Will more data help voters make better decisions?
More data doesn't help the voters because Western democracies currently use course-grained voting systems that bundle together hundreds of issues. The solution is to use fine-grained voting systems.
Someone on Reddit responded to what I wrote by suggesting that if voters had more data, then we would finally get responsive government. I wrote this in reply. I’m repeating things that I’ve written before, but perhaps this time I say them in a way that is easier to understand:
I think I said this before, but I'm worried about your focus on data. Mostly I'm worried that when you focus on data you don't focus on what behavioral science has taught us about the limits of data. You seem to think that more data would solve problems with our current political process. But why would more data make the current situation better? Research into decision making has amply demonstrated how much a person's pre-existing worldview will filter what data they are exposed to. Have you taken that into account?
The omnicompetant citizen does not exist. Even if you got a PhD in political science, and then another PhD in international relations, and then another PhD in world literature and then another PhD in religions of the world, you still would not know enough to make informed decisions. The world is simply too complex at this point for any one person to understand it. No amount of data will improve the situation.
We live in a world where the division of labor has advanced as the total amount of knowledge has expanded. In both business and government, the central challenge of our era is finding a way to organize large groups of specialists -- people who are useful thanks to their depth but dangerous because of their lack of breadth. We need to face the question, are there systems of voting that allow large groups of narrow specialists to collectively demonstrate the wisdom of those who have broad knowledge? In other words, can we get the benefits of both depth and breadth? Because that is the only way we can make progress.
The omnicompetant citizen does not exist. No amount of data will improve the situation. But what if, as a thought experiment, we imagined a person who had access to the world's most perfect database, something even better than Google, something that intuitively understands exactly what question you want to ask, and then it instantly gives you the data you need to answer that question. Such a database would still fail to teach you the questions you need to ask, and there would be urgent problems that you don't know about, should ask about, but don't enough to ask about.
But what if we go further, and imagine a magical database that both can provide you with all the data you want, instantly, and can also suggest to you questions you should ask, on subjects you didn't know about, but which turn out to be important. Could a person with this magical technology use the data to vote for the right candidate? Sadly no, because all of the current systems of voting used in Western democracies only allow course-grained voting that lump together a bundle of issues, whereas what is needed is fine-grained voting that would allow people to better parse the issues according to their needs. In their book "Democracy For Realists: Why elections don't lead to responsive government" the authors Achens & Bartels make clear that their is no equilibrium in a multi-issue context. And if that sounds abstract, I'll post a specific example (I posted this before, but here it is again):
Imagine a nation where there is only important issue: spending on education.
51% of the public wants more spending on education.
49% of the public wants less spending on education.
If you are a politician, it is easy to put together a 51% coalition that also supports 51% of all of the issues people care about, because there is only one issue that people care about.
What about a nation with 3 issues: education, environment, the military.
These are the factions that exist:
1. more education spending, more environment spending, more military spending.
2. more education spending, more environment spending, less military spending.
3. more education spending, less environment spending, less military spending.
4. less education spending, less environment spending, less military spending.
5. less education spending, less environment spending, more military spending.
6. less education spending, more environment spending, more military spending.
7. less education spending, more environment spending, less military spending.
8. more education spending, less environment spending, more military spending.
Now can you put together a 51% coalition that also holds the 51% popular position on all 3 issues? It is difficult. You’ll almost certainly have to get either some people who disagree with you on one issue, or you will have to advocate for an unpopular position, as part of putting together a larger coalition.
Now imagine a country that has 100 issues that the public regards as important. It is improbable, to the point of being close to an impossibility, that the 51% winning coalition also holds the 51% position on each of those 100 issues. In other words, the 51% winning coalition will advocate for the unpopular 49% side of at least some issues.
Put differently, for a politician to win, they must advocate for some unpopular positions. It is crucial to understand this, as it is one of the central reasons “Why elections do not lead to responsive government.”
Whoever wins an election, they come into power pushing for some unpopular policies. And that’s the average case. We can imagine extreme circumstances sometimes giving rise to the pathological case where the winning politician holds the unpopular position on the majority of issues (with a small handful of policies being popular enough that it still gives them 51% of the total vote).
There is no cure for this, it is purely a mathematical fact. Therefore, we must move away from any conception of democracy that has it representing “the will of the people.” While the benefits of democracy are well-documented, representing the will of the people is not one of those benefits. There is no aggregated will of the people that can be represented through elections. The benefits of democracy are caused by other factors (as we wrote about in the previous essay).
These issues won't be fixed by giving more data to the voters. These issues can only be fixed by ending the course-grained voting that we engage in now, and putting in place a system of fine-grained voting that would better surface voter preferences on each issue, with each issue being disentangled from the other issues.
While there are some complicated systems that could do this, I also suggested a simple system gives most of the benefits that we might want. Allow every voter 10 votes and then elect the 20 candidates with the highest vote totals. Or 5 votes and 10 elected candidates, or 50 votes and 100 elected candidates. The math here is a bit subtle, but the "10 votes elects 20 candidates" rule helps bring forth voter preferences, and if that sounds counter-intuitive, the basic math fact that makes this intuitive is that the the elected losers (the 11th, 12th, 13th, 14th, 15th, 16th, 17th, 18th, 19th and 20th vote getter) support the majority of the issues that are supported by the majority of the public. In other words, imagine a society where 51% of the public want more education spending and 51% want more military spending, the elected losers would support those 2 issues, but then hold an unpopular position on some other issue.
And if that sounds unlikely, I'm currently working on a software simulation that will allow people to see how these elections play out. But the math on this is solid.
To make the voting truly fine-grained, my suggestion is that the elections happen every month, for long terms, perhaps 10 years or 15 or 20 years -- the combination of monthly voting and very long terms means that the legislature would capture every mood of the public, and rather than being unbalanced by dark passions in any one particular year, the legislature would hold a beautiful average of the many moods that moved the public over many years. That averaged mood would help protect against extremists.
My point is, more data does not fix any problem. But replacing course-grained voting with fine-grained voting does solve real problems.