UK election dashboard (bad prototype)
A mediocre version of a good idea
Many discussions circle around, untroubled by reality, until finally it is too late.
Elections are often like this. We have months of surface-level discussion followed by an election, after which we silently know that much of the discussion was worthless.
Surely there is a better way?
My suggestion is Forecasting Dashboards:
Forecasts can tell us what is likely to happen after the election, depending on who wins. This is the information we actually want to know
It is cheap to put all the forecasts on one site and have them update live. You don’t need many articles, you look at one page,
every dayonce a weekSave your mental energy for just before the election when we are all engaged
The Prototype
Link: manifold.markets/dashboard/uk-2024
First, the obvious stuff:
Will Labour get a majority?
I like the politico poll of polls
If you just want to know what’s likely to happen, then these are the kinds of things you want to look at. In a better version, we might show these charts directly, but I made this in about two hours with
.Next, some decision markets:
I know these look overwhelming but they are basically a 2x2 grid:
Will Labour get a majority?
Will X happen?
So there are 4 options.
Labour majority and the thing happens
Labour majority and the thing doesn’t happen
Not a labour majority and the thing happens
Not a labour majority and the thing doesn’t happen
In each case, you can see what the market thinks a Labour majority (or not) will do to the thing in question1. And so you might get a sense of who to vote for.
For the dashboard we chose the following metrics:
Will more than 200k houses be built per year, on average, in England and Wales?
Will inflation on average be between 1.5% and 2.5%
Will illegal migration be above 20k per year on average in the UK?
Will net migration be above 400k per year on average in the UK?
Will there be more than 80k robberies, on average?
will fossil fuels (without imports) compose >33% of UK energy generation next Parliament?
Will UK aid be at or above the 0.6% of GNI by the end of the Parliament?
Will fast-growng broiler chickens still be legal by the end of the parlaiment?
(There should probably be a question on the NHS, suggestions please)
I’m not saying these are what you should care about, only that if you do, you can see how your vote might affect these things.
These markets are too new to be worth listening to, but, for the sake of argument at the time of writing, bettors thought that a Labour majority increased the chances of >200k houses per year in England and Wales2. This might encourage you to vote for Labour, or not! This may have changed by the time you see it, since these markets update live on the substack web app.
“Why always Labour majority?”
Why not ask about a Tory majority?
The betting markets say the two most likely outcomes are a Labour majority or no majority (either a Labour minority, a hung parliament, etc). Most future worlds are like this. The Tories win about 6% of the time. It isn’t worth asking Labour vs Tories, because the Tory win is so unlikely that people aren’t incentivised to bet on it much. Whereas a Labour non-majority is ~20%3.
Also the difference between a Labour majority and not is pretty consequential. If Starmer gets a majority, he probably begins a plan of long term reforms and there is a good chance (60%?) of 5 or more years of Labour Government. If he doesn’t get a majority, then we probably have a short Parliament focused on winning an actual majority next time. Those policies will be quick wins.
Talk to me more about this dashboard
In general, I think news rarely gets to the heart of an issue. We don’t talk about what will actually happen. We don’t need headlines daily about small boats or the border, we need numbers and what will happen under different policies.
You may say I’m a wonk, but I think most people agree that political discourse is bad, because they tune it out. They get on with their lives, because they can’t possibly follow this circus every day. It’s only the party faithful who actually want to defend their side day in day out and I think they exhaust almost everyone.
And so there is a better way. A very limited high quality source of news that is focused on the actual state of reality. How many excess deaths are there? What are UK NHS waiting times. We could poll people on the things that matter to them and show them what the numbers are and how they will change with different policies and politicians.
And the point of this article isn’t to convince you that forecasters can forecast this, but I am confident they can. Prediction markets are well calibrated, when they say a percentage, they mean it.
Let me know what you think
Dashboard link: manifold.markets/dashboard/uk-2024
So, do you like it? It’s very early so it won’t be for everyone.
Would you use it? Let me know in the comments or with a reply.
Why use 2x2 markets rather than conditionals?
Conditionals are hard to get your money (or tokens) out of. A market “Will England and Wales build 200k houses per year if Labour win a majority” resolves n/a some of the time and all trades get returned. This is has weaker incentives to bet and is a nightmare if real money is involved. Sure, you can also bet on the other conditional market but now you have to tie up twice as much money on a single return.
In my opinion, it’s better to use the correct underlying market structure and mess with website design than use the wrong market for the sake of ease.
Ideally we could make a bot that trades these market against a central “Will Labour get a majority” market and the display them as if each market is separate:
“Will England and Wales build 200k houses per year if Labour win a majority”
“Will England and Wales build 200k houses per year if Labour don’t win a majority ”
But crucially the underlying market would be unfied and if anyone tried to bet on it, it would take them there. (Or to a proper conditional market where bets cost a fraction of the condition rather than 1, but now we are really in the weeds)
For those without the substack web app, here is what this market looked like at the time of writing. manifold.markets/NathanpmYoung/2x2-will-labour-get-a-majorit-will
The numbers add up to not 100 because there are betting fees and probably other outcomes that don’t reach this graph. But it’s roughly right.