Once it became clear that Trump was going to become the president of the USA, my Facebook feed became cluttered with attempts to understand how that could possibly happen. How could a white supremacist, misogynist and utterly transparent snake oil salesman accumulate so many votes? Those on the left both inside and outside the borders of the USA struggled to understand what had happened.

A common conclusion in too many of these pieces is that the left needs to reach out, and listen to the concerns of, those who voted for him as a priority. In a similar fashion to how sections of the left evaluated Brexit, they see a working class anti-establishment rebellion in the Trump vote from what they term the ‘white working class’. They believe that component was won by Trump because it has been neglected by the left — often, they will assert, because the rest of the left was distracted by what they call identity politics.

This is a simple explanatory story that is particularly attractive to those sections of the left that have a nostalgic yearning for an imagined past of pure class struggle, shorn of internal concerns around oppression. But the concept of masses of otherwise progressive working class voters opting for Trump on economic grounds is a myth. The attractiveness of that myth and its promotion has more to do with the hostility of that section of the left towards the influence of intersectional feminism than anything more substantive. That hostility has caused them to seek out anecdotes and exceptional regions and present them as the typical story that defines the election just as liberal Hillary Clinton campaigners have focused in on Facebook false news stories as the cause of her defeat.

It’s this simple story that I’m examining at length in this piece, it’s not an overview of what the election means. It’s addressing a particular debate on the left that I have characterised as one between two poles of the modern left. A backwards looking Nostalgic Left dreaming of a mostly fictional past when a united (white male) working class was not distracted by what they term ‘identity politics’. And a forward looking Intersectional Left ambitiously trying to understand how class interacts with the intersections of white supremacy and heteropatriarchy. You can tell where my sympathies lie from that description but for the most part I’m going to be concentrating on the facts that can be gleaned from the exit polls after a few much needed opening disclaimers, in particular about the historical and current role of white supremacy in the US.

Below I show that while it is true that Trump attracted ‘white working class’ voters it’s also the case that the number is small in proportion to the US working class. And exit polling shows us that, in both opinions and demographics, those Trump voters are not our audience. For the most part they are not people who voted Trump in spite of his racist policies but people who share those racist policies.

The ‘sound on the economy but willing to vote for a racist’ fraction are not even that large a proportion of Trump voters, I’ll demonstrate this below. But what is really important to the story is that they are a tiny segment of the US working class. That is missing from almost all accounts of what happened that I’ve read, almost all proceed on the deeply misleading basis that half of the US population went for Trump. The error has arisen because most of the left, the electoralist left, is too used to speaking of voters as if they were representative of the population. So you will see lots of commentary that refers to the Trump vote as almost half the population when in fact it was about 20%. That’s because not everyone votes, more importantly not everyone can vote, and most importantly who votes is shaped by class, race and the intersections of both.

Some numbers to illustrate this. At the time of the 2010 Census the population of the US was 309 million, it has risen since, the UN now estimates it at 325 million. At the time of the election there were 231,556,622 eligible voters. Just 135 million of those voters actually voted and of those about 64 million voted for Clinton, 62 million for Trump and 7 million for other candidates. The slow count of the remaining votes and that some voters voted in referenda and / or local elections but did not vote in the presidential election account for the shortfall between the candidate votes so far and the total votes cast, it appears 133 million cast a vote for president.

Who doesn’t get to vote?

There is a shortfall of 94 million missing voters between the 231 million who could vote and the 325 million total population. The largest component of this is people under 18, around 75 million, but there are three other very important blocks

  1. 6 million people who are excluded from voting because they have been imprisoned for more than one year. Al Jazeera breaks the numbers down in their article on prisoner disenfranchisement One important figure is that while 1 in 13 African American are excluded from voting in this manner only 1 in 25 whites are.

  2. Homeland security estimates there are 11.4 million people in the US without the paperwork required for them to be there, about 74% of these are from central America with 5% from Europe and Canada.

  3. 14 million legal migrant residents can’t vote.

I was unaware of the size of this last group until I read the excellent piece on American Electoral Apartheid by Konstantin Kilibarda identifying 30–40 million in total excluded from voting, he includes those in de facto US colonies and recently deported, which is why his total is higher than my figures suggest, which are based on those inside US borders.

A very systematic detailed state by state breakdown which doesn’t count some of these exclusions has a shortfall of just under 20 million and in its detail and small print demonstrated why this is such a complex calculation. The final number being determined by political considerations as well as the difficulty of obtaining and grouping data. I’ve used their estimated final turnout figures as it includes where states release the information people who voted but did not vote in the presidential election.

Remember this is 62 million voters in a total population of 325 million and there were 75 million too young to vote. It’s hard to factor this in as we’d not expect a 1 year old to have an opinion on Trump but we would expect a 17 year old too have one. We do know from exit polls that he’d have massively lost the election if it was confined to under 25s, of which only 35% voted for him.

Another factor is that the votes cast in this election include a number of US citizens not living in the US, 5.7 million of whom were entitled to vote. So if anything this means the proportions of Trump voters in the adult population calculated below will be a slight overestimate.

67% of voters reported earning less than 100k per annum so of those 133 million voters about 89 million were in that 0-100k bracket which we will treat as enumerating the working class.

However workers were less likely to vote in the election. 2014 data suggests that in the US population overall about 74% of people (185 million) are in households that earn less than 100,000 which is 7% more than that found in the voter sample. The gap between votes and population is starker when expressed the other way around, only 26% of the population earn more than 100k but 33% of voters earned over 100k. So those earning over 100k were about 24% more likely to have voted.

The reasons why the voting population is more representative of wealthier people are important. The migrants and ex prisoners excluded from voting that we mentioned above are likely to overwhelmingly earn less than 100k. But it’s also the pattern in most countries that wealthy people turnout in elections at much higher proportions than poorer people because the wealthy know that politicians represent them and the poor know they will be ignored. 53% of the US working class did not vote (including the many who could not). Unusually the New York Times gave voice to some of them.

It’s worth noting in passing that just as Brexit could only pass because almost all the migrants affected by it were excluded from voting so too Trump only won the election because these millions of people were excluded from the eligible voters. This can be said with confidence because most of these people excluded from voting are people of colour, which is systematic of the structural racism that continues to be at the heart of the US system. This also illustrates why you can’t assume those who voted are somehow representative of the population never mind the working class. They are not, in terms of the legal exclusions above, barriers to voting, and the passive boycott by almost half the electorate, the voting population leans very much towards the more privileged section of the US population. In terms of this election and the Trump vote it’s almost certainly the case that that he pulled out additional voters motivated by white supremacy, misogyny, homophobia, transphobia and the sort of snobbery that hates others who receive state payments. And that some economic right wingers who found him too much on these issues stayed at home.

On the other side it’s clear that Clinton failed to mobilise millions of voters who turned out for Obama. Her campaign strategy was based on winning over moderate republicans horrified by Trump — they mostly held their nose and voted for him. She managed to increase the Democratic party share of votes among the wealthy — although Trump won the majority there. But her share of the working class vote fell sharply, including amongst black and latino working class voters whom the Democratic Party presumed the very presence of Trump would be enough to drive to the polls. Clinton was correctly seen as the representative of the neoliberal capitalism that has eroded wages, conditions and security for most workers in the US. She gambled that Trump would be seen as the same but that his racism and misogyny would mean that she’d maintain if not build on the vote Obama received from women and people of colour. It was a gamble that almost worked, she won the popular vote, but the nature of the US electoral system and Trump’s successful appeal to a bigoted slice of the ‘white working class’ meant he took the presidency.

On whiteness in the US and the use of ‘white working class’ by the left

This is a good moment to insert a disclaimer on the use of the term ’white working class’, in particular in relation to the US. The most brutal aspects of US capitalism were built on a premise of bringing poor migrants on board a colonial project that was genocidal towards the existing population of native americans. Many of these migrants did not have much in common with those running the colonial project, in some cases like the Irish they were victims of the same colonial masters back in Europe. Often they found themselves in America only because of the conditions of warfare, poverty and starvation the colonial project created that drove them from their homes. And significant numbers were forcibly transported after becoming prisoners in colonial wars or falling foul of a legal system that set traps everywhere.

In the initial period of colonisation in particular it was not that unusual for such colonists to desert and flee to the interior. In some cases common cause was made with the indigenous population although the fact that such colonists were illiterate means that we only have fragments of a record of the extent of this. For those interested in more The Many Headed Hydra provides an excellent introduction.

The entire project of creating the United States was one where the existing population were violently removed from the lands where they lived in a process covering hundreds of years, into the 1880s certainly and in terms of consequences right down to today and the Standing Rock struggle. Essential to the success of that project was the ability of elite Europeans to retain control over poor new arrivals. This wasn’t necessarily an easy task, often the relationship back in Europe between these two groups was hostile and conditions in the ‘new world’ were such that the settlers had a lot more power and autonomy.

The glue that allowed this was the deliberate development of white supremacy. Being white was what meant one set of people fought something close to a war of extermination against the ‘not white’ Native Americans. Being white meant you were given the land that the Native American population was living on and if they resisted the army turned up and killed them. Perhaps more insidiously being white was what the settlers had in common and, in often hard frontier conditions, the common mark they bonded under.

Within that project was the capture, transportation and enslavement of millions of black people, also over hundreds of years. There are moments in that history when poor whites stood with Native Americans and enslaved people, but for the most part the ideological bedrock of the USA was white supremacy and the granting of privilege to the section of the poor judged white enough. Particularly in later years becoming white was often a key breakthrough moment for populations including Irish catholics and European jews who were initially received with hostility and suspicion.

Where common rebellions happened the near universal lessons drawn by those who ruled was that they had become careless in ensuring that white privilege and needed to restore it to the lower orders. A lesson carelessly echoed in the current way some on the left talk of the abandoned white working class. The expected privilege of whiteness is a concept that often appears but as often is not remarked on in anecdotal accounts of working class Trump voters.

The big problem with that line cutting metaphor

A common metaphor used and reported on is the idea of a line of people queuing for the American dream. Working class white voters are reported as expressing their discontent that others are now being allowed to skip the queue, that they are stuck in place while ‘line cutters’ move in ahead of them. So their outrage is a product of this unfairness and resulted in a vote for Trump. But lets look at that line a bit more critically. It’s not a newly formed line, but one that has been shuffling along since the formation of the US. Those towards the front owe their position to white supremacy, it was their reward for denying access to the line to the black population, taking the line off the Native American population and keeping others including Chinese and Latinos at the back. Their outrage is that they have been ‘cheated’ of the reward because suddenly not everyone ahead of them is white, and they know that means those people have somehow got there from their rightful position behind them. And in neoliberal America the line hardly moves at all so there is a sense that perhaps they are being backed up.

The problem with the approach some on the left have argued to center this unhappy section of the white working class in order to avoid (in their terms) class disunity is that it amounts to the demand that everyone accept their current position in line on the promise that ‘come the revolution’ the line will be abolished. With no sign of the revolution in sight that’s a very unattractive proposition to anyone not near the front who has no hope of shuffling over in the meantime. And even the most pro revolutionary worker understands that ‘the meantime’ may be longer than their lifespan.

The exit polls tell us that Trump’s voters were overwhelmingly white, 55 million of his 62 million voters were white. If there was no bias in relation to the overall voting population it would instead have only been 43 million. They also tell us he did much better amongst higher earners, and in particular those earning over 100k, but there is no denying that many white people earning less than 100k also voted for him, mostly those in the 50-100k bracket. This is the Trump voting portion of the ‘white working class.’

The legal enforcement of the whiteness boundary was only removed in the 1960s and the structural basis of it remains strong, reflected in everything from police killings to average family wealth. So there are enormous dangers to the left in returning to a language that seeks to specifically single out and redress the grievances of the white working class as a distinct body, separate from the rest of the working class. For the purposes of this discussion I’m forced to reluctantly use the term because so many of the left rushed to uncritically embrace the idea of the rebellion of the ‘white’ section of the working class.

As far as possible I’m going to talk in terms of the working class as a whole as the other huge problem is that talk of the white working class so easily slides over into erasing the rest of the working class behind a ‘representative’ figure of a white male industrial worker. One excellent example of what I’m arguing against is provided in the article What so Many People don’t get about the US working class. It’s certainly not the case that the ‘white working class’ is the section of the working class hardest hit by neoliberalism, see the article A reality check on 2016’s economically marginalised.

Voters — a minority of the population

Within hours of polls closing it was clear that Trump would win the electoral college but two weeks later when I did these calculation Trump had got 62,513,667 votes and Clinton actually got more at 64,818,930. Roughly 7 million votes went to other candidates. Trump won because the election takes place on a state by state basis and in most states the ‘winner takes all’ of the votes that are then cast for president.

So of 251 million adults some 62 million or nearly 25% voted for Trump and nearly 26% for Clinton. 2% of the adult population voted for other candidates.

Who were these people voting for Trump? A section of the left has argued that they were white working class voters rebelling against neoliberalism. It’s then argued that these are people the left should be organising and fresh effort needs to be made to reach them. And finally some on the left argue that the left is too distracted on the way race, class and gender intersect and so has neglected the white working class voters. They want the left to ditch what they call ‘identity politics’ and return to focusing primarily on class.

In practise because they want to talk about class without centering how it is intersected by race and gender they end up focusing on the interests of largest segment of the US working class, white males, and treating those interests as typical rather than sectional. Far from creating unity this approach creates division as the fault lines imposed by white supremacist patriarchal capitalism reassert themselves. More on this later.

It’s this set of arguments from what I characterise as the Nostalgic Left that I address here. This piece isn’t about who is to ‘blame’ for the Trump victory. As an anarchist I wouldn’t have voted for Hillary even if she wasn’t such an establishment candidate so I’ve no interest in finding excuses for her failure to get elected. It’s clear there were multiple reasons for this, those who try to present it as a simple single factor story tend to be those who are seeking excuses on the one hand and hammers to batter their opponents with on the other.

Even in Democratic Party terms looking across the recent elections it’s very clear the reason she wasn’t elected was that 4 million people who found something to vote for in Obama in 2008 did not find the same in Clinton. It’s not hard to see why that might be, in 2008 Obama mobilised people on the promise that things were going to be better, Clinton mostly said things were already OK and relied on the horror of Trump to mobilise the vote. Trump marginally increased the Republican party vote, the number of additional voters he mobilised mostly being wiped out by the drop in the republican white women vote.

We can certainly say that the calculated gamble that the orientation the Democratic Party once had to industrial workers could be dropped and replaced by appeals for an identity plus ‘decency’ based vote didn’t work out. But to a large extent this wasn’t a tactical decision, Clinton’s neoliberalism was at the center of her politics and the central plank of neoliberalism in the US was the sacrifice of relatively well paid blue collar jobs. Clinton made rhetorical statements about defending such jobs as part of her election campaign but no one could take that seriously. She was widely and correctly understood as the Wall Street candidate. The only remarkable aspect is the possibility that Trump’s similar promises were taken seriously by some workers, but perhaps that is more a measure of desperation than anything else.

Both the main candidates were so unconvincing that a fair few people went to the voting stations, voted on referendums or local elections and then just didn’t vote at all for president. In the key state of Michigan MSNBC Morning Joe reported that 90,000 voters left the presidential slot blank on their ballots. And in the rustbelt most of the drop in the Democratic vote in comparison with 2012 saw a possibly equivalent rise in the vote for the third party candidates with the right wing Libertarian Party in effect offering an opposition to trade deals not coupled to Trump’s racism.

The Democratic party hoped that Trump was so terrible that who they saw as ‘their voters’ would turn out to vote for their less terrible candidate. They lost the election because too many of those voters either stayed at home or voted third party. Across the country there was a massive increase in the 3d party vote from 2 million in 2012 to close to 7 million in 2016, indeed in percentage increase terms the 3rd parties had a fantastic day with a trebling of their vote.

My interest here is also not to discover what best predicted whether or not someone would vote for Trump. For the curious the Economist has looked at this and it appears that the answer is poor health, even when controlled for obvious related factors like age.

The key question under examination here is what was the extent and nature of Trump’s working class vote. The nature of the vote is my key interest and to measure that we will look at what opinions did that section of voters espouse on questions that span the progressive V reactionary divide?

How big was the Trump working class vote?

Who were these people who voted for Trump? And more importantly what proportion of the working class are we talking about?

We can try to answer that question using the data available in the exit polls that asked voters a wide range of questions about their circumstances and their attitudes. I use the CNN presentation of the Edison exit poll data below. The same Edison data was used by the New York Times and a number of other major news sites. For these purposes it doesn’t matter that the exit polls failed to predict the result, its enough they were right within a couple of percent.

As with the Ashcroft vote poll taken during the Brexit vote we face the frustration that the exact questions we’d want are seldom asked so we have to make do with less precise approximations. The class positions of voters is particularly tough to capture but with the CNN poll the most useful indicator is how much voters earn and in particular the 0-50k, 50-100k and above 100k brackets.

Some left sources have used whether or not someone has a college degree to assign class but I don’t think that is as useful as income, not only because workers these days often have college degrees but also because older bosses of small and medium business often do not, even in 2003 almost half of managers didn’t have degrees. In that 2003 US census only 10% of those who were in the 25–29 age bracket in 1960 had college degrees, by the time of the survey nearly 30% in that bracket at that age did. Since Trump’s voters were very weighed towards voters over 45 this effect will be magnified further. College degrees in that age group are very rare.

That fact that across the board Trump did worse among those earning less than 50k subdivision of those earning less than 100k is further evidence that having a college degree does not map as well onto class as some have assumed. If it did we would see the reverse of that trend as those under 50k are less likely to have degrees. Income provides a far better approximation as few workers earn over 100k and few medium and high level managers or business owners earn below it.

A very detailed study that looked at incomes and education levels in comparison with how counties voted indicates that the Trump vote was predicted by low levels of education as distinct from low levels of income. That is that relatively high formal education but low income counties voted Clinton while high income but low formal education counties voted for Trump. / High income but low education Suffolk county, New York swung strongly for Trump (average income 88,000) but 22 low income, high education counties with average earnings of less than 50,000 swung to Clinton.

Of those voters earning under 100k only 45% said they voted for Trump, of those earning less than 50k his vote dropped to 41%. If we take that 100k/45% Trump voter segment this calculates his 62 million voters into 40 million that on income terms might be called working class, this is 64.7% of his voters. This is crude, there are certainly small business owners who earn less than 100k and some workers who earn more than this. But as an approximation of class positions it is the best available and tells us that no more than 40 million workers voted for Trump. While there are certainly managers and small business owners who earn less than that there are very few workers who earn more.

About 185 million adults are in households earning less than 100k so the Trump 40 million working class vote was about 22% of the working class in the US, slightly more than one worker in five.

Right away this figure on its own shows why a left concentration on winning over those who voted for Trump would be a mistake, Why focus on the 22% of the working class that voted Trump rather than the 78% of the working class who did not?

How reactionary were Trump working class voters?

But this ‘left’ strategy gets worse when we dig deeper. Being working class doesn’t automatically result in progressive attitudes, if it did we’d have won long ago, the working class is always the majority in the pyramid scheme that is capitalism. There are no shortage of working class reactionaries either scheming to get ahead in life by taking advantage of others or prone to scapegoating other more marginalised people for the real barriers they face. 22% is not a very big segment, in fact it’s small enough that it could almost entirely be composed of reactionaries, those who are the hardest and least likely to be won over.

There has been a lot of online speculation with those most inclined to put a silver lining on the Trump vote arguing that a lot of his working class voters were really voting against trade deals or the establishment. One measure of people voting for Trump despite what he has said and done is people who voted for him even though his treatment of women bothered them. The exit poll tells us almost half, 27 million, of Trumps 62 million voters were bothered by his treatment of women but still voted for him. So if we were just talking of misogyny there’d be an argument to entertain that half the working class Trump voted had done a bad thing against their better instincts.

Is this also true of his racism — did a lot of the ‘white working class’ who voted for Trump did so despite his racism? Or did they vote for him because they agreed with it, as an assertion of a need to ‘Make America great Again’ by enhancing the privileges expected from white supremacy?

A theoretical case can be made that because he said so many contradictory things he left space for people who saw themselves as progressives to self-justify what was actually a vote for racism. Listening to Trump by Christian Parenti provides a useful reading of the contradictions he provided in his speeches that might have allowed some to decide his racism wasn’t to be taken seriously.

At this point a lot of the left falls back on anecdote to prove their case. There are many many anecdotal accounts of Trump voters as unhappy but otherwise progressive blue collar workers or on the other hand as deep seated racists. My impression is that the best predictor as to which anecdotes any particular author relates is determined by their pre-existing approach to the intersections of class and race. I’ll look in more detail at some of this anecdotal material later but it’s clear that the stories of a few dozens votes among millions can’t answer the question as to what the nature of Trumps white working class support was.

It’s also the case that probably both stories are at least somewhat true, that’s why both sets of anecdotes exist. But we can do better than that statement, we can actually get a pretty good impression as to what extent Trump voters were straight up racists and to what extent some may have been people willing to sacrifice more marginalised people in the hope of economic advantage. The exit poll of people who voted included responses which tend to place people very firmly in the progressive or reactionary camp.

As an example 26% of voters described themselves as ‘white born-again or evangelical christian?’ From the Reagan years on in particular = these categories form the religious base of a range of reactionary politics in the US, in particular in relation to abortion and LGBT issues. 81% of those voters went for Trump. Unfortunately unlike the Ashcroft Brexit exit poll the questions as published by CNN aren’t also cross referenced against other responses. So we can’t see what percentage of people earning less than 100k went with each answer. But for the most part the Trump V Clinton differences in response is of such huge magnitudes that this wouldn’t change the story told. Incidentally I’d love to examine that breakdown if anyone reading this happens to have access to the data.

View of building a wall along the entire mexican border

Take Trump’s plan to build a wall along the entire Mexican border, it would be very hard to see anyone who supported such a scheme as being progressive. They clearly would be hard people to win over to a socialist view of the world. In fact 41% of those voting said they supported the wall and a massive 86% of these voters voted for Trump. That 41% is 55 million voters, 47 million of whom were Trump voters.

If their spread across his vote is not affected by class then 64.7% or 30 million of his 40 million working class voters support the wall, leaving only 10 million working class Trump voters who don’t support the wall or didn’t respond. 10 million is a little over 5% of the US working class. There are more undocumented migrants, about 11.4 million, than working class Trump voters who do not favour building the wall. If the left is going to listen to the concerns of the working class which bloc should we prioritise?

This illustrates another major problem with the ‘forgotten white working class’ left narrative, the inbuilt tendency to not see other sections of the working class so that the interests of the working class as a whole somehow become represented behind the figure of a white male worker who is taken as typical. Undocumented migrants are overwhelmingly going to be working class, in 2007 their median household income was 36,000. A later study published in 2013 estimated average undocumented migrant household income earnings state by state as averages from 23,000 for Kentucky to 44,000 for Alaska.

Deportation of undocumented immigrants working in the u.s.

A related question was whether these undocumented migrants already working in the US should be given legal status or deported. 25% of voters wanted them deported and 84% of these voters opted for Trump. So that’s 33 million voters wanting deportation and 28 million of these being Trump supporters. This would translate into 18 million of those 40 million working class Trump voters leaving 22 million working class Trump voters who are not openly racist on this issue. In this case that’s about 12% of the working class.

Does the country’s criminal justice system treat black people fairly

43% of voters thought that the criminal justice system treats all equally which is 58 million. 74% of these were Trump voters, which is 43 million. In this case 28 million of his 40 million working class felt all were treated equally by the criminal justice system, that at a time when police shooting and the #BlackLivesMatter protests have made the alternative viewpoint very visible. Again the number of working class Trump voters willing to recognise that the criminal justice system might be a tiny bit racist, 12 million, is roughly the same as the number of undocumented migrants.

Should the next president be more conservative

Perhaps most directly 48% of voters thought the next president after Obama should be more conservative and 83% off these opted for Trump. That’s 64 million voters, 53 million of whom voted for Trump and 34 million of whom were probably working class voters. That leaves a rather tiny 6 million working class Trump voters who do not want a more conservative president, just about 3% of the US working class. There are 2 workers facing deportation under a Trump presidency for every potential working class Trump voter who didn’t want a more conservative president.

Voters could also respond to that question saying they wanted a more liberal president. Only 17%, the equivalent of 22 million did but it is interesting that 23% of these were Trump voters. This is probably the most direct measure provided of how many ‘progressive’ Trump voters there were who voted for him despite his racism & sexism, there were a little over 5 million of them total and the working class component would have been a little over 3 million. Three million is a lot of people but it’s also not even 2% of the US working class.

Should the Trump voting working class really be our priority?

This really doesn’t leave much space for a left intervention directed at working class Trump voters as most of them are ideologically hostile to progressive politics. They are not progressives who voted for him despite his racism, they overwhelmingly agreed with his racism.

When sections of the left tell us we can’t assume all white working class Trump voters are racists they can’t be talking of those who want to build the wall or support mass deportations. So at best there appear to be 3–12% of the working class population of the US who voted for someone putting forward racist policies on grounds of economic self interest but may not have agreed with his racism. Yet a growing cacophony of voices on what I’ve called the Nostalgic Left insists that in order to win this section over the rest of the left has to abandon any major focus on other sections of the working class, including the equally sized segment that Trump was promising to deport!

It’s a minor aside but the Trump voting segment of the working class may also be the hardest demographic for the left to reach. As we’ve seen they are disportionately of retirement age meaning they are not even clustered in workplaces but they are also not in the cities. Only 16 million of Trumps voters lived in the cities as against 14 million in rural areas even though twice as many people live in the cities in the US. Most people leave in suburbs which are very often very segregated and most of Trumps voters were in the suburbs, 33 of the 62 million.

Much of rural America is very, very white and in that context the rural v urban divide in the vote provides a somewhat positive message. Trumps proportionally far larger rural vote may well indicate how much easier it is to sacrifice the interests of your fellow workers of colour when you don’t actually know many of them. In the cities, where workers were more likely to know those facing the racist reality of a Trump victory, it appears they were much less likely to ignore that cost to others in the hope of benefit for themselves. This is a common pattern in other countries where the far right vote tends to concentrate in areas where migrants are few or very recent.

This may also explain the strong link between low levels of formal education and voting for Trump. Higher level education institutions tend to be relatively diverse and often if you live in a rural area but have a degree this means you moved to a city in the past to obtain it. So rather than voting for Trump being a measure of stupidity, the liberal assumption from the education figures, it may instead once more illustrate that it’s much easier to sacrifice people you’ve never had any contact with in your own self interest than those you’ve studied beside.

Was the rust belt different?

The left has paid the vast majority of its electoral analysis attention to the ‘rust belt’ states and especially the ex coal mining belt of West Virginia. In part because some of this area swung the electoral college vote for Trump and in part because it provides the most fuel for the idea of an otherwise progressive ‘white working class’ that opted for Trump. Indeed when you dig into the numbers it’s very likely that this very small group nationally was concentrated in these areas and may well have given him the election in the key swing states.

Here the Republican vote increased in real terms (elsewhere it was pretty flat) and the Democratic Party vote fell further than elsewhere. This has been interpreted as white workers who were willing to vote for Obama — sometimes presented as proof they are not hard core racists — switching to vote for Trump. The asserted reason is that he opposes the trade deals that are blamed for decimating well paid employment in the rust belt or opposes the climate change legislation that has greatly reduced coal production.

But while this is a plausible story that can be constructed from the numbers it’s not the only one. For the most part the claim is made on the basis of previously Obama voting counties switching to Trump as if this was the equivalent of Obama voters switching. While it’s true the Democratic vote decreased and the Republican vote increased this doesn’t necessarily mean this was people switching from Obama to Trump. It could also be Obama voters staying at home while Trump energised a set of racists who didn’t vote last time around.

A close examination of the state by state exit polls in comparison with the national polls might provide clues. Do the rust belt states have a higher proportion of self described liberals voting Trump or union families voting Trump? Do Trump voters there have a better opinion of Obama? I had a peak at Ohio and the answers on what voters thought of Obama seem too close to the national average to favour the switching story. There is however more hostility to international trade 48% saying it took away jobs against 42% nationally. Another important difference is that while nationally 51% of union households voted for Clinton in Ohio it was only 42% with Trump taking 54%. Both these might indicate some Obama last time voters switching to Trump.

In the closest measure of what would actually be switchers 12% of voters in Ohio who identified as Democratic Party said they had voted for Trump as against 9% nationally. For comparison purposes the reverse republicans who voted for Hilary was 8%. In Pennsylvania it was 11% Democrats for Trump with 9% of Republicans flipping to Hilary. In Michigan 9% and 7%. These aren’t the huge percentages differences like those we saw for building the wall or deporting migrants but they aren’t insignificant and because they are in what because the swing states may have decided the election.

In the Ohio case where 23% of the voters were from union households this resolves to about 1.2 million union householders voting of whom just under 650,000 voted for Trump, 130,000 more than expected from the national average. This is not a large number in a voter turnout of 5 million but Trump only took Ohio by 400,000 so 650,000 union household votes would have been enough to give it to him.

Ohio is the example where the union household Trump vote was at its greatest in percentage terms. Where the responses to the union household question was reported on by CNN in other rust belt states Trump lost. On the income measure Trump won amongst voters earning more than 100k in all five states and generally lost or at best drew with those earning less than 100k so even in the rust belt working class voters did not give a majority to him.

Generally across the 5 rust belt states the Democratic party did lose close on 2 million working class votes in comparison with 2012 but many of these votes may have gone 3rd party or abstained rather than to Trump. The Myth of the Rust Belt Revolt looked in detail at the numbers. The small increase in the republican working class vote may include those who had previously voted for Obama but it’s also likely to be made up of of energised racists and misogynists who hadn’t voted last time but got out to vote this time. There isn’t strong data for the story that huge numbers of ‘white working class’ voters switched from Obama to Trump.

A final note on this, the argument that someone isn’t racist because they voted for Obama last time isn’t a very strong argument. Yes it indicates they are probably not a card carrying, hood wearing, KKK racist but having voted once for a black man isn’t a magic ‘not a racist’ card. It’s promotion by some on the left to excuse the Trump vote is no more than a update of the old ‘I’m not a racist but..’ to ‘I voted for Obama but ..’ In fact ‘I’m not a racist but I voted for Trump’ sounds like a ready made self-justification that many will hear at their family Thanksgiving dinner this year in the mouths of their relatives who they know to be a little bit racist.

Beware the power of anecdote

Journalistic anecdotal stories of the election do a lot to shape people’s understanding of what happened. By this I mean those countless pieces where a journalist goes somewhere and reports on what they are seeing and what they’ve been told. The problem with over emphasising these as a source is that journalists will tend to write about the most interesting story they can find. And interesting stories by definition are those that are unusual and at the extremes, so the opposite of being representative. A republican who has always voted republican and voted Trump this time makes for a poor headline grabber in comparison with a life long democratic party voter who voted Trump this time. Read that story enough times and it becomes proof of mass defections rather than what it is, the individual stories of some tens of voters in 135 million.

It is worth noting how many of the white working class people who voted Trump are described either as of retirement age or a bit younger but still described as retired because they are unlikely to find the sort of work they previously did. As with Brexit Trump’s election relied on the votes of those at and around retirement.

Coal mining country- McDowell county

The anecdotal approach is all the more powerful if rather than an individual the story can be about an entire geographic division. And sometimes with the smaller subdivisions you find results that do seem to reflect the Democratic voters becoming Trump voters interpretation. The most convincing one I‘ve found is for McDowell county, the southernmost county of West Virginia.

McDowell County is coal mining land which means, as with most coal mining regions in the global north, it’s story is of huge numbers of once well paid if dangerous jobs vanishing to be replaced with poverty, unemployment and alienation. The fossil fuel industry has done a powerful PR job in blaming this on Climate protection laws although the reality is that by the time any climate protection with teeth was introduced employment had already been devastated by automation and related new technologies to the extent that renewable energy tends to offer considerably more employment, but not in the same places or for the same people. Under neoliberalism the old mining communities were more or less abandoned.

McDowell County is typical of this story, the population peaked at 100,000 in 1950 but with the collapse of coal mining the population has also collapsed to only 20% of that peak. There is large scale unemployment with over ⅓ of the population below the poverty line and the 2nd lowest male life expectancy of any county in the US at 63.5 years.

In McDowell 2600 people voted in the Democratic Party primaries as opposed to only 860 in the Republican one but in the election itself the Trump/republican vote was just under 75% of the total. The story of Democratic voters switching appears all the stronger when you dig into the detail of the primary and see Saunders won the democratic primary by a very large margin indeed he had double the vote for Trump in the republican primary.

It gets a bit more complex though once you dig further into the story. The Democratic Party vote has declined election by election and the county flipped not in 2016 but back in 2012 when the vote went 64% republican. Trump did add 10% to the republican share but in the context of a very low turnout of 36%.

In 2012 2086 votes went to Obama, in 2016 this was just under 1429 to Clinton so there are 600 ‘missing’ votes to allocate. The Libertarian Party candidate Johnson got 90 which is perhaps some of those were the ‘missing’ 600 democratic party votes, the rest are therefore Democrats who stayed at home and Democrats who voted Trump.

At least 1100 of those who voted in the Democratic party primary didn’t vote Democrat in the election, did lots of Sanders voters opt for Trump in the hope of restoring employment or did they stay at home? As Sanders got 1488 primary votes to Hillary’s 817 and together that’s 86% of primary votes cast if you want to go with the Democrats voting Trump interpretation it has to mostly be Sanders voting Democrats switching to Trump for the election. Trump was promising to get coal mining going so a common left interpretation (often from afar) is that this is proof of an otherwise progressive working class voting for Trump on economic grounds.

There are two problems with the story. The first is that while McDowell used to be the heart of mining country today almost no one works in the industry, both because of closures and because modern mining no longer requires huge numbers of workers. The second is that Sanders wasn’t promising to reopen the mines. An anarchist familiar with the area told me it was much more likely that gun control was the key difference between Sanders and Clinton and the key common ground between Sanders and Trump. Unusually for the Democratic Party Sanders does not favour major restriction on gun ownership, he is elected from Vermont which is a rural state with a high rate of gun ownership and during the primaries Clinton repeatedly attacked him on just this issue. The story of white workers voting for the racist because they wanted to hold onto their guns is likely to have less purchase for a left seeking an ignored ‘white working class.’

Can the left win over Trump voters?

At this point it’s useful to remind readers that the purpose of this piece is not to give advice to the Democratic Party or to answer a ‘what Clinton should have done’ question. If it was then the nature of the winner takes all college system means the small percentages of rust belt workers might have been enough to tip the vote if Clinton had orientated towards their concerns. But that’s not my interest, my interest is whether they formed a large enough segment of the working class to take seriously the argument that listening to their concerns should be the primary concern for the left. My conclusion is no, for every worker who voted for Trump holding their nose to his racism there were four that embraced that racism but far more importantly 15 that didn’t vote for Trump, most of whom stayed at home. In numbers and attitude it makes far more sense to concentrate on that last group rather than sifting through the Trump tailings hoping for the gleam of the occasional overlooked progressive proletarian nugget.

There are some nuggets there so that said it makes sense for the left to make use of opportunities to win Trump voters over in the course of struggle. In particular if you are already living in smalltown USA it’s possible you will have little choice other than to focus on and win Trump votes in order to build anything that can last. In the past white supremacy coupled with a ‘red-scare’ has been very successful at smashing working class organisation which hadn’t inoculated itself against the reassertion of white privilege.

In particular workers with some reactionary opinions are often rapidly moved left if they are involved in strikes or other struggles that involve conflict with capital. So I’m certainly not presenting these numbers to argue against using making use of such opportunities, where and when they arise. Nor am I arguing against supporting comrades already living in areas where Trump voters are concentrated or working alongside them in their efforts to erode what is at best an indifference to racism.

The British film Pride provides an excellent example of how struggle can shift opinions radically.. If you haven’t watched it make sure to do so but essentially it’s about how during their 1984 protracted strike ‘Lesbians & Gays Support the Miners’ overcame homophobic attitudes amongst British miners to the level that Miners unions banners were subsequently sent to Pride marches.

But this was in the context of a titanic class struggle where the state was openly battering the miners off the streets week after week during their year long strike. And the media was publishing ludicrous hit piece after ludicrous hit piece on them. An experience that it would be almost impossible to go through without questioning all repressive aspects of the society around you and being open to anyone expressing solidarity. That sort of opportunity to challenge preconceptions though solidarity in struggle at that scale is exceptional. Finally while British coal miner in the 1980s may not have started off with great views of the LGBT community it’s not like they were open reactionaries across a range of issues in the manner of Trump voters. Finally those who were hardened reactionaries were where the ranks of those who scabbed on the strike were initially drawn from.

Outside such struggles how do those on the left advocating ‘take the concerns white working class Trump voters seriously’ see it working out. Our plan after all can’t start with ‘Trigger a massive year long strike with multiple police attacks on workers and a transparent media conspiracy against them’. Would the strategy be sending college students from the cities to the rural areas to leaflet the houses of the workers in their 50’s and 60’s who voted for Trump? It’s very hard to see that delivering very much, indeed it could very well be counter productive.

The marxist concept of ‘false consciousness’ is a barrier here as it often results in the idea that all that is needed is to expose such workers to reality to dispel these ideas. In fact white supremacy ties them into reproducing our racist reality because historically it rewards them in ways that are not just economic. What was the America Trump wanted to make ‘Great again’?

The cost of winning over Trump voters?

I am presenting the figures above to say a strategy of seeking out and convincing working class Trump voters would be an enormous waste of resources that could much better be used to convince and organise the 240 million who did not vote for him. And further that a Trump voter outreach strategy would be very very vulnerable to making compromises that would re-enforce the very problem of white supremacy and misogyny that helped him win the election in the first place. Beyond that a left that puts that section of the working class in the centre of its messaging will rightly lose everyone else.

As I concluded from my analysis of the similar Brexit vote the figures do not support returning to some sort of leftism centred on the feelings of the ‘white working class’ and unhindered by the considerations of ‘identity politics’. This false lesson is being promoted by the same sections of the left, what I’ve called the nostalgic left, who spun the same message out of Brexit. The disaster that is Trump was precisely born in part from the encouragement of resentment among the white working class, a sense that they were doing badly because those who once ‘knew their place’ were doing well.

His campaign slogan ‘Make America Great Again’ suggested a greatness located when exactly in a society that is only one generation removed from legally enforced racial segregation? By re-enforcing white supremacy Trump may well deliver some gains for some sections of the ‘white working class’ — if everyone else gets kicked downwards or deported then the sense of at least being at the front of the line once more will be restored.

The #alllivesmatter and #notallmen sections of the white working class need to be challenged and if need be isolated rather than normalised. Certainly reactionary workers don’t like being called racists but even more the rest of the working don’t want to be subjected to racism.
We can’t go forward together on the basis of some economic unity that involves turning a blind eye to white supremacy, misogyny and homophobia. And at its crudest this is what the Nostalgic left argument amounts to, for an example that will make you cringe see the Mother Jones article see ‘Let’s be careful with the white supremacy label.’

This isn’t to say all tactical nuance should be abandoned. On the one to one and small group level very often there is space and time to spend moving fellow workers without jumping straight into confrontation. But in many other circumstances, in particular where there are large audiences, rapidly calling out racist ideas for what they are is an essential part of isolating those trying to normalise them. Yes sometimes people are repeating racist myths out of ignorance and can be dissuaded but sometimes they are not and it’s a conscious strategy that can rapidly do harm.
.
What has to be rejected is the idea at the mass level that the left backs away from confronting prejudice in the interest of preserving what we have historically then found to be a weak class unity that lasts only as long as the high tide of struggle. That section of white workers who decided it was OK to vote for Trump on the grounds of economic self interest, even though his campaign was built on hate of people who don’t look like them, were not just terrible human beings, they also betrayed their class. After all a scab also chooses their own economic self interest above that of their fellow workers and their class. The momentary and hidden betrayal in the privacy of the ballot box can’t carry the lasting emotional impact of watching someone cross a picket line day after day but it’s still a betrayal that destroys the very class solidarity we need to build.

The changing nature of the working class

Fundamentally the position advanced by the nostalgic left that we have to focus on the concerns of ‘the neglected white working class’ has to be rejected. It has to be rejected because it is factually wrong but it also has to be rejected because it in no way can build class solidarity when it’s historic role is to destroy it. Focusing on the particular grievances of white workers is how the capitalist class created and maintained working class divisions.

For the working class to win it has to come together as a class despite the very real divisions in a common struggle to overthrow capitalism. That sort of unity demands a solidarity that is not restricted to those who look like us but that extends to everyone in the working class. In particular while white workers might well be the largest bloc in the US unity of the entire class can’t be built around the interests of the largest bloc. All the more so where there exists a long and deep history of the concerns of that bloc being deployed to undermine class unity through white supremacy.

Those sections of the left that have tended in practise to assume the interests of white workers stand in for the working class as a whole need to abandon that practise. This means getting away from an outdated approach that saw the working class vanguard as being that core of white male skilled workers located in the industralised belts of the planet. What became the rust belt in the US and the European equivalent, the Blue Banana running from North West England through Alsace, the Rhineland and Switzerland to Northern Italy.

That section of the working class had considerable trade union strength principally because before automation incredible numbers of workers were concentrated under very similar conditions in a few workplaces. Ford’s biggest factory concentration at the Rouge Complex once employed 100,000 workers, today although it’s still has the largest plant Ford plant but it only employs 6,000. The car assembly plants in particular were hotbeds of worker radicalism before the replacement of most of those workers by robots. Historically they were the ‘low hanging fruit’ for a union movement which could recruit large numbers at relatively low cost. But while that sort of focused organising did win real wage improvements for that section and some of that trickled down to other workers when it came to bigger struggles about the nature of society those sections proved to isolated too win. Even if at times like the British miners strike they put up incredible fights.

At the end of the 1960s and into the 1970s these huge concentrations of workers led many left organisations in the US to adopt ‘into the factory’ strategies. Under these they sent a lot of their members into the Detroit car plants in particular in order to try and radicalise large concentrations of workers whose working conditions already gave them power in confrontations with capital. This paid large dividends in the short term with Detroit becoming a hotbed of radicalism but in the medium and long term it isolated the left and capital was able to move the factories to the anti-union southern states.

‘Identity politics’ in the good old days of industrial struggle

It’s worth pointing out that this organising was not as the nostalgic left appear to imagine free of ‘identity politics’. Mass migration of both black and white workers from the south and racist hiring practises meant the car plants, the unions and the communities where the workers lived were dominated not just by pure economic struggles but also by the need to organise against white supremacy. At its height this saw a wave of wildcat strikes largely organised by black workers and opposed by a white union leadership in 1969 and in 1970 to the formation of the League of Revolutionary Black Workers.

In 2008 I spent a couple of days being shown around Detroit by an anarchist who had been active in the early 1970s. As well as telling me of the strength of the left he talked of the strength of the white supremacists of that period, often in the organised form of the KKK and the battles that took place when black families moved into white communities. Hundreds of racist pipe bombings of black family homes took place as the white supremacists fought to maintain segregation, sometimes with significant local support from white workers. I visited many of the major rust belt cities on that trip, staying with and talking to anarchists in each of them so the idea that there was a significant racist presence in the white working class wasn’t news.

Indeed for all the hatred the Nostalgic left has of the Intersectional left any examination of US history shows that ‘Identity politics’ developed out of the shortcomings of the left they are nostalgic for. The left’s failure in that period led to defeat of the working class in the Reagan period as he mobilised white supremacy as a counter weight to the influence of the left and the unions. Regan expressed that through the ‘war on drugs’ and a generalised panic about crime. The later development of the intersectional left is an attempt to put the pieces back together in a way that tackles white supremacy from the onset, rather than leaving it as a rotten foundation which will later bring down any new building.

The old (‘white’) working class identity a barrier to the ones we need

In any case conditions have changed, the mass factory where thousands or tens of thousands of workers carrying out near identical tasks have been abolished by automation where much smaller numbers carry out more differentiated work. In the traditional core this process has almost been complete at the level where tiny fractions of previously mass workforces produce as much steel or as many cars as hundreds of times more workers did in the past. Elsewhere, particularly in ultra low wage economies, mass factories still exist but the left that is concerned with the interests of the ‘white working class’ has not uprooted itself to the garment factories of Indonesia.

The working class in most of the global north and in particular the USA today is incredibly diverse and at the same time very fragmented in terms of workplaces and practises. This is not going to be overcome through Nostalgic left call’s that amount to recentering the white male industrial worker of the past. There is a need for the creation of a new class identity, a class identity that contains many identities. The needs exists because the revolutionary transformation of society requires that we move from a position where we are objectively workers because of our relationship to the means of production to where we also define ourselves as working class because of our common networks of solidarity & struggle. The networks create that identity and in turn the identity creates the networks. We already are workers but forging common identities as workers does not automatically arise from that, it has to be created, just as it was in the past.

We need to once more become aware of ourselves not as a stale throwback to a rose tinted past but as the present reality, a reality in opposition to the rulers whether they be the neo far-right around Trump or the neo-liberal elite around Clinton. Those identities will be forged where class exploitation intersects with the myriad of oppressions both the neoliberal and far-right false choices inflict on us, not just today but across our shared history.

It will be forged in opposition to much of that history, a history where as often as not they used the tools of white supremacy and misogyny to ensure some of us could be persuaded to police the rest. Those who would act as police have always been amongst us, the attitudes of the 22% of the working class that voted Trump reveal many of them to be in that category. Indeed in a literal sense Trump obtained the backing of police and prison warden organisations. There are a million police officers and half a million prison officers in the US almost all of whom would earn under 100k.

The answer to Trumpism and the equivalent far right forces unleashed by Brexit and festing across Europe is not a return to the neoliberal status quo. That status quo was killing us slowly through climate change and ruled over a society where police killings of black men were routine and deportations were escalating, Obama almost managed to deport three million people, more than any other US president and part of a pattern of escalation that Trump will certainly build on.

What we need is a coming together to tear down what has been and what they are trying to bring into being. There is a new world in our hearts and if for now a shadow has fallen over our dreams we have the power to banish it through solidarity and struggle.