Opacity of Transparency International Pakistan – by Imran Khan
Transparency’s opacity
Results from Transparency International’s “National Corruption Perception Survey 2009” are being presented in the media as an infallible judgment on the corruption of the ANP-led government of Khyber-Pakhtunkhwa. One would be tempted to accept this judgment because of the reputation of Transparency International. But a look into the details of the survey reveals significant inconsistencies which need to be taken into account before any conclusions are made.
Transparency International had had already conducted surveys in Pakistan in 2002 and 2006. Each new survey had a larger sample size: there were 3,000 respondents in the 2002 survey, 4,000 in 2006 and 5,200 in 2009. The 2009 survey differs from the other two when it comes to the provincial sample size. In 2002 and 2005, the sample size for each province was somewhat reflective of its overall population proportion in Pakistan. But in 2009, each province had the same sample size, 1,300.
A peculiarity about the sample selected from Khyber-Pakhtunkhwa was its higher monthly income. The survey had divided respondents into various income brackets, the lowest was “less than Rs5,000” while the highest one was “Rs31,000 and above.” If we assume that the respondents from the four provinces were overstating or understating their incomes by the same proportions, and then aggregating the higher income brackets, the results would be the following: for Khyber-Pakhtunkhwa more than 850 respondents, or approximately 65 per cent, belonged to the category of Rs16,000 and above. In Balochistan, there would be 40 per cent in the category, 50 per cent in Sindh and 60 per cent in Punjab. This higher income level of the sample from Khyber-Pakhtunkhwa is further borne out by respondents’ ownership of cars: 550 from the province reported owning a car, compared to 398 in Punjab, 383 in Sindh and 295 in Balochistan.
Considering the respondents from Khyber-Pakhtunkhwa to be financially richer, one would expect what can be called their bribery expenses to be higher than those for other provinces. After all, it is a common observation that someone with a higher income raises higher expectations among officials looking for bribes. For instance, tax evasion related to a higher income would be made possible by a larger bribe. Similarly, a car driver committing a violation would be expected to pay a higher bribe than a motorcyclist violating a traffic rule.
But when one looks at the actual bribery expenses as detailed on page 62 of the report, the sample from Khyber-Pakhtunkhwa reported the “lowest” total bribery expenses. If one is to divide the total reported expenses of each province by the sample size–i.e., 1,300–then we get the following measures of average expenditure on bribery for each province: Punjab, Rs17,791; Sindh, Rs16,885; Balochistan, Rs3,942; and Khyber-Pakhtunkhwa, Rs3,528.
This presents a perplexing situation. The survey clearly indicates that the expenditures of a relatively richer sample in Khyber-Pakhtunkhwa pays a far less amount in bribes, income-wise, than its comparable sample from Punjab. This, if anything, would reflect relative lack of corruption in Khyber-Pakhtunkhwa rather than the other way round.
So, what prompted this survey to declare the ANP government as the most corrupt of all provincial governments in Pakistan? The answer lies in a new question that was introduced to the survey this year. In this question the respondents were asked: Which provincial government was cleaner, the present (2008-10) or the past? If one is to rephrase this question in the context of Khyber-Pakhtunkhwa, then it would mean: Which provincial government was more corrupt, the ANP’s or the MMA’s? Seventy-eight per cent of the respondents declared the ANP to be more corrupt than the MMA.
This particular question is a verdict on a political party, and as an unqualified value judgment, it is bound to reflect the political leanings of each respondent. Given the introduction of this new question, Transparency International should have been extra-careful in making sure that the sample from Khyber-Pakhtunkhwa or, for that matter, from any other province, is a true representation of its political map.
But one is surprised to see that the four cities selected from Khyber-Pakhtunkhwa are Peshawar and all the urban centres of Hazara Division–i.e., Mansehra, Abbotabad, and Haripur. Anyone with even a rudimentary knowledge of the political history of Khyber-Pakhtunkhwa would understand the overwhelming verdict against the ANP from this sample.
The recent upheaval about the Pakhtunkhwa issue in Hazara Division is an indication of the unpopularity of the ANP in that area. If the respondents were equally distributed between the four cities then the three cities of Hazara Division would constitute 75 per cent of the sample size, almost the same as the respondents who declared the ANP to be more corrupt than the MMA. Given these facts, the selected sample from Khyber-Pakhtunkhwa is very unlikely to be truly representative of the political map of the province, and thus is very unlikely to produce an unbiased assessment.
The choice of the cities in this survey is a departure from the norms adopted in the previous two surveys. These deviations are peculiar, particularly when it comes to Punjab and Khyber-Pakhtunkhwa–i.e., the “cleanest” and “dirtiest” provinces.
A review of the cities selected in the earlier surveys would be useful here. The 2002 survey focused on the provincial capitals of the provinces. In the case of Punjab, Multan was chosen as well. The 2006 survey chose Rawalpindi and Faisalabad for Punjab and Peshawar and Nowshera for the former NWFP. The 2009 survey had room to be more diversified, as it aimed to choose five cities from each province. Interestingly, the report, on page 14, lists only four cities from Khyber-Pakhtunkhwa–i.e., Peshawar, Mansehra, Haripur and Abbotabad. Whether the survey was only carried out in four cities or the fifth was omitted because of a typo is not clear from the report. Also dubious is the selection for Punjab, which is Lahore, Sialkot, Daska, Gujranwala, and Chakwal. These are all areas where the PML-N performed particularly well in the 2008 election. The 2009 survey completely ignores southern Punjab.
The Corruption Perceptions Survey, 2009, blatantly asked about the corruption perceptions surrounding the ANP and the PML-N. For the ANP it did that mostly in areas where the party is unlikely to even field a candidate while for the PML-N it chose areas which are the party’s traditional strongholds. Therefore, the results cannot be considered an objective assessment, or even a ranking.
There are no doubts about the partiality inherent in this survey, whether this was done in error or by design. In any case, Transparency International needs to come clean on this issue. The last thing it wants is to be used as a tool in the domestic politics of Pakistan.
The writer is an economist working in Islamabad. Email: imran.khan.hks@gmail .com
Source: The News, June 09, 2010
A brilliant, well argued article. Based on the actual report itself, not hearsay.
The mainstream media is instinctively against the ANP because they are politicians and instinctively with Transparency because it is foreign-run.
Apart from the issue of statistical veracity, there is the other issue of impartiality and neutrality.
Someone at the Friday times wrote about it and its a really good read http://dashbored.wordpress.com/2010/06/19/neutral-impartial-and-independent/
“Transparency International has had a sound reputation for being impartial worldwide. We can’t say for certain, whether the problems with the Pakistan survey’s design are to serve a political purpose or are just flaws to be improved on. There is a lot of space TIP has to cover. Addressing corruption risks requires an understanding of the local political economy and the power structures that control access to resources or beneficiaries i.e. ‘gatekeepers’ who can be very influential on the research program being carried out!”