Yongjian Ke

Relational contracting in public projects: what four cities show

· 7 min

Public construction contracts are written to keep people apart. Open tendering, audit trails and rules against favouritism all work against the close working relationships that a private client can build up over repeat business. So the question is whether relational contracting, an idea developed mostly in private-sector settings, has anything to offer a public client who cannot promise anyone future work. Six papers answer it from four cities at once. They come from a study led by Professor Florence Ling at the National University of Singapore, funded by Singapore's Ministry of Education, on which I worked as a research fellow.

One questionnaire, four cities

The design is simple and it is what makes the set worth reading together. The same structured questionnaire was administered in Beijing, Hong Kong, Sydney and Singapore during 2011, producing 244 valid responses (59, 51, 30 and 104 respectively). Each respondent reported on one completed public project they had worked on: its cost, time, quality and client satisfaction outcomes, the extent to which each of 88 relational practices was present, and what drove or blocked collaboration. Holding the instrument constant and varying the institutional setting is the point of the exercise.

The systems the relationships sit in

The first paper maps the ground: how public projects are funded, approved, procured and evaluated in each city, checked against the survey data (Ke, Ling, & Ning, 2013). Beijing had the largest market and the least visible process, with project appraisal and post-implementation information generally unpublished and no formal accounting for contingent liabilities. Singapore and Hong Kong, having no sub-central tier of government, operated under considerably simpler regulation than Beijing or Sydney. Design-bid-build remained dominant in Beijing, Hong Kong and Singapore, while Sydney relied mainly on selective bidding.

The performance picture was uniform in a way the institutional picture was not. Around 60 per cent of projects in all four cities ran over budget and over schedule, with mean cost overruns between 5 and 13 per cent. Quality and client satisfaction were rated well in Hong Kong, Sydney and Singapore, and significantly lower in Beijing. Whatever the relational findings explain, they sit on top of a cost and schedule problem that is shared rather than local.

Which practices are present, and which useful ones are missing

The second paper takes the Beijing and Hong Kong data and asks which of the 88 practices actually track project outcomes (Ling, Ke, Kumaraswamy, & Wang, 2014). Thirty-two did in Beijing and 56 in Hong Kong. That the practices operate at all is the theoretical result: relational contract theory holds in projects procured by open tender, where the promise of future exchange, which the theory usually relies on, is absent.

Open tendering does not turn a project into a discrete transaction. It only removes the promise of future work.

The more useful move is a category the paper calls key practices: those significantly correlated with performance but not currently adopted to a significant extent. There were 8 in Beijing and 11 in Hong Kong, and they point at what to start doing rather than what already works. Beijing's list included the client's own knowledge of project processes and investment by both clients and consultants in collaborative working skills. Hong Kong's included equitable risk sharing, effective team building, and consultants' readiness to compromise on unclear issues.

What predicts relationship quality

The third paper narrows to Hong Kong and builds a model (Ling, Ning, Ke, & Kumaraswamy, 2013). Of 21 practices tested, 14 were adopted to a significant extent, and these grouped into three factors. Only one, preservation of relationships, was significantly associated with the strength of relationships at the end of the project. It comprises long-term commitment, mutual trust, aligned objectives, teamwork attitude, mutual understanding, information sharing, flexibility when situations change, and willingness to compromise on unclear issues. Trust carried the largest weight.

The two factors that did not reach significance are the interesting ones: harmonisation of conflicts (dispute resolution mechanisms, joint monitoring) and procedural fairness (performance appraisal systems, clear division of responsibility). Those are precisely the things a public client can write into a contract. The model was checked against the improvement works to Fuk Man Road Nullah in Sai Kung, Hong Kong's first NEC project, which finished about six months early with a shared saving. That is one case, and the model assumes the effects are linear, so I would read the result as an ordered checklist rather than a formula.

The Australian position, and one uncomfortable result

The Sydney paper asks a different question: does the contracting stance a party adopts affect what happens (Ke, Ling, & Zou, 2015)? Australian respondents sat in the middle on all three measures, leaning neither to rational choice nor to relational choice, which supports the view that contracting behaviour is a spectrum rather than a category. Relationship quality improved where decisions leaned relational and where trouble was handled through networks rather than contract terms.

Cost did not. Projects where decisions leaned relational reported worse cost overrun, which the interviews attributed to the extra staff, time and resources needed to align parties. With 30 responses and correlation only, that is a signal rather than a finding, but it is the kind of result worth keeping. Of the 20 practices correlated with contracting stance, 16 involved clients or consultants, which is where the paper lands: these two parties set the tone, because they are in the project from inception to completion.

Shared drivers, local barriers

The fourth paper compares Beijing and Sydney on motivation (Ling, Ong, Ke, Wang, & Zou, 2014). Eighteen factors drove the adoption of relational practices in both cities. Six were significantly stronger in Sydney, including reducing delivery time, improving design and quality, and supporting innovation. Two were stronger in Beijing: responding to competitors and seizing new market opportunities.

The barriers had no overlap at all. Beijing showed six significant ones, including a lack of training and guidance, a lack of client initiative, client representatives without authority to decide, and an industry culture that preserves the status quo. Sydney showed exactly one: public sector accountability concerns. That asymmetry is the most portable result in the set. The reasons to work collaboratively are broadly shared, so what has to be fixed is institutional, and therefore local.

Large projects, revisited

The last paper returns to the data with a size threshold, comparing 55 large public projects in Singapore with 33 in China (Ke, Ling, Ning, & Zhang, 2019). Both leaned contractual: decisions made on rational grounds, profit ahead of relationship, contract terms used to guard against trouble. Both also used relational practices at the same time, which is the substantive point. The two mechanisms run together rather than as alternatives. Singapore scored higher on commitment, teamwork and information sharing; China scored higher on prior familiarity and existing relationships between parties.

The recommendation runs against intuition. The paper suggests that a small, developed, high-transparency setting can afford to be more relational, while a large developing one should be clearer and more contractual, on the grounds that arm's length dealing moderates a heavily networked business culture. The differences are read through Hofstede's national culture scores, which is an explanation applied after the results rather than one the study tested.

What the set adds up to

Three things hold across the four cities. Formal and relational governance operate together, and public projects are no exception despite the constraints they work under. The motivation to collaborate is broadly similar everywhere, so the binding constraint is institutional. And clients and consultants set the tone, which puts most of the practical burden on the public side of the table.

What follows for a public client is unglamorous: build in-house knowledge of project processes, invest in collaborative working skills, share project information, define risk sharing equitably rather than defensively, and behave consistently across projects even when each new tender must be open.

What it does not settle

All of this rests on cross-sectional perception data collected in 2011, with one respondent per project and no verification of cost and time against records. The analysis is correlational, so the direction of the relationship is untested: relational practices may produce good outcomes, or projects that go well may simply be remembered warmly. Response rates were low outside Beijing, and the Sydney sample of 30 is at the floor of what supports statistical work. Procurement has also moved since. Alliance and collaborative contracting in Australia and NEC-style contracts in Hong Kong are more common now than they were then, so the baseline these papers describe has shifted.

Journal papers discussed on this page