Organizational structure, decentralization and responsibility centers

Control systems are built on top of an organizational structure. Before anyone can set a budget, measure a division's return or design a bonus, someone has to decide who is responsible for what. This lesson looks at how organizations divide up decision rights, why many choose to decentralize, the four kinds of responsibility center that result, and the controllability principle that governs how managers should be evaluated. ACCT-102 introduced responsibility centers briefly; here we examine the design choices behind them and the measurement problems they create.

Organizational structures

A functional structure groups activities by specialty: manufacturing, marketing, finance, engineering. It captures economies of scale and deep expertise, and it suits a single-business firm in a stable market. Its weakness is coordination: no one below the chief executive is responsible for the profit of a product line, and decisions that cross functions must travel up the hierarchy. A divisional structure groups activities by product, customer or region, with each division containing most of the functions it needs. Division managers can be held responsible for profit, which sharpens accountability and speeds decisions, at the cost of some duplication and possible conflict between divisions.

A matrix structure overlays the two, with staff reporting to both a functional manager and a product or project manager. It tries to capture both expertise and responsiveness but blurs accountability and can create conflict. Network or virtual structures rely heavily on outside partners for manufacturing, logistics or development, leaving the core organization to coordinate. Many large companies combine elements: Cormorant has three product divisions, but finance, information technology, legal and human resources are run centrally as shared services, and the Irish subsidiary is a separate legal entity organized by region.

Why decentralize?

Decentralization is the delegation of decision-making authority to lower levels of the organization. The degree of decentralization is a matter of extent, not a yes-or-no choice: a division manager may be free to set prices and hire staff but need head-office approval for capital projects above a limit.

The benefits are well established. Managers closer to customers and operations have better information about local conditions, and decentralization lets them act on it without passing it up the hierarchy, where it can be lost or delayed. Decisions are faster. Senior managers are freed to focus on strategy. Junior managers gain experience that prepares them for senior roles. Many people are more motivated when they have real authority. These benefits are greatest in organizations facing diverse markets, fast change and complex products.

The costs are equally real. Managers may pursue their own unit's interests at the expense of the whole, a problem called goal incongruence or dysfunctional decision making. Functions are duplicated across divisions. Divisions may compete with one another for customers or resources. Head office has less direct knowledge of what is happening and must rely more on reports, which managers can manipulate. Much of this course is about the measurement and incentive tools that try to keep the benefits of decentralization while limiting its costs.

Responsibility centers

A responsibility center is a part of the organization whose manager is accountable for a specified set of activities. The type of center depends on which financial results the manager can influence and is held accountable for. A cost center manager is accountable for costs only. An engineered cost center, such as a production department, has a clear relationship between inputs and outputs, so standard costs and flexible budgets (ACCT-222) can measure efficiency. A discretionary expense center, such as research, legal or human resources, has no clear input–output relationship; budgets are set by management judgment and the main control is spending against budget, which says little about efficiency or effectiveness.

A revenue center manager is accountable mainly for revenues, as in a regional sales office; the risk is that managers chase volume at any price or margin, so revenue centers are often given some responsibility for selling costs or for gross margin. A profit center manager is accountable for both revenues and costs, and therefore for profit. An investment center manager is accountable for profit and for the assets used to earn it, and often has authority over capital investment. Investment centers are evaluated with measures that relate profit to capital, such as ROI, residual income and economic value added, the subject of lessons 4 and 5.

The classification depends on authority, not on the unit's activities. A division that cannot set its own prices because head office negotiates all national accounts is closer to a cost center than a true profit center, whatever its organization chart says. A common source of frustration is a mismatch between authority and accountability: managers held responsible for profit but denied the authority to change prices, product mix or staffing.

The controllability principle

The controllability principle holds that managers should be held accountable only for the results they can control. It rests on fairness and on motivation: evaluating people on outcomes they cannot influence exposes them to risk without improving their decisions, and they may respond by setting easy targets, spending time on excuses or leaving. In practice, few results are entirely controllable. Merchant and Van der Stede identify several types of uncontrollable factors. Economic and competitive factors include changes in interest rates, exchange rates, customer demand and competitors' actions. Acts of nature include fires, floods, storms and pandemics. Interdependence arises when a unit's results depend on other units, as when Appliances' delivery performance depends on Components supplying motors on time.

Several techniques remove or reduce the effect of uncontrollable factors. Variance analysis separates volume, price and efficiency effects (ACCT-222), so that a manager is evaluated on efficiency even when volume falls for reasons beyond her control. Flexible budgets adjust cost allowances to actual activity. Relative performance evaluation compares a unit's results with those of similar units or competitors facing the same conditions, filtering out common shocks. Subjective, ex post adjustments let superiors modify results for events they judge uncontrollable. Each technique has costs. Relative evaluation needs a valid comparison group and can discourage cooperation. Subjective adjustments depend on the superior's judgment and can be seen as unfair or biased.

# Relative performance evaluation: remove a market-wide shock from a sales target
target_growth = 0.06          # Appliances' budgeted unit growth
industry_growth_budget = 0.04 # industry growth assumed when the target was set
industry_growth_actual = -0.02
actual_growth = 0.01
shock = industry_growth_actual - industry_growth_budget
adjusted_target = target_growth + shock
print(f"Market shock: {shock:+.1%}")
print(f"Adjusted target: {adjusted_target:+.1%}")
print(f"Actual growth: {actual_growth:+.1%}")
print("Beat adjusted target" if actual_growth >= adjusted_target else "Missed adjusted target")

Market shock: -6.0%
Adjusted target: +0.0%
Actual growth: +1.0%
Beat adjusted target

Appliances missed its original 6 percent growth target by a wide margin, growing only 1 percent. But the industry shrank by 2 percent when 4 percent growth had been assumed, a market shock of minus 6 percentage points. The adjusted target is zero growth, and Appliances beat it. Relative performance evaluation removes the common shock, though it raises a question of its own: should managers also be rewarded or penalized for how well they anticipated and responded to the shock?

Some organizations deliberately hold managers accountable for partly uncontrollable results. The argument is that managers can often influence how much an uncontrollable factor affects their unit, for example by hedging currency exposure, diversifying suppliers or preparing for storms, and that holding them responsible encourages them to do so. The principle is better stated as: hold managers accountable for results they can influence, after removing factors they clearly cannot.

Segment reporting for internal control

Internal reports for profit and investment centers usually follow a layered format that separates costs by controllability and traceability. Contribution margin is sales minus variable costs. Controllable margin, sometimes called short-run performance margin, deducts the fixed costs the division manager controls, such as divisional staff salaries, advertising and discretionary spending. Segment margin, or divisional contribution, deducts further fixed costs traceable to the division but not controllable by its manager, such as depreciation on assets bought by head office or the salary of the division manager. Head-office costs allocated to divisions come last, if at all.

Each layer answers a different question. Controllable margin is the best measure of the manager's performance. Segment margin is the best measure of the division's economic performance, its contribution to covering common costs and earning a profit, and is the right basis for deciding whether to keep or close the division. Profit after allocated head-office costs reflects an arbitrary allocation and should not be used to evaluate either.

# Cormorant's 2026 divisional performance report (illustrative, thousands)
div = {
    #             sales   variable  ctrl fixed  traceable non-ctrl  allocated HQ
    "Appliances": (40_000, 26_000, 8_500, 2_500, 800),
    "Components": (24_000, 14_400, 5_800, 2_000, 600),
    "Services":   (12_000,  6_600, 4_000,   500, 400),
}
print(f"{'':<11}{'Contrib.':>9}{'Ctrl margin':>12}{'Segment':>9}{'After HQ':>9}")
tot = [0, 0, 0, 0]
for name, (s, v, cf, nc, hq) in div.items():
    cm = s - v
    ctrl = cm - cf
    seg = ctrl - nc
    after = seg - hq
    for i, x in enumerate((cm, ctrl, seg, after)):
        tot[i] += x
    print(f"{name:<11}{cm:>9,}{ctrl:>12,}{seg:>9,}{after:>9,}")
print(f"{'Total':<11}{tot[0]:>9,}{tot[1]:>12,}{tot[2]:>9,}{tot[3]:>9,}")

            Contrib. Ctrl margin  Segment After HQ
Appliances    14,000       5,500    3,000    2,200
Components     9,600       3,800    1,800    1,200
Services       5,400       1,400      900      500
Total         29,000      10,700    5,700    3,900

The report shows Appliances with a contribution margin of $14,000,000, a controllable margin of $5,500,000 and a segment margin of $3,000,000 (figures in thousands in the program). Components shows $3,800,000 controllable margin and $1,800,000 segment margin; Services $1,400,000 and $900,000. Together the divisions contribute $5,700,000 toward the $1,800,000 of head-office costs, leaving $3,900,000. The divisional segment margins of $3,000,000, $1,800,000 and $900,000 become the operating income figures used for ROI and residual income in lesson 4.

Allocating head-office costs

Whether to allocate head-office and shared-service costs to divisions is a perennial debate. Arguments for allocation include reminding managers that these costs exist and must be covered, encouraging divisions to monitor and question central costs, and approximating what the division would pay if it bought the services outside. Arguments against include that allocations are arbitrary, uncontrollable by divisional managers, and can distort decisions if managers treat allocated fixed costs as if they were avoidable.

A middle path is to charge divisions for services they use and can control, based on usage, such as information technology help-desk calls, legal hours or payroll transactions processed, and to leave general corporate costs, such as the chief executive's office and investor relations, unallocated. Usage-based chargebacks, at a rate set in advance, encourage divisions to use services wisely and give the shared-service center a revenue figure against which to manage its costs. The rate should normally be based on budgeted cost at practical capacity, as ACCT-222 explains for support departments, so that divisions are not charged for idle capacity or for the service center's inefficiency.

Shared services and centers of excellence

Large organizations increasingly consolidate transactional activities such as accounts payable, payroll and information technology support into shared-service centers, sometimes in lower-cost locations, and some outsource them entirely. A shared-service center is usually run as a cost center with service-level agreements, or sometimes as a profit center selling to internal customers at market-based prices. The control issues mirror those of transfer pricing (lessons 6 to 8): what price to charge, whether divisions may buy outside, and how to measure the center's efficiency and quality. Centers of excellence, by contrast, concentrate specialized expertise, such as data analytics or tax, and are usually managed as discretionary expense centers.

Structure, segments and external reporting

Internal responsibility centers often, but not always, match the operating segments a public company reports externally. Under ASC 280, which ACCT-311 covers, operating segments are identified by the management approach: the components whose results the chief operating decision maker reviews regularly, and ASU 2023-07 now requires disclosure of significant segment expenses provided to that decision maker. Internal reports therefore shape external disclosure.