Budgeting for control and its behavioral effects

Budgets are the most widely used formal control in organizations of every size. ACCT-102 showed how to build a master budget and ACCT-222 how to analyze variances. This lesson asks a different question: what do budgets do to the people who prepare them and are judged against them? A budget is a plan, but it is also a target, a commitment, a basis for rewards and a weapon in negotiations over resources. The same document serving all these purposes creates predictable tensions and predictable games.

The purposes of budgets

Budgets serve at least six purposes. They support planning, forcing managers to think ahead about sales, costs, capacity and cash. They coordinate the activities of different units, so that Components plans to build as many motors as Appliances plans to need. They communicate senior management's priorities. They motivate, by setting targets for managers to strive toward. They provide a benchmark for evaluating performance. And they authorize spending, giving managers permission to commit resources up to an amount.

The trouble is that these purposes conflict. For planning and coordination, a budget should be the most likely outcome, an unbiased forecast. For motivation, research suggests a target should be challenging. For evaluation, managers prefer targets they are confident of meeting, because missing them carries penalties. For authorization, managers want generous amounts. When one number tries to serve all four, managers bias it toward whichever purpose most affects them, usually evaluation and rewards. Some organizations separate the numbers: a forecast for planning, a stretch target for motivation, and a lower threshold for bonuses.

How difficult should targets be?

Goal-setting research, much of it associated with Edwin Locke and Gary Latham, finds that specific, difficult goals lead to higher performance than vague or easy goals, provided people accept the goals and receive feedback. Beyond some point, though, goals become so difficult that people reject them, give up, or turn to manipulation to meet them. Budget targets in practice are usually set to be highly achievable. Merchant and Van der Stede report that many companies aim for targets that a competent manager will achieve most of the time, often cited as eight or nine years in ten, and use bonus schedules that reward performance above target to supply the stretch.

Highly achievable targets have advantages. They make the plan a more reliable basis for coordination and commitments to investors and lenders. They reduce the incentive to manipulate results, since a manager who expects to make target has less reason to cut corners. They protect managers from failure in bad years and maintain their confidence. They also give senior managers a signal: when a unit misses a highly achievable target, something serious has gone wrong. The cost is lower motivation for some people, which is why the bonus schedule usually continues to reward performance above target.

Top-down and participative budgeting

In top-down budgeting, senior managers set targets and pass them down. It is fast, ensures targets reflect strategy and can stretch units, but it ignores local knowledge and may produce targets that lower-level managers do not accept. In bottom-up or participative budgeting, the managers who will be held accountable prepare their own budgets, which are then reviewed and negotiated. Participation uses local information, increases acceptance and commitment, and develops managers' planning skills.

Participation has its own dangers. Managers with better information than their superiors, an information asymmetry, may use participation to build in slack. The negotiation can take a long time. And participation that is not genuine, where senior managers ask for input but then impose their own figures, can do more damage than no participation at all. Most organizations use a mix: top-down guidelines and targets for the company as a whole, followed by participative development of unit budgets within them, then negotiation to reconcile the two.

Budget slack

Budget slack, or padding, is the deliberate understatement of revenues or overstatement of costs in a budget so that it is easier to meet. Managers create slack to protect themselves against uncertainty, to make sure they earn bonuses, to look good, and to have resources available for things they want to do without asking. Slack is not always harmful: a modest buffer can absorb shocks and allow experimentation. But large or hidden slack misleads planners, ties up resources and rewards managers for negotiating skill rather than performance.

Slack is hard to detect directly, because only the manager knows what is truly achievable. Indirect signs include a unit that consistently beats its budget by similar margins, costs that rise just before the budget is set, and revenue forecasts that fall well below trends in the market. Comparing budgets with independent forecasts, benchmarks or past performance helps. So does a planning culture in which honest forecasts are not punished.

# Components: budgeted versus actual controllable margin, 2022-2026 (thousands)
years = [2022, 2023, 2024, 2025, 2026]
budget = [3_200, 3_350, 3_500, 3_600, 3_700]
actual = [3_520, 3_690, 3_820, 3_950, 3_800]
over = []
for y, b, a in zip(years, budget, actual):
    pct = (a - b) / b
    over.append(pct)
    print(f"{y}: budget {b:,}  actual {a:,}  beat by {pct:5.1%}")
avg = sum(over) / len(over)
print(f"Average beat: {avg:.1%}; years beaten: {sum(p > 0 for p in over)} of {len(over)}")

2022: budget 3,200  actual 3,520  beat by 10.0%
2023: budget 3,350  actual 3,690  beat by 10.1%
2024: budget 3,500  actual 3,820  beat by  9.1%
2025: budget 3,600  actual 3,950  beat by  9.7%
2026: budget 3,700  actual 3,800  beat by  2.7%
Average beat: 8.3%; years beaten: 5 of 5

Components beat its controllable margin budget in all five years, by about 10 percent in each of the first four and by 2.7 percent in 2026, an average of 8.3 percent. A consistent pattern like this is a classic sign of slack, although it could also reflect genuine improvement that the budget process failed to anticipate. The 2026 result, the lowest beat, may show that head office began to squeeze the slack out, or simply that 2026 was a harder year. A controller would look at the reasons behind each year's beat before concluding.

Budget games

Managers play a range of budget games. Sandbagging is understating what can be achieved. The ratchet effect describes how this year's actual results become next year's starting point, so a manager who performs very well is punished with a harder target; anticipating this, managers may deliberately hold back performance in good years. The hockey stick is a forecast that shows poor results in the near term and a sharp improvement later, often beyond the manager's tenure. Use-it-or-lose-it spending happens when unspent budget is withdrawn at year end, prompting a rush of spending in the last months. Managers may also shift revenues or costs between periods, through channel stuffing, delayed maintenance or early recognition, to hit targets in the period that matters for their bonus.

Truth-inducing schemes

If the problem is that managers have private information and an incentive to understate it, one response is to design rewards so that honest forecasting pays best. A truth-inducing scheme, sometimes called the Soviet incentive scheme or Weitzman scheme, rewards a manager both for the level of the forecast and for performance relative to the forecast, with a penalty for falling short that is larger than the reward for the forecast itself, which in turn is larger than the reward for exceeding it.

# A truth-inducing bonus scheme (Weitzman type), bonus in thousands
b1, b2, b3 = 0.004, 0.002, 0.006   # forecast reward, overshoot reward, shortfall penalty
actual = 5_500                      # the controllable margin the manager expects
def bonus(forecast, actual):
    gap = actual - forecast
    return b1 * forecast + (b2 * gap if gap >= 0 else b3 * gap)
for f in (5_000, 5_250, 5_500, 5_750, 6_000):
    print(f"Forecast {f:,}: bonus {bonus(f, actual):5.1f}")
best = max(range(4_500, 6_501, 50), key=lambda f: bonus(f, actual))
print("Best forecast:", f"{best:,}")

Forecast 5,000: bonus  21.0
Forecast 5,250: bonus  21.5
Forecast 5,500: bonus  22.0
Forecast 5,750: bonus  21.5
Forecast 6,000: bonus  21.0
Best forecast: 5,500

With an expected actual of $5,500,000, forecasting exactly $5,500,000 gives the highest bonus, $22,000. Forecasting low earns less because each dollar of forecast is worth more than each dollar of overshoot; forecasting high earns less because the shortfall penalty is steepest. The scheme works neatly on paper but is rarely used in its pure form, because outcomes are uncertain, managers may not understand it, and it can seem complex and manipulative. Its lesson is general, though: reward accurate forecasting, and do not punish managers for honest information.

Using budgets in evaluation

Anthony Hopwood's study in 1972 distinguished three styles of using budgets to evaluate managers. In a budget-constrained style, superiors focus mainly on whether the manager met the budget in the short run. In a profit-conscious style, superiors look at the manager's contribution to long-run effectiveness, using budget data flexibly. In a non-accounting style, budget data play little part. Hopwood found the budget-constrained style associated with higher job tension, poorer relations with colleagues and more manipulation of accounting data. Later research has refined this, finding that the effect depends on task uncertainty: a rigid emphasis on budgets is more harmful where outcomes are highly uncertain and less so in routine settings.

The practical lesson is that how budgets are used matters as much as how they are set. Flexible budgets that adjust for actual activity (ACCT-222), discussion of the reasons for variances rather than mechanical penalties, and attention to long-term indicators alongside the budget all reduce the dysfunctional effects.

Beyond budgeting and rolling forecasts

Critics argue that the annual fixed budget is so loaded with conflicting purposes and games that it does more harm than good. The Beyond Budgeting movement, led by Jeremy Hope and Robin Fraser from the late 1990s, proposed abandoning fixed annual budgets in favor of relative targets (against peers or competitors), rolling forecasts used only for planning, dynamic resource allocation on demand, and decentralized decision making within clear values and boundaries. The Swedish bank Handelsbanken, which abandoned budgets in the 1970s, is the best-known example; the Norwegian energy company Equinor, formerly Statoil, replaced its traditional budget with a process called Ambition to Action.

Few organizations have abandoned budgets entirely, but many have adopted parts of the approach, particularly rolling forecasts. A rolling forecast extends a fixed number of periods ahead, often four to six quarters, and is updated every quarter, so the planning horizon never shrinks as the year progresses. Because it is separated from targets and rewards, it can be a more honest view of the future.

# A four-quarter rolling forecast for Appliances unit sales (thousands of units)
history = {"2026Q1": 29.0, "2026Q2": 34.5, "2026Q3": 38.0}
season = {"Q1": 0.85, "Q2": 1.00, "Q3": 1.10, "Q4": 1.05}   # seasonal indexes
base = sum(v / season[k[-2:]] for k, v in history.items()) / len(history)
print(f"Deseasonalized run rate: {base:.1f}")
growth = 0.01   # assumed quarterly growth
quarters = ["2026Q4", "2027Q1", "2027Q2", "2027Q3"]
for i, q in enumerate(quarters, start=1):
    f = base * (1 + growth) ** i * season[q[-2:]]
    print(f"{q}: {f:5.1f}")

Deseasonalized run rate: 34.4
2026Q4:  36.5
2027Q1:  29.8
2027Q2:  35.4
2027Q3:  39.4

The forecast removes seasonality from the three quarters of actual sales to find a run rate of about 34,400 units per quarter, applies 1 percent quarterly growth and puts the seasonality back. It projects about 36,500 units in the fourth quarter, dropping to about 29,800 in the first quarter of 2027 before recovering. Next quarter, when actual fourth-quarter sales are known, the forecast will roll forward to include the third quarter of 2027.

Budgets in public and not-for-profit organizations

In governments and many not-for-profit organizations, the budget is a legal authorization to spend, approved by a legislature or board, and spending beyond it may be unlawful. Federal agencies operate under appropriations enacted by Congress, and the Antideficiency Act prohibits obligating funds beyond them. Use-it-or-lose-it behavior is common where annual appropriations expire at fiscal year end. ACCT-412 covers public sector and not-for-profit accounting in detail.