2020 marked the first time in the 21st century that a president failed to win re-election. It’s a fairly uncommon event in American electoral history, although it happened four times in the 20th century. What makes the 2020 election unusual is that Mr. Trump received more votes when he lost than when he won.
In 2016. Mr. Trump’s total vote from all 51 jurisdictions was 62,984,828. In 2020 he increased his national total to 74,223,251. That’s a gain of more than 11 million votes, or 17.8%.
The four incumbents who failed to gain re-election in the 20th century lost votes on their second try, some of them dramatically.
William Howard Taft received nearly 7.7 million votes in 1908. Thanks to the intervention of former president Theodore Roosevelt as a third-party candidate, President Taft received fewer than 3.5 million votes in 1912. That’s a reduction of 54.6%.
Herbert Hoover obtained 21,392,190 votes nationwide in 1928, when he carried 40 of the 48 states. In 1932, in the midst of a severe economic contraction, his vote total was reduced by 26.4%, down to 15,761,254.
Jimmy Carter won 40.825 million votes in 1976. Four years of pessimism, malaise, and stagnation combined with humiliation by the Iranian mullahs to reduce his vote total in 1980 to 35.480 million, a reduction of some 15%.
Finally, George Bush the Elder won 48,886,597 total votes in 1988 but only 39,104,550 in 1992. His total vote declined by 20%.
Those precedents make it all the more remarkable that Mr. Trump increased his vote total by more than one-sixth in a losing effort.[1]
There is another unusual quality to the 2020 election, perhaps not an anomaly but worth noting. In three of the four 20th century elections where the president failed to be re-elected, his party suffered significant losses in the Congressional elections held in that same year. Here’s a summary:
1912 – Republicans lost 62 House seats and a majority as Wilson (D) won the presidency.
1932 – Republicans lost a staggering 97 House seats and a majority as Roosevelt (D) won the presidency.
1980 – Democrats lost 34 seats but retained a majority when Reagan (R) won the presidency. (Republicans had a working majority on issues where they could count on conservative Democrats, a political category that existed in 1980.)
1992 – This election was different from the other three. Republicans gained 9 seats as Bush the Elder lost the presidency. The Democrats retained a substantial majority as Clinton (D) won the presidential election. That result is not so unusual when you consider that Mr. Clinton was able to attract only some 43% of the total votes cast. Mr. Trump received 46.9% of the 2020 vote.
2020 – Although the incumbent president lost, Republicans added 15 seats but did not gain a majority.
It’s numbers like those that give Mr. Trump and his supporters the idea that something was off in the 2020 election. It has become an article of faith on the left that the 2020 election was the fairest ever (unlike 2016). I have noticed that news coverage over this dispute often describes Mr. Trump’s position as “unsupported” “baseless” or “false”. The Trumpists make it easier for their critics to enlist those dismissive terms when they focus their attack on electronic voting machines, claiming that these machines helped to commit fraud. The charge about voting machines has been leveled before. Democrats claimed that electronic voting machines assisted in the commission of fraud in the 2004 presidential election that went to the Republicans, although I don’t think anyone sued in connection with those charges.
Journalists who use words like “unsupported” and “false” are usually careful to apply them to claims of “widespread” fraud. That is clever. As President Obama pointed out in 2016, when Mr. Trump expressed worries about fraud (before either he or Mr. Obama could know that Trump would win), the American system of elections is diffuse. Elections are managed at precinct, ward, city, and county levels. Any attempt at widespread fraud is unlikely to succeed. Fraud, if it occurs, is going to be local.
Evidence of electoral fraud is bound to be murky. Voting irregularities come to light only when the vote is close and tempers are made raw as party organizers have to weigh hopes of victory and fear of defeat as counts and recounts trudge forward. The side that is ahead can’t afford to admit that anything irregular has taken place. The side that is behind is desperate to find some basis to change the outcome. It’s not an environment in which clear thinking is to be expected. Here are two examples from close elections where it is easier in hindsight to conclude that fraud occurred than to prove it.
After the 2016 Michigan presidential election results were tallied, but before Michigan’s electoral votes were awarded to Donald Trump, a minor-party candidate named Jill Stein demanded a recount. This seemed odd. She had won around one percent of the vote. A recount might show that she was entitled to a few hundred more votes out of millions cast, but it wasn’t going to award Michigan’s presidential electors to her. Nevertheless, the recount went forward.
The recount confirmed that Trump won Michigan, which was the result reported on election night (or perhaps it was the next day). It also revealed something unexpected. There were some 248 precincts in Detroit where the number of votes counted exceeded the number of votes cast. The authorities investigated. They found 34 individuals who had cast more than one ballot. That closed the case, but it didn’t solve the problem. It would take more than 34 rogue individuals to manufacture the fraudulent votes that were counted in so many precincts of Detroit in 2016. Something irregular had occurred, but it was local to those precincts, not widespread. Whatever it was, it was swept under the rug.
Something similar happened in the State of Washington in 2004. The state routinely elects Democrats to statewide office. In 2004, the election night count was a surprise. It showed that the Republican candidate for governor had won by a few hundred votes out of some 2.8 million cast. A machine recount reduced the Republican lead to about 40 votes. A third count, done by hand, awarded victory to Christine Gregoire, the Democratic candidate, by 130 votes.
As the recounts unfolded, King County, where heavily Democratic Seattle is located, found previously uncounted ballots on several occasions. Republican critics pointed out that the total number of votes manually recounted in King County exceeded by about 3,500 the number of voters known to have voted in the county. That’s a significant number in an election where the final margin of victory was 130 votes statewide.[2] Even so, the irregularities were not widespread.
In both cases, there is good evidence that someone somewhere tampered with ballots. The problem is that it is not possible to determine which ballots were filled out by someone other than a voter. If the rogue ballots can’t be identified, then it’s not possible to adjust the count by debiting the candidate selected on those phantom ballots. Eventually, the results are certified by the official empowered by law to collect the election results and report them. A strong desire to finish the process and move on will outweigh claims of theft and fraud. Higher authorities like secretaries of state, governors, or judges are unlikely to look behind a certified vote. They may be prohibited by law from doing so.
My assumption, which I admit is not proved, is that excess ballots in situations like this are the product of individual election workers, unobserved and unsupervised, not acting in concert with a large group of conspirators, just doing what some workers at ballot-counting stations had done for decades past and are likely to keep doing. When things are slow, you fill out a spare ballot here and another one there as the day goes by.
As in most areas of life, the overwhelming majority of election workers are honest, conscientious, and reliable. The rogue actors are likely small in number. The key to running an accountable election is to limit the scope in which the small number of rogue actors can operate.
Unfortunately, the American system of elections has evolved in recent years to provide leverage to the dishonest. For example, ballots are often deposited in collection boxes and then transported to another location to be counted. It’s not hard to imagine a particularly zealous partisan adding to the load of ballots to be transported. Even before collection, ballots might be pre-loaded and filled out in favor of a preferred candidate and deposited for collection. There are many other ways to go about it.
Ideally, we would have voters go to a physical polling place on an appointed day, establish their identity, and fill out a paper ballot which is counted at the end of the day at the place where it was cast. Results of local hand counts would be aggregated at ward, city, and county levels, but the ballots themselves would not be transported and would remain in safekeeping and available for audit until the results can be certified. Bi-partisan observers at each level of the process would keep an eye on the count as it unfolds. That’s how they do it in western republics like France, Germany, Sweden, Britain, Canada, Australia, New Zealand and the like. Their citizens trust the results. They may not like them, but they accept them.
The trend in the United States has been to spread voting out over a period of weeks and to deliver ballots to individual residences, mailboxes, and residential institutions (such as nursing homes). Each of those destinations becomes a polling place. We deliver ballots to the voter rather than require voters to come to the polling place to collect and mark their ballots.
The American way of handling elections widens the scope for irregularities. The advocates of reforming the process by spreading it out over time and space may have had good intentions – some of them – but the methods chosen by the political leadership of both parties have created conditions that make cheating easier to practice and harder to detect.
I offer this postulate for further study: The potential for fraud in any electoral procedure increases as the square of the area over which ballots are distributed and as the cube of the time from ballot distribution to collection and counting. I look forward to reading the doctoral dissertation confirming this principle. Name it after me if you wish, but I do not insist.
I am not advocating a return to a mythical golden age. Urban machine politics has a long history of corruption in the United States. The Democrats have been in charge of most of the nation’s cities for the last three-quarters of a century or more, but their Republican predecessors as urban political bosses were no better. The principal mechanism for ensuring an urban political machine’s electoral success was the artificial expansion of the vote by making sure that supporters “vote early and vote often”.
The opposite approach was adopted by party bosses in the less urbanized South. After the Fifteenth Amendment was adopted, states could not deny the vote to a citizen on the basis of race. African-American citizens exercised the franchise in the South during Reconstruction and through the period of Redemption, well into the 1890s. It took the white South a couple of decades to develop the idea of using literacy tests, property qualifications, and poll taxes to suppress the black vote. C. Vann Woodward notes that there were more than 130,000 African Americans registered to vote in Louisiana in 1896. In 1904 there were 1,342. The Jim Crow system that dominated the South right through to the 1965 Voting Rights Act was the direct result of intentional disfranchisement of black voters.[3]
The Voting Rights Act was designed to destroy the voting system in the South that denied the vote to African Americans. Over time, the Act has beaten back the old system of voter suppression. Unfortunately, it has done little to address the types of manipulation that were practiced outside the South. These methods could be called “ballot augmentation”. They give a political operation scope to collect ballots that were not legitimate votes. Through a clever piece of rhetoric, today’s proponents of “ballot augmentation” accuse their opponents of favoring “voter suppression”. No one wants to be accused of preventing citizens from voting, so politicians who should know better acquiesce in a system that is subject to subtle but persistent efforts to augment the count, but not the legitimate vote.
I began by noting that the 2020 election was unusual because the incumbent received more votes in his unsuccessful second election than he did on his first, successful, try. It was unusual, but not unique. In 1884, Grover Cleveland was elected president and collected 4.914 million votes. Although the voters rejected him in 1888, he increased his vote total to 5.34 million (and won more total votes nationally than Benjamin Harrison, the winner in 1888). Mr. Trump’s many supporters may be cheered, and his even more numerous opponents appalled, to remember that Mr. Cleveland ran again in 1892 and won the rubber match to become the only president to fill non-consecutive terms. He picked up only a few more votes when he won in 1892 than he received when he lost in 1888, but those additional votes came from the right places to ensure success. As Jack Nicholson said in “Mars Attacks”, two out of three ain’t bad.
— Gerry Bresslour
[1] There were four elections in the 20th century where the incumbent had been elected vice-president in the preceding election and was finishing his predecessor’s term. In all four of those elections the incumbent won (T. Roosevelt 1904; Coolidge 1924; Truman 1948; Johnson 1964). That doesn’t include 1976 when the incumbent Gerald Ford had succeeded to the presidency from the vice-presidency but had not been elected vice-president. 1904 was the first time a vice-president finished his predecessor’s term and then won an election in his own right. The four 19th-century gentlemen who finished their predecessor’s term (Tyler, Fillmore, Johnson, Arthur) did not even receive their party’s nomination at the next election.
[2] In February 2006, about a year after Ms. Gregoire took office, the Seattle Seahawks went to their first Superbowl, number 40 (that is, XL), played in Detroit. The governor was there in one of the prestige suites at Ford Field. I think I recall that after a disputed play, the cameras cut to her suite while the officials were looking at the replay. I don’t claim to be a lip reader, but it seemed to me that the governor was asking her companion “Why was there only one review?”
The Seahawks could have used the extra help. They lost that game 21-10.
[3] In the interest of bi-partisanship, let’s note that the Jim Crow system was what kept the Democratic party in power in the South until the 1960s. However, the Jim Crow era could not have begun without assistance from the U.S. Supreme Court in a number of critical decisions. To mention only one example, in the 1883 “Civil Rights Cases” the Supreme Court held that the 1875 Civil Rights Act was unconstitutional. That act outlawed denial of service on the basis of race at private places of public accommodation. The 8-1 majority included four Democrats and four Republicans.
The Passing Scene 