hudebnik

OK, we all knew Trump was fond of bribery —- both giving bribes to public officials like Pam Bondi when he wanted them to not investigate his University and similar schemes, and even better, taking what appear to be bribes himself now that he’s a public official. But this Times interview lays out how mind-bogglingly, unprecedentedly, openly corrupt the Trump family’s crypto dealings are.

Yes, there have been lots of corrupt Federal officials before, but they usually tried to hide their corruption, while Trump is completely open about it: he knows that the only way he can be held accountable is through impeachment, which won’t happen as long as his party controls at least one house of Congress. And the scale is probably at least an order of magnitude beyond all previous Federal bribery scandals combined.

It raises some interesting questions: assuming Trump left office and we still had a vaguely functioning republic, how would we even begin to fix this?

(1) Overturn the Supreme Court’s immunity decision, and make clear that somebody who uses Presidential powers to commit a crime can be prosecuted for that crime. For example, the President has nearly unlimited pardon power, but granting a pardon in exchange for money would still be prosecutable as bribery. This would require either a SCOTUS decision or, better, a Constitutional amendment.

(2) Put teeth in the emoluments clause. We can’t expect the DoJ to criminally charge a sitting President, but we could impose a 100% income tax rate on the fair market value of any gift the President (or any other Federal officer) receives in violation of the emoluments clause. This could be done by act of Congress, and signed into law by one honest President.

(3) Make impeachment slightly easier —- say, a 3/5 majority in the Senate rather than a 2/3 majority. Again, this would require a Constitutional amendment. Trump’s second impeachment failed to convict by a vote of 57-43, not quite a 3/5 majority, but the margin would have been much less comfortable, and three more Republican Senators might have voted to convict if they had thought it possible to come out on the winning side.

(4) How to deal with cryptocurrency? The basic problem is that it enables anybody who wishes, anywhere in the world, to put large amounts of money directly into the President’s pocket, anonymously but verifiably if the donor wishes to prove it to the recipient in private. If you were designing a system to encourage bribery, you couldn’t do much better. And the Trumps are currently using it in several different ways, as the Times interview points out:

(a) direct gifts from the donor’s “wallet” to the recipient’s,

(b) purchases of assets from which the recipient receives a transaction fee,

(c) purchases of floating assets of which the recipient holds a lot, driving up their price, and

(d) purchases of stablecoins from the recipient, which must be backed by actual dollars so they amount to a zero-interest loan to the recipient (who can invest the money in something interest-bearing).

For all of these, the novel problem is anonymity: the public, the IRS, and law enforcement have no way of knowing who gave the President (or other public official) how much money when. If we could fix that, at least for a limited class of recipients (high-level Federal government officials), the usual mechanisms of accountability would at least have a chance to do their jobs.

Existing law requires certain categories of government officials (including the President and Vice President) to file financial disclosure forms, within 30 days of taking office and annually thereafter, identifying the sources and amounts of income, gifts, and reimbursements. In a crypto account, some income, gifts and reimbursements are likely to be truly anonymous, and you can't report what you don't know -- but you can still report the total amounts of anonymous transfers, and you can report the account numbers of any such accounts that you own, so the information can be verified. And the tax code could be amended to apply a 100% income tax rate to the total value of such anonymous receipts beyond an annual threshold.

That would work for dodges (a), (b), and (d), but not for dodge (c), the “driving up the price” trick, unless applied to every purchaser of the asset, even those with no interest in bribing public officials. Fortunately, this is the most unwieldy, inconvenient trick of the four: it costs the donor a lot of money, depending on market demand, to drive up the price of a publicly-held asset sufficiently to affect the recipient's behavior, and the recipient can't realize the gains except by selling large amounts of the asset, which pushes the price back down.

Thirty years ago or so a chiropractor X-rayed my hips and informed me I had a 6mm leg-length difference, which he said wasn't enough to see with the naked eye but was enough to put my pelvis at an angle and my spine into a slight S-curve, so he prescribed a heel insert. Which I used for a few years, and I think it helped with back pain, but over the years I lost the habit (particularly when wearing sandals), and then forgot which heel it went in. I've looked in a bunch of my old shoes for inserts, and found one pair in which the left shoe has a full-length after-market insole and the right shoe doesn't, so I'm guessing it was the left heel. So on my way home yesterday I stopped at a dollar store (they're really going to suffer from Trump's tariffs, since almost everything on the shelves is made in China!) and got some gel heel pads a few mm thick. I put one in the left heel of each of two pairs of shoes that I wear commonly; we'll see what difference it makes.

From this NY Times article ,

“Peter Navarro, the senior counselor for trade and manufacturing, said in a call with reporters Friday that “disastrous” timber and lumber policies “drive up construction and housing costs and impoverish America through large trade deficits that results from exporters like Canada, Germany and Brazil dumping lumber” into the U.S. market.”

Wait: if Canada, Germany, and Brazil are “dumping” lumber on the US market, wouldn’t that drive construction and housing prices down, not up? It might cost some US lumberjack jobs, in exchange for creating more US construction jobs. (Not to mention jobs making things that we export to Canada, Germany, and Brazil.)

Yesterday's Supreme Court decision in Allen v. Milligan, somewhat surprisingly, doesn't further eviscerate the Voting Rights Act of 1965 (as amended in 1982): by a 6-3 margin, the Court agreed with a Federal appeals court that the Alabama map of Congressional districts wrongly denied its 27% black population a fair voice in electing Representatives. I haven't read the details of the appeals-court or Supreme court decision, but the case appears to involve the usual "packing and cracking": putting most of the state's black voters in one district, which almost always elects a black Democrat, so that the remaining six districts have a white Republican majority and almost always elect white Republicans.

Justice Clarence Thomas, in dissent, complains that the Court cannot force a state to use race as a criterion in drawing district boundaries; racism is racism, even in the interest of a minority. Of course, his criticism on this point is one-sided and hypocritical: he seldom objects when a state uses race as a criterion in favor of a white majority, as long as they're not so foolish as to put it in writing. But what if the state didn't use race as a criterion in favor of a white majority? Would he have a legitimate point?

What would a truly race-blind redistricting system look like? It could use not geography, but something with no correlation with race or political party, such as your Social Security number or your birthday (as a number from 1-366). If your Social Security number, divided by 7, has a remainder of 1, you're in district 1; a remainder of 2, district 2; and so on. Absolutely and utterly race-blind, and disastrous for minority representation, because every district in the state would be 27% black; considering the importance of race in Alabama politics, it would be rare to ever see a black representative elected at all.

In Justice Thomas's dream world, the end of overt racism by government would lead to race becoming irrelevant to politics, people would win elections on individual merit, unaffected by race, and in the long run the percentage of black elected officials might resemble the percentage of blacks in the population. But in the real world, that could be a very long run indeed, and it's not clear that Alabama society would actually move in that direction at all; in the interim, black Alabamans would have even less representation than they have now.

The fundamental problem actually isn't about race; it applies to race, political party, sexual orientation and identification, and any other characteristic (call it X) that voters see as politically important. Single-seat, winner-take-all elections amplify majorities. If 60% of voters have characteristic X, an X-blind electoral map will generally give at least 60%, but usually far more than 60%, of the seats to candidates with characteristic X.

You can try to fix this with X-based redistricting, ensuring that only about 60% of the districts have pro-X majorities, but that has several problems. First, it leads to weirdly shaped, unnatural districts, such as the one that gave "gerrymandering" its name. Second, it requires government to explicitly consider X, and to discriminate on the basis of X, in its official proceedings, and that feels wrong, even if saying so puts me in agreement with Clarence Thomas (which also feels wrong). Third, it intentionally involves creating districts in which the general election is a foregone conclusion, and all the real action happens in primaries, which are dominated by more-extreme members of the respective parties. And fourth, what if political decisions depend not only on X, but on unrelated characteristic Y? Do we need to draw districts representing every possible combination of X and Y values?

Some democracies have tried to resolve this using proportional representation by party: if Party Z gets 30% of the votes, Party Z gets roughly 30% of the seats in a legislature (or one house of a legislature, or something). Unfortunately, this (a) fails to recognize any politically-important characteristic that isn't (yet) associated with a political party, and (b) gives political parties an official status in the (state or Federal) Constitution, whereas I would prefer to burn political parties to the ground.

A better answer is a voting system that tends to produce proportional representation (in a state's Congressional delegation or its state legislature), not only by race or party but by whatever characteristic voters feel is important. This cannot happen as long as each district elects a single winner independently; you need either a way for votes in one district to affect outcomes in another, or (more plausibly) multi-seat districts. And you need a way for each voter to express a preference, not only for a single candidate, but for a whole class of candidates who have characteristics that voter considers important (race, party, sexual orientation/identity, musical talent, business acumen, etc; the voter doesn't need to say or even consciously know what the characteristics are). That could in principle mean voters rank the candidates from first to last, or assign each candidate an independent numeric score from 1 to 10, or check off all the candidates they "approve" of, but you definitely need more than a single name per voter.

There are a lot of voting systems out there, each designed to solve a different problem and placing different priorities on the desirable properties of a voting system. I know of one designed specifically to achieve proportional representation, without pre-judging what characteristics voters will find important: Single Transferrable Vote, which is reminiscent of Instant Runoff but applied to multi-seat districts. Each voter ranks all the candidates in the district, and the system first looks at all the first-place votes. Any candidate with at least a threshold number of first-place votes is elected, and the ballots that listed that candidate in first place are removed from the pool, those voters having had their say. (There are variants involving redistributing the "surplus votes" over the threshold.) Of the remaining candidates, some with especially small numbers of first-place votes are eliminated, and we look at the second-place candidates on the ballots that chose them first. This continues in multiple rounds until all the seats are filled.

The system has many of the same problems as IRV: it pays more attention to whom you vote for than whom you vote against, thus favoring divisive candidates over consensus candidates; it's sensitive to "butterfly effects", in that a handful of votes changing hands between minor candidates can have a cascading effect on major candidates; and it's non-parallelizable, in that you can't do any substantial part of the counting work at the precinct level, you have to do everything globally. And it's even more complicated, difficult to explain, and difficult to justify than IRV. Among other things, there's no single obviously-right way to choose the thresholds for "elected" and "eliminated", nor any single obviously-right way to decide which surplus votes to redistribute. But it does a very good job of producing proportional representation, if there are enough seats per district (ideally, a single district for the whole state).

Is there a way to achieve something like proportional representation, but with a simpler, easier-to-understand, easier-to-justify voting system? I'm a big fan of the Borda count for single-seat elections; maybe that could be adapted? In the Borda system, if you vote for 7 candidates, your first choice gets 7 points, your second 6 points, and so on. So if there are 7 seats to fill, it seems natural to give them to the candidates with the top 7 point totals. The result won't be as nearly proportional as STV would produce, but you would expect a coherent, disciplined minority party to be able to elect a few top candidates over lower-ranked candidates of the majority party. But I don't think that works. I'm still working through the algebra, but it looks as though a majority party can fill all the seats, no matter what minority parties do, simply by telling its voters to rank the majority-party candidates randomly and ahead of all the other-party candidates. Furthermore, the majority party's advantage increases with the addition of minor, long-shot candidates -- not because they take votes away from either "real" party, but because they decrease the point ratio between a party's top-ranked and last-ranked candidates, and whenever this ratio is less than the ratio in party allegiance among voters, the majority party wins.

You could perhaps fix the above problem by doing a Borda count in which the number of candidates you can vote for (and thus the maximum number of points you give any one of them) is substantially less than the number of seats to fill. But that defeats one of the appeals of the Borda system: the fact that you can meaningfully downvote a candidate by ranking it last, and downvotes matter just as much as upvotes. I think Borda just doesn't work for multi-seat elections.

try this story about NYC rent (and a real-estate app that's used nationwide, so this isn't just about NYC).

The immediate impetus for the current post is this quotation from Florida's new anti-CRT law:

An individual, by virtue of his or her race or sex, does not bear responsibility for actions committed in the past by other members of the same race or sex. An individual should not be made to feel discomfort, guilt, anguish, or any other form of psychological distress on account of his or her race.

(taken from this opinion piece about Governor Ron DeSantis's political trajectory)

That sounds at first like an unobjectionable, anti-racist, anti-sexist sentiment. Individual blacks shouldn't be pre-judged based on what other blacks have done, individual whites shouldn't be pre-judged based on what other whites have done, individual Moslems shouldn't be pre-judged based on what other Moslems have done, individual Catholics shouldn't be pre-judged based on what other Catholics have done, individual men or women or NB's shouldn't be pre-judged based on what other men, women, or NB's have done, etc. Who could object to that? (Leaving aside the irony that this law uses exactly the same language of "discomfort" and "psychological distress" for which conservatives often mock liberals as "fragile snowflakes".)

The fly in the ointment is the word "responsibility", which in my experience has at least three different meanings (in the "something went wrong and needs to be fixed" context), which I explored here. In brief,

• Who's (most) to blame? Who caused the the thing to go wrong?


• Who's (most) capable of fixing the thing?


• Who stands to benefit (most) from the thing being fixed?

If all three of these are the same person (or corporation or whatever), the thing is likely to be fixed, and unlikely to recur because the person who caused it this time paid for that mistake. If they're three different people, the thing is unlikely to be fixed, and likely to happen again.

It occurs to me now that in some cases there's a fourth person:

• Who benefited (or expected to benefit) from the thing going wrong?

So, let's look at the history of race in America, in particular its effect on socioeconomic inequality (which is tied up with educational inequality, neighborhood crime rates, etc.)

Who's to blame? Obviously, I didn't personally cause slavery in the 17th-19th centuries, or Jim Crow and lynching in the late 19th and 20th, or real-estate red-lining in the mid-20th. Indeed, only the most recent and ongoing examples of institutional racism (which really are much less bad than fifty years ago, or a hundred years ago, or two hundred years ago) have anybody living who could be "blamed" for them. If we expect the problem to be fixed by those "to blame", it'll never be fixed because those people are mostly dead.

Who stands to benefit from the problem being fixed? One could argue that every member of society would benefit from reducing inequality, in that the resulting society would be stabler, more peaceful, with less crime, etc. But obviously those who stand to benefit the most are those on the losing end, those descended from many generations of people who were poor and/or uneducated largely because of their race. That's not me either.

Who's capable of fixing the problem? Again, almost everybody has something to contribute to solving the problem, but the people most capable of fixing the problem are, as always, those with the most socioeconomic power -- precisely the people who don't stand to benefit from the problem being fixed. And it's such a huge problem that no individual, no matter how wealthy, is likely to be able to make much of a dent in it, which allows most of us, even the relatively wealthy, to excuse ourselves from it.

And finally, who benefits from the problem? Even though it's "not my fault", I benefit from slavery and subsequent racial discrimination. According to Ancestry, I have several ancestors who owned slaves (mostly in eastern Maryland), and their wealth presumably affected the socioeconomic and educational status of their non-slave-owning descendants, including me. Both absolutely (in dollars and degrees) and relatively (to the rest of the US population, particularly the black population), I am almost certainly richer and better-educated than I would be if my distant ancestors hadn't owned slaves, and if subsequent generations hadn't been on the winning side of government policies and societal norms that explicitly favored whites over blacks, Native Americans, Chinese, etc.

So who's "responsible" for the socioeconomic problems of race in the US? If you think of "responsibility" in terms of blame, you're probably off the hook because you're alive today. If you think of "responsibility" in terms of ability to fix it, you're probably off the hook because you're not a billionaire or a government. If you think of "responsibility" in terms of incentive to fix it, most of the people reading this post are off the hook because they're above average in income and education. Only in qui bono terms do you bear any responsibility. So if your goal is to absolve yourself of responsibility, it's critically important to avoid any discussion of qui bono, and that's the purpose of the law.

On the other hand, if your goal is to fix the problem, discussing "blame" is irrelevant and useless: the only way to get a problem fixed is to somehow align the people most able to fix it (which in this case is those who benefit from the problem) with those with the most incentive to see it fixed.

With the start of the new US Federal fiscal year, we have new debt-and-deficit figures. And following the same methodology as for the past fifteen years, we get the remarkable result that the Federal budget deficit actually shrank during the Trump administration. It doubled during the first two years, was on track to double again for 2019-2020 before COVID hit, and actually quadrupled instead as a result of COVID. But the 2020-2021 deficit was completely erased by inflation: the national debt grew by slightly less than the rate of inflation. (It's possible that the deficit looks smaller than it "really" was because the Federal Reserve started taking "extraordinary measures", i.e. accounting tricks, towards the end of the fiscal year to avoid the debt limit. I haven't checked on that yet.)

That said, the Trump administration (including COVID) oversaw almost twice as much accrued debt per year as the G.W. Bush administration (including the 2008 recession). Likewise, the pre-crisis Trump administration (cutting off at December 2019) oversaw almost twice as much accrued debt per year as the pre-crisis G.W. Bush administration (cutting off at September 2008).

Among all the other news in the past week, the U.S. Federal Government started a new fiscal year on Oct. 1. So in keeping with Time Honored Tradition, I've updated my web page about Federal budget deficits.

In brief, before COVID-19 hit the U.S., fiscal 2019-2020 was already on track to have the second highest budget deficit in history (behind 2008-2009; all figures are inflation adjusted). That's what I get by extrapolating linearly from the first three months of fiscal 2019-2020, by the end of which China was getting hit hard but there wasn't yet a confirmed case in the U.S. And that deficit was already on track to be twice the 2018-2019 deficit, which was slightly larger than 2017-2018's deficit, which was twice 2016-2017's deficit.

But including the effects of COVID-19 (the tax revenue losses and the special spending), fiscal 2019-2020's deficit more than doubles again over the doubling it was already doing.

A great deal of conservative economics (the respectable kind, as practiced by conservative professional economists, as opposed to what Krugman calls "professional conservative economists") is based on the valid principle of "incentives": if you reward people for doing X and punish them for doing Y, then people (in aggregate) will do more of X and less of Y. (Liberal economists may retort that some individuals, through no fault of their own, can't do X, so a reward for doing so doesn't help them and merely puts them farther behind the curve than they already were.)

From this perspective it makes sense to decide what's in the public interest, and institute systems that reward people for doing that and punish people for doing the opposite. And if "the public interest" is defined largely in terms of GDP, then one activity in the public interest is to produce valuable goods and services, i.e. to work.

Now add the axiom that the value of your work is equal to what you get paid for it, and we conclude that to get people to do what's in the public interest, you merely have to let the market reward them with pay, and have the government interfere as little as possible -- in particular, tax their pay as little as possible, not progressively but flatly or even regressively, so they get as much reward as possible for producing more value.

Of course, that axiom is questionable (see here). In brief, if you're an ordinary employee of a large employer, the value of your work is an upper bound on what you get paid, but the less bargaining power you have, the farther below that value you'll actually get, and the more of that value will go to your employer instead. And if you're an upper-echelon employee of a large employer, you may actually get paid more than the value of your work, because you set your own salary (or you and a small cadre of similar people jointly set their own salaries -- it has to be a small cadre or they would bankrupt the company). But this phenomenon acts like a regressive tax and gives people even more incentive to work harder and move up in the company, so some might see this axiom failure as actually a good thing.

That axiom also largely ignores work not paid for in money, such as homemaking and caring for (sick, old, young) relatives. One can sweep this under the rug by assuming that every economic unit (let's call it a "family") has one person working for pay and one person doing the homemaking and personal-care tasks; if so, then this only applies a constant multiplier to everything and doesn't change the choices made by those families. This doesn't work so well when some families fit that mold and some don't, but let's leave that aside for now.

Anyway, the originally-utilitarian principle that, for the public interest, people should be rewarded for working-for-pay and punished for not-working-for-pay has taken on a moral quality: "he who will not work, he shall not eat." (Similarly, a lot of social practices that were originally considered in the individual or public interest, such as monogamy, not eating pork, or not eating other people, have become moral rules because it's easier to teach masses of people a moral rule than to have them all follow a utilitarian argument.) Those who believe deeply in this moral principle see anybody who doesn't work-for-pay, for whatever reason, as a "moocher", and any politician who wants to help unemployed people eat as simply "buying votes" with pan et circenses.

For some reason, conservative economists believe there is no such thing as involuntary unemployment: if you can't find work, it's because you're demanding too much pay, because there is always work available for anybody willing to work for sufficiently low wages. And if the value of your labor (= your wages, see above) isn't enough to cover the expenses (uniform, commuting, child care, etc.) of you going to work, you should probably go to work anyway because it's the Morally Right Thing To Do. It is of course true that if everybody doesn't work, no value will be produced and everybody will starve, not to mention having no clothes or videogames... but it's less obvious that if a few people don't work, perhaps for reasons beyond their control, those few people should starve.

In many economic circumstances, it's also in the public interest for investment capital to be available, which requires that people save money, i.e. spend less than they earn in the same period of time. This too has become a moral principle: everybody knows that it's right and moral and responsible to save money for the future, while those who do the opposite are irresponsible spendthrifts.
So government policy should reward people who save money with high real (after-inflation) interest rates. (The liberal economist will point out that for a substantial number of people living paycheck to paycheck, "saving money" is essentially impossible; even buying in bulk and buying higher-quality goods because they're more cost-effective in the long run are impossible because there's never enough cash available at one time.)

For some reason, conservative economists seem to think there's always a shortage of investment capital, and all economic problems can be solved by throwing more capital at them, so it's always in the public interest for people to save money and to be incentivized to do so with high real interest rates. The obvious counterexample is the recession of 2008, when lots of investment capital sat around with nowhere to invest because nobody was buying anything; what we needed then in the public interest was for people to spend money, even more than they earned, in order to kick-start the economy.

Which brings us to COVID. In a mass-contagion event, it's in the public interest for large numbers of people to stay home and not interact much with other people, and that means (for many of them) not working-for-pay. So it would seem to make sense for government policy, for the duration of the pestilence, to reward people for staying home and punish those who interact face-to-face with lots of other people. That's what several Asian countries (especially those that remember SARS) and most European countries have done: if you normally work an in-person job, we'll pay you to temporarily stop doing it in the public interest, and we'll fine people who congregate in large numbers. And that's what the US's "Paycheck Protection Act" was supposed to do: reward employers who keep their employees on the payroll even while they can't do their normal jobs. But the PPA didn't apply to small employers, and it didn't apply to large employers, and the amount of money authorized in it was an order of magnitude too small for the number of people who suddenly can't work, and it was implemented in a tremendous hurry with unclear rules for who could qualify and what they needed to do to qualify and with minimal oversight of how the money was actually used, and most of the money went to relatively-large employers who already had the clerical infrastructure to meet the qualifications, and and and.

Unfortunately, what's currently in the public interest collides with the moral principle of "who will not work, he shall not eat", based originally on the public interest in more-normal times. Naturally, the people who believe most strongly in this moral principle are vehemently opposed to helping anybody who's not working-for-pay, for whatever reason: if you're in a high-risk group, or your job involves a lot of physical contact with other people, or your employer has shut down because the mean old government has told it to, or you or somebody in your household are already sick, the fact remains that you're not working and therefore it would be immoral for you to be able to eat. Some Republican members of Congress are outraged that, with the temporary increase in unemployment benefits, some people might be better off unemployed than working -- which was exactly the point: we want them to stop going to work so they don't spread disease.

There ought to be a clever, penetrating conclusion to all these ramblings, but I don't have one and I need to go to work now :-)

So yesterday New York City's voters overwhelmingly approved an Instant Runoff Voting system. It only applies to primary and special elections, not general elections, and it only applies to certain city offices, not state or Federal offices (I presume those would require action at the State level, which is actually possible now that Democrats control both houses of the State legislature). And I'm sure it'll be an improvement over the current system.

It bothers me that so much media coverage calls this "ranked choice voting, also known as instant runoff voting." Those are not the same thing. Ranked choice voting, in a rational world, means that the voters cast ranked ballots. Instant runoff voting is one of the possible ways to count ranked ballots, and not a particularly good one; the other two well-known ways to count ranked ballots, the Borda system and the Condorcet system, are better in almost every way.

Vote-splitting

OK, why is a single-vote system bad? The most well-known reason is vote-splitting. Imagine there are six candidates: five nearly indistinguishable candidates from the Sane Party (which, being sane, gets 80% popular support for its positions) and one from the Loony Party (with 20% popular support for its positions). In a single-vote system, the five Sane candidates may split their 80% popular support, each get 16%, and the one Loony candidate wins with 20%. A runoff system avoids this scenario (at the expense of running an additional election), but is still susceptible to a scenario involving several Sane candidates, one Loony candidate, and one Bonkers candidate: the runoff can be between the Loony and the Bonkers, even though very few voters support either of them. The traditional solution to this problem has been political parties, which choose (whether by primary, caucus, or smoke-filled room) a single nominee from their field of possibilities, so there aren't several similar candidates running against one another. The downsides are substantial: the cost of running a separate election just to help the political parties decide whom they're going to put into the general election (and why do taxpayers pay for that, rather than the parties?), and the well-known phenomenon of candidates running to the extremes in primaries and suddenly moderating all their positions as soon as they win nomination, and the expense and old-boy corruption historically associated with party machines. Ranked ballots are an alternative solution to the vote-splitting problem, allowing several similar candidates to be in the same race without significantly hurting one another. In my ideal world, ranked ballots would make it unnecessary to have political parties or primaries at all: you just put all the candidates into the general election and have the voters decide. Yesterday's NYC referendum doesn't quite reach that desirable outcome, because it doesn't apply to general elections, but perhaps that can be a few years from now.

Strategic voting

Another problem with a single-vote system is "strategic" or "insincere" voting: it encourages voters to vote not for the candidate they most want to win, but the candidate they most want to win from among those they've been told have a realistic chance of winning, which in turn depends on which media outlets they watch/read and which public-opinion polls they believe. Every voter has to cast a vote based not on hir own preferences, but on hir own beliefs about other voters' preferences. How can a "democratic" voting system possibly do what the voters want if the voters aren't even telling it the truth about what they want? (The "realistic chance of winning" filter also makes it almost impossible for a third party to win, no matter how popular its positions are.) With a ranked ballot, there are fewer and rarer scenarios in which strategic voting helps you; lying about what you want is harder and more complicated, while telling the truth about what you want is easier, so people are more likely to do it. Instant Runoff is still fairly susceptible to strategic voting; Borda much less so; and Condorcet even less so.

Asymmetry of weights

A third problem with single-vote systems is that they weight a candidate's positives much more heavily than the candidate's negatives. (A hypothetical single-vote system in which you vote against your least favorite candidate, and the candidate with the fewest anti-votes wins, would have the opposite problem, weighting a candidate's negatives much more heavily than the candidate's positives.) In a single-vote system, the difference between your first choice and your second choice is everything, while the difference between your second choice and your tenth is ignored completely. The result is that a candidate with the support of a strong minority can win despite being despised (ranked last) by a solid majority of voters; a divisive, factional candidate will beat a boring, consensus candidate every time.

Instant runoff has the same problem, slightly less badly, because it doesn't look at the difference between your second choice and your tenth until your first choice is eliminated. If your first choice survives to the final round, IRV doesn't know whether the other candidate in that round was your second choice ("almost as good") or your last choice ("over my dead body"). Imagine an election in which Anne, Bob, and Charlie split the first-place votes evenly, while Denise is everybody's second choice. A majority of voters (all the Bob and Charlie partisans) prefer Denise over Anne; a different majority prefer Denise over Bob; a third majority prefer Denise over Charlie; and Denise is the most likely to be able to run a functioning government because nobody hates her; but Denise is eliminated at the first round because she didn't get any first-place votes.

By contrast, in the Borda and Condorcet systems, the difference between your first and second choices carries the same weight as the difference between your ninth and tenth choices; you can express your extreme disapproval of a candidate by ranking that candidate last, and it matters that you did so. As a result, Denise wins the above scenario, and a majority of voters are happier with the outcome.

Mathematical properties

Then there are the mathematical properties: common-sense things mathematicians would like to see in a voting system, such as "if your preferences and mine are exact opposites, your vote and mine cancel one another out." Due to the asymmetry of weights discussed in the previous two paragraphs, this isn't true in a single-vote system, nor in IRV, but it is true in both Borda and Condorcet. And single-vote and IRV are subject to "kingmakers": the presence or absence of one non-winning candidate can change who does win, which is almost impossible in Borda, and I think completely impossible in Condorcet.

Algorithmic efficiency

Finally, there are the practicalities of counting ballots. A single-vote system is easy to count: maintain a counter (initially zero) for each candidate, and for each ballot, increment the counter for the candidate chosen on that ballot; then see which candidate has the most votes. It's a linear-time algorithm (more precisely O(V + C), where V is the number of voters and C the number of candidates), and easily parallelized so you can do a count within each precinct, report a number for each candidate (C numbers to report), add them up and get valid city-wide, state-wide, or nation-wide results.

Instant runoff is much trickier. There are two obvious algorithms. In one algorithm, you loop through all the ballots, incrementing first-place counters as above in O(V) time, then in O(C) time find out which candidate has the most and fewest votes and whether the most votes is more than 50%, but if not, you then have to find all the ballots whose first choice was the least-popular, and go through them again, and perhaps do that again, and again, until somebody has 50%+1. Finding the number of first-place votes for each candidate is still parallelizable, but you can't do the "instant runoff" part locally because you don't know which candidate to eliminate until you've finished the global count; then you have to tell each precinct which candidate to eliminate and have them start all over, and possibly repeat this up to C-2 times. Total computation time O(C*(V+C)), with O(C) rounds of communication. If you want to know who came in second, third, etc. you need to do even more rounds of communication.

In the other algorithm, each precinct counts how many votes there were for each possible permutation of candidates, and reports this number to a central tabulator, which adds up the number for each permutation, and this is enough information to find a winner. But this requires reporting and adding up C! buckets, not only C buckets.

In the Condorcet system, you loop over each pair (X,Y) of candidates, and decide whether a majority of voters prefer X over Y or Y over X. This produces (in time O(C²V)) a bunch of pairwise win/lose choices, and a standard tournament algorithm (in time O(C)) finds who beat everybody else. Unfortunately, it's possible that nobody beat everybody else: it's possible for Anne to beat Bob, Bob to beat Charlie, and Charlie to beat Anne, resulting in a three-way tie which is the biggest disadvantage of the Condorcet system. (There are ways to break the tie by counting by how many votes Anne beat Bob, Bob beat Charlie, and Charlie beat Anne, but they make the system more complex to describe and analyze.) The system is moderately parallelizable: each precinct counts, for each pair (X,Y), how many voters preferred X over Y and how many Y over X; these O(C²) numbers can be reported to a central tabulator, which adds up the numbers for each pair and has enough information to find a winner.

In the Borda system, you maintain a counter (initially zero) for each candidate, and for each ballot, increment each candidate's counter by a number of "points" corresponding to the candidate's rank: if the voter ranked five candidates, give them 5, 4, 3, 2, and 1 points respectively. This takes O(C*V) time, but the O(C) factor is localized to a single ballot, not multiple rounds as in IRV. Then in O(C) additional time, find which candidate has the most points, and declare that candidate the winner. The system is easily parallelizable: each precinct computes a point total for each candidate, reports these C numbers to a central tabulator, the central tabulator adds them up and has enough information to find a winner (and indeed to rank all the candidates). There are only C numbers to report (although they're O(C) times larger), rather than C choose 2 or C factorial.

For a 10-candidate race, single-vote or Borda requires reporting 10 numbers from each precinct to central; Condorcet requires reporting 45 numbers; IRV with the first algorithm requires reporting 10 numbers but possibly having each precinct go through its ballots up to 7 more times; and IRV with the second algorithm requires that each precinct report over 3 million numbers (most of which may be zeroes) to the central tabulator, regardless of how many voters are in the precinct.

So, "yay, sorta". A definite step in the right direction, but if we were improving our voting system, why couldn't we have improved it more?

ETA: see today's XKCD.