Concessions and non-concessions in the 2020 and 2022 elections for Congress

Refusing to admit defeat?
Concessions and non-concessions in the 2020 and 2022 elections for Congress

Concession state­ments by defeated can­di­dates are an important political ritual. They signal the candidate’s adherence to the demo­c­ra­t­ic rules of the game, which depend on parties to settle who holds power via elections and to respect the legit­i­ma­cy of the system in both victory and defeat. 

Donald Trump’s attempt to overturn his defeat in the 2020 U.S. pres­i­den­tial election fla­grant­ly violated this norm, raising concerns about a potential wave of non-con­ces­sions among like-minded election deniers running for office in 2022. After hedging as to whether they would accept the results, some prominent statewide can­di­dates did refuse to acknowl­edge the results but the feared avalanche of election denialism never mate­ri­al­ized. Still, experts currently forecast that former President Trump will again refuse to concede if he loses the November 2024 pres­i­den­tial election (median forecast: 93%) and that one or more statewide can­di­dates in 2024 will refuse to concede (75%). 

This report offers the most com­pre­hen­sive account of con­ces­sions and explicit refusals to concede in elections for federal office since Trump’s landmark election denial. We measure the frequency of these state­ments among runners-up in the 2020 and 2022 con­gres­sion­al elections from both parties and also identify can­di­dates for whom no public statement on the outcome of the election can be found. Our key findings are as follows:

  • Defeated Republican candidates for Congress were more likely to make public statements explicitly refusing to concede (14%) than were Democrats (1%). 

  • The problem was worse with Trump at the top of the ticket. In 2020, only 45% of defeated Republicans acknowledged defeat and 20% explicitly refused to concede (the remainder made no identifiable public statement). By contrast, in the 2022 midterms, 59% of defeated Republicans conceded, though 8% still refused to concede. 

  • Those Republicans who conceded in 2020 also took longer to do so on average than did Democrats after the race was called by the Associated Press, although the difference between parties was smaller and not statistically significant in 2022.

  • Refusals to concede by defeated Democrats were vanishingly rare (2% in 2020, none in 2022). 76% of defeated Democratic candidates in 2020 and 73% of those who ran in 2022 conceded.

Data col­lec­tion

We iden­ti­fied all second-place can­di­dates in the contested races for the U.S. Senate and House of Representatives in the 2020 and 2022 general elections. In total, there were 884 such contests – 449 in 2020 and 436 in 2022. We gathered infor­ma­tion on each losing candidate including their par­ti­san­ship (we excluded the 31 cases in which the runner-up was not a Democrat or Republican), whether they had pre­vi­ous­ly held elected office, the vote margin by which they lost, and the precise time when the Associated Press (AP) called their race. For each race, three research assis­tants searched inde­pen­dent­ly for evidence of a public statement by the  defeated major-party candidate about the outcome of the race in social and main­stream media. When possible, we identify the time at which the statement was made to determine how long the candidate waited to make a statement after the race was called. Finally, we classify the state­ments as a con­ces­sion or an explicit refusal to concede. Some can­di­dates made one or more addi­tion­al state­ments after initially refusing to concede; in those cases, we measure whether a con­ces­sion was ulti­mate­ly made and, if so, how long it took for the con­ces­sion to take place. We describe the data col­lec­tion process further in the Appendix below.

Concessions by party and election

A majority of defeated major-party can­di­dates in 2020 and 2022 – 65% overall (77% of sen­a­to­r­i­al can­di­dates, 62% of house) – released a statement acknowl­edg­ing their defeat, which we call a con­ces­sion. Three in ten defeated major-party can­di­dates (31%) did not make a statement publicly acknowl­edg­ing the results of the election. However, 8% of losing can­di­dates across both elections issued public state­ments explic­it­ly refusing to concede or denying their defeat.

The frequency of con­ces­sion state­ments increased slightly from 63% in 2020 to 68% in 2022. Conversely, non-con­ces­sions decreased from 11% in 2020 to 4% in 2022. In total, 47 runners-up in 2020 refused to concede compared to 18 in  2022. (The share of races in which no statement was made slightly increased from 29% to 32%.)

However, con­ces­sion behavior differs dra­mat­i­cal­ly by party. Trump’s effort to overturn the 2020 election was emulated by many Republicans, who echoed his false claims of wide­spread voter fraud. By contrast, Democrats over­whelm­ing­ly attested to the integrity of the electoral system. 

The figure below sum­ma­rizes con­ces­sion behavior by party (between losing Democratic and Republican can­di­dates) and by election (comparing the 2020 pres­i­den­tial and 2022 midterm elections).

Losing Democrats con­sis­tent­ly conceded when they lost (76% in 2020 and 73% in 2022); only 2% (4 can­di­dates) explic­it­ly refused to concede in 2020 and none did so in 2022, though about a quarter did not release a con­ces­sion statement in each cycle (22% in 2020 and 27% in 2022).

By contrast, losing Republicans conceded at lower rates, espe­cial­ly in 2020, when only 45% issued a public con­ces­sion. In that year, one in five Republicans publicly denied the results of their election (43 can­di­dates), the highest of any year/party group. Defeated Republicans also made no public statement on the results at higher rates than did Democrats – 34% in 2020 and 32% in 2022. However, the rate of GOP con­ces­sions increased sig­nif­i­cant­ly from 2020, when Trump himself refused to concede, to 2022, when the most prominent election denier was Arizona guber­na­to­r­i­al candidate Kari Lake. In 2022, con­ces­sions among defeated Republicans increased to 59% and explicit refusals to concede decreased to 8%.

Concession timing

Concessions followed more quickly on the Associated Press’s announce­ment of a race than did refusals to concede in both the 2020 and 2022 elections. On average, major-party runners-up conceded 1.2 days after the AP call in 2020 and 1.6 days after in 2022. Runners-up who contested the results took on average 4.7 days after the AP call in 2020 compared to 5.7 days in 2022. 

The figure below sum­ma­rizes how soon can­di­dates conceded after the race was called by party or election. (These results are limited to races for which the time of con­ces­sion can be established.) 

Democrats not only concede more fre­quent­ly than Republicans but do so more quickly after the race is decided —  averaging 14 hours after the Associated Press called the outcome of the race in 2020 compared to 2.3 days among Republicans. In 2022, by contrast, the dif­fer­ence between parties was smaller and not sta­tis­ti­cal­ly sig­nif­i­cant (1.1 days for Democrats versus 2.1 days among Republicans). 

Notably, 80% of Democratic con­ces­sions in 2020 took place within one day of the AP election call compared to 61% for Republicans. The partisan gap in day one con­ces­sions also held in 2022 (83% of Democrats compared to 66% of Republicans). Refusals took longer on average; only 33% of explicit refusals to concede by Republicans took place within a day of the AP call in 2020 (25% in 2022).

Margin of loss

One might expect that can­di­dates would be most motivated to dispute the results of close elections, but support for this expec­ta­tion is mixed. We divided races into four cat­e­gories – the most closely contested races (with vote margins of 0% to 5%), com­pet­i­tive races (margins 5%-10%), uncom­pet­i­tive races (10%-30%), and land­slides (margins greater than 30%). In 2020, there were 38 closely contested races, 46 com­pet­i­tive races, 116 uncom­pet­i­tive races, and 93 land­slides. In 2022, these numbers shifted to 46, 35, 171, and 184, respec­tive­ly. While the overall con­ces­sion rate was lower in 2020, the relative trends across different margin cat­e­gories were con­sis­tent with those in 2022. Losing major-party can­di­dates in com­pet­i­tive races had the highest con­ces­sion rates (85% in 2020, 100% in 2022), followed by those in closely contested races (69% in 2020, 96% in 2022), uncom­pet­i­tive races (65% in 2020, 69% in 2022), and land­slides (50% in 2020, 48% in 2022). This trend was more pro­nounced among Republicans than Democrats.

Among losing Democrats, con­ces­sion state­ments were most frequent in the closest races in both elections (94% in 2020 and 100% in 2022). We then observe a steady drop-off in the like­li­hood of a public con­ces­sion as margins widen. The incidence of refusals to concede (shown at the bottom of the graph) is close to flat for Democrats, with 6% refusals in the closest races in 2020, then only a handful in uncom­pet­i­tive races (1%) and land­slides (2%) that election, and none at any loss margin in 2022. 

By contrast, among Republicans in the closest 2020 races, the share of runners-up who refused to concede (38%) rivaled that who did (48%). In general, the shares of refusals generally declined as margin of loss increased — 20%, 21%, and 15% in com­pet­i­tive, uncom­pet­i­tive, and landslide races, respec­tive­ly. In 2022, Republican refusals were less common overall in the closest (4%) and com­pet­i­tive (0%) races compared to uncom­pet­i­tive (13%) and landslide losses (8%). 

(In the figure, the omitted category of behavior is not making a public statement on the results of the election, which becomes more likely for both parties as loss margins grow.)

Experience in elected office

We measured whether a candidate pre­vi­ous­ly held elected office, a common measure of amateur status. In total, 91% (193) of defeated Republicans in 2020 and 92% (196) in 2022 had not pre­vi­ous­ly held office compared to 83% (185) of defeated Democrats in 2020 and 83% (169) in 2022. In total, 70% of can­di­dates for Senate were amateurs as well as 89% of can­di­dates for the House of Representatives (these figures were nearly identical across elections). 

In both parties, con­ces­sions are more common among can­di­dates who pre­vi­ous­ly held elected office. Among Democrats, in 2020, 75% of can­di­dates who had not pre­vi­ous­ly held elected office conceded compared with 85% among those who had. In 2022, the gap was 69% to 91%, respec­tive­ly. Similarly among Republicans, only 43% of amateurs conceded in 2020 compared to 68% of runners-up who had held office, though the rate of refusal to concede was similar across groups (21% to 16%, respec­tive­ly). In 2022, 58% of amateur Republicans conceded and 9% refused to concede, whereas 72% conceded and none refused among those with prior expe­ri­ence in elected office.

In races decided by fewer than 10 per­cent­age points, amateurs refuse to concede at a slightly higher rate (5%) than do non-amateurs (3%). As margins widen, amateurs are dif­fer­en­tial­ly more likely than prior office holders to contest the results rather than concede. For races decided by margins of 10 per­cent­age points or more, 8% of amateurs refuse to concede and only 57% concede compared to 0% and 71%, respec­tive­ly, among non-amateurs.

Appendix

We followed the steps below to collect the data in this report:

  • Scraped Ballotpedia to identify congressional runners-up in 2020 and 2022, including results for all House and Senate seats and the vote share and social media handles of the top two vote-getters. (Uncontested races and those in which the runner-up candidate was not a Democrat or Republican were excluded.)

  • We then scraped data on each runner-up, including whether they previously held elected office. 

  • Data on the timing of Associated Press calls was provided to us by the AP.

  • Using only social media accounts created before November 30th of the election year in question, we gathered each tweet from the runner-up candidate. 

  • Research assistants (RAs) searched the social media accounts of runners-up for text or video concession statements. Each candidate was assigned to three RAs. 

  • All identified statements were coded as a concession, a refusal to concede, or neither.  Disagreements between RAs were resolved by discussion with the supervising researcher. The timestamp is recorded where possible for each concession statement identified. All candidates for whom no concession or refusal to concede could be identified were coded as not having made a statement about the results of the election. 

  • Some candidates made multiple statements regarding the election results. We then classified each of them and made a judgment about the candidate’s behavior. In cases like these, we sought to give candidates the benefit of the doubt and minimize false positives for refusal. For instance, if a candidate refused to concede immediately after the AP call but later conceded, we would typically classify the candidate as having conceded based on the second statement and use the timing of that statement to measure when they conceded.