Around the time of the 2020 election, there was some discussion surrounding non-voters; FiveThirtyEight did a profile of non-voters leading up to the election and after the election NPR conducted a poll of non-voters with IPSOS. But I would argue that these are the exceptions, the bulk of election coverage in this country largely focuses on voters; who they are, what issues matter to them, which candidate do they support, and on and on. When discussing likely voters you might hear pundits say that voters are complex, their reasoning and motivations are complex, and that voting cohorts are rarely monolithic. But this same attention and analysis is rarely given to the non-voters.
I understand why this is the case, it is voters that select the politicians, not the non-voters. But non-voters are often labeled as apathetic and the conversation moves on. It is interesting that these non-voters are not also assumed to be complex in their reasoning and motivation for not voting, or that they are also not a monolith. Not looking at who non-voters are and why they are choosing to not vote misses important dynamics of our political system.
So, let’s take a look at why citizens are deciding not to vote. While the Federal Elections Commission does not include non-voters in their elections data at all, the Census Bureau not only tracks the total population, the citizen population, registered voters, and non-voters; but also surveys non-voters’ reasoning for being non-voters.
But before we dig into that data we should set our baseline. Depending on how you break down the figures, since 2000 on average anywhere for 13% to 43% of the US population is not participating in the Presidential electoral process, as shown in the table below.
|
TOTAL PRESIDENTIAL POPULAR VOTE |
TOTAL POP 18+ |
TOTAL POP CITIZEN |
REPORTED REGISTERED |
||||
|
2000 |
105,405 |
202,609 |
47.98% |
186,366 |
43.44% |
129,549 |
18.64% |
|
2004 |
122,295 |
215,694 |
43.30% |
197,006 |
37.92% |
142,070 |
13.92% |
|
2008 |
131,314 |
225,499 |
41.77% |
206,072 |
36.28% |
146,311 |
10.25% |
|
2012 |
129,085 |
235,248 |
45.13% |
215,081 |
39.98% |
153,157 |
15.72% |
|
2016 |
136,669 |
245,502 |
44.33% |
224,059 |
39.00% |
157,596 |
13.28% |
|
2020 |
158,383 |
252,274 |
37.22% |
231,593 |
31.61% |
168,308 |
5.90% |
|
AVG |
|
|
43.29% |
|
38.04% |
|
12.95% |
When calculating the percentage of reported non-voters the Census Bureau uses voting aged citizens as their base population, which I think is the most reasonable population to use as this population consists of all eligible voters (and as we will see voter registration can be a reason for not voting). With that, anywhere from 32% to 43% of the voting eligible population is not participating in elections.
Of this roughly 38% of the citizen population the Census Bureau has identified 12 standard responses why people do not vote, and one unique to 2020—spoiler, the reason is COVID. The table below provides the percentage breakout of each of these by year:
|
|
2000 |
2004 |
2008 |
2012 |
2016 |
2020 |
|
Illness or disability (own or family's) |
14.77% |
15.40% |
14.90% |
14.00% |
11.70% |
13.00% |
|
Out of town or away from home |
10.23% |
9.00% |
8.80% |
8.60% |
7.90% |
6.10% |
|
Forgot to vote (or send in absentee ballot) |
3.96% |
3.40% |
2.60% |
3.90% |
3.00% |
3.70% |
|
Not interested, felt vote would not make a difference |
12.24% |
10.70% |
13.40% |
15.70% |
15.40% |
17.60% |
|
Too busy, conflicting schedule |
20.92% |
19.90% |
17.50% |
18.90% |
14.30% |
13.10% |
|
Transportation problems |
2.37% |
2.10% |
2.60% |
3.30% |
2.60% |
2.40% |
|
Did not like candidates or campaign issues |
7.71% |
9.90% |
12.90% |
12.70% |
24.80% |
14.50% |
|
Registration problems |
6.86% |
6.80% |
6.00% |
5.50% |
4.40% |
4.90% |
|
Bad weather conditions |
0.64% |
0.50% |
0.20% |
0.80% |
0.00% |
0.10% |
|
Inconvenient polling place or hours or lines too long |
2.61% |
3.00% |
2.70% |
2.70% |
2.10% |
2.60% |
|
Other reason, not specified |
10.15% |
10.90% |
11.30% |
11.10% |
11.10% |
14.50% |
|
Refused or don't know |
7.54% |
8.50% |
7.00% |
3.00% |
2.70% |
3.20% |
|
Concerns about COVID-19 |
N/A |
N/A |
N/A |
N/A |
N/A |
4.30% |
Of these I have grouped them into four broad categories:
b. Not interested, felt vote would not make a difference
c. Did not like candidates or campaign issues
b. Too busy, conflicting schedule
c. Transportation problems
d. Bad weather conditions
b. Inconvenient polling place or hours or lines too long
b. Other reason, not specified
c. Refused or don’t know
d. Concerns about COVID-19
There could be some disagreement about these groupings; is someone who merely forgot to vote disengaged in the same way that someone who does not like any of the candidates running or issues on the ballot? Does it make sense to consider access issues separately from voter restrictions? These are open questions, but, as we will later see, there is broad overlap in the potential solutions irrespective of this grouping. The table below provides the percentage response by these broader groupings.
|
|
Disengaged |
Access Issues |
Voter Restrictions |
Other |
|
2000 |
23.91% |
38.69% |
9.47% |
27.93% |
|
2004 |
24.00% |
37.90% |
9.80% |
28.40% |
|
2008 |
28.90% |
35.20% |
8.70% |
27.10% |
|
2012 |
32.30% |
37.00% |
8.20% |
22.70% |
|
2016 |
43.20% |
28.60% |
6.50% |
21.70% |
|
2020 |
35.80% |
28.60% |
7.50% |
28.10% |
|
AVG |
31.35% |
34.33% |
8.36% |
25.99% |
Access Issues:
As the table above shows, access issues are, on average, the largest impediment to citizens voting. Improving accessibility should be relatively easy to address. Illness or disability (own or family’s), transportation problems, and bad weather conditions could all be solved by increasing the availability of mail-in voting. As 2020 showed, with the expansion of mail-in voting across many states in response to the COVID-19 pandemic, we are able to securely and effectively conduct large scale elections through the mail. As for citizens that are too busy or have conflicting schedules, extending early voting, increasing number of polling locations and polling hours, and making election day a federal holiday all would provide would-be voters additional opportunities to vote.
Now, given Republican legislatures across the country are adopting measures to roll-back the COVID-19 voting measure put in place for the 2020 elections , and their resistance to recognizing election day as a holiday, I think one could reasonably argue these should be grouped with voter restrictions.
One final, and very important point on voter access; there is little evidence that increased voter access benefits one party over another, so there should not be partisan motivations against increasing voting access.
Voter Restrictions:
Given the increased focus on voter restrictions the results of respondents to the Census’ survey were, I thought, surprisingly low at 8% on average. Further, the share of voter restrictions as a reason for not voting has been on a downward trajectory since 2000. That is not to minimize respondents' claims of voter restrictions, this number should be near zero. But because this figure was low I looked at this in a slightly different way.
The table below shows the percentage of the total citizen population, reported registered citizens, and reported not registered citizens in columns 2 through 7. Columns 8 and 9 show the number of people who reported registration problems and those individuals as a percentage of not registered voters. According to this, on average, only 9% of non-registered voters are unregistered because of registration problems. Again, this is a bit surprising to me. It may be the case that the number self-reported registration issues are too low as states conduct purges of the voter rolls resulting in a citizen being unknowingly removed from the voter rolls. It is also possible that the underlying causality for citizens not voting is not appropriately self-reported in the Census survey. By this I mean, a citizen could respond that they are too busy to vote, but does that also mean they are too busy to have registered to vote which could be a registration problem. Ultimately, I am not in a position to answer this with the data available here, so we will continue to assume respondents provided the most accurate and honest answers.
|
|
TOTAL POP CITIZEN |
Reported registered |
Reported not registered |
No response to registration |
Registration problems |
||||
|
|
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
|
2000 |
186,366 |
129,549 |
69.51% |
56,817 |
30.49% |
- |
0.00% |
6,294 |
11.08% |
|
2004 |
197,006 |
142,070 |
72.11% |
54,936 |
27.89% |
- |
0.00% |
6,117 |
11.14% |
|
2008 |
206,072 |
146,311 |
71.00% |
59,761 |
29.00% |
- |
0.00% |
4,496 |
7.52% |
|
2012 |
215,081 |
153,157 |
71.21% |
33,134 |
15.41% |
28,790 |
13.39% |
2,999 |
9.05% |
|
2016 |
224,059 |
157,596 |
70.34% |
32,622 |
14.56% |
33,841 |
15.10% |
2,370 |
7.26% |
|
2020 |
231,593 |
168,308 |
72.67% |
25,782 |
11.13% |
37,503 |
16.19% |
1,987 |
7.71% |
However, in actuality, the voting restriction numbers are much higher. If we were to include all citizens who cannot vote in presidential elections and either do not have representatives or have non-voting representatives in Congress—i.e. those citizens of DC, American Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the US Virgin Islands.
Increasing the number of polling locations, and extending polling location hours and the early voting period, mentioned above, would also address the restrictions caused by inconvenient polling locations and hours. However, for registration problems, a civics class added to public school senior curriculum that includes students registering to vote. This would not only ensure citizens are being registered to vote, but that they are beginning to learn and engage with our political and electoral systems at a young age.
Disengaged Voters:
Turning now, to our last grouping, disengaged voters. Based on my grouping of the Census data this grouping is the second largest reason for not voting in elections, at an average of 31% of non-voters responding that they were not engaging with the political process because they either “forgot to vote (or send in absentee ballot)”, “were not interested, felt vote would not make a difference”, or “did not like candidates or campaign issues”.
The voter that forgets to vote is probably truly not engaged and is never going to be. At an average 3% of non-voters and ranging from 2.6% to 3.9%, this seems to be a relatively stable population. Give that we shouldn’t expect 100% participation I think this is a reasonable percentage of truly disengaged voters. The other two response groups, however, I would like to take a closer look at.
If you have read my other essays you may be familiar that I believe voter engagement is fairly directly related to representation; if you increase representation through the viability of alternative parties more people will participate in the electoral system. I understand that “not interested” is half of the response “were not interested, felt vote would not make a difference”, but for my purposes here I am going to place more weight on the “felt vote would not make a difference” portion of the response. Votes cast for third-parties are often seen as throw-away votes, or spoilers for either of the two major parties. This could discourage would be voters. As such, in the table below I have grouped these respondents with votes cast for a third-party. Note, that these will be taken as a percentage of votes cast in the respective years’ Presidential general election, not of the total citizen population.
|
|
UNINTERESTED / VOTE DOES NOT MATTER |
THIRD PARTY |
DID NOT LIKE CANDIDATES / ISSUES |
NONE OF THESE CANDIDATES |
TOTAL |
|
2000 |
5.55% |
2.05% |
3.49% |
0.00% |
11.09% |
|
2004 |
4.46% |
0.58% |
4.13% |
0.00% |
9.18% |
|
2008 |
4.87% |
0.90% |
4.69% |
0.00% |
10.47% |
|
2012 |
3.98% |
1.07% |
3.22% |
0.00% |
8.27% |
|
2016 |
3.70% |
3.52% |
5.96% |
0.01% |
13.19% |
|
2020 |
3.08% |
1.22% |
2.54% |
0.01% |
6.84% |
If these groups were to decide to vote for a third-party instead of not voting, there would be a statistically significant portion of the voting population not voting for one of the two major parties. Almost certainly, this wouldn’t result in any election victories for third-party candidates, but it would make it harder disillusioned voters to be ignored. They would not even have to vote for the same third-party candidates for this to register as significant!
Another item of note in the table above, the “none of these candidates” column; this appears to not be a significant column, as it, at most, represents 0.01%. However, the option to cast a ballot for “none of these candidates” is only available in Nevada. In that state this response averages 1% of the total vote. If this was a ballot option in all states, and all non-voters who do not like the candidates or issues voter for "none of these candidates" then this too would be more statistically significant. Together, votes cast for a candidate other than the two major parties would be, on average, 10% of all votes cast.
Potential Solutions:
I have already mentioned a few above; increasing mail-in voting, increasing the number of polling locations, extending polling location hours, extending the early voting period, making election day a federal holiday, and creating a civics class in our public high schools to help students register to vote as they turn 18 years old. However, to unequivocally state what I had alluded to within my last section, individuals will not participate in politics unless they see themselves represented by political leaders. Take this however you like; increased diversity of political ideologies represented in Congress or increased minority representation—every American deserves the opportunity to vote for a candidate that represents them, in whatever way is important to them. Personally, I think the best way to do this is to adopt a proportional electoral system for the House of Representatives and Electoral College and adopt ranked choice voting for Senate candidates. It is also necessary that we extend citizens of DC full representation within the House and Senate, and offer US territories the option of full representation or full independence.
Conclusion:
So, with a quick look at Census data we were able to see the reasons that citizens are not voting. And, just as with voters, non-voters reasons for not voting are more complex than those individuals simply being apathetic. Understanding these reasonings would enable policy makers to better tailor solutions to address the causes of people not voting. As mentioned above, increasing voter access would not necessarily have a partisan benefit for either party, but that argument is beside the point since voting is a right and not one that is protected for voters of only one party.
