20 November 2020

U.S. Government Health Care Spending v. Military Spending

On Facebook, I recently saw the claim that:
Our taxes do not go for healthcare. Our taxes go for a bloated military complex that wages war on others.
My response was as follows (edited for this post which has more formatting flexibility such as the ability to use paragraph breaks):
We spend a lot more than we should on the military, but you are incorrect. 
Medicare ($644 billion/yr) paid for with payroll taxes and general purpose federal taxes, the combined state and federal Medicaid spending ($603 billion/yr) paid for with general purpose federal, state and local taxes, VA health care benefits ($76 billion/yr) paid for with general purpose federal taxes, and Obamacare tax credits ($52 billion/yr) which reduce federal tax revenues and are financed in part with an Obamacare tax on investment income, for a combined $1,285 billion/yr of the main line items of taxpayer funded government health care spending.

There are also other amounts. There are a few smaller state and local government health care programs and about $188+ billion of other health care related tax breaks. And, there is considerable spending on health care for current local, state and federal government employees (about one in seven U.S. employees in all). 

All of these figures exclude expenditures and tax breaks for medical research and charitable deductions for health care related non-profits.

Total taxpayer funded spending on health care exceeds the combined U.S. military and intelligence agency budgets and non-health care VA budget of about $740 billion/year, and is almost twice as large. 

Advocates often misleadingly look only at federal discretionary spending on health care (excluding Medicare, many VA benefits, tax expenditures, government employee benefits, and state Medicaid spending) to make the amount of tax dollars spend on health care look smaller than the defense budget (almost all of which is discretionary federal spending). But mandatory and discretionary spending at both the federal level and the state and local level is paid for with tax dollars.

Where Did Democrats Stumble In The 2020 House Races?

The Washington Post has the tally of U.S. House of Representatives election results in 2020 at 222 Democrats, 205 Republicans, and 8 races yet to be called. Democrats will control of House where a majority is 218 seats.

If one is a little less cautious than the Washington Post in calling races, the real tally so far (including the runoff that the Republican is sure to win in the GOP tally) is 222 Democrats, 211 Republicans and 2 races too close to call (voting counting has gone slowly in New York and Iowa). The Democrats can afford to have just 4-6 defectors on any given vote in the House (i.e. they need 97%-98% party loyalty on partisan issues).

Democrats had 235 seats immediately after the 2018 midterm elections and the special election in North Carolina (won by a Republican) when that election was invalidated due to election fraud. 

So, Democrats have lost at least 11 net seats so far, and could ultimately lose as many as 13 net seats. 

Roughly one in twenty seats held won Democrats in 2018 have flipped to the Republicans. Republicans flipped two seats in special elections to fill vacancies between 2018 and 2020. Republicans have flipped twelve to fourteen Democratic seats in the 2020 election, including two rated as "Safe Democratic" seats. Democrats have flipped just three Republican seats.

This is pretty horrible, given the circumstances. 

Voter turnout, which tends to favor Democrats was the highest it has been since 1900. Democrats performed better in campaign spending relative to Republicans in 2020 than in any other election in recent history. And, while there is no doubt that the number of seats that Democrats won was reduced by gerrymandering, this can't be blamed for the decline in the number of seats won in the 2020 election relative to the 2018 election which was conducted using them small Congressional district boundaries. And, every incumbent running for Congress in 2020 won a race in the same district just two years ago.

Also, there was ever reason to expect that Democrats would benefit from coattails in down ticket house races. 

Democrat Joe Biden won the Presidential election with 306-232 electoral votes and a roughly 3 percentage point lead in the popular vote (although he underperformed polling averages by about 4 percentage points, on average). Democrats will be hard pressed to secure the 50 seats in the U.S. Senate needed to control the Senate since they need to win two uphill runoff elections in Georgia on January 5, 2021 to do so, but they at least picked up three Republican seats while losing just one Democratic seat (in Alabama where their previous special election win was a fluke made possible by Roy Moore, the truly awful Republican nominee in that election). 

Where did the Democrats stumble?

Democrats lost 1% of their "safe Democratic" seats (as rated by Cook's Political Report) and about half (± one seat) of the "Vulnerable Democrat" seats. 

Five of the seats (three in California and two in Southern Florida) seem to reflect GOP strength in 2020 with certain Hispanic communities. 

Six plus the two uncalled seats (in Iowa, Minnesota, New York, Oklahoma, South Carolina and Utah) seem to reflect a return or possible return to a "natural" GOP lean from a blue wave in 2018 spurred by opposition to Trump. 

One (in New Mexico) seems to involve a bit of both of these factors.

Democrats won just 8% of the "Vulnerable Republican" seats and none of the "Safe Republican" seats. Each was an urban-suburban district in the South (one in Georgia and two in North Carolina) that was about 20% black.

A detailed district by district breakdown supporting the six races that I have called that the Washington Post has not yet called, is below the fold.

19 November 2020

A Defense Of Meritocracy And Its Shortcomings

A Defense Of Meritocracy 

The big problem with critiques of meritocracy is that the critics are rarely clear about they want instead. Mere calls to “be better” are just the Nirvana Fallacy, and aren’t worth addressing. And those contributing to the “meritocritique” aren’t usually even calling for a replacement for meritocracy, they’re just loosely suggesting some kind of redefinition of merit—often holistic approaches instead of SAT scores, even though elite schools and private employers already use a holistic approach. 

. . .

here’s one interesting paper, looking at CEO performance by using 2 measures of stock performance, published in a good management journal:

We found that an Ivy degree granted before 1960 did not confer any performance advantage; the opposite was true for degrees granted after that date. Thus, the value from an Ivy degree is derived not so much from the social capital conferred during the earlier era of social elite selection, but rather the talent associated with selection in the more recent meritocratic era.”

That’s not a clean test, since it really is just lumping all CEOs older than 55 in one group, and all under 55 in another, but by measures of both statistical significance and economic significance, the Old Boys’ Network looks at most half as valuable as the New Meritocrat’s SAT-Infused Diploma. The cynical “Matthew Effect” would predict that the value of an Ivy degree should increase with age as your network of Ivy insiders grows tighter and more powerful, but in real life it shrank—just what you’d expect if SAT-driven college admissions were actually a good way to find real-world, practical talent. . . .

the corporate governance literature has a relevant, non-utopian message: That the good old days of not-too-meritocratic corporations, the kind that JK Galbraith wrote about in The New Industrial State, were actually terrible, and that the relatively more meritocratic LBO era of Milken et al. was much better for the world, warts and all. 

. . . 

Nostalgia for the comfy old days when insiders ran cozy corporate clubs is wildly misplaced, and likewise the utopianism of a more comfortable, less rough-and-tumble, not-too-meritocratic elite selection process is likewise misplaced. I’d turn to Shleifer and Vishny’s influential paper, “Survey of Corporate Finance” on this. Searching for the words “family” and “insider” in their paper gets as some of this.

To exaggerate only slightly, if meritocracy imposed a huge cost, then I’d expect current Italian and pre-Asian-Financial-Crisis Thai firms to be ruling the world: familial capitalism and crony capitalism should be winning corporate models. 

. . .

A non-utopian critique of meritocracy would probably have to start by providing serious evidence that the spoils system and at-will employment in government helped cause prosperity and human flourishing, and that the Pendleton Civil Service Reform Act, on net, hurt prosperity and human flourishing.

A non-utopian meritocritique should probably also offer evidence that the Foreign Service Officer Test is, on net, bad for the State Department and that the U.S. military should stop using the Armed Forces Qualifying Test to screen out military recruits. If you’re serious about the meritocritique, I suspect you’ll have to be serious about dramatically reducing the role of test scores in U.S. government institutions. 

From here


On balance, I support meritocracy. 

This article, however, makes something of a straw man argument on behalf of its opponents. The alternatives suggested are basically "nepotism and clientelism." But the critics of meritocracy that I have encountered (e.g. arguing for the abolition of SAT requirements) aren't arguing for that.

I think that meritocracy can overcome stronger arguments which outlined below, in a lot of circumstances. But any argument should attempt to overcome the strongest arguments against it, rather than weaker arguments against it.

Affirmative Action As A From Of Anti-Nepotism And As A Remedy For Past Social Injustice

Sometimes, they are really usually arguing for some sort of affirmative action, which is really an almost polar opposite of nepotism, in part, because they believe that diversity provides intrinsic benefit to society and prevents it from becoming a harmful oligarchy, and in part, to undo past injustices.

For example, one way to interpret the rise and fall of the U.S. labor movement is that the standards of merit used prior to the Ivy League reforms of the 1960s, were also meritocrat but defined merit in culturally specific terms that excluded people ethnically different from the WASP elite for higher education and management positions. This resulting in talented capable people excluded from the ruling class putting their talents to work instead as union leaders and creating a powerful union movement, that undermined elites formed under the old standards of merit.

The economy-wide downsides of creating a class of under-recognized talented people is a macroeconomic externality that no one institution can make much of a dent in without collective action or some sort of policy mandate, even though this present a global, black swan threat to the stability of the system as a whole, if the system gets too far out of whack as a result and this leads to revolutionary responses.

When meritocratic reforms of college admissions like the SATs, and the emergence of much more robust financial aid removed those barriers, bright young people from the wrong background who otherwise would have joined the labor movement and tried to undermine the goals of managers of big businesses instead were co-opted to become management. As a result, the union movement suffered brain drain and eventually withered.

If you believe that few strikes and less unionization is good for the economy, as many business elites do, then anti-nepotism was a benefit to the economy. On the other hand, if you that a strong union movement led by talented union leaders who were not co-opted by management was beneficial to the economy, then perhaps the less meritocratic approach to college admissions and management hiring that was in place prior to the GI Bill and the Ivy League admissions reforms of the 1960s was beneficial to the economy in a backhanded way.

To connect the dots, affirmative action may be valuable to society and elites, because it co-opts talented people to work for the elites rather than against them.

The Risk Of Misdefining Merit And Thereby Wasting Promising Minds

Other critics of meritocracy tend to favor something of a laissez-faire approach to hiring employees and admitting students that argues that the meritocracy is flawed because fundamentally, as a matter of epistemology, it is impossible for all practical purposes, to accurately define what constitutes "merit". The harm to be prevented here is locking society in on an institutional basis that is hard to change, to an inaccurate definition of merit that excludes some truly exceptional people who have great potential to make huge contributions to society from a path that would allow them to do so. 

For example, at least some of the band members of The Beatles, were put on a dead end vocational school track by the British education system that would have crushed the contribution that they made to music if they hadn't defied its expectations for them.

Maybe high school academic ability isn't the best way to measure the larger contributions one can make to society, and to enterprises. But, if we mandate or prefer meritocracy according to universally accepted by flaws measures, we'll never get a chance to learn that another approach is better.

A laissez-faire approach to hiring employees and admitting students is that could lead to a more "natural" evolution-like way to determine which approaches work better and are worse, because those institutions that use the wrong approaches will disproportionately fail, and if they don't, the distinction is perhaps not all that important.

If meritocracy is really as important as its advocates claim, meritocratic enterprises will thrive and those that are less meritocratic will decline, and policy makers should neither encourage nor discourage meritocratic practices as a matter of law and policy. This is slower than mandating an outcome, and may not optimize the benefits of well done meritocracy. But it avoids the risk of locking in an inaccurate measure of merit.

On the other hand, perhaps meritocracy increases productivity and the benefit conferred by higher educational institutions and we have pretty good evidence already of ways to measure merit accurately. If so, by not taking policy actions to encourage meritocracy, we are allowing our educational system and economy to lag while meritocratic policies elsewhere cause those institutions to excel. If that happens, we might never catch up, because the economy increasingly has winner takes all incentive structures. And, there are lots of barriers to institutional change that can prevent subpar institutions from failing and can prevent better alternatives from thriving, even if the institutions with meritocratic policies are intrinsically better. The status quo can be very hard to dislodge without policy reforms, even when the status quo is an inferior one. 

The Case For Adding Value Rather Than Merely Sorting In Higher Education

In the higher education context, there is also something to be said for the notion that selective admissions are themselves misguided and in conflict with true educational mission of a college or university. Ideally, colleges and universities should be plays that add intrinsic value to their students, rather than merely engaging in a mission to sort the wheat from the chaff and to socialize young people into a college educated class identity. 

If selective higher educational institutions through their admissions processes are basically just sorting talented people from less talented people, without adding much value themselves, maybe these institutions are basically failures and we ought to focus on boosting institutions like the City University of New York (CUNY) system which boosts more working class kids into the middle class than the entire Ivy League combined, than on Harvard and Yale, which educates the best and the brightest, and whose successes have as much to do with being the best and the brightest in the first place as it does with the transformations that they experience in college.

Now, there's also a counterargument to that. Students with insufficient preparation and academic ability going into colleges and universities are most likely to drop out, and those who do graduate consistently show the less objectively measured signs of academic improvement and knowledge and thinking skills gained from the experience. Certainly, there is a decent case to be made for limiting higher education to those who benefit from it, or, in less black and white terms, to most strongly encourage rigorous valuing adding higher educational experiences for those most likely to benefit the most from the experience.

If our educational institutions are really working, we should care a lot about making sure that no one who can benefit from higher education is shut out of this opportunity entirely, something that affirmative action programs in higher education rarely do since these policies have far more impact in selective institutions than in open admissions colleges and universities. But, as a society, we shouldn't care much whether someone gets an opportunity to learn new things at a less prestigious institution like CUNY, or a more prestigious institution like Harvard. So, meritocracy in selective college and university admissions shouldn't matter very much.

On the other hand, maybe there are quality differences between what can gain from attending and getting a degree from CUNY and attending and getting a degree from Harvard. And, maybe the additional value added that a Harvard education provides is a result of a more rigorous curriculum that only the very most academically superior students are capable of benefitting from rather than being overwhelmed by.

In that analysis, selective institutions optimize the objective of providing the most benefit that can be obtained in the aggregate from offering an exceptionally rigorous course of study by having meritocratic admissions.

Also, even if selective college admissions mainly do serve a sorting function, perhaps the college admissions process, and the stress test of determining who manages to actually finish a demanding four year course of study, is a better way to do that sorting than the alternatives, which in turn permits employment hiring to be meritocratic more accurately and easily, which in turn makes the firms that employ college graduates more productive and efficient. This is certainly a very expensive and cumbersome way to engage in sorting, but if the students themselves pay for a large share of the costs of doing so that they are personally benefitting from, then maybe the expense is worth it.

In the absence of this kind of sorting, firms might have to hire and fire more entry level employees and might have to have long probationary periods to weed out new hires with self-discipline and behavior problems that would otherwise have been culled by colleges and universities before firms had to bear the downside risk of hiring such people.

The Case That Favoritism Towards Family And Others One Feels Kinship Towards, And Not Just Business Productivity, Is Part Of The Mission Of The Enterprise Especially In Small Enteprises

A third argument for a laissez-faire approach to hiring employees and admitting students is especially relevant in smaller enterprises. People start businesses to support their families. Nepotism arises, in substantial part, because for the owners or managers of the business it is often the case that providing livelihoods for family members is a more important objective than maximizing the productivity of the business. Nobody questions this self-serving attitude when the business distributes profits to business owners that the business owners used to improve the personal quality of life that they and their families can enjoy, rather than reinvesting the profits to make society in the aggregate more productive. 

Similarly, advocates of this rational argue, private educational institutions and their donors may legitimately include as part of their mission a goal to improve the lives of black young people, or young women, or members of a particular religion, even if this means providing educational resources and institutional prestige to people who further that mission rather than to people who would objectively benefit most from access to those resources and that prestige.

The kind of studies mentioned in the article below, in contrast, are to some extent victims of the implicit ethos of economists that the only thing that matters is maximizing per capita GDP and productivity.

This said, the case for nepotism and favoritism in lieu of meritocracy is much more palatable in the context of small enterprises in a world where there are few barriers to entry than in large publicly held companies, public educational institutions, and less particularistic mission driven private educational institutions, in which no one individual or group can legitimately argue that nepotism and favoritism should be part of the institution's mission.

Seen in this frame, nepotism and favoritism aren't inherently bad for all enterprises and institutions. Instead, they are simply a case of particular institutions, especially governmental and large private institutions, that don't have legitimate interests in preferentially promoting some families or communities over others, acting contrary to the missions of those particular institutions. 

The Case That Meritocracy Is More Important In Some Circumstances Than Others

Also, more supporters of meritocracy imagine that returns to merit are more or less linear. But as apologists for political patronage and nepotist systems often not, lots of hiring decisions involve satisficing and not optimizing.

Firms need someone capable of doing a job. But, in many positions, as long as the job gets done, there is no economically important distinction for the firm between someone who gets the job done well enough and someone who does a truly outstanding job. So, in those kinds of posts, as long as the job gets filled, it doesn't matter to the firm whether the hire was the most qualified candidate or was merely good enough. Indeed, if anything, a candidate who is merely good enough might be better than the most qualified candidate, as the latter is more likely to move on to bigger and better things eventually, rather than becoming a stable part of the firm's workforce.

On the other hand, there are some positions were merit can matter a lot, and there are disproportionate and exponential benefits to having someone who is just a little bit better, especially when the economy creates winner take all circumstances.

For example, maybe having a slightly better lawyer makes the difference between winning and losing a case that determines if an enterprise or institution or someone individual's future thrives or is utterly destroyed.

The same can be true in the case of scientists, senior managers, generals, engineers, or other employees, especially managerial and professional employees.

Again, the basic question is the domain of applicability were meritocracy makes sense. In many kinds of jobs in many kinds of firms, or for many courses of study in many kinds of educational institutions, the important thing that that the people selected are able to get the job done, or are able to complete the course of instruction offered, and there are rapidly diminishing returns to any level of merit that is very far beyond the threshold that is required. But in a few kinds of firms, and for a few courses of study in a few educational institutions, particularly large firms and governmental institutions looking for top level elites for highly demanding or high stakes engagements, maximal meritocracy is highly desirable. And, of course, the world isn't black and white and there are lots of circumstances where merit above a minimum threshold does confer some benefit, but not an exceptionally disproportionate benefit.  

If that is the case, and it probably is, we need to narrow our advocacy for meritocracy to the circumstances were it is particularly desirable, rather than trying to elevate meritocracy to a universal objective.

There is a reason that Ivy League professors and elementary school teachers have different inclinations towards the importance of meritocracy informed by what they do.

13 November 2020

Lessons Learned 2020

So, the dust has, mostly, settled. What have we learned?

The Presidential Race

The was a close election. 

This year, Biden won 306 electoral votes when he needed 270 electoral votes.

Biden's margin in the popular vote was 3.4 percentage points (5.35 million votes with 150.8 million votes counted so far). This will probably improve somewhat as California, New York and late ballots in Alaska and a few other states are counted.

Biden won the marginal state, Wisconsin, by a mere 0.6 percentage margin. Another state he needed to win, Pennsylvania, he won by a 0.7 percentage point margin. He won two states beyond the marginal state, Arizona and Georgia, by a 0.3 percentage point margin each.

Biden won Michigan with a 2.6 percentage point margin and won Nevada with a 2.5 percentage point margin. He won Minnesota with a 7.1 percentage point margin, New Hampshire with a 7.3 percentage point margin, and Maine by a 9.5 percentage point margin. All of the other states Biden won were by more than 10 percentage points.

Trump won North Carolina by 1.3 percentage points, Florida by 3.4 percentage points, Texas by 5.9 percentage points, Ohio by 8.1 percentage points, and Iowa by 8.2 percentage points. All of the other states that Trump won were by more than 10 percentage points, although Alaska had gotten much closer with 88% of the vote counted. Trump's margin in Alaska is currently 10.8 percent points and it could narrow significantly based upon the trend in late counted ballots, before the final count in Alaska (which has counted its votes more slowly than any other state) is in (although almost surely not closely enough to flip the state to Biden).

It is about as close as it could be without a serious threat of a recount or litigation or faithless electors or other subterfuge changing the outcome. The fact that Trump would have to three three separate states to change the outcome. It Trump has done better by 0.4-0.5 percentage points in the swing states, this election would have come down to a 0.1-0.2 percentage point margin of error in a single state.

Presidential polling across the board at both the state and national level, underestimated support for Trump by an average of about 4.5 percentage points (about 10%). The average error in the U.S. Senate polling was close to 7 percentage points. This was probably due to lower response rates to polls by Republicans than by Democrats.

In round numbers, Democrats need a 3 percentage point margin of victory in the popular vote to win the electoral college.

Trump's win over Hillary Clinton in 2016 was with an identical 306 electoral votes and likewise involved razor thing margins of victory in Pennsylvania, Michigan and Wisconsin following deeply erroneous polling biases against him. Biden won Pennsylvania, Michigan, Wisconsin, Arizona and Georgia and Omaha's one electoral vote in Nebraska, which went for Trump in 2016. The District of Columbia, the four other Congressional districts with their own electoral vote, and 45 other states went the same way in the 2020 Presidential election as they did in the 2016 Presidential election.

The House

After the 2016 election, Democrats had 194 U.S. House seats and Republicans had 241. 

After the 2018 election Democrats held 235 seats, Republicans held 199 and there was one vacant seat due to a voided election in NC-9 which a Republican later won. 

Going into the election there were 232 Democrats, 197 Republicans, 1 Libertarian, 1 Other, and 5 vacant. 

The Washington Post has called 219 House seats for Democrats (one more than the 218 seats required for a majority) and 203 House seats for Republicans so far. Democrats who have lost at least 5 seats with Republicans gaining at least 6 seats (5 from Democrats and one from an independent or third-party candidate).

There are 13 House seats that have not been called. Two of the uncalled seats are held by Democrats with a safe lead (NY-18 and NY-19). There are two seats where the Republicans have a safe lead (NY-2 and NY-24), one safe Republican Louisiana House seat headed to a runoff because two or more Republicans split the Republican vote denying anyone a majority (LA-5).

So, as I write, the real split is really 221-206 with 8 seats still in play. There are two very close races currently held by Democrats (CA-39 and NY-3) that had been deemed "safe" Democratic seats going into the election. There are two "vulnerable Democrat" races where the Republicans have a large lead (NY-11 and NY-22), and two vulnerable Democrat races where the Republican challengers lead by a nose (CA-21 and UT-4). There is one vulnerable Democrat race that is a tie and one vulnerable Republican race that is a tie (IA-2 and CA-26) Democrats have tended to do much better in late counted votes, so these seats in California, New York, Utah and Iowa are all likely to go to the Democrats. 

So the final margin in the House is likely to be 229 Democrats and 206 Republicans, a loss of six seats relative to 2018, but it could be worse.

The U.S. Senate

In the U.S. Senate races, Democrats held 47 seats to 53 Republican seats going into the election. So far, Democrats have 48 seats (picking up Arizona and Colorado while losing Arkansas) and the Republicans have 50 seats. This treats two independents, Angus King (ME) and Bernie Sanders (VT) who caucus with the Democrats as Democrats. 

Two more seats, a regular Senate seat and a special vacancy filling Senate seat in Georgia, went to runoff elections between Republican incumbents and Democratic challengers that will be held on January 5, 2021, because no candidate won a majority of the vote. Democrats are the underdogs in both of GA's runoff elections because all Republican candidates combined got more votes than all Democratic candidates combined in the first round and Republicans historically have had better runoff turnout than Democrats relative to the first round general election.

It is possible, but would take a minor miracle, for the Democrats to bring the U.S. Senate to a 50-50 balance which Vice President Harris would tip in the Democrats favor. But, the more likely result is that Republicans will end up controlling the U.S. Senate unless the Democrats can flip one of the moderate Republicans to their side.

Other Races

The Democrats have one fewer Governorship than they did before the election. State legislative election results and other races are still being sorted out.

The U.S. Supreme Court

The recent confirmation of Justice Barrett gives us a 6-3 conservative majority on the U.S. Supreme Court. Conservatives held a 5-4 majority in 2016 and the conservatives majority on the court now is farther to the right than the conservative majority was in 2016.

Bottom Line - Democrats Need To Grow Their Share Of The Population

Biden was the most moderate viable candidate in the 2020 Democratic Presidential primary.

Democratic fundraising was as good as or better than Republican fundraising this year and set record highs.

Voter turnout in 2020 as a percentage of eligible voters, was the highest the U.S. has seen in any federal election since the year 1900.

In short, given the underlying electoral system reality, Democrats did pretty much everything right in record breaking ways.

Yet, Biden won by the smallest margin that could not easily be contested in the electoral college. The Democrats lost several House seats, and the Democrats will pick up mostly likely only one and no more than three Senate seats. Democrats lost a Governorship.

Overall, Biden's coat tails were among the worst of a prevailing candidate in the U.S. history.

Democrats need a 3 percentage point lead in the popular vote to win a Presidential election. Tie votes in the electoral college in Presidential elections go to the Republicans even though the Democrats have a majority of the seats in the House. 

Reapportionment will make a new shift of electoral votes from states that backed Biden to states that backed Trump in 2022, although not by enough that it would have changed the results of this year's Presidential elections. Reapportionment will also put more pressure on the Democrats' thin majority in the U.S. House.

A lack of big coattails at the state and local level in 2020 means that the Republican gerrymandering that has held back Democrats for the last decade is likely to continue to remain in place through the 2030 election at levels just as severe.

The Democrats majority in the U.S. House is much thinner than the share of the popular vote that their candidates in U.S. House races receive due to gerrymandering in Republican controlled states.

Even though the Senate seats held by Democrats represent far more people (about 20 million) than the Senate seats held by Republicans, Republicans will either control the Senate or Democrats will control it by the thinnest of margins and will have a caucus that is hostage to a very conservative Democratic U.S. Senator (Manchin of West Virginia) that will limit the extent to which it can pass a liberal agenda.

Without control of the Senate, Democrats have no realistic possibility of fixing any of the systemic biases of the federal government's status quo against them. They are stuck with a deeply conservative 6-3 majority in the U.S. Supreme Court. They can't admit the District of Columbia or Puerto Rico or any territories as new states or divide any existing states like California into multiple new states. They can't end the electoral college by approving the National Popular Vote interstate compact even if enough states approve it. They can't propose any constitutional amendments that fix the situation. They can't pass significant legislation to revive key provisions of the Voter Rights Act.

The Democrats already have a huge tent, but if they are going to gain control of the federal government to make major reforms, they need to make their tent even bigger, winning over more voters from the Republican coalition to their side than they lose from their own coalition, without the monster that is Donald Trump to force the hands of Republicans and conservative leaning independents who are reluctant to make the switch.

Moreover, the Democrats have to do that in a way that doesn't alienate its base which is fed up with the party's excessive moderation already. The drift of the Republican party into a party of the far right helps the Democrats to hold this together, but there are limits to how broad of a coalition political ideology-wise the Democrats can hold together with their status quo approach.

Some of this can be done by riding demographic, religious and cultural trends with the existing coalition.

Demographic Trends

Not all of the trends are against them. 

Democrats have overwhelmingly won over the young, who will continue to be liberal as they get older and make up a larger share of the electorate, while  Republicans are reliant on an older base that is slowly dying. 

Democrats do better with non-white voters who are a growing share of the population, while Republicans do better with white voters who are a declining share of the population.

Democrats are very strong with secular and non-Christian voters whose share of the nation's population is growing rapidly due to religious de-conversion, while Republicans are wed almost entirely to Christian conservatism with is a shrinking share of Americans apart from the age gap.

Democrats are making great gains among the college educated who are slowly becoming a larger share of the population, while Republicans do best with people with no college education who are a declining share of the population.

Democrats do better with unmarried people who are a growing share of the population, while Republicans do better with married people who are a declining share of the population.

Democrats do better in places with higher population density which are a growing share of the population, while Republicans do better with rural people who are a declining share of the population.

The rise of younger voters and demise of older ones alone shifts 1-2 percentage points of the electorate from the Republicans to the Democrats with their current coalitions, every four years. Lots of the other factors above are partially captured by the rise of younger voters and the demise of older voters. 

But this doesn't capture all of the effects below as they pertain to existing voters who marry interracially and identify less strongly with white voters based upon a spouse or children, who become less religious, who get college educations as adults, or who move to more densely populated places, for example, and experience a change in political inclinations as a result.

Add those factors and perhaps the electorate is shifting across the red to blue divide at a rate of 2-3 percentage points per four years.

Democrats do better in more affluent places which have more economic power in lobbying and campaign finance, while Republicans are seeing their strength slip among big businesses and more affluent people.

On the other hand, the demise of private sector unions in blue collar occupations and industries has been steadily shrinking the Democratic Party coalition among non-college educated voters, especially in Rust Belt states that previously had lots of union members who were a core source of volunteer resources and voters for the Democratic Party. Some of this has been replaced by public sector union members outside law enforcement (who more often have at least some college and have white or pink collar jobs), but public sector union members are not as concentrated in states where Democrats are losing private sector union members as the manufacturing union members of old were and Republicans have effectively added these often ex-union members to their coalition pushing a protectionist, anti-immigrant agenda and distrust of higher education, science and the government.

Taken together with the reapportionment trends, the logical conclusion is that Democrats have a real shot at flipping some marginal red states due to demographic change. Georgia, North Carolina, Florida, Texas and maybe even Alaska and South Carolina, are ripe to flip from red to blue in demography driven shifts to the left in the coming years, just as Arizona, Colorado and Virginia have already in recent years.

But, holding onto Rust Belt states like Michigan, Wisconsin and Pennsylvania may get harder. They may follow in the footsteps of Iowa and Ohio that have already turned red. Fortunately, these prizes are losing electoral votes. But loses of those states still have to be made up elsewhere unless the Democratic coalition is reworked in a way that can bring them in again.


The shift of stated trending blue toward becoming more blue, and of states trending red towards becoming more red, is also driven by immigration and internal migration. Places that are growing economically see Democrats migrate to them from places that are declining economically, at least in relative terms. Rust Belt Democrats and Democrats from rural areas are tending to migrate to Sun Belt cities, leaving the Rust Belt that they left more Republican and making the Sun Belt cities to which they migrated more Democratic. These shifts are almost immediate.

Immigrant migration from abroad also tends to make places that are receiving those populations, some of which are already blue states, but others of which are economically healthy Sun Belt cities more liberal, in part by exposing the local population to more diversity and in part at a significant delay of a decade or so, by becoming citizens and voters themselves and at a generational delay by having children who become voters (with the second effect already captured by the rising young liberal voter trend).

Thus, Sun Belt cities should move left without intervention or action somewhat faster than the nation as a whole simply through natural drift, but at the cost of making rural and rust belt states more conservative.

Perhaps Sun Belt states are drifting left at a rate of as much as 3-4 percentage points a year, while internal migration depresses the overall leftward shift in Rust Belt and rural states to a mere 1-2 percentage points a year.

Changing The Party and Movement Politics

Everything so far in this post has basically reviewed where we stand now and the status quo trends whose inertia is already firmly established.

In the status quo, Democrats still risk losing Michigan, Wisconsin and Pennsylvania, or at least having to struggle to hold onto them each year. Ohio and Iowa are lost causes in this analysis.

Democrats have a good chance of holding onto Arizona, Georgia and Nevada going forward as they shift even more in a blue direction. North Carolina will be a likely Democratic pick up in 2024, and Democrats will have a real shot in Florida, even though it will be challenging.

Texas, Alaska (yeah, definitely not a Sun Belt state, but influenced by its Pacific State neighbors) and South Carolina, on the other hand are prizes that Democrats will only be able to seize if they change up the mix somehow.

To outperform mere drift, Democrats need to do one of two kinds of things. 

One is to adopt policies and make tactical decisions that bring more people into their coalition than they lose relative to the status quo. This is a pure case of changing the party's coalition. It isn't easy, because every change risks alienating some part of the existing party in a big tent party that is already very challenging to hold together.

The other is to adopt policies and engage in grass roots movement politics that cause people to change their attitudes in ways that make them more attracted to what the Democratic party already is or is becoming and to feel less favorable about what the Republican party already is or is becoming.

I'll save brainstorming about those two options for pro-active change to accelerate a natural demographic and cultural drift towards the Democrats for a future post.

12 November 2020

The Weak Foundations Of American Democracy

Election fraud was widespread in the late 18th century, the 19th century and the early 20th century in the U.S., and was often successful. The electoral process (and the legislative process as well) was sloppy, chaotic, unruly and disorganized in that era and easily manipulated as a result.

In urban areas (Chicago was particularly notorious), immigrant dominated "political machines" were often involved.

In the South, violent voter suppression and official misconduct were often at fault. As explained, for example, here:
Black male suffrage became national in 1870 when the Fifteenth Amendment prohibited states from discriminating against potential voters because of race or previous condition of servitude (but not sex). . . . For most of Reconstruction, blacks voted and often used their resulting political power to protect their other rights. Although white Democrats overthrew Reconstruction by forcibly keeping blacks from the polls, thereafter they generally eschewed violence. Instead, they relied upon discriminatory apportionment and election laws to limit black and poor white political influence, and in the 1890s and 1900s, imposed poll taxes and literacy and property requirements through constitutional revisions. Thus they disfranchised virtually all black men and many poor whites, and thereby ensured Democratic hegemony.
In that era, Congress frequently exercised its power to resolve disputed election for House and Senate and usually did so on a purely partisan basis, often allowing the wrong person to take office. From 1789 to 1908 the House of Representatives resolved 382 contested elections (keep in mind that the House of Representatives started out with just 65 members), of which only three were given to the minority party candidate in the House. The first such contest was in 1789 (in that case, upholding the election result), in the first first Congressional election under the current constitution. Since 1910, House of Representatives resolution of contested elections has grown much more rare, and state elections officials and courts almost always resolve the disputes before the House acts.

A more in depth review of the history of those disputes can be found in a 2007 academic journal article which looks at the question of why contests were initiated in the first place, rather than simply how they were resolved. As its abstract explains:
Recent studies of contested elections in the House have pointed to party electoral goals as motivating their resolution in Congress. However, little systematic research has been conducted on why such elections were contested to begin with. Using historical data and new statistical analysis of such elections from the late nineteenth century, I find that, in contrast to the claims of some scholars, political principles as well as electoral objectives mattered to parties seeking to contest elections. In addition, election conditions, particularly the means by which southern white Democrats attempted to repress the vote of southern blacks, independently influenced the probability of contestation. This finding has implications for our understanding of Republican Party strategies and electoral conditions in the South during the period, and the origins of contested elections at other times in American history.
In this era, Congressional elections were somewhat ad hoc, as were Presidential elections in the late 1700s and early 1800s. While the U.S. Constitution authorized it to do so in 1789, Congress only passed a law providing for a uniform nationwide date for choosing Presidential electors in 1845. This law did not affect election dates for Congress, however, which remained within the jurisdiction of State governments, but over time, although the States moved their Congressional elections to this date as well. As late as 1876–77, there were still 8 states with earlier election dates, and 1 state with a later election date.

There were no U.S. Senate elections, of course, until after the 17th Amendment was adopted in 1913, so few of its elections occurred when Congressional resolution of disputed elections was common. There have been 25 election contests in the U.S. Senate since its members began to be elected, 3 of which were successful and 22 of which determined that an incumbent whose re-election was supported by state officials retained his seat. (There were seven cases during the era of legislative appointment, however, where a legislative appointment was rejected as invalid or the state's purported nominee resigned in the face of a challenge.)

The electoral process that resulted in the adoption of the Texas Declaration of Independence in 1836 and formed the Republic of Texas was illegal under the then applicable law of Mexico, and was basically fraudulent.

The election of about 40 men in 1836 that gave rise to the very short lived California Republic was likewise basically fraudulent, although elections were secondary in the U.S. annexation of California which was part and parcel of a large war between the U.S. and Mexico that started in connection with Texas, that was underway before and after California's annexation.

The 1854 election in Kansas over whether it would be a free state or a slave state is another example from the same era.

The outcome in the Hayes-Tilden Presidential election of 1876 was probably a case of election fraud (see also here).

The sequence of events by which the Kingdom of Hawaii adopted its Constitution of 1887 through its annexation to the U.S. in 1898 involved multiple instances of what amounted to election fraud and other unlawful conduct by white businessman who had relocated to the islands. The U.S. government formally apologized for this in 1993.

The dynamics of the election process were also influenced in the early U.S. by the fact that ballots were not secret, although that goes more to duress than outright fraud and may have actually reduced election fraud. In the United States, most states had moved to secret ballots soon after the presidential election of 1884. However, Kentucky was the last state to do so in 1891, when it quit using an oral ballot. Therefore, the first President of the United States elected completely by secret ballots was president Grover Cleveland in 1892.

Needless to say, oral voting was also not auditable in the way that paper ballots are auditable.

In part, this happened because the federal government wasn't taken as seriously and wasn't as important in the early United States. According to political scientist Randall B. Riley, in his book "Congress: Process and Policy":
Until the 1880s the length of service of the average senator and representative remained at a low and fairly constant level -- representatives averaged two years (after a higher level of around three years during a peak of House influence on national policy in the 1810s and 1820s) and Senators around four years. Members of both chambers, especially the House, routinely left Congress for other opportunities, both governmental and private. Mid-term resignations were common.

Beginning about 1880 the average years of service rose dramatically in both houses, doubling in the Senate [to eight years] in less than two decades and almost tripling in the House [to six years] in three decades. The upward trend continued, with breaks for political turnovers, until the late 1960s.
Sessions of Congress were in any case short in the early Congresses because they had little to do. The first Congress met seventeen months out of twenty-four implementing the new constitution. The Civil War and Reconstruction Congresses met between ten and twenty-two months. In every other two year session of Congress until the one commencing in 1911, Congress met for less than twelve of the twenty-four months of the session.

And, the House of Representatives was a tough place. In the 19th century, physical violence on the House floor was not infrequent and guns and knives were occasional carried onto the floor. As an October 22, 2018 article in The Atlantic magazine recounts:
In the three decades before the Civil War, members of the House and Senate routinely threatened each other with violence, and often acted on it too. They brawled on the House floor; they faced off in duels; they fired shots in Congress. They beat each another senseless with canes. All told, members of Congress engaged in at least 80 acts of physical violence between 1830 and 1860, a remarkable fact uncovered by Joanne B. Freeman, a historian at Yale, in her superb new book The Field of Blood.

11 November 2020

State Legislatures Aren't Going To Override Their Voters Presidential Choices In 2020

States with complete Republican or Democratic legislative control (trifectas)

The U.S. Constitution gives states the ability to determine by law how their electoral votes will be allocated. Does this mean that there is a realistic prospect that Trump will be able to convince state legislators to overturn the outcome of the November 3, 2020 general election in states won by Biden to give Trump those electoral votes instead?

Not really. Why?

1. While legislators can determine a method before an election is held, once the election is over the law on the books controls. A law to change a state's allocation of electoral votes from being based on the results of the November 3, 2020 election to a legislative choice would be an unconstitutional ex post facto law. 

2. This legislation would also be contrary to the constitutional provisions that provide that the President must be chosen in some kind of election.

3.  The electoral college votes on December 14, 2020, in less than five weeks. No state legislator anywhere has introduced a bill to take this kind of step anywhere in the country yet. The logistics of building up enough support to get a bill like this passed would be challenging. Many state legislatures aren't even in session right now, so a special session would have to be called and it would have to move fast.

4.   State GOP governors and legislators know that it would be permanent political suicide to overturn the election results from their own state's voters in an attempt to thwart their choice for President. GOP support for Trump's challenges to the election outcome are a mile wide and an inch deep. They are humoring Trump but they know it is a done deal.

5.    The only states where it is politically possible to pass legislation like this, even assuming that conservative judges let partisan bias cloud their judgment and approved the constitutionality of a measure like this, are state where Republicans control both houses of the legislature and the Governorship, so that they could pass the law without Democratic opposition, but Trump won the state's electoral votes.

There are only three states that fit that description: New Hampshire, Georgia and Arizona.

The trouble for Trump is that even if all three of those states were flipped to Trump, contrary to the election outcome on November 3, 2020 in those states, Biden would still have 275 electoral votes and would win.

Republican state legislators and Governors aren't going to commit political suicide in furtherance of a pro-Trump tactic that isn't going to succeed even if they do it. And this attempt is obviously futile if you think about it for even ten minutes and do some minimal research.

So, this last ditch tactic isn't going to work even if Republican state legislators and governors, and the judges in courts asked to review the constitutionality of the legislation were shameless enough to try.

10 November 2020

Voter Turnout As A Percentage Of Voter Eligible Turnout In Presidential Election Years

Voter turnout in the Presidential election in 2020, as a share of people eligible to vote, is projected to be 66.5% and probably exceeded the 65.7% turnout in 1908, but not the record high turnout of 73.7% in the year 1900, which was 120 years ago. 

But this election defied the conventional wisdom that high voter turnout implies a blowout win for Democrats. 

Instead, Biden did only a little bit better than Hillary Clinton did in 2016, making his biggest gains over Clinton in the suburbs, smaller gains in urban areas, and sliding a bit relative to Clinton in rural areas. Specifically, "the biggest average shift was in suburban counties, which went 2.3 points more Democratic than in 2016, vs. urban ones that shifted 1.7 points in the blue direction. (Rural counties shifted just less than a point toward Trump.)."

But, this allowed Biden to win five states (WI, MI, PA, GA and AZ) that Clinton did not, although Biden's margin of victory in all but one of those five states was less than one percentage point. Trump won WI, MI and PA by the narrowest of margins himself in 2016. 

Colorado's 74.5% of eligible voter turnout was tied with New Hampshire for third in the nation behind number one Minnesota (79.5%) and number two Wisconsin (75.5%). The only states that did not set 40 year records were North Dakota, South Dakota, Idaho, Utah, New Mexico, Oklahoma and Mississippi, none of which had competitive Presidential or U.S. Senate contests, and all of them except Oklahoma and Mississippi had more than 60% of eligible voters turnout.

From the Washington Post.