
I discovered recently that assaults on planning officers are now, regrettably, a thing. I don’t think planning officers have become less tolerable, so I think this is quite a good bellweather for societal decline. It wasn’t a thing 30 years ago.
A few days later I listened to this podcast from Public Digital, which touches on dissatisfaction with government in a more general context. Leaving aside the obvious causes from late-stage capitalism – media-fuelled hatred of public servants, instant gratification culture, widening inequality etc. – I wondered whether there is also an element of frustration with inhumanly-automated and austerity-crippled public services. I found myself disagreeing with Tom Loosemore on whether public services have got better or worse over the past few decades. I have a lot of respect for Tom’s views, so I thought I’d do some research.
It turns out that the official measure of quality of public services is throughput. The ONS counts physical outputs – hospital treatment, classroom hours etc. – and weights them by relative cost. In absolute terms, the state delivers much more than it did 30 years ago – output grew 65% from 1997 to 2019, fell during the pandemic and then continued rising. But this tells us nothing about outcomes, just activity – and there are many who argue that the state should do less, not more. So I don’t think the official measure is helpful.
Looking at outcomes like pubic health, road deaths and carbon emissions shows a pattern of strong improvement up to 2010, and then plateau or decline after austerity kicks in. In fairness to fans of austerity, this period begins a noticeable improvement in public sector productivity (output per unit input) of about 0.7% per year, roughly triple that of the preceding period. Again you have to traverse the dip of the pandemic, but it’s essentially pick your poison: spend more to increase output at the expense of productivity or spend less to force productivity up at the expense of output.

Leaving productivity aside, Tom is right on two counts: the state delivers a much larger volume of activity, and in most areas outcomes are better than they were 30 years ago – the UK is cleaner, healthier and safer than it was. But to come back to my original issue, people don’t assault planning officers because they feel their life expectancy is shorter or their neighbourhood is less clean or less safe. I think there is a disconnect between macro indicators and people’s experience of public services. I think that has got worse, though I will leave the rigorous analysis to others – what follows are ideas and reflections.
One of the problems with automation is that while it dramatically increases efficiency for ‘normal’ cases, it can make abnormal cases much more difficult – especially if you have laid off or tied the hands of anyone with the experience and judgement to deal with them, in pursuit of the savings promised by automation. This is an error understood since at least 1969 when Michael Lipsky coined “street-level bureaucracy”,yet is still persistently made across public services to this day. This can create disproportionate failure demand – demand that exists only because of mistakes or delays by the organisation.

In response to this, one of the most pernicious developments in modern customer service is the deliberate enshittification of the telephone channel – tying people up in layers of menus designed explicitly to delay or prevent them from talking to a human being. This is euphemistically called “friction”, and it is by no means confined to the public sector. Octopus Energy win so many awards because they are one of the few large providers who have chosen to fund their customer service operation well enough to do away with friction.
In the very old days, before the advent of call centres, you interacted with the state by writing letters. You would have to wait for a response, which created periods when you felt less embattled, because the ball was in their court. Given the difficulty of communicating with a human being by telephone or over the internet, I have long championed this method. But last time I disagreed with a public body, it didn’t work – they simply ignored every letter I wrote, and sent me an endless stream of automatically generated garbage that did not acknowledge or respond to anything I had written. It took a Letter Before Action to elicit a human response. (13 months later I am still waiting for a court hearing date for that action, but even Tom doesn’t claim that the justice system has improved.)
Sometimes the system fails in cases barely outside the standard template. My brother is self-employed; one day he missed a tax payment deadline, having not thought about it before going on holiday. He got back to a notification that he’d missed it, and went online and paid it immediately. He then got a notification saying that he’d made an overpayment and HMRC were returning it to him – for some reason the system had not correctly matched the late payment he made with the payment he’d missed. HMRC returned his payment, and then pursued him for the missed payment with interest and penalties …. for two years. It took two years of letters and phone calls for him to get through to someone who understood what had happened and could fix it.
I once got an erroneous bill from British Gas (I know it’s a private company but it’s still a public service). I phoned them and explained, and they said they would correct it. The next month I got another bill with the same error, uncorrected. I phoned again, and they promised to correct it. This happened 12 times. After 12 incorrect bills I took them to Ofgem and won.
Based on what I hear from others, I don’t think my family’s cases are atypical – and I don’t think any of those three examples would have happened 30 years ago. The reliance on automation, and the removal of people capable of judgement has deprived public services of the ability to understand, engage with and solve problems.
This is what I mean by public services being worse – not at the macro outcome level, but at the level of individual experience. I think this is a non-negligible contributor to assaults on planning officers and other public servants. Ironically, planning is a sufficiently complex and varied discipline that it probably suffers less from one-size-fits-all workflows than many other services, but the people whose loss of self-control lands them in court can’t see that. They can just see the delay and unresponsiveness caused by excessive caseloads and understaffing.
Another problem is that when people successfully interact online with a public service and don’t need to engage with any humans, few are thankful for that. They don’t get any good feelings about the state, they just – on that particular day – don’t have any new bad ones. Paying your road tax or renewing your passport are vastly simpler today than they were, but that doesn’t generate anywhere near enough goodwill to offset the negative experiences of non-standard interactions.
So I think we have a mounting deficit – the more we automate, the more edge cases we create; the more savings we extract, the less we can deal with them. My personal view is that the state should do fewer things better, but I realise this isn’t possible in a political system which rewards doing more new things.