economics, psychology, policy: Summary of Lecture on Well-Being and Economics

I am currently giving a set of lectures as part of a module "Behavioural Economic: Concepts and Theories" in Stirling. I am posting brief informal summaries of some of these lectures on the blog to generate discussion. Thanks to Mark Egan for help putting these together online.

This section is on Well-being and Economics.

Introduction
A key theme in behavioural economics is what is the best measure of the welfare of individuals and countries. A huge literature has debated whether it is money that makes people better off or whether we need to look at an alternative measure.

The most famous paper on this area is by Brickman, Coates & Janoff-Bulman (1978). The results, described in Fig 1 below, showed that people who won the lottery and people who were rendered paraplegic by an accident initially experienced big changes in well-being in the expected direction but they expected to converge back to their base levels or thereabouts. Additionally, the lottery winners did not derive any more pleasure from everyday activities. This is in line with a big literature arguing that well-being is set to a fixed point determined by genetics, upbringing, disposition etc., and that it is not very malleable to changes in external factors like income.

Fig 1. General happiness: participants were asked how happy they at this stage of their life (Present), how happy they were before winning/the accident/6 months ago depending on the condition (Past) and how happy they expected to be in a couple of years. Mundane happiness: 7 questions about how much pleasure the participants got from everyday events such as talking with a friend, watching television, etc. All Qs answered on 1-5 scale. Figure adapted from original paper.

Can we predict what events make us happy?
People like Kahneman and others have argued that the focusing illusion effect might come into play when we think of what life would be like after some external change. For example, Schkade & Kahneman (1998) showed that people think they would be happier if they lived in California (nice weather etc.,) but, in fact, Californians are no happier than New Yorkers (same working stresses etc.,). We think it would be nicer to live in California because we focus on the most salient piece of information that comes to mind when we think about it, e.g. the nice weather. More succinctly, Kahneman explains the focusing illusion by arguing that "Nothing In Life Is As Important As You Think It Is, While You Are Thinking About It".

A more recent paper asks the question "would you be happier if you were richer?" again arguing that people overfocus on the benefits of money when making counterfactual judgments.

People like Daniel Gilbert have talked about affective misforecasting where we mispredict how good or bad we will feel consequent on life changes, in particular that we will overestimate how long we will have a change in well-being consequent on a change in circumstances (see this TED Talk). This can depend on what measure of wellbeing we are interested in - a paper by Kahneman and Deaton using the global Gallup data, showed that income has a much bigger effect on life satisfaction than emotional well-being. On a similar note, although the unemployed generally report lower life satisfaction than the employed, a recent paper by Knabe et al. (2010) using the Day Reconstruction Method found that average experienced utility did not differ between these two groups.

Conceptualizations of Well-being
So life-satisfaction and emotional well-being ("mundane happiness" in the words of the lottery study) seem to be tapping into two very different things. Most of the papers in the literature deal with levels of income rather than sudden once-off increases in income. It might be the case a that helicopter drop of money places so much exertion on the self-control of the beneficiary that it reduces their well-being, particularly if the person has no experience managing the various complexities (both social and financial) that accompany their new-found wealth.

In general, well-being can be conceptualised in a number of different ways. The standard concept is to think of utility deriving from consumption. In textbook microeconomics, individuals are assumed to have well-defined preferences and operate in well-functioning markets in conditions of strong information. Provided we make a set of assumptions about how people make decisions (known as the axioms of choice), then the behaviour of people in markets can be said to reveal their preferences and provide a measure of their welfare. A strong advantage of this approach is that it does not require us to make interpersonal comparisons of utility nor does it require us to directly measure well-being.

Measurements of Well-Being
Direct measurements of well-being take many forms. The most simple and widely used is the single-item global assessment of happiness (also called emotional wellbeing), which essentially asks people to rate their subjective happiness on a ten-point scale. This measure is simple and clear and has been included in many of the world’s largest survey exercises and has been studied substantially. Fig 2 shows the World Health Organization's measure for recent emotional well-being.

Fig 2. The WHO-5

A more cognitive approach is to elicit subjective evaluation of one’s position, generally conducted by asking life satisfaction on a scale of 1-10. You find this in many longitudinal data-sets; the NCDS, G-SOEP and World Values Survey all ask "How satisfied are you with your life, all things considered?" or something similar.

Happiness/emotional wellbeing and life satisfaction may seem similar and are indeed highly correlated, but they are measuring different facets of well-being. Happiness assesses an affective component and life-satisfaction a cognitive component. It is too restrictive to suggest that well-being is a uni-dimensional construct and there are many attempts to conceptualise the multi-dimensional nature of well-being. One of the most commonly used in the economics literature is the idea that one derives satisfaction in different life-domains such as finance, health, marriage and so on. As outlined by Van-Praag et al, overall welfare can be viewed as a sum of satisfaction in each of the life domains along with information about the relative weighting of each domain in the overall life satisfaction. Fig 3 gives an example of this approach for Ireland during the Celtic Tiger years.

Fig 3. Life satisfaction for three domains remained stagnant during the Celtic Tiger. Source: Liam's Barrington Lecture (2009).

Aggregation of these kind of measures allow for interesting comparisons between countries, such as Fig 4 which describes happiness around the world.

Fig 4. Happiness measured by various metrics using data from 2010-12. Data from United Nations 2013 World Happiness Report.

Determinants of Well-Being
The last ten years has witnessed an explosion in interest in the topic of the determinants of well-being. Part of this interest derives from suspicion that the use of GDP statistics provide a poor measure of the actual quality of life in society. Substantial work has been conducted arguing that a number of factors limit the use of GDP statistics as measures including:; failure of GDP to account for important non-market; relatedly, the substitution of market for non-market provision of goods; failure of individual rationality as a full model of consumption and investment; failure of GDP changes to account for changes in expectations and norms.
The most development statement of the view that economic growth does not imply well-being is that of the demographer Richard Easterlin. Easterlin first documented the Easterlin paradox (1974), a decoupling of well-being from economic growth particularly at higher levels of income.

Fig 5. The Easterlin Paradox

This paradox has been hotly contested in the literature and its ramifications are, of course, staggering. A recent paper by Stephenson and Wolfers (2013) calls the Easterlin Paradox into question. Using the global Gallup world health and well-being survey, they find strong evidence that economic welfare and well-being are strongly related, even at very high levels of income. The strength of their results led one commentator on the paper (Alan Kreuger) to suggest that the Easterlin paradox be renamed the Easterlin conjecture.

Eastern & Angelescu (2009) hit back strongly at these results, arguing that they confuse the implications of the original argument. They restate the Easterlin paradox starkly as: “Simply stated, the happiness-income paradox is this: at a point in time happiness varies directly with income, but over time happiness does not increase when a country’s income increases” (p2). Furthermore, they argue strongly that the Stephenson and Wolfers paper commits a logical fallacy: “As will be seen, the dissenting view appears to be largely the result of failing to distinguish between the short- and long-term temporal relationship between happiness and income. Over the short-term, when fluctuations in macroeconomic conditions dominate the relationship, happiness and income are positively related. Over the long-term, happiness and income are unrelated” (p2).

Individual determinants of wellbeing
As well as examining average national levels of well-being and their relation to macroeconomic aggregates, the literature on well-being is heavily focused on examining the individual determinants of well-being and how these can be influenced by policy. A recent review by Dolan, Peasgood and White provides a detailed review of the literature on the determinants of well-being. They highlight poor health, separation, unemployment and lack of social contact as factors that are substantially negatively associated with well-being. Their review also highlights the considerable evidence against income being a strong driver of well-being beyond a certain level.

Hsee and colleagues argue that money contributes to happiness to the extent that it can contribute to the purchase of what he calls "inherently evaluable" goods, which are goods that contribute to some basic primary need e.g. the need for basic social interaction, shelter, food, freedom from pain and so on. These are distinct from "inherently inevaluable" goods, where the consumption experience depends on the relative desirability of the good. The authors give the example that people need to learn that a higher karat diamond is more valuable than a lower karat one; it is not something they innately understand. Most financial improvements are evaluated relative to social benchmarks and many are subject to phenomena such as hedonic treadmills, habituation and social reference effects.

A further aspect of the relationship between economic possessions and well-being that is highlighted throughout the literature is that people value losses more than gains. The pioneering work of Kahneman and Tversky on prospect theory derived simple value functions that have been used to explain a wide range of economic behaviours. Related to loss aversion is the endowment effect (see part v for an explanation of both concepts), the tendency for people to value something more consequent on possessing it.

Unemployment and wellbeing
A 2008 review by Blanchflower points to a negative effect of unemployment on well-being both at individual and aggregate level. Wolfers (2003) shows that both unemployment and inflation decrease well-being and also finds a volatility effect, with more volatility independently lowering well-being. In general, I have not come across papers that argue against a causal effect of unemployment on well-being but would be interested to read any if I have missed them.

Case Study 1: Suicide and the Economy
While the debate might be shifting toward the side of gdp being a positive influence on well-being, the literature on suicide and gdp, in my opinion, is certainly not conclusive. In the Irish case, one only has to look at the last 20 years to know that suicide need not reduce during dramatic economic improvements and was, in fact, increasing at its highest rate in Ireland at the time of the most dramatic improvements in economic conditions we had witnessed. This is something that needs to be explained further.

Mary Daly and Daniel Wilson have written a number of papers about the economics of suicide. Fig 6 is from their paper "Happiness, Unhappiness and suicide" and it confirms the idea that time trends in suicide and overall well-being are apparently unconnected. I have talked about this in several seminars for Ireland and have raised it as a puzzle but am frequently told that this is unsurprising. The paper below finds that the determinants of suicide at the micro level are very similar to the determinants of well-being and thus, they argue that suicide should still be viewed as a strong measure of utility. I am still curious though as to why this relationship does not hold at time series level.

Fig 6. U.S. happiness distribution and the suicide rate over time.

Case Study 2: Well-Being and The Great Recession:
Given the continuous media and social coverage of the recession, it is worth asking the question again as to why it matters. The answer lies in the emotions we feel about different types of possessions. Without emotions, recessions would not have a human welfare component. Therefore, a good place to start in thinking about the negative effects of recession lies in these types of emotions. Fear, sadness and anger are three emotions that may be heightened in the context of economic loss. We need to understand further how these emotions are triggered in the context of recessions and whether they may "overshoot".

(i) Fear of the future. Some of this is rational concern about future reductions. A key question is the extent to which brain centres responsible for fear can "overshoot" in the context of continuous presentation of negative stimuli. In particular, it is worth considering whether continuous threatening images can lead to patterns of inertia. We do not have statistical evidence for this in Ireland but I think its plausible and worth testing that at least part of the negative effects of recession comes from a fear overshoot that inhibits search and active attempts to improve one's position.

(ii) Sadness due to the loss of possessions. This is a key one. What types of possessions loss make us feel most sad? The reduction in consumption of various goods is one consequence of a recession. Do we really believe though that this is the main driver of increased unhappiness during recessions? Sadness may also be triggered by the lost of one's job. The relation between unemployment and well-being has spawned a very large literature. My interpretation of this literature is that the loss of income is only one small component of the overall emotional loss from unemployment and that factors such as the loss of the social work environment and a loss of one's work identity are at least as important. Another loss that has been discussed in the literature is the loss of one's home security or home value. The feeling of seeing one's home values reduced is likely to affect people emotionally but we do not know enough about the factors at play here. It seems intuitive to me that people who purchased in different years will have different reference points and that a reduction may be less painful from a psychological perspective if its coming from an asset that inflated from a low base. Also, the emotional cost of home repossession has not been examined in depth though I have blogged before on some good papers in the UK context.

(iii) Anger and indignation, related to a perception that we are being treated unfairly by those with power over our lives. Such emotions may be genuine reactions to economic circumstances or strategic mechanisms used (perhaps unknowingly) by groups to influence political outcomes. In the face of a group that feel strong indignation about pay cuts, his task in doing this becomes harder. Knowing this, any group should ensure that its members are sufficiently susceptible to anger consequent on their position being reduced. Over time though, people may forget that this was the purpose of the emotion and actually begin to feel them. Its hard to act a part without taking some of it in, and others never know its an act.

Mental Health and Policy
The intersection of economics, wellbeing and mental health is clearly an important part of this for a number of reasons, including the extent to which mental health policy involves allocation of resources, the extent to which mental health reacts to economic circumstances and more generally the extent to which mental health considerations should be factored into how economic policy is conceptualised. Below are some broad areas that should be considered in developing a debate on mental health in Ireland that has an Economics input.

(i) How should well-being and mental health be used as indicators of societal progress? The Sarkozy commission has been influential in placing broader quality of life of indicators into the recent policy debate. It should be a prime feature of the current debate as to how this might be done in the Irish context, including how to make distinctions between overall well-being and the prevalence of acute psychological distress.

(ii) How is unemployment related to mental health? This is particularly relevant in the current economic situation. It is clear that there is a bi-directional relationship between unemployment and mental health, with psychological distress both being a consequence and cause of unemployment. There are many subtle interactions between the two that have important consequences for areas like job activation policy.

(iii) How does debt and poverty influence psychological distress? As with unemployment, the relationships are more subtle than just looking at the effect of being heavily indebted on mental health. Debt may influence mental health, but mental health response may also influence how people deal with their debt situation and this may again have important consequences for how debt policy is designed and implemented. Mullainathan & Shafir's new book Scarcity examines this topic.

(iv) How do economic fluctuations influence suicide rates? There is a large and somewhat inconclusive literature on the extent to which changes in GDP influence suicide rates. In Ireland, historically high rates of suicide were observed during the highest growth periods of the Celtic Tiger phase. This trend may have partly resulted from changes in alcohol and drug consumption that may mediate between economic fluctuations and suicidal behaviour. An approach to suicide that focuses on aggregate fluctuations themselves is not likely to be useful for policy. What is more interesting is to examine how different motivations for suicide change over the economic cycle and the cross-sectional variation in suicide. The data to do this in Ireland are, predictably, very weak and if the government are serious about mental health policy this is a clear priority in terms of research.

(v) To what extent does childhood mental health predict economic outcomes throughout life? A recent paper by Goodman et al. (2011), as well as others by James Smith, found dramatic effects of childhood mental health problems on outcomes all across the life-cycle. This research establishes a strong case for addressing childhood mental health from an economic and social perspective. However, far more research is needed on the economics of childhood mental health treatment. There is a paucity of research on the effects of standard psychotherapeutic and pharmacological treatments on the long-run outcomes of children. If we are going to make a large public economic investment in mental health treatments for children and adolescents it is important to know that these treatments will have a positive effect on life-time outcomes such as employment, life satisfaction and so on. This is clearly a complicated area but it should be put on a more inter-disciplinary footing in Ireland rather than being restricted to clinical discussion.

(vi) How does an aging population affect the relationship between economics and mental health? A number of studies including SHARE and TILDA in Ireland are examining the interactions between psychological, economic and social outcomes in aging. Issues such as the welfare cost of chronic illness pain, the extent to which psychological outcomes are better in home rather than hospital are, the influence of job environments among older people etc., are all big areas of research and will grow in interest over the next few decades.