Repatriation Under Conflict: A Review of the Encashment Programme for Afghan Refugees in Pakistan

EVAL/ENC/14

SCOPE OF THE REVIEW

This review of the repatriation grant (encashment) initiative in Pakistan examines the effectiveness of the programme in facilitating voluntary repatriation to Afghanistan. The evaluation analyses the programme's strengths and weaknesses, identifying the operational and contextual limits to its impact, as well as the key conditions for its replication in other repatriation operations. In conducting the review, the evaluation team undertook a mission to Pakistan to observe the programme at first hand and to discuss its implementation and impact with personnel from UNHCR and other international, governmental and non-governmental organizations. The team also reviewed the relevant documentation at Headquarters and interviewed key staff members in Geneva who have been involved in the development of the programme.

OVERVIEW

RETURN TO CONFLICT-AFFECTED COUNTRIES IS LIKELY TO INCREASE

(1) There has been an evolution in UNHCR's thinking on repatriation in recent years. It is now generally accepted that if refugees wish to return to their country of origin, the organization has a responsibility to assist them in doing so, even if conditions there do not appear favourable.

(2) At the same time, UNHCR has started to reconsider the artificial distinction that has traditionally been drawn between spontaneous return, in which UNHCR generally played little or no role, and organized repatriation, where the timing, pace and mode of return are commonly decided upon by UNHCR in response to the demands of the international community, leaving the refugee with minimal options.

(3) Prevailing conditions in Afghanistan provide a striking example of the opportunities and constraints confronting UNHCR in facilitating spontaneous repatriation to a conflict-affected country. The latest outbreak of fighting between rival factions in Kabul has caused the renewed displacement of hundreds of thousands of civilians and prompted the closure of the border by the government of Pakistan to prevent a further refugee influx. During the same period, more than 4,000 Afghans were monitored crossing the border from Pakistan into Afghanistan, returning voluntarily and spontaneously to their homes.

ENCASHMENT HAS FACILITATED SPONTANEOUS RETURNS

(4) Where government policy and market conditions permit, encashment can serve as an effective means of supporting refugee choices and facilitating spontaneous return to conflict-affected countries. Since its establishment in mid-1990, the repatriation grant (encashment) programme in Pakistan has facilitated the voluntary repatriation of between a third and half of the three million Afghan refugees in Pakistan.

(5) In addition to facilitating the voluntary return of large numbers of refugees, encashment has a number of important operational advantages which render it worthy of consideration for other repatriation operations:

• it is extremely cost-effective, with overheads at their lowest at times of peak returns;

• with minimum staffing requirements in the implementation phase, encashment provides one of the most direct means available for transfer of donor assistance to beneficiaries;

• providing refugees with a cash grant permits individual decision-making regarding durable solutions, whilst facilitating deregistration; and

• encashment comprises a number of components which can be readily adapted for use in other refugee settings.

THE SYSTEM PROVIDES A RANGE OF CHOICES

(6) While many of the characteristics of encashment are common to other repatriation systems, the range of choices offered to the refugee is exceptional. By providing a cash grant in exchange for a ration passbook at the time of the refugee's choice, encashment permits the refugee to choose not only whether, when and where to return, but also how to spend the cash.

(7) The encashment system in Pakistan also differs from more conventional repatriation systems in that it was never expected to benefit the entire caseload. It was intended, instead, to act as a catalyst to stimulate broader interest in repatriation to Afghanistan, building upon existing low-key, spontaneous returns.

(8) From the outset, it was acknowledged that there would never be sufficient funding to encash all 600,000 revalidated passbooks, and that in any case, a significant proportion of the caseload was unlikely to return for a variety of reasons. As such, the encashment programme may be seen as a gamble, one which paid off exceptionally well when conditions in Afghanistan were favourable in mid-1992.

ENCASHMENT DOES NOT OFFER A SOLUTION TO ALL REFUGEES

(9) Inevitably encashment has certain drawbacks, not the least of which is the fact that it can only be expected to benefit a part of the caseload. Whilst it is an attractively simple system, it has to be seen as a partial one, which only addresses the immediate needs of those who are ready and able to return. It does not provide the individual attention required to facilitate the return of vulnerable individuals, for example, nor of those with specific political, ethnic or security concerns.

(10) In a repatriation programme developed for such a massive refugee population, there is an understandable tendency to postpone the time-consuming work on more difficult cases until the population has been significantly reduced. In doing so, however, there is a considerable risk that some of the potential solutions or support networks required to provide a viable framework for return may no longer be available.

(11) One of the more negative consequences of the encashment system has been deregistration without return. Premature encashment, occurring either because of poverty or as a result of the upsurge of interest in return following the fall of the Najibullah government in April 1992, has led to considerable hardship. Once deregistered, refugees who do not feel ready or able to return to Afghanistan find themselves in the same disadvantageous situation as many of the unregistered refugees in Pakistan, commonly drifting towards major urban centres.

(12) To date, the Pakistani authorities have been tolerant of clandestine local settlement by Afghan refugees. This may not always be the case, however, and host government objectives must be taken into account when considering the potential negative consequences of an encashment programme.

MANY LESSONS LEARNED ARE APPLICABLE TO OTHER OPERATIONS

(13) Experience in Pakistan has once again highlighted a number of key operational and organizational weaknesses already identified in other repatriation operations. For example, major weaknesses in the registration system have had serious consequences, not only for ongoing assistance activities, but also for the planning and implementation of the repatriation operation.

(14) There is a growing recognition of the need to establish solutions-oriented and accurate registration systems at the beginning of an emergency programme. A significant step forward in addressing this perennial problem was made recently in the Burundi emergency operation where registration has been oriented towards eventual repatriation.

LINKAGES BETWEEN OPERATIONAL COMPONENTS REQUIRE GREATER ATTENTION

(15) An examination of the Afghan repatriation operation also points to the need to reinforce operational and organizational linkages at various levels. For a repatriation operation to be successful - regardless of the mechanism of return - UNHCR must begin to place much greater emphasis on the linkage between its repatriation activities in the country of asylum and its reconstruction and rehabilitation efforts in the country of origin. The traditional split between country of asylum and country of origin will continue to undermine operational efforts on both sides of the border until this linkage is achieved by bringing the two sides together into a single operation.

(16) Problems of coordination between the country of asylum and the country of origin were addressed to some extent in the Cambodia repatriation operation by the appointment of a special envoy. Unfortunately, in Pakistan, placement difficulties hindered the effectiveness of the High Commissioner's envoy. Attention needs to be paid in future repatriation operations to the placement of a special envoy in a way which most effectively brings together operations in the countries concerned. In addition, serious consideration should be given to the appointment of staff to an operation rather than to a particular country, giving staff in key functions a cross-border or regional responsibility.

(17) Within the country of asylum, much greater attention also needs to be paid to forging closer links between ongoing assistance and repatriation activities. One of the significant weaknesses of the Pakistan programme has been a lack of clarity, and in some cases directly conflicting objectives pursued by different components of the operation. When UNHCR is engaged in a facilitated repatriation operation, particular attention must be paid to the management of care and maintenance activities to ensure that refugees are not prematurely pressured into leaving the country of asylum.

(18) UNHCR should also attempt to ensure that major rehabilitation efforts carried out in the country of origin, in conjunction with other international and non-governmental organizations, are more effectively coordinated. Establishment of an inter-agency strategy for rehabilitation, and a careful definition of roles and responsibilities at an early stage, could have avoided many of the problems which have hampered reconstruction efforts in Afghanistan.

(19) The Pakistan programme has demonstrated the potential for facilitating spontaneous return of refugees to a conflict-affected country. It has also highlighted the benefits of providing cash for assisting the repatriation of a large caseload. The limitations and negative consequences of the programme have, however, once again exposed the need for UNHCR to confront and seek solutions to a number of important operational and organizational weaknesses which are, all too often, common features of its repatriation operations.

ORIGINS AND IMPLEMENTATION OF THE REPATRIATION GRANT PROGRAMME

(20) The recent outflow of refugees from Kabul and the subsequent closure of the border by the government of Pakistan highlight the risks and complexities of a programme which endeavours to assist spontaneous return to a conflict-affected country. Nevertheless, as refugees are increasingly faced with a choice between unsatisfactory conditions in the country of asylum and in the country of origin, UNHCR will be confronted with such challenges with growing frequency.

(21) The massive influx of refugees from Afghanistan into Pakistan began shortly after the Saur Revolution of April 1978, but it was not until late 1979, following the Soviet invasion, that the government of Pakistan requested the involvement of the international community. For the next three years, UNHCR's activities focused on providing emergency relief to the refugees, who, by mid-1981, were already estimated to number more than two million. By 1983, however, when the outflow had slowed down and the refugee population had started to stabilize, UNHCR's efforts were shifted away from care and maintenance towards self-sufficiency and income-generation.

UN PREPARED FOR REPATRIATION AFTER 1988 PEACE ACCORDS

(22) The Geneva Accords of April 1988 foresaw the end of a decade of war which had brought widespread devastation to the country. In preparation for an anticipated massive repatriation following the peace settlement, a Bilateral Agreement was signed between Afghanistan and Pakistan, followed by agreements between the UN and the two countries concerned, reaffirming the voluntary nature of the return and defining the general characteristics of the assistance that would be provided to returnees.

(23) In May 1988, the UN Secretary-General created the Office of the Coordinator for United Nations Humanitarian and Economic Assistance Programmes relating to Afghanistan (UNOCA), whose mission was to coordinate UN agencies and NGOs and to devise a comprehensive strategy for relief and rehabilitation. UNHCR established a presence in Kabul in the same month. The withdrawal of Soviet troops, which had begun shortly after the Geneva Accords, was completed by February 1989.

(24) At that stage, there was a widely-held expectation of mass repatriation. To ensure that such a massive movement of people would proceed in an orderly and sustainable manner, international, governmental and non-governmental organizations rallied their forces to address the vast range of needs in a coordinated manner. UNOCA launched its Operation Salaam, which was to provide country-wide rehabilitation assistance, with particular emphasis on the 14 provinces from which a third or more of the population had become refugees. Since some parts of the country were still affected by combat, it was envisaged that the work would proceed initially in so-called "zones of tranquility."

EXPECTED MASS RETURNS DID NOT OCCUR

(25) Despite this substantial mobilization of resources, the outcome proved to be anticlimactic as the anticipated mass returns failed to materialize. Instead of the Soviet withdrawal leaving behind a peaceful environment conducive to return, it left a well-armed and war-seasoned population, rife with traditional enmities based on ethnic, religious and political divisions. The pattern of internal conflict which followed has varied both in intensity and in geographical coverage over the intervening five years. It continues, however, on a scale which prevents the population of Afghanistan from finding a long-awaited peace.

REPATRIATION PLANNING HAD TO TAKE ACCOUNT OF FLAWED REGISTRATION

(26) Initial efforts towards developing a repatriation plan, in anticipation of mass returns following the final withdrawal of Soviet troops, had to take account of long-established weaknesses in the refugee registration system. The registration system had been established by the government of Pakistan at the beginning of the outflow from Afghanistan.

(27) The scale of the outflow had rapidly overcome the government's capacity to effectively register the population and a decision was made at an early stage to enlist the help of traditional community leaders (maliks) in the registration and the distribution of assistance. Although this system was subsequently abandoned, it opened the way to massive and uncontrollable abuse of the system whereby ration passbooks were openly traded and individual leaders gained enormous wealth through the proliferation of bogus passbooks. Abuse of the registration system has had serious consequences for the subsequent management of on-going assistance as well as for implementation of the repatriation programme.

FOOD SHORTAGES IN AFGHANISTAN HAD TO BE ADDRESSED

(28) In late 1988, WFP began to review its policy options for commodity distribution in Afghanistan on a sufficiently large scale to sustain the anticipated sudden increase in population size following mass repatriation. It was clear that the situation in Afghanistan presented a major challenge in this respect since none of the traditional basic requirements for food distribution were present. In the absence of a strong central government (following the Soviet withdrawal), and given a widely devastated road system and generally poor security conditions, the logistical support needed for large-scale food distribution was not available. Adequate manpower to provide the necessary monitoring of distribution and sufficient storage capacity to permit the required prepositioning and stockpiling were also lacking.

(29) In the face of such overwhelming constraints to a traditional approach to commodity distribution, the WFP Director of Operations devised an alternative model, dubbed "Project Market Place". The proposed system would involve the food allocation for the refugee population being sold to the government of Pakistan for re-sale to commercial traders on the border. By monetizing produce in Pakistan, the necessary quantities of wheat to sustain large numbers of returnees would find its way into Afghanistan along indigenous commercial channels. The profits made in the re-sale of the wheat would meanwhile be used to provide a fund for buying back refugees' ration books for their current market value (approximately US$ 100 per card at the time). This, in turn, would provide refugee families with sufficient cash to purchase wheat for the first 6-12 months after returning home.

EFFECTIVE DEREGISTRATION ALSO HAD TO BE ENSURED

(30) In developing an alternative model for food distribution, WFP was also concerned to find an effective system which would remove ration books from the system as people returned. Failure to do this would mean that ration cards would remain in circulation in Pakistan after their "owners" had repatriated with the result that WFP would be unable to make the appropriate reductions to the care and maintenance operation. In view of the large number of bogus passbooks known to be in circulation, as well as passbooks no longer in the hands of their original 'owner', deregistration was particularly important.

(31) In a separate initiative launched during the same period, UNHCR was investigating the possibilities for establishing a reliable system of deregistration in the likelihood of massive spontaneous repatriation. In addition, it was hoped that the existing registration system could be utilized to monitor movement out of the camps despite its weaknesses.

(32) A consultancy was established towards the end of 1988 to look into the available options. The study identified a number of serious constraints to the development of a strategy which would satisfy the major concerns of the international community. It was clear that re-registration of the refugee population was not feasible, given the size of the population and the vested interests of heavily-armed beneficiaries, and thus there was no way to include the unregistered in the proposed strategy.

(33) At the same time, it was evidently impossible to conduct any reliable forecasting of population flows in view of the unpredictability of the situation inside Afghanistan and the complexity of push and pull factors influencing the many different sub-groups in the refugee population. Furthermore, it was recognized that there was a very high probability of abuse, whatever system of deregistration was ultimately developed.

PMP PROVIDED CRUCIAL ELEMENTS OF A WORKABLE PLAN

(34) The consultants concluded that the Project Market Place concept provided a number of elements which could usefully be explored further with a view to developing a workable plan. The key elements identified at that stage were that existing passbooks should be validated and that the validated passbooks should ultimately be exchanged for cash. In the ensuing discussions a number of critical issues were debated at length. These included, for example:

• what should be the value of the package?

• should payment be made per passbook or by family size?

• should payment be made to family heads or actual passbook holders?

• should cash be used or vouchers? if cash, which currency?

• should the revalidation exercise be conducted separately from payments or at the same time?

• where should payments be made? in camps? in district centres? at the border? inside Afghanistan?

• who should make the payments? WFP? UNHCR? Government of Pakistan? NGOs?

(35) At the same time, concern was expressed over the many uncertainties and risks associated with the proposed system. For example, what would be the magnitude of the demand? Would UNHCR/WFP be able to keep pace with the demand if it was high? What would be the reaction of the unregistered refugees and how would new arrivals be dealt with once the system was operational? And perhaps most importantly, where would refugees go after receiving their payments?

(36) In conclusion, and in the absence of immediate answers to such questions, the consultants recognized that the situation presented very limited options and that complete risk-avoidance was not possible. They identified two critical areas where careful advance planning could serve to minimize a number of these risks, namely the production of a forgery-resistant document for use in the revalidation exercise, and the development of a security plan for cash disbursement. They also recognized that Project Market Place should be considered as one part of an integrated approach to repatriation, one of the most significant elements of which would be extensive reconstruction in Afghanistan.

REHABILITATION IS THE KEY TO AN INTEGRATED APPROACH TO REPATRIATION

(37) UNHCR had begun its cross-border rehabilitation efforts in Afghanistan, within non-government- controlled areas, before the end of 1988. During the first two years, more than $US 8 million were spent on some 88 sub-projects, most of which were intended to enhance food production through the repair of irrigation systems and the provision of agricultural inputs. Other projects included road repair, warehouse and hospital construction and local production of building materials, reflecting the priority needs in potential returnee areas.

(38) Prevailing conditions in Afghanistan required the development of new working methods. Unpredictable but widespread security problems hindered normal procedures such as on-site needs assessment and monitoring. UNHCR was obliged to rely predominantly on the skills and integrity of its international and local implementing partners. In an effort to develop a coherent response to the needs of returnees in such a generally chaotic situation, UNHCR commissioned a series of provincial profiles to provide detailed and specific information on the conditions affecting repatriation in the major returnee-impacted provinces.

(39) During this period, UNHCR also played a leading role among UN agencies involved in cross-border work. A long-established and substantial presence in Pakistan, familiarity with the Afghan refugee community, and relatively flexible project approval procedures permitted UNHCR to move quickly. Initially, UNHCR-sponsored cross-border projects were considered as remarkably successful given the tremendous constraints to implementation. Whatever their limitations, it was felt that such activities undoubtedly contributed to enhancing conditions for return.

(40) Ongoing conflict, however, frequently led to the temporary suspension of projects and occasionally their abandonment before, or destruction after, completion. Furthermore, with returns spasmodic and difficult to monitor, it was extremely difficult to identify a direct causal link between rehabilitation work and repatriation. By 1990, as other UN agencies, notably UNDP and FAO, were planning greatly increased involvement in the rehabilitation of Afghanistan, UNHCR began to seriously reconsider the appropriateness and relevance of its involvement in cross-border activities.

HIGH COMMISSIONER ADVOCATES WINDING DOWN OF ASSISTANCE PROGRAMME

(41) Throughout this period, there had been an ongoing debate regarding the need to wind down the assistance operation in Pakistan. Donors, disappointed that the Soviet withdrawal had not led to immediate mass returns, were eager to see a change in the existing care and maintenance stalemate. In March 1990, during a joint visit to Pakistan with the Executive Director of WFP and the Special Representative to the Secretary-General for Afghanistan and Pakistan, High Commissioner Stoltenberg made a surprise announcement indicating his plan for a three-year countdown to termination of UNHCR's assistance programme in the context of a political solution.

(42) The announcement fuelled the existing controversy. Donors were generally supportive of the proposal, depending on their political or economic interests. Some maintained that repatriation would, in itself, facilitate the peace process and should therefore be encouraged. The views of UNHCR staff in Pakistan varied greatly, ranging from full support to scepticism and, in some cases, strong opposition. There was particular concern at field level over mounting pressure to "produce" repatriation to a country in continuing conflict.

(43) Nevertheless, following the High Commissioner's announcement, it was clear that the time had come for UNHCR to launch its long-debated voluntary repatriation programme. After closer examination, Project Market Place was rejected by all concerned as being too expensive as well as logistically complex and unwieldy. Efforts then focused on identifying an alternative mechanism, which would at once facilitate return and ensure effective deregistration.

REVALIDATION OF 600,000 PASSBOOKS COMPLETED IN THREE MONTHS

(44) An essential first step, however, was to 'fix' the population, regardless of the repatriation mechanism finally adopted. The revalidation process was developed during a series of field-level brainstorming meetings in early 1990 and was modified to suit the particular circumstances of Baluchistan and NWFP. Forgery-resistant stickers, each bearing a discrete number (indicating the refugee village and district in Pakistan, plus a running serial number - see Annex 1) were then printed. In view of the considerable potential cash value of the stickers, a database of sticker numbers was established by OCM Islamabad, for monitoring purposes. This was subsequently used as a database for revalidated passbooks.

(45) During an intensive three-month period in mid-1990, UNHCR and the Commissionerate for Afghan Refugees (CAR) staff throughout Pakistan were involved in the revalidation process, affixing stickers to some 600,000 ration books representing more than 3 million registered refugees. Lists of revalidated passbooks were forwarded to Islamabad for data entry. After more than ten years of manual registration, the Afghan refugee population in Pakistan was finally recorded electronically.

ENCASHMENT PROGRAMME LAUNCHED USING SIMPLE CASH/WHEAT FORMULA

(46) In late July 1990, the Repatriation Pilot Project was launched, initially on a three-month trial basis. The project has continued, with minimal modifications, to the present day. Refugees presenting revalidated passbooks would receive a cash grant and a supply of wheat in exchange for their sticker. Agreement had been reached with the National Bank of Pakistan (NBP) to provide banking facilities for disbursement of the cash grant while CAR and WFP were to provide storage and distribution of the wheat encashment component.

(47) The grant level had finally been fixed at 3,300 rupees (approximately US$ 100) to cover average travel costs1, plus 300 kilogrammes of wheat, enough to feed an average family for the first few months following return. In view of the unreliability of available information on family size, it was agreed that the repatriation grant would be given at a fixed level to all passbook holders. Finally, it was decided to give cash in rupees since much of the expenditure was expected to take place in Pakistan before departure. This would allow refugees to arrange their own transport and determine their own way - and time - of going home. For refugees with long distances to travel, a subsidized transportation programme was established. This has been operated by IOM since mid-1992.

(48) Distribution of the 300 kilogramme wheat allowance for returnees under the encashment programme was carried out through WFP distribution centres established in collaboration with CAR near to the encashment centres. This arrangement permitted maximum flexibility, since distribution centres could be opened and closed in response to changing demands. WFP's role in encashment was primarily logistical and until early 1994, monitoring was limited to checking gross distribution of wheat and did not involve cross checking the encashment of wheat with that of the cash grant.

ENCASHMENT PROCESS COMPRISES SERIES OF STRAIGHTFORWARD STEPS

(49) The encashment process itself is remarkably straightforward, and has changed little since the beginning of the programme. It involves a series of carefully monitored steps:

• the refugee travels to the encashment centre nearest to his2 village (generally located within NBP premises) and hands in his revalidated passbook;

• the encashment assistant checks the sticker for forgery, notes down the refugee village code from the sticker, the family size and records the province/district to which the family is going; the numbered part of the sticker is then removed from the passbook thereby cancelling eligibility for continued assistance;

• the numbered half of the sticker is attached to the completed encashment data sheet (see Annex 2) and given to the refugee to take to the NBP cashier; the sticker with its attached data sheet, accepted by NBP as a negotiable instrument, is then exchanged for the specified amount of cash;

• having decided upon a departure date, the refugee negotiates with a trucker to provide transportation to his destination in Afghanistan; if he is eligible for subsidized transportation, he will advise IOM of his plans;

• once the refugee is ready to leave Pakistan, he takes his passbook and exchanges the remaining unnumbered half sticker for 300 kilogrammes of wheat at the most conveniently located WFP depot;

• the cancelled passbook is retained by the refugee to use as an identity document.

(50) Encashment figures are derived from reports submitted by NBP to OCM Islamabad, generally with a 7 day time-lag. The database is updated and reports generated for WFP to permit them to make the necessary reductions in their regular food distribution. The time-lag involved inevitably results in some financial losses but it has proved difficult to reduce this further.

(51) Data from the encashed stickers also provides important information for analysis of return flows, both in terms of anticipated or actual departures from refugee villages in Pakistan as well as anticipated or actual returns to specific destinations in Afghanistan. Analysis of data on destinations of prospective returnees provides valuable information for the planning and development of rehabilitation activities in Afghanistan.

ACHIEVEMENTS AND LIMITATIONS OF ENCASHMENT

(52) The impact of the encashment programme in Pakistan has been significant from a number of perspectives. It is important to grasp the sheer scale of the population involved; during its three and a half years of operation, more than a quarter of a million passbooks have been encashed, representing well over one and a half million individuals (see Table on following page). In the peak year of returns, 1992, more than 900,000 individuals were observed crossing the border into Afghanistan, the overwhelming majority of them during a six month period from May to October.

1992 RETURNS TO AFGHANISTAN DWARF OTHER REPATRIATION OPERATIONS

(53) Such figures dwarf the numbers involved in other major organized repatriation operations, such as Cambodia (362,000 returnees over twelve months), Nicaragua (73,000 returnees over two years) and Namibia (42,000 over six months). In addition, when considering such comparisons of scale, it should be recalled that in contrast to more costly, logistics-heavy operations, repatriation to Afghanistan has been entirely refugee-led and spontaneous, involving minimal inputs by the international community and providing the most direct means so far devised of transmitting donor assistance to beneficiaries.

(54) Furthermore, with the exception of a brief, and unfortunately abortive, experience in Somalia in 1990, the encashment programme as implemented in Pakistan was quite new. Considering its innovative nature, the programme has experienced surprisingly few operational difficulties. Even during the period of encashment frenzy in mid-1992, UNHCR staff and their partners were able, with considerable ingenuity, to find prompt solutions to the various difficulties which did arise.

(55) For example, the short-lived surge in demand stretched the system's processing capacity to its limits. Maintaining an adequate cash flow from New York to Islamabad and on to the individual bank branches was a major challenge and required NBP to shuffle funds between branches in order to meet the demand. Pressure on bank branches with limited working hours led to some security problems which, on a number of occasions, required the intervention of the police.

BORDER MONITORING CONFIRMED LARGE SCALE RETURNS IN 1992

(56) Since the early days of the encashment programme, concern had been growing over the difficulty in linking up encashment figures with actual physical return to Afghanistan. In response to this concern, systematic border monitoring was introduced at the beginning of 1992. The monitoring, conducted by UNHCR staff in conjunction with government of Pakistan border guards, focuses on principal crossing points. Virtually all returning families travel by road, in view of the large quantities of goods taken back. The system is therefore considered successful in tracking the bulk of returnees and in providing a sufficiently accurate picture of returnee trends to satisfy both planning needs and the information needs of donors.

(57) Nevertheless, since some crossings are inevitably missed by the monitoring process, the figures collected somewhat underestimate the total numbers returning. Furthermore, monitoring focuses exclusively on family group returns and does not record individual movements. Given the tendency for staged returns, whereby male family members return temporarily to rebuild their homes and make other essential preparations before the whole family repatriates, this system has been found to be the most reliable way to avoid duplication of recorded crossings.

(58) The current system does not permit a direct correlation to be made between border crossing and encashment figures since sticker numbers are not recorded at the border. Indeed, there are a number of justifiable reservations about the validity of both sets of figures which will be explored elsewhere in this report. Nevertheless, the fact that more than 900,000 people were observed crossing into Afghanistan during the course of just one year, with all the outward signs of families repatriating, should not be underestimated.

(59) Prevailing conditions in Afghanistan prevent systematic monitoring of returnees or an evaluation of the impact of large-scale returns on the local economy. There is, however, some evidence to suggest a positive linkage. For example, a senior official of a UN agency working inside Afghanistan noted a recent marked improvement in the local economy of Khandahar, as evidenced by the fact that the daily wage was currently higher than in neighbouring Quetta. In his view, this improvement could to a large extent be attributed to the flow of encashment money across the border impacting directly on local rehabilitation efforts.

SUCCESS OF ENCASHMENT ATTRIBUTED TO BOTTOM-UP APPROACH

(60) The success of the encashment programme in Pakistan can be attributed to a number of key operational factors and a system which combined cost-effective use of the private sector and low administration costs, with maximum delegation to the field and minimal deployment of staff.

(61) Recognizing the fact that the sheer size of the refugee population as well as continuing insecurity in Afghanistan precluded the use of more traditional repatriation strategies, UNHCR sought to develop a system which would be well adapted to the particularities of the local situation. To a large extent this was achieved by effective delegation and the involvement of both Pakistani and Afghan local staff during the planning stages. Sufficient flexibility was given to permit the basic system to be modified to suit the specific circumstances in Baluchistan and in NWFP.

ENCASHMENT IS WELL GEARED TO OPERATIONAL REALITIES

(62) By using a bottom-up approach to planning its repatriation grant programme, and thereby incorporating field and local staff members' experience and understanding of the local situation, UNHCR was able to effectively exploit a number of important local assets. The availability of a reliable banking system and plentiful commercial transport, for example, were ingeniously incorporated into the system, while recognition of the entrepreneurial spirit of the refugees ensured their positive response to encashment principles.

UNHCR ANTICIPATED TIMELY SHIFT IN GOVERNMENT OF PAKISTAN POSITION

(63) A good understanding of the local situation also permitted UNHCR to effectively exploit the interests of other key players, notably the Pakistan Government and the donors, in developing the repatriation grant programme. For economic and political reasons, the government had long been reluctant to consider any large-scale phasing out of the assistance programme. UNHCR staff in Pakistan correctly anticipated a shift in this position in the wake of significant geopolitical changes and effectively exploited this shift by ensuring that the encashment programme would bring substantial economic gains to the country of asylum. By deciding to distribute the repatriation grant in Pakistan and in rupees, UNHCR ensured the approval and the full cooperation of the government.

ENCASHMENT BRINGS DIRECT BENEFITS TO PAKISTAN

(64) By providing the cash grant in rupees through local banks in NWFP and Baluchistan, the encashment system has also ensured a healthy boost to the local economies. The scarcity of many commodities in Afghanistan requires returnees to purchase most of their basic requirements for initial reintegration in Pakistan before departure. Building materials, notably house-posts, household items, agricultural tools, clothing and livestock are all available in the flourishing markets en route to the border.

(65) The substantial potential for cross-border transportation upon which the encashment system depends, also impacts positively on the local economy of the border provinces. While largely Afghan-owned and run, the large fleets of trucks are predominantly purchased, repaired and fuelled in Pakistan.

(66) Another major economic benefit of the encashment system to Pakistan has been the handling of large sums of foreign currency through a national bank. While representing an unusual digression from normal banking business, the National Bank of Pakistan had quickly agreed to UNHCR's proposal to extend its banking facilities to participate in the encashment programme. The operation was of clear commercial benefit to the bank in view of the interest to be gained on the minimum 300,000 US dollar credit balance agreed with UNHCR, interest gained on all additional cash transfers from NBP in New York as well as gains made on converting dollars into rupees. It was agreed that these gains would cover the bank's administrative costs in running the programme.

(67) Between July 1990 and the end of 1993, more than 280,000 stickers were encashed by the National Bank of Pakistan, representing a turnover of almost 40 million US dollars. This figure represents virtually the entire programme costs for the same period. With the exception of the initial setting up of the programme, when a large number of staff were involved on a full-time basis, administrative costs have been minimal, ranging from 1.3% in the peak year of 1992 to 4% in 1993. Thus, even a year when encashment figures have been low, more than 95% of donor funds have gone directly to beneficiaries.

ENCASHMENT PROVIDES DONORS WITH NO-LOSE OPTION

(68) The encashment programme was also extremely attractive to donors, eager to find a solution to the Afghan problem and to wind down a programme which had dragged on for more than fifteen years and which had lost interest for many given the changes in international politics. For many donors, the principal appeal of the system was that it provided a reliable means of deregistration. Whether a refugee family repatriated or remained in Pakistan after encashment was of less importance; the key point for donors was their removal from the assistance register. Consequently, while other aspects of the humanitarian effort in Afghanistan have seen a dramatic reduction in donor interest in recent years, the encashment programme has until recently faced no serious shortage of funds.

(69) Whilst undoubtedly successful in terms of facilitating the return of very large numbers of Afghans from Pakistan, and ingenious in its adaptation to local conditions, the encashment system nevertheless has a number of significant shortcomings which need to be examined to provide a balanced assessment of its impact. Some of these shortcomings have already been touched upon: the lack of a clear causative link between encashment and physical return, the system's susceptibility to abuse and corruption, and the fact that however successful it has been to date, encashment has only involved about one half of the original refugee population.

LACK OF CAUSATIVE LINK BETWEEN ENCASHMENT AND REPATRIATION

(70) Although the figures indicate a loose correlation between encashment and observed border crossings, there is unfortunately no clear causative link between the two. Despite a highly-sophisticated electronic monitoring system, it has not proved possible to link observed border crossings with specific encashment events. Efforts are now being made to improve this link, but it remains impossible to estimate with any accuracy how many of the 900,000 individuals monitored crossing the border in 1992 had recently encashed, how many had encashed a year earlier, and - more importantly - how many had never been registered as refugees in Pakistan. This has important implications for assessing the long-term cost effectiveness of the programme, since it will be impossible to estimate the real number of beneficiaries and the actual amount of money spent on them.

(71) Most estimates place the pre-1993 unregistered refugee population in Pakistan at about 500,000. However, this figure is often disputed and has little concrete basis. On the other hand, it is generally agreed that as the pace of repatriation - and more importantly - encashment, accelerated in 1992, the concomitant reduction in assistance reaching the camps meant that many of the unregistered who had survived until then at the end of a long distribution line fuelled by surplus rations, were now obliged to leave.

(72) In contrast to the preceding year, monitored returns in 1993 were considerably more numerous than encashments. It is generally assumed that this indicated a pattern of delayed returns of families who had encashed prematurely in 1992. On the other hand, it is also acknowledged that the unregistered continue to constitute a sizeable proportion of returnees. While some efforts were made in 1993 to verify this particular point, in the absence of visual checking of passbooks at the border it is impossible to obtain reliable information.

SOME AFGHANS OPT FOR CLANDESTINE LOCAL SETTLEMENT

(73) A further factor which clouds the link between encashment and return is the environmental reality of a porous border, with the same ethnic groups on both sides and a long history of cross-border nomadic, seasonal and commercial movement. Given poor economic conditions in Afghanistan and a plentiful supply of wage labour in Pakistan, many refugees have opted for clandestine local integration following encashment.

(74) Finally, and most equivocal of all, is the extent to which the encashment programme itself has influenced people's decision to return. A question often asked, albeit rhetorically, is whether the scale of observed returns in 1992 would have been any different without the encashment programme.

(75) In the circumstances, it is not surprising that opinions vary greatly over how many of the refugees who have encashed have actually returned to Afghanistan. While UNHCR generally maintains that over two thirds of those encashing have so far returned, some Pakistan government officials place the figure at less than one third, arguing that given current conditions inside Afghanistan, instead of facilitating repatriation, encashment has instead promoted clandestine local settlement.

(76) Meanwhile, agencies working inside Afghanistan on reintegration projects report significant returns in certain areas. In some cases it has been possible to directly correlate encashment figures for specific districts or villages of origin, with empirical evidence of large-scale return. Indeed, some would argue that there is a much closer correlation between rehabilitation and return than between encashment and return. What remains undisputed, however, is that in conditions of continuing insecurity, and in the absence of a political settlement in Kabul, refugees and returnees will continue to maximize their options in a highly fluid situation.

ENCASHMENT IS BASICALLY A GOOD TIMES STRATEGY

(77) Indeed, it may be argued that a fundamental weakness of the encashment system is that it is basically a "good times" strategy; that ultimately it is constrained by conditions inside the country of origin, as recent events in Afghanistan have all too graphically demonstrated. Whilst many refugees, since the beginning of the encashment programme, have made the decision to return to unstable and insecure conditions in Afghanistan, in the end what determines the pace of repatriation is the refugees' perception of their survival prospects in the country of origin.

(78) In a decision-making equation, where conditions in the country of origin are pitted against those in the country of asylum, encashment can be seen as playing only a relatively minor role. Analysts generally characterize refugee decision-making as rational, with security and economic survival being the two principal concerns of potential returnees.

(79) It may therefore be assumed that it is only when conditions in the country of asylum become untenable that a refugee would opt to return to a fundamentally unsatisfactory situation in the country of origin. Paradoxically, the minimal role played by encashment in such decisions may also be seen as one its most significant strengths, as supporting refugee choices and facilitating - rather than promoting - repatriation.

ENCASHMENT DOES NOT PROVIDE THE ANSWER FOR ALL GROUPS

(80) With half the original registered refugee caseload remaining in Pakistan and continuing to receive assistance from the international community, it is important to identify the principal reasons for their decision not to return. Analysis of the approximately 1.6 million remaining registered refugees indicates that they include a number of distinct sub-groups each with their own particular reasons for remaining in Pakistan.

(81) Such sub-groups, with a few exceptions, are not unique to the Afghan caseload but are commonly found in most refugee populations in the process of repatriation. In some cases, there is little likelihood that members of the sub-group will opt to return either in the immediate or long term. These include, for example:

• those who are socially and economically integrated into the country of asylum and see no advantage in returning, and

• old border stragglers who had inadvertently been caught up with the refugee outflow but who would normally have remained legally in Pakistan.

(82) For other sub-groups, whether wishing to repatriate of not, return will not be feasible without specific targeted assistance. These include:

• vulnerable individuals who are without the protection of their extended families, including female heads of household, amputees, etc.,

• those who encashed prematurely and now lack the means to return, and

• those who had never been registered and lack the means to return.

(83) There are others for whom particular conditions in Afghanistan are the principal obstacle to return, and who are unlikely to be willing or able to repatriate until these conditions have been resolved. For example:

• minority Pashtuns from areas in northern Afghanistan which are dominated by other ethnic groups (Tajiks, Uzbeks, etc.) and who may have lost their lands in inter-ethnic disputes,

• those with specific security concerns regarding their village or district of origin, such as continuing fighting or the presence of mines, and

• recent arrivals who fear return for political reasons.

SUBSIDIZED TRANSPORT REQUIRED FOR REFUGEES FROM FAR NORTH

(84) Whilst recognizing that such clear-cut reasons for non-return do not apply to all of those remaining in Pakistan, it is nevertheless generally acknowledged that encashment does not address the needs of all potential returnees. This reality was recognized at an early stage by UNHCR, when specific action was taken to assist the Tajiks, Uzbeks and other ethnic groups from the far north.

(85) The repatriation cash grant was based on the average cost of return to towns and villages in the eastern provinces from which the majority of refugees in Pakistan originated. This was obviously insufficient for those with much greater distances to travel and it was therefore agreed that those going north of the Salang Pass would receive additional assistance.

(86) The subsidized transportation programme has, in fact, benefited more than just registered refugees from the North. Since eligibility for subsidized transport is not tied to the passbook, the programme has also facilitated the return of unregistered and deregistered refugees, provided they are able to pay the required 40% contribution to the fare.

(87) The subsidized transportation programme, which has been run since mid-1992 by IOM, has subsequently been expanded to include a limited number of individuals classified as vulnerable by UNHCR, for whom fully subsidized transport is provided. In late 1993, with encashment figures remaining exceptionally low, consideration was being given to broadening the programme still further as an incentive for others who might be hesitating over their decision to return.

DEREGISTERED ARE DIRECT CASUALTIES OF ENCASHMENT

(88) Of the various sub-groups remaining in Pakistan, it is above all the deregistered who may attribute their current predicament directly to the encashment system. In some cases, deregistration, or encashment without repatriation, has been part of a deliberate strategy by refugees with no intention of returning to Afghanistan, but who prefer to seek some other means to survive in Pakistan with the cash grant as a starting point. In many ways this can be seen as positively contributing towards a durable solution.

(89) In other cases, an unexpected change in circumstances in the province or district of origin may have prevented refugees from returning at the last minute. Perhaps the largest group of deregistered, however, are those who encashed in the months following the fall of Najibullah in April 1992. A number of factors contributed to large-scale panic selling of passbooks in mid-1992, not the least of which was the pervading atmosphere of encashment frenzy itself. Two additional factors coincided to persuade many refugees, who were far from ready to return, that they should nevertheless encash without delay.

(90) A temporary breakdown in the WFP wheat distribution which had deprived the camps of rations for several months had raised concern among the refugees that assistance was drying up. This fear was inadvertently boosted by a UNHCR appeal to donors to ensure that the growing demand for encashment could be met in a timely manner. Broadcast internationally, the appeal was interpreted by many refugees to mean that the money really was running out.

RE-REGISTRATION IS NOT AN OPTION

(91) For those who have encashed in panic or desperation, there is no immediate solution. A fundamental premise of the encashment system is that the population should first be fixed through the process of revalidation, as a means of avoiding, or at least, limiting abuse. To re-open registration at any stage after the encashment programme had begun would inevitably undermine this vital controlling mechanism. Other victims of the curtailment of registration are the growing number of new arrivals, many of whom may be in need of assistance, particularly during the first months, as well as the many unregistered refugees who are destitute.

(92) Although the immediate assistance needs of some new arrivals have been addressed, generally following an intensive screening process, they have not been issued with stickers which would enable them to benefit from assistance for repatriation. Consideration is, nevertheless, being given to providing them with assisted transportation when return becomes possible.

(93) Nevertheless, in the Pakistan context, where for so long the problem has been more one of over-distribution of assistance than under-distribution and where a vibrant informal sector provides ample opportunities for casual labour, it has generally been assumed that such casualties of the encashment system are basically not at risk. Even as rations have been progressively reduced in the camps, there has been an expectation that traditional social mechanisms would ensure that very few would fall through the cracks.

(94) This sense of security, perhaps peculiar to Pakistan, has hitherto obviated the need to find a practical and humane solution. Indeed, no such assumptions could have been made in other refugee situations, such as the Mozambicans in Malawi, where the population is highly aid-dependent, and where survival without prompt intervention would be inconceivable.

SAFETY NET BREAKS DOWN WITH MOVEMENT TO URBAN AREAS

(95) There is some evidence, however, to suggest that even in Pakistan, these long-held assumptions may no longer be totally valid. Afghan doctors at an NGO-run clinic in Quetta reported seeing a growing number of unregistered and deregistered refugee patients during 1993. They had also detected a modest increase in malnutrition rates between 1992 and 1993, mostly involving the dependents of widows or women heads of family whose husbands had returned to Afghanistan. In their view, these changes indicate that extended families are beginning to break-down under extreme pressure and, in some exceptional cases, some husbands are even abandoning their wives.

(96) This view is shared by a Peshawar-based anthropologist with long experience with Afghan culture. She argued that the integrity of the extended family depends ultimately on the amount of pressure placed upon it. Under normal circumstances, if a widow has no source of income, she will not be thrown into the street, but nor will she - or her children - be treated equally. Her brothers-in-law will use the family's resources first and foremost for themselves and for their own children. However, under extreme conditions, particularly under the pressures of urban living, the extended family may begin to give way and eventually be forced to resort to the abandonment of vulnerable individuals.

(97) Yet it is precisely to the urban areas that refugee families are compelled to move once they have lost their source of assistance - either having themselves encashed or as a result of the departure of a former benefactor. Rawalpindi, Sind, Lahore, Quetta and Karachi are common destinations of the deregistered and the unregistered. Although this rural-urban movement has long been tolerated, local Pakistanis are increasingly finding themselves in direct competition with Afghans in certain sectors and consequently demands for restricting such movement are mounting.

(98) As the political complexion of Pakistani-Afghan relations changes, so too may the tolerance of the government and the local population to such infringements on their daily life. Such changing attitudes may impinge on the government's view of the status of deregistered refugees remaining in Pakistan. Although at present the government still considers them as refugees, this view could quickly change when patience with a neighbour who has long overstayed his welcome begins to run out.

MOVEMENT INTO REFUGEE VILLAGES IMPEDES ASSISTANCE REDUCTION

(99) While many deregistered and unregistered refugees endeavour to survive in an urban environment, others have taken advantage of the large-scale repatriation movement in 1992 and smaller-scale movements in 1993 to move into established refugee villages. By doing so, they are able to benefit from the community assistance provided by UNHCR and its operational partners, notably access to water and to health and education services.

(100) As a consequence of this trend, UNHCR's efforts to rationalize its care and maintenance programme by reducing the number of refugee villages is impeded, since services can only be reduced, or cut totally, once depopulation has reached a certain level. While a reduction from 348 to 190 villages has been achieved to date, this has mostly been through "administrative mergers", and the link between encashment levels and reduction of services remains loose. Vested interests in keeping UNHCR-funded assistance programmes at artificially high levels are also a serious obstacle to this process.

ABUSE AND CORRUPTION HAVE UNDERMINED THE ENCASHMENT SYSTEM

(101) Perhaps the most serious weakness of the encashment system in Pakistan is the fact that its durability has been seriously jeopardized by its susceptibility to abuse. It should be recalled that the Project Market Place model from which the encashment system evolved, was designed specifically to take advantage of the entrepreneurial nature of the refugee and local populations and the cross-border potential for expansion of trade.

(102) Based, as it is, on a fundamentally faulty and unverifiable registration system, the encashment system has encouraged trading and dealing in registration cards and enables people outside the target group to benefit and profit from UNHCR assistance. Evidence of the target and local populations' responsiveness to the principles of Project Market Place was found only two or three months after the inception of the encashment programme, as refugees reported that the resale value of passbooks had already doubled during this period.

(103) The economic and financial environment in Pakistan provided ample opportunity for widespread manipulation of the system to maximize individual profit. For example, there have been frequent and significant fluctuations in the purchasing power of the cash grant, in the price of wheat both in Pakistan and in Afghanistan, and in the exchange rate between rupees and Afghanis. During the peak encashment and repatriation period in mid-1992 both the price of wheat in Afghanistan and the exchange rate of the Afghani tripled, bringing huge profits to tradesmen and money changers.

TENSION BETWEEN REFUGEE DECISION MAKING AND MAXIMIZING OPTIONS

(104) Some would argue that there is a fundamental tension between UNHCR's advocacy of refugee decision-making on the one hand and its reservations about their success in maximizing options on the other, particularly when the latter borders on exploiting the system. In fact, UNHCR's perception of refugee behaviour in Pakistan is, to a large extent, determined by the individual's intentions.

(105) If a refugee's intention is ultimately to return to Afghanistan, maximizing behaviour such as splitting his family, keeping his passbook current, doing seasonal work in Pakistan and returning periodically to Afghanistan to assist in the reconstruction of his home village, is considered quite appropriate and commendable. If, on the other hand, his intention is to remain in Pakistan, similar behaviour is viewed as unacceptable.

(106) In some instances, however, encashment has led to behaviour which goes far beyond simple exploitation of an individual's options. One official in Baluchistan remarked that in some camps there are now more bogus passbooks than real ones. In certain areas corruption is so pervasive that it is almost impossible to overcome, particularly when local officials are among the major beneficiaries. Fear of provoking a major break-down in law and order prevent UNHCR from promoting the implementation of many of the more obvious solutions.

(107) In the view of some observers, the break-down of the encashment system was inevitable, and in some ways even warranted, since from the outset it has been based on a gamble. Indeed, if all 600,000 passbook holders had decided to encash at the same time, the system would have quickly collapsed. Recognition of this reality has so far prevented UNHCR from advocating the introduction of a cut-off date for encashment, an initiative which might otherwise have been a possible means of resolving the current impasse.

ENCASHMENT SYSTEM IN BALUCHISTAN HAS BROKEN DOWN

(108) At the present time, at least in Baluchistan, the encashment system may certainly be considered as having broken down. As a result of misuse and abuse of the system, neither of its original objectives - deregistration and facilitation of return - can now be ensured. As we have seen, deregistration without repatriation has been occurring on a large scale. Now, thanks to the ingenuity of a number of group leaders in Baluchistan, a parallel system of encashment has been established, offering a more convenient, and in some cases, more lucrative, exchange rate than the official system. The passbook remains in circulation while the refugee may well decide to go home. Repatriation, therefore, is now occurring without deregistration.

REPLICATING ENCASHMENT: LEARNING FROM THE PAKISTAN EXPERIENCE

(109) Despite the limitations and the less positive consequences of encashment in Pakistan, the programme has nevertheless been successful in facilitating the return of an exceptionally large number of refugees. Since UNHCR is likely to be involved increasingly in operations involving return to conflict-affected countries, and since encashment provides the flexibility required to facilitate repatriation in such situations, closer attention to the question of replicability of the system is warranted.

(110) It is not so much replication of the Pakistan model of encashment that is proposed, but rather the approach - or operational principals - it represents: maximizing refugee flexibility and choice. As we have seen, in practical terms the approach comprises a number of distinct components which can be modified to suit each particular situation. These include:

• encashment and deregistration procedures
• implementing arrangements
• amount of assistance
• balance between assistance in cash and in kind
• currency of cash assistance
• location of encashment points

(111) Experience in Pakistan suggests that there are three or four key operational and organizational requirements which must be met if encashment is to function effectively. These are:

• clear, shared, operational objectives for activities in the country of a origin and in the country of asylum.
• an integrated strategy incorporating activities in both country of origin and country of asylum.
• effective coordination of activities on both sides of the border.
• solutions-oriented registration systems at the beginning of an emergency operation.

NEED FOR CLEAR OBJECTIVES

(112) Traditionally, UNHCR has promoted voluntary repatriation in situations where there is a general assessment that refugees can return safely to their country of origin, with exceptions being made for a minority of residual cases unable to return for political or other reasons. More recently, however, the concept of 'no-go areas' and 'safe zones' have been introduced in frank recognition that conditions in parts of a country of origin were still not safe.

(113) Such departures from the more traditional approach inevitably involve a certain conflict of principle and pragmatism. They are commonly justified on the basis of their support for an existing tendency to spontaneous return as well as for their potentially positive impact on the process of national reconciliation.

(114) These changes reflect an evolution which has been taking place over the past decade in UNHCR's approach to repatriation. The 1985 Conclusion of the Executive Committee was a turning point from a hitherto definitive stance in which repatriation of refugees could take place "only at their freely expressed wish" and under "conditions of absolute safety" to one in which repatriation might be encouraged "whenever it is deemed to be in the interests of the refugees concerned".

(115) The shift has been seen most graphically in the Gulf emergency operation and in Cambodia. In Iraq, the concept of 'safe zones' was developed for the first time, whereby refugees could return despite continuing hostilities elsewhere in the country. In Cambodia, the reverse idea - of 'no-go zones' - was used, also for the first time, whereby repatriation was promoted to all areas except those considered particularly unsafe.

DEBATE OVER VOLUNTARINESS VERSUS SAFETY ON RETURN

(116) The debate over the relative primacy of the principle of voluntariness versus the existence of safe conditions in the promotion of repatriation, continues unabated. The approach used in Afghanistan has been one of facilitated return, in the light of UNHCR's recognition that conditions in the country are not yet suitable for promoted return.

(117) With an increased focus on supporting refugees' choice to return spontaneously to their countries of origin, organized repatriation may well prove to be increasingly inappropriate for repatriation movements of the future. Indeed, a recent evaluation of the Cambodia repatriation operation pointed out that the return of the border population from Thailand could have been much less intensive, complex and costly if UNHCR and the international community had been more sensitive to indications that many among the border population were ready to return spontaneously, even to areas considered unsafe by UNHCR, in advance of the organized repatriation.

(118) The fundamental principle upon which the encashment system in Pakistan is based is that of maximizing refugee choice and supporting spontaneous returns which had been occurring since the late 1980s. Essentially this means giving the refugees a choice to return to their country of origin or to remain in the country of asylum. It is this choice which makes the difference between facilitating - as opposed to promoting - repatriation.

(119) While the encashment system could be used equally effectively in either situation, it is essential that all concerned - particularly the refugees - know which situation applies to them. Where repatriation is facilitated, and for as long as the unsettled conditions prevail, there should logically be flexibility in the time-frame for return. Facilitation should occur in a neutral environment where success does not depend on numbers, but on maintaining optimum conditions for refugees to make considered choices.

SUCCESS HAS BEEN LINKED TO MAINTENANCE OF HIGH NUMBERS

(120) The repatriation operation in Pakistan has suffered from a lack of clarity in its objectives. While presented by UNHCR as a facilitated repatriation operation, success has nevertheless been linked to and perceived in terms of the maintenance of high numbers. Similarly, the response to declining numbers has been an attempt to modify the programme to encourage them to rise again.

(121) The mixed messages created by this operational response, and the confused objectives which it implies, have not only increased the existing tendency to abuse the system, but have also made it difficult for UNHCR to determine the appropriate balance in other aspects of the operation. It is generally accepted that apart from conditions in Afghanistan, a significant retention factor for refugees in Pakistan is the continuation of the care and maintenance programme.

(122) While considerable efforts have been made in recent years to gradually wind down this massive assistance operation, it is generally acknowledged that any further significant cuts could change the delicate balance of go/stay factors and lead to more people deciding to repatriate prematurely or to migrate to the cities. While some in UNHCR argue against such action as being tantamount to encouraging return, others see little difference between cutting rations and offering a higher cash grant or more subsidized transport.

(123) There is, in fact, an important difference between them. What is significant about retaining assistance in the country of origin for those refugees who lack the potential to become self-sufficient is that it enables them to keep their options open. For encashment to function correctly as a facilitating mechanism for repatriation, this is an essential element in maintaining a neutral environment, in the same way as keeping the borders open and not imposing time limits.

(124) Inevitably a tension exists between the ideal of maintaining optimum conditions for refugee decision- making and the practical reality of running a multi-million dollar assistance programme which has lasted much longer than donors or the host government would have wanted. The future of the encashment programme is, however, unavoidably linked to the eventual phasing out of the care and maintenance programme. Planning for the progressive handing over of sectoral activities and for dealing with the residual caseload, will be greatly enhanced by harmonizing the objectives of repatriation and the care and maintenance programme.

NEED FOR AN INTEGRATED STRATEGY

(125) A second key operational requirement for the effective functioning of the encashment programme is that it be part of a comprehensive package involving activities in the country of asylum and the country of origin. Indeed, when the encashment system was first developed, it was perceived as being just one component in an integrated strategy which would not only provide the means for refugees to return at the time of their choice and under conditions of safety and dignity, but would also help to generate the conditions for such return. The principal components of this solutions-oriented strategy include:

• assistance in the camps
• open borders
• cash grant
• individual assistance/wheat
• subsidized transport
• demining
• reconstruction/rehabilitation

(126) A similar combination of components may be found in any repatriation plan, whether for Cambodia, Namibia or Nicaragua. Additional components might include:

• tracing/family reunion
• special assistance to vulnerable groups

(127) Tracing and family reunion have not been major issues for the Afghan refugee population since their movement into exile was generally well organized and included complete family groups. They may, however, be extremely important in other repatriation operations, particularly if the refugee population had experienced severe pre-flight social disruption and if the departure from the country of origin had been disorganized, as in Liberia and Cambodia. In fact, failure to take account of this important element was a significant weakness in the Cambodian repatriation operation as a result of which opportunities to strengthen the long-term reintegration prospects of many returnees were lost.

SPECIAL ASSISTANCE TO VULNERABLE GROUPS SHOULD NOT BE DELAYED

(128) In setting up the encashment programme, it was agreed that given the enormous numbers involved, encashment was the only feasible way to deal with at least part of the population. There was recognition, however, that cash grants would not meet the needs of everyone. In particular, vulnerable individuals and those with long distances to travel would need additional assistance. To date, efforts made towards providing additional assistance to individuals with special needs have been limited to the provision of subsidized transportation. Identification and registration of the vulnerable population has been postponed until the refugee population has fallen to a more "manageable" size (generally assumed to be at about 500,000).

(129) This delay may well prove to be self-defeating, since viable support strategies for vulnerable individuals are generally more readily identified at an early stage in a repatriation movement when their return with other able-bodied refugees can be facilitated. The Cambodian model of identifying support families from amongst the refugee population to accompany vulnerable returnees and assist in their initial reintegration may usefully be adapted to comply with Afghan social norms. Above all it can minimize the risk of a predominantly vulnerable residual population being left in Pakistan.

(130) The delay in identifying vulnerable individuals may also hinder UNHCR's capacity to identify those in need and to develop appropriate solutions for them. The number of support staff, including social services, is progressively being reduced as the overall refugee caseload decreases. The labour-intensive casework which will be required to develop appropriate strategies for individual vulnerable refugees has important staffing implications which might be difficult to satisfy when the programme is moving towards phase-out stage.

COMMUNITY-BASED RECONSTRUCTION WORK HAS NOT MET EXPECTATIONS

(131) One of the key elements in the integrated strategy for repatriation of the Afghan refugee population has been reconstruction and rehabilitation work in Afghanistan, including demining. It was recognized at an early stage that the provision of cash grants and wheat alone would not suffice to facilitate large-scale return to a country which had suffered such massive devastation. Community-based reconstruction projects would be required to provide the minimal environmental conditions for survival in the worst-hit areas and to enhance absorption capacity generally.

(132) Overall, repatriation-related reconstruction efforts in Afghanistan have had the objective of reinstating what was functional before the Russian invasion, including repairs to irrigation and water systems, road repairs and mine clearance. For a variety of reasons, reconstruction efforts in major returnee areas have fallen short of expectations.

(133) There are differing views over why reconstruction efforts have been so limited and the extent to which more could be done under current conditions. UNHCR staff in Pakistan, whose cross-border efforts came to an abrupt stop in the wake of a fatal incident in Jalalabad in February 1993, maintain that security conditions are still too precarious to permit much more to be done.

(134) UNHCR Kabul staff (evacuated to Islamabad in late 1992), who took over responsibility for cross-border activities from their colleagues in Pakistan in mid-1993, are hesitant about cross-border operations generally. For example, despite a number of requests, including those from refugee groups, they have been unwilling to consider funding large-scale cross-border road reconstruction. UNHCR Kabul maintains that such reconstruction efforts are the responsibility of other UN agencies, particularly UNDP. UNDP, however, has refused to fund the projects on the basis of their being capital- intensive.

LACK OF STRATEGIC PLANNING FOR REINTEGRATION AND RECONSTRUCTION

(135) The fundamental problem underlying this confusion is the lack of strategic inter-agency planning for reintegration activities. At no stage in the repatriation planning process have the roles and responsibilities of the various agencies been defined and agreed upon.

(136) The resulting inter-agency confusion belies the reality on the ground. UNDP/OPS and international NGOs working in the southern and eastern provinces of Afghanistan on a range of reconstruction projects, have a rather different perspective. Generally operating through indigenous NGOs, they have been obliged to adapt their working methods to suit the environmental realities and consequently view security as a constraint, which can be planned for, rather than a risk which jeopardizes their entire operations.

(137) These operational agencies generally agree that a great deal more needs to be done in Afghanistan and could be done since their respective capacities are far from exhausted. For example, a senior UN officer familiar with operations in Afghanistan estimated that the UNDP/OPS work in rehabilitation of irrigation systems could readily be doubled and still be far from meeting the potential needs.

(138) Efforts in road construction are similarly far from exhausted yet remain seriously under-funded. As noted earlier, UNDP is unwilling to fund such activities, despite their key potential role in reviving markets to the pre-war level. Non-funding of such projects represents opportunities lost in terms of the spin-off benefits they could bring to the local communities, in particular through cash flow as a result of local employment possibilities. More significantly, the stalemate has meant the loss of opportunities for creating conditions conducive to increased returns.

HESITANCY OVER RECONSTRUCTION UNDERMINES INTEGRATED STRATEGY

(139) Whilst supporting encashment as a means to resolving a long-term problem, there is generally less interest amongst donors in funding reconstruction work in Afghanistan from within their overstretched humanitarian budgets, particularly when there is ongoing conflict and therefore uncertainty over its impact. On the other hand, development funds cannot be utilized for such activities until a stable government is established in Kabul. It should be noted, however, that this stalemate situation is not unique to Afghanistan. Donors were similarly hesitant about funding reconstruction work in Cambodia until they were sure that the peace settlement would hold.

(140) However, as a result of the hesitancy of donors and UN agencies, including UNHCR, to treat security problems in Afghanistan as a constraint rather than a risk, a vicious circle has emerged: deteriorating security makes donors hesitant to commit more aid, so assistance decreases, generating increased conditions of desperation and rivalry which in turn generates further insecurity. In consequence, an essential component in the integrated strategy for repatriation is further undermined.

NEED FOR EFFECTIVE COORDINATION

(141) A critical organizational requirement for the successful functioning of the encashment system, like any other complex operation, is effective coordination. In repatriation operations, coordination is required at a number of levels, not the least of which involves different staffing components within UNHCR itself: those at headquarters and in the field, those in the country of origin and the country of asylum, and those working in the country of asylum on the annual programme and those involved in the repatriation operation. Coordination is also of fundamental importance between UNHCR and its governmental and non-governmental partners in countries of first asylum and origin as well as between UNHCR and the various UN agencies involved in reconstruction and development in the country of origin.

(142) The Afghan repatriation operation has suffered from significant weaknesses in many areas requiring coordination. To some extent these weaknesses may be attributed to the fact that at no time in the past fifteen years has there been a central government in Kabul which controlled the entire country. Long before the September 1992 outbreak of inter-factional fighting between rival Mujahideen leaders, Kabul-based agencies recognized by the government were denied access to many of the major returnee areas.

(143) There has consequently been a major split, from the outset of the repatriation process, between agencies working from Kabul and those working cross-border from Pakistan. Prior to the fall of Najibullah, the split had inevitably developed certain political overtones, in common with many other refugee situations. Although the political dimension has now changed, the split remains and is particularly marked within UNHCR.

(144) Despite the fact that UNHCR Kabul staff have been on evacuation status in Pakistan since fighting broke out in Kabul in late 1992, they remain reluctant to consider re-directing their efforts across the border. Thus, although UNHCR Pakistan handed over responsibility for cross-border activities to UNHCR Kabul in mid-1993, the latter has been cautious in its pursuit of such work, partly as a result of its reservations about working out of Pakistan and partly because of its serious lack of implementing capacity. Furthermore, UNHCR Kabul's perception of its role within an integrated strategy for repatriation is limited to meeting no more than the basic needs of returnees.

(145) Conflicting views over UNHCR's objectives for the Afghan refugee population in Pakistan are not limited to a split between the country of origin and country of asylum. Since repatriation was first placed on the agenda in the late 1980s clear lines were drawn, in Pakistan as well as in Geneva, between those who consider their responsibility to be the upholding of the annual care and maintenance programme and those who feel obliged to "produce" repatriation at any cost. Whilst the continuing political instability in Afghanistan has ensured that this will never be a simple black or white affair, failure to seek a compromise between the opposing views has inevitably undermined the integrated strategy.

OURA REINFORCED ORGANIZATIONAL FAULT LINES

(146) Various efforts have been made by UNHCR to coordinate its activities in the country of origin and the countries of asylum. Shortly after the Peace Accords of 1988, a Geneva-based coordinating mechanism - the Operational Unit for Repatriation to Afghanistan (OURA) - was established, with the objective of rationalizing and coordinating repatriation-related activities in Afghanistan, Pakistan and Iran.

(147) However, in the view of many UNHCR staff involved with the Pakistan operation at the time, rather than resolving existing conflicts, OURA complicated them further by reinforcing the organizational fault lines which had developed at all levels between those responsible for the ongoing assistance programme and those responsible for repatriation. The resulting split structure was described by one staff member as "two separate orbits of decision-making".

(148) In 1991, when the mass returns anticipated in Geneva failed to materialize, OURA was phased out as donor attention turned from Afghanistan to other, more dynamic, operations, notably the Gulf emergency. For the following year, the coordination function reverted to the Desk. It was agreed, however, that once repatriation movement began to increase significantly, any future coordinating mechanism would be based in the field rather than at Headquarters.

(149) Consequently, in May 1992, following the fall of the Najibullah government and the sudden rapid increase in encashment and physical returns, the High Commissioner appointed her Chief of Mission in Islamabad as Special Envoy. His responsibility included overall planning and coordination related to the voluntary repatriation of Afghan refugees from the neighbouring countries, as well as UNHCR's initial reintegration activities in Afghanistan. With the rapid decline in repatriation following the renewal of fighting in Kabul in September 1992, however, and the reassignment of the incumbent a year later, the function of Special Envoy has lapsed. The coordinating function is currently entrusted to the quarterly regional coordinating meetings.

STAFFING IMBALANCE UNDERMINES COORDINATION

(150) In recent years, as UNHCR has begun to play a more important role in the initial reintegration of returnees, the issue of coordination of activities between the countries of asylum and origin during a repatriation operation has become more critical. A complicating factor common to many operations involving large refugee populations, particularly those which have had long-established care and maintenance programmes, is that staffing levels in the country of asylum are likely to be significantly higher than in the country of origin.

(151) Consequently, in the early stages of planning for repatriation, the programme in the country of asylum will almost certainly be headed by a much more senior staff member than in the country of origin. This reality was no doubt taken into account in the decision to appoint the Chief of Mission in Islamabad as Special Envoy for the Afghan repatriation operation. Nevertheless, while every effort was made by the incumbent to clearly distinguish between his two roles, the perception of many players involved was that repatriation to Afghanistan was a Pakistan-led affair.

COORDINATION CAN BE ENHANCED BY INTEGRATING OPERATIONS

(152) This problem was avoided in the operational phase of the Cambodian repatriation by the appointment of a Special Envoy whose function was distinct from both that of the Representative in Thailand and of the Chief of Mission in Cambodia. While based in Cambodia, the Special Envoy had overall responsibility for the operation on both sides of the border and was able to go a long way towards combating long-established centrifugal tendencies.

(153) When more than one country of asylum is involved, the location of the coordinating mechanism is all the more critical. While mobility will inevitably be a key feature of a successful repatriation coordinator, residence in the country of origin, however symbolic, is essential. Furthermore, as recommended in the Central Evaluation Section's Review of the Cambodia Repatriation Operation, the coordination function can be significantly strengthened by appointing staff to an integrated operation rather than to a particular country, and by giving staff with key functions a cross-border or regional responsibility.

UNOCHA'S COORDINATION ROLE WAS SERIOUSLY FLAWED

(154) Coordination outside of UNHCR has also been fraught with problems in the Afghan operation. The overall coordinating mechanism for the United Nations system in the region, UNOCHA (formerly UNOCA), has experienced serious difficulties in fulfilling its mandate since its foundation in 1988. Given conditions in Kabul at that time, a decision was made to locate UNOCHA's focus for coordination in Geneva, rather than in Pakistan or Iran. The bulk of the professional staff were based in Geneva with limited staff placed in Kabul, Teheran and Islamabad.

(155) In the view of many of those working close to the operation over recent years, the inherent weaknesses in such "remote-control" coordination efforts were compounded at an early stage by UNOCHA's desire to guide and direct the work of the specialist agencies rather than simply to support them. Structural tensions between UNHCR and UNOCHA further compounded the problem. A senior WFP official went so far as to say that UNOCHA's failure to play a coherent coordinating role had seriously eroded the UN's potential for working within Afghanistan.

(156) The organizational split between staff in Kabul and those in Islamabad was not unique to UNHCR. A similar pattern existed in other UN agencies such as UNDP, FAO and UNICEF which had also been represented in Kabul during the Soviet occupation. Overlapping mandates and differing technical approaches, particularly in the critical area of water resources, further complicated matters. The provision of economic assistance under Operation Salaam also caused considerable tensions between UNOCHA and UNDP Kabul. Despite the termination of Operation Salaam and the change of UNOCHA's mandate in 1993, such tensions continue to impede effective coordination in an increasingly complex political environment.

(157) One of the most serious casualties of the generalized failure in coordination in the Afghan operation has been the integrated strategy for repatriation. The strategy has been systematically undermined by a lack of planning for reintegration activities at any stage in the repatriation operation. Little attention has been paid to the development of an inter-agency plan for the socio-economic reintegration of returnees to Afghanistan.

(158) Early efforts towards inter-agency cooperation in the Cambodia repatriation operation ultimately succeeded in avoiding many of the problems experienced in Afghanistan. An inter-agency mission to Cambodia was organised more than a year before the signing of the peace accords, setting the scene for later cooperation. At the beginning of the repatriation movement in early 1992, a Memorandum of Understanding was signed between UNHCR and UNDP, defining their respective roles and responsibilities with regard to the reintegration of Cambodian returnees. While this critical partnership still has considerable scope for improvement, concerted efforts have been made by both agencies to promote the reintegration of new arrivals and to rehabilitate the areas in which they have settled.

NEED FOR SOLUTIONS-ORIENTED REGISTRATION SYSTEMS

(159) Finally, at a more specifically operational level, the Pakistan experience highlights the critical importance of establishing solutions-oriented registration systems from the outset of an emergency programme. Recent efforts in the Burundi emergency, to include in the registration system the necessary details on refugees' place of origin to facilitate their eventual repatriation, represents a significant step forward to addressing this perennial problem.

FLAWED REGISTRATION ENCOURAGED ABUSIVE PRACTICES

(160) The registration system adopted in the Afghan refugee villages has had a significant impact on UNHCR's assistance activities and, more recently, on programmes established to facilitate voluntary repatriation. In 1979, when the growing numbers of new arrivals began to exceed the government of Pakistan's capacity to conduct systematic manual registration, a decision was made to enlist the help of traditional community leaders (maliks) in registration and in the distribution of relief items. After a promising start, the new system quickly ran into trouble as unscrupulous "ration maliks", recognizing an almost limitless potential for exploitation, took the place of the traditional leaders and declared hugely inflated entourages.

(161) In an effort to curtail this abusive practice, in 1980 the government banned the distribution of rations to tribal chiefs or other large groups and set up a new procedure whereby rations would only be delivered to individual family heads. The system was to be facilitated by security procedures which required each family head to obtain a certified identity card through one of the Afghan political parties based in Peshawar.

(162) A number of factors contributed to the general failure of the new system as a means to regulating ration distribution. Firstly, some of the ration maliks were reluctant to give up their newly-found power and were sufficiently well-armed to hold on to it. Particularly in Baluchistan and in the Tribal Areas of the North West Frontier Province (NWFP), where the influence of the Afghan political parties was very much weaker than in the rest of NWFP, (and the maliks consequently rather stronger), the "malik system" has managed to survive more or less intact. Moreover, the potential avenues for exploiting the system were so numerous, that multiple registration and trading of passbooks continued to flourish throughout NWFP even without the involvement of the maliks.

(163) The government's efforts to cope with this situation have, to an extent, been constrained over the years by its own ambivalence towards the problem. Multiple registration and a certain degree of inflation have always been tolerated for a number of reasons. They permit the trickling down of rations in sufficient quantities to the large population of unregistered refugees as well as new arrivals, the transfer of much-needed stocks into Afghanistan, and some limited stock piling in case of emergencies. Furthermore, while attempting to foil large-scale diversions by unscrupulous refugees and non-refugees, the authorities have remained determined to continue to allocate sufficient supplies to avoid depriving vulnerable refugees of their entitlements.

(164) Many efforts have been made over the years to reduce the gap between the distribution of assistance and the number of beneficiaries. Despite fierce and sometimes heavily-armed opposition, periodic verification exercises have been conducted in an attempt to identify and eliminate duplicate and bogus cashbooks. On a number of occasions, this resulted in the large-scale cancellation of hundreds of thousands of registrations.

(165) At times, these efforts have been linked with a freeze on registration, but organized in such a way as to maintain a large enough margin to ensure that those at the end of the assistance trickle still receive enough to survive. But each freeze has been followed eventually by a re-opening of registration in the wake of a new influx and in no time at all the temporarily- controlled inflation would once again run out of control. Thus, the registration and assistance delivery systems established in NWFP and Baluchistan have provided ample breeding ground for opportunistic practices. The consequences for UNHCR's repatriation assistance to Afghan refugees in Pakistan of the failure to address these fundamental problems are considerable.

(166) The experience in Pakistan has underscored the importance of clear objectives, an integrated strategy, effective coordination and solutions-oriented registration systems to a successful repatriation operation. These key organizational and operational requirements are interdependent and consequently weaknesses in one are likely to reinforce weaknesses in another. Many of the weaknesses identified in the Afghan operation are common to other repatriation operations. The success of any future operation will depend on our capacity to confront and resolve them.

---

1 The cost of hiring a truck, which will carry two or three families and all their goods to a destination within a day's travel of the border, is about 6,000 rupees. Alternatively, 3,300 rupees could purchase up to a year's supply of wheat for a family or cover the cost of constructing a simple house.

2 The male person is used throughout in recognition of the strict gender rules adopted by the Afghan refugee community which severely limit the movement and social interaction of women in public. In the context of the encashment programme, it is highly unlikely that a woman would contemplate travelling back to Afghanistan without an accompanying male relative.