Scarcity versus insecurity
Does poverty make it harder to make good choices, or does poverty change what counts as a ‘good’ choice?
Why do people on lower incomes die younger and get sick earlier than people who are better off?
Part of the reason is the fact that when you don’t have a lot of money, you are more likely to be exposed to hazards like damp, badly insulated housing; air pollution; crime; and poor quality, insecure work. These exposures lead directly to poor physical and mental health.
But an important part of the answer lies in the fact that behaviours like smoking, physical inactivity, harmful alcohol and drug use, and unhealthy diets are more common among people on low incomes. Here, unhealthy behaviours sit between poverty and illness. Because of this, the question of freedom comes up: are people choosing these behaviours freely, and with a full understanding of the risks? And if they are, is it right for the government to start trying to influence those choices?
This is a big question that touches a lot of contested philosophical questions including what the legitimate role of government is, through to whether free will even exists. There are lots of more prosaic responses to this question too, such as pointing to the fact that non-government actors (private companies, charities, religious organisations and so on) also try to influence people’s behaviour, and not always in ways that benefit the individual concerned. Or the fact that money buys choices and people living poverty simply don’t have the same options as people who are better off. I’m not going to try to address any of that here though. Instead, I want to focus on one idea about why poverty is so closely linked to unhealthy behaviour, and why I think it misses something.
Scarcity
A significant contribution to this debate has been the idea of ‘scarcity’, which Sendhil Mullainathan and Eldar Shafir explain in their book of the same name. This idea is that lack of money (or time, or other resources) reduces a person’s cognitive bandwidth, and this makes it harder to make the best decision (this might be about saving for a pension, or it might be about doing something unhealthy, like smoking a cigarette).
Scarcity is an important idea. Mullainathan and Shafir provide evidence from a range of studies that shows that people appear to perform worse on tests of cognitive performance when they are poor than otherwise. Noah Smith provides an excellent overview of the evidence here.*
This evidence is important because it is sometimes suggested that people on low incomes are poor and unhealthy because they are less intelligent (although rarely in such blunt terms). Mullainathan and Shafir suggest the opposite: poverty makes us less intelligent. This matters because it’s all too easy to start blaming people for their problems. It also suggests that our supposedly ‘free’ choices are heavily shaped by our circumstances.
The scarcity view still assumes that there is a consistent ‘right’ decision to be taken. But what if poverty and insecurity changes what the ‘right’ option is?
Healthy behaviours as investments
One way to think about healthy behaviours is as an investment in your future health. Just like saving for the future, healthy behaviours often involve effort or giving up some pleasure now in the expectation of some pay off in the future. This might mean not eating that tasty but unhealthy takeaway, or not lighting up that cigarette, or getting out for that run when you’re not feeling like it.
There are whole fields of study dedicated to these sorts of decisions. I don’t pretend to be an expert in this stuff, so bear with me. And if you are an expert in this, please correct me if I get things wrong.
In working out how people should choose between two options where the rewards happen at different times, economists use the concept of discounting. The rewards that come with each option are reduced (or discounted) according to how far in the future they happen. This assumes that other things being equal, we’d rather have our reward sooner rather than later, thank you very much. And the further away the reward is, the less it is worth. So if I offer you a reward of £10 now or £10 tomorrow, you’ll probably take that tenner now. But if I offer you a reward of £10 now or £20 tomorrow you might choose to wait. But if the choice is between £10 tomorrow or £20 in ten years’ time, you might not wait.
The details of how this works, both in theory and when actual people are involved, are fascinating** but not critical here. The important thing is that people differ in how much they discount future rewards. A high discount rate would mean a stronger preference for rewards sooner rather than later. Higher discount rates are associated with smoking, drug use, overeating, and other unhealthy behaviours that involve trading rewards now for more risk of getting sick later.
One example of this is the well known Marshmallow Test. In these studies, Walter Mischel and colleagues offered young children a choice between one of their favourite sweets now, or two sweets if they could not eat the first sweet until the researcher came back into the room. There are videos of children resisting the temptation to gobble the sweet in front of them. They are as funny as you’d expect.
Mischel and colleagues followed up the children in the original study later in life and found that the children who were able to wait were (i.e. those with lower discounting rates) more likely to do well in school, get good jobs, and so on. This has typically been taken as meaning that there are differences in willpower and impulsivity that are apparent in young people and which manifest through life in an ability to delay rewards, with better outcomes. Cue many parenting guides on how to develop such willpower in your children.
However, a later study found that this effect is strongly reduced if you adjust for the child’s socioeconomic status. Why should that be? Well, imagine if experience has taught you that that second sweet might not be forthcoming. Maybe the adults in your life haven’t been all that reliable in keeping their promises (maybe because they had other things to worry about). Or maybe there were other people around who might take away that sweet. In that case, does it make sense to wait for that second reward, if you’re not so sure that it’s going to arrive?
In other words, it might be perfectly sensible to adjust your discount rate depending on your understanding of how reliable and predictable the future is.
Some evidence for this comes from the literature on Adverse Childhood Experiences (ACEs). There is evidence that people who experience a lot of trauma and adversity in childhood are much more likely as adults to smoke, drink, have unhealthy diets, drink too much alcohol, take drugs, and a whole host of other unhealthy behaviours.*** Other studies have found growing up in poverty leads to higher discount rates in adulthood, and there is some evidence that thinking about a sudden drop in income increases discount rates.
Another way of looking at this is to think about what the minimum discount rate is that you would apply when making some kind of investment. I don’t know about you, but I’d quite like to be alive to see the pay off of my investments (I love my kids, but let’s ignore inheritance for simplicity). So my minimum discount rate should be related to the probability that I die between now and whenever the reward is due. Our subjective sense of safety is likely to be influenced by the extent to which we can control the risks that we’re exposed to. And that is likely to have a lot to do with poverty. Money can’t buy you happiness, but it can sort out a lot of your problems. That’s why happiness tends to increase with income but only up to a point.
A related idea has been explored by Daniel Nettle from an evolutionary perspective. In this wonderful paper, Nettle divides the risks that an organism faces into those over which the organism has no control (extrinsic risks) and those that the organism can do something about. Nettle shows that the more extrinsic risk an organism faces the less it makes sense to put effort into managing the risks that it can affect. Putting this in terms of health behaviours, it doesn’t make much sense to try to quit smoking if you’re facing a firing squad in the morning.
So what?
Hopefully, by now I’ve convinced you that at least part of the reason that people living in poverty are more likely to have multiple unhealthy behaviours is that if the future is bleak or highly uncertain, it makes sense to live for the moment. If tomorrow might not arrive, let’s have fun now. Or “let us eat and drink, for tomorrow we die” (Isaiah 22:13). This is in contrast to the scarcity view that the ‘right’ decision doesn’t change, but poverty makes it harder to choose well.
I should point out now that these are not necessarily mutually exclusive explanations. Both theories are at odds with the view that people are making free choices, well aware of the consequences, and uninfluenced by any external factors.
Why does this matter? Well, the implications for policy are different. The scarcity view may be taken to mean that people on low incomes just need extra help to make ‘better’ decisions. This might be in the form of advice or information, or more subtle nudges, such as shifting defaults. Whereas the view I’m arguing for (call it ‘precarity’ or ‘insecurity’) would put more emphasis on tackling the insecurity that makes people feel like there’s no point thinking about the future.
I’m not suggesting that we shouldn’t try to help people make longer-term choices for their health and financial wellbeing. And I’m also not saying that Mullainathan and colleagues aren’t well aware of the need to tackle insecurity. But I am suggesting that unless we recognise how insecurity affects our incentives and preferences, we will find it hard to understand why unhealthy behaviours follow such a pronounced social gradient. And this will make it very difficult to tackle the gaps in health between people on lower and higher incomes.
Policy implications
So, what to do? I think that there are a few areas of public policy where taking an insecurity-focussed approach might help.
First, and probably most importantly, we should use the tools at our disposal to protect people from chronic financial insecurity. Mark Gamsu, Woody Caan and I have argued elsewhere that financial welfare systems can be powerful tools for protecting people from the most extreme forms of financial insecurity, and the mental illness that goes with it. Evidence suggests that including an unconditional element (i.e. a guaranteed component of welfare payments that is not subject to conditions) is likely to improve health and wellbeing. The purest form of this policy is a universal basic income (UBI), but the evidence also suggests that unconditional payments need not be on the scale of a basic income to improve mental health. I would expect improvements in financial stability to lead to a greater tendency to value and focus on the future, and so to lead to healthier behaviours.
A similar line of thought could be applied to housing and employment laws. It’s hard to imagine prioritising your long-term health if you’re perpetually worried about losing your job or your home. In both areas, policymakers need to balance the needs of renters or workers against the needs of landlords and employers. In employment, policymakers have (in the UK at least) tended to emphasise the benefits of a flexible labour market, where it is relatively easy to hire and fire employees. An extreme example being zero-hours contracts. An insecurity-focussed view of this policy might put more emphasis on the costs that this imposes on workers in terms of insecurity, and the associated harms to physical and mental health. Likewise, laws that make it harder to evict tenants would protect renters from insecurity about where they are going to live.
The obvious riposte to this is to say that any such laws would reduce the demand for workers, or supply of rental accommodation, and that would increase unemployment and homelessness. I’m not suggesting that landlords should never be able to evict a tenant, or employers fire an employee. Rather I am arguing for a better understanding of the costs in insecurity imposed on renters and workers when we make work and housing insecure so that these are not forgotten when balancing the various interests involved. This might make, for example, policies like increasing the stock of social housing more attractive.
Another area where an insecurity perspective might help is crime. Crime rates regularly feature at the top of lists of public concerns. The impact of crime rates extends beyond the victims to other members of the local community. Higher crime rates are likely to cause anxiety in local residents. As well as the distress that this causes, the associated stress may contribute to higher rates of unhealthy behaviours.
So far, so obvious. And what government would say that tackling crime isn’t a priority? An advantage of taking an insecurity-focussed view of crime is that crime itself can be seen as a response to chronic insecurity. For one thing, if you feel that you are likely to be a victim of crime and the police aren’t there to protect you, then violent retaliation (or even pre-emption) starts to make more sense. For another, if your future looks insecure, you may not be all that deterred by the prospect of a prison sentence. And finally, an insecurity-centric approach to crime would emphasise the importance of tackling those things like child poverty, trauma, adversity and insecurity early in life that appear to have a big effect on setting our discount rates for the rest of our lives. Policies here might include reducing child poverty through transfers and providing services such as children’s centres, youth centres, and family support services that aim to provide security and head off insecurity.
These are just a few ways in which an approach focussed on minimising exposure to chronic insecurity could shape policymaking. By protecting people from insecurity, these policies would help to shift people’s incentives so that they have more reason to value the future.
There seems to be a George Orwell quote for any occasion, and this is no exception:
“And the peculiar evil is this, that the less money you have, the less inclined you feel to spend it on wholesome food. A millionaire may enjoy breakfasting off orange juice and Ryvita biscuits; an unemployed man doesn’t … When you are unemployed, harassed, bored, and miserable, you don’t want to eat dull, wholesome food. You want something a little bit ‘tasty’.”
George Orwell, The Road to Wigan Pier.
(thanks to Andy Turner for putting me on to this).
Footnotes
*Interestingly, Noah emphasises the role of stress, whereas Mullainathan and colleagues are careful to differentiate their concept of ‘bandwidth’ from stress. However, the research literature on the relationship between stress and intertemporal choice looks to be a confusing mess. Some studies find no effect. Some studies show that stress may decrease discounting rates. Others show that stress increases discount rates, but only acute stress, not chronic stress. Study designs vary, from experimental setups that induce stress by getting participants to give videotaped presentations, to others that cut out the middle-man and administer stress hormones directly. Some studies measure self-reported stress, while others use various biological measures of stress. An early draft of this blog had a whole bit about stress and the brain, but I took it out after a quick look at the literature gave me a headache. If someone knows of a coherent overview that makes sense of how the various studies fit together, I’d love to read it.
**For example, there’s plenty of evidence that humans don’t discount rewards the way that they should. For example, to be consistent, the difference in value between a reward now and the same reward tomorrow should be the same as the difference in value between a reward this time next week, and the same reward in one week and one day. But people tend not to behave that way — we tend to snatch at immediately available rewards more than we should. This is called hyperbolic discounting. Some have interpreted this as an example of the system 1/system 2 type thinking popularised by Daniel Kahneman in Thinking Fast and Slow. In this view, there is one region that deals with rationally comparing rewards that are distant, and another that is impulsive but is only activated for rewards that are particularly close. However, brain imaging studies have not found consistent evidence that different brain regions are involved in these different decisions.
***There’s a big debate about how much this is about adversity and trauma versus the wider socioeconomic circumstances that children grow up in, and whether the ACEs concept excludes other important sources of trauma. I’m not getting into that here. But whether you think the ACEs are the root cause or just one pathway between deprivation and illness, the evidence still suggests that early life experiences may be important in setting a person’s discount rate for the future.
