21

I'm not from the U.S. but someone told me that the Republican Party had an unfair advantage in elections for the House of Representatives. This advantage would be geographic.

The voters who usually vote for Democratic candidates tend to live in dense clusters, cities, urban districts. While on the other hand, the voters who usually vote for Republican candidates tend to live in rural districts.

So, much of the vote that Democratic voters cast are "wasted" on candidates who were going to win anyway. Usually, in districts where the Democratic candidate wins, he wins by a very wide margin (which is useless). But in districts where the Republican candidate wins, he wins by a small/narrow margin.

He also said something about the Republicans drawing the maps of the districts. The Republicans create one or two district extremely dense in Democratic voters, so that the Democrats win these districts with a (useless) gigantic margin. And the Republicans tend to win all the other district by small margins. But that looks more like some kind of conspiracy theory to me. Why would the districts get drawn by the parties anyway?

My question is: Is all of this really true?

And if yes, why doesn't the government do something about this? It looks totally unfair and ridiculous, so if this was true I would think that a lot of Americans would protest against this, no? Yet I have never heard of any such protest.

  • 3
    Redistricting is required by the Constitution. Each individual state controls the process for drawing its own congressional districts. The Congress was controlled by the democrats from 1960s-1990s Census. Was it unfair then? – user1873 Nov 14 '14 at 4:02
21

Republican Party had an unfair advantage in elections for the House of Representatives. This advantage would be geographic.

Well, "Fair" is in the eye of the beholder :)

But the geographic advantage in the House of Representatives is undisputably there. FiveThirtyEight covers it in detail in Feb 2103 post "Did Democrats Get Lucky in the Electoral College?":

However, much or most of the Republican advantage in the House results from geography rather than deliberate attempts to gerrymander districts. Liberals tend to cluster in dense urban centers, creating districts in which Democrats might earn as much as 80 or 90 percent of the vote. In contrast, even the most conservative districts in the country tend not to give more than about 70 or 75 percent of their vote to Republicans. This means that Democrats have more wasted votes in the cities than Republicans do in the countryside, depriving Democrats of votes at the margin in swing districts. Eliminating partisan gerrymandering would reduce the G.O.P.’s advantage in the House but not eliminate it.

However, to balance that advantage, the main topic of the linked article exists: Democrats have a similar, geography-based, electoral advantage in Presidential elections.


He also said something about the Republicans drawing the maps of the districts ... The Republicans create ...

Partly true, and partly false.

False: this is NOT a "Republican" thing. Both parties do it (see Tenthustice's excellent answer for details). FiveThirtyEight (not exactly a conservative bastion) says:

Democrats were no less aggressive about creating gerrymandered districts in states like Illinois.

However, as another answer noted, Republicans had more opportunity to do this in 2010 compared to Democrats, by virtue of winning more state legislatures. Meaning that between 2010 and 2020 censuses, they can gerrymander more states than Democrats can, despite equal desire by both to do so.


And if yes, why doesn't the government do something about this? It looks totally unfair and ridiculous, so if this was true I would think that a lot of Americans would protest against this, no? Yet I have never heard of any such protest.

Several explanations:

  1. Gerrymandering may be detrimental in the eyes of individual people... but it's a swell thing for every individual politician who is granted an enormous electoral advantage by it in "their" district, no matter what party. If you can guarantee your election by 30% margin, why would you put your election at risk by giving 20% of that margin to another district to let a fellow co-partier get elected?

  2. The "government" doesn't do anything about it because Constitution explicitly relegates the authority over how to do this to the state level.

  • 1
    Another measure is over-representation of the small states in Senate. Every state, regardless of the population size, has two senators. Senators representing around 10% of the total population can filibuster any activity - and those low-population states vote overwhelmingly Republican. So Republicans have similar unfair advantage in Senate too. – Peter M. Nov 18 '14 at 14:25
  • 12
    @PeterMasiar - That's a feature, not a bug. (At least, it was 227 years ago.) – Bobson Dec 17 '14 at 22:38
  • 1
    @PeterMasiar, in addition to Bobson's comment, the structure of the Senate is completely irrelevant to this question. – A Bailey May 4 '17 at 19:51
  • 4
    "If you can guarantee your election by 30% margin, why would you put your election at risk by giving 20% of that margin to another district to let a fellow co-partier get elected?" Actually, the whole point of gerrymandering is to do exactly that. The party drawing the lines gives a 10-15% margin to each of its members (more than enough in all but the biggest wave years) by making districts that have 30%+ margins for the opposition party. The representatives from the opposition party may appreciate running unopposed; it's the party drawing the lines that wants to win by "enough, but no more". – ShadowRanger Nov 8 '18 at 15:48
  • 1
    @ShadowRanger - all the Republicans who lost their seats 2 days ago because they gave up extra couple % of their gerrymandered districts to fellow districts send their regards and regrets. – user4012 Nov 8 '18 at 18:45
10

Right off the bat, it's important to note that since 1931, Democrats have won 31 of the last 41 House elections, historically by wide margins. So if they are playing with a disadvantage, they handle themselves pretty well.

It is true that people representing urban areas tend to get comically large margins of victory. But Republicans in rural areas tend to win by large margins as well. By way of comparison: Shiela Jackson Lee, the Democrat representing Houston, won by 72% of the vote this year. But Cynthia Lumis, the Republican representing the entire state of Wyoming, won with 68%.

As to the second part of your question, you're describing gerrymandering, a very real part of American politics. Technically, gerrymandering IS illegal. But it's damn hard to prove in court.

But gerrymandering is a bipartisan phenomenon. A Washington Post analysis this year found that the two most gerrymandered states were Maryland and North Carolina, one controlled by Republicans and the other by Democrats. Of the ten most gerrymandered states, Republicans controlled six.

Does it give Republicans an advantage? Signs point to yes... right now. Republicans had a wave election in 2010, which swept them into power in several state legislatures right when redistricting was set to occur. But if Democrats gain those legislatures back in 2020, expect retribution.

  • I think it's also worth pointing out that there's some natural gerrymandering going on: because democrats congregate in densely populated urban areas, that makes it such that districts naturally favor republicans a little. – Avi Nov 14 '14 at 5:23
  • Oh, for sure. Saying that Democrats suffer a disadvantage because their base voters congregate in certain areas says more about their electoral strategy than the "fairness" of a system of representation laid down 250 years ago. – TenthJustice Nov 14 '14 at 15:22
  • 5
    Eh, I think it's not worth getting hung up on the issue of "fairness" and just talk about how the system behaves. If you let that distract you it will be to the detriment of your answer. – Avi Nov 14 '14 at 16:24
  • One measure of gerrymandering could be ability to win on representative count even if losing total national popular vote. That happened (republicans won) in presidential elections in 2000, and 2012 House elections. – Peter M. Nov 18 '14 at 14:20
  • If districts were required to be divided by total population in them, rather than total area or geopolitical boundries, that would negate the differences between densely packed areas and sparsely settled ones - a city would just have far more districts than the rural areas. Of course, then you get legislatures dominated by city-dwellers, which might or might not be a new "fairness" problem. – Bobson Dec 17 '14 at 21:18
6

The Daily Kos reports that Donald Trump and a Republican Representative won 218 congressional districts in 2016. That's an absolute majority.

Trump-Republican   218
Trump-Democrat     12
Clinton-Republican 23
Clinton-Democrat   182

So we can see that Trump received a plurality of the vote in 230 congressional districts while the Republican Representatives won 241 congressional districts. Meanwhile, Hillary Clinton won a plurality of the national vote. Trump also won a majority of the states and counties. So Trump clearly outperformed his national vote share in geographic regions.

The Brookings Institution found that Democrats won a lower share of House seats than their national vote share in 2016. This was one of eight elections since 1946 when this happened. In the rest, it was the Democrats who received a higher share of seats than votes. And Democrats historically received a larger bonus than Republicans ever have.

Specifically in 2016, Democrats won about 47.3% of the national vote in House elections but only 44.8% of the seats.

TL;DR: historically it's been Democrats with the geographic advantage, but in 2010, 2012, 2014, and 2016, it was the Republicans.

Some blame this on gerrymandering. However, the most packed (partisan) districts in the country are not gerrymandered (at least not by Republicans). The ten most partisan districts are mostly located in New York City and California, where Republicans do not control redistricting. Only two of the ten are in states where Republicans redistricted after 2010: Pennsylvania and Florida. In all ten cases, the districts were located in major cities with enough population for their own congressional districts: New York City (5); Los Angeles; San Francisco; Chicago; Philadelphia; and Miami.

A larger problem is that Republicans have been outperforming Democrats in the districts that are not gerrymandered. For example, Trump won thirty states when Obama only won twenty-eight and twenty-six. Bush won at least thirty both times. Trump won about 80% of the counties. Republicans won 52% of the Senate seats, even though the Democrats had good years in 2012 and 2006. Republicans won 66% of the seats in 2016 and 2014.

3

In addition to Gerrymandering, there is the issue per-capita representation differences in the House of Representatives.

The variance ranges from roughly half-million constituents per representative (R.I.) to 1 million constituents per representative (MT): http://en.wikipedia.org/wiki/List_of_United_States_congressional_districts#Extremes

This interactive map shows you the breakdown by state: http://www.datamasher.org/mash-ups/people-representative

Does it favor a particular party? I don't know. But is another variable to consider.

At the state government level, the differences can be a bit more drastic between state:

http://ballotpedia.org/Population_represented_by_state_legislators

That of course doesn't have a direct correlation with the federal level, but there's likely some indirect correlations.

  • 1
    That's state legislators, not members of Congress. For members of Congress, the extremes are Montana with around a million per district, and Rhode Island with about 530,000 per district (Montana's the biggest state with 1 district, RI the smallest with 2) – cpast Nov 14 '14 at 23:20
  • @cpast ack! Yes, you're absolutely correct! Point still stands, but less so. I'll update the answer to fix that later! (or feel free to add that directly if you wish). – user1530 Nov 14 '14 at 23:56
2

The Republicans have an advantage insofar they control a large number of "small" (by population) Western states (the two Dakotas come to mind), that have only one Congressman based on population, but two Senators, hence three electoral votes. This compares to a large Democratic state like California, with 52 Congressmen, whose two Senators bring the electoral vote total to 54, or 4 percent (not three times) more than the number of Congressmen.

Whether or not that is "unfair" is another subject.

1

Summary: It's complicated. Results from the 2012 and 2006 elections show that, statistically, Democrats should have won all the seats, but results from the 2000 and 1994 elections show the Republicans should have won most (in 2000) or all (in 1994) of the seats. The problem cannot be resolved via simple statistical analysis (below).

According to https://en.wikipedia.org/wiki/United_States_House_of_Representatives_elections,_2012#Results_summary, of the 122,346,020 votes cast total, Democrats received 59,645,531 or 48.8% of the vote, Republicans received 58,228,253 or 47.6% of the vote, and the remaining 4,472,236 or 3.7% of the votes went to other candidates (%ages don't add to 100% due to rounding).

If Democrats received 48.8% of the vote in every district, and Republicans received 47.6% of the vote in every district, the Democrats would've won every seat.

However, probability doesn't work that way. If you flip a fair coin 100 times, it's highly unlikely to get exactly 50 heads: there's a 95% chance you'll get between 40 and 60 heads, a 99.5% chance you'll get between 30 and 70 heads, but there's even a small chance (0.5%) that you'll get fewer than 30 or more than 70 heads.

Let's apply this methodology (called the "binary distribution") to the problem at hand.

Because the binary distribution only applies when there are two choices, let's consider only major party (Democratic and Republican) votes. In this scenario:

  • There are a total of 117,873,784 cast for major parties.

  • Republicans received 58,228,253 or 49.40% of these votes

  • Democrats received the remaining 59,645,531 or 50.60% of these votes.

Let's now consider the House district with the fewest number of votes. According to http://uselectionatlas.org/FORUM/index.php?topic=255728.0 and backed up by http://history.house.gov/Institution/Election-Statistics/2016election/, this appears to be Texas's 33rd district with 126,369 votes.

NOTE: I didn't look at every single number if the PDF, nor did I confirm the validity of https://docs.google.com/spreadsheets/d/1oArjXSYeg40u4qQRR93qveN2N1UELQ6v04_mamrKg9g/edit#gid=0, but the argument below works even if the number is relatively close.

With 126,369 votes, you need 63,185 votes to win.

If the districts were evenly populated with Democratic and Republican voters nationwide, you'd expect 49.40% or ~62,425 voters in this district to vote Republican. Of course, the chances of that happening exactly is very low, so we also calculate the standard deviation which happens to be 125 in this case.

What are the chances of a Republican win? Since 63,185 is 760 or about 6 standard deviations above the average, the chances of a Republican victory are about one in a billion.

The chances are even smaller in districts with more voters. With roughly even distribution of Democrats and Republicans, the chance of Republicans winning even one seat are less than 1 in 2.3 million.

The results of the 2006 election are even more convincing: https://en.wikipedia.org/wiki/United_States_House_of_Representatives_elections,_2006, since the Democrats won by a larger margin.

However, the results of the 2000 election (https://en.wikipedia.org/wiki/United_States_House_of_Representatives_elections,_2000) tell a different story. Without going through the gory math (available at https://github.com/barrycarter/bcapps/blob/master/STACK/bc-geovantage.m), the Democrats had only a 1.3% chance of winning a given district, so their winning 212 seats would be next to impossible, statistically speaking.

Going back even further: statistically, the Democrats had virtually no chance of winning a single seat in the 1994 election (https://en.wikipedia.org/wiki/United_States_House_of_Representatives_elections,_1994), but actually won 204 seats.

Ultimately, the problem goes deeper than statistics. You could run the numbers on a state-by-state basis (allowing for state bias towards a candidate), but there's no good reason to believe states are a fundamental unit of bias: bias could occur at the county level, urban vs suburban level, etc.

  • I don't think this directly answers the question of geography...which does play into the house. Counties and cities are somewhat irrelevant as votes are based on districts. – user1530 May 1 '17 at 1:50
  • NOTE: for some reason, I thought House elections were only once every 6 years; my answer is still correct, but I'll try to improve it using the 2014 elections, which provide a more recent contrast. @blip I agree it doesn't really answer the question, but perhaps explains why the question is difficult. – barrycarter May 1 '17 at 3:04
  • Re "Since 63,185 is 760 or about 6 standard deviations above the average, the chances of a Republican victory are about one in a billion.": sorry, but it's not obvious how this result was arrived at. Please include more of the intermediate steps of this calculation. – agc Aug 18 '18 at 19:23
  • 1
    @agc I really need to improve my answer, and you are correct in saying I made a mistake in the math, although the general point remains the same. In the binomial distribution, the variance is n*p*(1-p), which is 126369*.4940*(1-.4940) or about 31587.7 in this case. The standard deviation is the square root of the variance, which is about 177.74 in this case. I mistakenly used 63185 as the population size. – barrycarter Aug 23 '18 at 15:23
1

This question popped up at the top again and it reminded me of this recent article:

http://www.governing.com/topics/politics/gov-democrats-urban-geography-demographics.html

While it doesn't specifically talk about the Federal House of Rep's, it does go into the broad geographic issue that democrats are facing: Over the past several decades, rural democrats have moved to the cities.

This has left very high concentrations of democrats in urban areas. This is where the geographic disadvantage happens. Since representatives not only represent people--but geography, vasts swaths of our country are in very rural, mostly purely red areas. These areas also have the lowest concentration of population, so there is a geographic discrepancy here in that mathematically, we can have more republican representatives even though there are fewer overall constituents that they are representing.

-1

The Voting Rights Act is a major factor in this issue. In many states and counties, the federal government has required that a quota of districts be set up that are easily winnable by blacks, or by hispanics, or by a coalition of the two. This quota makes it much harder for white Democrats to be elected.

Gerrymandering is most glaring when it turns an evenly split populace into districts that result in a legislature that is very unbalanced. The opposite of gerrymandering is most glaring when it deprives a uniformly distributed minority party of any legislative seats. The Voting Rights Act seeks to prevent the latter phenomenon from happening to "protected" minorities.

Take a hypothetical jurisdiction. Suppose:

  • 30 % are in minorities "protected" by the Voting Rights Act.
  • 70 % are not in these minorities.

Suppose that jurisdiction-wide elections are competitive -- Close to a 50 / 50 split between Democrats and Republicans.

For many years, Democrats have won about 60%-70% of Hispanic votes, and 85-95% of black votes. Suppose that in our hypothetical jurisdiction, the "protected" minority vote is split 25 - 5. For the jurisdiction-wide elections to be competitive, this implies that the remaining vote is split 25 - 45.

As implemented (though not necessarily as written), the Voting Rights Act causes about 30% of the legislative seats to be easily won by protected-minority Democrats. Notice that a disproportionate number of Democrats are in these districts. This means that unless the remaining Democrats concentrate in enclaves, non-minority Democrats are both unlikely to win in the minority districts (by design), and are unlikely to win in the non-minority districts (because they are a spread-out political minority).

For two generations, ambitious white politicians (in jurisdictions subject to Voting Rights Act quotas) have been noticing the handicap that white Democrats face in getting elected to legislative positions. There has been a tendency (once Republicans become a majority of the white population) for white politicians to switch from the Democratic party to the Republican party. This tendency causes affected states to "snap" from having legislatures that are dominated by Democrats (with the minority seats plus about half of the non-minority seats) to being dominated by Republicans (with most of the non-minority seats).

  • The last paragraph is an interesting theory but is there any data to back up that theory as being a broad trend as opposed to 1-2 random figures? – user4012 May 5 '17 at 13:36
  • Please cite this "quota" you reference. – user1530 Jun 21 '17 at 4:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.