To determine the rate of heat transfer between a hot and cold zone, for example to consider how an air conditioning system works, an engineer creates a system boundary, and defines the conditions that hold true in their calculations at the system boundary. Drawing a system boundary makes a given thermodynamic problem solvable. For a government policy decision, system boundaries are more subjective and are chosen at the policymakers’ discretion. So how do they choose?
A prime minister introducing a policy, say to increase education spending, can only reap the political goodwill from the short-term benefits brought during their reign. For most western democracies, the average democratic term is 4 to 5 years. The longer term benefits due to a better educated, higher skilled workforce are reaped by a successive government. Unsurprisingly, there is broad consensus that investing in children’s education does produce better societal and economic outcomes - higher earnings, less crime, and lower expenses on social services, with one study estimating a rate of return of between 7-10%. Notably, the return is measured 2 to 3 decades after the initial investment, when those children are 27 and 40, but policymakers are not drawing system boundaries to consider a return so far in the future.
With six different prime ministers in the past 10 years, the UK is notoriously bad for government stability. Liz Truss served heroically for 49 days; the game of parliamentary musical chairs is dynamic, to say the least. The time for a policy decision to become legislation, and a post implementation review be conducted, typically 3 to 5 years after commencement, far exceeds the 1.9 year median term of a prime ministerial tenure. Each prime minister assembles new ministers, further degrading the continuity in the policy making process. Without scrutinising any of the policy decisions being made, it is clear that the ruling prime minister who enacted a bill will long be out the door before there is a discussion of the impact, let alone any accountability for their decisions.
Concretely, the figure shows examples for the shortest, median, and longest timeframes for when a post implementation review would be due, from 239 bills passed through parliament between 2017 and 2026. This sample excludes bills connected to the treasury department, as treasury policies do not undergo a post implementation review process. Bills have been matched to their corresponding act by matching the titles. The data was obtained using the governments’ Bills API.
The timespan for the best case scenario still surpasses the median prime ministerial tenure from the past 10 years, meaning the impact assessments are consistently delivered to a prime minister who did not introduce such legislation. The extreme example is the Smart Meters Act 2018, first brought in by Theresa May, with the impact to be assessed five prime ministers later by Andy Burnham. Whilst it is natural for some legislation to require a phased roll out, if this straddles so many prime ministers premiership, who really remains accountable for the delivery in the end?
Such timescales incentivise passing legislation rather than considering the long term consequences. If we can’t rely on these timescales to drive accountability through the system of policy creation and implementation, it must come from elsewhere. Journalism, general elections and select committees, or independent watchdogs bring a semblance of accountability, but they are structurally external to the process of policy making itself. Approval from these institutions is not a requirement to pass new legislation, a slap on the wrist will do.
If the structures of government, democracy and the media don’t require any long term thinking, there is no reason to draw a system boundary for a policy decision longer than your own tenure in parliament. Good engineering practices teach us that poorly chosen system boundaries lead to inaccuracies: the problem will be modelled incorrectly, and your solution will be wrong.