Capex vs Opex: A Decision Guide for Operational Managers

What opex and capex actually mean

Two short definitions, intentionally non-technical:

Operational expenditure (opex) is spend that’s consumed during the financial year — salaries, rent, utilities, materials, services, training. It hits the P&L in the year it’s incurred.

Capital expenditure (capex) is spend on assets that generate value over multiple years — equipment, property, vehicles, infrastructure. It hits the balance sheet first, then flows into the P&L gradually through depreciation over the asset’s useful life.

The accounting test most commonly used: will this spend create or improve an asset whose useful life exceeds one year? If yes, capex. If no, opex.

Most businesses also impose a R-value threshold (often R10,000 to R50,000) below which items are expensed regardless — because the administrative cost of capitalising a R3,000 item exceeds the savings.

Why the classification matters to the CFO

Three concrete reasons:

Cash impact pattern. A R5m capex item and a R5m opex item have completely different cash profiles. The capex is one cash outflow up front, then no further cash impact (only depreciation, which is non-cash). The opex hits this year’s P&L in full. Working-capital decisions and dividend decisions depend on this distinction.

Tax treatment. Opex is generally deductible in the year incurred. Capex is deductible via depreciation allowances over the asset’s life (the exact mechanics depend on the asset type and the relevant tax rules). The timing matters for cash tax in the current year.

P&L volatility. A business that runs heavy capex sees stable annual depreciation expense; a business that runs the same spend as opex sees a spike. The CFO is managing the P&L’s trajectory across years — the classification affects how that trajectory looks.

A proposal that doesn’t have an explicit opex/capex breakdown forces the CFO to do the classification themselves. The course manual’s business case template requires the breakdown as a mandatory section in cost analysis. Skipping it is a structural mistake.

Opex: what counts

The bulk of most departmental budgets is opex. Typical items:

• Salaries, contractors, freelancers, professional services
• Rent and utilities
• Materials, supplies, consumables
• Travel, accommodation, meals
• Training (more on this below)
• Software-as-a-service subscriptions (in most cases)
• Maintenance and routine repairs
• Marketing campaigns
• Audit, legal, insurance fees

The test: would the spend stop generating value within twelve months of being stopped? If yes, opex.

Capex: what counts

Capex items are typically:

• Property purchase or substantial improvement
• Vehicles
• Machinery and equipment (above the R-value threshold)
• Major IT infrastructure (servers, networking equipment, on-premise installations)
• Furniture and fittings (if material)
• Long-life intangibles in some cases (patents, certain software licences)

The test: does the spend create or substantially extend the useful life of an asset that will generate value beyond twelve months?

The R-value threshold matters. A R8,000 office chair is typically opex regardless of how long it lasts; a R250,000 piece of equipment is typically capex regardless of whether the business intends to replace it sooner.

The “wish list” trap

This is where most operational managers get into trouble.

The temptation: the opex budget feels tight, the capex budget seems to have more headroom (it’s often discussed as a separate process, with longer planning horizons). So managers pile wanted items into capex on the assumption that capex is “free-er” money.

This is wrong in two ways.

First, it usually isn’t. Capex requires approval through a different process — typically with stronger justification, longer review, and tighter scrutiny. The capex budget is harder to access for the items that should be in it.

Second, it corrupts the capex budget into a graveyard. The JSF course manual’s framing is direct: opex, capex, and the wish list become a tangled mess where wanted-but-not-funded items pile up under “capex” indefinitely. The capex budget loses its discipline as a strategic-asset-funding mechanism and becomes a complaint queue.

The cleaner discipline: classify honestly. Items belong where they accounting-correctly belong. If the opex budget is tight, that’s an opex conversation. If the capex budget is restrictive, that’s a capex conversation. Mixing them by misclassification doesn’t solve either.

The decision tree for borderline cases

Several common cases that don’t fit the simple test cleanly:

SaaS subscription vs on-premise software licence. A R2m on-premise software licence with perpetual usage rights is typically capex. The same R2m spent on a three-year SaaS subscription is typically opex. Same business outcome, different classification — and the choice often is about classification preference (alongside other factors). Worth being deliberate about, not accidental.

Major refurbishment vs repair. Replacing a roof to extend the useful life of a building is capex. Patching a leak is opex. The grey area: a “renovation” that’s part one part the other. Generally: if the work substantially extends the asset’s useful life or materially upgrades its capability, it’s capex; if it merely maintains existing capability, it’s opex.

Internally-developed software or tools. A piece of internal software that took 300 person-hours of internal development time to build — capex or opex? Depends. The labour costs of building an asset that will have multi-year value can be capitalised (with caveats), but the rules are narrower than people assume. Worth a conversation with finance rather than assuming.

Long-lived intangibles. Brand work, training programmes, R&D, market-development costs. The default in most accounting frameworks is opex, even though the value is multi-year — because the future value is hard to verify reliably. Some specific cases (registered trademarks, certain capitalised development costs) can go to the balance sheet, but the default is opex.

Replacement vs upgrade. Replacing a like-for-like piece of equipment is typically capex (continued asset existence). Replacing it with a substantially superior model is capex (new asset). Repairing the existing one is opex. The threshold question: did the spend extend the life or merely maintain it?

In every borderline case, the practical move is: classify the way your finance team would classify, then state the classification explicitly in the budget proposal. The CFO can override it if they disagree; what they can’t do is mind-read your assumption.

Is training capex or opex?

For South African businesses asking this question: training is opex in nearly all cases. Even where the training builds long-term capability, the accounting treatment is to expense it in the year incurred. There are narrow exceptions in some sectors (specific apprenticeship-style programmes with capitalisable elements) but the default is opex.

The SDL/BEE angle is separate and important. Training spend is also claimable as Skills Development under the Amended B-BBEE Codes — Category F training (which includes general management courses) is claimable within the 25% Cat F+G cap. The BEE positioning page on jhbfin.com sets out the mechanics. This is additional to the opex treatment — not a replacement for it.

What this piece is not about

A note on scope, because the topic invites scope-creep.

This piece is about the classification decision — whether an item belongs in opex or capex, and the consequences. It’s not about capital budgeting techniques — net present value (NPV), internal rate of return (IRR), discounted cash flow (DCF) analysis. Those are valuable skills for evaluating individual capex investments, but they’re a different course territory and JSF’s Budgeting and Cost Control course doesn’t cover them. (For that, look at investment-appraisal training.)

The reason this matters: classification gets harder when managers think they need to “do the NPV” before classifying. They don’t. Classification is an accounting question; NPV is an investment-decision question. Different skills, different sequence.

How this fits in the bigger picture

Opex/capex classification surfaces in several places in the budget conversation:

• The business case template requires an opex/capex breakdown in the cost analysis section.
• The budget defence conversation goes differently depending on which budget the items hit.
• The variance reporting tracks opex and capex separately, with different cadence and thresholds.

For the wider context, see the main budgeting guide. Cost analysis is also skill #5 in the Financial Skills Every Manager Should Have framework — opex/capex classification is part of that broader skill.

What to do next

If you’re preparing a budget proposal and the opex/capex breakdown isn’t explicit, take ten minutes to add it. The discipline of going line by line tends to surface several items that were classified by reflex rather than analysis.

For training that covers opex/capex classification, the business case template, and the wider budgeting toolkit, JSF’s Budgeting and Cost Control course runs four days, Monday to Thursday mornings, live online via Teams. The team training page covers cohort sizes and in-house options. The opex/capex/wish-list framing is covered in Section Four, with the business-case-level cost analysis discipline in Section Six. To talk through fit, request a callback.