Figure 1: Picture trade union ANAB left panel
Happiness economics is still in full development. The present column elaborates on several controversial points in this discipline. Various types of utility are studied, among others of the cognitive and affective type. The phenomenon of unemployment is an illustration of this. All in all the monistic utility seems yet to be most useful. Observations of feelings have a rate-scale, with a fixed reference point. Happiness increases with the individual income, also at the national level. So there is no reason for state paternalism, such as a policy of braking growth.
The Gazette has paid attention to happiness economics in various columns, because it reveals the themes, that are important for the human well-being. It can even principally express happiness in money. The columns show that happiness economics is a useful complement of the common economics. It shares this position with behavioural economics. However, since happiness economics is rising, it still contains many controversies. The present column wants to solve a number of these controversies. The arguments are mainly taken from the book Geld macht doch glücklig (in short Gmdg) by the German economists J. Weimann, A. Knabe and R. Schöb. But other sources are also consulted1.
The English philosopher Jeremy Bentham is convinced that human behaviour is determined by the pursuit of the maximum quantity of utility. Individuals try to enjoy their life, and to avoid suffering pain. Pleasure and pain can compensate each other, so that their nett effect can be expressed by the utility function uj(xk). In this function, j refers to the concerned individual, and the variables xk represent the various state variables, with for instance k=1, ..., K. The Americans sociologist Amitai Etzioni calls such an all-encompassing function a mono-utility, whereas the Dutch economist J.J. Graafland uses the expression monistic happiness. The philosophy of Bentham is also called hedonism. It assumes that man is an egoistic being. However, the view of Bentham is controversial.
Several millenia before, the Greek philosopher Aristotle has already equated happiness to the full individual unfolding (in the Greek language eudaimonia)2. Man must strive for a high level of virtue and meaning. A virtuous character requires a sense of duty, and thus the adherence to morals. This clashes with hedonism, because sometimes pain must be endured for morals' sake. The English economist J.S. Mill has tried to reconcile both views by introducing the bi-utility, that is to say, happiness of a lower and higher order. For, virtue is not short-lived, but gives a lasting pleasure. Unfortunately, then utilities become incomparable (incommensurable). Graafland calls this pluralism. This has the consequence, that there no longer is a nett utility uj. Moreover, in this way the egoistic man is replaced by a virtuous and social type.
Loyal readers will not be surprised by the philosophy of Aristotle, because it is related to communitarianism, which is decribed in a recent column. Etzioni has built his I&We model on it. Egoism has a logical-empirical character, whereas virtue is normative-affective. In other words, material considerations are cognitive, whereas moral considerations are emotional3. On p.106 in Gmdg the short terms cognition and affect are used. However, it is controversial whether models of the I&We type are fruitful. For, in such models the human behaviour becomes difficult to predict, because morals differ in space and time. Models that assume individualism do allow to make concrete predictions. Therefore it is worth trying to keep the morals out of the model.
A hallmark of eudaimonia and of the I&We philosophy is, that the importance of outcomes is seen as relative. Like the German philosopher Immanual Kant argues: the intention of activities itself makes already happy. The hedonistic consequentialism is rejected. Greed and jealousy are vices, whereas reversely soberness is advocated. All these commands of the eudaimonia encourage to find happiness in acquiescence, even under poverty or miserable circumstances. This philosophy agrees with human nature. Namely, behavioural economics shows, that the human mind remains fairly stable, even under adversity (see p.128 in Gmdg). People have a striking mental capability to adapt4. Apparently the affective happiness, virtue and the eudaimonia do not incite to improve the quality of life (p.143 in Gmdg).
The quality of life is in essence material, and is therefore related to the cognitive happiness. This is also the domain of happiness economics. The cognitive happiness turns out to be a good indicator, especially when policy proposals must be formulated for increasing welfare. It is true that the affective happiness must also be taken into account, but it must not get the dominating position, that for instance Etzioni wants to give it.
It is known from studies in happiness economics, that unemployment makes unhappy. This is measured by asking the unemployed about the satisfaction uj with their life. The measured happiness is cognitive, because a well-considered answer is demanded. Next the anwers are compared with those of employed in otherwise similar circumstances. It is especially important, that in the comparison the incomes of the employed and the unemployed are equal. In this manner the researchers eliminate differences in satisfaction due to income effects. The reason for the discontent of the unemployed is that a job yields, besides the income, also several social benefits. They have already been summarized in a column of several years ago. Notably, work offers an identity, and comradeship5.
In view of the preceding paragraph it is now interesting to analyze the affective happiness of the unemployed. This is studied in chapter 6 of Gmdg6. Incidentally, this study ignores the normative component, so that here the theme is not the eudaimonia. The unemployed simply answer the question how they feel emotionally at a certain time. During the whole day this question is repeated regularly, so that the feelings during various activities are measured. This approach is called the day reconstruction method. The feelings are multiplied by the duration of the activities, and subsequently the happiness in all these periods is summed. This yields a daily or weekly total Gj. In formula this is
(1) G = Σt=1T τt × gt
In the formula 1 there are daily T activities, each with a duration of τt. The corresponding affective happiness per unit of time is gt. Depending on the method that is used to measure gt, G is called the periodical nett affect or the summed episode satisfaction. It turns out that the results for G are quite similar, irrespective of the selected method. As an illustration, the table 1 gives a summary of empirically found values for the nett affect, albeit not for unemployed.
| t | τt in hours | gt (values between 1 and 6) |
|---|---|---|
| sex | 0.2 | 4.8 |
| leisure | 2.2 | 4.0 |
| diner | 0.8 | 3.9 |
| sports | 0.2 | 3.9 |
| watching television | 2.2 | 3.7 |
| house-keeping | 1.1 | 3.0 |
| raising children | 1.1 | 3.0 |
| work | 6.9 | 2.7 |
| commute to work | 0.5 | 2.1 |
It turns out that gt is also lower for the unemployed than for the employed, just like is the case for the cognitive satisfaction. This effect is called saddening (p.140 in Gmdg). On the other hand, the unemployed are free ti plan their pursuits, and so can choose enjoyable activities. In other words, they have some freedom in their choice of τt. The empirical results show that this can compensate the saddening effect. Thanks to their leisure time the unemployed have, all in all, about the same daily Gj as the employed8.
It turns out that this study displays the remarkable phenomenon, that the weekly nett affect Gj hardly correlates (depends on) with the cognitive satisfaction uj of the interviewed persons. Here it must be stressed once more, that these two measures relate respectively to an emotional value and to a rational evaluation. Since the cognitive happiness uj correlates strongly with material matters such as income and health, apparently the nett affect Gj is mainly immaterial. The unemployed can reasonably maintain their weekly nett affect in the daily routine, thanks to an emotional adaptation, which perhaps consists partly of acquiscence. This measurement suggests that the misery of unemployment is less than expected. In other words, when one wants to get an impression about the objective quality of life, then the nett affect Gj is not a particularly useful indicator.
What does the difference of uj and Gj tell about the models of the I&We type? Apparently emotions and rational evaluations can indeed not be summed simply. They are not commensurable. In the emotional (affective) domain individuals succeed better in adapting to set-backs, than in the rational (cognitive) domain. In the remainder of this column it will be argued, that precisely the cognitive happiness is more important for the social well-being. Therefore a bi-utility or plural utility is not very meaningful. This is the conclusion in Gmdg. Your columnist notes, that apparently the happiness research is stil developing. The most fruitful indicators must yet be found. As long as consensus about standards of measurement is lacking, happiness economics is not a mature science9.
In economics it is common to measure the social progress by means of the gross domestic product (in short GDP) per capita. More production is better for the general well-being. For, according as more is produced, the number of consumptive alternatives of the households is larger. This is indeed true, provided that some nuances are taken into account. Namely, the realization of growth requires that the labour productivity is raised, and that can have negative side effects for people. A sound measure of progress must also take into account these side effects.
This is exactly one of the reasons, why happiness economics has recently become popular. Happiness economics does not just measure the positive effect of the income on the well-being uj, but it can also quantify other sources of well-being. Already two years ago a column in the Gazette has shown that the Dutch economist B. van Praag couples well-being to for instance housing, health, leisure, and social life.
However, for the moment the measured happiness (or satisfaction) is not universally accepted as an indicator of the general well-being. One prefers to try calculating composite indicators from various sound economic and social variables. Such attempts are already made since the sixties of the last century. Here there is naturally the gigantic problem, that both the choice of the variables and their weighing is purely subjective. It is almost impossible to justify these choices in a scientific and objective manner (p.79 in Gmdg). Thus the danger threatens that an uncontrolled growth of composite indicators occurs, which are all advanced in political clashes.
For the present, actually just one indicator has received a wide recognition, and that is the Human development index (in short HDI), which dates from 1990. The value of the HDI depends for a third on the GDP (per capita), for a third on the national health (namely the life expectation at birth), and for a third on the education (its average and expected duration), The well-being is higher, according as the HDI has a higher value. The HDI is promoted by some, because it can rise significantly, notably in Third World states, even when the economic growth remains modest. The figure 2 expresses this10. Nevertheless, the reader will acknowledge, that the HDI is not ideal as an indicator of policy.
It is interesting, that in the past decades politicians have also attempted to find composite indicators for the Well-being. Very well-known is the commission, that the French president Sarkozy installed in 2008 (p.77 in Gmdg). This commission consisted, among others, of the economic heavy-weights Stiglitz, Sen, Kahneman, Arrow, and Putnam. The commission obviously concludes, that the GDP per capita is an incomplete indicator. A problem is that it is difficult to value the public services within the GDP. A composite index would be better. The commission indeed mentions several variables, that can be important besides the GDP. However, a recent study has shown, that these variables correlate strongly with the GDP (p.79 in Gmdg). Therefore it is questionable, whether such an index is worthwhile11.
Now your columnist puts aside the composite indicators, and elaborates on the (monistic) mono-utility uj of an individual j. First it will be explained how the measuring scale of uj has been constructed. Among others, the physiologist E.A. Weber has discovered, that sensorial observations such as listening have a lower and an upper boundary. In order to distribute the observations regularly along the range of measurement, the senses register relative changes, and not the absolute changes (p.104 in Gmdg). A familiar example is the human hearing, which registers the level L of sound pressure. This L is not the sound pressure p itself, but its relative change. In formula the relation between L and p is
(2) L = L0 × ln(p / p0)
In the formula 2, p0 is a reference sound pressure, and L0 is the level of the sound pressure, which corresponds e×p0. Such a measuring method is called ratio scaling. The advantage of such a scale is that the senses can also accurately observe signals, that are much smaller than p0. Later the psychologist G.T. Fechner has applied the relation in the formula 2 also to the observation of psychical feelings (p.105 in Gmdg). Perhaps he believes that feelings are also physiological, because they stem from the nervous system. Modestly, he has called this relation of observed feeling the law of Weber-Fechner. Since happiness or satisfaction is a psychic phenomenon, it is logical that uj also reacts to external stimuli p in accordance with the law of the formula 2. Nowadays, this is indeed commonly assumed12.
Measurements of the individual utility uj commonly restrict the measured values to a bounded interval, say from 1 to 10. Then the interviewed persons j can attach a natural number between 1 and 10 to their experienced satisfaction. Thus they are limited in their choice, because the formula 2 can principally yield all rational numbers. Notably at the extremes this restriction could lead to a distortion of reality (p.105 in Gmdg)13. Therefore it is necessary to choose the reference point p0 for the stimuli in such a manner, that extreme values of uj are very rare. However, this is somewhat problematic. Suppose for instance that the stimulus p consists of the income y. Due to the economic growth, y will continue to move away from the reference income y0, according as time advances. The income becomes extreme, and therefore the measuring scale would become ineffective.
Fortunately, the human nature has supplied a solution for this problem. Namely, according as a new situation becomes durable, man gets used to it. He adapts to the changing life, and therefore that becomes his new reference point. For instance, y0 can shift upwards with time. Van Praag calls this the preference drift. Kahneman uses the expression experienced utility, which due to the changed reference point deviates from the earlier decision utility. This is quite convenient for the effectiveness of the sensory observation, but also has the disadvantage that the individual measuring scale changes. When a time series of the individual utility is measured, then the researcher must always be aware of possible shifts of p0 (p.29, 102 in Gmdg)14.
It has even been said, that due to the upward shift of p0 (often in the context of the individual income y0) each improvement in the situation is compensated in such a manner, that the individual does not become happier. This supposed phenomenon is called the Easterlin paradox. The authors of Gmdg have in part written their book, because they want to refute this statement (p.127, 130 in Gmdg). The statement indeed seems a bit weird. It is difficult to imagine, that people would have no recollection of their earlier situation, or would be completely cut off from situations beyond their own reference group.
It is worth mentioning, that on p.145 in Gmdg the Rayo-Becker model is explained. This model relates the shifts of human reference points to the evolution. People take their existing situation as the historical reference point, because then they are stimulated most to realize improvements. Each individual has the potential to make his environment more pleasantly. That is to say, individuals with such a reference point have better chances of survival than others, who havy difficulty in comparing their well-being with something. The evolution model is an interesting find, although for the moment your columnist does not see its practical relevance.
In 1974 the economist R.A. Easterlin analyzed socio-economic data, that were collected in the World value survey. This is an international study, where various states repeatedly collect the same data. The goal is evidently to mutually compare the performances of the various states. Easterlin has notably studied, how the general satisfaction uj varies with the GDP (say, the national income y) per resident j. The analysis of Easterlin suggests, that above a national income per capita of $15.000 (expressed in the dollar value of 2009) the extra income does not add anything to the individual happiness. The figure of Easterlin has been shown before in a column of the Gazette, and it is repeated here schematically in the figure 4 (red curve, see p.121 in Gmdg). It seems that at least above $15.000 the economic growth is meaningless for the general well-being. It has just been stated, that this discovery looks like a paradox.
The authors of Gdmg have studies the Easterlin paradox again, this time with data from the Gallup world poll. This collection has a better quality than the World value survey, among others because more states are included, such as many poor states from the Third World. The result is also presented in the figure 4 (green curve). It is clear that now the satisfaction does continues to rise with the income per capita. This refutes the Easterlin paradox15. The green curve is explained by two phenomena. On the one hand, the rising income yields more satisfaction, in agreement with the formula 2. On the other hand, the rise of the national income will also raise the reference point y0 of the individuals somewhat. This slightly tempers the satisfaction, but it does not eliminate her.
Note that all these phenomena can become a potential problem for the scaling of happiness measurements. For, when the income rises faster than the reference point, then according to the formula 2 the happiness (or utility) will also continue to rise. Then people can be impeded by the upper boundary of the measuring scale (say, 10) in the realistic numerical valuation of their great happiness. However, at present this seems still to be a hypothetical problem. The figure 4 shows that even yearly incomes of $50.000 give a general satisfaction of merely 7.5. It seems obvious that at still higher incomes the relative increase of happiness flattens further. Apparently the existing scale will remain usable for the near future.
The preceding discussion is essentially scientific. However, happiness economists like to apply their findings for improving the political policies. Especially the English economist R. Layard has attracted the attention with his policy proposals, which sometimes look almost draconian16. His starting point is that the decisions of individuals take insufficiently into account the rise of their reference point. They want to work hard in order to raise their income, without realizing that other individuals have the same urge. Indeed, they will probably succeed in their ambition, but their reference point will increase correspondingly, because their environment becomes richer as well. A race to the top would emerge, which makes nobody any happier. This is called a zero-som game. Apparently, the people themselves sometimes do not succeed in making rationally optimal choices.
Now, it is known from behavioural economics, that indeed people often impulsively take wrong decisions. They have trouble in realizing their individual autonomy in a responsible manner. This is called irrational behaviour. Then the state must sometimes intervene in a paternalistic manner, such as by means of the compulsory education, or saving for pensions, or social security, or the interdiction of using drugs. In the same vein Layard now suggests the state to tax the high incomes so heavily, that hard work is discouraged. Reversely, Layard states that individuals become happier, according as they participate more in social life. Also here, he propagates state paternalism, by means of subsidies for voluntary activities. The opportunities for advertising products are diminished. The education must invest more in meaning and morals. Health care must focus more on mental problems17.
Although nobody will deny the need for some state paternalism, happiness economics has hurt its own good reputation somewhat by radically propagating paternalistic politics. It has indeed provoked resistance among other economists. The German economist H. Beck stresses, that paternalism hurts individuals, when they do decide rationally, for instance to work hard. Consider also the compulsory cooling-off periods before a divorce is allowed, or perhaps even compulsory marriage counselling, etcetera. Paternalism could of course be abused by politicians, in order to increase their own well-being in this manner18. Also p.85 and further in Gmdg are reluctant to accept the proposals of Layard. On p.153 it is stressed, that competition within society leads to better performances, which increases the general well-being. Reversely, acquiscence stifles innovation.
Yet it is somewhat bitter, that Layard wants to curb the economic growth by means of state interventions, considering that growth does add to the general well-being. Thanks to growth, the health care, education, and the physical infrastructure can be improved. That undeniably raises the quality of life (p.131 in Gmdg). Nevertheless, the call for more paternalism, and even for a purposive reduction of the economic growth, meets with a wide response among some politicians, who like to introduce strong morals. In the Netherlands a typical representative is Femke Halsema, who rejects "hyper-consumption" and "agitation". She refers to the conclusions of Layard for the support of her own ideology. Incidentally, here the aim is not to present Halsema as a scape-goat, but merely to give an illustration. For, may others defend similar views as she does19.
Nor does your columnist want to state, that the green curve in the figure 4 is definitely more accurate than the red one. He merely concludes, that the red curve is scientifically quite controversial, and thus the Easterlin paradox as well. Moreover, she is intuitively illogical. Then it is somewhat disconcerting to see, that some happiness economists (with Layard as the shining example) want to execute a radical reform of society, and justify it with the supposed paradox. Then science turns into a dogma. It is an arrogance of science, which irritates and even alarms, especially when the dogma acquires a broad political support20.