How (not) to manage your civil service, or what is “stepping down” anyway? - Prof Perri 6, Chair in Public Management, Queen Mary

Times photographer Matt Lloyd

So he’s going. Sir Bob Kerslake, that is. “Sir Bob who?”, you ask. He has been head of the civil service in the United Kingdom since 2011. “And why does it matter?”, you want to know. Oddly, the commentariat doesn’t agree on just what the really big issue is in Sir Bob stepping down, only that there surely is one. In fact, five very different debates have been rolled together. Let’s disentangle.

Sir Bob has been a part time head of the civil service because he has kept his other job as Permanent Secretary in the Department of Communities and Local Government. The Cabinet Secretary has been and still is Sir Jeremy Heywood. For much of the last century, those two jobs have been combined. The argument for combining them was that the Cabinet Secretary captures ministers’ decisions in cabinet and its committees and communicates them as instructions to the civil service to carry them out. Therefore, he (so far, all eleven cabinet secretaries have been men) has to sustain the overview of how the development and implementation of policy is going across government, and both those roles require management authority. The argument for separating them was that the skills and requirements of being the policy integrator and advisor, and being the chief executive focused on systems, civil service capabilities and structures are very different, although the people doing those things should work closely together. Now, it’s been announced that Sir Jeremy Heywood will become head of the civil service as well as cabinet secretary. Nevertheless, a new job will be created of “chief executive” for the civil service. That person will report to the Cabinet Secretary and Head of the Civil Service rather than having equal status as Sir Bob did.

One issue is how, when and at who’s say-so top civil servants should go. First, “how?”

Was he pushed or did he jump? The Cabinet Office press release says coyly that he is stepping down. Some journalists have suggested that the minister for civil service reform, Francis Maude, got him sacked for obstructing his reforms. But which ones? Most of Mr Maude’s programme remains government policy. Maude was annoyed recently that the job description prepared by the civil service for permanent secretaries, who head each department of state, says that they should “‘balance ministers’ or high-level stakeholders’ immediate needs or priorities with the long-term aims of their department, being shrewd about what needs to be sacrificed, at what costs and what the implications might be’. Mr Maude thinks that sentence defies the constitutional rule that civil servants carry out minister’s instructions. But that document has been around since 2009, from before Sir Bob’s appointment. Although some journalists claim that Kerslake lost ministerial confidence generally. But if he and Maude were on bad terms, it doesn’t follow that this must explain his exit. The prime minister’s office denies losing faith in him, of course, and Kerslake himself says he was ready for retirement. Even if he has been pushed, it would have taken more than just Maude’s decision to get him out. If prime minister thought that relationships had broken down irretrievably, as some journalists claim most likely Kerslake would have to have been offered the chance to return to his departmental job. It’s unlikely that what lies behind this is a straightforward return to the days before the 1870s of crude ministerial patronage and its withdrawal.

Nor is the story credible that he had to go simply because he revealed to the Public Accounts Committee that the Treasury had still not signed off the budget for the Department of Work and Pensions’ Universal Credit scheme. The project wasn't Kerslake’s show, in either of his jobs. If any top civil servant might have had to walk the plank for that debâcle, it would have to have been Robert Devereux, the permanent secretary at DWP. But unlike some months ago, when Devereux was under pressure, there has been no suggestion of it now. The fact that this story is being bruited about matters only because it serves to remind us that the criteria used to assess the performance of the head of the civil service are so little known and understood outside the civil service itself that it is hard to have a well-grounded discussion of whether Kerslake did his job well or not.

Next, “when?” The civil service unions are very upset that Kerslake’s departure was announced at the same time as the ministerial reshuffle making it look like a political decision, so compromising the constitutional rule that top civil service jobs are not given or withdrawn on ministerial favour and disfavour. This is probably overstated. The simultaneous timing wasn't adroit. But the announcement did not come from the prime minister’s office. It came from the Cabinet Office. If this is “politicisation” it is a very feeble form of it and not the main issue to be worried about in this whole story. And given the timetable for the rest of the programme of civil service reform, if Sir Bob was going to go, July is probably a fairly appropriate time to make the announcement.

Another issue is whether separating the top policy job and the top management job was ever a sensible thing to have done, or whether the problem was the particular way that the British government arranged the separation. There’s also the interesting question of whether the separation is really so completely ended as the announcement of Sir Bob’s departure might lead us to believe.

A chorus made up of the Institute for Government, former cabinet secretaries such as Lord (Gus) O’Donnell and leading commentators such as Sue Cameron of the Daily Telegraph have all welcomed the reunification of the jobs of head of the civil service and the cabinet secretary, and reaffirmed the old rationale for it. And even I have some sympathy with the conventional view. Yet it’s worth recalling that it was Heywood himself who urged the split, and who thought the skills of policy integration and service management were quite different. It’s true that providing the PM with daily policy advice, servicing the cabinet and its committees, compiling policy briefings and synthesising decisions into useable instructions and undertaking occasional policy inquiries all requires very different skills and uses of time than does the business of reforming structures, integrating professions, managing spending budgets, developing skills frameworks and cross-service promotion systems and providing inspiring leadership to staff. The problem has been that the policy works still holds the higher status and is what most civil servants aspire to do, while implementation interests too many of them mainly as a source of information about feasibility for use in policy advice.

Indeed, arguably, creating the new “chief executive” role will likely reflect that argument for separation, but without even the equality of status that Kerslake enjoyed. We can leave to writers of dictionaries the question whether it’s appropriate to call the Number Two a “chief executive”. Nonetheless, applicants might wonder whether they would have the kind of power that they’d expect from that title. Since this person won’t be responsible for operations – they’re done in departments and agencies – she or he can’t be called a chief operating officer. Perhaps she or he should really be the “chief management officer”?

A third issue is what, if anything, his departure means for the coalition government’s programme of reforming the civil service to make it more accountable to ministers.

Despite off-the-record briefings hinting that unnamed ministers thought Kerslake and some other permanent secretaries have occasionally obstructed ministers’ plans, nothing in Sir Bob’s departure really changes the position. In fact, some permanent secretaries have grumbled off the record to journalists that Kerslake was too close to ministers and didn’t protect them enough. The real problems about accountability to ministers lie elsewhere. Sometimes they lie in ministers not thinking through exactly what their plans would really involve and being reluctant to take advice. That, in the Public Accounts Committee’s view, may explain some of the continuing problems with the Universal Credit programme for integrating cash benefits for people who are out of work. In other cases, the problem is not unwillingness but incompetence in the civil service. The management of the information technology in Universal Credit may be an example of that. It appears that the fiasco over the letting of the franchise for the west coast main railway line a few years ago was little more than incompetence. If the real problem had been civil service unwillingness to carry out ministers’ instructions, then we should have expected many more cases in which permanent secretaries had asked for published written directions from their ministers. In fact, what is remarkable about the coalition government is that so few have done so.

Indeed, one consequence for Sir Jeremy Heywood of the fact that he now has the title of the head of the civil service as well, is that he will have to worry, as he has not needed to since 2011, about just who should carry the can for failures within departments and agencies. If the civil service comes under greater fire for such failures as the Home Office’s apparent loss of key files from the 1980s and 1990s on alleged paedophiles in the period before the general election, Sir Jeremy may sometimes wish Sir Bob were still around to share the flack and help him deal with his departmental permanent secretary colleagues. Or maybe the new “chief executive” will be scapegoat-finder in chief and, herself or himself, scapegoat of last resort?

A fourth question is where all this leaves the principle which has had constitutional status since, roughly speaking, the 1870s, that the civil service has a measure of independence in providing impartial policy advice to ministers, whether or not they want to hear it.

On BBC Radio 4’s flagship “Today” programme, Sir Gus O’Donnell seemed to imply that the new structure would better enable civil servants fearlessly to offer policy advice before ministers make decisions. It is not obvious why this should be so. Mr Maude has abandoned neither his plan for a greater role for political appointees in policymaking, nor his aspiration for more policy advice from outside bodies under contract. If the Conservatives win an outright majority in the 2015 general election, we might well see that agenda pursued more vigorously. In that case, with or without a chief executive at his side, Sir Jeremy Heywood not could resist a firm government commitment. Anyway, civil service policy advice continues to be required on all sorts of practicalities, even after French-style “cabinets” of political advisers and external think tanks or consultants have provided a direction to which ministers are committed in principle. As Sir Jeremy Heywood told a “Civil Service Live” event, when arguing for civil service policy work even when ministers emphasise delivery responsibilities, “The worst policy mistakes are made by those who don’t takeinto account the views of the implementers”. Sir Bob’s departure does not really affect the issue of policy independence very much either way. For that, what matters is really what the government elected in 2015 wants.

The fifth question is whether the formal reunification of the top policy and top management job will or even could really do anything much better to integrate the civil service.

In respects of many inputs, the service is more integrated than it ever was. It now has common grades, more service-wide professional structures (including a new one for a policy profession), common skills and competency frameworks, more routine promotion by moving around departments, more common systems of financial appraisal and reporting. Where it remains weak is in integrated thinking about outputs and outcomes where they require work which spans several departments and agencies and tiers of government. That’s an ancient problem of course. Lloyd George, for example, was much exercised by it a century ago, just as Tony Blair was almost two decades ago (at least until the end of his first term). If a century of having the cabinet secretary and head of the civil service roles combined didn’t solve it, it is unlikely that adding a subordinate chief executive to a combined head will do so, at least on its own. Of course, some people will argue that the 2015-2020 government should not make this its priority, because instead it has to deliver on huge spending reductions to which all the main parties are more or less signed up, and worrying about integration of what gets achieved will be a distraction at best and an obstacle at worst, to the task of cutting government back to the bare bones. But if those cuts are to be achieved with some measure of public consent, there will be no alternative. The civil service will have to help ministers show that government is nevertheless achieving many of the things that citizens care most about, even with less cash being spent to do it.

So what then really does matter, for the quality of government, about Sir Bob’s departure and the combining of the roles? In itself, rather little. But in its timing, it achieves something rather useful. Announcing this change just ten months before a general election has stirred debate afresh about these five issues at just the right time, exactly when the country should be debating how to run and reform civil service, what we most want it to do well, how the performance of top civil servants should be measured, what skills they should have, and about what we want the relationship to be between ministers, their political appointees and civil servants. In that regard at least, announcing Sir Bob’s departure at the same time as the reshuffle may not have been at bad thing. These are exactly the right debates for Britain to have just before this election.

Acknowledgement: I am very grateful to Janice Morphet, David Walker and Colin Talbot for their advice and suggestions on an earlier draft of this article. They can be completely exonerated of blame for my errors and should not even be assumed to agree with my arguments.