We are less than a year and a half from the Census date of April 1, 2010 and with the Census comes the decennial reapportionment of seats of the House of Representatives among the several States. This event affects more than just how many seats each state gets in the House, but also determines how many electoral votes each state will have in the 2012, 2016 and 2020 Presidential elections. The importance of the apportionment should therefore be self-evident, however it is one of those things that is seemingly taken for granted without much fuss, at least not for nearly 100 years; however, the parameters of reapportionment ought to be reexamined, especially the size of the House, which, despite a more than tripling of the U.S. population, has remained fixed since 1910.
First I'm going to give a brief history of apportionment before comparing our ratio of population per representative to other nations of the world. I'll then delve into some analysis of how apportionment has effected recent Presidential elections. Finally I'll provide a preview of what the 2010 reapportionment have have in store and which states will be the winners and losers in the process. If I get a bit too technical or overwhelm you with the numbers and charts, my apologies ahead of time, I am a Math geek.
I. The History of Apportionment
(if you're not interested in the history, feel free to jump to the meat of the diary starting in Part II)
The first apportionment of representatives among the States was dictated in the Constitution (Article I), however the document provided that every ten years a census be conducted and that the apportionment of Representatives thereafter would be based upon the census. The Constitution though is silent as to the method of apportionment and to the size of the House aside from a minimum population to Representative ratio of 30,000.
When the first apportionment based upon the first census was made, it was vetoed by George Washington, at Thomas Jefferson's urging, technically because it violated this quota, but it was also because Jefferson wanted his proposed method of apportionment rather than that proposed by Alexander Hamilton. In both methods, a standard divisor was computed by taking the total population divided by the number of seats in the House. The fractional remainders were discarded and each state assigned this lower quota value of seats. Jefferson and Hamilton differed in how they handled the situation if the sum of the lower quotas didn't equal the number of seats. Hamilton wants to assign the remaining seats to the states in descending order of the largest fractional remainders. Jefferson wanted to modify the standard divisor (adjusting it down) until the sum of the rounded down lower quotas equaled the number of seats. The House size was set at 105 seats.
Jefferson's method would continued to be used for the next four reapportionments with the House size being increased to 141 in 1800, 181 in 1810, 213 in 1820 and 240 in 1830. In 1840, however, a method first proposed by Daniel Webster in 1831 was adopted. In the Webster Method, like the Jefferson and Hamilton methods, a standard divisor was calculated (total population/# of House seats). Each state's population was then divided by this divisor. If the fractional remainder was ≥ 0.5, the state was assigned the upper quota (rounded up) and if it was < 0.5, the state was assigned its lower quota (rounded down). If the sum of the quotas didn't add up to the number of seats, the standard divisor was adjusted until they were equal. This was also the first and only time the size of the House was reduced dropping its membership to 223.
An argument developed over the 1850 apportionment with the Hamilton and Webster methods dividing the Congress deeply. The matter was only solved when the Census statistician found a size of the House (234 seats) in which the Hamilton and Webster methods both yielded the same results, but the Hamilton method was the method codified by the apportionment act. The Hamilton Method was used again in 1860 with 241 seats and 1870 (283 seats), but in both years the apportionment was later modified however the 1872 modification violated every apportionment method (Jefferson, Hamilton and Webster) and had enormous ramifications on the election of 1876 four years later. In 1876, New York Governor Samuel J. Tilden defeated Ohio Governor Rutherford B. Hayes...or at least he should have. Under the original 1870 apportionment, Tilden would have won the Electoral College regardless of apportionment method. If the 1872 modification of the apportionment adding 9 seats to the House was done in accordance with any of the three methods previously used, Tilden would have won. But the 9 seats that was added in 1872 were done haphazardly and to the benefit of states Hayes would subsequently win in the 1876 election. Hayes was elected by the electoral college 185 to 184 in 1876 after a contentious battle over the electoral votes of four states, a race he should have lost by 1 to 3 electoral votes.
The 1880 and 1890 apportionments were again done with the size of the House specifically chosen so the competing Webster and Hamilton Methods would be in agreement (325 seats in 1880 and 356 seats in 1890), but a paradox discovered in 1880 in Hamilton's method began to erode support for that method (see here for more info on the paradox). By 1900, Webster's Method had won out and the new House was increased to 386 members. In 1910, the House size was fixed to 433 members with two extra seats reserved for the soon to be admitted states of Arizona and New Mexico to give the House the 435 members it has today. In 1920, it was apparent that the nation was changing. The urban areas of the country were growing much faster and the rural House members saw they would lose clout under any reapportionment. They successfully blocked any apportionment from taking place and as a result, the 1910 apportionment remained in effect for another 10 years.
For the 1930 apportionment, Congress adopted an apportionment method that had been developed by Harvard Mathematics professor Edward Huntington and Census Bureau Statistician Joseph Hill in 1911. In this Huntington-Hill Method, also known as the Method of Equal Proportions, the initial quota created by the division of a state's population by the standard divisor is rounded up if the initial quota is greater than the geometric mean of the upper and lower quota (i.e. the square root of the product of the lower quota and upper quota) and rounded down if it is lower than the geometric mean. A 1940 law makes the reapportionment of seats automatic every decade and keeps the House fixed at 435 seats unless the admission of new states temporarily increases the number (as occurred from 1959-1961 due to the admission of Hawai'i and Alaska). The Huntington-Hill Method is still in place nearly 70 years later and the House is poised to remain at 435 seats for the tenth consecutive apportionment. Three challenges to the apportionment are notable however. In 1990, both Montana and Massachusetts challenged the apportionment in court after each barely lost a seat. Montana argued that the Huntington Hill method was unconstitutional and "suggested" another apportionment method that would "coincidentally" would have retained Montana's second house seat. Massachusetts on the other hand argued that the inclusion of overseas military personnel in each state's resident population had deprived it of its seat. Both lost their challenges. It was Utah's turn to challenged the apportionment in 2000 when it failed to gain a fourth House seat. Utah argued that since the Census counts overseas military personnel and their overseas citizen dependents, the census should also include overseas citizens, such as Mormon missionaries living abroad. An increase of just 857 people to Utah's population would have moved Utah ahead of North Carolina in the race for the 435th House seat. The Supreme Court ruled against Utah.
II. Comparing our Population per Representative Ratio
From the beginning of this country until 1910, the size of the House was, for the most part, increased in each apportionment to atleast partially keep up with increases in population and prevent existing states from losing representatives as newer states were added. However since 1910, the size of the House has remained fixed at 435 members. In the intervening time, the population of the United States has more than tripled, having a deteriorative effect on the ratio of population per representative. As more and more people are served by a single person, the ability of that person to in a sense be connected with his or her constituents diminishes. For a comparison, here is a look at the ratio over time:
The Least Square Regression slope of the pre-1910 part of the graph is 1570. The post-1910 slope is 5525! Maintaining a fixed House size has exploded the disparity as the population has increased.
Increasing the size of the House would help ameliorate this situation some, though the days when most people could know their member of Congress have probably long passed. Compared to most other large countries, including those democratic allies with whom the U.S. shares a sacred bond, the U.S. is woefully undemocratic in terms of how many people our Representatives represent. For comparison here is a look at other nation's ratio of population per elected representative in the lower chamber of their legislative body.
Only India has a worse ratio (and far worse at that) and we are just marginally better than Pakistan (though Pakistan's ratio is somewhat higher in the chart as I excluded seats not up for general election as about 70 are reserve for women and non-Muslims). Our allies in Europe, Canada and Australia has ratios one fifth or less that of the U.S. Granted, the U.S. is substantially more populous, however other large nations like China, Indonesia and Brazil have better ratios with large populations.
One idea that has been bounced around has been to index the size of the House to the size of the smallest state, the so called Wyoming Rule. Under the Constitution, each state is guaranteed at least one representative. In the last Apportionment, it was only because of this rule that Alaska, Vermont and Wyoming have a Representative and North Dakota barely made the cut coming in with the 434th seat if the rule weren't in place. If such a rule were in place for the 2000 reapportionment, the House would have been 570 members.
III. The Effect of Apportionment on the three most recent Presidential elections.
The 2000 Presidential Election represented one of the closest elections in American history based on the electoral vote count. Vice President Al Gore actually won the popular vote, but lost the electoral vote. Given Gore's greater vote popular count, there is a point at which increasing the size of the House would have given Gore a sufficient number of electoral votes to overcome the net advantage Bush had by winning more states (and thus more electoral votes owed by each state's Senate seats); however, the margin bounces back and forth for a while before the weight of population of the Gore states takes over. Ignoring the Maine and Nebraska proportional rules, here is a graph of the Bush-Gore margin of victory for House sizes from 400 to 700:
A House size of 596 is the last point at which Bush would have won and 655 is the last House size that would have resulted in a tie. Gore would have won with all House sizes greater than 655.
Under a similar analysis of 2004, permuting the size of the House John Kerry would have come no closer than 21 electoral votes to George Bush and the House would have to be reduced to 220 for that to have happened. In 2008, McCain would have come no closer than 27 electoral votes...with a House size of 50 (i.e. only one rep per state). If the Wyoming rule had been in effect for all three elections, the electoral margins would have been 325-324 Bush in 2000, 353-320 Bush in 2004 and 471-202 Obama in 2008.
At any rate, increasing the size of the House increases the chances that the winner of the popular vote will win the election since the apportionment of electoral votes will more closely align with the apportionment of the population among the states. Moreover, the advantage of this is that there is no constitutional amendment necessary. Reforming and/or getting rid of the Electoral College would take a constitutional amendment, however, increasing the size of the House thereby reducing the negative effects of the College such as
- Having states so small that they would not be entitled to having a Representative were it not for the constitutional rule and thereby depriving larger states of Representatives to which they would otherwise be entitled
- Having the strength of small states artificially inflating by their two extra votes by virtue of their U.S. Senate seats
While the smaller states will still have disproportionately more power in electing the President, the system will be more democratic in that the popular vote will be more important than it is now. How would this look in practice? Continue reading.
IV. A look at how the 2010 Apportionment might play out
To calculate how the House may be divvied up among the states in 2010, I ran the number based on the Census Bureau's 2007 state-by-state population estimates and their 2010 population projections. I have also included the Wyoming Rule based apportionment for the 2010 census projections. Here is the breakdown:
| 2000 | 2007 | 2010 | ∂ 2007 | ∂ 2010 | 2010 Wyo |
Alabama | 7 | 7 | 6 | 0 | -1 | 9 |
Alaska | 1 | 1 | 1 | 0 | 0 | 1 |
Arizona | 8 | 9 | 9 | 1 | 1 | 13 |
Arkansas | 4 | 4 | 4 | 0 | 0 | 6 |
California | 53 | 53 | 54 | 0 | 1 | 73 |
Colorado | 7 | 7 | 7 | 0 | 0 | 9 |
Connecticut | 5 | 5 | 5 | 0 | 0 | 7 |
Delaware | 1 | 1 | 1 | 0 | 0 | 2 |
Florida | 25 | 26 | 27 | 1 | 2 | 37 |
Georgia | 13 | 14 | 14 | 1 | 1 | 18 |
Hawaii | 2 | 2 | 2 | 0 | 0 | 3 |
Idaho | 2 | 2 | 2 | 0 | 0 | 3 |
Illinois | 19 | 19 | 18 | 0 | -1 | 25 |
Indiana | 9 | 9 | 9 | 0 | 0 | 12 |
Iowa | 5 | 4 | 4 | -1 | -1 | 6 |
Kansas | 4 | 4 | 4 | 0 | 0 | 5 |
Kentucky | 6 | 6 | 6 | 0 | 0 | 8 |
Louisiana | 7 | 6 | 7 | -1 | 0 | 9 |
Maine | 2 | 2 | 2 | 0 | 0 | 3 |
Maryland | 8 | 8 | 8 | 0 | 0 | 11 |
Massachusetts | 10 | 9 | 9 | -1 | -1 | 13 |
Michigan | 15 | 15 | 15 | 0 | 0 | 20 |
Minnesota | 8 | 8 | 8 | 0 | 0 | 10 |
Mississippi | 4 | 4 | 4 | 0 | 0 | 6 |
Missouri | 9 | 8 | 8 | -1 | -1 | 11 |
Montana | 1 | 1 | 1 | 0 | 0 | 2 |
Nebraska | 3 | 3 | 3 | 0 | 0 | 3 |
Nevada | 3 | 4 | 4 | 1 | 1 | 5 |
New Hampshire | 2 | 2 | 2 | 0 | 0 | 3 |
New Jersey | 13 | 13 | 13 | 0 | 0 | 17 |
New Mexico | 3 | 3 | 3 | 0 | 0 | 4 |
New York | 29 | 28 | 27 | -1 | -2 | 37 |
North Carolina | 13 | 13 | 13 | 0 | 0 | 18 |
North Dakota | 1 | 1 | 1 | 0 | 0 | 1 |
Ohio | 18 | 17 | 16 | -1 | -2 | 22 |
Oklahoma | 5 | 5 | 5 | 0 | 0 | 7 |
Oregon | 5 | 5 | 5 | 0 | 0 | 7 |
Pennsylvania | 19 | 18 | 18 | -1 | -1 | 24 |
Rhode Island | 2 | 2 | 2 | 0 | 0 | 2 |
South Carolina | 6 | 6 | 6 | 0 | 0 | 9 |
South Dakota | 1 | 1 | 1 | 0 | 0 | 2 |
Tennessee | 9 | 9 | 9 | 0 | 0 | 12 |
Texas | 32 | 34 | 35 | 2 | 3 | 47 |
Utah | 3 | 4 | 4 | 1 | 1 | 5 |
Vermont | 1 | 1 | 1 | 0 | 0 | 1 |
Virginia | 11 | 11 | 11 | 0 | 0 | 15 |
Washington | 9 | 9 | 9 | 0 | 0 | 13 |
West Virginia | 3 | 3 | 3 | 0 | 0 | 4 |
Wisconsin | 8 | 8 | 8 | 0 | 0 | 11 |
Wyoming | 1 | 1 | 1 | 0 | 0 | 1 |
Total | | 435 | 435 | | | 592 |
---|
Now I think ultimately the House should be expanded to the size that would be dictated by the Wyoming Rule as it has a basis in reality, insuring that each state is mathematically entitled to a representative. The current situation unfairly stunts the power of larger states reducing their power in a body that was intended to give power equally based on population. I can understand however that an increase of approximately 157 U.S. Representatives overnight would be a bit too dramatic. I do think that there is a good solution to more slowly grow the House without being too arbitrary in how much growth takes place and when. Let the numbers decide the growth. An in between solution is a no-loss rule. In 2010, the size of the House ought to be increased to the point that no state losses a representative in Congress, but other states get their proper allocation based on that House size. Again using the 2010 Census projections, the new House would be 467 members. The 467th seat goes to Ohio, its 18th seat, same as it has now. All states would either hold their current number of seats or gain seats. This process could be repeated again in 2020 and so on until the House has grown to the level the Wyoming rule would dictate. From then on that rule could control the size of the House. A breakdown of the 2010 House using this no-loss rule idea would be as follows:
| 2000 | 2010 no-loss | ∂ 2010 |
Alabama | 7 | 7 | 0 |
Alaska | 1 | 1 | 0 |
Arizona | 8 | 10 | 2 |
Arkansas | 4 | 4 | 0 |
California | 53 | 58 | 5 |
Colorado | 7 | 7 | 0 |
Connecticut | 5 | 5 | 0 |
Delaware | 1 | 1 | 0 |
Florida | 25 | 29 | 4 |
Georgia | 13 | 14 | 1 |
Hawaii | 2 | 2 | 0 |
Idaho | 2 | 2 | 0 |
Illinois | 19 | 20 | 1 |
Indiana | 9 | 10 | 1 |
Iowa | 5 | 5 | 0 |
Kansas | 4 | 4 | 0 |
Kentucky | 6 | 6 | 0 |
Louisiana | 7 | 7 | 0 |
Maine | 2 | 2 | 0 |
Maryland | 8 | 9 | 1 |
Massachusetts | 10 | 10 | 0 |
Michigan | 15 | 16 | 1 |
Minnesota | 8 | 8 | 0 |
Mississippi | 4 | 5 | 1 |
Missouri | 9 | 9 | 0 |
Montana | 1 | 2 | 1 |
Nebraska | 3 | 3 | 0 |
Nevada | 3 | 4 | 1 |
New Hampshire | 2 | 2 | 0 |
New Jersey | 13 | 14 | 1 |
New Mexico | 3 | 3 | 0 |
New York | 29 | 29 | 0 |
North Carolina | 13 | 14 | 1 |
North Dakota | 1 | 1 | 0 |
Ohio | 18 | 18 | 0 |
Oklahoma | 5 | 5 | 0 |
Oregon | 5 | 6 | 1 |
Pennsylvania | 19 | 19 | 0 |
Rhode Island | 2 | 2 | 0 |
South Carolina | 6 | 7 | 1 |
South Dakota | 1 | 1 | 0 |
Tennessee | 9 | 9 | 0 |
Texas | 32 | 37 | 5 |
Utah | 3 | 4 | 1 |
Vermont | 1 | 1 | 0 |
Virginia | 11 | 12 | 1 |
Washington | 9 | 10 | 1 |
West Virginia | 3 | 3 | 0 |
Wisconsin | 8 | 9 | 1 |
Wyoming | 1 | 1 | 0 |
| | | |
| 435 | 467 | 32 |
---|
Increasing the size of the House will allow and increase the chance for a greater diversity of members, more African-Americans, Hispanics, Asian-Americans, Native Americans, GLBTs, Jews, Muslims, etc. It will help make the House look more like America and that most certainly is not a bad thing.