Judge the manager on what she controls; then watch what ROI makes her refuse

The controllability principle

A manager should be evaluated only on the revenues, costs and assets she can influence. That is the controllability principle, and it shapes the document every responsibility centre receives each month: the responsibility report. The report compares budget with actual for each line the manager controls, shows the variance, and flags the lines that need attention. Anything the manager does not control — an allocated share of head-office rent, the interest on debt head office raised — is shown below the controllable subtotal, or not at all, so that it informs without being judged.

Management by exception is how the report is read. Nobody reviews every line; the flag goes on lines outside a tolerance — Boreal uses ±5 % of budget — and those are the ones the manager explains. The rest are assumed to be under control until they cross the line.

Boreal’s Trail division, one month (CAD):

                              Budget     Actual   Variance    Flag
Revenue                      420,000    401,000   −19,000   −4.5 %
Cost of bikes sold           231,000    222,000    +9,000   (moves with revenue)
Store wages                   58,000     64,000    −6,000  −10.3 %  ← flag
Marketing                     20,000     18,500    +1,500   +7.5 %  ← flag
Other controllable costs      35,000     34,000    +1,000   +2.9 %
Controllable profit           76,000     62,500   −13,500  −17.8 %  ← flag
──────────────────────────────────────────────────────────────────
Allocated head-office cost    24,000     26,000   −2,000   not controllable
Division profit               52,000     36,500

The Trail manager answers for wages (10 % over on 4.5 % less revenue) and for the marketing underspend that may explain the revenue miss. She does not answer for the 26,000 of head-office cost, which rose because head office hired.

ROI and residual income

An investment-centre manager controls assets too, so her measure has to involve them. Two measures, on average operating assets (opening + closing) ÷ 2, at Boreal’s required return of 12 %:

$$\text{ROI}=\frac{\text{operating income}}{\text{average operating assets}}\text{RI}=\text{operating income}-12\%\times \text{average operating assets}$$

                     Road          Trail
Operating income     360,000       275,000
Average op. assets   2,000,000     2,500,000
ROI                  18.00 %       11.00 %
RI (at 12 %)         120,000       −25,000

Road earns 18 cents per dollar of assets and creates 120,000 of income above the 12 % the company needs from those assets. Trail earns 11 cents — below the hurdle — and its residual income is negative: it uses 2.5 million of assets and does not cover their cost. ROI, as a percentage, lets the board compare divisions of different size; residual income, in dollars, favours the larger division and cannot rank them.

Where ROI goes wrong

Road’s manager, whose bonus is paid on divisional ROI, is offered a project: 500,000 of new assets earning 75,000 a year. Project ROI = 75,000 ÷ 500,000 = 15.00 % — well above the 12 % the company requires. Project RI = 75,000 − 60,000 = +15,000: the company wants it. But Road’s ROI after the project = 435,000 ÷ 2,500,000 = 17.40 %, down from 18.00 %, because 15 % is below Road’s current average. A manager paid on ROI refuses a project that creates value. That is goal incongruence produced by the measure itself, and it is the reason residual income exists: RI rises by 15,000 whether Road’s average is 18 % or 30 %, so a manager paid on RI accepts every project above the hurdle. Trail’s manager has the mirror problem — anything above 11 % raises her ROI, including projects below the 12 % the company needs.