Ann Minogue - Some of the touted alternatives to retention are not all they're cracked up to be – but there are other ways to overcome the ills that they seek to remedy
Whole books-worth of prose have been generated in this magazine on retention. Most recently, it reported the escalation in the Constructors Liaison Group's campaign against it, with up to 60 MPs had signing an early day motion proposing that it be outlawed.

James Bessey, in his column in Building (27 April), offers a number of arguments as to "why we should ditch retention". He argues, correctly, that even 5% retention is unlikely to cover the cost of making good any substantial defects. He suggests that the value of snagging works should be "totted up and settled by a payment notice" at practical completion. But, of course, this ignores the fact that payment is only due for work that has been "properly executed". Work that is incomplete or defective should not be included in valuations – there is no need for any payment notice. Most professional teams are already very wary of over-certifying payments due to the contractor and are frequently criticised for too cautious an approach.

James goes on to suggest "more attractive alternatives to retention" – he lists "performance bonds, guarantees or insurance". How far will the industry really thank him for these suggestions? The issue of performance bonds is of course old hat: "on default" bonds issued by surety companies are hardly a substitute for a cash retention given the protracted arguments that usually follow a call. The "bond in lieu of retention" published by the JCT – essentially an on-demand bond – has not proved widely popular with the contractors, probably because of the cost of securing them. "Guarantees", presumably from the contractor, raise the question rather than providing the answer. And latent defects insurance cover does not provide a comprehensive solution to the problem of latent defects in buildings (and provides no solution at all to defects or snagging items known at practical completion).

James concludes that construction lawyers might borrow from computer software agreements, and he advocates the use of stage payments. Again, experience shows that most contractors and subcontractors shy away from them, largely because, if operated strictly, they deprive them of payment where the "stage" is not completed. In other words, they can only operate to defer payment compared with a valuation-basis of payment. Likewise, a large stage payment triggered by completion of "snagging" is no more than a retention in disguise.

How should the industry seek to solve the ills for which retention is such a crude and defective cure?

So how should the industry seek to solve the ills for which retention is such a crude and defective cure? One answer obviously lies in the closer commercial relationships that exist between companies that do repeat business with each other. Serious defects are not usually a good recommendation for further work.

The emphasis on off-site production, the elimination of waste and "right first time" strategies will all help to reduce defects lists at practical completion. Clients' (and therefore their architects') tolerance of lengthy lists attached to the certificate of practical completion is undoubtedly diminishing – the books are being replaced by pamphlets. And the incentives are there – zero defects at practical completion will encourage more clients to abandon retentions, and will mean that the full contract sum will be paid on the next certificate after it.

The more troublesome issues are latent defects or, perhaps more accurately, apparently minor defects that are revealed to be much more significant during the defects liability period. In these cases, neither valuation nor retention can help. Commercial relationships might, but even the most reconstructed contractor will argue about liability where remedial costs are high. Surely, in this case, the Construction Act provides clients with a remedy – adjudication against the contractor with an enforceable decision in 28 days.