Sometimes we are so focused on big changes that we miss the mundane ones. But we shouldn’t. Sometimes a mundane change can tell us something about the character of a set of arrangements. 

The new UK Government has just issued a new version of the Ministerial Code (‘the Code’). The Code has a quasi-constitutional status, serving partly as an ethics code for Ministers (see eg., Part A. Ministers’ Standards of Conduct) and partly as a set of administrative procedures that Ministers should follow (eg., Part C. Ministers’ Procedures of Government). The Code is periodically reissued whenever a new Prime Minister takes office; and there have now been twelve iterations since 1997 (for previous versions, see here). 

Others may talk about the big picture. I want to focus on one particular section- paragraph 6.2. The key part states: 

‘Cabinet ministers may appoint special advisers to support them and their department. … All appointments require the prior written approval of the Prime Minister…’

Compare this with the equivalent clause in the previous 2022 version (para 3.2—hereafter, ‘the cap’):

‘With the exception of the Prime Minister, Cabinet Ministers may each appoint up to two special advisers. … All appointments, including exceptions to this rule, require the prior written approval of the Prime Minister…’

So what? We have a rather mundane paragraph which removes the cap on the number of special advisers (‘spads’) for Ministers. But what makes the change interesting is that the cap had not changed in substance in at least 25 years (and potentially half a century): it was an example of constitutional boilerplate.

This requires explanation. In the UK system, the vast bulk of the civil service is non-partisan and impartial in character: civil servants support the incumbent government, regardless of their own personal beliefs. Spads are temporary civil servants, personal appointments of a minister, who abide by the Civil Service Code but are exempted from the requirement to be politically impartial. Their role is to provide the partisan angle and sometimes the expertise that ordinary civil servants cannot. 

Spads have a long history: they have been around for decades. The key statute governing the UK civil service, the Constitutional Reform and Governance Act 2010, recognises them as part of government (see s15). But they are often treated with suspicion because they are exceptional, being partisan and personal appointments of the Minister. They remain controversial because they are thought to have too much power, and much more than some ministers—Dominic Cummings, Jonathan Powell and Ed Balls are obvious examples. 

The cap on spads has a history almost as long as spads themselves. It was there in the first 1997 Code (see here, para 48), published over a quarter of a century ago, and has remained, albeit with slight tweaks over the years, with successive PMs unwilling to explicitly raise the cap for fear of political fallout. But the cap’s origin lies with PM Harold Wilson (PM 1964-70; 1974-76), who did more than any other PM to institutionalise spads into British government. He introduced the cap of two spads per Cabinet Minister in 1975 because he worried about senior ministers using their personal appointments to undermine him and his government. 

So the cap was partly about political management; but it also embodied a fundamental and ongoing tension of Cabinet government: the need to balance the autonomy and needs of senior ministers against the responsibility of the PM to maintain the collective nature of government. And spad choice could and can really matter for how government is run, and its character. We can think of Dominic Cummings, for instance. At least one senior minister, Sajid Javid (Chancellor, 2019), made headlines after he resigned over an attempt to force him to dismiss his own spads. And in the past Damien McBride (a spad of Gordon Brown’s) made smears against Labour Ministers seen as hostile to his Minister, causing much discontent within government. 

Previous iterations of para 6.2 acknowledged that there could be exceptions to the cap—most notably, the Prime Minister, who has always had a much higher number of spads. Tony Blair (1997-2007) had 28 in his heyday; Rishi Sunak (2022-24) 41. But other ministers have also had many spads—successive Chancellors, for instance. Gordon Brown (1997-2007) had an annual average of 8-9 spads; George Osborne (2010-2016) five; Philip Hammond (2016-19) six. In the Coalition, Deputy PM Nick Clegg (2010-15) had an annual average of around 17-20 spads. Some offices and office-holders were more important than others, so the cap gave way to expediency, even though its form did not change.

Still, high numbers were meant to be exceptional. The median number of spads per minister under Blair was probably about two; and just over 80 in total. Matters progressed, and by the time of Boris Johnson’s government, there were just under 140 spads, with the majority of Cabinet ministers having three or more spads. Practice outpaced text.

The cap is not merely of historical significance. The Civil Service Commission has recently launched a reviewinto appointments to the civil service by exception, because of reports (here and here) that the Starmer government appointed a number of people with Labour connections to senior positions in the civil service. One reason for this may have been to circumvent the cap and avoid explicitly increasing the number of spads in government. The Starmer Government’s workaround appears to have been to appoint people via the exceptional appointments route.

Constitutional Boilerplate

Perhaps, then, we can see the previous versions of para 6.2 as constitutional boilerplate, akin to commercial boilerplate (see the wonderful piece by Scott, Choi and Gulati). Commercial boilerplate are phrases or clauses in contracts, copy-pasted from previous iterations. Repeated over time, they become almost mundane, taking on a ‘stickiness’: they are resistant to revision in spite of changes (sometimes quite radical) in context. The result is that boilerplate can become ‘landmines’: problems that lie dormant but may explode at any time.

But there is something odd about applying the term ‘boilerplate’ to constitutional documents. First, aren’t all constitutions ‘sticky’? A key characteristic of constitutional documents is that they are meant to be resistant to change. Second, in constitutional interpretation (as with statutory interpretation), there is a presumption against surplusage: we presume that every word and phrase in a constitutional document is chosen for a reason. ‘Boilerplate’ suggests that some phrases or words are superfluous; that so-called ‘framers’ have not paid attention to the text.

But these points are valid only if we presume that constitutional documents are the result of ‘one off’ transactions, with a single set of ‘framers’. This may be so for ‘big C’ constitutions, but there is also the ‘small c’ constitution—the constitution outside a core text, the rules, practices and understandings that shape state power. Some small ‘c’ constitutional documents have had multiple iterations: the Ministerial Code is one example (the Standing Orders of the two Houses of Parliament are another). 

Small ‘c’ constitutional documents may also be the product of an iterative process between multiple actors. The Ministerial Code is primarily an executive document, and the executive is plural, not unitary: it contains multiple actors (the various ministers; civil servants; centres/ departments) with multiple interests. Aspects of the Code may embody the tensions between these various actors. 

It is also worth noting that the Code, while being ‘owned’ by the PM, is in practice usually drafted—well, ‘revised’ is the better term, since the Code is never drafted from scratch—by civil servants in the Cabinet Office’s Propriety and Ethics Unit (which in the past Sue Gray used to head), with contributions from other officials across government. Given the input of multiple actors whose interests may not necessarily coincide, and the intensely political nature of the Code, there was a strong incentive to be risk-averse. After all, if the cap had worked for this long, there must be a reason; and textual changes draw attention to themselves. Best to make as little change as possible—or no change at all.

Thus, it is plausible to see the previous version of para 6.2 as boilerplate. It was mundane; it was sticky; and it was a landmine. And in examining the history of para 6.2 and the Code we have learnt that the world of constitutions is not quite as we imagine it. Some constitutional documents are regularly revised, by multiple actors with multiple interests; they may embody tensions in government; and aspects of such documents may be problematic, but remain unamended in a highly risk-averse environment. 

So why the revision? Arguably, the debacle over exceptional appointments has forced the hand of the Starmer Government. The cap is more trouble than it is worth. The new para 6.2 renders any further recourse to ‘workarounds’ such as exceptional appointments unnecessary. The removal of the cap has ‘disarmed’ the landmine. 

Constitutional Minimalism

So while the new para 6.2 of the Code is a welcome change, it is a forced change. It is also a minimal change. It needed to accompanied by a full-blooded defence of spads—affirming that they do have a role in government and explaining when and why their use is appropriate. Successive governments—Labour, Coalition and Conservative, and now Labour again—have struggled to explain why spads are useful, preferring instead to shrink or expand spad numbers without any explanation at all. The number of spads a Cabinet Minister has should not depend on a set limit but rather on how many are needed, depending on the nature of the office, and the needs of the office holder. This may mean more than two, but it may also mean less.

Para 6.2 says none of this. It simply notes that Cabinet Ministers can appoint spads; that appointments require the PM’s approval; and (in para 6.3) that spads have a responsibility to the government as a whole as well as to their Minister. In short, the current formulation continues a hallowed British tradition of risk-averse constitutional minimalism. The result is that spads and their growing numbers will remain a perennial distraction from far more important issues—like woeful infrastructure and weak capacity—that bedevil the British state. We deserve better.  

Ben Yong is Associate Professor of Public Law and Human Rights at Durham Law School.

(Suggested citation: B. Yong, ‘The quiet demise of constitutional boilerplate; or How special advisers can get their groove back’, U.K. Const. L. Blog (13th November 2024) (available at https://ukconstitutionallaw.org/))