The Registration System

Individual family members are recorded on the registration card with their name, gender, month and year of birth, and "rank"; a serial number of two digits. The family registration number followed by these two digits may thus constitute a 'personal identification number'. However, such numbers contain information about the Head of Family, not the individual. The family remains the unit of registration in UNRWA's system, cards are issued in the names of family heads.7

It is important to note that initial registration with UNRWA, reporting of "changes", such as marriage, births and moving, and utilization of UNRWA services are all voluntary. Because UNRWA registers descendants of refugees in the male line in addition to those who originally fled, information recorded on the registration cards may be rendered obsolete if changes in the family's situation are not reported.

One group of changes are those related to the "natural" life cycle of the family, such as births, marriages and deaths. In many cases the effort involved in reporting changes tends to outweigh the benefits. As a result, births and deaths both seems to be under-reported.

Assuming relatively stable fertility in the Fields of Gaza and the West Bank the total numbers of births among refugees recorded by UNRWA's health statistics for 1993 are much higher than the number of persons aged 0 to 1 recorded in the family registrations cards for 1992 (Figure 1, and Appendix 5, table 1). The under-registration seems to continue to a lesser degree up to the age of three or four years.8 Automatic transformation of the recorded event "birth" into a recorded "person" in the family registration card system would represent an improvement of the latter data base.9

It has been claimed that UNRWA registration statistics are inaccurate due to under-reporting of number of deaths. A possible indicator for this phenomenon would be the share of registered refugees aged 60 years and older.10 In 1992 this age group comprised 9.3% and 6.8% of the refugees registered by UNRWA in the West Bank and Gaza respectively. In FAFO's living conditions survey in the same area (Heiberg and Øvensen 1993) the corresponding percentages were 6.3% and 5.8% (Appendix 5, table 2). Using the share of refugees aged 60 years or older as an indicator, there is thus no clear support for this claim.

In principle, under-reporting of births and deaths, however, both contribute to an upward bias in the age of the registered refugee population. In the West Bank and Gaza, those aged 14 years or younger comprised respectively 32% and 41% of the refugees registered by UNRWA.11 In FAFO's living conditions survey (Heiberg and Øvensen 1993) the corresponding percentages were 39% and 49%. The difference mainly seems to be caused by under-reporting of births. (Appendix 5, table 3)

Upon the marriage of a child the refugee cardholder himself must come to UNRWA at camp or Area level to register and announce the "splitting" of his family. UNRWA's family size statistics show remarkable differences in family size varying with host country, Jordan having the largest families. This phenomenon can be explained with cases where the Head of Family does not report the marriage of his (or her) children, but keep them on his registration card, adding his grandchildren to it as well.

Marriage frequently also trigger off moving which is another source of population registration errors. The least problematic, and possibly most common, type of moving is movement within one of UNRWA's Fields of operation. More problematic are movements between Fields, but in case UNRWA's services are utilized by the families at their new Field of residence, the movement will be recorded by the registration system.

If a family moves outside UNRWA's areas of operation, however, it has no incentive to notify UNRWA, and substantial registration errors may occur. Movements from UNRWA's five Fields to the Gulf Countries and Western Countries are the most common example. When the net migration balance was negative in the 1970s and 1980s, the registered refugee population probably exceeded the number of refugees actually living in the respective Field, possibly with the exception of Syria.
During and after the 1991 Gulf War a large number of UNRWA refugees left the Gulf Countries and returned to UNRWA Fields. If their original move to the Gulf Countries was not reported and registered, neither did their return show up in the registration records. Is it so that as a result neither their original moving, nor their return did show in the UNRWA refugee registration system?
The Jordan Field provides one example. Before the Gulf War, in September 1990, the population of UNRWA registered refugees in Jordan was 936 000.12 After the return of Palestinian refugees from Kuwait in September 1992, the number of registered refugees in Jordan was 1 025 000.13 If we subtract an 3% annual natural increase in the refugee population in Jordan at that time, the net increase caused by moving to Jordan during this two year period was merely 30 000 persons, far below the number stated by field reports.14

The discussion above shows that lack of reporting of births, deaths and movements renders the number of registered UNRWA refugees less valid as an indicator for the number of UNRWA refugees actually living in an area. It is not clear whether the resulting effect is an over- or underestimation of the number of UNRWA refugees actually living in each Field. To the extent under-reporting of births and deaths occur, however, they both yield an upward bias of the average age of the registered refugee population.

The problem of registration errors could, in principle, be reduced if appropriate systems for assessing biases were introduced, for example by household interviews for checking family size in a random sample of registered refugee families. Checking up registration records in the field would, however, be controversial, and would imply a change of the UNRWA policy of voluntarily reporting among refugees. There is however one important exception to the UNRWA policy of voluntarily reporting among refugees, namely the practice used for "Special Hardship Cases", (SHC).

In Syria UNRWA has recently introduced the so-called Unified Registration System (URS). In the URS changes in the family status like for example births are entered on the computer in the field. This system eliminates or shorten the period from changes are reported in the field till they appear in the records and allows for immediate overview of all transactions that have taken place for a given family name. The system in itself does not routinely capture changes, though, as reporting of changes remain voluntary.

7.) UNRWA announced in 1982 new registration cards which would be issued for every refugee indivdual, thus substituting the present family registration number with a personal identification number. This meet resistance both from the refugees and from the Jordanian government, and UNRWA cancelled the plan (Viorst 1989;40)
8.) Source: UNRWA Statistical Yearbook for Education; 1992-1993 table 1. (Registration records for 1993 has not yet been available to us.)
9.) However, UNRWA registration systems would then develop into a full scale vital registration system.
10.) The validity of this indicator is weakened by the fact that people die at all ages, also before the age of 60.
11.) West Bank includes UNRWA refugees in Arab Jerusalem
12.) Source: UNRWA Statistical Yearbook for Education 1990-1991, table 1
13.) Source: UNRWA Statistical Yearbook for Education 1991-1992, table 1
14.) Source: Guide to UNRWA, April 1993; 6

al@mashriq                       960428/960613