Media Briefings

‘INTERIM PERFORMANCE EVALUATION’: New analysis of its impact on the eventual outcomes of investment projects

  • Published Date: June 2013

The value of assessing how well a project is going before it is completed depends on whether the intention is to continue whatever the results of the ‘interim performance evaluation’. That is the one of the findings of research by Bin Chen and Stephen Chiu, published in the June 2013 issue of the Economic Journal. Their results have implications for how organisations give feedback to employees and suppliers – and how they design compensation packages.

The study notes that in most real-life situations, the successful production of a good requires a sequence of investments. Building a bridge, for example, involves planning and a feasibility analysis at an early stage and then the actual construction of the bridge at a later stage. Likewise, the completion of a doctoral degree requires the student to take course work, to pass a qualifying examination, to conduct research and to write a thesis of reasonable quality.

In many cases, interim performance evaluations (IPEs) are performed. In the case of bridge building, the results of the feasibility analysis may suggest the impracticality of building the bridge. In the case of doctoral studies, the qualifying examination may lead to the student being dropped from the programme.

The merit of such IPEs is clear, as the optimal continuation actions vary depending on the IPE outcome. A less clear issue is whether they are still valuable when the principal intends to continue with the same course of action regardless of the IPE outcome. These issues are also related to two main purposes of conducting IPEs: providing performance feedback; and designing compensation packages.

Despite the popularity of IPEs in practice, both management researchers and economists warn against the use of IPEs. Many companies that have experimented with reward systems using informative feedback have found such systems to be failures. A leading expert also warns that ‘subjective measures and milestones may provide more effective incentives for innovation than do the accounting numbers, but using them to provide very intense incentives is certainly problematic.’

This research studies the issue as a two-stage ‘principal-agent’ problem. The principal solicits the agent’s help to produce a final good, which may turn out to be a success or a failure. The probability of success depends on two non-observable sequential efforts that the agent makes.

The analysis focuses on situations where the two efforts exhibit complementarity – that is, a higher level of effort in stage one makes the effort in stage two more productive and vice versa. It is also assumed that while both parties are risk-neutral, the agent is subject to limited liability, a reasonable assumption in many applications. The benchmark model is a so-called traditional contract, where the payment made to the agent is dependent only on the quality of the final good.

Given this benchmark model, the new study looks at a contract in which payments to the agent depend not only on the quality of the final product but also on the outcome of an IPE. For simplicity, it is assumed that the IPE is conducted at a low cost, and that the IPE informs the principal about the quality of the interim product, which is correlated with the probability of success of the final good.

A positive IPE outcome means a high probability of success and, given the assumption of effort complementarity, the increase in the success probability due to the second-stage effort will be large and this thus boosts the agent’s morale – the beneficial effect.

On the other hand, a negative IPE outcome means a low probability of success, implying the increase in success probability due to the second-stage effort is small and the agent’s morale is damaged – the harmful effect.

There are two main results. First, the researchers show that when the same course of continuation effort is intended regardless of the IPE outcome, an IPE contract might be more desirable than a traditional contract only if the beneficial effect dominates the harmful effect; otherwise, conducting an IPE will worsen the principal’s payoff. This is consistent with a typical remark in the profession that feedback does ‘tremendous damage to employee morale and self-esteem’.

Second, even if the IPE outcome is subjective, the researchers still find parameter values over which conducting IPEs is more profitable than not conducting them. But because the principal may have an incentive to lie about the findings, IPEs are less likely to be desirable. As a result, concealing feedback or revealing limited information in IPEs may be preferred. This result resonates well with ‘targeted inaccuracy’ in rating, as documented in the research literature of organisational behaviour.


Notes for editors: ‘Interim Performance Evaluation in Contract Design’ by Bin Chen and Stephen Chiu is published in the June 2013 issue of the Economic Journal.

Bin Chen is at Sun Yat-sen University. Stephen Chiu is at the University of Hong Kong.

For further information: contact Stephen Chiu on +852 2859 1056 (email:; or Romesh Vaitilingam on +44-7768-661095 (email:; Twitter: @econromesh).