The Ratio Nobody Has Defined
The ratio the board quotes is a fraction with a numerator most SACCOs cannot defend — how to build operating cost properly, why the income half will never live in an operations system, and what a defensible number actually requires.
Cost-to-income is the one operating ratio a SACCO board can be relied upon to know. It appears in the strategic plan, it is compared against peer societies at every forum, and it is quoted at the AGM as evidence of prudence or as an apology. What it almost never comes with is a definition. Ask three officials in the same SACCO what sits in the numerator and you will get three answers, and the honest reading of a ratio that moved from 68% to 61% is often that somebody reclassified a cost.
This is not a criticism of SACCO finance teams. It is a structural consequence of where the two halves of the fraction live. Income — interest on loans, investment income, fees and commissions — sits in the core banking system, which is regulated, audited and reasonably disciplined about classification. Operating cost is spread across procurement, payroll, branch petty cash, fuel, rent and a general ledger that was set up by whoever set it up. The denominator is governed; the numerator is assembled.
From the trial balance to a number you can defend
Illustrative figures, and the point is the direction of the adjustments rather than their size. Two of the five adjustments are classification decisions where either answer is defensible; the other three are simply right or wrong. A SACCO reporting 97.9% on the unadjusted total and 63.1% on the adjusted one has not improved anything — it has finally defined its terms. The number matters far less than whether the same five decisions were made last year.
What makes a numerator defensible
Three properties, and none of them is precision. A defensible operating-cost figure is consistent — the same classification rules applied in the same way to consecutive periods, so a movement is a movement rather than a reclassification. It is traceable — every line reduces to transactions somebody can open, with an approver and a document behind each. And it is complete — the categories that habitually escape, chiefly cash spending at branches and payments to people who are not on payroll, are inside it rather than lurking in a suspense account.
Consistency is the one that gets sacrificed, because it is invisible. Nobody notices that this year's ratio excludes depreciation and last year's included it, and the improvement gets attributed to management action. The protection is dull and effective: write the definition down, in the finance policy, naming which ledger accounts sit inside the numerator and which are excluded and why. One page. Then the argument at the next board meeting is about the policy rather than about the number, which is the argument you want.
How much of the numerator an operations system can actually see
Procurement spend
Requisition, approval, order, delivery, invoice — every line traceable to a document and an approver. This is the strongest part.
Staff costs
Payroll with statutory deductions on date-effective rules. Attributable by department from an export, though not from a report.
Standalone expenses
Rent, utilities, bank charges, subscriptions — coded to a category and a project, with approval where a SACCO has switched it on.
Assets and their upkeep
Repairs traceable as expenses or orders against a vendor. Depreciation is not posted by us at all — that stays in your accounts.
Branch cash spending
Only as visible as your discipline in entering it. Nothing forces a branch to record what it spent from the till.
Cost of funds
Interest on member deposits. Belongs to the core banking system and should never appear in an operations tool.
The income half
Interest, fees, commissions, investment income. Entirely absent here, and correctly so.
The useful thing about laying it out this way is that it tells you where the ratio can be automated and where it cannot. Everything left of the middle can be assembled from transactions. Everything right of it is a figure you carry in from somewhere else, which means the ratio itself is a spreadsheet — and there is nothing wrong with that as long as nobody is under the impression it is a dashboard.
Building it once, properly
-
Write the definition before you build anything
One page in the finance policy naming the ledger accounts inside the numerator, the accounts excluded, and the reason for each exclusion. Have the audit committee adopt it. This single page is worth more than any reporting work that follows, because it is what makes next year comparable to this year.
-
Fix your expense categories to match it
Categories are the join between the transactions and the definition, and most SACCOs have accumulated forty of them with overlapping meanings. Collapse them to the fifteen or so that map cleanly onto the policy. Do this before volume builds up, because recoding history is miserable and recoding it consistently is worse.
-
Decide how branch attribution will work, once
A purchase order carries a department, so procurement spend attributes to a branch properly. A standalone expense does not. If you want per-branch operating cost, run a standing project per branch and code expenses to it — a project is the one dimension that expenses, orders, stock issues and invoices all carry. Decide this on day one; retrofitting it means touching every historic expense.
-
Close the cash gap
Branch cash spending is the single largest source of an understated numerator, and it is the subject of our SACCO expense control guide. The control is not software, it is a float with a fixed ceiling, replenished only against coded receipts. If the float is not reconciled, the ratio is guesswork with a decimal point on it.
-
Pull the numerator monthly, not annually
A cost-to-income ratio produced once a year is a historical fact. Produced monthly against a budget, it is a management tool, and the classification errors surface in a month when someone still remembers the transaction. Monthly is also the only rhythm at which the cut-off problem stays small.
-
Reconcile to the audited accounts, deliberately
At year end, tie your operating-cost figure to the audited financial statements and document the differences. There will be some, and the documented ones are fine. The undocumented ones are how a board discovers at an AGM that management and the auditor disagree about the ratio.
Score your own numerator before you compare your ratio to anyone else's. Anything under 4 on the heavier criteria means the ratio is not yet a management number.
A written definition exists and was adopted by a committee
Make them prove it: Ask for the page. If the answer is a conversation rather than a document, the definition changes with whoever prepares the report.
Last year was computed the same way
Make them prove it: Recompute the prior year under the current definition and compare it to what was reported at the time.
Every line traces to transactions with an approver
Make them prove it: Pick the third-largest cost category and open it. Then pick a transaction inside it and find who approved it.
Branch cash spending is inside the figure
Make them prove it: Compare total float replenishments for the year against coded branch expenses.
Payments to non-payroll people are captured
Make them prove it: Look for sitting allowances, casual marketers and field recruiters. Are they in the numerator?
The ratio is produced monthly
Make them prove it: Ask to see the last three months, not the last three years.
Cost of funds and provisions are excluded
Make them prove it: Check whether interest on member deposits and loan-loss provisions are in the numerator.
One caution on peer comparison, since it is the reason most boards care about the ratio at all. A comparison is only meaningful if the other society defined its numerator the way you defined yours, and there is no reason to believe it did. A SACCO with a 54% ratio may have a leaner cost base or a looser definition, and from the outside those look identical. Comparing your own trend against your own definition is a genuinely useful discipline. Comparing your number against a figure quoted at a conference is entertainment.
What AWRA OpsHub does today
- Procurement spend, fully traceable — requisition, threshold-based approval, RFQ comparison, order, delivery and invoice, with an audit trail on each step. The strongest single input to an operating-cost figure.
- Standalone expenses coded to a category and a project, with optional approval workflow, and rejected or unapproved claims excluded from every spend figure the system reports.
- Payroll with Kenyan statutory computation — PAYE, NSSF, SHIF and the housing levy on date-effective rules — plus payout runs for people who are not on the payroll, so field recruiters and casual marketers land somewhere real.
- Budgets per department and category over a period, with the budget-versus-actual comparison reporting procurement spend and standalone expenses in separate, labelled columns.
- Categories and a chart of accounts you control, which is the mechanism that makes a written cost definition enforceable.
- Reporting with export on all of it, which is what you need — because the ratio itself is assembled outside any operations system.
What it does not do
- No income side whatsoever. Interest income, fees, commissions and investment income live in your core banking system. There is no member, no loan and no deposit here, so the denominator is not something we can see, let alone compute.
- No cost-to-income ratio report. We do not produce this number, and a vendor who offers to is either integrating with your core system or guessing.
- No depreciation posted to the ledger. Nothing here charges depreciation. If you see a written-down value on an asset dashboard, check the rate before quoting it — a single flat rate across every asset class is not book value and will not agree with your audited accounts.
- No branch dimension of its own. A branch is a department, a location or a warehouse by convention. Per-branch operating cost requires a standing project per branch, because a standalone expense carries a project and not a department.
- No cost of funds, no provisioning, no impairment. Correctly outside our remit, and named here because they are the two lines most often wrongly left inside a numerator.
- Nothing enforces your written definition. Categories make it possible to apply consistently; nothing stops somebody coding a cost to the wrong one.
The shape of the fit is easy to state: we can make your operating cost complete, traceable and consistent, which is the half that is usually the problem. The ratio is still a fraction you assemble, with our numerator and your core system's denominator, in a spreadsheet with a documented method. Any vendor telling you their operations platform produces your cost-to-income ratio is describing an integration they have not built or a number they should not be computing.
Our take
Stop trying to improve the ratio and start trying to define it. Almost every SACCO we have seen could move its reported cost-to-income by five points in either direction without changing a single business decision, purely through classification — which means the number currently carries no information. Write the one-page definition, collapse your categories to match it, close the branch cash gap, and recompute last year on the same basis. The ratio you get will probably be worse than the one you have been quoting, and it will be the first one worth acting on.
Make the cost half defensible
Governed procurement, coded expenses with approval, payroll with statutory computation, and budgets with spend reported by channel — a numerator that traces to transactions. The income half stays in your core banking system.
See AWRA for SACCOsFrequently asked questions
What is a good cost-to-income ratio for a Kenyan SACCO?
The honest answer is that the question is less useful than it sounds, because the figure depends heavily on your definition, your size, whether you are in an expansion phase, and how much of your income is interest versus fees. A society opening two branches will report a worse ratio than one that is not, and be making a better decision. Watch your own trend against your own definition over three years, and treat any externally quoted benchmark as a conversation starter rather than a target.
Should depreciation be in the numerator?
Either treatment is defensible provided you apply it consistently and state which you chose. Including it captures the real cost of the asset base, which matters for a society investing in branches and ICT. Excluding it makes operational efficiency clearer year to year. What is not defensible is switching, or having different people prepare it differently. Pick one, write it down, and recompute the comparative year on the same basis.
Our branch expenses arrive late and land in suspense. How much does that distort things?
Enough to change the answer. Late branch costs are the most common reason a ratio looks good in December and worse after the audit, because cut-off errors all push in the same direction: costs incurred in the period arrive after it. The fix is not accounting, it is the float discipline — a fixed ceiling replenished only against coded receipts, reconciled monthly. Once the float reconciles, cut-off becomes a rounding issue.
Can we get operating cost per branch out of this?
Yes, if you set it up deliberately. Procurement spend carries a department, so that channel attributes to a branch directly. Standalone expenses do not carry a department, so those need a standing project per branch. Payroll cost per branch is derivable by exporting the run and joining it to the employee list, as our [SACCO staff costs guide](/blog/sacco-staff-costs-payroll-kenya) sets out, since employees carry a department. None of it is one click, and all of it is repeatable once configured. Decide on the convention before volume accumulates.
Why exclude interest paid on member deposits?
Because it is a cost of funds rather than a cost of operating, and mixing the two makes the ratio meaningless. A society that raises its deposit rate to attract savings is not becoming less efficient, but a numerator including deposit interest will say it is. Cost of funds belongs in your interest margin analysis, where a rate change is supposed to show up.
Does SASRA prescribe how this is calculated?
Supervisory reporting has its own prescribed formats and definitions, and those are what you file — do not assume your management ratio and your regulatory return are computed identically, and do not let one silently replace the other. Confirm the current return requirements with SASRA directly rather than relying on an article. The management ratio described here is for internal decision-making, and its value comes from being consistent with itself over time.
We have forty expense categories. Is collapsing them worth the disruption?
Yes, and the disruption is smaller than you expect if you do it at a period boundary. Forty categories with overlapping meanings means every preparer makes judgement calls, which is exactly the inconsistency that destroys comparability. Fifteen well-defined categories that map onto your written cost definition will be coded the same way by three different people, which is the entire objective. Map old to new once, apply it from a clean period start, and keep the mapping for the auditor.
