Is there complete social mobility over time?

Over two generations, socioeconomic class tends to be ‘sticky’. Statistically speaking, the occupation of a parent influences the occupation of a child. But what about the very long term?

Robert Wiblin has drawn my attention to this very interesting paper by the economic historian Gregory Clark, which argues that over multiple generations there is a class ‘regression to the mean’, with the inequalities of one generation washing out over time.

Clark’s method is to use English records of surnames, which can be used to roughly trace the class progress of people with different family names. Some surnames reveal class backgrounds because they are taken from medieval occupations (eg Smith, Clerk/Clark, Shepherd, Cooper, Carter). Clark furthers his study by examining the names in records of wills, tax payments, and court appearances. Over time, the share of names appearing in lists of those with large estates or criminal defendants can roughly track class progress.

What Clark finds is that England over the 800 years from 1200 was without persistent social classes. The handful of aristocratic families who can trace their family trees back centuries are outliers.

I will be very interested to see what other historians make of Clark’s thesis. Over the last 250 years I find it intuitively reasonably plausible that there are very high levels of multi-generational class mobility. The modern industrial world created many opportunities for social and economic advancement, and relatively speaking devalued the traditional source of wealth, holdings of agricultural land. But high mobility in the earlier period conflicts with my (admittedly superficial) understanding of the pre-modern world.

23 Responses to “Is there complete social mobility over time?

  • 1
    Peter T
    February 12th, 2010 09:36

    Very interesting paper – and I think Clark’s A Farewell to Alms an excellent book.

    But I would not buy the argument without further tests. For one thing, Clark does not consider the effects of inheritance through females – the surname changes but the wealth goes on. Nor does he consider upward adoption of names. Or, in the Irish and Highland Scots cases, Anglicisation of names.

    David Cannadine (Decline and Fall of the English Aristocracy) shows major swings in the open-ness of the elite – more open in the 18th than the 19th centuries for example, and a strong tendency to concentrate more wealth in fewer hands (an astonishing 75% of English land was held by a small fraction of the population by 1850).

    In short, the history is rather messier than Clark’s methods will allow. But still a good read.

  • 2
    Andrew Norton
    February 12th, 2010 09:58

    ‘Clark does not consider the effects of inheritance through females – the surname changes but the wealth goes on. ‘

    How prevalent was primogeniture in medieval England? That would keep wealth with the surname. On the other hand, it would also explain downward mobility, as second and later sons received little or no inheritance.

  • 3
    derrida derider
    February 12th, 2010 11:20

    Interesting, but absolutely irrelevant for any policy purposes.

    The notion that my children might be hobbled all their life by their childhood disadvantage and by class barriers is not made any more palatable by the knowledge that my great great grandchildren may have surmounted these.

  • 4
    Andrew Norton
    February 12th, 2010 11:46

    DD – True, though nobody is suggesting it is policy relevant. Clark notes some apparent contemporary examples of groups who are not very mobile.

  • 5
    Robert Wiblin
    February 12th, 2010 15:57

    Robin Hanson has a few thoughts on how this evidence gels with A Farewell to Alms:

    http://www.overcomingbias.com/2010/02/no-english-gene-classes.html

    This is a fascinating paper – but I find it hard to fully swallow such an extraordinary conclusion.

  • 6
    Michael "Lorenzo" Warby
    February 12th, 2010 16:25

    Primogeniture was very common in medieval England. Once social status became so dependent on having a certain benchmark levels of property (to sustain a castle, to sustain knightly training and equipment, to sustain dynastic power) it became very widespread very rapidly. Creating lots of second, etc sons who had to make their own careers (including going into the Church).

    A big difference between England and the Continent was that, in England, only the holder of the title was noble, the rest of his (almost always a his) family were commoners. On the Continent, the entire family were noble. The different incentives that created had major differences down the centuries (including the frequently attested as notable English/British sense of “fair play” despite being a very socially stratified society).

    DD: Property has almost nothing to do with that problem. Any Leninist state rapidly turns into one with hereditary elites. It is because we are few+specific+high investment-offspring primates, not private property. I venture to suggest no contemporary society is as dominated by the hereditary principle at all levels as North Korea (and I do not mean just the Kim dynasty at the top).

  • 7
    Terry Barnes
    February 12th, 2010 19:10

    An interesting point Andrew. I’m reminded of when Sir Alec Douglas-Home was asked about being called the 14th Earl (of Home) by Harold Wilson and he said, “I suppose when you think about it he’s the 14th Mr Wilson”.

    I’m not sure that Clark’s basic thesis is watertight, but you and he touch on perhaps a better historical mobility indicator, and that’s looking at the English and Irish nobility (less so the Scottish) and the lengths of the creations. The vast bulk of hereditary peers and baronets would have titles that date back 150 years of less, and many of these families came from very humble or ordinary origins before someone distinguished themselves.

    And of course, those with money but no breeding (so to speak) have for centuries bought their way into British society, gentrifying their descendants along the way. Even newspaper barons acquired titles and status. Wealth is a great leveller, whether acquired by effort or inheritance.

    We also tend to forget that in England the equivalent of the Roman novus homo runs strongly through mediaeval and modern history, as ability even then was a trait highly valued, if often resented by a man’s social “betters”. An excellent example is Thomas Cromwell, Henry VIII’s de facto prime minister, who was of vague but rough and humble origins. Hilary Mantel’s Booker-winning novel Wolf Hall compellingly tells his story.

    In other words, Australian society isn’t the clean break from class-ossified Britain the Manning Clark mythology would have us believe. I’d also add that Michael is technically correct as saying that only the legitimate holder of a title and his wife are officially “noble” in an aristocratic family (indeed, the heir to an English peerage with a courtesy title can sit in the Commons before succeeding), but culturally the whole family traditionally was treated as noble. Of course, as a recent poll showed, everyone in the UK now thinks themselves as middle class, or aspiring to be! Jolly good show!

  • 8
    Andrew Norton
    February 12th, 2010 19:36

    Terry – Your knowledge of English social history far exceeds mine. For more recent times, I know that people (including my former CIS colleague Peter Saunders) have argued that at least post-WW2 Britain was essentially meritocratic, with ability as a young person by far the best predictor of life outcomes. The presence of people with fancy titles gives a misleading impression of how class-ridden Britain is.

    Nevertheless, at least early Australia was a big opportunity for people at the bottom of British society to advance themselves far further in one generation than would have been likely at ‘home’.

  • 9
    skepticlawyer
    February 12th, 2010 20:13

    Class in Britain has always been subtle and complex, with a great deal more to it that just inherited wealth. Much of it revolves around such diffuse things as taste and education, which led — among other things — to wealthy industrialists sending their children to Oxbridge and wanting a spot in the Lords as much as it did to the production of more industrialists. That said, people noticed from very early on the English ‘fair play’ and the relative lack of control English people exerted over their children’s choice of marriage partner, as long as the person came from (broadly) within the same class.

    A good read on the above issues is Kate Fox’s ‘Watching the English’, which among other things draws interesting comparisons with aspects of ancient Roman and Japanese society.

    What can I say? The English are different.

  • 10
    Michael "Lorenzo" Warby
    February 12th, 2010 20:43

    Terry is actually just restating my point. Yes, England was very socially-stratified, but law operated in a significantly different way which also had social effects. Interestingly, Venice was also a country with equality before the law despite the sharp difference between citizens (i.e. the nobility, who could vote, etc) and subjects (the vast majority). Venice was further renowned for the quality of its lawyers. Something about being stuck together on a bit of land surrounded by water, perhaps?
    Also, medieval society was rather more meritocratic than people grasp. Even without all those younger sons looking for opportunities, nobility and knights were functional social roles and decision-makers at all levels of society were on the look-out for talent. Recommending some clever fellow to your superiors was part of the coinage of influence. (The career of Thomas Wolsey being a case in point.) It was when nobility became less functional it tended to become more privilege-conscious and exclusory. (The notion of some medieval lord turning his nose up at a way to make a nice profit is hilarious–that is rather a C19th affectation.)
    A borough is, after all, medieval-speak for “enterprise zone” with medieval concerns for basic law and order added in.

  • 11
    Peter T
    February 13th, 2010 21:09

    Clark has failed to consider a couple of basic points, so I am not convinced by his argument. Names are simply not a good proxy for wealth over so long a period – their relationship to status changes too much.

    That said, the medieval and early modern English class structure was not as rigid as often imagined, but it was very much centred on the family, not the individual. Kinship was enormously important, not just socially but psychologically. So mobility was mostly generational, and involved kin groups and patronage networks. This worked downward as well as upward.

    The House of Lord is a bad example- there is no limit to how many lords can be created, so from the late 19th century it became a form of party patronage (mostly Tory, but also Liberal – “Lloyd George knew my father…”). The 19th century saw the top of English society – the 10,000 or so established gentry families – achieve an unprecedented degree of control over most sources of landed and state wealth in Britain – land, the Church, the Army, positions of state and the rest. A large proportion of the Commons was drawn from their immediate families. This was higher in 1850 than in 1750, and broke down only after 1870.

    Also, the more modern research on family structures and much else tends to find England being pretty much the same socially as France, the Low Countries and western Germany – not surprising in view of the close links between them.

  • 12
    Jack Strocchi
    February 13th, 2010 21:57

    Andrew Norton said:

    The modern industrial world created many opportunities for social and economic advancement, and relatively speaking devalued the traditional source of wealth, holdings of agricultural land.

    Land ie real estate is again becoming the predominant source of wealth in the western world. Two thirds of AUS wealth is held in properties, one third in equities, a reversal of the normal ratio.

    Most likely property will be transferred by inter-generational transaction ie inheritance. Which is a quasi-feudal system.

    So we are entering a new feudal era of entrenched privilege, with likely much reduced social mobility based on position, position, position esp proximity to tony schools in leafy suburbs.

    There might be some social mobility based on wealthy trust fund males marrying down to sexy chicks from the western suburbs. Cinderella style.

    But assortative mating preferences by the higher status males will probably mean that such women will be relegated to escort/lap dancer/zoo model status as casual sex providers. Whilst they breed with higher-status females.

  • 13
    Robin Hanson
    February 13th, 2010 22:07

    I’m jealous of the quality of your commentators!

  • 14
    JC
    February 13th, 2010 23:35

    What Clark finds is that England over the 800 years from 1200 was without persistent social classes. The handful of aristocratic families who can trace their family trees back centuries are outliers.

    Doesn’t that sort of indicate what ought be obvious in a sense. That we’re all in a sort of competition and you or your offspring and down the line can’t hold on forever.

  • 15
    Social Mobility in the Long-Run
    February 14th, 2010 04:42

    [...] Norton: Is there complete social mobility over time?, by Andrew Norton: Over two generations, socioeconomic class tends to be ’sticky’. Statistically speaking, the [...]

  • 16
    Andrew Norton
    February 14th, 2010 05:58

    “Names are simply not a good proxy for wealth over so long a period – their relationship to status changes too much.”

    Peter – Isn’t that Clark’s point? In a rigid class system, names should be a reasonable proxy for social position. The fact that they do not seem to be a proxy is his finding.

  • 17
    Andrew Norton
    February 14th, 2010 17:27

    “Land ie real estate is again becoming the predominant source of wealth in the western world. Two thirds of AUS wealth is held in properties, one third in equities, a reversal of the normal ratio.

    Jack – I very much doubt this is true of household wealth, ie real estate has long been the major household asset but due to superannuation it is probably not as high as in the past. In any case I’d like to see your statistics.

    Also, human capital is now far greater than in earlier years and has a much greater impact on socieconomic status.

  • 18
    Peter T
    February 14th, 2010 18:07

    Andrew

    No – primarily because names are inherited in the male line only, but also because the pattern of surname creation, adoption and usage varied widely over time and over the different regions of Britain.

    No class system is so rigid that names can be a good proxy – class is simply too fluid a concept, varying over time.

    Just one example – I saw a piece on Normandy in the 11th century, where an influx of new names was taken as a sign of greater social mobility. A detailed study showed the reverse – old families took new names, probably to emphasise their loyalty to a new regime. My own genealogy includes an estate that has been in the one family (not mine, alas) since Tudor times – but the surname has changed at least 4 times, and some names have also been kept through a couple of female inheritances.

  • 19
    Andrew Norton
    February 14th, 2010 18:23

    Peter – Though as I understand it surnames, though not necessarily their spelling, have been standard in England for many centuries. Until relatively recently, the male line has been by far the best way of measuring social mobility, since they acquired the most common public markers of class, such as occupations and typically any family wealth was held in their names. I can’t see any reason to believe that a study focused on males only would give misleading results (especially if we presume that in a stable class society a few female inheritances would mainly see wealth transfers within a class). Perhaps there is name changing going on, but this would not explain why people with ‘common’ names (for which there is presumably a low change-to incentive) rise among the well off.

    Tracking social mobility isn’t easy even with all the tools of modern social science, but provided the argument doesn’t turn on precise numbers I haven’t yet seen arguments in comments that convince me that Clark’s methodology leads to fundamentally misleading results.

  • 20
    Michael "Lorenzo" Warby
    February 14th, 2010 20:47

    Most likely property will be transferred by inter-generational transaction ie inheritance. Which is a quasi-feudal system.
    No, it is not. There was nothing much feudal about Roman law, and land transfer operated much the same in Roman society as ours. Land ownership is also hugely less concentrated in contemporary Australia than it was in either feudal England or later Roman society.
    The role of land in wealth in Australia is largely a function of regulation-induced scarcity in preferred land use. That this operates inequitably, I entirely agree, but that does not make it “quasi-feudal” in any interesting sense.

  • 21
    Peter T
    February 15th, 2010 08:19

    Andrew

    Consider two groups of people, one small, the other larger. They are mostly endogamous, but there is some some mixing. Surnames are more likely to die out in the small group, just because the chances that there are no males in one generation are higher. When this occurs, there is a chance that a some of the females will marry out. Since names in the larger group are subject to the same process (but more slowly), the chances are that the new name will be much more common than the old name. So over time common (lower class) names will replace uncommon high-class ones. Even aside from pure chance, there are some factors reinforcing this among elites – elite males were more likely to be killed in war, or suffer the consequences of poor political choices, than females or common people.

    This process is quite distinct from the inheritance of wealth, but would produce the appearance of upward mobility as measured by names. Clark’s study does not control for this.

    English surnames were in widespread use by the 14th century. But the elite changed names (cadet branches adopted locality names eg Henry Bolingbroke), minorities adopted – or were forced to take – English names, change of name on inheritance happened in the 18th and 19th centuries and so on. In short, it’s messier than Clark allows.

    Clark may be right, but I would want better evidence before taking this too far.

  • 22
    Andrew Norton
    February 15th, 2010 10:03

    Peter T – Thanks for that, though it seems a better theory for the downward mobility of the upper class than the upward mobility of the lower classes, unless we give a lot of weight to brotherless girls marrying down.

  • 23
    Peter T
    February 15th, 2010 14:44

    Partly – but it’s actually just a recipe for less common names to die out, and more common ones to spread.

    Names are not distinct – some names will be shared by elites and masses. These will propagate among elites (eg, Stewart and Butler are two venerable aristocratic names, and location names can be “from x” or “owns x”) – sideways mobility masquerading as upwards mobility.

    Also interesting that Clark did not consider the mechanisms of mobility. For the British elite, these have frequently included civil war, forcible dispossession attendant on regime change and punitive inheritance taxes. “Upwardly mobile” sounds so much better than “massacre the rich”.