United States midterm election

A United States midterm election is an election held halfway through the president’s four-year term. Its principal federal contests determine the membership of Congress: all 435 voting seats in the House of Representatives and approximately one-third of the 100 Senate seats are regularly contested. Midterms therefore form part of the biennial congressional election cycle rather than a separate type of federal election. They do not elect a president. (usa.gov)

Midterm elections can change which political party controls either chamber of the national legislature, affecting the institutional relationship between Congress and the presidency. The president’s party has frequently lost congressional seats at midterms, although the historical record includes exceptions and different outcomes between the two chambers. Understanding midterms requires attention to constitutional arrangements, electoral administration, participation, campaigning, and the distinction between individual election results and nationwide patterns. (usa.gov)

Constitutional foundations and the electoral calendar

The midterm cycle results from the interaction of presidential and congressional terms under the United States Constitution. Article II, Section 1 establishes a four-year presidential term, while Article I, Section 2 requires members of the House of Representatives to be chosen every two years. Consequently, regularly scheduled congressional elections alternate between presidential-election years and years halfway through a presidential term. The 2022 and 2026 elections belong to the latter category; 2024 was a presidential-election year. “Midterm” describes this position in the electoral cycle, not a distinct constitutional category of election. Odd-numbered years are commonly called “off-years”: they fall outside the regular federal congressional cycle, although state and local elections and congressional special elections may occur then. (archives.gov)

The two congressional chambers follow different constitutional schedules. Article I, Section 3 establishes six-year Senate terms and divides the Senate into three classes, with approximately one-third of the seats coming up for election every two years. The Seventeenth Amendment, ratified in 1913, retained six-year terms while establishing direct popular election of senators. Responsibility for congressional elections is divided between the states and Congress: Article I, Section 4—the Elections Clause—assigns state legislatures the initial responsibility for prescribing their times, places, and manner, but authorizes Congress to make or alter those regulations, except the places of choosing senators. This allocation underlies the combination of state-administered elections and binding federal electoral requirements. (archives.gov)

The November election date is statutory rather than constitutional. Under 2 U.S.C. § 7, the regular election for representatives takes place on the Tuesday following the first Monday in November of every even-numbered year: therefore, between November 2 and November 8, never November 1. Under 2 U.S.C. § 1, regularly scheduled Senate elections coincide with the corresponding House election preceding the expiration of a senator’s term. These provisions establish the date of the general election, not a uniform nationwide schedule for every stage of the electoral process. (uscode.house.gov)

Party primaries are separate contests used to select nominees, while early in-person and absentee voting are methods of participating in an election before its designated election day. Special elections fill vacancies and may occur separately from, or concurrently with, a regular general election. A subsequent runoff can extend a contest beyond November: Georgia’s 2022 calendar, for example, scheduled its general primary for May 24, the general election for November 8, and a general-election runoff for December 6. These were distinct stages, rather than additional nationwide midterm election days. (sos.ga.gov)

Election day does not itself terminate congressional service. Section 1 of the Twentieth Amendment, ratified in 1933, fixes the end of expiring congressional terms at noon on January 3 and begins successors’ terms at that moment. The outgoing Congress therefore retains its constitutional authority after the November election until the term boundary; a session during this interval is commonly called a lame-duck session. Section 2 ordinarily requires Congress to assemble at noon on January 3, but permits a different meeting date by law. The constitutional beginning of a term is thus distinct from the chamber’s actual convening and members’ oath-taking. Continuing senators do not begin new six-year terms merely because a new Congress begins, and the presidential term boundary remains separately fixed at noon on January 20. (archives.gov)

House, Senate, and territorial contests

All 435 voting seats in the U.S. House of Representatives are contested in each regular midterm election. Representatives serve two-year terms, so no voting House seat is excluded from the regular cycle because its incumbent has time remaining in a longer term. Each representative is elected from a single congressional district; in a state entitled to only one representative, that district encompasses the entire state and the election is described as at large. Under the allocation following the 2020 census, six states have one representative each: Alaska, Delaware, North Dakota, South Dakota, Vermont, and Wyoming. Their statewide House contests remain elections for one of the 435 voting seats, not a separate category of congressional membership. (usa.gov)

The U.S. Senate has 100 seats, two for each state, with senators elected statewide for six-year terms. Its seats are divided into three classes: Classes I and II contain 33 seats each, and Class III contains 34. One class reaches the end of its term every two years, producing either 33 or 34 regularly scheduled Senate contests rather than an election for the entire chamber. A state’s two seats belong to different classes, so they do not ordinarily face regular election together; some states therefore have no Senate contest in a particular midterm. The rotation also crosses presidential and midterm cycles: the same class is regularly contested every six years, not at every midterm. (usa.gov)

Special elections can add contests to this regular schedule without increasing either chamber’s authorized membership. A vacancy in a voting House seat must be filled by election, not gubernatorial appointment; the winner serves the remainder of the existing term. A special House election may occur before, or alongside, the regular election for the following term. For Senate vacancies, the Seventeenth Amendment permits state legislatures to authorize temporary gubernatorial appointments pending an election, with the election arrangements determined by state law. A special Senate election fills an unexpired term and does not move the seat into another class. In Oklahoma in 2022, for example, voters reelected James Lankford in a regular full-term contest and elected Markwayne Mullin in a separate special contest for the unexpired term associated with Jim Inhofe’s seat. Thus, two Senate elections in one state need not mean that both seats’ regular terms are expiring. (archives.gov)

Separate from the 435 voting representatives, the House includes five delegates, representing the District of Columbia, American Samoa, Guam, the Northern Mariana Islands, and the U.S. Virgin Islands, and a resident commissioner representing Puerto Rico. Delegates serve two-year terms and consequently face regular election in both midterm and presidential-election years. Puerto Rico’s resident commissioner serves a four-year term, with regular elections aligned with presidential-election years rather than midterms. These six offices must therefore be distinguished both from voting House seats and from one another when counting congressional contests. (house.gov)

Delegates and the resident commissioner can introduce legislation, speak in House proceedings, and serve on committees, but cannot vote on final passage of legislation in the full House. Their elections provide congressional representation without conferring the same floor-voting authority as a representative elected from a state. Accordingly, a count of midterm House contests may include the five delegate elections, but those contests are additional to—not part of—the elections for 435 voting seats. (clerk.house.gov)

State and local elections and ballot measures

Gubernatorial elections coincide with midterms in 36 states, although their alignment results from state electoral calendars rather than a requirement governing congressional elections. Governors serve four-year terms except in New Hampshire and Vermont, where two-year terms place the office on the ballot in both midterm and presidential-election years. In 2022, both states reelected incumbent governors—Chris Sununu in New Hampshire and Phil Scott in Vermont. Other calendars separate gubernatorial contests from midterms: Kentucky, Louisiana, and Mississippi elect governors in the odd-numbered year preceding a presidential election, while New Jersey and Virginia do so in the year following it. Kentucky nevertheless elects its legislators in even-numbered years. (nga.org)

State legislative elections are more extensive than gubernatorial contests. In November 2022, regularly scheduled legislative elections took place in 46 states and 88 of the nation’s 99 legislative chambers; Louisiana, Mississippi, New Jersey, and Virginia held none because their regular legislative elections occur in odd-numbered years. Most state lower-house members serve two-year terms, whereas Alabama, Louisiana, Maryland, Mississippi, and North Dakota use four-year terms. State senators serve either two or four years, depending on the state. Four-year terms may be staggered or run concurrently: California elected 20 of its 40 senators in 2022, while Alabama elected all 35 senators and all 105 representatives. Consequently, a midterm can renew an entire state legislature, only one chamber, or portions of both chambers. (ncsl.org)

Other offices appearing alongside congressional candidates depend on the jurisdiction. California’s November 8, 2022, election included attorney general, secretary of state, and other statewide executive offices, as well as judicial retention questions. In its appellate judiciary, voters mark “yes” or “no” on retaining individual justices rather than choose between competing candidates; Supreme Court retention questions appear statewide, while appellate-court questions are restricted to the relevant districts. County and municipal ballots may include county executives, council members, mayors, and other local officers. Their cycles need not match those of state officials: Texas’s 2022 election calendar expressly combined federal, state, and county elections with regular elections held by participating local governments and special elections on propositions or vacancies. (vigarchive.sos.ca.gov)

Ballot measures ask voters to decide a legal or policy question rather than fill an office. Their subject and route to the ballot are distinct characteristics. A state constitutional amendment changes the state’s fundamental law; an initiative allows petitioners, after satisfying applicable requirements, to submit a proposed statute or, in some states, constitutional amendment. A popular referendum allows voters to approve or reject legislation already enacted, while a legislative referral places a question before voters through legislative action. California’s 2022 Proposition 28, for example, was a petition-qualified initiative statute proposing additional state funding for public-school arts and music education, with an estimated annual state cost of approximately $1 billion. Local propositions operate within a narrower jurisdiction: New York City’s 2022 ballot included citywide charter questions alongside a separate statewide environmental bond proposition. (ncsl.org)

These overlapping contests connect electoral cycles without creating a single nationally administered election. State law determines which state and local offices and questions coincide with congressional voting, and territorial boundaries determine which contests appear on a particular voter’s ballot. Election administration remains decentralized: state officials exercise oversight or advisory responsibilities, while counties—and, in some states, cities or townships—usually conduct elections. Sharing an election date therefore does not give a municipal proposition statewide effect or place state candidate selection under a national election authority. (sos.state.tx.us)

Eligibility, nomination, and voting procedures

Voting in congressional elections requires United States citizenship, attainment of age 18 by election day, and satisfaction of applicable state residency requirements. Registration procedures and deadlines vary; North Dakota does not require voter registration. States also differ in restrictions associated with criminal convictions and legal determinations of incapacity. These qualifications operate within protections of the United States Constitution: the Fifteenth Amendment (1870) prohibits racial discrimination in voting; the Nineteenth (1920) prohibits discrimination based on sex; the Twenty-fourth (1964) bars poll taxes in federal elections; and the Twenty-sixth (1971) protects citizens aged 18 or older against denial of voting rights because of age. (usa.gov)

Candidate selection is separate from voter eligibility. In party primaries, voters select nominees under rules governing participation: closed primaries generally require party affiliation, while open primaries permit participation without prior party registration; intermediate arrangements admit some unaffiliated voters. Alternative systems narrow the field rather than select a nominee for each political party. California’s congressional primaries place candidates on a common ballot and advance the two highest vote-getters regardless of party preference, potentially producing a general election between candidates preferring the same party. Independents must also advance through that primary. Elsewhere, separate petition routes exist: Virginia requires an independent United States House candidate to submit 1,000 qualified voters’ signatures, together with the required candidacy filings. Ballot access therefore depends on state-specific nomination routes, documentation, and deadlines. (eac.gov)

The rule determining the winner also varies. Under plurality voting, the candidate receiving the most votes wins without necessarily obtaining more than half. A majority-runoff system instead requires a further election when no candidate reaches the prescribed majority. Georgia’s 2022 Senate election proceeded from the November 8 general election to a December 6 runoff between Raphael Warnock and Herschel Walker. Alaska used a different arrangement in 2022: an August 16 nonpartisan primary advanced four candidates, followed by a November 8 general election using ranked-choice voting. Voters ranked candidates; if nobody initially secured a majority, the lowest-ranked candidate by vote total was eliminated and ballots transferred to their next continuing preference. Successive rounds continued until a candidate obtained a majority of the votes remaining in the count. Ballots without a continuing preference became inactive, so that majority need not represent a majority of all ballots originally cast. (axios.com)

Ballot-casting methods do not themselves establish eligibility. Depending on jurisdiction, voters may vote at election-day polling places, during an early in-person period, or through absentee or mail voting. Some states require an accepted reason for absentee voting; others permit it without an excuse or routinely mail ballots to registered voters. Mail ballots may be returned through authorized postal or in-person channels. Identification requirements, early-voting periods, return deadlines, and verification procedures differ. A provisional ballot records a vote when registration or another qualification cannot immediately be confirmed; it is kept separate and counted only after officials determine eligibility under applicable law. (eac.gov)

Military and overseas voting has additional federal protections. The Uniformed and Overseas Citizens Absentee Voting Act of 1986 covers qualifying uniformed-service voters, eligible family members, and overseas citizens in federal elections. As amended by the Military and Overseas Voter Empowerment Act of 2009, it generally requires transmission of requested absentee ballots at least 45 days before a federal election when requests arrive by that deadline. It also provides for electronic delivery of blank ballots and, under specified conditions, a Federal Write-In Absentee Ballot. These arrangements facilitate absentee participation rather than create a separate electorate. (justice.gov)

Administration is decentralized. State chief election officials exercise oversight or advisory responsibilities, while local officials generally manage registration records, polling operations, ballot handling, and counting, subject to state arrangements. The Election Assistance Commission, established by the Help America Vote Act of 2002, distributes grants, develops voluntary voting-system guidelines, administers testing and certification, and supplies research and administrative guidance. The Federal Election Commission instead administers and enforces federal campaign-finance law, including financial disclosure and contribution restrictions; it does not operate polling places or count congressional votes. Neither agency constitutes a single national authority administering midterm elections. (eac.gov)

District boundaries and the geography of representation

The decennial population census affects House elections through two distinct processes: apportionment distributes the 435 voting seats among states, while congressional redistricting establishes district boundaries within them. Each state receives at least one seat; the remaining seats are allocated through the method of equal proportions. Apportionment uses population, not the number of registered voters or ballots cast. Following the 2020 census, seven seats shifted among 13 states: Texas gained two; Colorado, Florida, Montana, North Carolina, and Oregon each gained one; and seven states each lost one. These allocations first applied in the 2022 midterms. Even states retaining the same number of representatives must account for population shifts within their borders when drawing districts. (census.gov)

House representation is therefore population-based, although allocating whole seats and guaranteeing each state one representative prevents exact equality across state lines. The Senate instead embodies equal state representation under the United States Constitution: California and Wyoming each elect two senators despite their different populations. Congressional redistricting does not alter Senate constituencies, which are entire states. Moreover, the Senate’s three staggered classes mean that only approximately one-third of its seats undergo regular election in a given cycle; the states voting in one midterm consequently differ from those voting in the next. (congress.gov)

Districting institutions vary by state. State legislatures generally exercise the districting power, but some states assign congressional mapmaking to commissions. In Arizona State Legislature v. Arizona Independent Redistricting Commission (2015), the Supreme Court upheld Arizona’s use of an independent commission established through a voter initiative. Whatever institution draws the map, federal constitutional constraints apply. In Wesberry v. Sanders (1964), the Court required congressional districts within a state to have populations as nearly equal as practicable. In Karcher v. Daggett (1983), it rejected the argument that New Jersey’s maximum population deviation of approximately 0.7 percent was automatically permissible: avoidable differences required adequate justification. Equal population concerns persons represented, not equal turnout in each district. (constitution.congress.gov)

Voting-rights protections impose additional constraints. In Allen v. Milligan, decided June 8, 2023, the Supreme Court affirmed a preliminary injunction against Alabama’s congressional map after the lower court found a likely violation of Section 2 of the Voting Rights Act of 1965. The map contained one majority-Black district among seven in a state whose population was approximately 27 percent Black. The decision applied an inquiry into minority population size and geographic compactness, political cohesion, majority bloc voting, and the totality of circumstances—not a requirement of proportional racial representation. Separately, racial classifications in districting face constitutional scrutiny. Partisan gerrymandering presents a different legal question: Rucho v. Common Cause (2019) held that partisan-gerrymandering claims were beyond federal courts’ jurisdiction, without eliminating state-law remedies. (supremecourt.gov)

Residential geography also affects how votes translate into seats. Concentrating a party’s supporters in a few districts can produce large winning margins there without yielding additional representatives; distributing supporters differently can produce more narrowly won seats. Jowei Chen and Jonathan Rodden’s 2013 study, Unintentional Gerrymandering, used simulated districting to show how residential clustering could generate partisan electoral bias without deliberately partisan mapmaking. Boundary choices can reinforce or counteract these geographic effects, so a difference between a party’s vote share and seat share does not by itself identify its cause. (sites.tufts.edu)

Nationwide vote totals also combine contests with different turnout and candidate choices. An uncontested district provides no direct measurement of how voters would divide between two major-party candidates; the Federal Election Commission’s 2022 compilation notes that some unopposed candidates did not appear on ballots and therefore received no recorded votes. Each such winner nevertheless occupied one seat. Senate aggregates are still less comparable with the chamber’s overall composition: they combine differently sized statewide electorates, exclude seats not contested that year, and can include additional special elections. A national vote–seat comparison must therefore distinguish House from Senate results, the particular contests included, and seats won that year from seats held afterward. (fec.gov)

Parties, candidates, and campaign finance

Midterm candidacies include incumbents seeking another term, challengers opposing sitting members, and contenders for open seats where no incumbent is running, including after retirement. Democratic and Republican campaign organizations recruit candidates and assist campaigns, but competition also occurs within each party. In Alaska’s 2022 Senate election, incumbent Republican Lisa Murkowski and fellow Republican Kelly Tshibaka both reached the general election, which Murkowski won. Independent and minor-party candidates also contest congressional elections: the Federal Election Commission’s (FEC) 2022 results record independent Angus King’s Senate reelection in Maine in 2018 and independent Bernie Sanders’s reelection in Vermont that year, illustrating that congressional competition is not exclusively between the two major parties. (fec.gov)

The principal congressional campaign organizations are the Democratic Congressional Campaign Committee and National Republican Congressional Committee for House elections, and the Democratic Senatorial Campaign Committee and National Republican Senatorial Committee for Senate elections. Their activities include recruitment, fundraising, campaign assistance, and advertising. The NRCC’s account of its 2010 campaign, for example, describes regional teams responsible for recruitment, fundraising, and incumbent retention. State and local party committees conduct voter identification, registration, and get-out-the-vote operations through personal contact, telephone calls, mail, and digital messaging. Volunteers supply unpaid services; under FEC rules, uncompensated personal services to a party committee are not reportable contributions, although associated payments or expenses may receive different treatment. (fec.gov)

Federal reporting distinguishes organizations from the activities they finance. A candidate’s authorized committee receives and spends funds on that candidate’s behalf; party committees support the party and its candidates. Political action committees (PACs) are a separate category encompassing organizations connected to corporations or unions and nonconnected committees. Traditional PACs may contribute to candidates subject to applicable restrictions. Independent-expenditure-only committees, commonly called super PACs, may accept unlimited contributions from permitted sources for independent political activity, rather than make contributions to candidates. A leadership PAC associated with an officeholder is not that officeholder’s authorized campaign committee. (fec.gov)

An independent expenditure pays for communication expressly advocating a clearly identified candidate’s election or defeat without coordination with that candidate, the campaign, or a political party. It is not a contribution received by the beneficiary’s campaign. “Electioneering communication” is a narrower statutory reporting category: broadcast, cable, or satellite communication referring to a federal candidate within 30 days before a primary or 60 days before a general election and targeted to the relevant electorate. For congressional candidates, targeting means the communication can reach at least 50,000 people in the relevant district or state. Express advocacy is not required. Noncommittee spenders report qualifying electioneering-communication disbursements exceeding $10,000 in a calendar year on Form 9; political committees report such spending under their expenditure requirements instead. (fec.gov)

Campaign communication combines television and radio advertising, direct mail, paid digital placements, social media, and voter-contact operations. These paid or campaign-produced messages are distinct from independent news coverage and candidate debates. FEC debate rules permit qualifying nonprofit staging organizations to use pre-established objective participation criteria, provided the organizations do not endorse, support, or oppose candidates or parties. Campaign advertisements and debate statements communicate candidates’ arguments and allegations; their inclusion in campaign reporting does not establish those claims as facts. (fec.gov)

Financial comparisons require explicit boundaries. For January 1, 2021–December 31, 2022, the FEC reported congressional candidate receipts of $3.7561 billion and disbursements of $3.7267 billion, covering 2,656 candidates whose committees reported financial activity. These figures include primaries, general elections, and specified special elections—not merely advertising before November 2022—and exclude separate party and outside-group totals. Receipts are not expenditure: cash balances, borrowing, repayments, and refunds affect the accounts. Comparisons should identify the reporting period, offices, committees, and spending categories, distinguish operating expenditure from transfers, and avoid counting both an intercommittee transfer and the recipient’s subsequent purchase as separate final spending. Independent-expenditure notices likewise should not be added again to committee disbursements already containing the same transactions. (fec.gov)

Turnout and the composition of the electorate

Midterm participation is lower than presidential-election participation in the Census Bureau’s Current Population Survey (CPS) series. Using the citizen voting-age population (CVAP)—citizens aged 18 or older in the civilian, noninstitutionalized population—as the denominator, reported turnout was 41.9 percent in 2014, 53.4 percent in 2018, and 52.2 percent in 2022. The increase between 2014 and 2018 was therefore 11.5 percentage points, followed by a 1.2-point decline in 2022. By comparison, the same survey series estimated presidential-election turnout at 66.8 percent in 2020 and 65.3 percent in 2024. These comparisons use reported participation rather than combining survey estimates for some years with administrative ballot counts for others. (census.gov)

The denominator substantially affects the meaning of voter turnout. The voting-age population includes adults irrespective of citizenship; CVAP excludes noncitizens but does not comprehensively remove citizens legally ineligible to vote. Voting-eligible population (VEP) estimates make further adjustments, notably for disenfranchisement associated with felony convictions; the University of Florida Election Lab also adds eligible overseas citizens to its national estimates. Registered-voter turnout instead measures participation among those registered, excluding unregistered eligible citizens from the denominator. Thus, the Census Bureau’s finding that 69.1 percent of CVAP was registered in 2022 describes registration, not turnout. Comparisons require an explicit population base, not simply a percentage labeled “participation.” (census.gov)

Participation differences alter the electorate’s demographic composition. In 2022, citizens aged 65 or older constituted 23.8 percent of CVAP but 30.4 percent of voters; those aged 18–29 constituted 19.6 percent of CVAP but only 11.7 percent of voters. Geography qualified this national pattern: young adults were not statistically underrepresented among Wisconsin voters, whereas their voter share in Texas was 10.2 percentage points below their CVAP share. Such comparisons distinguish a group’s share of all voters from its turnout rate: 11.7 percent was the young-adult share of the electorate, not the proportion of young adults who voted. (census.gov)

Education, income, race, and ethnicity also differentiated participation. Citizens with at least a bachelor’s degree accounted for 36.1 percent of CVAP but 46.1 percent of voters in 2022. Among people living in families, those with annual family incomes of at least $100,000 represented 36.7 percent of the relevant CVAP and 45.0 percent of voters. Non-Hispanic White citizens constituted 66.4 percent of CVAP and 73.3 percent of voters; corresponding shares were 12.4 and 10.9 percent for non-Hispanic Black citizens, 5.0 and 3.9 percent for non-Hispanic Asian citizens, and 13.4 and 9.7 percent for Hispanic citizens. These are descriptive associations, not estimates of independent causal effects. (www2.census.gov) Regional turnout also varied: the South’s CVAP turnout was 48.9 percent, compared with 54.7 percent in the West, 54.1 percent in the Midwest, and 53.8 percent in the Northeast; the latter three estimates were not statistically distinguishable. (census.gov)

CPS figures are weighted survey estimates, subject to sampling uncertainty, nonresponse, reporting errors, and questionnaire effects; numerical differences are not necessarily statistically significant. They measure reported voting, not verified individual ballot records. Administrative figures answer somewhat different questions: total ballots cast need not equal votes recorded for a particular office, because voters may leave that contest blank or vote only in state or local contests. The Census Bureau consequently cautions that its estimates can differ from official counts because of survey coverage, population controls, respondent misreporting, and the definition of the administrative total being compared. The CPS also asks no questions identifying respondents’ candidate choices or party affiliations, so its demographic findings do not directly establish partisan voting behavior. (census.gov)

Historical development and selected elections

The first midterm cycle occurred in 1790–1791, during George Washington’s first presidential term, and selected the Second Congress. States chose their own election dates rather than voting on a single national day. Congressional terms and actual meetings were also distinct: the Second Congress’s term began in March 1791, but its first session opened on October 24. These arrangements could leave substantial intervals between voting and legislative business. An act signed by President Ulysses S. Grant on February 2, 1872, established a common November date for House elections, although implementation was not immediate; the House historian identifies 1876 as the first congressional election under the uniform-date arrangement. (en.wikipedia.org)

Senate selection initially operated through state legislatures, so nineteenth-century midterms did not ordinarily include direct popular votes for senators. State legislative elections could nevertheless influence subsequent Senate choices. The Seventeenth Amendment, ratified on April 8, 1913, replaced legislative selection with direct election while preserving six-year terms and staggered classes. Existing senators’ terms were unaffected. Consequently, the 1914 midterms became the first regular nationwide Senate election cycle under the amendment, adding direct Senate contests to midterm ballots wherever seats were scheduled for election—not placing the entire Senate before voters simultaneously. (archives.gov)

Nineteenth-century results reflected party alignments that do not map neatly onto later two-party classifications. In 1862, during the American Civil War, Abraham Lincoln’s Republicans lost House seats but retained control with Unionist support. The House historian separately classifies Republicans, Unionists, Unconditional Unionists, and independent Republicans; the Senate likewise records multiple Unionist affiliations and the absence of Confederate-state representation. In 1874, during Grant’s second term, Democrats captured the House while Republicans retained the Senate. The election also occurred amid documented violence against Black and Republican voters in the South: a congressional investigation examined intimidation, destroyed ballot boxes, and fraudulent counting in Alabama. These circumstances establish historical context without reducing the national result to a single cause. (history.house.gov)

The twentieth century supplied contrasting outcomes. Franklin D. Roosevelt’s Democrats expanded their majorities in both chambers in 1934. Harry S. Truman’s Democrats lost both chambers in 1946, producing the Republican-controlled Eightieth Congress. Later transfers also differed by chamber: Republicans won both houses in 1994, and Democrats won both in 2006; Republicans captured only the House in 2010 and added the Senate in 2014. In 2018, Democrats captured the House while Republicans enlarged their Senate majority. In 2022, Republicans captured the House, but Democrats retained Senate control and gained one seat. (presidency.ucsb.edu)

The table reports net House-seat changes for the president’s party, calculated from the House historian’s initial general-election totals for successive Congresses—not from membership immediately before election day. D means Democratic and R Republican; control is listed as House/Senate. Vacancies remain unassigned, and subsequent special elections, contested-election resolutions, and party switches are excluded. Senate control includes caucus alignment, so Democratic control in 2006 did not require a majority of senators formally identifying as Democrats. (history.house.gov)

These snapshot-based calculations deliberately differ from election-night or seat-transfer measures: the 2010 House figure compares 257 Democrats in the preceding general-election results with 193 in the new results. The 2018 Republican figure compares 241 with 199; the historian’s initial results left one district unresolved. Neither calculation assigns later special-election outcomes retroactively to the preceding midterm. (history.house.gov)

Patterns of presidential-party performance

The president’s party lost House seats in 20 of the 23 midterm elections from 1934 through 2022. The exceptions were 1934, 1998, and 2002. This is a recurring historical tendency, not a constitutional rule or a prediction of a fixed-sized loss. House and Senate outcomes can diverge: in 2018, Republicans lost 41 House seats while gaining two Senate seats; in 2022, Democrats lost nine House seats while gaining one Senate seat. (presidency.ucsb.edu)

The exceptions also distinguish seat gains from chamber control. Using the American Presidency Project’s comparison of party divisions immediately following successive elections, Franklin D. Roosevelt’s Democrats gained nine House seats and nine Senate seats in 1934, retaining both majorities. Bill Clinton’s Democrats gained four House seats in 1998, with no net Senate change, but Republicans retained both chambers. George W. Bush’s Republicans gained eight House seats and two Senate seats in 2002, retaining the House and regaining the Senate. This baseline differs from counting seats held immediately before election day, because intervening membership changes can alter the comparison. (presidency.ucsb.edu)

One explanation is surge and decline, associated with Angus Campbell’s 1960 study. Presidential campaigns mobilize occasional voters and generate short-term support that benefits the winning presidential candidate’s congressional allies. At the following midterm, some of those voters abstain and some temporary supporters return to their customary partisan choices. Later versions distinguish changes in turnout from changes in vote choice and presidential coattails. Evidence is contested: Robert S. Erikson’s 1988 analysis of elections since 1902 found that standard surge-and-decline and withdrawn-coattail explanations did not adequately account for the historical regularity. James E. Campbell’s subsequent research gave presidential-year forces a larger explanatory role. The disagreement concerns mechanisms and model specification, not the existence of frequent losses. (doi.org)

Retrospective voting instead emphasizes judgments about governing performance. Voters may reward or punish congressional candidates of the president’s party for economic conditions and presidential conduct. Edward R. Tufte’s 1975 study connected national midterm voting with presidential approval and changes in real disposable personal income per capita. Such models explain variation in performance without necessarily explaining why losses recur so consistently: Erikson found that favorable economic conditions and presidential popularity did not eliminate the historical midterm penalty in his sample. Approval and economic indicators therefore enter as explanatory variables, not independently established causes of a particular result. Their estimated effects depend on the elections studied, the measurement period, and whether the outcome is votes or seats. (cambridge.org)

A third hypothesis, electoral balancing, treats opposition voting as an attempt to moderate policy rather than simply punish failure. In Alberto Alesina and Howard Rosenthal’s models, voters between the parties’ policy positions can strengthen the congressional opposition once the president’s identity is known. Their work predicts a midterm shift against the presidential party even without deteriorating economic performance. Alesina, John Londregan, and Rosenthal’s empirical analysis found support for a model combining policy moderation with economic assessments, but its tests also supported conventional retrospective voting. Thus, aggregate results need not uniquely identify voters’ motives, and balancing does not imply that every opposition voter consciously seeks divided government. (nber.org)

Empirical analyses also consider the starting distribution of seats, incumbency, retirements, open seats, and challenger quality. A party holding numerous marginal districts has different exposure from one concentrated in secure districts; national conditions may also influence strong candidates’ decisions to enter races. These factors limit the predictive value of a historical average. Describing midterms as a presidential “referendum” is consequently an analytical metaphor, not a direct confidence vote: voters choose congressional candidates under differing local conditions, rather than decide whether the president remains in office. Nor does the result determine the next presidential election, which occurs two years later with different candidates, participation, and political conditions. (edwardtufte.com)

Counting, certification, and interpretation of results

Election-night reporting produces unofficial returns, not certified results. Local election officials publish available counts, while state systems aggregate returns for congressional and statewide contests. Even “100 percent of precincts reporting” does not necessarily mean that every eligible mail or provisional ballot has been counted. Media projections identify an expected winner but have no certification authority. The subsequent canvass reconciles ballots cast with voter-participation records, checks precinct returns and tabulation reports, resolves discrepancies, and incorporates additional valid ballots. Certification is the legally prescribed attestation that the resulting totals accurately account for the votes cast. Certifying authorities and deadlines differ by state and may involve local boards, state canvassing boards, secretaries of state, or other designated officials. (eac.gov)

The time required depends partly on which ballots remain to be processed. Absentee and mail ballots require administrative checks before tabulation; some states also accept specified categories of ballots after election day if statutory conditions are met. Pennsylvania’s Department of State explains that officials cannot begin opening mail-ballot envelopes and preparing their contents for scanning—the process called pre-canvassing—before 7 a.m. on election day. Provisional ballots remain separate while officials investigate registration, eligibility, or possible duplicate voting; they may subsequently be counted fully, counted partially, or rejected under state law. These procedures explain why reporting can continue after polling places close and why election-night totals may omit valid votes. (eac.gov)

A post-election audit checks election evidence or procedures rather than automatically repeating the entire count. A tabulation audit commonly compares a sample of paper ballots with recorded totals; its scope may expand when discrepancies require further examination. A recount counts votes again in a specified contest and may be triggered by a narrow margin, an authorized request, or a legal order. Administrative reviews, such as checking ballot-accounting records or determining whether a provisional voter was eligible, are distinct from both procedures. Timing also varies: a recount can follow certification and produce revised official totals. For example, Georgia’s published rules for its June 2022 primary runoff allowed the second-place candidate to request a recount within two business days after certification when the difference was no more than 0.5 percent of votes cast in the contest. (eac.gov)

An election contest is a formal legal challenge to an election’s validity or outcome, not simply another name for a recount. Depending on the applicable law, it can concern ballot eligibility, counting decisions, or alleged irregularities and seek judicial relief. Congressional contests also involve a constitutional distinction: Article I, Section 5 of the United States Constitution makes each chamber the judge of its members’ elections, returns, and qualifications. In Roudebush v. Hartke (1972), the Supreme Court held that this authority did not prevent a state recount of a Senate election, because the Senate retained its independent judgment. State certification therefore does not displace the relevant chamber’s constitutional authority. (ncsc.org)

A runoff postpones the selection of a winner for a different reason: an additional election is required, rather than merely additional counting. Georgia’s November 8, 2022 Senate election produced no majority winner, so Raphael Warnock and Herschel Walker contested a runoff on December 6. Published congressional election statistics consequently identify the runoff totals as the decisive result. A results account must distinguish “runoff pending” from “count incomplete”; neither status establishes a certified winner of the seat. (clerk.house.gov)

Results tables likewise need separate categories for certified winners, projected winners, unresolved contests, and vacancies. Party affiliation should be recorded independently of caucus alignment: the Senate’s historical record lists the opening division of the 116th Congress in 2019 as 53 Republicans, 45 Democrats, and two independents, both caucusing with Democrats—not 47 Democratic senators. Election outcomes and subsequent membership changes also require separate dates and baselines. A later resignation, appointment, death, or party switch changes chamber membership without retroactively changing the original election result; a vacancy is not equivalent to an unresolved election. (eac.gov)

Consequences for governing institutions

A midterm election can transfer control of the House’s organization without changing the presidency. The incoming House elects its Speaker; party caucuses nominate candidates, but the election is decided by the chamber, not automatically by the national election result. The majority normally supplies committee chairs and controls committee majorities, giving it substantial influence over hearings, investigations, and which bills advance. The Speaker, majority leader, and Rules Committee also shape floor scheduling, debate, and amendment opportunities. These powers determine which proposals receive consideration, although they do not guarantee enactment. Following the Democratic House victory in 2018, Nancy Pelosi was elected Speaker on January 3, 2019, replacing Republican Paul Ryan. (house.gov)

Senate control affects both legislative scheduling and functions not assigned to the House. Party conferences choose their floor leaders, and the majority leader works with committee leaders to schedule business; priority recognition gives that leader an important procedural advantage. Control therefore influences the consideration of presidential nominees for executive offices and the federal judiciary, although confirmation remains a Senate decision rather than a leadership prerogative. Confirmation generally requires a simple majority, and precedents established in 2013 and 2017 permit majority cloture on executive and judicial nominations, including Supreme Court nominations. Treaty consent, by contrast, requires two-thirds of senators present. The House neither confirms these nominees nor gives constitutional advice and consent to treaties, although implementing legislation or appropriations may require both chambers. (senate.gov)

Unified government exists when the president’s party controls both chambers; divided government exists when the opposition controls at least one. The 2018 midterms produced divided government in January 2019: Democrats controlled the House, while Republicans retained the Senate and presidency. Such configurations alter bargaining over legislation, appropriations, and oversight, but do not themselves determine cooperation or deadlock. Under the United States Constitution, ordinary legislation must pass both chambers in identical form and be presented to the president; overriding a veto requires two-thirds support in each chamber. Appropriations likewise require legislation, so House control alone cannot establish a federal spending policy. Unified control does not eliminate these requirements or ensure agreement among legislators, and most Senate legislation remains subject to a three-fifths cloture threshold. Committee oversight, meanwhile, can proceed without enacting a new law. (history.house.gov)

Congressional defeat does not remove a president or shorten the presidential term. Removal through impeachment requires a separate constitutional process: the House approves articles of impeachment by a simple majority, and the Senate conducts a trial. Conviction requires two-thirds of senators present; when an incumbent president is tried, the chief justice presides. Conviction entails removal, whereas House impeachment alone constitutes the formal accusation. Consequently, a party’s acquisition of a House majority does not by itself confer the power to dismiss the president. (senate.gov)

Concurrent state elections change authority over state legislation, budgets, and executive administration when the successful candidates take office. Legislative majorities can revise a governor’s proposed budget, while gubernatorial vetoes and legislative override thresholds structure subsequent negotiations. Longer-term consequences depend on the institutions responsible for redistricting: elections can change the officials who later draw or approve maps, but some states assign that work to commissions. Michigan’s November 2018 electorate changed the institution itself by approving a constitutional amendment establishing a 13-member Independent Citizens Redistricting Commission—four Democratic affiliates, four Republican affiliates, and five members affiliated with neither major party. Its authority over congressional and state legislative boundaries illustrates how a midterm ballot can reshape a later redistricting cycle independently of immediate changes in elected officeholders. (ncsl.org)

References

1. Congressional elections and midterm electionsusa.gov
2. The Constitution of the United States: A Transcriptionarchives.gov
3. The Constitution: Amendments 11-27archives.gov
4. 2 USC 7: Time of electionuscode.house.gov
5. 2 USC 1: Time for election of Senatorsuscode.house.gov
6. Voting and electionsusa.gov
7. Absentee voting and voting by mailusa.gov
8. Representativeshouse.gov
9. U.S. Census Bureau Today Delivers State Population Totals for Congressional Apportionmentcensus.gov
10. U.S. Senate: Senators Receive Class Assignmentssenate.gov
11. U.S. Senate: About the Senate & the U.S. Constitution | Senate Classessenate.gov
12. Congressional Research Service: House Vacanciescongress.gov
13. U.S. Senate: About Electing and Appointing Senatorssenate.gov
14. Federal Elections 2022fec.gov
15. Office of the Clerk: Member FAQsclerk.house.gov
16. Constitution, Jefferson’s Manual, and Rules of the House of Representativesgovinfo.gov
17. Elections - National Governors Associationnga.org
18. Governors’ Powers & Authority - National Governors Associationnga.org
19. Consolidating Election Datesncsl.org
20. 2022 Legislative Races by State and Chamberncsl.org
21. Number of Legislators and Length of Terms in Yearsncsl.org
22. Candidate — Official Voter Information Guidevigarchive.sos.ca.gov
23. Supreme Court — Official Voter Information Guidevigarchive.sos.ca.gov
24. State and local electionsusa.gov
25. Nov 8, 2022 Election Law Calendarsos.state.tx.us
26. Initiative and Referendum Processesncsl.org
27. Proposition 28 — Official Voter Information Guidevigarchive.sos.ca.gov
28. The Contest List — General Election 11/08/2022vote.nyc
29. Who can and cannot voteusa.gov
30. Primary Election Typeseac.gov
31. Voter-Nominated Offices Informationsos.ca.gov
32. Primary Elections in Californiasos.ca.gov
33. Code of Virginia Code - Article 2. Independent Candidateslaw.lis.virginia.gov
34. Voter FAQseac.gov
35. Overview of Federal Election Lawseac.gov
36. 2020 Census Apportionment Resultscensus.gov
37. U.S. Census Bureau Today Delivers State Population Totals for Congressional Apportionmentcdn.www.census.gov
38. Congressional Redistricting: Key Legal and Policy Issuescongress.gov
39. U.S. Senate: Equal State Representationsenate.gov
40. Karcher v. Daggett, 462 U.S. 725 (1983)govinfo.gov
41. Allen v. Milligan (June 8, 2023)supremecourt.gov
42. Unintentional Gerrymandering: Political Geography and Electoral Bias in Legislaturessites.tufts.edu
43. Congressional Candidate Table 5fec.gov
44. Statistical Summary of 24-Month Campaign Activity of the 2021-2022 Election Cyclefec.gov
45. Looking Back at the 2010 Cyclenrcc.org
46. Campaign Guide for Political Party Committees (February 2024)fec.gov
47. Volunteer activity for party committeesfec.gov
48. Committee type code descriptionsfec.gov
49. Political Action Committees (PACs)fec.gov
50. Understanding independent expendituresfec.gov
51. Making electioneering communicationsfec.gov
52. Making independent expendituresfec.gov
53. Public communicationsfec.gov
54. Voter Registration in 2022 Highest in 20 Years for Congressional Electionscensus.gov
55. 2020 Presidential Election Voting and Registration Tables Now Availablecensus.gov
56. 2024 Presidential Election Voting and Registration Tables Now Availablecensus.gov
57. Frequently Asked Questions (FAQs) About Voting and Registrationcensus.gov
58. 2024 General Election Turnoutelection.lab.ufl.edu
59. 2022 Voting and Registration Data Now Availablecensus.gov
60. Voting and Registration in the Election of November 2022www2.census.gov
61. Census Bureau Releases 2022 Congressional Election Voting Reportcensus.gov
62. 1790–91 United States House of Representatives electionsen.wikipedia.org
63. 1st to 9th Congresses (1789–1807)history.house.gov
64. Election Firsts & Notableshistory.house.gov
65. 17th Amendment to the U.S. Constitution: Direct Election of U.S. Senatorsarchives.gov
66. Party Divisionshistory.house.gov
67. Party Divisionsenate.gov
68. The Demise of Reconstructionhistory.house.gov
69. 80th Congress (1947–1949)history.house.gov
70. Seats in Congress Gained/Lost by the President's Party in Mid-Term Electionspresidency.ucsb.edu
71. Coattails, Raincoats, and Congressional Election Outcomescambridge.org
72. The Puzzle of Midterm Lossuvm.edu
73. Determinants of the Outcomes of Midterm Congressional Electionscambridge.org
74. Partisan Cycles in Congressional Elections and the Macroeconomynber.org
75. A Model of the Political Economy of the United Statesnber.org
76. Midterm congressional elections: ET papersedwardtufte.com
77. The 2002 Midterm Election: A Typical or an Atypical Midterm?acsu.buffalo.edu
78. Midterm Elections and Divided Government: An Information-Driven Theory of Electoral Volatilityfaculty.ucmerced.edu
79. Election Results, Canvass, and Certificationeac.gov
80. Why do election results change after election night? When are election results final?eac.gov
81. Covering PA Elections 101pa.gov
82. EAC Best Practices: Provisional Votingeac.gov
83. Post-Election Tabulation Audit Guideeac.gov
84. Election Law Manual Fall 2022ncsc.org
85. Statistics of the Congressional Election of November 8, 2022clerk.house.gov
86. U.S. Senate: Party Divisionsenate.gov
87. Leadershiphouse.gov
88. Committee FAQsclerk.house.gov
89. House Leadershipclerk.house.gov
90. About the Committee on Rulesrules.house.gov
91. Speakers of the House by Congresshistory.house.gov
92. About Parties and Leadership: Majority and Minority Leaderssenate.gov
93. About Traditions & Symbols: Leader Recognitionsenate.gov
94. About Judicial Nominations: Historical Overviewsenate.gov
95. About Votingsenate.gov
96. About Treaties: Historical Overviewsenate.gov
97. Constitution of the United Statessenate.gov
98. Party Government Since 1857history.house.gov