A few weeks ago, as Chris kindly noted, I spent a little time hacking through the detailed report of the Conservative Social Justice Policy Group on family breakdown (pdf – 2Mb), which was produced by Iain Duncan Smith’s Centre for Social Justice, and cross-referencing its analysis and recommendations against a number of key pieces of statistical trend data obtained from the Office of National Statistics.

Why?

Well, largely because the current Tory mantra on all manner of different aspects of social policy is that all roads lead to the allegedly pernicious effects of family breakdown, which they claim they will tackle if they form the next government; even to the extent that their standard rhetorical reply to any proposals for changes in social (or criminal justice) policy by any other party has become the spectacularly boneheaded claim that whatever the policy is, it won’t work because it doesn’t, in their eyes, address family breakdown. Now, if that’s what any party is going to claim then I think it only reasonable that they produce evidence both to back up the assertion and to suggest that they understand the dynamics of family breakdown sufficiently well to deliver effective policies to address it and many the issues it raises.

I am, of course, sceptical, not only for partisan political reasons but, on a much more fundamental level, because I don’t believe that the kind of legislative and fiscal levers that are available to governments, of any political persuasion, are likely to be effective in tackling such a complex, individual and deeply personal set of issues – not without resorting to nakedly coercive mechanisms, such as a complete prohibition on divorce. Governments (and politicians) in my view, don’t possess the ‘tools’, or moral authority, necessary to intervene in such matters in any kind of meaningful or productive manner.

While researching some of the statistical evidence on trends in marriage and divorce I came across rather a curious thing.

Conventional wisdom would, I think, suggest that the earlier in their adult life a couple marries, the more likely it is that the couple will eventually divorce and if they do divorce then its likely to be sooner rather than later when compared to couples who marry when they’re rather older. Most would consider that factors such as emotional maturity and financial stability, which tend to increase as people get older, will have a significant impact on the chance of a marriage being successful; the more you have of both when you marry the more likely it is that you’ll make success of things or, at least, be in a position to sustain your marriage through the difficult times that come to pretty much all couples without heading for the divorce court.

And yet, buried in the archives of the ONS is a set of statistical evidence covering the period from 1990-2000 showing, for each year, the median duration of marriages ending in divorce in each year by the age group (in five year intervals) of woman at the time they married, data which paints an interesting and very different picture to that which conventional wisdom suggests.

What this data shows is that:

For women who married for the first time between the ages of 16 and 20, the median duration of marriage at the time of divorce in 1990 was a touch over 12 years. By 2000, this had risen steadily over the ten years that the data covered to 18 years.

In the next age group (21-25) the starting point in 1990 was around 9.5 years and this, again, rose over the next ten years to just under 12 years.

For the remaining three age groups (26-30, 31-35 and 36-40), however, not only was the starting point much lower – about 7 years – but the trend over time was broadly static. The median duration of a first marriage at the time of divorce for women who married between the ages of 26 and 40 was around 7 years in 1990 and still around 7 year in 2000.

So, according to the data, women who marry earlier in life are not only likely to stay married for longer that women marrying in the late 20s and in their 30s but, over a ten year period when the general marriage rate was in decline (and, broadly, its still falling) while the annual number of divorces moved from stability to a slow decline (it fallen by about 15% over the last 15 years or so, although the picture provided by calculating the divorce rate is rather more mixed, rising throughout the 1980s and early 90′s and between 1998 and 2003 and falling in between and since to a current level that is the lowest since 1987), the gap between the median duration of marriage at time of divorce for women who married earliest in adult life and those marrying at or above the average age for a first marriage across the whole population grew by around six years. The median duration of marriages was get longer in the youngest age group but barely altering at all for older women.

Why?

The Tories report on family breakdown neither acknowledges this data nor provides any substantive answers. The closest it come to providing anything that might account for this curious trend is the suggestion that a component of the slow decline in divorce rates since 1992 could be attributed to the introduction of the Child Support Agency which, for all its manifest failings, may – according to the Tories – have made some men a little less inclined to walk away from a marriage for fear of the financial consequences of such a decision, i.e. that male behaviour and attitudes to marriage have been altered to some small extent by the knowledge that its now much more difficult for them to walk away from their financial responsibilities towards their children (if they have any, of course) thus providing them with an incentive to work a little harder at making their marriage work.

Sorry, but I’m just not convinced either that the CSA has had such an effect, at least not one that would be statistically significant, or that, even if it has, this is necessarily a good thing for either married couples or, particularly, for children, who could easily find themselves growing up in a family home with parents who are constantly at ‘war’ with each other but trapped and unable to separate because of the financial impact of such a decision.

So how else might we account for this trend?

1. One possibility may be that what we’re seeing here are generational effects.

We’re looking at differing trends across different age groups, which suggests that women in these age groups may have different social attitudes to marriage and divorce. Except that a woman who married at 18 only to divorce in 2000 at the median duration of marriage (18 year) for her age group would have divorced at 36 years of age, having been born in 1964, while one who married at 27 and divorced in 2000 at the median duration (7 years) would have divorced at the age of 34 (i.e. born in 1966). A two year age gap is not enough produce a generational effect, particularly when there is no effective difference in median duration between women in 26-29 and 36-39 age groups, despite their being a potential age difference at the point of divorce of anything from 5 to 15 years.

2. Another possibility is that we may be seeing a purely statistical/demographic effect?

We know that the general marriage rate has declined steadily over the last 40 years and that it has declined most rapidly in the youngest (16-20) age group, which means that we’re dealing with an ever decreasing ‘pool’ of potential divorcees in which the smallest ‘pool’ is that in which women married earliest in their adult life*.

* Before this gets mentioned, we can safely use the ‘pool’ analogy here because we’re dealing only with first marriages. If one were to factor in data that included re-marriage then we would, as when considering the labour market, have to use the flowing rivers analogy or fall into the trap of relying on the lump of labour fallacy.

In statistical terms, the smaller the sample you have to work with, the more likely it is that the presence of a small number of exceptional results with skew the average, although this can be less of a factor when one uses the median rather than the mean as the basis for calculating averages. In theory, therefore, the rising trend in median duration of marriage at divorce in the 16-20 age group (and to the lesser extent in the 21-25 group) could simply be the product of the trend data be based on an ever-decreasing size of data set over time, giving rise to a statistical error. This, in social policy terms, would mean that the increase in the length of marriage in the youngest age group is no more than a reflection of the growing trend for women to cohabit with a partner before marriage, which would have the effect of winnowing out the least successful partnerships before they entered into marriage and could, therefore, show up on the statistics as having divorced.

The problem with this is that a six year upwards shift in the median duration of marriage at times of divorce over ten year is simply too large a change to be accounted for by statistical error when your looking at a steadily rising trend, rather than a short-term spike in the data, which could well be such an aberration. So its not an error, but a real trend we have here.

3. What about cultural effects that might be related to changing demographics?

Going back to the idea that we have an ever decreasing pools of potential divorcees who married between the ages of 16 and 2o over a period in which the general trend amongst women was toward later marriage – the average age of a woman marrying for the first time is currently 27.7 years of age, then one has to ask why some women still buck the trend and marry at such an early age?

One of the more plausible explanations is that many of those who do marry at such an age do so for cultural/religious reasons – there is a strong general correlation between religiosity and early marriage which stems from the fact that many, if not most, orthodox religions do tend to favour not only marriage but, particularly, early marriage for women on, broadly, what they regard as moral grounds. Factor in demographic changes in population arising out of migration from, in the main, the Indian subcontinent, and its plausible to suggest that, increasingly, the pool of women marrying in their teens is being dominated by those who marry at such an age for cultural/religious reasons and, therefore, that both the longer median duration of marriages before divorce in that age group and the rising trend over time are the result of such marriages being buttressed by both strong religious and cultural views about the nature of marriage and equally powerful cultural, religious and social injunctions against separation and divorce.

This is certainly a possibility although, again, the scale of the rising trend over such a relatively short period of time suggests that even if this has a contributory effect it is still unlikely to be large enough to account for the entire trend, particularly if one considers that the buttressing of marriage by such factors may well be strong enough to negate the possibility of their producing a significant statistical effect in this data by severely limiting the numbers who do eventually divorce and, therefore, turn up in the statistics. In simple terms, those who do marry for life and stay the course will have no effect at all on statistics that measure the duration of marriages at the point of divorce.

None of these factors seems likely to account for anything but a very modest variation in median duration of marriages at the point of divorce, one that would be nothing like as large as that revealed by the data, which means that we have to look for something else to account for this trend, something that could produce the observable difference across both the different age groups and in the specific age-related trends over time – but what?

What we can say for certain is that whatever it is, it ceases to have any significant impact once a woman reaches her mid-20s, after which point both the median duration of marriages that end in divorce and the trend over time are stable at around 7 years.

We can also say that its something specific to women, because the trends we’re looking relate only to women, that that it must be something that makes younger women different, somehow, from older women and also that it must be related to some sort of wider general change in demographics and or social conditions and attitudes.

Looking at a range of different data, the most plausible explanation for this trend lies in the realm of economics and in changes in the UK economy and labour market over the last 40 years.

The big economic shift in the UK economy over that period has been the decline of Britain’s manufacturing base and the growth of service industries. What was once primarily a major manufacturing economy has, over time, turned into a service economy, changing the nature of the labour market in the process to one that, in terms of its demand for certain skills, particularly ‘people/social’ skills, is widely held to be much more favourable toward women. Add into the mix a measure of government intervention, in the form of legislation covering equal pay and gender discrimination in the work place together with changes in social attitudes toward women and employment and the net result is that far more women are, today, economically active than they were twenty years or so ago and, therefore, far less dependent on marriage (or co-habitation with a partner) for their economic well-being. Even allowing for gender difference in working patterns – around half of all women in employment are working part-time compared to only 1 in 6 men – in 1985 there were 2 million more men in work than there were women where, today, the numbers in work are about even.

This has had the effect of, quite obviously, reducing the financial penalties associated with divorce/separation that women, generally, face when considering whether to stick with an unhappy marriage/relationship or move on and chance their arm at finding happiness elsewhere. It doesn’t remove those penalties entirely, or course, but it does reduce them sufficiently to make divorce a much more viable option for women than it used to be. It’s a simple enough equation, greater economic independence for women equals less dependence on marriage (and men, of course) for their economic well-being, reducing the financial disincentives associated with divorce.

Access to welfare benefits, is course, part of this equation and the part that many on the political right tend to emphasise most in criticising what they term ‘welfare dependency’. However its not at all clear that the available data supports such a clear contention. For one thing, the Tories attempts to put the squeeze on welfare payments to lone parents during the 1980s appears to have at best a very limited effect on both the general divorce statistics and on the median duration of marriages. One might argue that it could be a minor factor in both the stabilisation of and reversal in the general divorce rate from the early 90s onwards and also, perhaps, contributed to the rising trends in median duration of marriage at divorce in the two youngest age groups.

However, as the trend data shows, if such effects are present then they are strictly confined to the two lowest age groups (i.e. women marrying before their mid-20s) and it doubtful, even, that one could argue for that when one considers that the introduction of tax credits, which the Tories (and Frank Field) argue positively favours lone parents and should, therefore, favour divorce by reducing the financial disincentives attached to it, produced only a short-lived spike in divorces between 2001 and 2004, since when both the number and rate of divorces has fallen sharply back below pre-spike levels. The most one can say about tax credits is that, like the majority of financial levers open to government, it may have generated a short-term effect but otherwise has failed to create any significant changes in long-term trends.

By way of contrast, the impact of increasing levels of economic activity and, therefore, economic independence amongst women can be seen in the long term trends and, more to the point, also provides a plausible explanation for why the age-group related trend data on the duration of marriages ending in divorce flies in the face of conventional wisdom’s suggestion that later entry in marriage provides greater prospects of success.

The argument is a fairly straightforward one. Women who marry (and have children) at an early age, particularly before the age of 20-21, leave themselves very little time to build up any significant personal economic capacity (i.e. educational qualifications, skills, workplace experience) and are consequently considerably more dependent on their marriage for their own economic well-being, and that of their children, than women who delay entry in marriage and motherhood to their mid-to-late 20s.

This accounts for the marked differences between both the median duration of marriages at divorce between the respective age groups. Women who marry at a younger age are less able to support themselves and their children, without relying entirely on welfare benefits, should they divorce and are, therefore, much less likely to seek a divorce, even if the marriage is an unhappy one due to their limited, if not near non-existent ability to offset their financial penalties associated with divorce by their own efforts. By way of contrast, women who marry later, and after having built up their economic capacity in terms of educational qualifications, skills and workplace experience, are both less dependent on marriage for their economic well-being and less susceptible to the financial disincentives that come with divorce and, therefore, less constrained in their options should the marriage begin to fail. They may also have married ‘better’, of course, and have greater expectations of financial support from their ex-husband, which will go some way further towards reducing the financial disincentives of divorce.

It also, when one looks at general economic trends, offers a plausible explanation as to why, during the 1990s, the median duration of marriages ending in divorce rose significantly in the youngest age group, while the trend in older age groups remain static.

One of the obvious consequences of moving from a manufacturing to a service-based economy is that this significantly alters the types of skills and experience that are most in demand in the labour market, favouring not only people/social skills but also knowledge-based skills and experience, many of which are likely to be obtained via formal education and training which takes place outside the workplace. These are precisely the kind of skills that women who marry and have children early in their adult life are less likely to acquire than those who marry/have children later in life.

This view is consistent with research from New Zealand which shows that amongst women who do combine motherhood with a job/career, the age at which a woman becomes a mother has only limited impact on their income due to the effects of part-time working – those in better paid employment tend to work fewer hours to maximise the time they have with their kids than those in lower-paid jobs, who work longer hours at the expense of losing time with their children in order to maximise their income – but has a major impact on both their educational attainment and the types of jobs that are open to them. Women who marry/have children in their teens are less likely to complete their education or go on to higher education, less likely to return to education in later life, which translates into to taking on low paid jobs and working longer hours than their better educated counterparts.

This, in a service economy where demand is skewed towards skills gained through education and non-workplace training, puts those women who leave the labour market to marry and have children at an early age at a significant disadvantage compared to those who delay those decisions until their mid 20s and beyond, and, again, causes those who marry early in their adult life to be more dependent on their marriage for their economic well-being than those who marry later, although what the data also tells us that once a woman reaches her mid 20s these dependency effect disappear and there are no general economic advantages to be gained from further delay.

For women, the lesson here is that the first 9-10 years of adult life are critical in determining both their capacity for economic independence and, by extension, their ability to escape from a marriage or committed relationship that turns sour without finding themselves trapped by their economic circumstances and their dependence on the marriage for their, and their children’s, economic well-being.

The lesson for politicians and policy makers is altogether more problematic, especially when they set themselves the objective of ‘tackling’ family breakdown because what the evidence tells us is that, given the kinds of levers they have at their disposal, the kinds of policies they could adopt that would have an impact are those which limit or reduce women’s capacity for economic independence and which force them out of the labour market and into situations where they are so dependent on a marriage/relationship for their economic well-being that its becomes near impossible for them to contemplate divorce for fear of the severe financial penalties that go with it.

Put simply, if you have only the levers of legislation and financial incentives/disincentives to play with – and that’s all that any government has to hand – then the only policy you can adopt that stands a chance of operating to prevent families breaking down is one based on the old misogynistic adage, ‘keep ‘em barefoot and pregnant’; a policy that, when women are a key and hotly contested electoral demographic, would amount to the most spectacular act of electoral suicide in modern political history.

Having read IDS’s report in full, one of the most intriguing things about it is that, at several points, not only is there the evidence that should lead you to precisely this conclusion, but the analysis and argument on offer almost tiptoes right up to the point of admitting that the changing role and status of women in the labour market and their growing economic independence is a major causal factor in family breakdown – because women no longer have to tolerate an unhappy marriage for purely financial reasons, naturally – only for the report to shy away from stating the obvious at the last minute either by changing the subject entirely or by inserting a claim that the use of financial levers would impact on male behaviour and male attitudes, as they did in trying to link the CSA with the falling divorce rate post 1992.

Now, obviously , I’m not about to advocate such a policy myself, not least because I would view it as morally and ethically reprehensible, unremittingly reactionary and a return to the dark ages…

…but then I’m also not running around blaming all the social ills of the world on family breakdown while simultaneously trying to suggest that I have policies that will do something about it and that will work to prevent it happening – the only viable suggestion the Tories actually have is that of providing couples who do hit a rough patch with greater access to counselling, which is less a policy than an admission that organisations like Relate are badly underfunded.

Instead what I hope I’ve demonstrated here is that the ability of governments to intervene in such matters to good effect is heavily limited and that claims to be able to tackle such issues by way of changes to the law or using crude financial levers without resorting to undue coercion, if not outright misogyny, are no more than a rhetorical sham and are definitely not to be trusted.

About the author
'Unity' is a regular contributor to Liberal Conspiracy. He also blogs at Ministry of Truth.
· Other posts by Unity


Story Filed Under: Blog

Reader comments

Reactions: Twitter, blogs

Sorry, the comment form is closed at this time.

35 Comments


48 Comments


21 Comments


49 Comments


4 Comments


14 Comments


27 Comments


16 Comments


34 Comments


65 Comments


36 Comments


17 Comments


1 Comment


19 Comments


46 Comments


53 Comments


64 Comments


28 Comments


12 Comments


5 Comments


NEWS ARTICLES ARCHIVE