Divisional Performance Measurement

1 Introduction

A divisionalised (or decentralised) organisation is split into units whose managers are given authority to take decisions for themselves, and who are then held accountable for the results. Those units are called responsibility centres.

This chapter deals with four questions that follow from that structure, and they are the four questions the syllabus asks. What are the different types of responsibility centre and what should each be measured on? Which costs and revenues may fairly be laid at a manager's door? How is the budget used to evaluate a centre's performance? And for an investment centre — where the manager also decides how much capital to employ — which measure relates the profit earned to the capital used?

Throughout, keep the second question in view: a performance measure is not only a measurement, it is an instruction. Managers do what they are measured on, and a badly chosen measure will get them to do the wrong thing perfectly well.

2 Divisionalised structures

YouTube videoLoad YouTube video

2.1 What makes a division

Having a branch office does not make an organisation divisionalised. What does is autonomy — the manager can take decisions without asking head office first, and runs that part of the business much as though it were a separate company.

Divisions are usually organised by product line, customer type or geographical location. Head office keeps strategy, the allocation of capital, and the design of the performance measurement system; everything else is delegated.

2.2 Advantages of divisionalisation

• Decisions are taken by the people closest to the market, the product or the customer, who know more about them than head office does.

• Decisions are taken faster, because they do not queue for head-office approval.

• Divisional managers are more motivated: they are appraised on factors they largely control, and they are running something of their own.

• It develops general managers, because a divisional manager has to take the whole range of decisions rather than one function's.

• Objectives become clearer and narrower — one division rather than the whole group.

• Divisional results can be compared with each other and with the plan.

2.3 Problems of divisionalisation

• Loss of central control, and difficulty in co-ordinating divisions whose decisions affect each other.

• Dysfunctional decisions — a manager doing what is best for the division may do what is worst for the group. This is the goal congruence problem, and it runs through sections 5, 7 and 8, and through the whole of chapter 6.

• Where divisions trade with each other a transfer price has to be set (chapter 6).

• Duplication of activities that could have been carried out once centrally.

• Comparison between divisions can be unfair — different markets, different asset ages, different degrees of central support — and unfair comparison demotivates.

3 Responsibility centres

A responsibility centre is a unit of an organisation whose manager is accountable for a specified set of results. The 2027 syllabus names four types, and they differ in exactly one respect: how much the manager controls.

3.1 The four types, their objectives and their measures

Two consequences follow from the table and both are examined.

1. A revenue centre is not a profit centre. Its manager can raise revenue by discounting, and revenue will rise while profit falls. A revenue centre must therefore always be measured on revenue and on the selling costs of earning it, and usually on mix as well, or the measure invites exactly the wrong behaviour.

2. Only an investment centre is measured on a return. Applying return on investment to a profit centre punishes a manager for assets they did not choose.

4 Controllability, cost variability and attributable costs

4.1 Controllable and uncontrollable costs and revenue

A cost or revenue is controllable by a manager if that manager can significantly influence it within the period being reported on. Everything else is uncontrollable by that manager — which does not mean it is uncontrollable by anybody.

The classic uncontrollable item in a divisional report is the apportioned head-office cost: a share of central salaries, group insurance or the group audit fee, charged out on a formula the divisional manager did not choose and cannot change. It must be shown, because it is a real cost of the group, but it must never be inside the figure on which the manager is appraised.

4.2 Cost variability

Before any cost can be compared with a budget, its behaviour has to be known — how it responds to the level of activity:

This is not bookkeeping detail. It is the reason a division that produced 15% more than planned will have spent more than the budget, and why comparing the actual spend with the original budget tells nobody anything. Section 5.2 does the comparison properly.

4.3 Attributable costs and revenue, and the divisional report

A cost or revenue is attributable to a division if it would disappear were the division closed — whoever happens to control it. Attributable and controllable are different tests, and the difference gives the divisional performance report its structure:

The same distinction decides whether revenue belongs to a centre: revenue is attributable to a division if it would be lost were the division closed. A sale that would simply be picked up by the division next door is not.

5 Budgets and performance evaluation

A performance measure needs something to be measured against. For a responsibility centre that yardstick is normally the budget: the quantified plan for the centre for the period, agreed in advance.

5.1 What the budget does for a responsibility centre

1. It sets the target — what the centre is expected to achieve, in the terms the centre is measured in (cost for a cost centre, revenue for a revenue centre, profit for a profit centre).

2. It authorises — it tells the manager what may be spent, and therefore what autonomy actually amounts to in figures.

3. It co-ordinates — one division's sales budget is another's production budget.

4. It provides the control loop — actual results are compared with it, differences (variances) are reported, and the significant ones are investigated. That is management by exception.

5. It motivates — but only if the target is seen as demanding and attainable, and only if the manager believes the comparison will be made fairly.

5.2 Fixed and flexed budgets

The budget set before the period began is the fixed budget — fixed in the sense that it was drawn up for one particular level of activity, not in the sense that its costs are fixed costs. If the centre ends up operating at a different level, comparing actual results with it mixes two quite different things: the fact that volume differed from plan, and the fact that spending per unit differed from plan.

The flexed budget separates them. It restates the budget for the activity level actually achieved — variable costs are re-costed at the budgeted rate per unit, fixed costs are left alone — and the comparison is then like with like:

Flexed budget cost = budgeted fixed cost + (budgeted variable cost per unit × actual units)

5.3 What the report contains for each type of centre

5.4 Best practice in budget-based evaluation

1. Flex before you compare. Section 5.2.

2. Apply the controllability principle. Report uncontrollable items, but outside the line the manager is judged on.

3. Let managers participate in setting their own budgets. Participation improves the quality of the estimate — the manager knows the division — and improves acceptance of the target. Its cost is budgetary slack (section 5.5).

4. Make the target demanding but attainable. A target believed to be impossible is not a target; it is an excuse to stop trying.

5. Use more than one measure. A budget is a financial yardstick and financial yardsticks are backward-looking. Pair them with non-financial indicators (chapter 8).

6. Re-budget when the world moves. Rolling budgets, updated each quarter for the following four quarters, keep the yardstick current in a volatile market. The cost is the budgeting effort itself.

7. Separate the uses. The same figure cannot always be the best plan, the best forecast and the fairest basis for a bonus.

5.5 Behavioural consequences

Because the budget is the yardstick for reward, managers manage the yardstick as well as the business:

6 Reporting and data analytics for responsibility centres

Section 5 assumed a periodic report — a table of budget, actual and variance, arriving after the period has closed. That is still the core of responsibility accounting, but it is no longer the whole of it, and the syllabus asks how divisional managers use analytics, visualisation and self-service technologies to manage performance.

6.1 What a responsibility-centre report is for

A report exists to support a decision. If nothing a manager could decide would change because of a line in the report, that line is costing money and buying nothing. Three tests apply to every line: is it controllable by the person receiving it, is it timely enough to act on, and is it material enough to be worth acting on?

6.2 Four kinds of analytics

The progression matters for performance management: descriptive analytics tells a manager the period was bad, and only diagnostic and predictive analytics let the manager do something about the next one.

6.3 Visualisation and dashboards

A dashboard presents a small number of measures for a responsibility centre visually and continuously, rather than as a monthly table. Well designed, it does three things a printed variance report cannot:

• Exception first. Measures outside tolerance are highlighted, so attention goes where management by exception says it should.

• Drill-down. A divisional total can be opened up to the branch, product or customer behind it during the conversation, rather than in a request that takes a week.

• Trend, not point. A twelve-month line shows the direction of travel; a single month's variance does not.

The risks are real and examinable. A dashboard shows what is easy to measure, which is not always what matters; a chart drawn on a truncated axis exaggerates; and a screen refreshed hourly invites managers to react to noise. The remedy is to fix a tolerance before looking, and to keep the number of measures small.

6.4 Self-service and what it changes

With self-service reporting tools the divisional manager queries the data directly instead of asking the finance function for an analysis. That shortens the loop from days to minutes and frees the finance function from producing routine reports. It also moves two things into the manager's hands that used to sit with finance:

1. Definition. If two divisions each build their own “gross margin”, the figures will not be comparable. Self-service needs a governed set of agreed definitions behind it, or divisional comparison quietly stops meaning anything.

2. Interpretation. The finance function's role shifts from preparing the numbers to challenging them — and to guaranteeing the integrity of the data underneath. Where that data comes from, and how its integrity is assured, is dealt with in chapter 18.

Chapter 8 takes the reporting half of this further: which non-financial indicators belong on the dashboard, and how the balanced scorecard organises them.

7 Investment centres and return on investment

An investment centre's manager decides how much capital the centre employs. Measuring such a manager on profit alone is useless: an extra $10,000 of annual profit is good news if it came from $40,000 of new investment and bad news if it came from $10m. The measure has to relate profit to the capital used.

The first of the two measures is return on investment (ROI), also met as return on capital employed:

ROI = (Controllable divisional profit ÷ Divisional capital employed) × 100%

• Quoted as a percentage, which is why it is the more popular of the two with managers.

• Use annual profit. If the data is monthly or quarterly, annualise it first.

• Use controllable profit and controllable capital employed — the same principle as section 4.3, applied to both halves of the fraction.

So far so good. Now change one figure.

8 Residual income

Residual income (RI) measures the same thing in absolute money rather than as a percentage. The profit is charged with a notional — imputed, not actually invoiced — interest cost on the capital employed, normally at the group's cost of capital:

Residual income = Controllable divisional profit − (Cost of capital × Divisional capital employed)

Whatever is left is the residual income. If it rises the manager has performed well; if it falls, badly.

9 Return on investment compared with residual income

Two criticisms apply to both measures, and both follow from their being accounting measures:

1. They favour divisions with older assets. Accumulated depreciation reduces the carrying amount of the capital employed, so the denominator shrinks year by year and the measure improves for no better performance. A manager can also improve the measure by not replacing assets — which is the same short-termism as section 5.5.

2. They are based on profit, not cash. Both are therefore affected by the choice of accounting policies and by estimates such as depreciation, provisions and the treatment of development expenditure. That is precisely the criticism economic value added sets out to answer.

10 Economic value added

YouTube videoLoad YouTube video

Economic value added (EVA) is a trade-marked measure developed by Stern Stewart & Co. It is residual income with two refinements: the profit is adjusted to bring it closer to a cash, economic figure, and the capital charge uses the weighted average cost of capital applied to a similarly adjusted capital employed.

EVA = NOPAT − (WACC × Adjusted capital employed)

where NOPAT is the adjusted net operating profit after tax and the capital employed is the opening figure for the year — on the reasoning that it was the assets in place at the start of the year that earned the year's profit.

The principle is that a business creates value only if its profit exceeds the minimum return that its shareholders and lenders could have obtained elsewhere on investments of comparable risk. EVA also gives one consistent language — value added to the investors' capital — in which every management decision can be modelled, monitored, communicated and rewarded.

10.1 The adjustments

The corresponding adjustments to capital employed are to add non-capitalised leases, the research and similar expenditure now treated as capitalised, goodwill previously written off, and provisions — and to treat debt as part of capital employed.

10.2 The weighted average cost of capital

WACC = (k_e × W_e) + (k_d × (1 − t) × W_d)

The cost of debt attracts tax relief, so it enters the average net of tax; the weights are the target capital structure, not necessarily the actual one.

10.3 Problems with EVA

• It is an absolute measure, so it suffers the same comparison problem as residual income: a large division reports a large EVA.

• It can still encourage short-termism, because the capital charge falls on investment made now and the returns arrive later.

• Economic depreciation is difficult to calculate and departs from generally accepted accounting principles, so the figures are not the audited ones.

• The adjustments are numerous and to a degree arbitrary; two analysts can produce two EVAs for the same company.

• Like every financial measure it ignores what is not in the financial statements — brand, capability, customer satisfaction. Chapter 8 addresses that.

11 Test your knowledge

Two quick checks before you move on: work through the flashcards to fix this chapter’s key terms and definitions, then sit the objective questions for exam-style practice. Both mark themselves and explain the answers as you go.