Virginia Governor’s Race a Clean Sweep for Warner, but not for Democrats

--Mark Rush, Washington and Lee University

(Stay tuned—subject to revision and update. Last update, 14 November 2001)

Mark Warner’s victory represents a striking reversal of partisan trends throughout the 1990’s that had favored the Republicans. Warner gained an average of 10.9 percentage points across all counties and cities. The Democrats lost ground in no county or city—at the gubernatorial level.

Warner made significant gains not only in areas of Democratic strength, but also in areas where outgoing Governor Gilmore and the Republicans had dominated four years ago (such as Augusta, Rockingham and Shenandoah Counties).

This year’s vote indicates a major swing towards the Democrats statewide. The color-coding of both maps indicates Democratic strength in yellow, Republican strength in dark blue.

Clearly, the election represents a rising tide of support for the Democrats. The following graph indicates the scope of the Democratic Party’s gains. Notice that the party gained support in every part of the state. Warner’s message struck a popular note with all voters.

Turnout Trends

At this point (until the results are certified), only a rough, aggregate analysis is possible. But there are some intriguing (and somewhat disturbing) trends in Virginia voter turnout. While voter registration has skyrocketed throughout the last decade, turnout has not done so. That is, while the number of people registering to vote has increased (due, no doubt to the motor voter law), the ease of registration has not lead to an increase in people actually voting.

In 2001, the dramatic increase in voter registration continued. But, the actual number of voters continued its less dramatic, but steady decrease. The following graph indicates the actual number of voters who came out to vote since 1992. Turnout in gubernatorial and presidential races is more or less constant, but off-year congressional and state legislative races manifest a clear downward trend.

The downward trend becomes more pronounced when we look at actual turnout percentages. While the raw number of voters has not dropped all that much, the number of voters as a percentage of registrants clearly has dropped across all legislative levels.

Turnout rates varied markedly across the state’s political subdivisions. Yet, there was no pattern between the turnout rate and the partisan split. This, again, confirms Warner’s broad appeal.

The turnout patterns did vary between the 1997 and 2001 Gubernatorial races, but this was, again, due in large part to the increased number of registered voters. Thus, while the average turnout rate across the state’s political subdivisions dropped, the graphs above indicate that the actual raw number of voters participating was more or less comparable between the two elections.

In general, there was no relationship between the turnout rate or the actual increase in raw numbers of voters and either candidate’s success. This confirms the patterns above that demonstrate the statewide appeal of Warner’s candidacy

Turnout patterns: House of Delegates Races

There really were two elections this year, and the voting and turnout patterns manifest this clearly.

The House of Delegates election was drastically affected by the newly-drawn legislative districts. Strong Democratic incumbents such as Dick Cranwell were forced to with draw from seeking re-election because their districts were drawn in a manner that would have forced them to challenge fellow Democratic incumbents in a primary or to run as an independent.

The overall field contrasted starkly with the field in 1999. This year, 23 Republican and 14 Democratic incumbents were unchallenged. As well, in two open seats, two new Republican candidates ran unopposed. This created an odd and perhaps disturbing dynamic: before any votes were cast, 40 of the 100 seats in the House of Delegates had been won!

This actually was an improvement over 1999, where half of the seats were uncontested.

This may actually be due, in part, to the fact that redistricting does “destabilize” electoral districts. That is, in 1999, the district had been in place for 5 electoral cycles. Incumbents were familiar with and in their district constituencies and correspondingly harder to challenge.

Redistricting alters district lines and forces incumbent candidates to appeal to new and possibly unfamiliar voters. This creates an incentive for potential challengers to run against the (relatively) less secure incumbents.

This is reflected somewhat in the turnout figures for the 1999 and 2001 House of Delegates elections. As the graphs above indicate, while registration has increased, the turnout rate has actually decreased over the last decade. However, the pattern of turnout is by no means uniform across the state.

While 2001 turnout was hardly very high (it ranged from 23.3% in the 3^rd to 55.6% in the 68^th District), it is important to note that average turnout was 10 points higher in contested districts. Where races were contested, turnout averaged 44.6%. In uncontested districts, it averaged 34.5%.

In 1999, the turnout pattern was even more disturbing. Absent the gubernatorial statewide race to increase interest, overall turnout was much lower. In contested races, it averaged 36.7% while in uncontested districts it averaged only 26.7%!

What is especially disturbing about these data is the extent to which the “price” of a house seat (measured in votes necessary to win) varies from one district to the next.

The US Supreme Court has ruled that legislative districts must have approximately the same population. The Court permits some deviation from strict population equality to allow states to avoid splitting municipalities and communities of interest.

So, while all districts have roughly the same population (the average district population after the 2001 census was 70,785 and the average deviation from that was about 1400 voters or less—see the website at the State Division of Legislative Services: http://dlsgis.state.va.us/House/EnrolledHB1/EnrolledHB1.htm), the actual number of voters that voted in each district varied remarkably.

The following maps and graphs indicate the extent to which variance in turnout actually results in the “cheapening” of legislative seats. In 1999, the average number of votes cast in each house district was 12186. Yet, in some districts, as few as 2849 voters bothered to vote. The following table shows the three districts with the lowest number of voters (who actually voted) in 1999—they were all uncontested seats.

Cheap Seats-1999

District	Total Votes Cast	Registered Voters	Turnout%
69	2849	28638	9.96
70	3576	26326	13.6
71	4132	30218	13.7

In contrast, the districts with the most voters turning out were, not surprisingly, all contested.

Expensive Seats—1999

District	Total Votes Cast	Registered Voters	Turnout%
55	20324	50666	40.1
58	20451	46068	44.4
14	21206	41474	51.1

The correlation between contested seats and turnout is not exactly perfect. But, nonetheless, this presents a disturbing problem: Some seats in the house are cheaper than others. To win in the 69^th district, a candidate needed only 1450 votes. To win the 14^th, a candidate needed 10,603!

This same pattern is evident in the 2001 election, even though turnout was higher statewide.

Cheap Seats—2001

District	Total Votes Cast	Registered Voters	Turnout%
100	8545	30537	28.0
50	9190	36103	25.5
3	10154	43475	23.4

Expensive Seats—2001

District	Total Votes Cast	Registered Voters	Turnout%
8	24982	46829	53.4
65	25157	47665	52.8
68	26236	47164	55.6

In sum, while redistricting made a difference regarding the caliber or, perhaps, experience of the candidates running (because some of the more formidable Democratic incumbents lost their districts), it important to note that redistricting notwithstanding, there is a great variation in the nature of the campaigns that must be conducted from one district to the next. The nature and cost of an election in which 3000 votes are at stake is radically different than one in which 25000 are at stake.

More Analysis to Come…