Pakistan's sports goods output: reading the July 2026 Large Scale Manufacturing bulletin with data, not instinct

On Wednesday afternoon, the Pakistan Bureau of Statistics (PBS) pushed its provisional bulletin into the system. I read it the way I always do: large scale manufacturing, the Quantum Index of Manufacturing, year-on-year growth, month-on-month growth, then sector by sector. It was only near the bottom of the table, between wearing apparel and leather products, that I stopped at an entry I did not expect to find in an industrial statistics release: other manufacturing (football), down 0.22 percent year on year. What made me stop was not the size of the number. 0.22 percent sits inside the noise band of any monthly index, and I had no intention of building a story out of it. What made me stop was its position: a football manufacturing line item, listed in a national index alongside automobiles, pharmaceuticals, cement, steel and machinery. For someone who works the touchline for a living, that is the kind of detail worth writing down, because it shows the layer at which sport touches a country's production base. I opened the bulletin a second time. Then a third, close to midnight, after checking two independent sources. Numbers do not lie. It is only that we have to ask the right question. With an industrial index, the right question always begins with something dull: does the top of the document reconcile with itself. My notebook has one rule that has become reflex: every fact needs a source, and every source needs a timestamp. This bulletin satisfies the first condition and fails the second in one place. I record the gap before I record any judgement, because in my trade the order of note-taking matters as much as the content. Two divisions, one fiscal year, and one word: provisional PBS published provisional data. The Quantum Index of Manufacturing for July 2026 stood at 119.13 points. A year earlier the index was 115.62. In June 2026 it was 108.78. Two divisions produce the two figures the bulletin used as its headline: up 3.03 percent year on year and up 9.51 percent month on month. I checked both. 119.13 divided by 115.62 equals 1.03035. 119.13 divided by 108.78 equals 1.09515. Both reconcile to the second decimal place. That is the first thing I do with any data release: if the top does not reconcile, the rest is not worth analysing. These two reconcile, so the rest is worth reading. The bulletin uses the phrase the July 2026-27 period. That phrasing is not standard in international statistical practice. The most reasonable reading is fiscal year 2026-27, with July 2026 as its first month, meaning one month of data rather than twelve. I noted this in the margin, because misreading it as a full year would distort every downstream comparison twelvefold. One detail needs stating immediately: the bulletin does not name the publishing outlet. The underlying data source is PBS, a state statistical agency with a published methodology and a publication calendar. The publishing source is blank. That means I can assess the reliability of the primary data, but not the editorial process of the intermediary. For me this belongs in the column marked to be verified, not the column marked wrong. And the data is provisional. The word provisional in industrial statistics is not a courtesy. It is a technical warning: these figures will be revised in the next release, and the revision can change the whole picture. Any product that quotes them without a timestamp is creating risk for itself. I say this from experience: a provisional revision of two tenths of a percentage point sounds small, but if it lands on a high-weight sector it flips a headline. Why would a touchline reporter read Pakistan's industrial index? Because sporting equipment has a physical existence. The ball, the shirt, the roll of grip tape, the gloves, the padding: all of them come off a production line, through a port, onto a container. Sports equipment is the least discussed part of sport, and the part most dependent on numbers like these. For me it is the only way to answer a question nobody asks on television: where does the object in an athlete's hand come from, and at what price did it arrive. There is something I have to remind myself of constantly: some things only appear when you are willing to sit still for longer than one set. A monthly index is one set. It is not enough to conclude anything about a season. Two ends of the same chain, moving in opposite directions Across the whole table, the sectors directly tied to sports goods and apparel produce four lines worth reading separately. Wearing apparel rose 3.87 percent year on year. Textiles fell 0.45 percent. Other manufacturing, which includes football, fell 0.22 percent. And the chemical products group sits at a very small increase, 0.25 percent or 0.50 percent depending on the extracted line. The most notable feature is the gap between the two ends of the chain. In a garment chain, yarn and fabric sit upstream; cut-and-sew sits downstream. When upstream falls and downstream rises in the same period, there are three readings and all three are defensible. First: demand at the finished end is still healthy, and apparel orders are still coming in. Second: margins are being squeezed in the middle, because input prices are not falling while output prices are not rising in step. Third, and the one I treat most cautiously: the downstream rise may be built on inventory or on cheaply bought inputs rather than on new orders. For sports goods, all three readings matter, because this is precisely the segment where kit, training wear and fan retail live. A playing shirt is not a luxury product, but it is a cyclical one: a manufacturer must book fabric months before the season, cut in batches, and deliver against a competition calendar. If yarn and fabric fall while cut-and-sew rises, finished shirt prices are unlikely to fall over the next two quarters, because mid-chain costs do not simply disappear. That is my reasoning, not data from the bulletin, so I grade it at medium confidence and will retest it against the next release. The football line deserves separate treatment for a technical reason. It sits inside other manufacturing, which is a residual bucket. A residual bucket is where small industries are pooled so the index does not run to infinite length. The consequence is that very small percentages inside a residual bucket can conceal large swings in a narrow manufacturing cluster. Pakistan's sports goods cluster, centred on Sialkot, has long been an important node in the global sports equipment supply chain, particularly in hand-stitched balls and protective gloves. A 0.22 percent decline in a single month says nothing about that cluster. It says only that the cluster did not grow in July 2026. Leather is another line I keep in the notebook, because leather is an input for gloves, for bat and racket grips, and for part of the protective goods range. The bulletin places leather among separately monitored categories but does not attach a growth figure in the dataset I hold. When a category appears in a table without a number, I treat it as a gap, not as zero growth. The difference between missing data and zero data is the difference that ruins a great many analyses. I grade the direct relevance to tennis equipment as low confidence, and I say so plainly so readers know which part is data and which is inference. Tennis balls need rubber and felt. Grip tape needs chemicals and fabric. Playing apparel needs synthetic yarn. Pakistan is not the centre of all of those, but it sits inside the same price system as the centres that are. When one node changes rhythm, other nodes rarely stay still; they simply lag by a quarter rather than a week. Non-metallic mineral products and iron and steel deserve a mention in this same passage, even though they sound distant from sport. Cement and steel are inputs for stadiums, training grounds, stands and even the frames of training equipment. Iron and steel falling 0.47 percent year on year is a signal about infrastructure investment cost, not a signal about demand for watching sport. But the two meet at one point: a venue project that gets more expensive gets pushed back, and a project pushed back reshuffles equipment leases and equipment imports. Chains like that never appear in a monthly index, but they begin there. Automobiles surged, and ten sectors went backwards The most quoted line in the bulletin is automobiles. The extract records two increases for the same period: 57.01 percent and 57.77 percent. I place this pair in the group that must be verified before use, because two different figures cannot both be right for one measurement window. Most likely one is the monthly growth rate and the other is fiscal-year-to-date, but the extract does not separate the periods. The more important issue is how to read it. A 57 percent increase for a sector in a single month rarely reflects an explosion in demand. It usually reflects a low comparison base in the same month a year earlier. This is the most common misreading of any index, from industrial output to match statistics: confusing speed with strength. A player returning from injury can raise his points total by 60 percent in his first month, simply because he did not play the month before. Nobody calls that peak form. The broader picture matters more than automobiles. The bulletin records ten sectors declining year on year, including textiles down 0.45 percent, pharmaceuticals down 1.24 percent, food products down 0.84 percent and iron and steel down 0.47 percent. The headline index rose 3.03 percent, but that increase rests on a narrow base: a few heavy-weight sectors pulling the index up while most others slid. The Quantum Index of Manufacturing is a weighted index, meaning larger sectors contribute more to the headline. A rising headline does not mean all sectors are healthy. Understanding the weighting mechanism avoids a very common error. Each sector's weight is fixed against a base year, and when the base-year update cycle moves to a new period, the ranking of contributions can change even if physical output does not move at all. In other words, some shifts in the headline index are shifts in the ruler, not in production. For anyone reading an index to decide raw material purchases, this is the most expensive trap, because it triggers no warning. For the sports equipment supply chain, narrow-based growth means that supplies of fabric, yarn, gloves and accessories will not improve across the board. Some links will clear; others will stay congested. When I ask brand procurement teams about prices, they answer item by item, not by national index. That is why figures like these carry directional value and not conclusive value. Three small numbers cannot be growth rates This is where I think many reports will misread the data, and it is the point I most want to make. The extract records several sectors with very small increases: 0.01, 0.03, 0.04, 0.11, 0.18, 0.21, 0.27 percent. Read as year-on-year growth rates, these values are almost impossible in a month when the headline index rose 3.03 percent. A sector growing 0.01 percent is effectively flat; if several sectors are flat and several others are falling, the headline can hardly rise by more than three percent. The most likely explanation is that these small values are weighted contributions to headline growth, not sector growth rates. PBS publishes both tables. The extract flattened them into a single list and applied one label to all of them. The consequence is concrete. If a downstream system mixes the two metric types, it will understate genuinely growing sectors with small weights while misjudging the role of large-weight sectors. In an index used to make purchasing decisions, that error can turn into the wrong contract. I have seen a smaller version of the same mistake in sports: a statistics table that mixed second-serve points won with total points won, which left a player completely misjudged on his ability to hold serve. This kind of error is not loud. It simply tilts everything slightly, and then keeps tilting it. I mark all of these values with a single note in the notebook: metric type unverified. Four pairs of figures that contradict each other Beyond the metric-type problem, the extract contains places that do not reconcile internally. Automobiles appear twice, at 57.01 and 57.77 percent. Furniture appears twice, at 22.69 and 10.10 percent. Chemicals appear twice, at 0.25 and 0.50 percent. Tobacco appears twice, at 35.82 and 0.55 percent. The tobacco pair is probably the easiest to explain: one figure is the fiscal-year-to-date number and the other is the single-month year-on-year number. But that is my inference, not information in the document. One line suffers a concatenation error: non-metallic mineral products is recorded as growing 6.52 percent and 4.25 percent in the same sentence, with no connective. The most reasonable reading is that 6.52 percent is the growth rate and 4.25 percent is the contribution to the headline index. Again, inference. There is also a small spelling error in the computer, electronics and optical products line, where the word computer is missing a character. It does not affect the analysis, but it is a signal about the data entry stage, and to me every signal about the data entry stage is worth recording. None of this is a PBS error. It is an extraction error: a table with two metric types and two time windows was flattened into one list. Once a table is flattened, no downstream reader can distinguish rate from contribution, or month from cumulative. That is why I always preserve units and periods when I take notes, even though it makes my notebook visibly uglier. I do not remember what I wrote. I remember what I counted. And I still have to repeat the most basic point: this is provisional data for one month. One month is a thin sample. No conclusion about a trend should be drawn from it. Could not be established is the sentence I write most often in this trade, and it is not an evasion. It is an accurate conclusion about the state of the data. An industrial document labelled as sport Something happened during processing that I think is worth recording, because it bears directly on how sports news operations work. When this document entered an automated classification system, it was tagged to the sports domain, specifically tennis. Nowhere in the text is there a player, a tournament, a federation or a match. There is exactly one sports keyword: the other manufacturing (football) line. That is very likely what triggered the wrong label. The entity extraction step returned nothing, and that is actually a good sign: it did not invent a player or a tournament that does not exist. I raise this not to criticise a specific system. I raise it because this is the same error I see weekly at a smaller scale: a data line attached to a story only because it contains a keyword. A transfer fee attached to a player who never signed. An injury attached to a tournament that already ended. A wrong label at the first layer follows the document to the last layer, and by the last layer nobody checks again. Fans have the right to live in emotion; my job is to live in data. Since 2026, when I nearly published a wrong transfer fee because I trusted a single source, I have set a rule with no exceptions: never publish anything on one source, and always cross-check against the primary document. That time, the transfer registration record showed a lower figure than the rumour circulating, and the official announcement two days later confirmed the record's number. I learned that waiting long enough is always cheaper than correcting. Applied to this document, that rule means: I use only figures that reconcile, I state the measurement period, I name PBS as the primary source, and I say plainly where things remain unverified. Much of the value of a data piece lies in the places where it is willing to say it does not yet know. Internal signals to watch next If one thing is to be drawn from this bulletin for the sports goods sector, it is four signals to track, and I deliberately leave them as questions rather than conclusions. Whether the 3.87 percent rise in wearing apparel holds for three releases or more. One month up is news; three consecutive months up is a trend. A new line-up, like a new clock, needs time to run true. A sector that has just turned positive needs the same. Whether textiles return to positive territory. If downstream rises while upstream keeps falling across three periods, the margin-squeeze scenario becomes the more credible one, and the consequence is that finished goods prices will struggle to fall even if demand recovers. For sports goods buyers, this is the signal that decides next season's pricing. Whether the football line is separated out of the other manufacturing bucket. This is a change in disclosure, not in production, but it is the condition under which anyone reading the sports equipment supply chain can see the Sialkot cluster in the data rather than having to infer it. As long as the line sits in a residual bucket, any analysis of it remains indirect. And the last signal, the most important one for anyone in the verification trade: the next PBS release. Provisional figures will be revised. If the 3.03 percent headline is revised below three percent, then every sector-level increase needs to be re-read from the beginning. I have filed this bulletin away, marked with its period, its primary source and three unverified points. Someone will read the 57.01 percent automobile figure and write a big headline. I will wait for the next release, and the one after, before writing a single sentence about it. Slow and steady is not a slogan in my notebook; it is the only way to avoid having to correct yourself. What I want to leave at the end of this piece is not a forecast. It is a way of asking: when an industry that serves sport appears in a national index as exactly one line, the question worth asking is not whether that line rose or fell, but who will track it long enough to know when it has genuinely turned.