I define happiness, subjective well-being, and welfare of an individual as essentially equivalent terms, with only two minor differences in usage. First, we tend to use ‘happiness’ to refer to the current feeling. If I see you singing and dancing gladly, I tend to say that you are happy now. If I know that you are healthy and have a loving spouse, nice children, good income, etc., I tend to conclude that your welfare must be high, because you will likely be happy for a long time. Second, we tend to use ‘happy’, ‘happiness’ in a less formal way and ‘subjective well-being’ and ‘welfare’ in a more formal way. For example, as mentioned by Diener and Tay (2017, p. 90), ‘Subjective wellbeing, often called “happiness” in layperson terminology, refers to peoples’ sense of wellness in their lives’. If we ignore the formality in usage and holding the time period as given, I find that my happiness, subjective well-being, and welfare must be exactly identical. If a person is unhappy through her whole life, her welfare cannot be high; her subjective well-being cannot be positive.

Since most, if not all individuals care much about their own welfare, preference and welfare normally go together (Benjamin et al. 2012). However, the preference of an individual may differ from her happiness or welfare for the following three reasons.

First, preference may differ from welfare due to ignorance and imperfect foresight. While an individual may prefer x to y, believing he will be better off in x than in y, it may turn out to be the other way around. This is the question of an ex-ante estimate versus ex-post welfare. While the ex-ante concept is relevant for explaining behavior, it is the ex-post one which is the actual welfare.

Second, the preference of an individual may not only be affected by her own welfare but may also be affected by her consideration for the welfare of other individuals. Thus, it is possible for a person to personally have a higher happiness level in y than in x, yet herself prefers and chooses x over y because she believes that other people are happier in x than in y. While it is true that the belief that other people are happy may make her happy, these positive feelings may not be strong enough to outweigh the loss that she has to suffer for changing from y to x. For example, a person may vote for party x, knowing that she herself will be better off with party y in government. The reason she votes for x is that she believes that the majority of the people will be much better off with x. This itself may make her feel better (affective altruism). However, this external benefit may not be important enough to overbalance, in terms of her subjective happiness, her personal loss, say in income, under x. She may yet vote for x due to her moral concern (non-affective altruism) for the majority. To give an even more dramatic example, consider an individual who expects to lead a very happy life. When her country is being invaded, she may volunteer for a mission which will bring her the certainty of death. The prospect of being a citizen of a conquered nation especially with the guilt conscience of failing to volunteer for the mission may not be too bright. But overall she may still expect to be fairly happy in leading such a life. Yet she chooses death for the sake of her fellow countrymen. In this case, she is not maximizing her own welfare.

Some economists have difficulty in seeing the above distinction between preference and welfare, saying that whenever an individual prefers x to y, she must be, or at least believe herself to be happier in x than in y. This difficulty completely baffles me. Clearly, a father (or mother) may sacrifice his (her) happiness for the welfare of his (her) children. I cannot see why similar sacrifices cannot be made for a friend or a relative, and further for a countryman, any human being, and finally, any sentient creature.Footnote 1

It may be doubted that the existence of true non-affective altruism is inconsistent with Darwinian natural selection. However, as preferences are the result of both cultural and genetic inheritance, one can demonstrate that pro-social traits could have evolved under the joint influence of cultural and genetic transmission.Footnote 2

If some readers still doubt the existence of truly non-affective altruism, they are likely to be convinced that in fact they themselves possess some degree of non-affective altruism by considering the following hypothetical choice. Like Einstein’s thought experiments, such hypothetical exercises cannot be dismissed on the ground of being unrealistic. Suppose that you are asked by the Devil to press either button A or B within 2 s. You know with certainty (for simplicity of comparison) that one of the following will happen depending on which button you press. Within these two seconds, you will be so preoccupied with pressing the right button such that your welfare will be zero whichever button you press. After pressing, you will lose memory of the present world and hence will not have feelings of guilt, warm-glow, or the like, related to which button you press.

  • A: You will go to Bliss with a welfare level of 1,000,000 trillion units. Everyone else will go to Hell with a welfare level of minus 1,000,000 trillion units each.

  • B: You will go to Bliss Minus with a welfare level of 999,999 trillion units. Everyone else will go to Niceland with a welfare level of 999 trillion units each.

  • C: If you do not press either button within the 2 s, you and everyone else will go to Hell.

By construction, choosing A will maximize your welfare but most people will choose B out of non-affective altruism. If you still think that you will choose A, change Bliss Minus into a welfare level of 999,999.999 trillion units. If you still opt for A, I have to concede that you are not altruistic non-affectively. But how could you have the heart to condemn all others to Hell for a fractional increase in your own welfare? (In my view, the existence and degree of non-affective altruism marks true morality.)

Third, an individual may have irrational (or imperfectly rational) preferences. The preference of an individual is here defined as irrational if he prefers x over y despite the fact that his welfare is higher in y than in x, and his preference is unaffected by considerations of the welfare of other individuals (any sentient creature can be an individual here), or by ignorance or imperfect foresight. The definition of irrationality here is so as to make the three factors discussed here exhaustive causes of divergence between preference and welfare.

While few, if any individuals are perfectly ignorant and irrational, some degrees of ignorance (or imperfect information) and imperfect rationality clearly apply to most individuals.Footnote 3 However, some alleged irrationalities could be simply due to errors, computational limitations, and incorrect norm by the experimenters (Stanovich and West 2000). There are a number of causes that may make preferences differ from happiness other than ignorance and a concern for the welfare of others, and hence they are irrational according to our definition here. The following two (may not be completely independent) causes may both be explained, at least partly, by some biological factors (On the biological basis of social behaviour, see, e.g. Wilson 1975; Crawford & Kreps, 1998; Nicolosi and Maestrutti 2016).

First, there is a tendency for many people to discount the future too much or even to ignore it completely. This is widely noted, including by economists. For example, Pigou (1912, 1929, 1932, p. 25) called it the “faulty telescopic faculty”, Ramsey (1928, p. 543) called it “weakness of imagination” about the future, and Harrod (1948, p. 40) regarded it as the “conquest of reason by passion”. A discount on future consumption, income, and any other monetary value is rational as a dollar now can be transformed into more than a dollar in the future. A discount on future utility may still be rational if the realization of future utility is uncertain. (For healthy people, this uncertainty is usually very small per annum.) Discounting the future for more than these acceptable reasons is probably irrational. A manifestation of this irrationality is the insufficient amount of savings for old age, necessitating compulsory and heavily subsidized superannuation schemes. I came across an extreme example of such under-saving during a survey regarding how much people would be willing to save more if the rate of interest were higher (Ng 1992). The question implicitly assumed that everyone did some saving, as the answers were in terms of how many percentages more one would save. One subject declared that he did not save anything. I then asked him to change the answers to be chosen from “saving 20% more” into “saving $20 more per month”. He said he could not be persuaded to save anything at whatever interest rates (500% was mentioned). He only conceded willingness to save when I said, “What if a dollar saved now will become a million dollars next year?” I was careful enough to find out that this healthy-looking young man was not expecting early death from a terminal disease or the like.

The behaviour of most other animals is largely determined by pre-programmed instincts rather than the careful calculation of the present costs versus future benefits. The storing of food by ants, the burial of nuts by squirrels, etc. are largely, if not completely, instinctive. If calculated choices are made by animals, they are largely confined to sizing up the current situation to decide the best move at the moment, like fight or flight. The ability to anticipate the rewards in the fairly distant future requires much more ‘reason’, ‘imagination’, and ‘telescopic faculty’ than normally cost-effective to program in most other species. However, we know that we are endowed with some such faculty. Nevertheless, since this advanced faculty is almost completely absent in most other species, it is natural to expect that it is not fully developed even in our own species. Moreover, different members of our species may be endowed with different degrees of such faculty. The existence of a significant proportion of members of our species which do not possess a full telescopic faculty is thus not surprising.

Secondly, there are the excessive temptation of pleasure (especially present pleasure vs. future costs, hence related to the preceding cause) and the powerful biological drives. After the evolution of flexible species (defined as one where the behavior of its members is not completely determined by the automatic programmed responses but also by choice), natural selection ensured that the flexible choices made were largely consistent with fitness by endowing the flexible species with the reward-penalty system. Thus, eating when hungry and mating with fertile members of the opposite sex are rewarded with pleasure, and damages to the body are penalized with pain. (This makes the flexible species also “rational” as defined in Ng 1996 which shows that complex niches favor rational species which in turn make the environment more complex, leading to a virtuous cycle that accelerates the rate of evolution. This partly explains the dramatic speed of evolution based mainly on random mutation and natural selection, a speed doubted by creationists.) On top of the ex-post rewards and penalties, we are also endowed with inner drives to satisfy the fitness-enhancing functions like mating. On the whole, these powerful temptations and drives work in the right direction, making us do things that both enhance our biological fitness and psychological welfare. However, since evolution is largely fitness-maximization and the welfare-enhancing aspect is only indirectly to enhance fitness, some divergence between our behavior and our welfare is unavoidable, as our behavior is not completely determined by rational calculation, but also partly by programmed inclination, including the drives. (See Ng 1995 on the divergence between fitness and welfare maximization especially with respect to the number of offspring.)

It has also been shown that, ‘wanting’ or preference and ‘liking’ or welfare are mediated by different neural systems in the brain and are psychologically dissociable from each other. In other words, an individual may prefer something without liking it or prefer something more strongly than could be justified by his liking for it, and vice versa. In particular, neural sensitization of brain dopamine systems by addictive drugs may create intense ‘wanting’ way beyond that which could be explained by ‘liking’ and the relieving of withdrawal symptoms. (See Berridge 1999 for a review.) As an example of these powerful drives, adolescent girls and boys often engage in careless sexual acts propelled by their sexual drive and the temptation of sexual pleasure, even if there is a high risk to their long-term welfare, such as in the case of unwanted pregnancies or the contraction of AIDS. While this is partly due to ignorance, the role of biological drives cannot be denied.

Consider a specific example. Suppose that a person agrees that, for choices involving risks, the correct thing to do is to maximize expected welfare (assuming no effects on the welfare of others) and also actually do so for most choices. However, for choices concerning sexual activities, he chooses x over y, although his expected welfare is lower with x than with y and he knows this to be the case. Here, x may involve having sex with many persons without clear knowledge (this knowledge is assumed to be not feasible to obtain and hence not relevant) about whether they have AIDS. His (expected) welfare-reducing choice of x may be due to his biological inclination to seek many sexual encounters. He knows that doing so has a non-insignificant chance of contracting AIDS and hence is welfare-reducing. He has all the relevant feasible information and yet chooses (due to the powerful sex drive) x that he knows to be of lower expected welfare. (This is not really a hypothetical example. I am confident that, out of 100 normal adult males, at least 10 have actually made such choices. If one wants more solid evidence, one may look at the frequency of prostitution and extra-marital sex.) Should we call this preference informed as the person has all the relevant feasible information, or uniformed because it is not in agreement with his real interests?

The above two causes of irrational preference illustrate the point that, due either to imperfection in our endowed faculty or the biological bias in favour of reproductive fitness, we may do things not quite consistent with our welfare. The issue here is that, for normative purposes, should we use welfare or actual preferences/behavior. Clearly, we should use welfare instead of behaviour dictated by biological fitness. An old Chinese dictum says, “Out of the three un-filial acts, not having offspring is the greatest”. However, for the human species as a whole, we are certainly not getting smaller in population size. Moreover, a long-run social welfare function accounting for the welfare of future generations should account for that. If we go for biological fitness, we will prefer unlimited procreation even if that means that we will all be suffering as compared to a smaller population with a higher aggregate welfare. ‘We’ are the feeling selves that care ultimately about our welfare (positive minus negative affective feelings). We are not them, the unfeeling genes that, through random mutation and natural selection, programmed us to maximize fitness. Unlike other species who are almost completely controlled by their genes and the environment, we have learned to change our fate by using such measures as birth controls. For normative issues, it is our welfare, rather than the selected random dictates of the unfeeling genes, that should count. (On a survey of different concepts of individual welfare, see Ackerman et.al 1997; Diener et al. 2018.)

On top of the above two biological causes, there is another source of imperfect rationality. An individual may stick rigidly to some habit, custom, ‘principles’, or the like even if he knows that this is detrimental to his welfare and the welfare of others even in the long run, taking into account all effects and repercussions. Customs, rules, moral principles, etc. have a rational basis as they may provide simple guides to behaviour which may be, at least on the whole, conducive to social welfare. It would be too cumbersome and time-consuming if an individual were to weigh the gain and loss in terms of social welfare or his own welfare each time he has to make a decision. Thus he may stick to his routine, rules, principles, etc., without thinking about the gain and loss. If this results occasionally in decisions inconsistent with promoting his welfare and the welfare of others, it may be regarded as a cost in pursuing generally good rules. If, say, there is a change in circumstances, the adherence to some rules may result in persistent net losses in welfare, taking everything into account. An individual may stick to these rules without knowing that they are no longer conducive to welfare. Then the divergence between preference and welfare can be traced to ignorance. If he knows this and yet sticks to the rules, he is irrational.

Many readers may disagree with the definition of irrationality adopted here. For example, suppose a man sticks rigidly to the principle of honesty and would not tell a lie even if that would save his life and contribute to the welfare of others, taking everything into account. According to our definition here, he is acting irrationally. To those (like Kant) who are willing to accept honesty as an ultimate good in itself, he may not be irrational. (For our case against Kant’s categorical imperatives, see Chap. 5.) But let us consider such questions as: Why shouldn't a person tell a lie? Shouldn’t one lie to protect his people from a cruel invading army? If we press hard enough with such questions, I believe that most people would ultimately rely on welfare as the justification for any moral principles such as honesty. Personally, I take the (weighted or unweighted) aggregate welfare of all sentient creatures, or a part thereof, as the only rational ultimate end (my basic value judgment; more in Chap. 5), and hence define irrationality accordingly. I know the controversial nature of this definition. But fortunately, one does not have to agree on the definition of irrationality given here to agree with the arguments presented in this book. If preferred, the word ‘irrational’ as used here could be taken to read ‘irrational according to the objective of welfare maximization’.

However, are moral principles really fundamental? Before the evolution or development of morality and the like, we (perhaps still in the form of apes or even earlier ones) had no moral or other principles, no concept of commitments and justice, etc. Self-interest dominated entirely, although this does not exclude genetically endowed apparent ‘altruism’ for the maximization of inclusive fitness. As we evolved and increasingly relied on our high intelligence and social interaction for survival, the instinct for moral feelings also evolved which helped our survival by enhancing cooperation. This was enhanced by learning the importance of such moral practices as honesty in improving our struggle against nature (including wild animals) and against competing human groups. No one can deny that the initial evolution/development of morality must be purely instrumental (in enhancing either our welfare or our survival and reproduction fitness) as there existed no morality to begin with. We then learned and taught our children and students to value moral principles; this was first done to increase the degree of adherence to these principles, and consequently, our welfare. Eventually, some, if not most, people came to value these principles in themselves (i.e. regarding them as of intrinsic values) by learning and probably also by instinct. The evolution of such commitment enhancing devices as blushing can be fitness-enhancing; see Frank (1987). Failing to see the ultimate value is a kind of illusion fostered by learning (I dare not say indoctrination) and perhaps genetics. However, I personally have great moral respect for people with such illusions. They most probably make better citizens, friends and colleagues. But illusions they are nevertheless, at least at the ultimate analytical or critical level. While these illusions are on the whole positive (in maintaining moral standards), they do have costs, for example, in delaying the rejection of certain outdated moral principles.

While we recognize the three sources of divergence between welfare and utility discussed above, it is convenient to ignore the divergence except when we come to discuss problems (such as merit goods and the materialistic bias; Ng 2003 and chapters below) where the divergence is important. In other words, in the absence of specific evidence/considerations to the contrary, we assume that, as a rule, each individual is the best judge of her own happiness/welfare and chooses to maximize her welfare. Then the question of happiness/welfare measurability coincides with that of utility measurability, as discussed in Chap. 6.

One real-world example where the violation of the preferences of people actually improved their welfare happened decades ago (in mid 1960s) in Singapore under Lee Kuan Yew’s government. Lee decided to expropriate a piece of land used as a cemetery for certain public development without sufficient compensation. Existing tombs there had to be evacuated for reburial elsewhere. Such an excavation is regarded as an extreme disturbance of the peace of the dead and most survived children would not take millions of US dollars to accept such excavations. Even if the government had only to pay a small fraction of the amount people were willing to accept, the public development would certainly have turned out to involve net negative benefits. However, I certainly agree with Mr. Lee that the government should look after the welfare of existing (and future) people rather than that of the dead, even if this has to be in violation of the preferences of the people now. This welfare-improving decision in favour of development would not only certainly fail to pass the traditional cost-benefit test based on preference, it would also likely fail to pass the public choice test of democratic voting (also based on preference).

It is interesting to examine why preference fails in this case. First, it is partly due to the external costs created by the tradition of excessive respect for the ‘peace of the dead’. An individual failing to show due respect would run the risk of social disrespect. Some due respect for the dead may serve some useful function but it has become excessive due to a complex process of interaction, including the individually rational but socially harmful strategy of pretending to be very respectful. If this failure can be explained in the traditional analysis in terms of external costs, the next failure cannot. Secondly, even abstracting away the danger of social disrespect, individuals may have genuine preference for showing extremely high respect for the peace of the dead due to cultural influence. They may genuinely feel the importance of avoiding the excavation of the remains of their ancestors. However, if the decision for compulsory acquisition was made by the government, they would accept it as unavoidable and beyond their control and hence would suffer little loss in welfare. It is thus more than a publicness problem. If the decision were put to a vote, most of them may feel compelled by the respect for the dead to vote against excavation and development. However, if the decision were made for them by the government, most of them would not feel too distressed. Thus, Lee’s decision almost certainly increased social welfare despite being against the preferences of the people. (However, this example has some degree of exceptionality and does not justify autocratic decisions against the will of people in most cases.)

The Singapore example above is similar to the situation in ancient China (much less so now, but still applicable to some extent). It was a compelling duty of children to give a deceased parent a decent burial. Thus, one often reads in novels or watches in films how a poor man willingly sold himself to become a slave for a few years in order to give his deceased parent a decent burial. If the prevailing law does not allow such servitude, such a person will not become a slave, improving his welfare, if not his preference. If the law allows such a sale, he will feel unease without giving his parent a decent burial; if the law does not allow it, he will regard it as beyond his option and will have to contend with a simple burial, without much misgiving.

An interesting question arises as to the way we should classify the second factor accounting for the divergence between preference and welfare discussed above, according to our tripartite classification (imperfect knowledge, concern for the welfare of others, and imperfect rationality). It may be thought that, provided we include the dead under ‘others’, it should be classified as a concern for the welfare of others. However, until we have more evidence to convince us otherwise, I think that the dead are not capable of having welfare. Hence, it should be classified as imperfect rationality. According to our definition of rationality, it is not (perfectly) rational to have respect for the dead over and above contribution to the welfare of existing and future sentients, and apart from such concerns as the fear of social ostracization.

As an individual typically cares greatly about her own happiness, an individual’s satisfaction with life is highly correlated with her happiness. But, again, as one may also care about one’s contribution to others, the two may differ, as further discussed in Chap. 4. Many researchers use subjective well-being (SWB) as encompassing both happiness and life satisfaction (e.g. Adler 2017, p. 119). This is one of the reasons that make them think of happiness as multi-dimensional. Throughout this book, I use SWB or (individual) welfare as synonymous with happiness, and use life satisfaction separately. I believe this is less confusing and more consistent with the meaning of the various terms.