Is there really a ‘glass floor’? Or can the children of the elite be genuinely downwardly mobile?

In the previous post I explained why, in order to get a full picture of downward mobility in Britain, we need to consider the prestige of people’s occupations. I asked whether people who had apparently been downwardly mobile from advantaged backgrounds might, in many cases, have actually fallen sideways into highly prestigious jobs.

In this post I use data from the UK Labour Force Survey to try to answer this question.

The Labour Force Survey

The Labour Force Survey (LFS) is the UK’s largest labour market survey – interviewing around 40,000 households every quarter. In 2014 the LFS added a set of questions about parental occupations – specifically the occupation of the ‘main wage earner’ in the household when the respondent was 14. From 2014 onwards, this makes the LFS the largest, most representative source of data on occupational mobility in the UK.

The parental occupation questions are only asked in the Summer wave of the survey (see here for an explanation of the LFS wave structure). So I combined the Summer waves of the LFS from 2014 to 2019 (stopping before the pandemic). Because of how the LFS waves work, I dropped all respondents from 2015 onwards for whom the survey was their fifth wave (because they would have been in the dataset once already). I also dropped everyone under 30 (and over 59) to avoid mis-classifying people based on temporary early (or late) career jobs. Finally, I dropped the small number of people for whom the primary wage earner in their household at age 14 was not a parent.

In the last post, I reported the basic figures for occupational class mobility from this dataset. To recap:

• Around 40% of men from higher professional and managerial backgrounds (NS-SEC Class 1 – see the previous post for a primer on NS-SEC) have ended up in this top class themselves. Around 30% have fallen one rung, to Class 2 (lower professional and managerial). The remaining 30% have fallen further, to Class 3 or below.
• For women from Class 1 backgrounds, downward mobility to Class 2 is much more common. Only a quarter of women from the most advantaged backgrounds have ended up in Class 1 themselves, with almost 40% being downwardly mobile to Class 2. The remaining 35% have fallen to Class 3 or below.

So this is the basic class picture. But what about status/prestige? Are these people who are downwardly mobile in terms of class also genuinely downwardly mobile in prestige terms? Or are they quite often ending up in high-status, sought-after jobs – for example in the creative sector? To answer that question, we are going to need a way to measure the prestige of someone’s occupation.

How do you measure prestige?

I actually looked at three different measures of prestige in my analysis, but I’m going to focus on one here for the sake of simplicity – The Cambridge Social Interaction Scale, or CAMSIS for short (the results weren’t radically different regardless of which measure I chose). CAMSIS is a scale of social advantage constructed from patterns of marriage and friendship. It is based on the idea that jobs which often marry or befriend each other are probably quite close together in the social hierarchy. For example, university lecturers are much more likely to befriend civil servants than car mechanics. This suggests that civil servants and lecturers are at a similar social ‘level’, while car mechanics occupy a different level.

This yields a measure that is entirely separate from class as captured by something like NS-SEC (which, remember, is almost entirely based on employment conditions). There are plenty of occupations that are quite likely to associate with each other, but which have different employment conditions and therefore different NS-SEC classes. Think of, for example, journalists and lawyers. You would not be remotely surprised to find a journalist married to a lawyer – but the former is NS-SEC Class 2 and the latter is Class 1. Or, even more starkly, a civil engineer is unlikely to raise an eyebrow if she marries a graphic designer, despite the former being Class 1 and the latter being Class 3.

While NS-SEC places these occupations in different buckets, CAMSIS looks at their patterns of marriage and friendship and places them at a similar level in the social hierarchy. You will find some other examples if you look at the graphs below. These very rough graphs (apologies) show the distribution of occupations (2010 Standard Occupational Classification codes) by CAMSIS score, within each NS-SEC class. (Click on the images for the high-res versions – there’s still a lot of overlapping text but you should be able to pick out the occupations at the top and bottom end of each class).

Occupation CAMSIS scores by NS-SEC class – Women

Occupation CAMSIS score by NS-SEC class – Men

You’ll notice that I’ve been using words like ‘advantage’ and ‘stratification’ here, rather than specifically ‘status’ or ‘prestige’. That’s because the CAMSIS authors very explicitly do not claim that CAMSIS measures occupational prestige. They only say that it measures social stratification in a much more non-specific way. Nevertheless, Tak Wing Chan and John Goldthorpe (among others) have made a compelling case that the patterns of marriage and friendship that lie behind CAMSIS (and behind their own Chan-Goldthorpe status scale) primarily reflect equality and inequality of social status. Their argument, essentially, is that marrying or becoming close friends with someone is a good sign you consider them your approximate equal in status terms. It is very hard to establish and maintain an intimate relationship across a large gulf in social status – and most people don’t really want to.

And so, in common with a lot of other research, I used CAMSIS as a direct measure of occupational prestige. It’s far from perfect, but it is close to as good as it gets, given how difficult it is to measure ‘prestige’ objectively.

Genuine downward mobility among the children of the elite

So what do the CAMSIS scores of the downwardly class mobile tell us about the social positions they have ended up in? Are they really falling down the social hierarchy compared to their parents, or are they just falling sideways? To my surprise, the data seem to favour a story of genuine – rather than simply technical – downward mobility.

You can see this in two main ways. First, by comparing the CAMSIS scores of those from NS-SEC Class 1 backgrounds who i) ended up in Class 1 themselves, versus ii) ended up in Class 2 or below (i.e. were downwardly class mobile). I expected that those who had fallen below Class 2 might have suffered some genuine loss of prestige. However, Class 2 includes many high-status jobs – almost all creative and media workers are in Class 2; as are teachers, therapists, and PR and advertising professionals. Plenty of opportunities here for the downwardly class mobile to fall sideways (or even up).

Those who had ‘fallen’ from Class 1 backgrounds to Class 2 destinations were slightly more likely to enter, for example, the creative professions. However, they still had meaningfully lower status jobs on average than their peers who had remained in Class 1. The average CAMSIS score for a man from Class 1 origins who had a Class 1 job themselves was 66. Men who had fallen from Class 1 to Class 2 had an average score of 61. For women, these figures were 71 and 60, respectively. To put this into context, a difference of five CAMSIS points (men) is the same as the difference between a chartered accountant (67) and a librarian (62). A difference of 11 points (women) is roughly the same as the difference between a musician (71) and a special needs teacher (61). These are not precipitous falls, but they are not negligible either.

My second approach was to examine whether people who had been downwardly mobile from Class 1 had a status advantage over their peers in their destination class. For example, do those who enter Class 2 from Class 1 backgrounds have higher status jobs than those in Class 2 whose parents were Class 2 or below?

Here the story is a little more complicated. For men and women from Class 1 backgrounds who had fallen to Class 2, there is really no clear pattern. Whether you see anything depends very heavily on how you slice the data. From some angles it looks like there is a slight status advantage for people from Class 1 (versus Class 2) backgrounds, but it is certainly nothing you could hang your hat on.

For people in Class 3 or below, the picture is a little clearer. It does seem that people from Class 1 backgrounds really are occupying higher status jobs in these classes than their peers who are ‘class natives’ – but again the differences are not very large.

The most substantial and consistent differences are when the downwardly mobile from elite backgrounds are compared with those who are upwardly mobile to a given destination class. For example, those who are downwardly mobile to Class 2 have higher status jobs on average than those who have risen to Class 2 from backgrounds in Class 3 or below.

The overall picture here is muddy. But we can confidently conclude that those who are downwardly mobile from the top of the class tree are not falling, en masse, sideways into prestigious jobs. They may fare a shade better than their peers in their destination class – especially compared to those who have been upwardly mobile from lower class backgrounds – but they have, on average, genuinely moved down the social hierarchy.  

We can, however, isolate a group who have fallen sideways. These are people from Class 1 backgrounds who have ended up with highly prestigious jobs (in the top 15% of the status distribution) despite being downwardly mobile in class terms. They are a minority, but not an insubstantial one – representing about 15-20% of the downwardly mobile from Class 1. This group are much more likely to be the children of traditional professionals (architects, lawyers, doctors, academics) than the other jobs in Class 1 (senior executives, ‘New’ professionals like software developers or project managers, etc.). They are also, notably, substantially more likely to be white or of mixed ethnicity. People from Class 1 backgrounds who are of Black, Bangladeshi or Pakistani ethnicity are much less likely end up in the highest status jobs.

So is there a ‘glass floor’ which protects the children of the elite from substantial downward mobility? In one sense yes: the majority of children from the most advantaged backgrounds either won’t fall down the class ladder at all, or will fall only a short distance. They are largely protected from precipitous falls into more disadvantaged class destinations. However, this glass floor is neither as robust, nor as high-up as I expected. A small minority of the downwardly class mobile do end up in highly prestigious jobs. However, on average, the downwardly mobile children of the elite are not particularly distinguishable from their peers in their destination classes.

Obviously, there are other ways in which children from the most advantaged backgrounds are protected from the vicissitudes of the economy and the labour market – for example, help with buying a home, or just ongoing financial support more generally. But status is important in and of itself. You can be living comfortably in a home your parents paid for, but if your job is looked down on compared to the jobs your friends and family have, you will still feel a real sting of downward mobility.