New research provides some surprising evidence regarding a perennial issue that bedevils capital projects: how to avoid cost underestimates that can lead to budget-busting cost overruns.
Commonly proffered advice for avoiding overruns addresses decision-makers’ tendency to escalate their commitment to undertakings that they themselves launched, even when they later face soaring expenditures. Avoid the escalation problem, the counsel goes, by choosing someone other than the original manager — someone less attached to the project — to decide on continued funding.
That would seem to be plain common sense. After all, if the party responsible for providing an undertaking’s original dollar estimate knows that someone not involved in the launch decision will later be deciding the project’s fate, wouldn’t that constrain a tendency to understate its cost?
“On the one hand, the prospect of facing a new superior at the continuation stage could make subordinates more cautious about understating costs in their initial budget proposals,” acknowledge the co-authors of the new study, Alexander Brüggen of Maastricht University and Joan L. Luft of Michigan State University
“They might anticipate that superiors who originally funded the projects would be sufficiently committed to them to overlook moderate cost overruns when deciding about continuing funding, but [that] new superiors would be more critical and more likely to cut off funding if the projects’ performance diverges too much from the subordinates’ predictions.”
On the other hand, the professors continue, “the prospect of a new, non-committed, critical superior making the continuation decision could decrease subordinates’ expectations of continuation funding and/or increase their uncertainty about this funding. As a result, subordinates under changing superiors could discount possible later-period reputation payoffs from accurate initial forecasting and focus primarily on gaining initial funding by understating costs.”
Indeed, an experiment the professors conducted to test the two alternatives suggests the latter to be the case. It involved 84 business students who interacted in groups of three over a computer network. The subjects were randomly assigned to be either subordinates or superiors throughout the experiment, which consisted of seven rounds and lasted about an hour.
In each round, two subordinates were directed to separately propose a capital project to the same superior, who chose to pursue one or the other. Both subordinates and superiors were compensated principally through a percentage of the realized profit per funded project, as determined by the difference between project revenue and cost.
Of primary interest to superiors, then, was to select projects that would yield the greatest profit. Of primary concern to subordinates was to get their projects chosen, which would motivate them to forecast the lowest possible cost (since cost would largely decide whether they were funded).
Unknown to superiors, subordinates each received private information on what their project would likely cost, an estimate that would guide their prediction to the superior. Superiors and subordinates also learned, midway through each round, how much the project was actually costing, at which point a decision would be made on whether to continue funding it.
Half of the subordinates had been informed at the outset that mid-course determinations would be made by the same superiors who originally chose to fund projects. The other half had been told that the mid-course determination would be made by a different executive.
Unsurprisingly, the more subordinates understated their costs, the more likely superiors were to choose their projects. But much more surprisingly, the biggest understatements occurred among subordinates who’d been told that a different executive would make the continued funding decision. In fact, the average gap was 45% higher for that group than for the group that was told that the original executive would make the decision.
The results “provide an important caveat to the previous literature’s observation that the practice of changing decision-makers improves capital budgeting performance by limiting the continuation of poorly performing projects,” the professors wrote.