Roger Joby and Professor David Bryde
The research question that we asked was “Does Principal Agent Theory provide an explanation for different levels of project success?” Principal Agent Theory can be used to describe and explain the behavioural problems that you may experience in outsourced projects, where the Principal is the Project Sponsor and the Agent is the Project Supplier(s).
We focused on looking at:
- Conflict over goals – Sponsor and Suppliers have different business models and therefore different business goals. The aim is to take steps to either align these goals or at the very least recognise the differences and try to accommodate them within your project.
- Opportunistic behaviour – both Sponsors and Suppliers may seek to gain a personal or an organisational advantage which may be to the detriment of the project.
- Asymmetry of information – is there one shared version of the truth?
The first issue that we faced was defining “project success”. Just relying on the iron triangle of time cost and quality did not seem adequate, so we chose Sponsor/Supplier satisfaction. This definition obviously includes the issues of time cost and quality, though not in a rigid and easily quantifiable way. Satisfaction also takes account of the how good the relationship was. A project that is a few days late or a little bit over budget is not necessarily perceived by the Sponsor as a failure and a project that doesn’t make the projected project margin is not necessarily perceived by a Supplier as a failure.
We looked at four projects which had varying levels of success and were in or close to the handover phase. Two projects involved clinical research and two were major construction projects. All four were classified as Type 1 projects according to the Turner & Cochrane’s Goals-and-methods matrix, which means they had relatively well-defined goals and methods to achieve those goals.
The source data were collected as recorded interviews of the key personnel in both the Sponsor and the Supplier organisations. We also collected secondary date from project documentation. The interviews were transcribed and the comments were coded as either demonstrating a positive response to an Agency issues or demonstrating a negative response to an agency issue. We then calculated the % of positive and negative comments for each of the three agency problems described above: conflict over goals, opportunistic behaviour and asymmetry of information.
Below we provide a summary of the analysis of the interview data. Behaviour linked to Agency problems can be displayed by both Principal and Agents so in the results below we have not attempted to differentiate between Sponsors and Suppliers in terms of positive v negative comments.
Project A: “Clinical 1”
This project was a global clinical study looking at the treatment on non-small cell lung cancer. This project was considered by both Sponsor and Supplier as a success. This is supported by the results in the above graph. There were virtually no negative comments at all in terms of Agency issues. In fact the relationship between the key people from the Sponsor and the Supplier was so good that the thought of agency problems being present did not even occur to them. There was little goal conflict and a complete absence of demonstrable opportunistic behaviour (or the suspicion of it by one party) and, whilst there was some withholding of information (so asymmetry was present) this was not perceived to be a problem (trust was high).
Project B: “Clinical 2”
Project B was a complete contrast to Project A in terms of levels of satisfaction. It involved the treatment of a rare disease and was considered a total disaster by both sides. If the first project was an example of a symbiotic win-win relationship then this would be classified as toxic lose-lose one. As you can see from the above graph the negative comments far outweigh any positive ones. Agency issues were largely commented upon from the point of view of the harm they were causing. They did use tools like Earned Value. Yet from the start they did not agree on the baseline and as the project progressed and the baseline proved to be widely inaccurate in terms of the costs and timings their positions just became more entrenched. Goals diverged even further, opportunistic behaviour was either practiced or suspected and information was increasingly not shared, leading to ever high levels of information asymmetry.
Project C: “Construction 1” Tis Project C was a large construction project where the Sponsor employed a project management (PM) consultancy company to manage the project. There was awareness from the start of the potential for agency problems to be present, even if that awareness was not articulated in those terms. The PM consultants put a lot of effort into mitigating for Agency issues i.e. they spent a lot of time making sure that their Suppliers were happy about the baseline, they developed a highly visible information system incorporating earned value analysis (asymmetry of information), and they made sure that suppliers were paid what they were due and in a timely fashion (showing an appreciation of the different business goals of their suppliers). The Sponsor considered this project to be a success, whereas some of the suppliers were less enthusiastic. The graphic is predominantly positive but not quite at the level of Project A. It is a Qualified win-win.
Project D: “Construction 2”
Project D was both late and over budget. The main contractor (Supplier) went into administration during the build phase, which increased the Sponsor’s oversight costs as well. The Sponsor did not consider this as a wholly successful project whereas the Supplier was at least pleased that they were retained until the end of the project despite their financial problems. The analysis shows that there was a high degree of negative agency behaviour although it has not reached the levels displayed in the previous clinical example. It is probably a lose/win situation, though with some areas of satisfaction for the Sponsor, such as the fact that they displayed their Corporate Responsibility credentials by supporting a local business that was in trouble.
So, to come back to the original research question: “does Principal Agent Theory provide an explanation for different levels of project success?” on the evidence gathered so far we would say yes.
The next question that we ask relates to whether these finding have any practical use, and our answer to this is also yes. For example, the findings could help from both a predictive and a diagnostic point of view. At the project planning stage a company could carry out a “Principal/Agent Health Check” – as part of wider project health checks that are routinely undertaken. This health check could include analysis of key project documentation and documented practices i.e. the contract, the communication plan the procurement plan, the engagement practices to see if there are elements of these documents and practices that are likely to lead to Agency-related problems, and of course recommending solutions. From a diagnostic point of view a similar exercise could also be carried out on a project that is underway and recommendations also made to rectify identified problems and to hence increase the probability of project success. Perhaps the activity of a Principal/Agent Health check could be established as a recognised project management process linked to the project planning and control activities?
Our next step is to continue the research and see if the Health Check idea is effective and practical. We are looking for companies to collaborate with in terms of testing in practice our “Principal/Agent Health Check” tool and if you would be interested in helping us please contact myself or Professor David Bryde at one of the email addresses below.