Cost Escalation at V&A Dundee


How much does an ‘iconic’ building cost? Dan Anderson looks at the results of the recent investigation into the escalation of capital costs at the V&A Dundee and tries to answer this question. 

In January 2015, Dundee City Council appointed an independent procurement expert to investigate how the construction cost for the much anticipated V&A outpost in Dundee – designed by Japanese architect Kengo Kuma – exploded from £27 million at design competition stage to £80.1 million at last count.

John McClelland’s report has just been unveiled and it is a fascinating insight into all that can go wrong in planning and budgeting for a new destination. McClelland’s inquiry drew 15 major conclusions about the project, usefully summarised in the August 14 edition of The Courier. They describe a series of systemic failures in procurement, governance, project management and cost control.

The standout conclusion, however – which has been seized upon by the Architects Journal, The Scotsman and most of the broadcast media – is that the chosen design was never going to be deliverable within the original £27 million budget.  McClelland himself calls attention to this fact by saying:

“This project had, from the beginning, little prospect of being delivered for the original budget due to an accumulation of factors that mitigated against that being achieved. The largest single cause of the increase was an understatement of the original budget.”

It all paints the picture of a design competition that went awry. The project started with a reasonable budget, but the design competition produced a solution that exceeded the original aspiration. It suggests that jurors were so enamoured with the Kengo Kuma concept that they threw caution to the wind and selected it ahead of other, more modest and appropriate options. Any subsequent failure of project management or cost control may have magnified the problem, but – in effect – the die was cast from the moment the jury selected the iconic Kuma design.

The danger in how this conclusion is phrased or, more importantly, in how it is interpreted, is the implicit suggestion that an ‘iconic’ building was never going to be deliverable within a meagre £27 million budget.  Inasmuch as the iconic cultural building is now a mainstay of any self-respecting waterfront regeneration programme, it begs the question: just how much does a building have to cost to be iconic? Was it always fanciful to believe that a gallery of global renown could be built for £27 million? Was it the client’s ambition that escalated or was it a failure of project management that let the design process run amok?

At first glance, it is easy to see how Design Dundee Ltd – the partnership composed of The V&A, the City Council, Scottish Enterprise and two universities – fell into this trap. We entered the key size and cost metrics into our own database of 86 worldwide museum and art gallery capital projects – adjusted for relative exchange rates and inflation – and found that £27 million should have been enough to build a new gallery of international standard. As shown in Figure 1, this original estimate would have put the V&A Dundee in the company of some genuinely world class buildings, including – on a unit cost basis – the Aga Khan Museum in Toronto (architect: Fumihiko Maki) and the Centre Pompidou in Metz (architect: Shigeru Ban).


If we consider only those museums and galleries of commensurate scale – between, say, 5,000 sqm and 15,000 sqm – and look at them on a unit-cost basis, then it becomes even more apparent that the original budget should have been sufficient to deliver the signature building to which the partnership aspired. It just could not deliver that particular building.


The McClelland report suggests that this is precisely the type of international benchmarking process that the partnership went through in advance of the design competition.

“In summary, the review has concluded that the international benchmarking approach which formed the basis of the original project cost estimates was not sufficiently robust to address the unique challenges associated with Kengo Kuma’s iconic winning design for the V&A in Dundee”.

Indeed, the type of benchmarking described above is a very blunt instrument. It makes no allowances for the specific challenges of a particular site or design.

More importantly, it points to an important issue about design competitions that too many people often overlook. There is a world of difference between the appointment of a designer and the selection of a design. Jim Roberts, my colleague and fellow Fourth Street founder, can sometimes be evangelical about this. It is always his first question whenever we are asked to advise on design team procurement. Do you want a designer or a design? The V&A Dundee debacle illustrates just how important this distinction is.

If the purpose is to select a designer – with whom the final design brief will be finessed, agreed and budgeted – then it is not such a big deal if a competition entry is overly elaborate or ambitious. In most cases it isn’t even necessary for architects to produce anything more than notional ideas about a potential solution.  The point is to identify a designer that shares the client’s vision and values and who demonstrates a compatible approach and the right sensibility for the project. Is this somebody that the client can work with? Does he or she understand what the client is trying to achieve? The sometimes difficult process of pinning the architect down to a specific budget can follow.

This applies to commercial schemes as much as it does to cultural ones. A client of mine is a hugely experienced development director for a major London house builder. His approach is always to put very little restriction on the architect for the first iteration of any project. The way he sees it, he is paying for the architect’s imagination and creativity. He doesn’t want to stifle that too early in the process, so he does his best to not prematurely ‘shackle’ the architect to a budget. He very deliberately wants to see how the architect reacts to the site when freed of any budget constraint. Invariably, the first output of that process is something far more elaborate and expensive than he is ever going to pay for. But that’s okay.  He takes from that first draft the two or three kernels of an idea that he likes and builds a new brief – with a firm spatial plan and budget – around those qualities.

All of this works if the competition is set up to select a designer or a design team. It is fair to say, in that situation, that £4,500 per sqm is enough to buy an ‘iconic’ building. There is plenty of evidence from around the world to demonstrate that that kind of budget – in the right hands – can deliver some stunning architecture. If, on the other hand, the competition is set up to select a design or a fully formed concept, then the McClelland Report is absolutely right in suggesting that the costing of different schemes had to be much more robust than a simple international benchmarking exercise. Averages can, after all, be deceiving. While it may be okay to say that £4,500 per sqm is enough to deliver an iconic building, it is not okay to assume that that is enough to pay for any particular building. If the competition was meant to select a design solution rather than a design team, then the entries should have been much more rigorously scrutinised from a capital cost perspective before any decision was made.

photo credit: Heading In via photopin (license)