From 86 to 50: An Audit of the 9pm Shop-Closing Rule and Its Three Broken Links

In a cloth trader's ledger in Sindh, two numbers sit one below the other — 86 and 50. The upper one belongs to the years when the shutter came down at eleven; the lower one to the period after closing was fixed at nine. This is not one shopkeeper's diary. Across a survey of 300 retailers, average daily transactions in the top ten product categories slid from 86 to 50 — a fall of roughly 42 percent. What the Ipsos survey has placed on the table is the first hard invoice for a policy sold as an energy-saving measure.

This is not a pitch report. There is no formation here, no pressing trigger, no half-space. What exists is a regulatory intervention and the machinery built to measure it. Three decades of reporting sport have left me with one stubborn habit: before any claim, ask for the measuring instrument, and let the statement wait. So I treated this the way I treat a match — grid first, trigger second, consequence third. And drawing the grid immediately exposed three broken links: the transaction link, the energy link, and the consent link.

Evidence tier first, argument second

One thing before the work begins. The material supplied dates the policy change to 16 May 2026 and places fieldwork in August–September 2026. Both dates sit ahead of the current date. That does not mean the numbers are invented; it means I will not treat them as settled present-day fact. Read everything below under that condition: the figures come from the survey summary available, and their chronology remains unverified. Wherever I am inferring, I will say so — inference, not data.

Context: what the rule was, and what it promised

Sindh and Khyber-Pakhtunkhwa lifted the requirement that shops close by 9pm on 16 May 2026. The rest of the country kept it. That asymmetry is the spine of the study, because it created a comparable group: provinces where the rule went, and provinces where it stayed.

The stated objective was energy saving. The logic was linear — earlier closing, lower consumption, smaller bills, less strain on the grid. On paper it holds. On the ground it needed three conditions satisfied: consumption actually falls, the fall can be measured, and the fall can be attributed to the rule. All three are examined below.

The survey frame was 300 retailers, 1,046 consumers and 25 key informants. That is not a large sample, but it is enough to read a direction. A comparison of retail practice across fifteen countries is also supplied — which I read as a policy benchmark, not as proof, because climate, energy mix and working-hour patterns differ too widely for clean comparison.

Link one: the collapse in trade

The heaviest figure is the one this piece opened with. Average daily transactions across the top ten product categories fell from 86 to 50 — roughly thirty-six customer events lost per shop per day. Over a month that is about eleven hundred transactions. At the scale of a small shop, that is the slice of margin that pays rent, the electricity bill and two salaries.

Profitability ratings fell alongside it, and traders responded along three routes: opening earlier, expanding online sales, and cutting staff. None of the three is a solution; all three are relocation of the problem. Opening earlier shifts the working day forward — same hours, same wage, different fatigue. Going online adds cost: packing, delivery, platform commission and returns. And the survey itself cautions that online sales should not be assumed to fully replace in-person demand. That caution is honest and important, because convenience purchasing belongs to the customer with a card, data and trust. The customer paying a couple of hundred rupees in cash does not migrate online. He exits the market.

A structural finding sits here that most coverage will skip. Neighbourhood shops gained share while malls lost it. Many will read that as a small-trader victory. I read it in reverse. When shopping time compresses, the customer cuts travel cost — the mall is far, the neighbourhood is near. That is not the small shop winning; it is the customer adapting under constraint. Gains won from constraint vanish at the next rule change. Whether total retail spending rose or fell is not established by the summary — and until it is, treating format substitution as market growth is a mistake.

Link two: where is the energy saving

This is the sharpest turn in the whole story. Asked whether the policy's objective was met, 56 percent of traders said yes. But asked about their own electricity bills, only 41 percent reported a fall. In provinces where the rule remained in force, that figure was 31 percent.

Place the three numbers together. If lifting the rule lowers bills, the lifted provinces should show a higher figure than the others — and they do, 41 against 31. The direction is right. The magnitude is too small to call a policy success. A ten-point gap, in a country where tariffs, weather and equipment efficiency can move a bill far more within months.

And the real difficulty sits in the fifteen-point space between 56 and 41. Among a group where only 41 percent saw their own bill drop, 56 percent believe the objective was achieved. A fifteen percent bloc is calling the policy successful while its own bill did not fall. That gap is belief, not accounting.

Let me be plain, because this is the knot: a bill falling and consumption falling are different events. A bill is units, tariff and slab. If the tariff rose and consumption dropped, the bill can stay flat, or climb. The trader saying honestly that his bill did not fall is not lying — he is measuring what is within reach. That is not a flaw in the survey; it is the survey's boundary. And that boundary decides how much weight energy policy should place on this study. Meter-level consumption and bill perception do not belong in the same column, and putting them there is the error that has held this debate back for years.

Link three: consent that was never built

Ten of twenty-five key informants said traders were not consulted. Ten. Forty percent. That is not a rounding error.

An honesty clause is required here, because the easiest move in this kind of reporting is to trace every gap back to bad intent. I will not. A rational explanation exists: an energy crunch does not wait for a calendar. When grid pressure peaks in summer, administrators face pressure to act fast — and consultation is the first component to be cut from a fast decision. That may be incapacity rather than unwillingness.

But the cost hides inside it. A policy the market does not trust is complied with under compulsion, not habit. One detail from the informants points there: traders needed four to six weeks to adjust. Four to six weeks, every time timings change. Multiply that by the repeated revisions mentioned in the material. Every timing change forces roster changes, delivery-slot changes and security-guard hours to be redrawn — that is not inventory cost, it is uncertainty cost. And uncertainty cost never writes its name on a balance sheet.

Ten informants also questioned consistency of enforcement — uneven or heavy-handed by location. When the rule is uniform and inspection is not, the compliant trader is made a fool and the non-compliant one is vindicated on appeal. In that condition the policy's effect cannot be measured at all, because every district has become its own treatment.

Consumers: divided opinion, uneven burden

The consumer data is the most telling, because no public consensus formed. Thirty-five percent support early closing, thirty-two percent oppose it, thirty-three percent are neutral. Three equal parts. That is not a mandate; it is a stalemate.

Reported experience leans clearer: thirty-seven percent said shopping worsened, twenty-three percent said it improved.

But the figure that forces thought hides in the average. Women's monthly shopping trips fell 19 percent; men's fell 3 percent. A sixfold gap.

No single cause explains that, and I will not pretend one does. Evening hours open after housework, cooking and children; daytime hours are consumed by work and distance. When the shopping window shuts at nine, it shuts on the person whose only available slot was the evening. I say this as inference, not data — the summary does not explain the gap. What the summary does establish is enough on its own: the burden of the policy was not distributed evenly, and the group with the least time flexibility paid the most. Any policy declared on the basis of a single variable that produces a sixfold differential should have addressed that in its evaluation frame. It was foreseeable, not hypothetical.

One correction, in a fixed slot

Now the correction, which appears once in this piece and will not be repeated.

The debate generally assumes that closing shops at nine means cutting peak demand. That assumption may be wrong. In summer months, peak electricity demand typically arrives in the evening window when homes return and air conditioners and fans switch on — often before nine. If so, a nine o'clock closing rule does not touch the peak; it trims the half-hour after it, where the saving is lighting and cooling, not production. I state it plainly: this is inference, not verified data. Verifying it needs one input — hourly load curves by province. The summary does not carry them. Until it does, the policy is applying a lever that may not be positioned at the point of pressure.

Contrarian angle: the blind spot is not 'did it work' but 'what counts as working'

The argument ahead is predictable — supporters will cite energy saved, opponents will cite trade lost. That argument is sterile, because both sides are fighting over a policy that never fixed its own success metric.

That is the real blind spot. It was designed, implemented, lifted, and evaluated afterwards. When a policy's success threshold is not set before implementation, whatever verdict is published afterwards is not a verdict on the policy — it is a verdict on the framing. Read the trade data and you conclude failure; read the energy data and you conclude a near miss. Both readings will be internally correct.

The second blind spot is causal. A fall from 86 to 50 is a serious shock, but assigning all of it to the rule is the lazy explanation this trade has taught me to distrust. The same period carried inflation, cash stress, platform growth and seasonal shift. The control group spans two provinces with different economic structures. Some of the decline is demand displacement, some is reduced frequency, some is pure reporting bias — a trader absorbing a loss remembers it larger. The summary separates none of these.

And the third, most uncomfortable: the alternatives. The material references peak-hour tariffs and efficiency programmes — meaning other jurisdictions solve this with price signals rather than orders. Raise the tariff and let the individual decide when and how much to consume. That instrument bends the market; it does not drop a shutter. A closing order stops a person; a tariff signal makes a person think. The first looks fast, the second holds. Choosing between them is impossible without the weakest number in this survey — hourly consumption.

What to watch: three verifiable checkpoints

If this case advances, weigh it on three things, all checkable. First, does meter-level consumption data appear anywhere, controlled for tariff and weather? If yes, the argument moves to new ground immediately. Second, when timings change again, do transaction figures climb back? If they do, the rule's effect is proven; if they do not, the cause lay elsewhere. Third, does the 19 percent gap in women's shopping trips repeat in a later survey? Repetition converts it from coincidence into finding.

The lesson here is about governance rather than energy: a policy that does not fix its own yardstick ends up adopting its opponents'. And the timeline from that 16 May 2026 notification to the August–September 2026 fieldwork, still awaiting verification, will decide whether this audit belongs to a live argument or a post-mortem. Until verification arrives, treat the gap between 86 and 50 as a number, not proof. But do not ignore it, because it is written in someone's ledger — and ledgers do not lie.