John Mcternan, a former Labour adviser, wrote a brilliant piece in the FT on making Downing Street work. His advice? Know what you want to do, argue for it, have simple tests that drive everyone in government’s work (his example was Thatcher’s ‘is there a more market-based solution to this problem?’), and focus on ‘go big or go home’ interventions.

On these measures, Liz Truss looks promising, and her team structure looks reasonable.

The truth is, the Downing Street Policy Unit rarely functions well. It tries to do too much: man mark departments, come up with electorally compelling policies for conference or spending reviews, and think ‘blue sky’ thoughts. No one has the breadth of abilities, or enough time, to do all three.

There’s a bigger problem, which is that as a policy adviser your only real lever – the Prime Minister – cannot be used very often. You’re constantly hinting you might press a nuclear button. But you have much less time with the Prime Minister than a normal special adviser does with their Secretary of State (and most policy advisers pretend they have the PM’s ear more than they do, both during and after the role). Julian Le Grand, a health adviser under Blair, put it well: ‘you have to enjoy being a god on the outside, and a slave on the inside. I didn’t.’ (Neither did I).

As a unit, you are too big (the PM can’t spend lots of time on each of the areas covered) and too small (you can’t actually man mark departments effectively, or impose your will). The latter is always easier to manage in government – they like hiring people. They add delivery units, implementation units, analysis units. The former is ignored.

Yet the most successful policy advisers have a single obsessive priority which is shared by the Prime Minister. Andrew Adonis was a good example. I would say my mother is another!

I always felt a more sane structure, therefore, was one a bit more like Liz Truss has chosen (or I think she has chosen). The Prime Minister should have a very small number of policy advisers, each in areas they really care about, who are senior and experienced enough to drive through a proper priority. In an ideal world, they’d have an outside network in their area.

One of the few ways I think I added value when in Downing Street was through my network – I knew more heads and education leaders than most people in the Department for Education (because civil servants are moved around too often, but that’s another topic). I could find out what was happening, get ideas and advice, and see where the real world differed from what the Department told me was happening.

A good version of Liz Truss’s structure would be if she said ‘I am choosing not to care about most departments. Don’t man mark the DfE, don’t worry about culture. I only want you to focus on these areas, and these policies within those areas. Make them happen.

Given she has chosen relatively inexperienced people (and, for what it’s worth, I don’t think youth is a bad thing – I’m pretty sure I was much quicker and cleverer a decade ago) they will need to add networks and experience another way. Her advisers will have to figure out who the really experienced people are who fundamentally agree with the Prime Minister, and use them as additional capacity (lots of ways you can do that, formal and informal). In other words, to make this function well, you are going to need to co-opt in the Julian Le Grands, or Alison Wolfs, or Gerard Lyons of this world and use them.

To be clear, this kind of Downing Street structure – which Truss has chosen, and I am advocating, is unquestionably a trade off, for three reasons. 

First, there will be more media disasters. You can ask your private office to man mark for insanity from departments, but you also just have to accept things are more likely to blow up. There will be bad news stories you might have stopped, and that’s ok – because you’re advancing the core agenda. It’s not a great media strategy, but it might be a great policy and electoral strategy.

Second, you are trusting your departments and empowering your cabinet much more. You’re likely to get more variance in performance (I think this is fine, even good, and the Theresa May micromanagement didn’t work, but you have to accept it).

Third, you are empowering your treasury, because they will have the capacity to man mark. If you trust your chancellor implicitly, that’s fine, but they’ll have more sway over most departments, even if on basic economic policy you’re beefing up your capacity slightly. 

This is a long way of saying – I don’t think that her Downing Street policy structure is bad. In fact, it could be better, and more rational, than her predecessors’ – if she uses it well.

P.s. I should note lots of my colleagues disagree with me!