Band
Is a sleeve outside its mandate?
Chapter 04 · weekly model-book allocation
KCP transformationEvery model-book decision. Including wait.
A compact public ledger of model-book allocation, restraint, terrain and the lesson each chapter leaves behind.
Evidence
Did the terrain materially change?
Liquidity
Could a real market absorb the action?
Attribution
Capital, allocation, or market P&L?
Hypothetical allocation record
Allocation and restraint, side by side.
Every chapter keeps its date, terrain, model-book allocation and lesson.
CH 24Model-book allocation2026-10-17 · Equities (liquidity concentration + breadth) + fixed income (provenance) + the capital ledger
Nine-tenths of the money, in two names that fell
Four verified sessions, because 14 October was Mwalimu Nyerere Day. Across them CRDB and NMB together carried 91.2% of all equity turnover value; CRDB 65.2% and leading all four sessions, NMB 26.0%; and both fell, which is why 10 of 28 counters rose and the All-Share still lost 0.54%; Chapter 12 recorded 85.4% for the same pair ten weeks ago, so the concentration deepened rather than resolved. Turnover was TZS 12.944bn, −28.2% on the headline but only −10.2% per session, and net foreign selling ran TZS 1.843bn across the three sessions the participation record covers, about TZS 614m a session against TZS 626m. The model book fell 0.03% to TSh 611.03m, a third consecutive weekly fall and thirty times smaller than the first: equities lost TSh 0.580m, the ballast accrued TSh 0.390m and covered 67.3% of it against 23.7% and 6.2%, on an accrual inside rounding of both; the hole shrank, the ballast did not change. Three of five holdings rose and the two that fell are the two banks, matching a banking index down 1.62% while Industrial & Allied rose 1.74%. No bill auction settled, so 1208 stands untested; a 25-year reading of 10.84% dated 15 October appears in the curve store without an auction number, locator or row hash and is published as unconfirmed with no decision resting on it. The largest allocation drifted to 18.20%, further below the 20% cap on price alone for a third week, leaving it untested; the two banks are 26.06% of the book and the pair question is dated to 1 November. October’s TSh 7.0m is still not entered in the tracker, sixteen days on, and is disclosed a third time rather than back-dated
→CH 23Model-book allocation2026-10-10 · Equities (breadth + concentration) + fixed income (the short end) + the capital ledgerNine rose, nine fell, and the bill curve moved from the wrong end
Across the 28 counters with a verified close on both Fridays 9 rose, 9 fell and 10 were unchanged, an exactly even board, and the All-Share still fell 0.74% because CRDB and NMB were both among the fallers and carried 83.8% of 8 October’s turnover value; turnover fell 35.4% to TZS 18.019bn with foreign sales above purchases in all five sessions, TZS 3.131bn net. The model book fell 0.19% to TSh 611.22m, a second consecutive weekly fall and the first back-to-back pair this ledger records: equities lost TSh 1.559m and the ballast accrued TSh 0.370m, covering 23.7% of it against 6.2% last week on the identical accrual, so the hole was smaller and the ballast no cleverer. BoT auction 1208 of 7 October raised the 35-day rung 84 basis points to 3.36%, its second-largest single-auction rise since January 2023, and cut the 364-day by one to 6.2839% for a sixth straight fall; both short rungs still pay less than September’s 4.3% inflation, so November’s savings were not promised to the bill ladder. The largest allocation drifted to 18.38% of the book, further below the 20% cap on price alone for a second week, which leaves the cap untested rather than vindicated. October’s TSh 7.0m is still not entered in the tracker, nine days on, and is disclosed again rather than back-dated
→CH 22Model-book allocation2026-10-03 · Equities (concentration + breadth) + fixed income (curve) + the capital ledgerA rule nobody could break, and a deposit the ledger has not taken
The model book fell 0.90% to TSh 612.41m, the largest weekly fall this ledger records, while the All-Share fell 1.08% on turnover that barely moved at TZS 27.874bn and one counter led turnover in all five sessions. The single-name cap published on 1 October could not bind in its first week, because it binds only on new savings and none arrive before 1 November; the largest allocation fell 4.23% and its weight drifted from 19.26% to 18.61% on price alone, so the cap stopped nothing and was neither vindicated nor refuted. Two bookkeeping disclosures are published rather than smoothed: October’s TSh 7.0m is not yet entered in the tracker, and Chapter 21 paired a 30 September return with post-deposit capital. Market P&L of TSh 91.41m and a money-weighted +17.89% are identical either side of the entry; a deposit is capital, not a gain
→CH 21Model-book allocation2026-10-01 · T-bonds + income fund + cash + gold; equity single-name capTwo rules, and both of them bind on the way in
The 364-day bill rung has fallen at five consecutive auctions while the bond curve still pays a real term premium, so 5m goes to bonds and nothing to bills; 1m gives the collective income fund a job because the ballast carried 86.9% one issuer; equities got nothing for a third month. The single-name cap promised in chapters 18 to 20 is published: 20% of the book, binding on new savings only, never forcing a sale, with a drift breach reported weekly rather than traded. A pair cap was considered and declined, with the reason given
→CH 20Model-book allocation2026-09-26 · Equities (concentration + breadth) + fixed income (curve)Turnover halved, and the book set a record anyway
Equity turnover fell 48.2% to TZS 27.471bn and the market rose anyway; the model book marked its highest reading yet, TSh 620.28m on 22 Sep, and closed the week at TSh 617.96m with contributed capital unchanged, so the whole TSh 10.02m rise is market P&L. The top two equity allocations reached 36.52% of the book, up for a third straight week on price alone, and BoT auction 1207 cut the 364-day rung for a fifth consecutive auction. A record on thinning volume is the worst moment to move a dated rule, so the single-name cap and the duration question both go to the 1 October allocation
→CH 19Model-book allocation2026-09-19 · Equities (concentration + breadth)The market moved the number before the rule could
The largest allocation fell 5.48% on 66.0% of market turnover with no issuer filing, taking its weight from 20.2% back under 20% without a decision; Vodacom’s 9.91% rise cancelled it and the book ended about TSh 1.0m up on a week with 15 of 28 counters lower. A cap that only binds after the price has done the work is not a cap, so the 1 October date stands
→CH 18Model-book allocation2026-09-12 · Equities (concentration) + fixed income (curve)All five rose, and this time nothing was deposited
All five equity holdings rose and no savings landed, so the whole of the NAV rise is market P&L; the largest allocation passed 20% of the book, so the single-name cap is dated for the 1 October allocation rather than postponed again
→CH 17Model-book allocation2026-09-05 · Equities (concentration) + fixed income (curve)Four rose, one fell, and the book still lost money
Four of five equity holdings rose or held flat and the book still gave back about TSh 3.3m of market P&L; the NAV rose only because September savings landed, so hold and read the deposit apart from the return
→CH 16Model-book allocation2026-09-01 · T-bonds + T-bills + cash + goldThe tenor was the problem, not the ladder
The 364-day rung pays a real yield where the 91-day does not, so the bill ladder returns one year out; equities got nothing for a second month
→CH 15Model-book allocation2026-08-29 · Equities (comparability)The measuring stick changed mid-week
Keep the allocation unchanged, report the split-adjusted holding outcomes and preserve the official DSEI and TSI weekly comparison as unavailable across the temporary discontinuity
→CH 14Model-book allocation2026-08-26 · Equities (NMB basis)Ten shares replaced one. The allocation did not change.
Apply the one-for-ten unit and cost-basis adjustment, withdraw the false mechanical loss, and leave the allocation unchanged
→CH 13Model-book allocation2026-08-15 · Equities (breadth) + fixed income (curve)Four of five paid, and the front end finally printed
Breadth improved and four of five equity holdings paid, but the front end printed a narrower reward for duration; hold, and don't mistake a better week for a solved problem
→CH 12Model-book allocation2026-08-08 · Equities (breadth) + fixed incomeTwo banks carried 85% of the money
Two banks supplied 85.4% of the week's turnover and one holding 92% of the book's; study the concentration, and let the duration idea wait
→CH 11Model-book allocation2026-08-04 · Equities (concentration)Five rose, fourteen fell; and the book had its best week
Best week since inception, ~85% of it from one holding; examine the concentration, do not feed it
→CH 10Model-book allocation2026-08-01 · T-bonds + gold + cashWhere the month's 7m went; and the sleeve that got nothing
The bill ladder stopped paying a real yield; equities got nothing, by choice
→CH 09Model-book allocation2026-07-25 · Equities (breadth)A record with an asterisk, and the book holds
An all-time high on zero-volume marks and negative breadth; hold until the board confirms it
→CH 08Model-book allocation2026-07-18 · Equities (banks)A narrow recovery, and the book holds
The index clawed back near its record on bank-led, negative-breadth strength; no drift to correct
→CH 07Model-book allocation2026-07-11 · Equities (banks)A record cools, and the book still holds
The record cooled; the book still holds; no drift to correct
→CH 06Model-book allocation2026-07-04 · Equities (banks)Holding into a record tape
Three record closes in a row; do nothing, on purpose
→CH 05Model-book allocation2026-07-01 · Equities + T-billsWhere the month's 7m went; and the signals behind it
Feed the under-weight sleeves that had room
→CH 04Model-book allocation2026-06-20 · Commodities (gold)Opening a gold sleeve on the reserve-accumulation thesis
Open a small reserve-accumulation hedge
→CH 03Model-book allocation · Equities (banks)The Journey; the story so far.
Let the bank winners run rather than take the easy profit
→CH 02Model-book allocation · CommoditiesCommodities
Set the gold rules before opening any position
→CH 01Model-book allocation · Fixed incomeThe Journey; the story so far.
Patience in T-bills while waiting for a decision worth making
→CH 00Model-book allocation · MethodWhat The Ghost is; and isn't.
The premise, the mandate, and the rules it lives by
→