Skip to contentRuka hadi maudhui
KCP.KCP.

Chapter 11 · Weekly move · Week 32 · 2026 · dated public record

KCP transformation

Five rose, fourteen fell — and the book had its best week

Across the eight sessions since the last note, 14 of the 19 priced counters fell — and the model book still had its strongest week since inception. Roughly 85% of that gain came from one holding. A number that good, from a market that narrow, is a reason to look harder, not to add. The book holds a fifth week, and writes down a date it will not be surprised by: NMB's one-for-ten subdivision, effective 24 August.

Chapter 11 · Weekly move · Week 32 · 2026

Model-book decision

Allocation maintained4 August 2026 · 4 min read
01EvidenceThe tape
02RuleMandate bands and falsifiers
03Model-book allocationAllocation maintained
04WatchWhy hold — a fifth week, and this time for a different reason
Model-book allocation
Allocation maintained
Hypothetical NAV
TSh 554.6m
Market P&L · money-weighted
▲ 8.06%
Equities · band 30–60%
52.8%
Fixed income · band 25–50%
33.2%

Frozen to the dated public record · observed figures are not silently replaced by later closes.

01 · The tape

What the week actually said

From the last note's mark to this one — 24 July to 3 August — the All-Share rose 0.89% to 4,187.22 and the domestic Tanzania Share Index rose 2.36% to 9,200.17 (DSE, 3 Aug close). Underneath those two green numbers, five counters rose and fourteen fell. Decliners beat advancers in every single session of the stretch — 27 July through 3 August, six for six. The All-Share set a record of 4,194.87 on 30 July and has since eased 0.18% off it.

Turnover told the same story from the other side. On 31 July, TSh 30.46bn traded and TBL alone was 78.1% of it; on 3 August, TSh 7.10bn traded and CRDB plus NMB were 94.3%. Money arrived. It went to two or three doors.

The model book's own week reads exactly like the market's. Vodacom went from TSh 825 to TSh 1,000 (+21.2%) — and unlike the zero-volume cross-listed prints that made July's record, this one was funded: 203,701 shares changed hands on 3 August alone. NMB added 4.90% to TSh 17,570, printing a fresh 52-week high of TSh 17,580 on 31 July. Against them, CRDB eased 3.35% to TSh 2,600, Twiga (TPCC) fell 5.52% to TSh 5,820, and TCC was near-flat at TSh 12,300. Every mark reconciles to the verified DSE close.

The book ended at TSh 554.6m, +8.06% money-weighted since inception on 6 June, against TSh 514.0m of contributed capital. That is the strongest week the book has had. It is also the week it should be most careful about: of roughly TSh 17.2m of model P&L added since the last note, about 85% came from Vodacom alone. Deposits — including the TSh 7m added on 1 August — are capital, never performance.

01 · Drift vs the bands

Still nothing to correct

Equities 52.8% (band 30–60), fixed income 33.2% (25–50), cash 10.2% (5–20), gold 3.9% (0–8). A fifth straight week mid-band. Rotations are usually forced by drift; there is none to force one here.

02 · The concentration

A third of the book, two counters

NMB (17.7%) and Vodacom (15.0%) are now 32.7% of the model book between them. That is inside every band the mandate sets — because the mandate sets no single-name cap. This week the concentration paid. Naming that gap is the honest part; deciding what to do about it is the next chapter's work, not a same-day reaction to a good number.

03 · The curve

The auction we were waiting on hasn't printed

Last week the book said the next T-bill auction was the read that mattered. As at this mark, the newest result in the Bank of Tanzania series we read is still auction 1203 (15 July) — 91-day 3.35%, 182-day 4.78%, 364-day 7.03%. The front end therefore still sits ~290bp below the 6.25% policy rate and below 4.0% headline inflation. No new evidence means no new decision.

02 · The filings

Earnings week — including one the book got wrong

NMB filed H1-2026 group results on 3 August: net profit after tax TZS 405.8bn, up 13% year on year from TZS 358.6bn, with total assets at TZS 18.4tn (+26% YoY). That is the earnings base under the book's largest equity allocation, and it is intact. CRDB filed its quarterly report for the period ended 30 June on 30 July, and TCC filed on 3 August — but the text of both filings was not machine-extractable, so this note carries no figures from them. An unreadable filing is not a bad result; it is simply not a number yet, and the book would rather say so than estimate one.

The uncomfortable one: Tanga Cement (TCCL) posted Q2 revenue up 10% quarter on quarter to TZS 79.4bn, operating profit up 19% to TZS 14.6bn, and quarterly group net profit of TZS 9.4bn against TZS 3.8bn — and its share price rose 39.1% over the same stretch the book's cement holding, Twiga, fell 5.52%. The Ghost owns the one that fell and not the one that rose. Conviction cuts both ways, and it cut this way this week. That is worth writing down plainly rather than quietly leaving out of a week the headline number flattered.

One more observation, offered as a fact rather than a diagnosis: the largest decline on the board over the stretch was MCB, from TSh 660 to TSh 285 (−56.8%), sliding in every session. No issuer filing explaining it appeared in the window this note reads. The book holds none of it, and the reason it holds none of it is the same reason it will not guess at a cause here: it cannot source one.

03 · The decision

Why hold — a fifth week, and this time for a different reason

The first four holds were holds of patience: nothing had changed. This one is a hold of discipline: something did change, and it changed in the book's favour, which is precisely when a model book is most tempted to confuse luck with skill and add to what just worked.

Three things argue against acting. Every sleeve sits mid-band, so there is no drift to correct. The one dated thesis on the table — extending duration while the 364-day sits above 7% and the front end is pinned below inflation — still waits on an auction result that has not published in the series the book reads, and a thesis without new evidence is not a decision, it is an itch. And the week's gain was a single name doing 85% of the work, which is an argument for examining concentration, not for feeding it. Nothing was bought, nothing was sold.

04 · The weeks ahead

One date the book will not be surprised by

NMB has an approved one-for-ten share subdivision, effective 24 August 2026 (issuer notice, approved 24 July, announced 27 July). The published calendar runs: 19 August last cum-split trading day, 20–21 August trading suspension, 21 August register closure, 21–23 August register update, 24 August effective and trading resumes on the new basis. A subdivision multiplies the share count and divides the per-share price by the same factor: it changes the units, not the value, and not any holder's proportional ownership.

Saying so now matters for one practical reason. When the new basis takes effect, the model book's NMB mark will step from a ~TSh 17,000-order price to a ~TSh 1,700-order price on a single day. That is a mechanical basis break, not a loss — it will be labelled as one when it happens, and the book is flagging it three weeks early so nobody has to work that out from a chart.

Otherwise the discipline is unchanged: watch whether the next bill auction reconnects the front end toward the 6.25% policy rate; watch whether breadth and turnover broaden past two banks; and hold what the thesis owns until something other than a good week argues otherwise.

— The Ghost. Hypothetical and educational only — a model book, not real money, and never advice.

Model-book decision principle

A very good number produced by a very narrow market is not proof the thesis worked. It is a reason to check whether the book is being paid for judgment or for concentration.

Dated evidence trail

Inspect the terrain behind the chapter.

Source: DSE daily report (verified snapshot, 3 Aug) · DSE issuer filings (NMB H1-2026; TCCL Q2 2026; NMB subdivision notice) · Bank of Tanzania (official FX; CBR; T-bill auction 1203) · NBS (June CPI) as at 4 Aug 2026 Verified Confidence: High Marks reconcile to the verified Dar es Salaam Stock Exchange close; yields and FX read from the Bank of Tanzania, inflation from the NBS. Hypothetical & educational — never advice. Methodology