OVER A COFFEE : Perils of sloppy use of IT — Dr Haider Shah
Just as in posh restaurants where we often remain unaware of the substandard hygienic conditions of the kitchens where food is cooked, our senior managers take nicely printed reports with colourful graphs on them at face value
If manmade energy is the distinguishing line between modern and old eras, it is information technology (IT) that separates the knowledge economy of today from the industrial economy of yesterday. There is, however, also a need for appreciating the limitations and risks of IT-based decisions, especially when there exists a wide disconnect between policymakers and IT professionals. While there is a wide body of literature on this issue, I am making this suggestion on the basis of some personal experiences in the field. Today, when the National Database and Registration Authority’s((NADRA) database-based voting verification debate is making headlines, I thought it appropriate to share my concerns.
In 2009, I was entrusted with the job of looking after preferential trade agreements in the Federal Board of Revenue (FBR), Islamabad. One morning, I was asked by the then member customs to prepare a summarised report of Afghan transit figures for the last five years within an hour. It transpired that a senior official in the commerce ministry needed the report as a part of strategic dialogue related discussions, including the new Afghan Transit Treaty. I hurriedly went to the office where such data was maintained. I asked the concerned clerk to give me the required report but he did not seem very enthusiastic as he was already up to his eyeballs in getting printouts from an aging computer on directives received from all imaginable quarters. Faced with the ‘baarbaarteredar per jaatahoon’ (I keep coming at your door) situation, I finally prevailed upon him to print out annual figures and an overall summary report. When I examined the figures and the summary report I noticed that the column format had been changed in one particular year and hence the summary report was based on wrong figures added together. I mentioned this to the clerk who, after some unsuccessful defence, agreed that the annual figures were not comparable and hence the summary report was misleading. I reported my concerns to the member who told me that there was no time for such Sherlock Holmes queries and to fax the report to silence the guns that were constantly firing for receipt of information.
In the more distant past, when I was a young officer in the customs department in Peshawar, I shared my proposal of generating a database of imported vehicles so that the verification of genuinely imported vehicles could be made easier. I visited the Karachi customs house and discussed the idea with the chief executive of the automation project. Later, when I was posted in Karachi I took up the initiative again with the new and energetic head of the automation project. After a few months,I was given a CD of the database of all vehicles imported through Karachi along with an assurance that the database was very authenticas no single entry had been left out. I was very happy to see a dream coming true and thought that now it would be very easy to single out smuggled vehicles with the help of the database contained in the CD. The proof of the pudding is in the eating. I entered the chassis number of my official vehicle and a few more that were known to me as legitimately imported vehicles. Surprisingly, they didnot show up in the database that had been prepared by the IT experts of one of the most modern organisations of Pakistan. I went into Sherlock Holmes mode and finally found the reason: the import documents from which data had been copied were not uniform in making declarations about imported vehicles,resulting in chassis numbersappearing at different places in the documents. This was not picked up by the data entry clerks. I shared my observation with the young executive who quickly understood the blunder. As was required, he promised redoing the whole project from scratch to make the database reliable.
There are countless examples that demonstrate how poor and untrustworthy our data collection and reporting standards in public sector organisations are. There is no culture of maintaining data integrity with the help of regular internal and external audits of IT systems, data collection and reporting methods. Just as in posh restaurants where we often remain unaware of the substandard hygienic conditions of the kitchens where food is cooked, our senior managers take nicely printed reports with colourful graphs on them at face value. It is therefore important that use of IT-based verification systems should be subjected to strict adequacy tests before they are used for important public policy matters.
In this backdrop, the present euphoria about fingerprints verification is potentially disastrous. Even in a small-scale research survey, it is always advised that some pilot testing should be done before launching the survey questionnaire fully. Using IT-enabled techniques for verification of votes is a good idea and must be keenly pursued. However, this should not be applied unless it has been found working satisfactorily at a pilot testing stage by independent auditors. The election commission should have used the system in a carefully chosen test case. It should have been applied in one constituency as a pilot study. Once all teething problems were removed and the system had proved its authenticity, only then it should have been allowed to rollout as a trustworthy verification system. Without this homework, using a fingerprint database amounts to opening a Pandora’s box. With multifarious challenges of internal and external security, we can hardly afford new controversies.
The writer teaches public policy in the UK and is the founding member of the Rationalist Society of Pakistan. He can be reached at email@example.com