00 - OverviewThe Short Answer: Five Drivers Behind the Regional Credit Gap

Banks in four of Tanzania's six economic zones lent more in 2025/26 than they collected in deposits there: Southern Highlands (loans equal to 153% of deposits), Central (139%), South Eastern (128%) and Lake (127%). Dar es Salaam, by contrast, holds 62.8% of all deposits but only 51.5% of loans, and carries a loan-to-deposit ratio of 75%. This report takes the Bank of Tanzania zonal data and asks why the gap exists, what it does to the real economy, and how it is likely to move between October 2026 and March 2027.

Why the gap exists: five drivers, ranked by how well the data support them

Source: BoT Tables 1.1, 1.2, 3.x, 6.1-6.3 and Annex 5; TICGL calc. "Measured" means the figure is in the report; "inferred" means it follows from the figures but the report does not state it.

What it means

• A loan-to-deposit ratio above 100% is not the same as over-borrowing. Loans are only 8-20% of GDP in the four borrower zones, compared with 59% in Dar es Salaam. These zones are under-banked, and the gap reflects missing local savings as much as heavy borrowing.
• The risk sits in two zones. Central and Southern Highlands combine a ratio above 135%, deposits that barely grew, inflation above 5% and loan books dominated by household and farm credit. Together they are the only zones where all four risk flags in Tab 5 are switched on.
• Six-month path. If each zone repeats its 2025/26 growth rates, Southern Highlands moves to about 187% and Central to 154% by March 2027 (Tab 6). If deposit growth halves, Lake also rises to 146%. Dar es Salaam stays below 80% in every case.
• Policy angle. The issue is not lending that is too high; it is deposit mobilisation that is too low in the zones where lending is rising fastest, and the absence of published credit-quality data by zone.

01 - The Gap in NumbersDeposits, Loans and the Net Funding Position by Zone

The six zones in the Bank of Tanzania report hold TZS 50.0 trillion of deposits and TZS 45.8 trillion of loans (both exclude Zanzibar). The national loan-to-deposit ratio is 92%, which hides two very different groups: Dar es Salaam and Northern, where deposits exceed loans, and the other four zones, where the reverse is true. The table sets out the stocks, growth rates and net position.

Source: Banks via BoT Tables 6.1 and 6.2; ratios and net positions are TICGL calc. on zone-booked stocks at end-June 2026 (provisional).

1.1 How the gap has built up

The ratio has risen in Central from 91% to 139% and in Southern Highlands from 51% to 153% over four years, while Dar es Salaam has stayed around 75-83% and Northern has crept up from 72% to 81%. South Eastern and Lake started above 100% and have stayed there (123% to 128%, and 119% to 127%), so for them the gap is a long-standing feature, not a new one. The widening is therefore a Central and Southern Highlands story: loans in Central tripled (x2.95) while deposits did not double (x1.93), and in Southern Highlands loans rose x3.19 while deposits rose only x1.07.

Source: TICGL calc. from BoT Tables 6.1 and 6.2. Negative = deposits exceed loans.

1.2 The Southern Highlands break in 2024/25

Southern Highlands loans jumped from TZS 1,060.8 billion to TZS 2,268.8 billion in 2024/25 (+114%) while deposits fell by TZS 96.2 billion, taking the ratio from 53% to 118% in a single year. A move of this size looks more like a change in where large loans are booked, or a reclassification, than a sudden rise in household borrowing. It should be confirmed with the Bank of Tanzania before it is used to argue that the zone has become over-indebted. The 2025/26 ratio of 153% is the result of that step plus a further 28.1% rise in loans.

02 - Driver 1A Thin Local Deposit Base

The first and best-supported explanation is that outside Dar es Salaam there is simply not much deposit money booked locally. Depositing needs income left over after consumption, a bank within reach, and an institution that holds its accounts in that zone. All three are weaker in the four borrower zones, and in two of them the deposit base has barely moved in four years.

Source: BoT Tables 1.1, 1.2, 6.1, 6.2, 6.4; population implied by GDP / GDP per capita; per-person and growth figures are TICGL calc. Excludes Zanzibar. Deposits and loans are booked by bank location, not by the residence of the customer.

2.1 Income and savings: richer zones do not automatically deposit more

Deposits per person range from TZS 5,347,000 in Dar es Salaam to TZS 203,000 in Southern Highlands, a gap of about 26 times, against a gap in GDP per person of only 1.7 times (TZS 6.78 million against TZS 3.99 million). Southern Highlands and Northern are richer per head than Lake, Central and South Eastern, yet Northern holds nearly three times the deposits per person of Southern Highlands (TZS 560,000 against 203,000). Income alone does not explain where deposits sit; the structure of the local economy does. Much of the output of Southern Highlands and Lake is farm, forest and mining value that is sold to buyers, processors and exporters, and where their accounts are held is not shown in the report.

2.2 Reach: bank agents and access

Agent banking is the main way banks extend deposit-taking beyond branches. Dar es Salaam has 102 agents per 10,000 people; the other zones have between 18 (Lake) and 29 (Northern). Lake, with 21.1 million people and the largest economy, has the lowest agent density. The agent network grew 25.6% nationally in 2025/26, but more points of contact do not guarantee more money held locally: in Central the number of agents grew 29.7% while bank deposits grew 0.9%.

03 - Driver 2Where Money Is Booked: Dar es Salaam as the Deposit Centre

The second explanation is about accounting geography. The Bank of Tanzania reports deposits and loans by the zone in which the bank records them, which is where the branch, head office or treasury sits, not necessarily where the customer lives or earns. Dar es Salaam is where the Government, large companies, many banks' head offices, the port and the Revenue Authority are located, so deposits from all over the country are booked there.

Source: TICGL calc. from BoT Tables 1.1, 1.2, 3.x, 4.1, 5.x, 6.1 and 6.2. Population is implied by GDP / GDP per capita.

3.1 Deposits are worth 3.7 times more to Dar es Salaam than its economy suggests

Dar es Salaam produces 17.0% of mainland GDP and holds 62.8% of deposits: a deposit-to-GDP weight of 3.7. Lake produces 26.1% of GDP and holds 10.5% of deposits (0.40); Southern Highlands 15.9% and 3.8% (0.24); South Eastern 10.3% and 3.2% (0.31); Central 13.5% and 9.0% (0.66); Northern 17.2% and 10.7% (0.62). Loans are less concentrated: Dar es Salaam has 51.5% of loans, a weight of 3.0. The 11-point gap between its deposit share (62.8%) and its loan share (51.5%) is the mirror image of the borrower zones' funding gap, together with the TZS 4.2 trillion that is not lent anywhere.

Source: TICGL calc. from BoT Tables 1.1, 6.1 and 6.2. A weight of 1.0 means the zone holds deposits (or loans) in proportion to its GDP.

3.2 Productive credit also sits in Dar es Salaam

Table 6.3 of the report splits lending by activity and zone. Dar es Salaam accounts for 70% of manufacturing credit, 81% of transport and communication, 80% of building and real estate, 82% of health, education, utilities and finance, and 89% of "other" lending, but only 30% of personal loans and 13% of agriculture. The credit that funds factories, trucks and buildings is booked where the head offices are; the credit that reaches households and farms is the part left for the other zones.

3.3 What the report cannot show: the inter-zone funding mechanism

The arithmetic is consistent with a simple story. Dar es Salaam and Northern collect TZS 8.87 trillion more than they lend; the four borrower zones lend TZS 4.65 trillion more than they collect; so a flow of about TZS 4.65 trillion must be moving from deposit-rich booking centres to branches that lend locally, most likely inside banks through head-office treasury and inter-branch accounts. The report has no bank-level or inter-branch data, so this cannot be confirmed from it. What can be said is that the ratio above 100% is a bookkeeping outcome of a network that collects centrally and lends regionally, not a sign that those zones are lending money they do not have.

04 - Drivers 3 and 4Farm Incomes, Cross-Border Trade and What the Borrowing Is For

The third and fourth explanations are about who is borrowing and how much they earn. Two zones, Central and Southern Highlands, combine rising loans with weak farm incomes and, in the case of Southern Highlands, a collapse in recorded cross-border trade. The remaining borrower zones, South Eastern and Lake, borrow mainly for consumption and trade respectively.

4.1 Farm incomes and local deposits move together in the zones that matter

Food-crop production fell 8.8% in Central and 11.3% in Southern Highlands in 2025/26, the only two zones with a decline. In the same year nominal GDP grew 10.3% and 11.0% there, so recorded activity kept pace while physical output did not. The deposit data tell a clearer story: Central deposits grew 0.9% and Southern Highlands deposits fell 1.4%, against 22.6% to 32.3% in the other four zones. Loans kept growing, by 16.0% and 28.1%.

Source: BoT Tables 3.1 (food crops), 1.1 (GDP), 6.1, 6.2; Annex 4 (inflation). Growth rates are TICGL calc.; food inflation is year-on-year, June 2026.

4.2 Southern Highlands: trade, deposits and a loan jump

Recorded cross-border exports from Southern Highlands fell from TZS 6.45 trillion in 2021/22 and 6.59 trillion in 2022/23 to TZS 1.28 trillion in 2023/24, and have stayed near that level (TZS 1.09 trillion in 2025/26). The zone's deposits, which had been growing slowly, stopped growing in the same period: TZS 2,016.9 billion in 2023/24, 1,920.7 billion in 2024/25 and 1,893.8 billion in 2025/26. The timing is consistent with a link between lower trade earnings and a smaller local deposit base, but the report does not prove one. Loans moved the other way after 2023/24 (see the 2024/25 step in Tab 1).

4.3 What the borrowing is for

Within each zone's own lending, personal loans take 49% in Central, 49% in Southern Highlands, 66% in South Eastern, 40% in Lake, 44% in Northern and 22% in Dar es Salaam. Agriculture takes 33% of Central lending and 23% of Southern Highlands lending, against 4% in Dar es Salaam. The report does not say what the personal loans fund (salary-backed credit, housing, consumption or small business), so we cannot tell how many are productive. What the data do show is that credit outside Dar es Salaam is mainly household and farm credit, which depends on incomes that were under pressure from food output and prices in 2025/26.

Source: BoT Table 6.3; TICGL calc. Zone mix is computed on each zone's own column total (which does not equal the Table 6.2 loan share; see the Tab 3 source note).

4.4 Loans are growing faster than the economy, which is partly a good sign

Loans grew between 16.0% and 39.0% in every zone, against nominal GDP growth of 9.8% to 11.0%. Where credit starts from 8-11% of GDP, as in Lake, South Eastern and Southern Highlands, faster growth than the economy is what financial deepening looks like. Lake is the clearest case: loans +39.0%, deposits +32.2%, with trade (22% of its lending) and mining (7%) both active and the zone's recorded cross-border exports up 75.6% to TZS 12.8 trillion. Deposits there are growing nearly as fast as loans (32.2% against 39.0%), so its ratio has risen only 6 points since 2024/25.

05 - Economic EffectsWhat the Gap Does to Growth, Prices and Risk

What does a persistent gap between local lending and local deposits do to the economy? The effects run through four channels: funding dependence, the squeeze on borrowers when prices rise, the allocation of credit between productive and consumption uses, and the unevenness of financial deepening. We then combine four observable conditions into a simple risk screen for each zone.

5.1 Four channels

• Funding dependence. A zone that lends TZS 0.5-1.7 trillion more than it collects is relying on funds booked elsewhere. If banks tighten liquidity or reprice head-office funding, the first branches to ration credit are those with a thin local deposit base. This makes regional credit more cyclical than credit in Dar es Salaam.
• Price squeeze on borrowers. Fuel prices rose sharply in April-June 2026, and inflation reached 8.0% in Southern Highlands and 5.2% in Central by June. Household and farm borrowers on fixed instalments see real incomes fall just when loans are largest relative to local deposits. The report contains no zonal data on arrears, so the effect cannot be measured, only flagged.
• Allocation of credit. Productive lending (manufacturing, transport, building) is concentrated in Dar es Salaam, so the capital that raises output per worker is concentrated with it. Credit to the other zones is dominated by personal and farm loans, which support consumption and seasonal working capital more than they raise capacity.
• Uneven deepening. Loans are 59% of GDP in Dar es Salaam and 8-20% elsewhere. Even the four borrower zones are under-banked relative to the economic centre; the problem is not too much regional credit but too little regional saving to support more of it.

5.2 A four-flag risk screen

For each zone we check four conditions that can be read directly from the report: loans above deposits; loans growing at least 10 percentage points faster than deposits; June 2026 inflation above 5%; and personal plus agricultural lending above 70% of the zone's lending. The number of flags is a count, not a credit rating, and it says nothing about actual repayment performance.

Flags: (1) loans/deposits above 100%; (2) loan growth exceeds deposit growth by more than 10 points; (3) June 2026 inflation above 5%; (4) personal plus agriculture above 70% of zone lending. Source: BoT Tables 6.1-6.3, Annex 4; TICGL calc.

• Central and Southern Highlands (4 flags). High ratio, loan growth far ahead of deposits, inflation above 5% and a household-and-farm loan book. These are the two zones to watch first in the next BoT zonal release.
• South Eastern (2 flags). Ratio of 128% and a 77% personal-and-farm book, but deposits grew 32.3% (faster than loans) and inflation is the lowest of any zone (2.6%).
• Lake (1 flag). Above 100%, but deposits and loans are growing together, inflation is 3.0% and trade and mining are well represented in lending.
• Dar es Salaam (1 flag). The only flag is inflation (6.2%). Its funding position is the strongest, which is why it can subsidise the rest.
• Northern (0 flags). A balanced zone: 81% ratio, deposits and loans growing in step, inflation 3.3%.

06 - Six-Month OutlookScenarios to March 2027, Watch-List, Policy and Entry Screen

This tab projects each zone's loans and deposits to end-March 2027, nine months after the June 2026 base (a factor of (1+g) to the power 0.75). The three scenarios are mechanical and transparent, so that readers can reproduce them from Tab 1: Momentum (each zone repeats its 2025/26 growth in loans and deposits), Parallel growth (loans and deposits both grow 20% a year in every zone, so ratios stay where they are) and Savings squeeze (loans keep their 2025/26 pace but deposit growth halves, for example because higher living costs cut saving). They are scenarios, not forecasts of record.

Source: TICGL calc. from BoT Tables 6.1 and 6.2 (June 2026 base). Excludes Zanzibar. Momentum extends 2025/26 growth rates for nine months and is a mechanical extrapolation, not an expectation: the Southern Highlands result in particular reflects the 2024/25 loan step and a falling deposit base.

6.1 Reading the scenarios

• Under Momentum the ratio rises where deposits are weakest. Southern Highlands moves from 153% to about 187% and Central from 139% to 154%; Lake goes from 127% to 132%. South Eastern drifts down from 128% to 123% and Dar es Salaam from 75% to 69% because their deposits are growing faster than loans.
• Parallel growth freezes the picture. If banks and savers grow together, no zone gets worse and the four borrower zones stay above 100%. This is the benign case; it needs deposit growth in Central and Southern Highlands to jump from near zero to 20%, which the 2025/26 data do not suggest.
• A savings squeeze is a Lake and Northern story as well. If deposit growth halves, Lake rises to 146% and Northern from 81% to 87%, so the fuel-price shock can push the funding gap into zones that look comfortable today.
• Nationally the effect is small. The all-zone ratio would be 89% under Momentum, 92% under Parallel growth and 97% under a squeeze, against 92% in June 2026. The story is about distribution between zones, not the system total.

6.2 What to watch between October 2026 and March 2027

Source: BoT zonal report and TICGL calc.

6.3 Policy options

• Publish zonal credit quality. Non-performing loans, average loan size and interest rates by zone would show whether a ratio above 100% is a risk or just a funding pattern.
• Mobilise deposits where lending is growing. Link agent-banking incentives to balances held, not only to transactions, in Central, Southern Highlands and Lake.
• Share the risk on farm credit. Credit guarantees and crop insurance for lending in zones with falling food output (Central, Southern Highlands), so that a poor season does not turn into arrears.
• Route crop and trade proceeds through local accounts. Cooperative and warehouse-receipt payments that settle into zone-based accounts would keep more of the value chain's deposits where the output is produced.
• Watch consumer credit. Personal loans are two-thirds of South Eastern lending and about half of Central and Southern Highlands lending; affordability checks matter most when inflation is rising.
• Disclose inter-zone funding. A simple statistic on bank funding flows between zones would settle the question the report cannot answer.

6.4 What it means for investors and lenders: a zone entry screen

Source: BoT Tables 6.1, 6.2, 6.4 and TICGL calc. as in Tabs 2 and 5. The reading column is TICGL/TERI's interpretation of these indicators, not a recommendation to lend or invest.