Rushed today, so I think I will deal with one short [POSTSCRIPT: not exactly short, sorry!], but I think significant point today and write the longer piece on Friday, with respect to one of this blog’s constant bugbears, the UK’s Office for Budget Responsibility[1]. Basically, the general theme is that some messengers are really bad, passing on incorrect messages and accidentally or purposefully manipulating the system they’re meant to be communicating for. Not shooting the messenger is definitely a tactic rather than a principle.
On Friday, I’ll try to put down my explanation of why the OBR simply can’t work and some suggestion of how it could be reshaped to have a different mission that might be achievable. But today, let’s look at the shorter and easier question of why it hasn’t worked, and the demonstrable historical fact that since it was set up in 2010, we have had much worse fiscal policy outcomes, not better.
The reason for this is what might be called the “Pirates of the Caribbean effect” – every six months, the OBR publishes a report, and it’s either a “war chest” or a “black hole”. Because we have a fiscal rule based on a five year forecast, we have a control system that is based on a forecast residual, which is bound to swing around randomly (because the predictable part is what’s picked up by the model – forecast errors have to be random, it’s a major diagnostic if they’re not!). Since cutting current running costs and firing people is difficult, this volatility tends to be fully absorbed by the capital budget, and one thing we do know about investment is that it’s very inefficient to do it in a stop-go fashion that doesn’t facilitate long- or even medium-term planning.
According to the Treasury Select Committee, though, this is terribly unfair on the OBR, particularly in the version which was put to them by Jo Michell. Meg Hillier put the blame for the war chest/black hole phenomenon back on the Chancellor:
“The fiscal rules, and therefore the OBR’s forecasts, become more important when government spending operates on razor-thin margins. That is a feature of decision-making by chancellors, not overreach by the OBR.”
To be honest, I don’t really think that this picture of the OBR as a purely neutral and academic forecasting body, which wasn’t meant to have any control over the democratically elected representatives … I don’t think it passes a laugh test. OBR forecasts are meant to constrain Chancellors, when he set it up George Osborne was pretty clear that what he was doing was putting in place an institution which would make it harder for future governments to reverse his policies. But let’s take this at its merits.
The idea here is that like the speed limit on a road, the OBR “financial capacity” measure ought to be thought of as a ceiling rather than a target. Most of the time, chancellors should be operating well below the maximum level of borrowing, so that the OBR forecast residuals don’t affect their policy.
I’m going to continue the policy of honesty and say I don’t think this passes a laugh test either. For one thing, OBR forecast residuals can be pretty bloody large – there would need to be a really significant margin of safety if this were to work. OBR forecasting errors also, like any other economic model errors, tend to be clustered around crises and cyclical turning points, which is just the moment when you don’t need another sudden constraint on policy.
I also think that the real laugh test here is the one that Keynes described as “shouting fire! fire! in Noah’s Flood”. That is to say – knowing all we know about the UK economy today, are we really going to argue that the biggest problem is that Her Majesty’s Treasury are too keen to support unnecessary public investment? Did we really need to be running a tighter structural fiscal policy for the last fifteen years? A policy of constantly running a much tighter budget than the cyclical position would allow is a policy of permanent stagnation, and the UK already seems to have one of those.
But at a deeper level, this is a representative of a kind of policy that’s superficially attractive but impossible in practice. If you say that “everyone should drive 10% below the speed limit”, then that’s the new speed limit. If policy were to be set that the Chancellor should have a margin of safety, then immediately it becomes a requirement to have a margin of safety to make sure you don’t put the margin of safety at risk. This is a live issue in bank regulation at the moment, where regulators have belatedly realised that if you publicise the “buffers” of liquid assets that you expect banks to hold in order to slow down a bank run, then all you have achieved is to start the bank run a few weeks earlier when the buffer levels are reached.
If you’ve got a system that can be controlled by limits and targets, you need to set the limits and targets and control the system. If that doesn’t work, then what you’ve got is a system that can’t be managed in that way and needs to be redesigned. You can’t find a clever way around Goodhart’s Law by saying “when a measure becomes a target, it can still be a good measure if you pretend it isn’t a target”.
[1] I mean, come on. One thing that definitely has to change in any reform is the name. I just don’t believe it’s possible to do your best, politically neutral work in an environment where there’s a huge, dumb politicised slogan hanging over the door and printed on the letterhead. If you want to make the case for independent fiscal forecasting make it, but really, “Budget Responsibility”, my eye.
This regulatory stuff is probably a bit above my pay grade, so instead I’m sharing this video if you want a slightly more entertaining justification of sometimes shooting the messenger: https://youtu.be/foT9rsHmS24?is=TNltL693Uud0Qzeg
In my youth I frequented a sparklepets website called Chicken Smoothie. (A website where you can adopt art, basically). They came out with new pets (art) every month, and you could adopt a litter of each species.
One of the features of Chicken Smoothie was that every pet had a rarity bar beneath it. The rarest pets were mostly ones from the beginning of the website, where there were only a few people were online to adopt pets.
But not all these pets were created equal. Some of them (like the Rose Dog) were much more desirable than others.
Another feature of Chicken Smoothie was that you could trade with people. Everyone had a trading page where they could write their own trading rules, like "don't offer me anything below an uncommon". People cared a lot about "a fair trade".
Some prominent users (including some mods, but not the site's actual owners) noticed that new users who got their hands on rare pets through the site's redistribution methods didn't always understand the value of their pets, and got taken advantage of. They also noticed that people had strong but differing opinions on what fair trades were. So they created the Rares List, which described what pets actually traded for. The Rose Dog, for example, traded at a level equivalent with notably rarer dogs.
People thought of the Rares List as a prescription. Someone would offer a trade and justify it on the grounds that this fox and that cat were equivalent by the Rares List. But everyone wanted that cat! And no one wanted that fox. So no one would accept this 'equal' trade without something extra along with the fox.
Because the Rares List was descriptive, eventually it was updated to reflect these dynamics. You can see where this went. The Rares List was discontinued 9 years after its creation.