Would a good society seek to maximize a social welfare function?

 This article is about my personal experience in attempting to understand social welfare, the concept of a good society, and my role as an economist involved in the processes of social choice. I decided to write about this topic after writing an article for Savvy Street on the related topic, “Can social planning enhance individual flourishing?”

When economists talk about maximizing social welfare, they are referring to a concept that appears to have something to do with the well-being of people. However, the concept is best viewed as a signaling device to suggest that the social planner claims to have obtained insights about society from studying an abstract mathematical model. Such signaling is not helpful to consideration of the merits of policy proposals.

Maximizing social welfare can encompass policies that would enlarge the economic pie (national product) so that there is potential for everyone to be given a larger slice. In that case, it might be reasonable to argue that the policy would receive widespread support among citizens. A good society - one that is good for the people who live in it – could be expected to adopt such policies. However, claims about pursuing social welfare objectives make such policies no more attractive than if they are advocated to simply expand opportunities for individual flourishing.

Maximizing social welfare can also encompass policies to redistribute the economic pie in a manner that advocates believe will somehow enhance the collective well-being of citizens.  When maximizing social welfare is said to require redistribution of the cake, some citizens will be advantaged at the expense of others. It is possible for some policies of this nature to receive widespread support (e.g. provision of a basic social safety net) but that is less likely when extensive redistribution is proposed to equalize the utility that different individuals obtain at the margin from additional income.

Whose welfare function should we maximize?

The idea of social welfare maximization implies the existence of a social welfare function reflecting insights about determinants of collective well-being and expressing the “general will’ of the people. It was over 50 years ago that I began to realize that this idea is highly problematic. My libertarian friends might find this hard to believe, but it happened while I was studying welfare economics.

An article by Francis M Bator influenced me greatly, although perhaps not in the way the author intended. As I was reading Bator’s article - ‘The Simple Analytics of Welfare Maximization, The American Economic Review, 17(1) March 1957 - I remember feeling that this was an object of great beauty. I suppose the article seemed beautiful for the same reasons that abstract art can seem beautiful. Bator provides a geometric presentation of the derivation of a production possibilities curve, then proceeds to derivation of the utility possibility frontier, which he then crowns with a social welfare function, as shown in the diagram above.

Bator’s description of that diagram left a lasting impression on me. He tells us that BB represents the grand utility possibilities frontier, showing at each point the maximum utility for person X given any feasible level of utility for person Y, and vice versa. He then proceeds to explain the “bliss point”, Ω, in the following words:

“To designate a single best configuration we must be given a Bergson-Samuelson social welfare function that denotes the ethic that is to “count” or whose implications we wish to study. Such a function – it could be yours, or mine, or Mossadegh’s, though his is likely to be non-transitive – is intrinsically ascientific.”

What Bator meant by ascientific is that the function involves ethical valuations. However, the point that has stuck in my mind is that despite the heroic assumptions Bator was making in constructing his beautiful geometric edifice, he did not try to pretend that it could be crowned with a social welfare function aggregating the preferences of all citizens. The function depicted “could be yours, or mine, of Mossadegh’s”. (Mohammad Mosaddegh was an Iranian prime minister who held office from 1951 until 1953, when his government was overthrown - apparently in a coup orchestrated by M16 and the CIA.)

Is it possible to make sense of the diagram? 

As I look at the diagram now, the idea of choosing between the utility levels of different people seems problematic. It would also be problematic to some modern utilitarians whose social welfare function is defined simply in terms of maximizing average life satisfaction (making the implicit ethical judgement that everyone deserves to have the same life satisfaction). In that case, if the axes measure the life satisfaction of X and Y, the bliss point would be defined by the intersection of the possibility frontier and a 45⁰ line drawn from the origin. The 45⁰ line would represent all points where X and Y have equal life satisfaction – X and Y would each have maximum life satisfaction at the bliss point.

However, I reject that modern utilitarian view. It seems to me to reflect an inadequate understanding of the determinants of individual flourishing. As argued in Freedom, Progress, and Human Flourishing, even though average life satisfaction may be a reasonable indicator of the average psychological well-being of large groups of people, psychological well-being is just one of the basic goods of a flourishing human. In my experience, when people are encouraged to offer more than perfunctory responses to questions about how they are faring, they tend to talk about a combination of different things such as their aspirations and the choices they have made, their health, and their personal relationships. Satisfaction is relevant, but does not encompass all relevant aspects of human flourishing.

To make sense of the choices represented in the social welfare function depicted, I would need to replace “utility” with “opportunity to flourish”. Even then, I would need good reasons to make an ethical judgement about whether X and Y deserve to have their opportunity to flourish enhanced or restricted.

What are the implications for social choice?

While Bator’s description of the social welfare function let the cat out of the bag for me, I remember reading about Kenneth Arrow’s impossibility theorem at about the same time. I think the main lesson I took away was that the processes of government must inevitably be somewhat dictatorial. That makes it important to have constitutions that protect liberty and electoral processes that are capable of kicking tyrants out of office.

While studying welfare economics, I also took a course in public choice in which I had my first exposure to The Calculus of Consent, by James M Buchanan and Gordon Tulloch. That book and other writings by Buchanan have had a profound impact on my views about the good society and the role of economists.

Buchanan and Tulloch noted that when individuals are considering constitutional rules that they expect to be in place for a long time, they are uncertain as to what their own interests will be in any of the whole chain of later collective choices made according to those rules. Such uncertainty may enable people to set aside their current economic interests in making constitutional choices. One implication is that individuals will tend to choose somewhat more restrictive rules for social choice-making for areas of potential political activity that could involve violation of liberty.

Buchanan and Tulloch link liberty directly to the concept of a good society:

“The acceptance of the right of the individual to do as he desires so long as his action does not infringe on the freedom of other individuals to do likewise must be a characteristic trait in any “good” society. The precept “Love thy neighbor, but also let him alone when he desires to be let alone” may, in one sense, be said to be the overriding ethical principle for Western liberal society.” (p 217).

 Buchanan later warned that the norms that underlie democratic institutions are under threat when politics is allowed to become little more than a ‘commons’ through which competing coalitions seek mutual exploitation.  (For further discussion of this please see Chapter 6 of Freedom, Progress, and HumanFlourishing).

What should economists do?

Economists who advise on public policy often view themselves as social planners who are advising benevolent despots. They are frequently disappointed to find that those whom they advise give higher priority to political and personal goals than to publicly stated economic objectives, or lack the political power to implement recommendations.  

James Buchanan suggested that economists should adopt a contractarian approach, with a focus on the consequences of rules and, in particular, on the question of what rules of the game individuals might accept voluntarily as participants in an authentic constitutional convention. In providing an example of this approach, Buchanan suggested that such a convention would be unlikely to endorse rules of the game which allow majorities in a single generation to impose public debt burdens on subsequent generations of taxpayers. (Nobel prize lecture).

My career

The focus of my career in public policy advice was partly contractarian. For most of my public service career I had the good fortune to work in agencies of the Australian government (predecessors of the Productivity Commission) which undertook research and published reports on the economic implications of changing the rules of the game for economic development. The focus of much of this work was assessing effects of barriers to international trade and other forms of industry assistance.

I note that my career was only partially contractarian because the agencies were required to make recommendations to the government according to specific terms of reference for individual inquiries and more general guidelines. The specific terms of reference were sometimes designed to ensure that governments received politically palatable recommendations, but the research and policy analysis published in inquiry reports, and in annual reports, informed policy-making processes in ways that led eventually to adoption of rules of the game more favourable to free trade.

The advisory agencies were given general guidelines including having “to have regard to the desire of the Australian Government …  to improve and promote the well-being of the people” and to “improve the efficiency with which the community’s productive resources are used”. I do not believe that the collectivism reflected in the reference to people and privately owned capital as “the community’s productive resources” had one iota of influence on the research and policy analyses conducted by the agencies.

I have endeavored to maintain a focus on the implications of different “rules of the game” in the public policy aspects of my subsequent consulting career and my writing on freedom and flourishing on this blog and in my books. There have been some lapses, but I hereby forgive myself 😄 . It has not always been easy to avoid falling into the trap of viewing oneself as a social planner advising a benevolent despot. 

How should voluntary organisations make decisions?

 

Should voluntary organisations make decisions by seeking consensus or by majority rule?

In my experience, the usual practice is for members to seek consensus before putting issues to a vote. Formal meeting procedure is normally used to record decisions and meet legal requirements. The steam roller of majority decision-making is available where consensus is not possible but that way of making decisions is usually considered to be a last resort option.

Why is that so? Formal meeting procedure is part of our democratic heritage, so why should we reluctant to use it to make decisions in voluntary organisations? I see two reasons - potential divisiveness, and the potential for members to vote with their feet.

The potential for divisiveness is obvious to anyone who observes parliamentary debates. When people take sides on issues, they seek to demolish the views of their opponents and, in the process, implicitly (or explicitly) cast aspersions on their reasoning abilities. That is not conducive to the mutual respect among members that is required for voluntary organisations to function effectively.

The potential for members to vote with their feet is inherent in the nature of voluntary organisations. If individual members feel that insufficient efforts are being made to accommodate differing interests and views of members, they are free to leave.

These issues must be arising in many voluntary organisations all over the world as groups decide whether to conduct meetings online or in-person during the Omicron wave of COVID19.

Some groups are asking members whether they support a specific proposition (to meet online on in-person). That is equivalent to putting the matter to the vote straight away. People who have no strong preference will agree to either proposition, depending on which one is put to them. Those who have a strong preference opposed to the proposition that is put to members have a good reason to feel that they are being steamrolled.

The alternative I favour is to survey members to ask whether they are willing to attend meetings that are held online, in-person, or both forms of meeting. If a substantial proportion of the membership is willing to attend both forms of meeting, that opens up the possibility for compromises to be made to reach consensus. For example, one way forward would be for some meetings to be held online and some meetings to be held in-person.

What have you been thinking about this year?

 

I expect that many readers of this blog will have spent some time this year thinking about the response of governments to the COVID-19 pandemic. That is a topic I have been thinking about, but I have not previously blogged about it this year. I wrote about it on this blog in March and October 2020. With the benefit of hindsight, I think that what I wrote then is defensible, although not particularly illuminating.I set out to write something about the costs and benefits of lockdowns a few weeks ago, but got sidetracked into considering the WELLBY approach to assessing the value of a human life. I thought I might write on that topic in this article but after some additional reading I have decided to adopt less ambitious objectives. My objectives are to consider:

• why there is disagreement on such basic issues as whether lockdowns work;
• whether it would be desirable to have a uniform regulatory approach in all jurisdictions; 
• what we should learn from policies adopted in East Asia; and
• how we should be thinking about government intervention.

Do lockdowns work?

I don’t think disagreement about the effectiveness of lockdowns can be attributed solely to the ideologically blinkers of the participants in policy debates. Some people who are not ideologically opposed to much other government regulation – including Paul Frijters, Gigi Foster, and Michael Baker (authors of The Great Covid Panic) claim that lockdowns do not prevent deaths. On the other side of the debate, some classical liberals who are opposed to much government regulation, nevertheless saw merit in lockdowns - at least during the early stage of the pandemic - to buy time to enable hospitals to prepare for an influx of patients requiring treatment.

The reasoning behind lockdowns is that if you can get people to stay far enough apart from each other, they cannot infect each other. The most obvious problem in getting people to stay at home that is that they need to go to shops to buy food and, in some instances, to deliver health and other “essential” services.  

Lockdowns seemed to suppress virus transmission in Australia in the first half of 2020. In October 2020 I suggested that the combination of self-isolation, shutdowns and lockdowns had worked well in April and May of that year. I have become more pessimistic about the efficacy of lockdowns in Australia this year.  Lockdowns seem to have become less effective in Australia in presence of more infectious strains of the virus, and a decline in public support for lockdowns which was particularly evident in Melbourne - the world’s most locked down city.

Some evidence from other parts of the world suggests that lockdowns have never been effective in reducing death rates. For example, despite its relatively elderly population, Florida did not experience higher death rates than other regions of the United States after abandoning lockdown policies.

The chart shown above (based on a survey conducted by YouGov, an international research data and analytics group) suggests an important reason why the effectiveness of lockdowns is likely to depend on context. Willingness to comply with such regulations is much higher in some countries than others. I think the relatively high compliance level in Australia reflects strong public support for the regulations rather than the substantial penalties that applied if non-compliance was detected. The regulations were difficult to police even in the presence of strong public support, and would have been impossible to police if blatant non-compliance had become widespread.

Would it be desirable to adopt a uniform approach in all jurisdictions?

Differences in support for regulation and associated differences in willingness to comply, are good reasons for different approaches to be adopted in different jurisdictions.

Frijters et al suggest a more fundamental reason why a diversity of approaches is desirable. After noting the value of state-level experiments in the United States, including the minimalist policies adopted in South Dakota, the authors suggest:

“The provocative takeaway is that the intelligence of a whole country is enhanced when it contains communities adhering to truths completely opposed to those of the intellectual elites. That takeaway is, moreover, a deep lesson from history that Western countries have embedded into their institutions over centuries. It has been remarked upon before by historians that competition between radically different systems leads Western countries to learn faster than more centralised places like China.”   

What should we learn from the policies adopted in East Asia?

The experience of East Asian countries in preventing deaths from COVD-19 has been held up as example for others to follow. For example, an article by Mingming Ma,  Shun Wang, and Fengyu Wu, published as Chapter 3 of World Happiness Report 2021, concludes as follows:

“In general, we find that the relatively successful story of the five East Asian regions, compared with the six western societies, can be attributed to the stronger and more prompt government responses and better civic cooperation. Except for Japan, all of the East Asian governments implemented more stringent mobility control and physical distancing policies, as well as more comprehensive testing and contact tracing, especially at the early stages of the outbreak. A summary of the government interventions and anti-COVID measures in the East Asian regions indicates that a combination of strong government response systems, early and rigorous mobility control, extensive screening, testing, contact tracing and isolation, coordinated resource allocation, clear communication, enforced self-protection practice, and supportive economic measures are important in fighting COVID-19 outbreaks and resurgence.”

The five East Asian jurisdictions referred to are China, Hong Kong, Taiwan, Japan, and South Korea. The six Western countries included in the study for comparative purposes were France, Germany, Italy, Spain, the United Kingdom, and the United States.

 The authors seem to be suggesting that all the East Asian jurisdictions adopted stringent policy responses.

Frijters et al reach a different conclusion using the same data on policy stringency in a study published in Chapter 3 of The Great Covid Panic. The authors group countries and regions into three categories, Minimalists, Pragmatists, and Covid Cults, on the basis of the stringency of the average stringency of their policies during 2020. They found that the minimalists had far fewer claimed Covid deaths than either the pragmatists or the cults and that the pragmatists accumulated only a little over half the death rate of the cults.

It is interesting that most of the East Asian jurisdictions referred to in the first study were classified as either minimalists or pragmatists. Taiwan and Japan were classified as minimalist, and South Korea was classified as pragmatist. The United States and most European countries were classified as cults, along with China, Australia and New Zealand.

It is certainly difficult to maintain that stringency has been a major factor explaining the relative success of policy responses in the East Asia region.  I am not sure what other conclusions can be drawn, except that further study will be required if we are to learn from the experiences of countries adopting different policy responses.

How should we be thinking about government intervention?

Many politicians and other commentators seem to imply that apart from lives lost (or saved) the only other factor that needs to be considered in evaluating policy responses to COVID-19 is their impact on GDP. Far too little account is taken of the future consequences of increases in public debt that have been incurred to support people during lockdowns and the psychological impacts of restricting social interactions for long periods.

When freedom is mentioned by advocates of stringent regulation, it is often viewed as something frivolous that must be sacrificed to prevent deaths from Covid. That is the way a bureaucrat might view the options if given prevention of deaths from Covid as a key performance indicator (KPI). Within that mindset, freedom must be sacrificed to a sufficient extent to ensure that lockdowns work. Prevention of deaths from Covid is seen as being of utmost importance. Just as some soldiers have claimed that they had to destroy villages in order to save them, the single-minded advocates of lockdowns seem to be willing to destroy people’s lives in order to save them.

I am not implying that freedom is more important than health, or that liberty is more important than human flourishing. I am just suggesting that it is unhelpful to view the issues in that way.  

About 15 years ago, after reading some of the writings of Douglas Rasmussen and Douglas Den Uyl, I realized that it makes no sense to think in terms of a need to choose whether priority should be given to liberty or to human flourishing. Human flourishing does not exist apart from the flourishing of individuals, and the flourishing of individuals is not possible without opportunities for self-direction. Once we recognize the importance of self-direction to individual flourishing, that poses the question of what rules of the game – or political / legal order - would allow greatest opportunities for individual self-direction. Liberty is the answer! The protection of individual liberty – or the natural rights of individuals – provides the context in which individuals can flourish in different ways, provided they do not interfere with the rights of others. (You can find further explanation and links to the works of Rasmussen and Den Uyl in my book, Freedom, Progress, and Human Flourishing.)

Recognition of the foundational role of liberty doesn’t tell us what rules of the game should apply in a pandemic. However, it does tell us that we should be looking for rules of just conduct that would provide an appropriate balance between the different interests of individuals in getting on with their lives and avoiding exposure to infection.

The discussion earlier in this article suggests:

• The most appropriate rules in any society must depend, to a large extent, on the degree of support for them.
• A diversity of approaches in different jurisdictions is highly desirable to provide greater opportunities to learn from the experience of others.
• Different interpretations of the East Asian experience suggests that some caution is required to ensure that we learn the right lessons from the experience of others.

How are life satisfaction ratings related to living standards evaluations?

 

It is well known that in wealthy countries, further improvement of average incomes has only a small impact on average life satisfaction. Diametrically opposed explanations have been offered.

On the one hand, there are those who say that if rising incomes have little effect on average life satisfaction, that must mean that their apparent impact on living standards is a mirage – rising incomes do not count as progress.

On the other hand, there are those who say that average life satisfaction numbers are garbage – you can’t expect to get useful information by asking people to rate their satisfaction with life on a 10-point scale. They say that rising average incomes provide an accurate picture of progress.

In my view, those opposing explanations are both unhelpful to an understanding of the relationship between living standards and life satisfaction. In my book, Freedom, Progress, and Human Flourishing, I explain that rising incomes result in actual improvements in living standards, and count as progress because that is what people aspire to have. Since self-direction is integral to human flourishing, it is obvious that progress is inextricably linked to conditions that enable individuals to meet their aspirations more fully. In the book, I also explain why I think average life satisfaction is an appropriate measure of psychological well-being at a national level. I suggest that psychological well-being, along with wise and well-informed self-direction, is one of several basic goods that a flourishing human could be expected to have.

The moving benchmark problem.

The failure of life satisfaction to reflect improved living standards is explained as follows in my book:

“The happiness surveys behind this puzzle, often referred to as the Easterlin puzzle, ask respondents to rate their lives relative to benchmarks such as the best possible life. Let us assume that when a person in a high-income country, call him Bill, answered that question in 1990, and he gave a rating of 8/10 for his life. Since then, Bill’s income has increased at about the same rate as the average for the country in which he lives, and there have been no abnormal changes in the circumstances of his life. In January 2020 … he again rated his life as 8/10. …

Bill’s income has risen, but his rating of his life has not risen.

The problem is that the survey prompted Bill to rate his life against a moving benchmark. Bill’s view of what constitutes the best possible life is likely to have risen over time. The people he sees living such a life have obtained access to better communication technology, and other things that have potential to enhance the quality of life. If you ask people to rate their current lives relative to a benchmark that is moving upwards over time, measurement error is inevitable.”

Ways to avoid the moving benchmark problem include the ACSA approach previously discussed on this blog (here and here). For reasons best known to themselves, happiness researchers have not shown much interest in using that approach to test the extent to which life satisfaction measures are distorted by moving benchmarks.

Living standards comparisons

The moving benchmark problem does not arise when people are asked how their standard of living compares with that of their parents when they were about the same age. Surveys of that kind have tended to provide information consistent with perceptions of ongoing progress with rising incomes in wealthy countries.

There is no plausible reason why such inter-generational comparisons should be viewed as less credible than life satisfaction ratings, or vice versa. As I see it, they are cognitive evaluations of different things. The intergenerational comparisons are measuring perceptions of progress, and the life satisfaction ratings are measuring current psychological well-being.

Merging life satisfaction and living standards evaluations

In order to obtain a better understanding of the linkage between perceptions of progress and current life evaluations, it is necessary to bring those different cognitive evaluations together in some way. That has been made possible by inclusion in the latest round of the World Values Survey of a question asking respondents whether their living standards are higher, lower, or about the same as those of their parents when they were about the same age. The graphs shown above were prepared using the excellent Online Data Analysis facility of the World Values Survey. Information is shown for the United States and Australia, but similar pictures emerge for other high-income countries.

The most obvious point illustrated by the graphs is that people tend to be much less satisfied with their lives if they perceive that their living standards are lower than those of their parents at a comparable age. Their perceptions that their living standards have fallen tends to make them feel grumpy about life.

The second point to emerge is that the life satisfaction ratings of those who perceive that their living standards are better than those of their parents are not much higher than for those who perceive that their living standards are about the same as those of their parents. Their perceptions of progress are not reflected to any great extent in their satisfaction current lives. That result is consistent with my view that life satisfaction is a poor indicator of the extent to which people meet their aspirations for higher living standards.

Implications

Perceptions of change in living standards that emerge from intergenerational comparisons are related to the recent history of economic growth in different countries. The greatest percentage perceive that their living standards are higher than their parents in countries that have sustained high rates of growth in per capita GDP over several decades. Of the 54 countries for which data are available, Vietnam has the greatest percentage in that category (90%) and Iraq has the lowest (21%). The corresponding percentages for Australia and the U.S. are 56% and 48% respectively.

Percentages who perceive that their living standards are lower than their parents follow a broadly similar pattern, but in most countries are within the range of 10% to 25%. Of the 54 countries, Zimbabwe is the only one where more than half of respondents perceived that their standard of living was lower than that of their parents. The corresponding percentages for Australia and the U.S. are 15% and 19% respectively.

The age structure of people who perceive themselves to be worse off than their parents suggests that this source of grumpiness is likely to pose a greater problem in Australia and the U.S. in the years ahead. The incidence is lowest among the 65+ age group (7.6% for Australia and 8.4% for the U.S.). The highest incidence in Australia is in the 25-34 age group (20.1%) and in the U.S. in the 35-44 age group (26.4%).

Conclusions

Average life satisfaction provides useful information on psychological well-being at a national level, but is a poor measure of the extent to which people are meeting their aspirations for higher living standards. As expected, people who perceive their standard of living to be higher than that of their parents, do not rate their life satisfaction much higher than those who perceive their standard of living to be about the same as that of their parents. However, people who perceive their standard of living to be lower than that of their parents have markedly lower life satisfaction than the other groups. The percentage of grumpy people in countries such as Australia and the U.S. seems set to rise in the years ahead unless opportunities improve for young people to meet their aspirations for higher living standards.