The Week

The Week 16 February 2024

Charlotte Pickles

Too late to make last week's edition of The Week, last Friday the Department for Science, Innovation and Technology published Lord Willetts' 'Independent Review of the DSIT business case and approvals process'. And with a title like that, it was bound to be a cracker!

In fact, we couldn't recommend it more highly. At just 22 pages plus a summary list of the recommendations, it is, for a government publication, uncharacteristically succinct. Lord Willetts clearly and compelling lays out the problems with the current model and then puts forward clear, implementable recommendations.

Right from the off the former cabinet minister doesn't mince his words:

“The [business case] BC was developed for conventional public spending — buying services and building stuff — not for promoting inherently uncertain Science and Innovation. It assumes more capacity to forecast costs and benefits than is possible when the Government is funding innovative and risky R&D. Moreover, the BC became in BEIS and DCMS a slow, time-consuming, labour-intensive bureaucratic process, producing long reports which make it hard to extract the key information and issues for decision – a source of frustration to the Treasury itself. Such processes are a significant barrier to delivering the Government’s Science Superpower strategy.”

So the department is using a one-size-fits-all model that's an inappropriate match for the task. It can't handle risk, and the process is cumbersome and time-consuming. The final product is barely useable and an ineffective tool for decision-making. Worst of all, the whole thing is actually counterproductive in achieving the government goal it's designed for.

Thank goodness DSIT commissioned the review, and hats off to permanent secretary Sarah Munby for backing it in her foreword — “I am personally committed to making this change real in DSIT”. She also argues, rightly, that the proposed reforms should act as a blueprint for business cases across government.

But the short report is particularly powerful because so many of the problems identified in DSIT's business case and approvals processes are systemic issues that plague Whitehall's everyday operations. It's a microcosm of the whole of Whitehall. Here's just four examples:

1. The universal business case model, underpinned by the Treasury's Green Book, doesn't work for this inherently risky area of public investment. Just like the rigid civil service recruitment and HR processes don't work for certain roles & issues. Or the slow, siloed, bureaucratic approach to problem-solving doesn't work for urgent crises (and so inevitably has to be dismantled, but despite everyone reflecting how much better that more streamlined model is, the system immediately reverts to the norm).

2. Business cases are onerous and counterproductive and actively work against good decision-making. The template is 40 pages (seriously?!), and a sample of business cases found they averaged 249 pages. Much info is “superfluous” and the prose of the generalist could often be replaced by a simple table (but isn't) — what non-expert, super busy senior leader is going to see that as a helpful decision-making tool? Just like the copious pages of paper slipped in ministerial Red Boxes, stuffed with unnecessary information and lacking the clarity of decision-making thinking (ask any minister trying to make their way through that box, bleary eyed after a 10+ our day). Or some of the government procurements requiring bids stretching to thousands of pages.

3. The work of producing these mammoth documents is often “farmed out to new recruits who then work on them with all the thoroughness of the novice trying to impress but with little experience of what is actually needed.” A common frustration across Whitehall where bright young generalists are tasked with everything from technical analysis to project delivery. Which is not to say there isn't a role for the generalists — there very much is — but that when specialist knowledge or experience is needed, specialists should be deployed.

4. The approvals process for business cases can involve more than 20 people across at least 10 approval points. Layer upon layer of clearance helping to explain why the process can take “well over a year” (by which time specific tech may be obsolete). A familiar tale for anyone whose worked in Whitehall. In our publication 'Civil Unrest', the multi-layered clearance process was identified as “just ridiculous”… “‘check with somebody who checks with their manager who checks with their manager’ — you know, four or five lines of checking... if they are not happy with the people they’re paying 45 grand to make decisions, [then] get shot because you are wasting public money.”

Lord Willetts puts forward six sensible recommendations for addressing this farcical and damaging business case model in DSIT (have a 12-page cap, use business cases at a programme rather than a project level, increase delegated authority levels, produce supplementary guidance specifically for tech in the Green Book, use business cases to inform spending review decisions, and work with Cabinet Office and the NAO).

Now we just need equivalent solutions for the rest of the machine.