The curious case of Punjab’s missing cotton bales

In July 2019, the federal tax bill imposed 10 percent sales tax on ginned cotton. Later same year (FY20), variance between cotton arrivals, as reported by Pakistan Cotton Ginners Association - a 1,200+ members strong association - and, cotton output as reported by official publications - shot up to 24 percent in Punjab, up from just three percent the previous year. Coincidence?

BR Research will not shy away from admitting that the insinuation above runs counter to the position argued earlier in this space; that Pakistan is witnessing one of its worst cotton output years in history. That position remains intact, considering federal Cotton Crop Assessment Committee’s projection of 7.7 million bales (second estimate, announced in Dec-20), is worst since at least 1987. But consensus market estimates, based on discussions with farmer representatives and seed companies place the output at much lower. How lower? Market expectations find an anchor in the form of USDA, which places national output for the 2020-21 season at 5.8 million bales, nearly 2 million bales lower than official figure!

That figure is corroborated by PCGA’s fortnightly cotton arrival reports, which places cotton bales arrivals at 5.7 million bales thus far, with ginning season basically over. PCGA’s biweekly report highlights that the estimates presented are supplied by member ginning factories but are also ratified by Karachi Cotton Association (KCA), as well as All Pakistan Textile Manufacturers Association (APTMA), the mothership of all lobby groups within Pakistan’s largest industrial segment.

It is pertinent to note that the 2 million bales difference between official and PCGA estimates is unlikely to have been finalized yet. Official enumerators may further revise total cotton output down by an odd 10 percent, before the national GDP estimation exercise concludes in April. This may take final (officially) reported figure down to 7 million bales. That would still leave 1 to 1.3 million missing bales unaccounted for – a gaping hole between (a possibly further revised) official estimate, and figures reported by private sector bodies.

Ask representatives of the private sector, and they would point to the embarrassing incompetence across public sector bodies, whether it is the Provincial Crop Reporting Services Departments (CRD), Pakistan Central Cotton Committee (PCCC), Textile Commissioner’s Organization (TCO), or the office of Cotton Commissioner. These bodies lack capacity to conduct any scientific surveys of crop acreage it is argued, let alone ability to calculate yields achieved, or losses due to extreme whether events or pest attacks. Private sector estimates are all that may be relied upon, we are told, as official estimates simply apply percentage changes to historic data, and use averages to fill in for missing figures in between.

But if official estimates are really that bad, why is it that the variance between PCGA’s and CRD (Punjab’s)* reported figures used to be negligible until just two years ago? To understand the anomaly better, BR Research conducted a district-wise analysis of aggregate cotton arrivals as reported by PCGA by season-end over past 6 years and compared it with cotton output (final) estimates as reported by Government of Punjab’s Agriculture Department. Note that the official figures are initially reported by provincial CRD but are eventually communicated through several official publications by other bodies such as PCCC, TCO, MNFS&R, MoF’s Annual Economic Survey, and SBP’s State of the Economy. (Note: The district wise analysis is restricted in its scope to Punjab, as micro-data for Sindh is released by several years’ delay; similarly, the analysis could not be conducted for years FY15 or earlier as PCGA arrival reports are unavailable for the period).

For four years between FY16 – FY19, variance between final cotton arrival figures as reported by PCGA and CRD-Punjab averaged at just 5 percent, dropping to as low as 1 and 3 percent in FY17 and FY19, respectively. In FY20, variance shot up to nearly one-fourth overnight, the same year that the federal government brought cotton ginning under the purview of GST regime. Although rumours of under reporting have been rife ever since, the same have been – rightfully - overshadowed by the massive decline in national cotton output, which, after all, poses a far more ominous and urgent national challenge.

Out of 36 districts in the province, data from CRD-Punjab shows that cotton production takes place in as many as 24 districts of the province, while 12 districts of the province have nil cotton production. Out of the 24 cotton producing districts, 11 districts of southern Punjab (Multan, Bahawalpur, and D.G. Khan divisions) account for over 90 percent of provincial cotton output, while remainder 10 percent is sourced from 9 districts in Sahiwal, Faisalabad, and Sargodha divisions. Four districts of the province report negligible output. In contrast, PCGA only reports cotton arrivals for 21 districts, as only these are home to ginning factories in the province (four districts with negligible output are not included). To avoid visual clutter of a district-based infographic, year-wise data has been clubbed at division level in the accompanying illustrations. Now that the context is set, some interesting outcomes of the exercise:

One: cotton arrivals in ginning factories of Bahawalpur division (Bahawalpur, Rahim Yar Khan, and Bahawalnagar districts) have historically far outstripped indigenous cotton production in the region. However, that variance has reduced from historic average of 24 percent more arrivals than regional cotton output, to just 5 percent in FY20 (CRD-Punjab’s district-wise data for FY21 is so far unavailable).

Two: historic trend from Multan division seems to reverse-mirror the trends in Bahawalpur division, but only until FY19. Back then, while ginning factories in Bahawalpur reported more cotton arrivals than indigenous output (until FY19), factories in Multan reported 42 percent less arrivals than regional output. This indicated that cotton output from Multan division used to find its way to ginning factories in Bahawalpur. So far, all kosher. Except, during FY20, Multan’s arrivals versus output variance substantially widened, even as factories in Bahawalpur region noted a visible decline in arrivals. Where is all the extra cotton produced in Multan division (as per CRD-Punjab) now headed, if not to factories in Bahawalpur area?

Three: on pan-province basis, the decline in cotton arrivals (as reported by PCGA) is far more severe than fall in cotton output (as reported by CRD) and seems to have only worsened in FY21. Ginning is a cottage-based industry, thus under rising input costs and increased competition for higher priced cotton, incidence of under reporting may not come as a big shock, worsened only by the hammer of GST and documentation drive. But how high really is the extent of under reporting? There is no telling.

Why is it important to find answers to these questions? Because from public to private sector bodies, nobody seems to have a clue as to the market size of national cotton consumption, and whether the same is on an exponential rise due to revival of textile exports or has undergone a substitution due to greater production of mixed yarn. Pakistan has already imported 3 million cotton bales in 8MFY21 thus far. The missing 1 to 2 million bales can make a world of a difference in reaching a decision whether the textile sector even requires Indian-origin cotton on an urgent basis. Before the federal cabinet makes more mockery of itself by taking a U-squared on its earlier U-turn, some research, please?