The Mental Health puzzle, part IV: the economic hypothesis.

In three previous parts, I posed the puzzle of the measured increase in mental health problems (depression, anxiety, and obesity) across the Western world since the 1950s and briefly discussed the pros and cons of the main cultural explanation doing the round. Here I want to discuss the mainstream ‘economic explanation’.

The mainstream economic explanation is to simply take for granted that people are rationally choosing their risks of becoming mentally unhealthy later in life and hence that the increase in mental health problems must reflect increased benefits of those risks and reduced costs. People are then obese because they want to be obese and they are depressed and anxious because they got unlucky in that they took decisions that entailed a high risk of these problems and lost.

There is a lot to be said for this kind of brutal cost-benefit rationale.

For one, the health system has become inclusive in that many of the costs of mental health problems are borne by the community.

In 2006 for instance, I already calculated that the average obese American cost 2000 dollars more in terms of health costs than non-obese Americans and that these costs primarily came at the expense of others, ie they were not borne by the obese themselves. Furthermore, the health effects of obesity and in particular reduced length of life has since the 80s been overcome, mainly by the widespread use of statins. Hence the obese now live about as long as everybody else, a clear reduction from the point of view of the individual in terms of the negative consequences of obesity.

Similar things can be said about anxiety and depression and other mental health problems: sufferers are no longer told they are crazy and locked up, but are now much more looked after with much more resources flowing towards them. They are still not pleasant conditions to suffer from, but the private costs have clearly come down, increasing the payoff for those who would rationally take risks that might lead to depression and anxiety. Prozac and other medical interventions have made these mental health problems more bearable, thus increasing the incentives to risk them.

If you think about the direct costs and benefits, the same story emerges. The actual food costs of becoming obese has of course declined, and so has the payoff to being physically fit since less jobs than before demand physical fitness. Similarly, labour laws now make it more difficult to fire people who are depressed or anxious, and generous government welfare programs take in millions of people in these categories, effectively reducing the monetary costs on individuals and their families from these mental health problems.

Within this approach, there are a variety of multipliers that create a long-run lag between changed monetary incentives and behaviour. One of those multipliers is for instance the marriage market, which would initially penalise the few who are mentally unhealthy (a thin market problem) but in the longer run adjusts as the market is flooded by the mentally unhealthy. Similarly, adjustments in terms of the design of buildings and consumer items to cater for the mentally unhealthy (such as clothing lines for the obese or convenience outlets for those too anxious to go out in the open) take time, again creating a lag between initial changes in monetary incentives and the behaviour of whole groups.

The policy prescriptions of this mainstream economic approach to mental health is basically the exact opposite of where policy is going: from the mainstream economic perspective, one would advocate a ‘tough love’ approach to all of these diseases: one would allow health insurers to charge the obese more for their insurance; one would reduce the monetary compensation flowing to sufferers from depression and anxiety; and one would encourage the use of fitness and mental health tests as a valid selection tool for employers. The policy reality is clearly in the exact opposite direction so from a mainstream economic perspective one should expect nothing but worsening mental health outcomes in decades to come as our societies reward the mentally unhealthy more and more.

The problems with this economic approach are again in terms of plausibility and policy prescription.

In terms of plausibility, the main problem is to find some benefit to these mental health problems that makes it rational to risk them. Which choices that lead to higher risks of depression, for instance, have a possible payoff making the risk worthwhile? I dont know of any such choices, since everything that is good for economic outcomes (education, savings, fitness, mental discipline) is usually associated with lower risks of mental health rather than higher risks.

Indeed, to depict obesity as a rational choice maintained for decades by individuals is rather odd. You see, whilst life is no longer shorter for the obese, it is not pleasant either. Obesity is still associated with reduced physical fitness, reduced libido, erectile dysfunctions (particularly if lots of medicines are involved), diabetes, and social stigma. In which weird world could that be a choice that a fully rational and calculating individual would take? Not the world we live in, and the same can be said for the other mental health problems; the model of rationality simply doesn’t fit them.

The economic approach also has great difficulty rationalising the cross-sectional variation; there is for instance little reason why mental health problems should be higher in the cities than in smaller communities, why the same change in economic incentives should have played out so differently over countries, etc.. Via ad-hoc trickery one might fill in the cross-sectional puzzles, but it’s a stretch.

In terms of policy, the basic prescript of course fails the democratic test: with large proportions of the population now suffering personally or indirectly (via family members) from mental health problems, the point where politicians can advocate a tough economic line on mental health sufferers has long since past. It’s a non-flier. So from a mainstream economic perspective one would not hold out much hope for reducing the mental health decline seen in recent decades. Indeed, one would expect worse to come.

This entry was posted in Uncategorized. Bookmark the permalink.

42 Responses to The Mental Health puzzle, part IV: the economic hypothesis.

  1. derrida derider says:

    Shorter Paul: We minimise deadweight losses by killing off the weak, so we should do it. Ban statins and Prozac now.

    Seriously Paul, you seriously need to think about the normative limitations of a marginalist framework.

    • Paul Frijters says:

      Why dont you read the whole post so that you realise I actually agree with you about the limitations of the rational choice paradigm in this case?
      The deeper point though is that the rational choice approach is one story ‘in town’. Which other do you have to offer, apart from disliking that one?

  2. Gummo Trotsky says:

    The mainstream economic explanation is to simply take for granted that people are rationally choosing their risks of becoming mentally unhealthy later in life and hence that the increase in mental health problems must reflect increased benefits of those risks and reduced costs.

    That idea is, quite simply, unworthy of any extended exploration or elaboration. You’d have to be crazy to believe that anyone rationally chooses to risk becoming mentally ill later in life. For example, what rational choice has someone who becomes mentally ill as the result of repeated sexual, physical or psychological abuse in childhood exercised?

    If that is the mainstream economic explanation of mental illness then it’s time for mainstream economics to shut up on the subject.

    • Paul Frijters says:

      Why dont you read the whole post so that you realise I actually agree with you about the limitations of the rational choice paradigm in this case?
      The deeper point though is that the rational choice approach is one story ‘in town’. Which other do you have to offer, apart from disliking that one?

      • Gummo Trotsky says:

        OK, so maybe I’m a little more emotionally labile and frustration intolerant today than I thought. My bad.

        Here’s a story I picked up from RB, subject of my most recent post. While working as a nurse he was asked by someone – either a visitor to a patient the hospital he was working at, or a patient waiting in the emergency department – to look after her daughter while she slipped out for a smoke. He agreed to do so.

        Then he had one of those awkward conversations that adults have with stranger’s kids- how old are you? Do you like school? … And finally, ‘What does your father do?’

        ‘Oh, he beats me with a stick.’

        A response which he rightly reported to the proper authorities.

        That little story provides an example of why I might feel a little affronted by a ‘rational life choice’ explanation of mental illness and why I might – perhaps over-vehemently – reject it. If the idea that people make a rational choice to become mentally ill is a popular story in town we’re all in deep doo-doo. Because it just ain’t true.

        Other stories I’d offer – possibly genetic predisposition, patterns of family life, adult abuse of children, badly managed schools more geared to cultivating academic achievement than student welfare – in short a complex interplay of personal, familial and social factors which escape analysis by any single academic discipline.

        And that’s enough – probably already said too much, too badly.

  3. conrad says:

    You could look at “rational man” in reverse where you offer incentives or make the costs of some things less. This is what the government has been trying to do for some time. The reason for this is that for some (many) disorders and other negative things for that matter (e.g., marital problems) if you look at the amount time people take to actually do anything about them, it often hugely lags the emergence of the problems, and by the time people do anything it’s often worse. This includes some things like social anxiety where I think the lag between I’m-a-hopeless-case and getting anything done is over a decade and PTSD where most people are divorced etc. . by the time they do anything.

    So you really need to do a number of things. These include:
    1) You need to de-stigmatize the problem (reduce the social cost). This is a long term thing that can and has been done.
    2) You need to convince people to get help early, potentially with incentives (treatment is really expensive, and often doesn’t go to the groups that really need it).
    3) You need to work out some way of stopping organizations trying to cover-up the problems because it costs them money (e.g., the army is an obvious example). This is open to exploitation so is some balancing act.

    Number (2) is really hard (and 3 is almost impossible and really political, so I’ll leave it here), and there are lots of problems with the governments schemes. These include that (a) the services are mostly used by people that could already pay rather than the people that need them but can’t get them; (b) the providers simple charge over the medicare rate because they can, so poor people can’t afford to pay anyway; (c) the type of services are very limited; (d) it’s very hard to evaluate service providers for how good they are, and the current situation in Australia is almost entirely focused on a narrow range of clinical stuff yet these people often pretend to solve all problems; (e) it’s likely to become more limited as the clinical-psychologists have vastly too much power, have turned themselves into a medical-style mafia, and think they can solve everything via a fairly limited array of techniques that are certainly not appropriate to everything; and (f) I’m led to believe that GPs, who often know SFA about these disorders and what to do about them, are actually the worst for ripping off the current system (that’s probably biased because of who I know).

    The other problem, which I think you ignore too much, is that different disorders probably need different approaches, so I don’t think there is a one grab-bag story, with some things being very hard to treat. For example, most programs to do with aggression get bad outcomes, so it’s not clear what you can do about it (jail unfortunately a lot of the time), whereas other things like anxiety programs are generally much better.

  4. Michael says:

    A very interesting series of posts.
    I find the mainstream economic explanation of “rationally choosing” doesn’t stand up to much analysis – perhaps it should be relabeled the “lone wolf story of choice”. How many people really make choices about behavior, education, jobs, living arrangements, diet etc in isolation of their peers, families and communities? What would account for so much difference in historic and cross-cultural differences? There are narratives and fads that influence people both in terms of accepting or rejecting ideas and narratives – herding etc. Why have so many people switched to SUV’s (obese cars) when road surface quality has improved and the chances of crossing a stream on your way to the supermarket must be at a historic low. Likewise other socially influenced behaviors like drinking – do people make mass rational decisions to engage in anti-social behavior when drunk in larger numbers in one country than another? Diet and easting habits, it hardly needs to be pointed out are highly influenced by culture.
    I wonder if there is an element of contagion with things like depression? Environmental or cultural factors might initially exacerbate predispositions in individuals, which then get reinforced among social groups. I wonder if there is any evidence of geographical clustering in metal health or obesity stats.

  5. TN says:

    My problem, Paul, is with your labelling tof he applicaton of the simplistic marginal choice model to matters such as obesity or mental health as “mainstream”. I don’t know many economists who would subscribe to it, and I now of some fairly well-respected economic institutions that have rejected it. The Productivity Commission in its reports on gambling, smoking and obesity, for instance has recognised that cognitive limitations, along with other forms of market failure, would render such an application, well, simplistic.

    This is of course not to say that the marginal choice model might not still give some interesting and indeed policy relevant insights – such as the idea that taxing calories/cigarettes would reduce obesity/smoking, ceteris paribus – but I doubt that the naive model is really ‘the mainstream economic position’.

    • Hi TN,

      To see the dominance of the rational choice model in health economics, look no further than the Grosman health production approach and the rational addiction literature. The number of applications of either of them, which are at their core the application of the principles of rational maximising investors/consumers to any health problem, runs into the thousands. Indeed, I know of no other theoretical tradition in health economics that gets anywhere near the same journal space. For sure, that approach is not dominant outside of economics and health economics has lots of atheoretical scholars working in it who champion no theory in particular. But in terms of the dominant cause-effect stories in health economics, there is currently a clear dominant storyline in town and is that of the rational anticipating planning economic agent who chooses his actions to balance the future health risks with the other benefits of his current actions. Usually this includes perfect foresight, such that in the canonical model people even know the date on when they will die, but in more sophisticated models there is some uncertainty about that.

      • TN says:

        Thanks for your response Paul.
        Yes the ‘naive’ model gets a fair run in the HE literature, but do any of its users believe that it ‘explains’ people’s behaviour in the way you say – for example, do any of them believe that “People are then obese because they want to be obese”, as you put it?
        I would encourage you to differentiate between the theoretical games academics play and what they really believe, at least when the policy rubber hits the road. For instance, re: rationale addiction, the PC explicity said rejected it as an explanator of gambling addiction; nor did they endorse it w.r.t. obesity or smoking. Do you think serious economists – even academic economists – would disagree with the PC’s take?
        I guess my problem is that you present a model economists know to be naive, though still useful in some applications, as the ‘mainstream economic explanation’, without qualification or explanation.

        • Paul Frijters says:

          Hi TN,

          I come from the world of research so naturally I am presenting you with what I perceive as the mainstream line (and am happy to defend such a statement). The PC and other more government-alligned institutions of course inform themselves from a wide variety of sources.

          As to whether the academic who tow these lines ‘really believe’ it, my answer is ‘yes, when it comes to their role as referees, editors and teachers of the next generation of economists’ and ‘no, of course not when it comes to their own private lives’. In many ways the first one is the one that matters more than the second. I have always been amazed at the incredible disconnect between what social scientists (not just economists) say they believe about the world and what they reveal to believe by their private actions.

          Yet, to truly want to reconcile professional beliefs with private beliefs is a major burden: it would need the social scientist in question to truly try to have a more or less consistent view of how an enormous array of problems and empirical regularities hang together. Its a big ask. Just ask yourself: do you even try to truly make sense of it all and leave no inconsistencies between your professional opinion and your daily behaviour? Honestly? Do you know anyone without inconsistencies?

  6. MT Isa Mine says:

    So do we just forget the 30% of the pop that have IQ under 100 and may not be able to make the rational choice we need them to make about food to use your example. What about the any number of people that have a genetic predisposition to impulsive decision making and that plays out either in food or THC which sets off mental illness or those that have an undisclosed genetic predisposition to insulin sensitivity.

    My point is that the more I read around these areas the more I believe that genetics and IQ interrelate much more than many academic would like to believe. Why would they not affect health and longevity? Many people shear away from facing the inescapable animal natures of ourselves. The more we know of the interactions between neuropsychiatry and chemistry and genetics the more we will understand mental illness.

  7. Ken Parish says:

    Surely a relevant factor regarding increases in the incidence of mental illness and related phenomena would be the fact that in modern post-industrial societies there are less jobs for those with lower skills, lower IQs and/or borderline mental health issues from the outset. They experience prolonged unemployment, demoralisation and social exclusion. You would expect to see increasing mental health issues among these groups over time, as lower-skilled labour-intensive industries move offshore to low-wage countries and are replaced by “knowledge industries”.

    Then there is the phenomenon that Alvin Toffler labelled “future shock” almost 50 years ago. The sheer pace and pressure of change and an increasingly “dog-eat-dog” culture impact more heavily on some people who are in no sense stupid or mentally ill in the beginning, but who are more sensitive/artistic and therefore more vulnerable to those pressures of rapid change than the norm. The article I linked above has some interesting observations about future shock and ts manifestations, including the following:

    Not only individuals but society as a whole is likely to undergo these negative effects of too rapid change. The three basic attitudes are easily recognized in current patterns of social behavior. Aggression directed at no one in particular seems to underlie phenomena like vandalism and hooliganism. Individuals running amok and shooting scores of innocent bystanders may suffer from a more extreme version of this condition. Helplessness and despair can be recognized in the increasingly common “burn-out” syndrome, and in the ever so frequent depressions. Drug addiction may be another one of its symptoms.
    But perhaps the most common neurosis in present society is anxiety. This is illustrated by the record use of anxiolytic drugs (e.g. benzodiazepines) that suppress anxiety symptoms such as sleeplessness, worrying, irritability, tension and digestive upsets. Anxiety also shows in the many irrational fears and scares, where far-away threats trigger disproportionate reactions.

    The phenomenon Toffler observed is exacerbated by the “information overload” engendered by the Internet, ubiquitous mobile telephony and ICT in general. Occasional attacks of “stop the world I want to get off” are things all of us experience from time to time, but I suspect they are a permanent reality for those who are inherently intellectually or emotionally ill-equipped to cope with that world. Do we just write them off and invent patronising rationalisations? Or should we try to find solutions?

  8. Ken Parish says:

    There are also clear parallels between the points that I made above and Rob Bray’s post about the national minimum wage. It seems very clear to me that there is a significant cohort of people for whom there is no economic place (at least without subsidy in some form) in our modern post-industrial society. What we should do about it raises some very tricky issues, but labelling these economic casualties as dole bludgers or weaklings who are choosing mental illness is at the very least unhelpful in my view.

    • Hi Ken,

      Yes, the issue of the underclass who lacks the skills to compete in our society is very real, though I would say that there is nothing new about this: full employment never really existed and the main difference over time is the visibility of those at the bottom and how welfare for them has been organised (they were sent to Australia once upon a time!).

      Your burn-out explanation is certainly a story you hear, but comes with its own plausibility and prediction problems. Perhaps I should make it the subject of the next instalment, but in short: the onset of mental health problems is now earlier than before, hence amongst people who will often have no work experience and hence been subject to that phase of the rat race. You again have the problem of the cross country evidence in some countries to explain (low rates in Japan and Korea). The is the question whether all jobs really are that more stressful than the ‘old jobs’ which of course were much dirtier and actually more dangerous than the current jobs. There is the general fact that levels of violence are at historically low rates and that we thus do not see the prophesised increase in aggression. In terms of policy, there is little hope of outlawing competitive pressures. And of course at the top end, the most stressed people are doing well. Finally, one then wonders what it is about different cultures that brings in a difference in how people cope with life’s pressures.

      It’s a real puzzle Ken. No truly obvious cause or remedy in sight.

      • Ken Parish says:

        I haven’t analysed the Japanese and Korean evidence in any systematic way, but my intuitive suspicion is that lower rates of mental health issues there might have something to do with the fact that their societies remain somewhat more communal and based on extended family and other mutual obligations, compared with the West’s more atomised, individualistic, nuclear family organisational structure.

        • Paul frijters says:

          Yep, that was the story of part III. Came with its own problems…..

        • conrad says:

          I must admit that I’m still yet to believe there are less mental health issues in Japan and Korea — having a mental health issue is like having the plague in these countries, so no-one is going to tell you if they do. Alternatively, if you’re depressed in Australia you can moan to your friends and they’ll probably be sympathetic.

          As for obesity, the difference there is obvious. They eat good food. Isn’t it nice being able to buy rice balls at the train station or airport instead of Hungry Jack?

          You might also want to look at (a) drug use, especially for things like ecstasy, and whilst it isn’t popular to say, genetic differences.

  9. Ken Parish says:

    In addition there are parallels with Nicholas Gruen’s post about Terry Eagleton and atheism. We need some modern day secular social and psychological mechanisms that do some of the positive work religion once did. Reminders and a sensibility that our good fortune is not solely a result of our own personal striving and inherent excellence. Being slow to judge others as lacking in backbone or moral fibre. Most of all, a secular equivalent of the sentiment “There but for the grace of God go I”.

  10. TN says:

    Hi Paul
    There’s no REPLY button option to reply above to your last (7.41am) response, so let me do so here.
    I am interested in (and inclined to challenge) your characterisation of economists and social scientists as maintaining a set of professional beliefs that differ from what they personally believe.
    As an economist myself, I am unaware of holding different professional and private beliefs on particular matters (which is different from whether the belief set I hold contains any inconsistencies within it). Rather, as I see it, my discipline gives me a bunch of tools and models – some quite ‘naive’, others more realistic – that I deploy as best I can to address the questions I face. (Of course, I also use and/or incorporate tools and models from other disciplines where they appear more ‘fit for purpose’.)
    Sometimes I use (or, at least, start with) naive models because there is nothing better, and the answers those models yield are probably “fairly right”, or probably “more right than just guessing”. But that does not mean that I believe the simplifying behaviourial or motivational assumptions in those naive models. Nor do I think those assumptions/models necessarily “explain” the matters they are dealing with in the way you suggested – ie “obese people want to be obese”.
    Obesity is of course the culmination of a series of actions, and one can model each as resulting from a rational choice within its context. But being able to model something in a particular way is surely not the same as saying you have explained it.

    • Paul Frijters says:

      Hi TN,

      a topic close to my breast!
      I think it best to refer you to my writings on this topic first for many examples of common inconsistencies and self-delusions. I ran a blog series on ‘Can you handle the truth’ a few years back and more recently on ‘Why politicians lie’ (they are in the troppo archives). My recent book delves in-depth into this area.

      To examine your case I would have to know what it is that you have taught students and that you have advised others. Something to chat about on some suitable occasion. You are at the PC, right?

    • Paul Frijters says:

      yep, the high suicide rates there are one of the problems with the cultural explanation. One ‘quick fix’ to rationalise the apparent puzzle of the high suicide rates amongst teenagers with the low mental health problems recorded later on in life is to invoke a difference between short-run effects of high social pressure leading to suicide amongst those who cant comply with the norms and long-run effects leading to stronger communities. This was already talked about at the end of part III……

      • conrad says:

        I think I have a lower belief acceptance criterion than you do. You seem happier to believe in somewhat untestable observations than me (actually, I think the above one could be tested with some effort — you could try and find short-term blips in social pressure in groups within countries and see if it really does increase suicide rates — serious life pressure at least does have a big effect — e.g.,, and the extent that translates to social pressure who knows, and so something like, say, the effect of being an Arab in the US after 9/11 would be interesting. According to you mental health problems with this group should have risen and then fallen after things normalized a bit), so data that goes the opposite way of your current hypothesis is rationalized away but data that goes the correct way is taken as evidence.

        As it happens, if I had to weight your evidence, then I probably would fall on your side, but I think the effect sizes for the things you are looking at are going to be pretty small (i.e., they’re just one factor).

        The other reason I’m less believing is there are a lot of ideas in this area that are basically propagated by confirmation bias that are basically false. For example, any number of people love to believe that the reason people get depressed in those miserable countries in the Northern Hemisphere in winter is due to lack of light. Indeed, there is a whole industry doing “light therapy” where you sit beside a lamp for an hour or two a day. But actually, if you look at the evidence that it is light that causes people to be miserable at this time of year (cf., being stuck inside with kids that scream all day, getting colds etc. ), you’ll find is pretty scant apart from a few poorly controlled studies. Properly controlled ones show light therapy is basically a placebo effect (you get peace and quiet for an hour or two a day sitting next to a light in a clinic with people being nice to you), and the actual neuro-data looking at the effect of light and how this should interact with anti-depressants turns out not to support it either. This is pretty much the history of understanding mental health — it’s a theory graveyard and so you really do need solid evidence to support arguments as what seems intuitively obvious is often later shown to be incorrect (or at least the power of the effect is tiny).

        • Paul frijters says:

          Hi Conrad,

          I think you misunderstand my intentions and beliefs here: I truly have no clear ‘front runner’ as to an explanation for the mental health decline. I have many theories, each more unlikely than the last one, but none I truly see as clearly the most likely, as should be clear from these four posts. The uncertainties pertain not just about the problems with each of candidates but also about the basic data. However, that is precisely what attracts me to this puzzle: we researchers have no idea what is going on and are not even absolutely sure anything is going on. My kind of puzzle.
          Your guesses are spot on though. There have been quite a few economic studies using 9/11, for instance looking at the birthweight of babies in uterus at 9/11 from parents with Arab names. Their birthweight was markedly lower, indicating a direct effect of the stress experienced by the mother on the unborn.

          Let me take away even the little certainty you seem to think you have in the realm of suicide data: it is notoriously messy as data. For instance, suicide rates in catholic countries used to be zero because it was considered a sin to commit suicide and hence they were recorded as something different. Measured changes in suicides in Spain are hence not all that reliable. In other cultures, like the Japanese one, it is considered more noble to commit suicide and they are historically thus more often recorded as such. Do we know whether observed increases reflect real increases of changes in measurement? No we don’t. Are we even sure we know in which countries there are more suicides! no. Are there deaths of which one cannot be sure whether it is suicide or accidental? Of course there are. Does the method of suicide differ by gender and country? yes. Do countries differ in official reporting standards? Of course, but they change over time. On and on the interpretation issues go.

          So yes, Conrad, I see all the problems and am not really buying into any theory. But I do see the value and need for simple stories so am definitely shopping around. Your need is perhaps not so strong?

  11. conrad says:

    “So yes, Conrad, I see all the problems and am not really buying into any theory. But I do see the value and need for simple stories so am definitely shopping around. Your need is perhaps not so strong?”

    Paul, I think there is no simple story in terms of explaining different disorders or in terms of finding factors even for single disorders.

    In terms of the first one, there are probably clusters of disorders that are causally affected by the same factors (notably depression, stress, and anxiety).

    I see obesity as something that would better fit into addiction models since you have two time-frames that are important — the short term one where you can get gratification (the booze and turkey was great!) which trades off with the long term (I’m getting fat and unhealthy). Trying to change this behavior then involves not only thinking about the short versus long term tradeoff, but trying to get rid of habits accrued over a life time. So there might be some simple story for part of the causal relationship between obesity and other addiction-like behaviors.

    As for whether one factor is going to be out there for most things? There’s any number of changes since decades ago that could all play some role. These range from the simple, like ecstasy abuse causing more depression to the more complex, like styles of parenting and that kids don’t get taught much about HTFU these days, so they are not learning anything about resilience (as of your own posts note). I don’t see why any single one of these should explain most of the data — if it did it should be pretty obvious. Why, for example, would a small increase in individualism (if it’s really happened) outweigh large differences in learning to be resilient? I can’t think of any reason, but I think the only way to know is to get the data (both probably have some role).

  12. conrad says:

    Just to show that I’m not just whinging here that there is a simple solution to every complex problem, which is usually wrong, I decided to come up with 20 different factors (some of which are mentioned here). I actually came up with 21.

    1. Higher individualism breaking down community structures (less support causing stress)
    2. Less community/family structures in general (less support causing stress)
    3. Less religion (less support causing stress)
    4. Increased ecstasy use (brain damage across the lifespan)
    5. Increased binge drinking (brain damage across the lifespan)
    6. Increased marijuana use (brain damage if smoked before full brain development)
    7. More pollutants in the environment (any number of bad effects)
    8. Dysgenic effects (low IQ people, who are more likely to suffer mental disease, have more children)
    9. Dysgenic effects (More people with mental disease likely to be able to have children thanks to medical advances passing on a genetic susceptibility)
    10. Higher recognition rates for the same level of disorders
    11. Parenting without emphasizing bad things happen in life (lack of resilience to minor events)
    12. Schooling without negative/hard events (lack of resilience to minor events, unrealistic expectations of ability likely to be destroyed in the job market)
    13. Wider social factors (e.g., media) that give people unrealistic expectations of possible life achievements and abilties (negative self comparisons)
    14. Less parental discipline for negative behavior (lack of development of inhibitory mechanisms)
    15. Less discipline at schools for negative behavior (lack of development of inhibitory mechanisms)
    16. Increased parental coddling (lack of reality, lack of development of self-regulatory mechanisms)
    17. Higher populations (more competition for resources, more increased daily life stress due to increased traffic etc.)
    18. Higher awareness of negative events (the world must be a bad place, negative perception of the world)
    19. Pessimism about the future (first generation in ages that probably won’t be richer than their parents)
    20. Older parents/fewer children (more stress caused by sick/dying/dead parents and fewer people to look after them. See also (2))
    21. Fewer social face-to-face contacts (poorer development of relationship skills/theory of mind)

    • Paul Frijters says:

      Hi Conrad,

      thanks, this is helpful. Let me in return say how I currently think about each element in your list in terms of their value/importance for creating a simple story that fits most of the mental health decline:

      1. Higher individualism is probably real (but itself not an easy concept to define) and I do see it as a likely part of the puzzle. The nice thing about it is that it has policy levers so the more individualism is involved, the more we could potentially do about it.

      2. Same as 1.
      3. Same as 2.
      4. I doubt it matters much, maybe in extreme cases.
      5. Same as 4: as far as I understand this part of the literature, in the long-run there is a lot of repair (ie evidence that young binge drinkers have a bad subsequent life does not mean there is long-run physical damage that is irreparable). Indeed, I see this one as a bit of a moral-story (people want to scare others into not drinking for reasons that are interesting themselves).
      6. Bit more convinced of the potential for damage, but doubt it explains more than the odd case.
      7. Doubt it and probably goes the wrong way: our food standards and safety systems are pretty good. local environments have cleaned up in recent decades, such as via unleaded petrol.
      8. The old malthusian worry (the wrong people get kids and not enough of the ‘weaklings’ die young). Its an important possibility, for sure, though obviously highly politically incorrect to talk about so you dont see much open discussion on it. In terms of explaining the recent increases though, it is not quick enough to explain much, it violates the cross-country variation (because this mechasnism should hold for all countries that have seen large reductions in infant mortality) and we dont actually know yet which genes are supposedly coding for stupidity nor whether they are passed on (there are alternative theories for why some people are not so smart and why there is a limited degree of heritability in those). I thus think it likely that this effect is small, though there is bound to be some of this.
      9. See 8.
      10. Clearly true. Probably an important part of the measured increase, though not for obesity which convinces me that there is probably also a real increase in the other ones (exaggerated by reduced social barriers to measurement).
      11. I normally fold this one into 1, ie a change in group formation which itself comes with a change in ideals and parenting styles.
      12, 13, 14, 15, 16. Same as 11. This is a group of possibilities relatively easy to capture in simpler stories. It is precisely those factors that I tried to capture in the catch-all cultural story of part III).
      17. Unlikely. No obvious relation between population density and mental health in the cross-country patterns. Likelier that someone else happens to co-move with urban versus rural communities (see the comment I made in one of the comments to Part I).
      18. Same as 13: the media story. I generally see the media as reacting to changed demand rather than a truly exogenously moving factor of its own.
      19. Unlikely. Optimism rules in the behavioural data. Indeed, over-optimism is seen as one of the hallmarks of the current generation. And of course, in many ways life is still getting better on average.
      20. A possibility, but doesnt fit the cross-country evidence (lots of countries have ageing populations), so this one is more a ‘back of the mind’ worry with which the story has to be consistent.
      21. Same as 1: part of the changing mosaic of groups in our society.

      As to the use of simple stories, I see it as my job to come up with simple stories that capture a lot. For me, that’s what a general social scientist, and particularly an economist, produces for his/her society: simple stories that help people ‘understand’ a lot of their surroundings and helps them to find solutions to their problems more quickly and with less error. It is not important that these stories are imperfect for that is inevitable. What matters is whether they are good enough to be useful. Of course, you probably see yourself as having a different role to play.

      ps. Are you sure about that light-thing? I thought winter-depression was real for a minority of people based on stories of vitamin D and serotonin-via-sunlight. Is the best-guess now that it is probably not real?

  13. Michael says:

    “I view obesity as a mental health problem, ie the result of a lack of willpower.”

    I’m still troubled by this classification. Are all causes of obesity the same? What about super-tasters (it’s real and it does contribute to obesity in some individuals – you can look it up). Obesity is better classified as a symptom not an illness. You could lump all kinds of failures of willpower as a mental illness but that would be ridiculous.

    Other failures of willpower that reduce optimal health and achievement:
    — alcohol use
    — general laziness/procrastination/ blog reading
    — co-habitation leading to negotiated outcomes
    — car dependency and suburban living reducing enforced exertion
    — general failure to maintain a motivating sense of shame at your current state, etc

    • Paul Frijters says:

      Hi Michael,

      Taking for granted that obesity is not a desirable thing that people consciously choose to be in full recognition of the consequences (which already places us outside of the usual assumptions in mainstream economics), I am forced to think of obesity as arising from an inability to resist food and exercise more. I view it as a threshold-thing: one’s ability to adhere to behaviour that one on reflection recognises as better for oneself in the long run (=willpower) is not sufficient to withstand the temptations and possibilities available in the short run, nor is it sufficient to seek out different circumstances and habits under which the short-run temptations and possibilities change.
      Yes, I thus also see obesity as a symptom. A symptom of poor mental health that can in turn come from either side of the threshold. And yes, your other examples can be argued to also be such symptoms, though it is less clear that they really are all undesirable (blogging, alcohol use, co-habitation? I do all those believing I roughly know the long-run pros and cons and am happy with the tradeoffs. Arent you?).

      • Michael says:

        OK – I was being facetious with some of those points.

        I was just trying to point out that “willpower” is an attribute that is mainly understood in individualistic terms rather than groups (although it can be applied). Mainstream economics places too much emphasis on “individuals rational choice” and not enough on group norms as influences of behaviour. Willpower isn’t that useful in considering children. Many obese people start out as obese children living in families that are also obese and I think looking at the information asymmetry between food manufacturers, retailers and marketers vs busy people balancing the demands of family life and work seems more useful in conjunction with other environmental and cultural influences.

        I’m happy with the trade offs I have been able to make that allow me to eat reasonably healthily and exercise a fair bit. This has been part willpower but it also happens to be mostly luck in that I can ride to work and exercise nearby doing something I enjoy and for which I don’t have to pay to much for. I doubt I could easily find another employment situation that would offer me this.

        • john r walker says:

          The kind of ‘calories’ you eat is a big factor, refined sugar (and highly refined carbohydrates in general) and hydrogenated and partially hydrogenated oils have weight gain effects that are not simply down to their calorific value. They also seem to have ‘addictive’ effects .

          Generally, high levels of consumption of the above group of ‘junk’ maps very closely to the ‘disease’ cluster you are disusing.

        • john r walker says:

          Generally, high levels of consumption of the above group of ‘junk foods’ maps very closely to the ‘disease’ cluster you are discussing.

  14. conrad says:

    Ok, we agree then we have a number of different stories even at the general level:
    1) Individualism stuff
    2) Parenting
    3) Better recognition

    I don’t know if you could really come up with a simple story amenable to the public for either of 1+2 (they’re complex). If you say “bad parenting”, for example, people will ask you want you mean, and you will get into all the sub-threads related to this, and you can also see how hard a simple cultural story is to make convincing just from these posts (now try convincing people that don’t read or think).

    One’s we disagree on but would certainly be arguable in the public sphere:
    1) Ecstasy abuse (I personally wouldn’t care too much about any of the other common drugs at least in terms of brain damage, but I certainly wouldn’t would have lots ecstasy, and that’s not because I’m ill-informed. This is a bit of an epidemic just boiling away in my books, but there isn’t any truly super data).

    One’s I’m more on the fence about are:

    1) Population size — obviously there is big cross-cultural variation, but Sydney must be the most stressful PIA city in Aus to live in (Hong Kong, alternatively, was fine), and so I don’t live there! Surely within countries things like urban amenities and urban planning make a difference (isn’t this part of the reason why people think people get obese in the US and Aus — they live in places you can only drive to?)
    2) Malthusian stuff. This is really a more interesting argument. The obvious and non-interesting one is people of low-IQ breed with other people of low-IQ, and these people will have a tough time in modern societies. I agree this is a slow effect.

    The more interesting one which no-one really thinks about much because people have been stuck in disorder vs. normal land but is probably exceptionally important for getting a grip on what causes these disorders is that historically, with some disorders, people have had terrible outcomes. These are mainly thought disorders which I think are excluded from your list (even through I think Aspeger’s/Autism is supposed to be on the rise also). So if people with Schizophrenia, for example, tend to die before they have children yet there is a genetic component, how does it stay in the population? The answer must be that it is some combination of generally normal genes that constantly allows the re-emergence of it. So by better medications and better outcomes, you will inevitably end up with a higher-base. I’m not sure how to model this over time (no-one is!) since we just don’t know enough about the genetics. But if you want data that isn’t controversial to show that bad genes can get into a population quickly or at least have a higher prevalence to start with (at least for susceptibility), I simply point to the Asthma rates in Australia or the short-sightedness rates in Southern China (90%).

    “As to the use of simple stories, I see it as my job to come up with simple stories that capture a lot”

    Sure, but stories should not be too simple. For example, if there are multiple pathways to these disorders, and they are affected by different things (which seem uncontroversial), you are going to have to have multiple stories if you want to make smart suggestions on real issues.

    Psychologists are worse than economists for this incidentally — Everyone wants a single story, and often don’t even care about effect size (probably half of them don’t know what that is!). So they do a t-test between a normal and abnormal population, find a difference, and think that explains the story. This is why fixing things like education boils down to “bad teachers” in the public mind.

    • Paul Frijters says:

      more stuff, thanks!

      – On exctasy, I will now keep your prediction in mind. To be honest, I always thought it was a fairly innocuous one from its places on lists of ‘most harmful addictive drugs’ (where alcohol usually wins that title, though really ‘power’ should be on top by a mile).
      – Population size: not just do you get the ‘wrong prediction’ problem that there are high-density low-problem places like Singapore and urban Japan, but you also get the historical issue that for centuries much higher populated areas like the Netherlands have been doing fine in terms of relative mental health. For sure, stress is important but modern societies are remarkably good at packing millions of people close by without necessarily leading to problems.
      – I have much less understanding of schizophrenia or autism so purposely left it out at the start of the series. Rates are rising, yes, but since I have much less of an idea how to view them, am not yet game to lump them together into a ‘probably connected’ group.
      – I thought the reason for poor eyesight in South-East China was that children stay indoors too much to study when young and thus their eyes deteriorate relative to Western kids who get sunlight by working less hard and playing more outside.
      – I thought the Astma story was now thought to be related to over-cleanliness and lack of exposure to particular diseases very early on in childhood. The first is a social thing (for which I am happy with the stories I have on that), the second an unintended consequence of improvements in medical technology for which there are obvious technical fixes (simply give babies particular diseases to make sure their immune system orients itself on the right thing). On both, policy reactions are thus fairly obvious though it will probably take a while to get socially accepted.
      – IQ heritability. This is obviously ‘Bell curve’ territory (I just reviewed a book on this). Until they find the genes supposedly responsible for high IQ, my best guess is that IQ has little to do with genes (at least not genes that have a high chance of being passed on). What then explains the heritability? It might be womb quality; it might be a combination of interactions between genes, food, and environment (without any gene under which low IQ is truly inevitable); it might be gene expression and as such quite persistent from generation to generation (but not indefinite). There are clear early-childhood effects here and quite likely cross-generation lock-in effects too (a poorly fed, stressed, and educated mother makes a worse womb and thus a baby with less life chances already at birth, perpetuating accidental economic disadvantages). The strong reversal to the mean one finds in abilities between generations is my main reason to doubt the gene-story on IQ.
      – Having said the above, it is possible that IQ is far more heritable than I currently think, but that is a truly scary proposition because it gets us policy-wise into Untermensch-Ubermensch territory, so I fervently hope IQ is not truly genetic for all but a minority, but have to acknowledge that it might well be proven to be so eventually.
      – funny to hear that ‘statistical significance’ is a craze in psychology now too. In economics, Deidre McClosky for a while made a living pointing out that it is ‘economic significance’ that should matter more than statistical significance.
      – It is my current understanding that teacher quality is the main policy-sensitive input in schooling outcomes. We looked at some economic papers showing it to have a large economic effect a few year ago in our reader group (cant remember on the top off my head who the authors were). Are you telling me that conceived wisdom has been overthrown by newer and better studies showing the effect to be small?

      • conrad says:

        “I have much less understanding of schizophrenia or autism so purposely left it out at the start of the series.”

        I don’t think anyone has a good answer for this, so you are certainly not alone!

        “I thought the reason for poor eyesight in South-East China was that children stay indoors too much to study when young and thus their eyes deteriorate relative to Western kids who get sunlight by working less hard and playing more outside”

        I assume this is true (I should ask a vision friend next time I see him), but there are certainly genetic predispositions. If you look at white/mixed race kids growing up in HK (and you could probably get similar data in Vancouver and Hawaii), for example, they get short-sighted less (funnily enough, glasses first appeared in Southern China and Northern Italy at about the same time, and I can’t help but notice that both groups seem to have poor eyesight!).

        ” I thought the Astma story was now thought to be related to over-cleanliness and lack of exposure to particular diseases very early on in childhood”

        I’m not happy with this story. Australians are far dirtier than Northern Europeans. All the Germans I know are also pretty fastidious about cleanliness (and I’m told the Swiss are worse) yet they don’t suffer such bad rates of Asthma. So it’s either just some random genetic drift (which no-one ever thinks about) or something in the Australian environment. And it’s not just Aus — rates go up worldwide, although by different amounts, and no-one knows the answer (there is nice article here: so there is another problem for you.


        I agree it’s a terribly complicated story, and the best you will ever find is correlates of genes, and some of the correlates are going to be things like “poor processing of alcohol”, and so you end up with populations that might have one or two points more because they don’t drink much, not because the genes have anything really to do with thought at all.

        That being said, I don’t think anyone believes that there isn’t some heritability of IQ. People are just too nice now to say it. For example, we know various genetic correlates of working memory and other cognitive tasks (although no-one knows what the genes actually do). But we call them working memory and not IQ even though working memory is a big part of IQ scales.

        “funny to hear that ‘statistical significance’ is a craze in psychology now too.”

        It’s always been a craze. I blame journals for taking “whizz-bang” finds, and not intelligent work. Social psychology, which pretty much is the same as economics at least in terms of the questions asked, is the worst offender for this (Daniel Kahneman recently complained about this).

        “Are you telling me that conceived wisdom has been overthrown by newer and better studies showing the effect to be small?”

        No — this is mainly my opinion — but in my books the biggest factor is the curriculum, but this is to hard to measure and disentangle from other factors, and hence it is almost never considered in these studies. For example, how do you measure the effect of, say, what’s taught from grade 1-4 in mathematics on Year 12 achievement? At least in part, what is probably going on with good teachers is that they are compensating for bad curriculum and I can think of well known changes that have had hugely detrimental effects (e.g., graphical calculators) but most changes are fairly non-obvious.

        For example, you were complaining about language courses in Australia some months ago. Now let’s just assume that your kid had all good teachers in these areas but they were just teaching dross as that’s what they’ve been told to do (and probably forced to thanks to “quality” measurements designed to make all students happy).

        Now let’s say you had streamed classes and a curriculum with a nested structure so they could go from level to level, and let’s also say that the teachers they have are a full SD worse at teaching than their current ones.

        Are you predicting that your kids would have learnt more in the first or the second condition? I know where my money lies, and the reason obvious. Even with really poor teachers, kids will still learn by themselves, will get taught by their parents (another huge factor that generally isn’t mentioned), and so on. But if you force them to learn dross, this won’t be very helpful. So a full one SD difference in terms of teaching quality is destroyed by bad courses, so in this case clearly course quality is far more important than teaching quality.

        • john r walker says:

          This is really interesting!
          A question about Asthma – are Australia’s high rates constant across Australia… are they the same in the cities as in the bush and the same in ‘penrith’ and ‘Woolahra’?

        • conrad says:

          John, I don’t know what the data is in Aus apart from the fact that it is high. Alternatively, there are very good data sets from other countries, and people still don’t know the answer. Indeed, it appears they arn’t even close to knowing what causation is. see e.g., . As you notice these guys implicate “degree of modernization” not cleanliness. It’s worthwhile noting here that even this article doesn’t explain why Singapore (apparently 27% of the population) and HK (11%) differ so much despite both having genetically similar populations, similar weather, somewhat similar food, lots of pollution (HK obviously has more), and similar living conditions. What’s the difference? Beats me.

        • john r walker says:

          I am not sure;
          Asthma is a sort of auto inflammatory condition, no?

          It seems curious that we do not seem to have a mapping of its rates of occurrence across Australia.

          Do you know if cases of life threading- needing hospital treatment- attacks and/or deaths, have also increased in proportion to overall rate of increase of diagnosis of asthma?

          Asthma is an odd thing, my mother, at 70, suddenly had a severe attack that needed 7 days in hospital, it was the first and only attack in her whole life.

        • conrad says:

          John — there’s sure to be that data somewhere, both for locality and for hospitalizations (the latter of which is most likely to have gone down due to better medications etc.) as people report on it sometimes — I just don’t where it is.

        • john r walker says:

          Conrad my brother-in law is a prof of pharmacology, we were discussing this a few weeks ago … He did not know of any geographical distribution data , I will ask him if he has found anything.

  15. conrad says:

    “ps. Are you sure about that light-thing? I thought winter-depression was real for a minority of people based on stories of vitamin D and serotonin-via-sunlight. Is the best-guess now that it is probably not real?”

    Winter depression is definitely real, it’s just that the cause is disputable. For example, if it was just vitamin D, all you would need to cure it was a few vitamin D tablets. Why bother sitting in front of a lamp? But this doesn’t work. The same is true of the hormone stories. These stories clearly predict that certain types of anti-depressants will work better than others, but if you look through studies looking at this, it isn’t true. So the real story is that being dark cold places is depressing for any number of reasons, and that placebos that cause changes in mental states work really well.

Leave a Reply

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.