“Benefits ‘wrecked the British work ethic,’ new study claims”, ran the Daily Mail headline I cited in the previous post. But when I looked into the paper by Michau 2009 (based on this LSE paper) that prompted the article, I found some interesting results, but ones that fall short of providing persuasive support for Michau’s claim.

What makes his claims seem particularly odd, though, is that it is countries with the most generous welfare states that have the STRONGEST work ethic (measured by attitudes to benefit fraud). This figure is again taken from Michau 2009:

[According to Michau, these numbers are marginal effects - so that being British (vs. French) increases the chances of saying 'never justifiable' by 24%.  Interestingly, as noted in the full paper (p33), Britain is an outlier here from the general relationship]

Alongside this, we can look at Ingrid Esser’s work (inc her PhD thesis), particularly in her chapter in the British Social Attitudes report from 2010 (which also has a helpful review of previous research).  She uses ISSP 2005 data and measures work ethic through two questions:

• I would enjoy having a paid job even if I did not need the money
• A job is just a way of earning money – no more

This also isn’t exactly what lots of people would mean by ‘work ethic’ – as Esser discusses, this is about non-financial employment commitment, or in plainer English, it measures what you get out of work beyond money. To my mind, this does NOT include feelings of a moral compulsion to work.

Still, like Michau, this shows a clear association of welfare generosity with employment commitment. The highest levels are among 13 countries being are in the generous welfare states of Denmark and Norway, and the weakest employment commitment is – contra the Daily Mail – in the UK with a relatively ungenerous welfare state.

Trends in different countries

So how does Michau square this? Well, he argues that we would EXPECT countries that have a stronger work ethic to have higher benefit generosity when the system is introduced – it makes the problem of benefit fraud smaller. BUT over time, he argues we would expect the work ethic to decline most in those countries with the most generous welfare states (p31) – and this is what he claims to show below  (all changes expressed relative to a 1930 baseline).

However, remember again the points from the previous post: this is not a direct measure of work ethic, and we don’t actually know if this is a trend or just a difference in ages.

More importantly, even if these trends are true, then the fact that the work ethic is STILL highest in the more generous welfare states – even after this sharper decline than less generous welfare states- is even more surprising.

And if we get around some of the problems in Michau’s analysis and instead look directly at trends using Esser’s measures (p95-6), then we instead find completely the reverse picture.  From 1989-2005, employment commitment in Norway and Germany rose.  However, in the US and Britain – and particularly among British men – it fell.

Err… So what does this all mean?

Michau purports to show that the welfare state leads to a declining work ethic, while Esser claims the complete opposite. To my mind, both studies are incredibly interesting, yet struggle to justify the claims they make – not least because neither study actually measures what I would understand by the term ‘work ethic’.

Still, on some level I’m convinced that both accounts are true:

• One of the effects of welfare states is Esping-Andersen’s ugly term ‘decommodification’ - that is, people don’t have to rely on the labour market quite as much, which enhances their ability to reject truly crap situations that they would otherwise have to accept out of desperation.

• On the other hand, as I’ve covered on the blog before, this doesn’t mean that people have no work ethic.  And we see that more generous welfare states – where workers are generally treated better – will have a GREATER commitment to the work that actually exists.

So the evidence does tell us something about the work ethic and the welfare state – but it doesn’t conclusively come down in favour of one political perspective or another. Once more, then, this is an area where the debate has run for centuries past and is likely to run for centuries more…

Here are two charts on injury fatalities over time between non-Hispanic white adolescent males and adolescent males in four major racial/ethnic minority groups. The first chart shows annual motor vehicle fatalities per 100,000, the second shows annual homicide fatalities per 100,000 (the data file can be found here).

In their own way, each set of numbers is really staggering. Take disparities in motor vehicle fatalities: there has consistently been an almost threefold difference between the group with the highest mortality (Native Americans) and the lowest mortality rate (Asian Americans). Some of these differences are easily accounted for. For instance, Native American youth spend more time in motor vehicles because they tend to live in more rural areas with more dangerous roads.

How should we explain the spectacular decline across groups? I would be interested to see any relevant literature on this topic. My hunch is that it is a combination of factors: declining rates of drunk driving among teenagers, safer cars, safer roads, and better emergency medicine that keeps youth alive after car accidents. Still, we can do much better. Consider that Native American youth in 2007 still died at rates higher than blacks and Asian Americans from twenty years prior.

How about homicides? If you were a researcher or journalist in the early 1990s, it would have been reasonable to extrapolate that black males would continue to kill each in a situation that resembles today’s northern Mexico. Not surprisingly, homicide rose to the number one cause of preventable years of life lost in the late 1980s among black youth, eclipsing the AIDs epidemic. The increase among Hispanics in the early 1990s was also staggering in relative terms.

Things changed in the late 1990s. We know part of what changed: the crack epidemic subsided, the economy improved, policing strategies changed, and community leaders began to organize and mobilize against urban violence. The world is safer for black adolescent males than it was in the early 1990s, but the data are still very sobering. Even in 2007, the black adolescent male homicide mortality rate eclipses the Hispanic rates from the height of the homicide epidemic.

Some of the solutions to this problem are very easy and very obvious: first, stop the sale of automatic weapons and clamp down on secondary markets for handguns. Period. Some of the change will be harder to realize. We need to deal with persistent unemployment among minority males in urban areas. Still, these data should prompt a discussion about ways to bring about changes like those seen in the late 1990s.

Young people are more affected by the outcomes of the democratic process than other cohorts: their youth means that by and large they will live with the consequences of political decisions for longer. Furthermore, young people are at a crucial life-stage where the impact of political decisions will have a decisive and cumulative effect on their socio-economic circumstances across their lifecourses. The growing power of ‘the grey vote’ appears to have had real consequences for the ability of young people to make themselves heard within the democratic process, but more worryingly, may begin to undermine the legitimacy of democracy itself.

How old are the voters

Simply, cohort size matters. Analysis of the British Election Survey by Andy Furlong and Fred Cartmel shows that generations tend to be selfish when they get to the ballot box. This of course does not mean that age is the only or main determinant of voting behaviour, or that age-based political inequality matters more than any other form of inequality. That generations can act, more or less coherently, to bring about change in social structures was a proposition first put forward by Karl Mannheim in 1923. Mannheim, one of the founding fathers of modern sociology, believed that generational change was one of the main driving forces of political change. Strangely, this key precept of the discipline of sociology seems to have been largely overlooked by the study of democracy by political scientists.

At the 2010 British general election, 40-somethings were dominant at the ballot box. The youngest voters, and voters in their early-30s, were particularly disadvantaged. But the voting power of people approaching retirement, whose life chances will be affected by electoral outcomes to a far lesser extent than younger voters, was also highly significant. For example, there were more voters aged 50, 51, 52 or 63 than any single age between 31 and 36.

This inequality will accelerate in coming decades. Due to increasing survival rates, and the ageing of the members of the large baby booms of the immediate post-war era, the overriding trend is towards an older electorate, with greater concentrations of potential voting power among people in their 50s and 60s. There will be only 700,000 18 year-old potential voters, compared to a single-year age cohort average size of 902,000 for 50-somethings). Thirty years later, in 2051, there will be a particularly powerful set of cohorts aged around 60. The average single-year cohort size for people aged 58-62 will be 937,000, yet there will be only 825,000 18 year-old voters, and no smaller cohort up to age 68.

The median potential voter was 46 in 2010. In 2021 this will rise to 47. The median potential voter will be aged 50 by 2041, and 51 in 2051. It is worth reiterating that this is a relatively recent phenomenon, or more accurately, one we are yet to fully experience. The median potential voter in 1981 was already aged 46; this fell to actually fell 44 in the ten years to 1991, before rising to 45 in 2001.

Who votes

Taking voter turnout rates into account shows that the democratic process was even more skewed towards older cohorts. The median ‘actual’ voter was aged 49 in 2010, three years older than the median ‘potential’ voter. The median actual voter will be 52 by 2021, rising to 54 by 2051. OECD research shows that the gap between young and older voter turnout in the UK is three times the OECD average – leading to calls, for instance, for compulsory voting among first-time voters.

At the 2010 general election, relatively high turnout meant that 40-somethings were largely successful into converting their potential power into actual votes. But older cohorts had closed the gap significantly. Excluding 40-somethings, there were more actual voters aged 63 than any other age. Given their lower propensity to vote, 18 year-olds exercised less actual voting power at the 2010 general election than 73 year-olds, while 45 year-olds exercised 84 per cent more actual voting power than 18-year olds. Similarly, Scott Davidson’s research demonstrates that at the 2010 general election, more than half of MPs (319 seats) were elected by constituency electorates within which more than half of actual voters were aged 55 or over, with a further 102 MPs elected in constituencies were more than 40 per cent of voters were 65 or over.

My research also projects future inequalities. In 2021, 18-year olds will exercise less actual voting power than 79 year-olds. 55 year-olds will exercise more than double (115 per cent) the power of 18 year-olds. By 2051, if turnout rates persist, 18 year-olds will exercise less actual power than a typical single-year cohort in their late-80s. (The full results of this research are available from the Intergenerational Foundation website.)

These inequalities do not mean, in any straightforward sense, that young people’s vote should somehow be worth more. Democracy’s first principle is, and must remain, ‘one person, one vote’. But that representative democracies with near-universal franchises have only ever existed within populations with pyramid-shaped age distributions may be one of the hidden foundations of representative democracy. Representative democracy without this demographic bias towards young people is entirely untried – it is into this uncharted territory that we are heading at a rapid pace.

Generals are always fighting the last war

Clinton’s top-down approach may have reflected different political necessities, however, and Obama’s gamble to leave reform in the hands of Congress could not be sustained when he lost a super-majority in the Senate. In his very lucid and even-handed book, “Remedy and Reaction,” Paul Starr revisits much of the conventional wisdom on American health reform. Starr has the right credentials – he wrote arguably the most important history of American medicine (the “Social Transformation of American Medicine” published in 1982), and was a policy advisor on Clinton’s reform team. Starr’s book corrects some important misconceptions about the apparent reason for the failure of Clinton’s reforms in 1993 (another example – the insurance industry’s notorious “Harry and Louise” ads probably did not substantially undermine public support). He also traces the emerging consensus that led to reform in the next Democratic presidency.

Here are some important lessons about social policy that stood out for me in the book:

 Incremental programs builds consensus: between the signing of Medicare and Medicaid by LBJ in 1965 and the signing of the Affordable Care Act by Obama in 2010, Congress passed many smaller scale health care laws, and several states also overhauled their insurance programs (most notably, Mitt Romney’s Massachusetts in 2006). The legacy of smaller scale programs – expanded eligibility for Medicaid in the 1980s, the Children’s Insurance Program in 1997, the Medicare Modernization act of 2003 – is difficult to quantify. However, as Starr points out these smaller reforms both showed the viability of tweaking insurance markets and expanding public programs, and also helped to form consensus among centrist and liberal reformers that had previously embraced much different approaches (from single payer, to expanded Medicare, to tax credits).

…But incremental policies can also undermine systemic reforms: the “health policy trap” in the United States has always been that reforms mainly serve vulnerable subgroups such as the uninsured using resources from populations that are already protected. Ironically, this protection often is the legacy of previous public programs (like Medicare for seniors). Since leaders need to gain reelection, it is especially difficult to take steps that would improve health policy overall without alienating the elderly and those in the middle class that like their coverage. The new taxes and the minimum coverage requirement (“the individual mandate”) in the Affordable Care Act are good examples. These steps were needed to create viable insurance markets that pool risks between the healthy and the sick, but they also rock the boat for some populations that now feel that they are being coerced into helping others pay for medical care.

“Bipartisan politics” takes many different forms: the Affordable Care Act passed with a strong majority of Democrats in the House and Senate, but not a single Republican voted for the legislation. This is more than a bit ironic, since the stamp of Conservative ideas and policies is all over the legislation. That odious individual mandate, for one, was first championed by the rightwing Heritage Foundation and the overall legislative blueprint comes from Romneycare. They may deny it now, but several Republicans in the Senate – Olympia Snowe, Charles Grassley, and Mike Enzi to name three – had a heavy hand in drafting the legislation (which they ultimately did not vote for). Although this backchannel contribution of Republicans has not softened the partisan bickering about the law, it signals that there may be more bipartisan agreement around social policy than roll call votes reveal. 

Thinks tanks build the foundation for social policy: Finally, it was heartening to see Starr single out several policy research organizations – like the Urban Institute, Family’s USA, and the Center for American Progress – as playing a central role in shaping ideas and conducting analysis for health reform. At their best, the policy research and advocacy organizations in Washington provide imaginative and original solutions to public problems. This role does not come automatically, however, and often think tanks are not in synch with the current constraints and ideologies of Congress. Still, the fact that many of the ideas that ultimately entered into the ACA were formed several years before the law was passed shows that even ideas that are on the shelf do not automatically expire, but can be revived when the window of opportunity opens.

I don’t think that the claims of ANY of the papers stand up to critical scrutiny (!). But in the process of taking these claims apart, we can see something more interesting about the work ethic and the welfare state over time and across countries. In this first post, I look at trends over time; next week, I come back to looking at the work ethic in different countries (including trends in different countries).

A declining work ethic?

So what has been happening to the work ethic in welfare states in the recent past?  Well, underlying the Daily Mail headline “Benefits ‘wrecked the British work ethic,’ new study claims” is a study by Michau 2009 (based on this LSE paper). This makes claims about the ‘work ethic’ based on a single question in the World Values Survey:

“Please tell me whether you think it is always justified, never justified or something in between to claim government benefits to which you are not entitled”

So even as we start, it’s clear we’re not talking about the ‘work ethic’ in the way that we’d usually understand it – this is about attitudes to benefit fraud.

Now, Michau wants to make claims about changes over a long period of time, but only has data in 1980, 1990 and 2000 for 18 countries. So to find a ‘trend’, he looks at different attitudes to benefit fraud by age. The results are shown in the Figure below – and show that younger people think it is more justifiable to cheat on benefits.

The problems with the killer figure…

There are three problems with these claims though:

• We are not actually seeing trends here. Instead we are ASSUMING them, based on attitudes among people of different ages. In the full paper, Michau (p30) uses the 1980-2000 change to try and disentangle this, but it’s still not absolutely clear whether benefit fraud attitudes deteriorated even over the period we can observe.

• This is not a measure of ‘work ethic’.  Again from the full paper, Michau (p28) finds that this correlates with more usual work ethic questions (e.g. job satisfaction, work should come first, look forward to work after the weekend) – but the correlations are actually quite weak. He does find that the cohort effect can also be seen when looking at the question, ‘work should always come first, even if this means less spare time’ (p31). But this is only observed at one point in time (2000), and moreover, shows a different TIMING of the cohort effect (declining among cohorts born in the 40s and 50s vs. the 30s, and then staying stable – which suggests a possible WW2 role, as Michau mentions but then ignores).

• We don’t know WHY attitudes changed. Michau claims that “the rise in European unemployment can be explained by a generation-long lag between the introduction (or expansion) of unemployment benefits and the behavioural responses of workers”. But the actual evidence for this is weak to non-existent (see also below).

We know, for example, that there have been huge changes in attitudes completely unrelated to the welfare state – such as widespread rises in criminal behaviour in the late 20th century and increased individualism – and these could easily explain any trend in attitudes. When noting the benefit cheat attitude trend in countries with weak welfare states (below), Michau notes that “work ethic is certainly affected by many factors beyond the generosity of the welfare state” (p33), but he never really considers this properly.

(This seems to use the integrated Ljunge 2011:614 makes similar if more muted claims based on the same analysis for 95 countries using the same data – but also makes some inferential leaps in seeing increased sickness benefit claims as a sign of changing benefit attitudes, rather than (say) fundamental changes in the nature of work).

Next week I’ll show how Michau’s claims seem a bit puzzling when we look at the work ethic in different countries – and try to draw all this together into a coherent picture.

In a 2011 paper (here as a 2010 working paper), Anne Case and Christina Paxson take a different approach to the Whitehall data. They make three observations that I believe should be garnering more attention.

First, they argue that the study is not ideally set up to study the impact of early health on adult health:

“We find that, because the population from which this cohort was drawn consisted almost exclusively of white collar civil servants, the Whitehall II sample is not well suited for quantifying the importance of childhood conditions for the population as a whole. Children from poor backgrounds who find white collar positions in Whitehall are different in many dimensions from other poor children, and these differences lead to a systematic underestimate of the impact of early-life health and circumstances on later-life health and social status.” (Pg. 2)

They quantify the potential bias in a clever way. They take the two largest cohort studies from the general population in Britain: the NCDS 1958 Cohort and the BCS 1970 cohort, and examine how much socioeconomic status in childhood (measured by father’s occupational category) predicts adults height, first in a full sample and then in a sample restricted only to social servants. If social servants from low-status fathers have some unobserved differences that compensate for early health disadvantages, then we would expect the association to be weaker between parental education and height. Indeed, this is what they find.

Second, focusing on the Whitehall II cohort, they show that childhood socioeconomic status (father’s occupational status, and family car ownership during childhood) has a lasting impact in adulthood both on initial occupational category, and also on subsequent promotions. Not surprisingly, much of this effect is mediated through education. Importantly, however, they find that early health exerts a strong independent effect on these same outcomes (captured by height and whether the person was ever hospitalized before age 16). These are important results, especially since as the above results indicate, the Whitehall study is likely to be positively selected.

Third, these early health effects also exert an independent effect on adult health. For example, in regressions that control for a variety of current and previous characteristics, spending time in the hospital as a child is associated with a 7 percentage point lower likelihood of being in “excellent” or “very good” health as an adult. Controlling for current occupational grade does not dampen the association between child health and adult health. To try to disentangle whether occupational grade is high because health is good, or whether health is good because occupational grade is high, they also estimate regression models predicting current health on future occupational grade using earlier waves of the survey. Future grade is unlikely to cause current health, but they find that current health does predict future grade. Moreover, in fixed effect models, they find that current health is a much stronger predictor of future promotion than any previous occupational grades.

The authors conclude: “despite the downward biases that are likely to result from selection into Whitehall II, we still find evidence that health in childhood influences occupational status in adulthood. Adults who had better childhood health—as measured by adult height and hospitalizations in childhood—start at higher grades in the civil service, and are promoted to higher grades after they enter the civil service. The association between height and occupational status in adulthood is robust to controls for education, implying that childhood health does not operate solely through its effects on educational attainment” (Pg. 22).

If we take this study seriously, and I believe we should, it invites serious reconsideration of the apparent finding in Whitehall that occupational status causes health in adulthood. This does not mean that inequality in social status does not matter. To the contrary, the Case and Paxson results highlight the centrality of early life health exposures on late life outcomes. This is a different causal mechanism, and would result in a greater focus on interventions focused on poor children than on promoting greater workplace equality later in life as a means for improving adult health.

And then there’s costs that probably never occur to you – but which were cleverly revealed in a paper that came out earlier this year.

Getting fleeced: A field experiment

The paper is by Uri Gneezy (lovely name) and colleagues earlier this year (£), and looks at a series of different aspects of discrimination in the US – finding, for example, that young black men are less likely to be helped by strangers. But that’s not what concerns us here.

The experiment that shows these hidden costs of disability is refreshingly simple. Take 12 middle-aged men in Chicago, and get them to approach a set of 36 garages (i.e. car repair shops, for US readers lost in translation) using  the same set of slightly battered cars. Half of the customers are in wheelchairs, and half aren’t – each pair of people approach a randomly selected garage with the identical car, and ask for an identical quote.

So what happens?  The people who weren’t using wheelchairs were quoted $1212 – but the people in wheelchairs were quoted $1425!  When they take account of some quirks of the design, they estimate that wheelchair users are charged 30% more than non-users to fix an identical car.

What’s happening here?

This was a bit puzzling to me – these were all wheelchair-accessible garages, after all. But Gneezy et al had a hunch, and then showed that their hunch was right. (Academic papers are designed to tell stories like this – people don’t usually tell you of the blind alleys they went down…).

Their hunch was that wheelchair users don’t shop around as much as non-users – so that when it comes to bartering prices, wheelchair users have a weaker hand. To check this, Gneezy et al first did a survey of customers and of mechanics. Their story here checked out: on averaged disabled customers visit half as many garages than non-disabled people when they’re getting their car fixed, and what’s more, mechanics knew this.

But still – how do we know this is the reason that wheelchair users were getting charged so much more? To really hammer the point home, Gneezy et al sent their volunteers out to 24 Chicago garages again. Everything was pretty much the same as before, except that the wheelchair users made a point of saying, “I’m getting a few price quotes”.

Bingo. In this case, there was absolutely no difference in the prices quoted to the wheelchair users and non-users (in fact, non-wheelchair-users had slightly higher quotes, but this wasn’t statistically significant). In other words, the reason that wheelchair users were getting charged more was ENTIRELY because they weren’t able to shop around as much.

The moral of the story

There’s several things I take out of this. Firstly, it shows to me (yet again) the power of social experiments – this really is much more convincing than the endless, increasingly complex analyses of people’s responses to questionnaires without any manipulation. I hold a guilty hand up here for past and future sins.

Secondly, and more importantly, it shows the continuing, hidden nature of disability discrimination. The US – like the UK – has anti-discrimination laws, and sometimes non-disabled people seem to suggest this has levelled the playing field. But the sheer spectrum and range of disadvantages goes beyond the situations that the law can deal with – not just in this particular case, but also in many others which I’ll come back to on the blog.

Third, it highlights the logic of having benefits to pay for the extra costs of disabilities, separate to benefits that are paid to people who are not working due to their disability. There’s a vast literature looking much more broadly at the extra costs of disability (I won’t review this all here, but see my comment below the post). But aside from affirming the importance of this benefit, there’s a couple of things to note:

• The numbers claiming this extra cost benefit in the UK is being cut by 20%, as it moves from being called ‘Disability Living Allowance’ (DLA) to ‘Personal Independence Payment’ (PIP) – two genuinely bad names, and a change that will be devastating for those affected.

• And amazingly, while we have this benefit to cover extra costs, it’s treated as ‘income’ in our official poverty measures, rather than as meeting extra needs. So the real levels of poverty in disabled people are much higher (for example, see Chanfreau & Burchardt 2008, p9).

And as a last word – if you use a wheelchair and have a car, then let us know if this tactic works with garages near you!

[As a general disclaimer: I've used the term 'disability' very loosely here, because this is a blog post, and repeatedly talking about 'people with impairments' has a tendency to put off some readers. But the social model of disability is underlying this, and hopefully readers familiar with these debates can pick this up. Apologies if not, and I promise to use terms more carefully in academic papers!]

An ambitious new paper by Orazio Attanasio, Erik Hurst, and Luigi Pistaferri attempts to address measurement error in the Consumer Expenditure Survey (CEX), and comes to a different conclusion. The CEX is the most comprehensive source of information on spending and saving in the United States, but it has become increasingly worse at measuring certain types of non-durable purchases. If this bias is worse at the top of the spending distribution – a real possibility since high spenders may not track certain purchases as well as those living on tighter budgets – then the unadjusted trend in the CEX may bias the inequality trend downward.

The authors deal with this in a few innovative ways (consult the paper for exact details). One is that they focus on categories of consumption that are known to be measured with less error, such as money spent on food at home versus money spent on entertainment. The second is that they examine the value of durable purchases such as car ownership (which is very well measured in the CEX). The third is that they bring in other data sources, the Panel Study of Income Dynamics, which is thought to measure certain purchases with more accuracy (although the PSID covers fewer items in the family budget). They also consult another CEX data source that is independent of the quarterly survey – the daily diary (respondents keep a diary over two weeks and enter in their purchases every day) – because it covers a smaller time-slice there is more variance in the daily diary, but it is a good supplement for examining trends.

All of these data sources lead to the same conclusion: the trend in consumption inequality mirrors the trend in income inequality much more closely than previously thought. 

 Here are a few figures that emphasize the trend:

This figure just shows how much more variability there is over time in people’s consumption (although it doesn’t tell us where the variance is growing in the income distribution).

Similarly, we can see that the ratio of spending on food (which is income inelastic) to entertainment (which has an elasticity greater than 1.5) has been increasing over time. This is interesting circumstantial evidence, but not conclusive. The figure below I think makes the best case.

Here we see that there is a divergence in consumption measured in the diary component at all points in the income distribution.

This paper makes an important contribution from a policy standpoint, because it mutes the argument that the “real” effect of inequality on people’s standards of living is much lower than its impact on their incomes. Changes in income are transitory, but what people actually purchase reflects something about their permanent income (since people can smooth their consumption across time by borrowing and saving).

The figure above tells an important story. Twenty one percent of respondents had been to prison, but an additional 53 percent had a close family member or friend in prison. Two thirds had a close friend or relative who was murdered. Almost half of all respondents had been arrested at some point. Consistent with other national studies, just over ten percent had an alcohol or drug problem (it’s commonly believed – wrongly – that drug and alcohol addiction is more common among African Americans than among whites).

The majority of respondents reported at least one kind of discrimination. Most common, more than half believed that they have been unfairly stopped by the police, and a substantial proportion said they experienced other forms of discrimination such as poor service in a restaurant, people acting afraid of the person, or people thinking they are not smart. These more subtle forms of everyday discrimination can accumulate and lead to heightened stress and lower self-esteem.

Despite all this, fully 96 percent of young black men say that parents should tell their sons that anyone in America can be successful if they work hard, although more than half (52 percent) believe that black men need to work harder to get credit, and 49 percent believe that the police are looking for any reason to give black men a hard time.

As I have reviewed before with data from the GSS, America has made important strides in reducing racial prejudice in many areas of life. African Americans are better integrated into social life than thirty years ago, but an important barrier to future progress is a system of law enforcement and mass incarceration that is perceived as biased against black men.

I previously reviewed the subjective wellbeing (SWB) concept, and described some interesting papers that exploit exogenous shocks in wealth to examine subjective wellbeing. SWB is really several concepts: a person’s happiness at a point in time, a person’s assessment of how well their life is going (for example, the “Cantrell ladder” which asks people to place themselves on an 11 point life satisfaction ladder), and the experience of different feelings such as anxiety, smiling, anger, or crying. Across countries, the relationship between logged GDP per capita and average life satisfaction scores is positive and linear at any point in time (Americans average around “7”, very poor countries like Benin and Togo average around “3”, the Danish top the scale at around “8”). By contrast, there is no strong relationship between the hedonic experience of happiness and GDP per capita: people in poor and rich countries are about as likely to say that they feel happy at a point in time.

The Bitter Taste of Politics

Deaton examines changes in all of these dimensions during the recession. As one might predict, the average ratings on the Cantrell ladder scale drop considerably after the collapse of the financial markets in 2008, however, they massively increase for no immediately obvious reason on April 6^th 2009 (see picture below). As it turns out, the reason is due to a change in survey design: on that date survey respondents were no longer asked about their political opinions on various issues before the SWB questions. Deaton says: “People appear to dislike politics and politicians so much that prompting them to think about them has a very large downward effect on their assessment of their own lives… the effect of asking the political questions on well-being is only a little less than the effect of someone becoming unemployed, so that to get the same effect on average well-being, three-quarters of the population would have to lose their jobs. Not everyone becomes unemployed, but either half or all of the respondents are asked the political questions.” The bitter taste of American politics evidently makes people feel momentarily feel that their own lives are going worse.

There are lots of gems related to survey design on this point in the paper, and I encourage you to read it for that alone. One interesting finding is that putting in a transition question, that is, asking people to not focus on politics and focus instead on their own lives appears to effectively neutralize the question ordering effect. In any case, Deaton provides a fairly interesting and convincing methodology for repairing the series of observations so that he can consistently evaluate the trend in SWB over time.

How do people feel in a rollercoaster economy?

Deaton looks at differences in SWB by age group. Consistent with other research, there was a clear gradient between age and life satisfaction: older people rated their wellbeing higher at all time points. There was no real secular trend for older people, but those of working age (especially 18-59 years old), had the most pronounced dip during the height of the economic downturn. Deaton says: “The oldest group, whose pension income is unlikely to depend on the market, are the least affected by the crisis.”

Focusing on trends by income, most income groups dipped in SWB during the height of the downturn in 2009, but the trend was most pronounced among those with lower incomes. This might be surprising considering that these households tend to not have much wealth tied up in the stock market. While one might predict it is because of higher unemployment, it actually turns out that SWB started to increase overall for this group several months later when the unemployment reached its lowest point.

Deaton in part believes that some people may have undergone some adaptation in their positive emotions over time, so that people acclimated to a bad economy and did not report as large decreases in happiness as might otherwise be expected:

 “Hedonic experience, particularly worry and stress, but also physical pain, deteriorated during the crisis, becoming rapidly worse during the summer and fall of 2008, recovering briefly during the year-end holidays, only to reach their worst values around the time that the stock market was at its lowest. There is a (small) increase in hedonic affect in all of these measures on St Valentine’s Day, and a much larger one around the Christmas holidays. As the stock market revived, negative affect fell. By mid-2010, there is very little trace of the crisis in these measures—though admittedly it is hard to detect small trends among the variance—even though the crisis continued in terms of lower incomes, employment, home and stock prices. These results are consistent with hedonic adaptation, especially in the positive measures (happiness, enjoyment, smiling, not being sad). Worry and stress (which behave similarly to one another) are particularly sensitive to the crisis, at least in terms of the increase in the fraction of the population reporting them. Hedonic adaptation is somewhat less clear for worry and stress than for the positive emotions.”

This gets to the heart of the debate about SWB – indeed the heart of a debate that has been argued by philosophers for centuries – how well does subjective experience represent people’s overall wellbeing? If people are adapting to bad situations by changing their expectations, then we may end up giving legitimacy to these experiences and undervaluing how bad life truly is for such people. This would matter if, for example, we wanted to design policies that compensated people for their loss in wellbeing (and not just income or other resources) from experiencing unemployment. As it turns out, the experience of being unemployed matters, but the effects of large scale unemployment are minuscule in the population overall. The stronger association appears to be with future expectation, as represented by the stock market: “today’s calculation of lifetime utility reflects not only what is happening to today’s instantaneous utility, but also the expected utility of future outcomes, predictions of which are changed by today’s news.” If we care about improving people’s short-term expectations, we should pay more attention to their SWB, if we care about improving their lives overall we may need to look elsewhere.