the eighth year of the scheme, the outstanding loans of any foundation heifers are paid off. Based on the Initial Model, for 100 participants receiving one foundation heifer each, 29 will need to claim insurance. Thus, to cover the direct cost of insurance, each participant will need to pay a premium of at least 29 percent of the value of the foundation heifer loaned to him or her.
Bull Calf Model 1 assumes that bull calves can be exchanged for heifer calves with the result that, at the end of the scheme, the number of cows free from debt is four times greater than in the Initial Model. This is because the extra heifer calves born or bought in exchange for the bull calves can also give birth, and all their offspring are taken into account. As a result the average number of cows per participant free of debt increases to 5.7. In addition to the scheme being sustainable for participants, only 8 percent of participants fail to repay the required number of heifers. However, while the number of heifer calves repaid increases to 92, this is still less than the 100 needed for project sustainability. When Bull Calf Model 1 is run with the requirement that two calves are repaid, the scheme is sustainable at both the project (172 calves returned) and participant levels (80 percent of participants who successfully repaid their loan have on average 3.5 cows).
It is somewhat doubtful that one bull calf would be worth as much as a heifer calf; therefore, in Bull Calf Model 2, the value of a bull calf is considered to be half that of a heifer calf. Compared with the Initial Model, the number of cows free from debt is more than doubled, but sustainability at the project level is still not achieved. This model was run again requiring the repayment of two calves. As in the Initial Model and Bull Calf Model 1, the scheme becomes sustainable at both the project and participant levels.

DISCUSSION AND CONCLUSIONS

Clearly, the assessment of the sustainability of alternative HIT schemes will depend on estimated calving and mortality rates. There are a number of reasons to believe that, at least in the initial years, HIT schemes will experience relatively low calving rates and high mortality. Not the least important of these factors will be the inexperience of the participants with these relatively high-management animals, and problems arising from the transfer of pregnant heifers.

The results from the Initial Model showed that the scheme is unsustainable at the project level when one heifer is repaid, simply because some heifers will die before giving birth to a heifer calf. It has also been demonstrated that, by requiring the repayment of two heifer calves, the scheme is sustainable at the project level but not quite so at the participant level. The addition of an insurance scheme ensures both project and participant sustainability.
For HIT scheme organizers, insurance should be an attractive option for ensuring the sustainability at the project level. However, for such a scheme to work, participants would need to be willing to pay the insurance premium before receiving the heifer. In the example given earlier, the premium would be in the range of 29 percent of the value of a heifer, which would appear to be a prohibitively high sum if, indeed, HIT schemes are to benefit poor farmers.
When the bull calves were given an equal value to heifer calves in Bull Calf Model 1, the sustainability of the HIT scheme was ensured at the project level if two calves were repaid per participant. The sustainability of the scheme at the participant level was also greatest in this model. The only failing of this model was that, unlike the Insurance Model, not all participants repaid the loan fully. When a bull calf was worth half as much as a heifer calf, as in Bull Calf Model 2, the sustainability of the HIT scheme was also ensured at both participant and project levels if two calves were repaid per participant. However, even fewer participants fully repaid the loan.
Without an insurance scheme the importance of participants being able to utilize bull calves is apparent. If a bull calf is worth half as much as a heifer calf, and the participant needs additional finance, this could then come from some of the proceeds of the dairy production. However, if participants have to use up the income generated by the foundation heifers just to ensure they have a replacement heifer, the whole purpose of having the heifer to improve income is lost. The only benefit would be free manure, and possibly draught power when the animal is not heavily pregnant.
A potentially interesting modification would be if participants were to be loaned two foundation heifers and asked to repay three calves. This would reduce the need for an insurance scheme since, if one heifer died, the participant would still have one heifer remaining. It is also more likely that sustainability would be achieved at the project level, again reducing the need for an insurance scheme. However, it may be even more unlikely that poor farmers would be able to maintain and manage successfully two dairy animals.
Currently, the model does not differentiate between the quality and size of young born. The TSDDP Management Team (1995) reported that 20 percent of female HIT offspring are stunted, resulting in increased age at first calving, decreased adult body weight and lower milk production potential, all of which will have a negative impact on project and participant sustainability.
The models show that standard HIT schemes are very unlikely to be sustainable at both project and participant levels. If schemes require the repayment of just one heifer, they will quickly run down the initial capital represented by the 100 foundation heifers. If schemes require the repayment of two heifers, a significant number of participants will not be able to repay their loan and others will have no replacement heifer by the time the foundation heifer dies.
If HIT schemes are to become effective in benefiting the rural poor they will need to implement an insurance scheme, or allow repayment in cash which will let farmers fully realize the value of bull calves, as well as making significant improvements in calving rate and mortality. If a scheme is to be run without insurance, more than one heifer calf must be repaid for each foundation heifer received. Unfortunately, in the absence of an insurance scheme, not all participants will repay the loan and have a heifer free of debt.
It is important to note that this article has dealt only with the impact of calving rate and mortality on the sustainability of HIT schemes. However, there are a number of other factors not directly associated with poor reproductive performance or mortality which affect their success, such as loan defaulting. While peer pressure may be one way to address this problem, it is seldom altogether successful. Another issue, particularly in more isolated rural areas, is the limited size of the milk market, and the fact that only a certain number of productive animals can saturate this local demand. It is because of this that some HIT schemes eventually become directly involved in milk collection, processing and marketing. Finally, because of variability in the productivity, growth and reproductivity of foundation heifers and their offspring, returns to participants can vary significantly, which will in turn affect motivation and investments in management.
To summarize, while some steps can be taken to ensure that HIT schemes are more sustainable, these steps cannot address the more fundamental question associated with such schemes. Is it possible or desirable to insist on sustainability when attempting to distribute high-value, high-management animals to farmers who have neither the experience nor the resources to guarantee their survival and reproduction? In this light, perhaps it is time to re-examine Heifer-in-Trust schemes as a poverty alleviation strategy.

Bibliography

Chamberlain, A. 1989. Milk production in the tropics. Harlow, UK, Longman. 242 pp.
de Waal, A. 1990. A re-assessment of entitlement theory in the light of the recent famines in Africa. Dev. Change, 21: 469-490.
de Wolff, J.T. 1996. Some reflections on HIT. In W. van Weperen, ed. Proceedings of the Dairy Development Conference held at Karibuni Centre Mbeya on 16 and 17 May 1996, p. 83-93. Iringa, United Republic of Tanzania, CTA-SHDDP.
Dreze, J. & Sen, A. 1989. Hunger and public action. Oxford, UK, Clarendon Press. 373 pp.
Havers, M. 1996. Financial sustainability in savings and credit programmes. Development in Practice, 6(2): 144-150.
Houterman, J.F., Lutalala, E.M., de Wolff, J.T. & Dummer, R.M. 1993. Quantitative aspects of Heifer-in-Trust credit schemes: the Kagera experience. TSAP Proceedings, 20: 168-188.
Johnson, S. & Rogaly, B. 1997. Microfinance and poverty reduction. Oxford, UK, Oxfam. 134 pp.
Kinsey, E. 1994. Aspects of credit in dairy development: the HPI Tanzania experience. TSAP Proceedings, 21: 58-68.
Kinsey, E. 1996. Heifer Project International's twelve cornerstones of just and sustainable project development in the Tanzania context. In W. van Weperen, ed. Proceedings of the Dairy Development Conference held at Karibuni Centre Mbeya on 16 and 17 May 1996, p. 76-82. Iringa, United Republic of Tanzania, CTA-SHDDP.
Matthewman, R.W. 1993. Dairying. London, UK, Macmillan. 152 pp.
Pelant, R.K. 1992. Chain of life. Little Rock, Arkansas, USA, HPI. 39 pp.
Rutamu, I.B. 1996. The features and future of Heifer-in-Trust credit schemes in Tanzania. In W. van Weperen, ed. Proceedings of the Dairy Development Conference held at Karibuni Centre Mbeya on 16 and 17 May 1996, p. 94-100. Iringa, United Republic of Tanzania, CTA-SHDDP.
Rutamu, I.B., Minja, F.N., v. Munster, B.H.M., Schoonman, L., Swai, E. & Zylstra, L. 1994. The sustainability of Heifer-in-Trust credit schemes: the Tanga experience. TSAP Proceedings, 21: 96-113.
Sen, A. 1981. Poverty and famines: an essay on entitlement and deprivation. Oxford, UK, Clarendon Press. 257 pp.
Swift, J. 1989. Why are rural people vulnerable to famine? IDS Bulletin, 20(2): 8-15.
TSDDP Management Team. 1995. Tanga Smallholders' Dairy Development Programme 1995 progress report. Tanga, United Republic of Tanzania, TSDDP. 16 pp.