What is state failure? See my conceptualisation of state failure on the right flank below.

Tuesday, August 28, 2007

Those mysterious footprints, I mean footnotes, of UNODC

MStFB Poppy Series update, lost in basic calculations
There's plenty to reflect on in UNODC's Afghanistan Opium Survey 2007. And I will try to digest it more thoroughly in the future. Right now I will just note that I'm somewhat puzzled by what I read on pages 6 and 7*.
Why is the definition of a poppy-free province different for 2007? I see no conspiracy here, I just don't understand why it is so. With the 2007 definition you would have had eight provinces falling into the poppy-free category in 2006. That's because according to footnote 4 on page 7* of the 2007 report, the "Definition of opium poppy free status in 2007: province with less than 100 ha of opium poppy", and so, beside Ghazni, Logar, Paktia, Paktika, Panjsheer and Wardak, in addition Kabul (with only 80 ha) and Bamian (with only 17) would also have been poppy-free in 2006. Of course Kabul would be lost by now (at 500 ha for 2007), and that means that instead of the "we have seven more poppy-free provinces" success story, one could only say it's a "we have five more poppy-free provinces all in all, but that in fact is six more, while we have of course lost one from last year, and it happens to be the most central province, you know" sort of success story. Definitely, the latter doesn't sound so good, but let me repeat that I really don't see any conspiracy here. I just mention this because I don't know what's the point in writing a paragraph into a report saying there are 13 poppy-free provinces this year, as opposed to 6 last year (see this on page 7*), when the thing we are talking about was defined differently for the two years surveyed.
And I have one more question in connection with the data on pages 6-7. All thirteen provinces that qualify as poppy-free for 2007 are said to have had zero area with poppy plantations. None of these thirteen fall just anywhere in the interval between 0 and 99, which is of course the interval statedly relevant by definition, instead they are all right at zero. So for example Kunduz registered a 102 ha drop in poppy cultivation, and went from 102 ha in 2006 to 0 ha in 2007. Similarly, Parwan went from 124 to zero, and Khost went from 133 to zero. So has a drop below 100 automatically allowed a province to be booked as recording 0 ha of land under poppy cultivation? You know why I'm asking this. If some of the thirteen provinces registering zero ha under poppy cultivation are actually somewhere between 0 and 99 rather, that might mean a very different figure for the number of newly poppy-free provinces according to last year's definition (absurdly, that may even be a negative number, using last year's definition). I don't think this is deliberate manipulation, since one can fabricate any number of success stories out of almost any kind of figures - it only takes arranging them into a desired order - but it may be necessary to zero in on those zeroes to arrive at refined conclusions.
It's not the first time a footnote creates a bit of a mystery about what the UNODC says. Joel Hafvenstein noted in a comment to this post by Carl Robichaud at Afghanistan Watch that while UNODC earlier this year alarmed people that opium processing now takes place largely in Afghanistan, unlike earlier, in fact in their 2005 opium survey on the country they already stated that "seizures in Afghanistan and neighbouring countries (2002-2004) suggest that – expressed in heroin equivalents – 71.5% (of poppy) is transformed into morphine and/or heroin in Afghanistan". So the change can't be assumed to have been that great for the period 2005-2007 - if all these estimates would be right, the change would have been a rise of 18.5%, given the claim that now some 90% of processing takes place in Afghanistan (my quote from the 2005 survey is from page 54 which is where I found what the above mentioned commenter remarked; the latter person claimed to remember this from a footnote that I didn't manage to bump into, so that footnote remains even more of a mystery to me).
You could say that a change of 18.5% is significant change, but consider this. AFP reported, citing UN sources, when it heralded the publication of the 2007 World Drug Report, the following (with this particular AFP report currently available on the UNODC website): "Sophisticated laboratories inside Afghanistan are now converting 90 percent of the country's opium into heroin and morphine before smuggling it around the world, the United Nations said Monday. Afghanistan, the world's biggest producer of opium, had until two years ago exported the illicit drug almost exclusively in its raw form, said the United Nations Office for Drugs and Crime (UNODC)." (Highlighting by me - P.M.)
As I said, it's really interesting to trace all these footprints, I mean footnotes, at times.
* The referenced page numbers are from the executive summary, and not the full version of the Afghanistan Opium Survey 2007.


bonnie boyd said...

Dear Peter,
I too have been reading through the UNODC report and wondered about the changes. However, the one I noted most was that poppy cultivation and distribution was reported "stable", although poppy crops have gone up and more states are seriously affected.
For the "stable supply" result, I think the problem isn't exactly a conspiracy but more that many of the states receiving the crop are not states that report--and therefore, stats look "stable" but aren't.
For the problem you mention, which is related by inattention at the least: Afghanistan's provinces "under cultivation" estimates may be based upon levels of effort expended rather than an actual survey. In other words, Kabul has x square ha of field, minus y square ha of eradication equals less than 99. But poppies will re-seed naturally, or be re-sown by human agency. Therefore, simple subtraction formulae are useless. But for institutional purposes, one can say that there is less opium in one report and acknowledge that there is more opium overall, without feeling internally inconsistent. And yet--if my guess is true--it is in reality a statistic which does not take into account the nature of its subject.
Thank you for writing on this, because I thought I was alone in thinking the UNODC report does not give universally useful data.
Oh, and another thing: they have not updated the stats in the index for Central Asia at least: therefore Kyrgyzstan in text is given a higher addiction rate than it is in the index.

Péter MARTON said...

I see, thanks for all the insight, Bonnie.
Actually, I suspect that UNODC's staff might be quite overburdened writing these highly detailed reports, so I guess it is just normal to expect errors creeping in at times.
In the known circumstances, e.g. what they have to deal with in Afghanistan, methodological perfection and 100% attention to detail just has to be forgotten, one has to suppose. Still there is this feeling in me, that the suggestion of a more decisive change from last year in the north, claiming to have now seven instead of just five more 'poppy-free' provinces does fit nicely with one obvious way to interpret UNODC's message that the south is now the central location for Global Drugs Inc., i.e. that the south is ripe to be hit now with something like a precision strike. UNODC - they stress there being a clear north/south contrast in their press release on the survey, which is available here.
And the other thing that came to the surface in the discussion over at Afghanistan Watch, the apparently mistaken assumption that the ratio of raw opium processed inside Afghanistan grew so significantly over the last couple of years, may also fit with that thesis.
I do think these are inconsistencies by mistake (a conspiracy would need more of a distortion of facts than claiming there to be seven instead of five more new poppy-free provinces), but in the end they may turn out to have consequences in reaffirming assumptions and currently perceivable impulses regarding what to do.