Another byte

a contrary look at project funding for gov

Joe Hill’s Bite-Sized Budgeting from the Reform thinktank is interesting.

The current mechanism isn’t fit for purpose. The VC funding process is built around a discovery-based decision flow. A change to the government equivalent would produce the required changes as a side-effect. So what would that look like?

An analytical framework

The research for Foundations of the Digital State identified the different decision flows for functional and non-functional questions. (functional is what the system does, non-functional (or infrastructural) is how it does it - slightly clunky old fashioned terms.)

Functional requirements in government are in legislation and non-functionals are a mixture of standards, guidelines, best practice, blog posts, internal position papers. (See this working paper)

Failures to digitise in government are now mostly non-functional failures: clunky, hard to navigate, confusing and unjoined-up systems. By and large the state now delivers software that does what the law requires and fulfils its statutory duties.

Writing software involves trade-offs and compromises. But when one party is a must-have law and the other is a nice-to-have blog post, there is only one winner. We drive the state with the steering lock on. You can’t fix this problem with financial controls.

Decisions, decisions, decisions

In order to consume finance like a VC backed firm you need the decision flow that comes from exploratory and iterative development of system, incrementally building, testing and measuring. It’s well understood that this has to be pre-legislative. Here’s Jerry Fishenden and Cassian Young wrote about it 8 years ago.

Once the legislation is written you are locked in. The report quotes Byrne’s Law “most government technology projects could cost 10% of what they do, and still provide 85% of the functionality”. The problem is that they have to provide 100% of the legal requirements. To get that cost reduction you have to eliminate the damaging 15% of requirements before or during the legislative process.

I have recommendations to do that - albeit in a Scottish context.

You need a legislative process that addresses oversight of non-functionals, early runaway programme catch and kill, iterative development of policy, systems and legislation, learning through building, co-design and so on.

None of these changes can be effected via a budgeting process.

Raw Power

So what does the financial controls process do? To understand that you need to understand power. At Westminster, the Parliament, the Government, the PM are extra-statutory, they are creatures of convention and not written law.

Individual ministers have lawful authority granted to them in two ways. Firstly by virtue of holding a named seal of office (the President of the Board for Trade for example). There is a collective set of powers granted to an unspecified Secretary of State which is exercisable by anyone appointed as a Secretary of State.

The PM’s power comes from the convention that the King will dob in any minister at his request because he holds the House. The Chancellor on the other hand has actual legal powers over the money. The core of government is a dyarchic dance between confidence and money.

The Minister for joined-up government is a buttonman for the PM. The Chancellor of the Duchy of Lancaster has no statutory powers, but if a pol crosses him, they sleep with the Min. of Ag. & Fish.

Due to the firewall between Ministers and the civil service Pat McFadden has no legal powers to co-ordinate government digital policy by instructing civil servants except by coercing their Minister.

Lacking statutory power to do joined-up government, public sector transformation is driven by financial controls on loan from the Treasury. If you want to dethrone the financial controls process you need to put joined up government on a proper legal, organisational and constitutional footing.

You gotta have some standards

The Foundations of the Digital State proposes such a footing (for Scotland) which creates a co-ordination mechanism based on a standards regime (like the internet has).

The core thrust is a weak co-ordinating centre , where departmental teams are given maximum autonomy - and that needs to include financial autonomy. Designing it needs a theory of state.

What would that look like at Westminster? The Treasury split into a Finance Minister and and Office of Budget Management (a girl can dream!). The OBM would work under the new SoS for Infra/Digital/Joined-UP Gov (ie Pat McFadden) in a PM’s office with a standards creation organisation and appropriate parliamentary oversight.

I appreciate this is the liminal space between “think bold thoughts, Oh Silicon Valley flavoured guru!” and the inevitable “but not like that”. Ho-hum.

The year of public affairs

Now the research is done and the report published, the hard work begins - the public affairs necessary to make this a top topic in Scottish politics and to work across political boundaries to get consensus and action on what is after all constitutional change. Looking to talk to people interested in this, work with thinktanks, appear on panels, publish other people’s work on this SubStack, collaborate and generally do stuff. HMU