Actual success in laws to reduce corruption and money in politics

At this week's Singularity U Global Summit, I got a chance to meet with Josh Silver and learn about his organization, I have written often in My New Democracy Category on ways to attack the corruption and money in politics. is making a push for the use of laws to fix some of these issues, through ballot propositions. In the past, I have felt this approach to be very difficult, because for every step that could improve democracy, one of the major parties is benefiting from the flaw, and will fight any effort to fix it. Fixes in congress or the statehouses are difficult, and many of the fixes people like (like campaigning restrictions) violate the 1st amendment.

This organization is trying for a few specific measures in a bipartisan effort to pass ballot resolutions. To make it bipartisan, they are doing it in pairs of "red" and "blue" states. The core changes they are looking for are:

  • Public campaign finance through vouchers. Every voter gets "vouchers" they can hand to the candidates they wish
  • Rules to fix the nightmare of gerrymandering, primarily by having non-partisan committees draw the district boundaries, as has already happened in some states
  • Preferential ballot systems to allow minor parties to participate in elections without risk of "spoiling" the battle between the 2 main parties, as Nader did in Florida 2000 and Perot did in 1992.
  • Improved voter participation though improved registration (another common approach in place in some districts.)
  • Limitations on revolving door lobbying and favours for donors.

RU's plan is a surprising one -- that all 4 of these together might have a better chance of passing than the individual components do. Polls show that voters often have strong support for this full package, even if they don't like one of the items. So they have this on the ballot in South Dakota and Washington, though the ballot language in Washington is not superb. They are looking for money and support in their campaigns, and I have offered to be on their advisory board. They have already passed versions of their anti-corruption bills in several cities.

Their strategy might work on me (if I were a voter.) I have my own preferred versions of these approaches, but I would rather see this package pass than fight for the perfect version of either one. Nonetheless a few things I would tweak:

Gerrymandering is one of the great cheats of political systems, and it got a lot worse in 2010 through a deliberate effort of the Republican party to massively overspend national money on key statehouse races, allowing it to control those statehouses and redraw the lines to both assure continued control of the statehouses and a control of the House of Representatives in spite of getting a serious minority of the popular vote. Non-partisan redistricting committees are a start, but we need more, and parties that have gained control this way will be unlikely to give it up. I have advocated a rule of convexity to prevent even partisan groups from gerrymandering. But the only hope I have hear is finding a constitutional principle -- such as the basic right of franchise -- that can get this stopped.

Preferential ballots are good, but sadly the "instant runoff" (also known as Hare, Single Transferable Vote and Australian ballot) is actually the worst of the systems. The problem is not just the chaotic conditions in that simulation article, but that it is one of the harder systems to explain to the voters. If the voters are not immediately clear on how their system works, it causes lack of confidence and probably less voting.

From a purist standpoint, my favourite is Condorcet. It gives good results and can be explained reasonably easily.

Rank your choices in order. To decide the winner, all candidates are compared against all other candidates as though they were in a 2-way race, deciding if more people liked A over B or B over A. The winner is the candidate who beats all the others in these 2-way comparisons. In the very rare case where this doesn't happen, a tiebreaker is done among the candidates with a claim for the top.

On the other hand, the Appoval system is even simpler. Its instructions can be understood quickly by all:

Check the box next to all candidates which you support as suitable for the role. You can check any number from one or all but one. The candidate with the most votes wins.

Approval throws away the fact that you like one candidate more than another, but in reality it seems to work just as well as the systems that don't do that, and it's much simpler to understand. The real flaw is that with Approval, if you have candidates who are close in support, you can get a little "strategy" where voters might not vote for their 2nd choice candidate (even though they like them) out of fear of hurting their first choice. You can't hurt your first choice in Condorcet and instant-runoff, which is a plus, but in reality, this sort of situation doesn't occur in the USA, where there are 2 strong major parties and much weaker minor parties. (Ie. in 2000, every Nader supporter who also liked Gore, and many Gore supporters who liked Nader would have voted for both, even though it was sure Gore would handily defeat Nader.)

Improved participation -- diminishing the value of GOTV is also a good plan, though we need much more here. Even with high registration, voter turnout remains low in the USA, which means that elections are actually won and lost mostly on GOTV.

If you support these plans, then give some money to Represent.US and vote for their measures if you live in Washington or South Dakota.


In 2016 the US Federal Government spend about 4 trillion dollars, taxed about 3.5 trillion, and had a regulatory affect of similar proportion.

So to pick a round number the Federal Government had about a 10 trillion dollar impact on the economy. Those are the stakes. How do you expect any small number of detailed rules to impact forces of that size? It's a bit like squeezing an elephant sized balloon with two fingers.

This is not even taking into account that for politically motivated people its not really about money, its about power.

We're discussing that in Canada, and I read it as "To win, a cantidate needs 50% of the vote. You specify the order you prefer, and losers are dropped one after another until someone has more than 50%"

[formerly Dave Brown when we were at Waterloo, but I married a Collier]

A decent description, but the question is do people understand it when they see the breakout of how it actually went down? When you describe it as a runoff, people who have seen a runoff should understand that but they don't.

Approval's big advantage is that you can publish the result of the ballot counts, and everybody can immediately see how it worked. No algorithm needed at all. With Condorcet, you can also do a fairly decent job, in that you publish something like, "128 voters preferred A to B than B to A. 900 voters preferred A to C than C to A. 1200 voters preferred A to D than D to A. A beats everybody and is the winner." Here you are not publishing the full result but that subset tells you all you need, and you can then look into the full result.

To understand IRV you need to see the chart of the iterations, or the math for Borda. For plurality/FPTP, it is easy to see the result (and easy to see times when it is wrong) and I think that is one of the things people like about it.

All of these schemes are better than first past the post, but all are much worse than proportional representation.

(when there are more than two options, at least).

In case you don't know Arrow's Impossibility Theorem you might want to check it out. It's a fool's errand to find the perfect voting system, and they all seem to potentially fail in meaningful ways in real life. In fact, there was a timely discussion of problems with different voting systems on the most recent episode of EconTalk:
(voting systems were discussed towards the end)

Yes, that theorem is cited in the links I provide, no need to point it out. However, it is very false to say that the theorem says that all systems are equivalently good! Some have many more problems than others.

Also, the theorem only is true if the election has to have one winner. When ties are allowed, there are systems which work very well. We have a strange fear of ties in our electoral systems, and many people have the ridiculous belief that when it is close, the right thing to do is to have a single vote flip how the country is governed. In fact, it seems obvious to me that when there is a near tie, it means that neither side should get power, and there should be some sort of compromise. And there are various systems which do that, including proportional representation systems.

Condorcet has the 3-way (and more) tie situation where A > B, B > C and C > A, but it's very rare. As such, you can do radical things if it happens, like a random tiebreaker, or a second election if you prefer it that way (though in theory that should not differ, in practice it might.) Or a tiebreaker based on margins, such as if if A beat B by more than B beat C and C beat A, then A is the winner. It doesn't actually matter -- it is a tie, and the right answer is not to choose any one of them, but the world seems to insist on that.

Hey Brad, did you read the Atlantic's recent article about how political reform since the 1970s has caused the gridlock and disfunction of today? The article argues by removing seniority and accountability, we have caused a self-interest free-for-all. It's an interesting idea that a legislature might need long-term professional members with a vested interest in order to function effectively. What do you think of this idea?

By removing accountability, do you mean the idea that those term limited out or otherwise not seeking election next time are not accountable? Anyway, I can certainly see term limits as having downsides.

I am interested in radical changes and entirely different systems, but they are not something that can actually happen in the USA or any other place with a well established constitution. What interests me about is they are finding a way to at least do minor but real change within that system. Removing term limits is something that could be done (though for President it's in the constitution) and so could be discussed.

Add new comment