Wednesday, November 09, 2005

ID-ing the Poor

Although I agree that we do need a single identification number (or a SSN-type of ID) for administrative purposes, I would be very cautious in advocating the use of this type ID for any other purposes. In fact I think we need to make a clear distinction between creating ID for administrative purposes and collecting data for policy or research purposes.

The privacy concerns with linking a national ID to any sorts of individual or household data are well justified. There is no quicker way to undermine the credibility of a data collecting agency such as BPS than to ask them to make a transparent list of individuals and households with all of their characteristics. The agency and its activities will immediately be politicized by the government, political parties, researchers, media, businesses and very soon the public will stop answering any questionnaires. Mayling Oey Gardiner wrote an excellent column on this in Kompas which I wished more people have discussed. She said,

“…Patut disayangkan bahwa di kalangan pemerintah pun berkehendak mengetahui rincian orang miskin dengan nama, alamat yang rinci, keterangan tempat tinggal dan aset yang dimiliki… Ada dua hal yang patut disayangkan dari usaha ini. Pertama, pemerintah dengan sadar memberi tugas pada BPS agar melanggar kode etik statistik internasional. Kedua, daftar orang miskin yang akan dihasilkan oleh BPS, bagaimanapun canggihnya, hanya akan memuat nama dan alamat orang yang pada waktu pencacahan dinyatakan miskin…”

If BPS or any other agencies decide to link these IDs with data gathered from individual or household surveys, not only would they run the risk of violating international ethics on statistics, they might also violate the Indonesian law on statistics.

For our practical purposes as academic researchers, having these types of data will put us under enormous responsibility and may immediately jeopardize our chance to actually be allowed to do any research (or get any research funding). While rules governing research on human subjects varies between countries and between institutions, almost all of them include strong requirement to respect and protect the rights of the subjects, including the rights for privacy and to withdraw from the research.

So clearly there will always be a trade off between knowing with certainty who the poor are and respecting the rights of the human subjects, and in my view the latter should trump the former. Besides, isn't this the sort of challenge that econometricians and economists strive for?

Aside: There are several well known micro data sets that allow one to match individuals' administrative records with the records from individual or household surveys. Some well known examples in the US are the Survey of Income and Program Participation (SIPP), and HRS (Health and Retirement Survey). Numerous safeguards are usually in place to protect the confidentiality of the respondents. The restrictions include: who are the researchers, what research questions can be asked, what can be published, at what level the data can be merged, what information should be parsed out from the data before they are used, and so forth. In some cases, noises are even added to the data as extra protection.

No comments:

Post a Comment