Life‐Style and the Standard of Living
Life‐Style and the Standard of Living
Abstract and Keywords
In a famous, although dubious, definition of economic welfare, Pigou (1952: 11) defines it as ‘that part of social welfare that can be brought directly or indirectly into relation with the measuring rod of money’. The aims of this definition, and the almost insuperable difficulties of implementing it, are both echoed when we try to establish a clear and sensible meaning for the term ...
1.1. The Meaning of the Standard of Living
In a famous, although dubious, definition of economic welfare, Pigou (1952: 11) defines it as ‘that part of social welfare that can be brought directly or indirectly into relation with the measuring rod of money’. The aims of this definition, and the almost insuperable difficulties of implementing it, are both echoed when we try to establish a clear and sensible meaning for the term standard of living.
This issue is in part semantic. Provided that we make our meaning clear, we can call what we please by the name standard of living. Great departures from standard usage may cause confusion, however. It is perhaps unfortunate that economists, and popular usage, have suborned the term standard of living to a narrowly economic meaning. Because they have done so, we should probably invent a new term if a different meaning is intended. Quality of life already serves that role.
Unfortunately, not all the issues are simply semantic. The question of whether it makes sense to treat the economic quality of a life apart from more general considerations is a real issue. Pigou favoured that separation because he judged that the correlation between what his measuring rod would show, and other, wider, aspects would not be negative. His ‘measuring rod of money’ turns out in fact to be an estimate of real purchasing power.
This is much narrower than Pigou's definition seems to allow, and gives rise to the question of whether there is anything that might not in principle be brought into relation with the measuring rod of money. If a healthy man would consent to have a leg amputated for $1 million, or change his religion for $200,000, does this provide us with monetary measures of the value that he places on his leg, or on freedom of confession?
A purely practical reply would say that it is a matter of measurement. Bread is traded a great deal and its cost of production may be gauged quite accurately, so the consumption of bread comes within Pigou's definition. The same cannot be said of selling legs or religious allegiance. The common‐sense determination of the scope of a standard of living index leaves open the question of whether we would want the standard of living to include nearly everything if we could measure whatever we pleased.
To illustrate some of the issues, consider a rich individual who suffers from an untreatable disease which seriously interferes with the enjoyment of life.1 (p.418) In this paper I propose not to describe the said individual as having a low standard of living. I shall instead say that he has a high standard of living but he is in poor health.
I recognize and respect a different view argued vigorously by Amartya Sen (see Sen et al., 1987) according to which, to put it simply, the standard of living should embrace all aspects of the quality of life. Not only has this view great appeal but it also enjoys the advantage that it evades awkward delineation problems.
In a developing country, much of the population is in poor health and the endemic bad health is an aspect of the country's low standard of living. In one case, that of the rich man, the ill health is accidental. In the other case, that of the poor country, it is a consequence of material deprivation. In the first case it is excluded from the standard of living, in the second case it is included.
1.2 Measurement and Objectivity
As part of the justification for a narrow focus definition of the standard of living I have mentioned measurement. This leaves open the question of what counts as measurement and of who is to do the measuring. The principle of positivism, applied to economics, holds that quantities must be defined in terms of observables.2 This view is unfashionable at the present time. However, even if it were agreed that quantities should be defined in terms of experiments, some experiments are impossible to implement, or difficult to interpret even if carried out. So we must consider magnitudes that cannot be directly measured.
The positivist approach says that measures should be defined in terms of observables. This is independent of another idea, often encountered in economics, according to which the measure of welfare, and presumably of the standard of living, should be the preferences of the actors themselves. My argument will take that idea seriously but will show how it runs into difficulties when preferences are required to bridge large culture or, as I call it, ‘life‐style’ gaps.
The measurement of the standard of living by preferences seems to lead to an objectionable subjectivism, in which what should be a scientific measurement is reduced to whatever people think it to be. The opposite view would hold that there is an expertise concerning standard of living measurement, so that there might be an objective sense in which the participants' preferences would be fallible, and the economist would know best.
I believe that expertise in standard of living measurement should be admitted, but only in two instances. First, individual agents can only evaluate small changes in living standards, so the economist's advantage lies in his ability to integrate to obtain a global measure. Second, the individual may not be fully (p.419) informed of his own, or his group's, state. The inhabitants of a poor country, for example, may not realize how unhealthy they are, and the consequences of that ill health, whereas an expert will know. Both these cases are really the same: they involve imperfect vision on the part of the individual. They are important, but they should not persuade us to let go of the fundamental point that man, if not the measure of all things, is at least the measure of the standard of living.
1.3. Preferences and the Standard of Living
Above we encountered the idea that the standard of living should be measured by the preferences of the actors themselves. Ignoring the difficulties of finding out those preferences, there remains the important issue of how preference, which is at bottom an ordinal notion, is to quantify the concept of the standard of living, which is surely, to an extent at least, a cardinal concept.
One route out of the difficulty just explained takes the direction of arguing that preference is not simply an ordinal notion. Rather, according to this view, strength of preference is a natural and perfectly legitimate notion. We then have what I shall call measured preferences—a preference ordering enforced by values for the power of the preference.
There has been much discussion in the social choice literature of strength of preference, often related to interpersonal comparisons of utility.3 Strength of preference is different from and weaker than strong cardinality. Thus, if utility may be defined subject to linear transformations, differences are ordered and strength of preference is measurable, say between two changes affecting an individual. Full cardinality, on the other hand, involves cross‐calibration of settings across individuals.
In my view, strength of preference has meaning only in terms of a fixed frame. Hence it makes no sense within a system subject to Arrow's unrestricted domain, for then nothing is fixed and there is no bench‐mark for strength of preference. Strength of preference can only be measured in terms of what one is, or is not, willing to surrender to obtain something else.
This is where the argument leads on naturally to choice under uncertainty because that is the field in which von Neumann and Morgenstern (1944), established their well‐known utility index. That index is invariant once the utility level 0 has been attached to the worst possible outcome and the utility level 1 to the best possible outcome.4
Strangely, the use of the von Neumann—Morgenstern utility index as a measured preference indicator has not proved popular. Thus Sen (1970: 94–9) notes a number of objections to this procedure (see also Harsanyi, 1955). Sen's leading (p.420) objection claims that gambles may not provide uncontaminated information on preferences. The case of the agent who will not reveal his preferences in a simple lottery experiment, because he regards gambling as sinful, is cited. Much more importantly, how do we know that the curvature of the utility function—which is the specific information which cardinalization provides—does not reflect risk aversion as opposed to strength of preference?
In the modern theory of choice under uncertainty, due to Ramsey, Savage, and von Neumann, there is no room, it would be argued, within a set of consistent preferences over certain and uncertain outcomes, for separate and non‐identical quantifications of measured preference5 and of risk aversion. The two must correspond. Unfortunately, consistent preferences over certain and uncertain outcomes may very well be precisely what actual agents do not have.
2 Fairness, Contract, and the Standard of Living
2.1. The Standard of Living and Fairness
Fairness may be defined in more than one way, for example, as the absence of envy, or in terms of the existence of a consumption bundle such that each individual is indifferent between his own consumption and the said bundle.6 For our purposes the absence‐of‐envy definition will serve. Absence of envy directly establishes only a partial ranking, that is, sets of positions evaluated as representing an equal standard of living. However, that partial ranking may be extended to cases of positive envy, and a complete ordering of positions established.
When agents have different tastes, the envy criterion exhibits paradoxical properties. For example A may envy the bundle enjoyed by B, while B envies the bundle enjoyed by A. This can happen when A and B have different tastes. Also, when different tastes are involved, absence of envy does not equate to justice. In a classic example, A divides a cake between himself and B; B greatly prefers nuts to raisins, while A is indifferent between the two. The cake has most of the nuts on one side. A can then cut himself a hugely greater slice than he offers to B, even if he is constrained not to take a slice that B envies. However, B does envy A's positional advantage in being the one who cuts the cake. This example reminds us of the general point that it may be the opportunities that the rich enjoy that are enviable, not necessarily the consumption that they choose.
The concept of fairness was developed within a framework in which tastes, although differing between individuals, are constant for each individual. With standard of living comparison we sometimes need to allow for tastes which are endogenous to the allocation of individuals to positions and consumptions. An illustration would be the migrant. He considers the life of the city factory worker, (p.421) and as a sophisticated observer he might not envy him unambiguously. He is disgusted by certain aspects of city life, yet he supposes, rightly perhaps, that he would get used to them, that migration would change him.
2.2. Contract Models, Lotteries, and Inequality
There is a connection between fairness and Rawlsian distributive justice7 but the two are not the same. Consider the ‘original position’. in which Rawls's social architects agree on the design of the social system. For simplicity, assume that only the economic aspects of the society are at issue. The Rawlsian social architects, yet to be clothed in their personal features and positions, so arrange things that they maximize the lowest standard of living.
It does not follow that the Rawlsian outcome would be free of ex post envy. If the president of a major company has to be paid $2 million per annum to give her the incentive to sweat the last drop of efficiency (and with it enormous social benefit) out of her company, then the payment of that salary will command universal support in the original position. After actual positions have been determined, however, the person who has been made company president will be the object of everyone else's envy. Yet ex ante there is no envy, for all are in the same position. The question of how the notion of fairness should be applied to processes involving uncertainty is discussed by Diamond (1967).
Rawls's social architects will not surrender the slightest decrease in the welfare of the worst‐off one for a large general increase in all others' welfare levels. It seems that this extreme aversion to risk is seen as a typical feature of being in the original position. Imagine that in the original position all have the same tastes and know that they will retain these tastes after the veil of ignorance has been removed. Meanwhile all have an equal probability of occupying any position. Why would they not all vote for arrangements that would maximize the unweighted sum of utilities, since this would measure each one's expected utility prior to knowledge of position?8
This solution allows inequalities which are not justified by the interest of the later‐to‐be‐identified worst off. I have no space here to make a case for such a radical rewriting of Rawls's method. Notice, however, that the alternative approach leads naturally to a helpful way of looking at inequality within units whose standards of living are to be compared.
Does an inhabitant of Jamaica envy an inhabitant of the USA? It obviously depends upon which inhabitants, for there are certainly many North Americans who are worse off than the Jamaican middle class. Yet it is absurd to suppose that this fact prohibits us from claiming that the USA has a higher standard of living than Jamaica. One solution to the difficulty is to use a notional lottery to compare positions of inequality. Country A has the same standard of living (p.422) as country B if an agent with an equal chance of occupying any position in one country or the other is indifferent as to which country it should be.
One appealing implication of this idea is that national inequality automatically reduces the national standard of living. If country A has the same population and the same total national income as country B, but in country B that income is more unequally distributed, then an agent about to play the lottery that will allocate him a position in country B will envy the agent about to play a lottery that will allocate him a position in country A. For the purpose of this exercise inequality should be measured by the Atkinson index (see Atkinson, 1970).
3 The Classical Theory and Extensions
3.1. The Classical Theory
The classical theory uses the model of consumer choice with a budget constraint to compare different situations. The subject has been treated extensively elsewhere.9 Here I present only the results required to bring out particular points. A brief but more formal treatment of certain topics will be found in the appendix.
At the heart of the classical model is the theory of the cost of living index number. This measures how much money income in one position would make the subject as well off as he (or she) is in another position. Our starting point is similar. We take two agents in different positions. But then we do not ask, what change would make them equal? Rather we ask, how far apart are they in standard of living space? This small extension leads to large differences.
An agent is defined by a list of features: his income, the prices that he faces, his tastes, etc. We shall refer to a complete list of the agent's features as his station. The image of an observer looking out at other stations and forming judgements concerning them, and of each station being occupied by a similar observer, captures well the idea intended. The classical theory assumes uniform tastes in all stations, so a station is fully characterized by prices and a level of income. As tastes are uniform, the stations can be unambiguously ranked in terms of their desirability.
We start with the case of two stations of equal standard of living. As all individuals have the same tastes, the level of utility enjoyed at the two stations must be equal. If utility is not directly observable, we could ask an agent to choose between the stations and count indifference as standard of living equality. The classical index number is an approximation that bypasses the need for that experiment.
Standard of living equality tests side‐step an important issue. Is standard of living to be measured in terms of real purchasing power? If it is, then it must be silent on the issue of whether high standards of living in terms of a high real income are proportionately effective in buying economic quality of life. (p.423) For example, if we define an index number measure of the difference between the standards of living of two stations as the difference between two price indices, then the difference is expressed in terms of real purchasing power.
Does a real‐purchasing‐power‐index measure capture what we want the standard of living to mean? Suppose, for example, that prices at two stations are the same but that income at the first station is twice as high. Then an agent at the first station may, if he wishes, consume twice the quantity of goods consumed by the agent at the second. When we measure a standard of living difference in terms of quantities of goods we make it a matter of definition that the rich enjoy a hugely higher standard of living than the poor. Perhaps the standard of living should be measured in terms of the utility that it makes possible, rather than the quantities of goods the purchase of which it permits.
A problem with defining the standard of living in terms of utility is that utility is not directly measurable. In order to use it we have, therefore, to establish a procedure for calibrating it. To make precise the common notion of a standard of living, we need to be able to say something about differences in living standards, and on this question ordinal utility is silent. It may be clear that the standard of living in Zaire is lower than the standard of living in Ghana, which in turn is lower than the standard of living in Switzerland. There must also be some meaning to the idea that the standard of living in Switzerland is much higher than in either Zaire or Ghana.10
3.2. The Lottery Measure
The obvious objection to measuring standard of living by utility level is that utility is not cardinally measurable. We noted above (Section 2.2), however, that a lottery of stations provides a measure of utility with a genuine cardinal property. Why should it not provide a cardinal measure of the standard of living?
A sensible objection to a lottery index of standard of living would argue that it makes little sense to ask most agents to evaluate a lottery involving a gamble between the highest and the lowest standard of living in the world, as the original definition would entail. Most people are far from either, and to pin everything on such a decision is seemingly to invite a confused and meaningless result. Fortunately, if agents can accurately evaluate lotteries involving local stations, then a chain method can link together their choices to provide a globally accurate measure. This is shown in the appendix.
A standard of living measure on this interpretation is like a map from which the distances of long journeys can be read off. However, no map‐maker will have directly measured the distance concerned. Rather, the map will have been constructed from numerous local, line‐of‐sight measurements.11 The chain method (p.424) works in the classical model because there is a global map to be reconstructed. To run ahead of our argument, note that when the implications of life‐style are taken into account, discontinuities emerge which make it impossible to make local measurements in all directions. This has radical implications for standard of living measurement.
3.3. The General Equilibrium Model
The specification of the consumer in standard expositions of the classical theory is narrower than the treatment of the consumer in modern general equilibrium theory, the locus classicus of which is Debreu, 1959. In Debreu's general equilibrium model the consumer is confined to a consumption set. This device is important. Indeed, unless it is included the theory of the consumer is just nonsense. The provision of a labour service is, for example, treated as negative consumption in the general equilibrium model. Then only the constraints of my consumption set, and not my preferences, can explain why I am writing this paper, rather than conducting a symphony orchestra, which my preferences certainly rate higher.
A consumption set is an important component of a life‐style because the consumption set varies with life‐style. In the standard general equilibrium model the consumption set is a fixed characteristic of the agent. As long as it is fixed, including a consumption set makes little difference, and the version of the classical theory developed in the appendix needs only small amendments to cover the general case.
Where comparisons, and especially international comparisons, of standards of living are concerned, we do need to allow for different consumption sets. The consumption set of a typical American and the consumption set of a typical Indian villager may both be fixed, but they will surely be different. Notice that care must be taken to distinguish between the consumption set and prices. Much more is obtainable in the USA, even in a village, than is obtainable in an Indian village. We might regard the unobtainable as having an infinite price, but this makes comparison almost impossible.
In any case, obtainability is not the only issue. The populations of rich countries are better educated and trained than those of poor countries. Even when faced with identical prices for goods and labour services, their standard of living would thus be significantly different. The prices used in the usual cross‐national income comparisons do not include wage rates. Income is taken as a proxy for earning power, whether the latter is determined by skill level or wage rates. However, differences between consumption sets can have significant consequences, even when income is given.
Some of these consequences have been explored by Usher (1968) in the context of a comparison of the UK and Thailand. UK consumers, for instance, spend much more heavily on transport services than their Thai counterparts. This cannot be accounted for either by differences in income or by differences in tastes. The difference is in the consumption sets. This is not the way that Usher (p.425) expresses the point but it is implicit in his argument.
An estimate of the standard of living based on a comparison of expenditure outlays by households which fails to note that much expenditure on transport is consumption set‐imposed is seriously biased. The streams of commuters who pour into London every day do not reveal a preference for moving around. Their consumption sets allow them no combination of a good job and affordable housing which is not jointly consumed with a large quantity of transport.
This last case is close to an instance of life‐style demonstrated from within the classical model extended to include consumption sets. The commuters of the last example are not forced to commute. They could perhaps obtain poorly paid jobs in their home towns. If they decided to do so it would profoundly affect how they lived and would influence the tastes which they revealed, and not just because of the income shift. The introduction of life‐style in Section 5 below is therefore a novelty, but one which springs naturally from received theory.
4 The Determination of Tastes
4.1 Endogenous Tastes
The idea that preferences are fixed and given is probably the least appealing idea that economists have ever come up with. Some writers, including Marxists, have made the criticism of this idea a leading point in their assault on economic theory. Generally speaking, however, economists have not been much concerned by this case. Most would take the point that preferences are governed by upbringing and culture, but would still feel that they can be taken as fixed for many purposes of economic analysis.
Many writers have looked at endogenous preferences.12 To a remarkable degree, however, this literature neglects to examine the welfare implications of endogenous preferences. As the whole basis of the welfare evaluation of change is unchanging preferences, this is understandable. However, Fisher and Shell (1972) face the problem when they discuss the introduction of new goods in cost of living comparisons. New goods and changes in tastes raise similar problems.
There are further exceptions to my claim that the endogenous preferences literature ignores welfare evaluation. Especially relevant here is von Weizsäcker, 1971, because this paper looks at an example in which migration is what affects the endogenous preferences of the model.13 In the von Weizsäcker example, a farmer has to consider whether to migrate to the city or not. Relative prices are different in the city but tastes only adapt slowly.
(p.426) How convenient it would be if we could invoke cardinal utility comparisons between positions. Imagine the farmer saying: ‘I moved to the city because I thought that I could earn $3,000, which would have given me 100 miwigs,14 but in fact I could only earn $2,000. However, I discovered that once accustomed to city life I liked it more than I had expected, so that the $2,000 could buy me 120 miwigs.’ Unfortunately, as we know, and as von Weizsäcker agrees, the meaning of an absolute cardinal utility quantity is questionable.
The strictly ordinal approach mostly favoured by economists has been questioned but no clear alternative has been worked out. The problem is a genuine one. Intertemporal comparisons of happiness do not translate readily into orderings or into even hypothetical decisions. What do people think they mean when they assert that one's school days are the happiest days of one's life?
In von Weizsäcker's account the farmer who might migrate applies a method for evaluating the change which is clearly invalid. specifically, he assumes that his preferences will not change when he takes up residence in the city, when in fact they will. This is the result of applying the short‐run indifference curve and it is invalid as a matter of ordinal fact. The farmer's ordinal preferences will change.
However, imagine a farmer who is more clear‐sighted. He understands that he will adapt. He might even have a very good idea of the form that this adaptation will take. What is he to do? To be specific, how is he to form intertemporal preferences with which to evaluate such a history? It is hard enough to know how to weight one's future self against one's present self, even if they are the same character. With a fundamental change, the problem looks impossible. Imagine that you are told that you will shortly undergo a profound religious conversion that will radically alter your personality and your values. Meanwhile you have to decide how much to save. If such exercises have meaning, it is only by virtue of the preferences of the present and actual. Forced to weight one's changed self against one's present self one can only allow the present self to act as a benign dictator.15
The problem discussed by von Weizsäcker is echoed in Hollis's discussion of the Ant and the Grasshopper.16 The Grasshopper consumes during the summer while the Ant invests. When winter comes the Ant asks the Grasshopper, who is hungry and depressed, whether he is not now sorry. Hollis retails the story as follows: ‘ “Are you not sorry that you sang all summer?” “Very sorry”, the (p.427) Grasshopper sighed, “just as I knew that I would be. But now is now—I acted rationally then. It is you who are irrational in resisting your present desire to help me”’ (Hollis, 1987: 95).
The Ant's criticism of the Grasshopper seems to amount to an insistence on the priority of long‐term preferences over short‐term ones. However, there is a crucial difference between this last story and von Weizsäcker's construction. Weizsäcker assumes myopia; indeed he freely employs the term, and myopia is used to explain, for example, the catastrophic path chosen by the lover of variety who switches back and forth, though consuming less and less, until he reaches starvation point. Hollis's Grasshopper, however, is far‐sighted. He expects to feel painful regret.
Regret can mean more than one thing. Genuine regret should amount to a preference, if one could be transported back in time, to be constrained to take a different decision. The Grasshopper feels no such regret. He wishes that he had food in the cold winter, as how could he not, but if transported back in time he would not wish to be constrained to choose differently.
Are preferences concerning situations of sheer impossibility meaningful? Many feel that they have such preferences. ‘If I could be young again I would not marry.’ Such preferences, if they mean anything, reflect current views. Fisher and Shell (1972: 4) argue provocatively that the question ‘Would you like to live your youth over again having the tastes that you do now?’ is more meaningful than the question ‘Were you happier when young than you are now?’, on the ground that the former question elicits a preference concerning an, admittedly fantastic, choice, while the latter does not.
It is on a test of the form: If you could go back having your present knowledge, would you wish to be constrained? that the non‐migrant in von Weizsäcker's example can be convicted of non‐optimality. Present or final tastes are allowed to legislate the optimality of past actions.
Eventual non‐regret is, however, a dubious criterion, as the Grasshopper understands. The Grasshopper is rigidly rational and consistent. How much worse when changes of taste are involved which are so radical that they amount to a change of personality. We do well to turn away he who says to us: ‘Try this, man, it will blow your mind, but I promise that you will never regret having taken it.’ But can we prove that we are right to do so?
5 The Concept of Life‐Style
5.1. The Meaning of Life‐Style
We come now to the heart of our argument, the idea of life‐style. As it stands, the term is not precise, and some of its overtones may be out of place. Bear in mind that formally a life‐style involves both a consumption set and preferences.
The first examples that spring to mind are those that correspond most closely to a life‐style in the popular sense of the term. The Bohemian (or in modern (p.428) terminology the ‘Hippy’) life‐style incorporates a happy‐go‐lucky life of a non‐acquisitive kind, which rates contentment and personal fulfilment above the dictates of the work ethic. We can also consider styles corresponding to ways of organizing social life—tribal or patriarchal—as opposed to the nuclear family and individualistic, for example. Or we might consider life‐styles that are generated by means of organizing production—rural self‐sufficient or nomadic, for example. The Western life‐style might be characterized as acquisitive and consumerist, with an emphasis on work as a valuable and fulfilling activity. It might be regarded as embodying a high level of ‘rationality’—in the sense of Max Weber rather than the economist's sense.17 Ultimately the list of life‐styles is coextensive with the number of separate types of station, the differences between which complicate the comparison of standards of living.
Life‐style, as a joint specification of the consumption set and preferences, is closely related to the productivity of the economy. The strict work discipline of industrial societies may be regarded as a feature of their predominant life‐style. This life‐style in turn allows for the high productivity which results in enlarged consumption choices as well as narrowing choice with regard to how and how much the subject works.18
We can now contrast the formal content of the idea of a life‐style with the assumptions of the classical model. The differences are twofold. First, when different life‐styles are considered, preferences, prices, and consumption sets are not independent. Second, the choice of a station involves the choice of a consumption set, and also of the prices that will apply for the chosen station. Hence the consumption set is to some extent chosen by the agent but chosen jointly with prices.19
In assuming that preferences are determined by life‐style, I am neglecting the purely individual and idiosyncratic component of preferences. Something like this view has wide support. The classical model adheres to it in assuming tastes to be uniform. Those who stress the social and cultural determinants of taste also set aside the merely individual. Stigler and Becker (1977) argue that many transcultural differences in tastes can be accounted for by relative price differences. I agree with them but would go further. Even more differences in tastes can be accounted for by differences in relative prices and consumption sets. The residual element of strictly individual quirk then looks quite unimportant.
5.2. Life‐Style and Community
People are not simply individuals. They live socially and their views, their (p.429) values, and even their beliefs, as well as their abilities, are formed and sustained within social groupings, families, and communities. Living a life‐style and inhabiting a community are not equivalent but they can be closely interrelated. Perhaps the consideration of life‐style offers an alternative to the methodological individualism that has been held to be a weakness of orthodox social science.
For the case of self‐contained, even isolated, collections of similar people, to inhabit a community is to occupy a place together with others and to participate in their social and economic activity. However, even common usage recognizes more complex and ambiguous cases, for example, the Jewish community in France. Often the community does not conduct its economic activity in isolation from the rest of the nation. Neither are the values maintained by that group independent of those that prevail elsewhere.
The attempt to prove that the smaller community is viable within the larger defines the life and the history of certain groupings: the gypsies, the wandering Jews, or the monastic orders. This is partly an issue of economic viability, but value and identity dilution are often more acute problems, as the history of the Jews—both as recorded in the Bible and in later history—well illustrates. These questions are powerfully focused by migration, whether forced or chosen.
5.3. Switches of Life‐Style
If life‐style never changed we should all still be hunter‐gatherers. The environment is never static, however, and life and ideas are constantly changing. In the course of this process life‐styles undergo translation, sometimes gradually, sometimes abruptly. Some switches reflect a conscious decision to switch, some are unintended implications of other decisions. A family decides to give up farming, or to send a son to college, and thus embarks upon a road which will change their horizons and alter their life‐style whether they realize it or not.
Life‐style may be changed by force, when on account of changes in prices and the environment a point is reached when the existing consumption set is no longer attainable. We call the last outcome life‐style breakdown. Even when unforced, the decision is typically made in a cloud of uncertainty, as the forward extrapolation of taste changes is assessed in terms of present tastes.
5.4. Standard of Living Comparisons Across Life‐Styles
Comparisons of the standards of living of different stations are judgements, and judgements concerning which there is often no higher authority than the concerned actors themselves. That sounds simple, but its application must confront some serious problems. Compare comparisons of standard of living, on the one hand within the classical model, and on the other across life‐styles. In the first case we are dealing with judgements, but the content of the judgement is fairly clear. To judge A to be a better standard of living than B is equivalent to deciding that one would rather enjoy the income and prices of A than the income and prices of B.
Of course prices and incomes are not everything, which means that on my (p.430) narrow definition of standard of living a person need not wish to move from a lower to a higher standard of living. If standard of living is a partial measure of the value of a life, then a high standard of living is something that a rational agent will not invariably pursue.
When we turn to standard of living comparison across life‐styles, what an evaluative judgement means becomes a very difficult issue. The monastic life‐style may involve a vow of poverty and a positive commitment not to seek a high standard of living. Even in this extreme instance, however, it is not clear that a judgement cannot be made. One does not have to want to live in a fine house to judge a dwelling to be such. However, an enclosed monk might be incapable of appreciating what a rich man's life‐style entails, so that not only does he not want it, but he cannot even judge it. Such differences of viewpoint, sometimes in subtle guises, will always intrude themselves into comparisons across life‐styles.
In the classical model we ‘solved’ the problem of myopia by stitching together a chain of locally small steps, each step involving comparison of the relatively adjacent. Then it was not a preposterous act of faith to assume that meaning attached to the integrated sum of such steps. With standard of living comparison across life‐styles, however, we may encounter discontinuities, and across boundaries between life‐styles there may be nothing adjacent in the relevant sense.
It does not necessarily follow from the lack of clarity in the comparisons that there will be disagreement—at least as far as the ranking of stations in terms of standard of living is concerned. Perhaps the ordering will command agreement, and it may even correspond to one defined by a classical real income measure. Even then, however, we should still lack a measure of how far apart are the standards of living of different stations. And when we consider the special nature of the assumptions needed to provide such a quantification in the classical case, we might very well wonder why anyone should suppose that the difference between the US standard of living and the standard of living of Burma admits of exact quantification.
6 Migration and the Standard of Living
6.1. The Decision to Migrate
I shall deliberately take a somewhat narrow view of migration, concentrating on migration motivated by economic considerations. The migration motivated by personal safety, or to avoid maltreatment, may be called flight. The distinction between economic migration and flight is notoriously difficult to administer. Governments may wish to turn away people in flight, by claiming that would‐be migrants are economically motivated. However, migrants do have multiple and ambiguous motives.
When we consider migration defined as movement resulting from a desire for economic betterment, we are almost by definition looking at people who are (p.431) trying to improve their standard of living, at least as they perceive it. The literature on the economics of migration takes a rather simple view of the economic gain which drives migration.20 This is broadly a higher income, although the most influential model21 includes uncertainty concerning the realization of the improved income which the city promises.
The subtleties of the difference between a city and a rural standard of living are difficult to quantify. Close investigation has also tended to undermine the original Todaro view that urban earnings are simply higher and it is only the risk of unemployment that deters migrants. The comparison of standards of living in the country and in the city encounters many conceptual problems. For example, the city may offer better education, on the one hand, but more expensive, if not frankly inferior, housing, on the other. How to weight these differences is in part an issue of the city life‐style versus the rural life‐style itself.
Migrants are critical assessors of standard of living because they vote with their feet. Yet they may make mistakes; as von Weizsäcker's farmer is supposed to do, by underestimating city life in that case; or as a Todaro migrant, unfortunately unemployed in the city, finds out that he has done.
The British Industrial Revolution experienced mass migration from the country to the town, the evaluation of which is every bit as complicated as the evaluation of the, in some respects similar, migrations of the present day.22 The majority of migrants do not regret the decision to move. Not all have burnt their boats, but return to the country is fairly rare.23 Yet tastes are altered by the city, and the possibility of return to the old life‐style can be forfeited. As Hobsbawm, referring to the British Industrial Revolution, puts it:
Whether the Industrial Revolution gave most Britons absolutely or relatively more and better food, clothes and housing is naturally of interest to every historican. But he will miss much of its point if he forgets that it is not merely a process of addition and subtraction, but a fundamental social change. It transformed the lives of men beyond recognition. Or, to be more exact, in its initial stages it destroyed their old ways of living and left them free to discover or make themselves new ones, if they could and knew how.24
6.2. Life‐Style Breakdown
I explained life‐style breakdown as the arrival of a point at which the existing consumption set is no longer attainable. More common is the less drastic life‐style erosion. In the latter case a life‐style does not become unviable, but it fails to retain sufficient adherents to survive. The two are closely connected. As a life‐style approaches the point of forced extinction it can become harsher and less attractive, and people will leave it.
(p.432) History and the contemporary world provide examples of life‐style erosion or breakdown. Thus, nomadic pastoralism as a way of life is under acute stress, if not in the course of breakdown in various places (e.g. much of the Sahel), due to overgrazing, soil erosion, and the invasion of traditional pasture land by the crops and animals of settled farmers.
Changes such as these pose acute problems for moral or welfare evaluation. It is easy to say that large‐scale social transition imposes powerful stresses on those who undergo it, even if it is in some sense a movement towards a better way of life, or to a higher standard of living. This is to look at the moral issue too microscopically. A human life‐style is quite like a species of animal in some respects. It can be regarded as a resource, a valuable thing. It would be absurd to take this view to extremes, to want to preserve any way of life for its own sake, however horrible or brutal it might be. The danger in the modern world, however, is of an extreme concentration on a narrow range of life‐styles.
The motives that lead people to attack life‐styles include the simple and usual motive of self‐betterment. The loggers attack the forest to obtain for themselves and their families a better standard of living, longer life, more capabilities, etc. They might not recognize the terminology, but that is what they are doing. Consider the motivation of governments, World Bank officials, and various do‐gooders. The intentions with which they are armed are impeccable: better health, longer life, less hunger, etc. Yet there exist life‐styles that cannot receive these goods without disappearing. Can we be sure that it is good that they should disappear?
A station is characterized by a normalized price vector p, and a level of income in terms of the numeraire. Tastes are uniform and the stations can be unambiguously ranked in terms of their desirability. We assume them indexed by k, with station 1 the best station, and station N the worst.
Income might be simply money income and prices money prices. In that case, assuming money not to be desired for itself, we can normalize by setting income to equal 1, and expressing prices as costs in terms of income.25 Then hours of work, working conditions, etc., are not taken into account. However, if there is a fixed income at each station, we can normalize as described, and a station will be fully characterized by a vector of already normalized prices, p.
Classical standard of living equality exists when the same level of utility is associated with two stations. Equality between stations i and j may be conveniently expressed in terms of the equality of two values of the indirect utility function:
where the prices normalized on income equal 1. A direct method of testing for the satisfaction of (A.1) follows.
By the mean‐value theorem, (A.1) implies that there exists a value z between 0 and 1 such that:
where H k is the marginal utility of income in station k, x k is the vector of net demands in station k, and k is a station defined by prices:
with z chosen to satisfy the condition of the mean‐value theorem.
Rearranging (A.2), the unobservable H k value disappears and we are left with:
Standard of living equality corresponds to an index of normalized prices being the same. The price index is precise, but as its weights are obtained from a point determined by the mean‐value theorem, an approximation may have to be employed. If prices p i and p j are not too far apart, a good approximation is:26
where x k is again determined by the requirement of the mean‐value theorem. The move from equality to inequality makes an essential difference. The marginal utility of income cancels out of an equality measure, so that utility equality is perfectly measured by a real income index. A real income index of differences measures differences in real income and is silent on the question of how far utility levels differ.
The lottery index of the standard of living is defined as 100 times the probability of being translated to the best station (station 1), as against being translated to the worst station (station N), which would leave the agent indifferent between staying in his existing station and entering a lottery as specified. Let T k be the value of the lottery index in station k(k = 1, . . . , N). It can thus be specified that T N = 0 and T 1 = 100.
A chain method is employed to link together local choices to provide a global measure. There are N station (i = 1, . . . , N), and N − 2 probability values p j (j = 2, . . . , N − 1), such that the agent at j is indifferent between station j with certainty and station j + 1 with probability p j together with station j − 1 with probability (1 − p j). Then we have:
Arrow, K.J. (1963). Social Choice and Individual Values, 2nd edn. New Haven: Yale University Press.
Atkinson, A.B. (1970). ‘On the Measurement of Inequality’. Journal of Economic Theory, 2, 224–63.
Baumol, W.J. (1986). Superfairness: Applications and Theory, Cambridge, Mass., and London: MIT Press.
Debreu, G. (1959). Theory of Value: An Axiomatic Approach. New York: Wiley.
Diamond, P. (1967). ‘Cardinal Welfare, Individualistic Ethics and Interpersonal Comparisons of Utility: A Comment’, Journal of Political Economy, 75, 765–6.
Dixit, A., and Norman, V. (1978). ‘Advertising and Welfare’, Bell Journal of Economics, vol. 9, no. 1, 1–17.
Fisher, F. M., and Shell, K. (1972). The Economic Theory of Price Indices, New York and London: Academic Press.
Foley, D. K. (1967). ‘Resource Allocation and the Public Sector’, Yale Economic Essays, 7, 45–98.
Hammond, P. J. (1976). ‘Endogenous Tastes and Stable Long‐run choice’, Journal of Economic Theory, 13, 329–40.
Harris, J. R., and Todaro, M. (1970). ‘Migration, Unemployment and Development’, American Economic Review, 60, 126–42.
Harsanyi, J. (1955). ‘Cardinal Welfare, Individualistic Ethics and Interpersonal Comparisons of Utility’, Journal of Political Economy, 63, 309–21.
Hobsbawm, E. J. (1968). Industry and Empire. London: Weidenfeld and Nicholson.
Hollis, M. (1987). The Cunning of Reason. Cambridge: Cambridge University Press.
Marglin, S. (1974). ‘What Do Bosses Do? The Origins and Function of Hierarchy in Capitalist Production’, Review of Radical Political Economics, 6.
Peleg, B., and Yaari, M. E. (1973). ‘On the Existence of a Consistent Course of Action when Tastes are Changing’, Review of Economic Studies, vol. 40, no. 123, 391–401.
Phlips, L. (1974). Applied Consumption Analysis. Amsterdam: North‐Holland.
Pigou, A. C. (1952). The Economics of Welfare. London: Macmillan.
Pollack, R. A. (1970). ‘Habit Formation and Dynamic Demand Functions’, Journal of Political Economy, 78, 748–53.
Rawls, J. (1971). A Theory of Justice. Oxford: Oxford University Press.
Runciman, W. G. (1972). Critique of Max Weber's Philopsophy of Social Science. Cambridge: Cambridge University Press.
Samuelson, P. A. (1950). ‘Evaluation of Real National Income’, Oxford Economic Papers, 2, 1–29.
—— and Samy, S. (1974). ‘Invariant Economic Index Numbers and Canonical Duality: Survey and Synthesis’, American Economic Review, 64, 566–93.
Sen, A. K. (1970). Collective Choice and Social Welfare. Amsterdam: North‐Holland.
—— —— (1979). ‘The Welfare Basis of Real Income Comparisons’, Journal of Economic Literature, 17, 1–45.
Sen, A. K., et al. (1987). The Standard of Living, Cambridge: Cambridge University Press.
Stigler, G. J., and Becker, G. S. (1977). ‘De gustibus non est disputandum’, American Economic Review, 67, 76–90.
Todaro, M. (1969). ‘A Model of Labour Migration and Urban Unemployment in Less Developed Countries’, American Economic Review, 59, 138–48.
Usher, D. (1968). The Price Mechanism and the Meaning of National Income Statistics. Oxford: Clarendon Press.
—— (1980). The Measurement of Economic Growth. Oxford: Basil Blackwell.
Varian, H. (1974). ‘Equity, Envy and Efficiency’, Journal of Economic Theory, 9, 63–91.
von Neumann, J., and Morgenstern, O. (1944). Theory of Games and Economic Behaviour. Princeton: Princeton University Press.
von Wright, G. H. (1972). ‘The Logic of Preference Reconsidered’, Theory and Decision, 3, 140–69.
Williamson, J. (1989). ‘Migration and Urbanization’, in H. Chenery and T. N. Srinivasan (eds.), Handbook of Development Economics. Amsterdam: North‐Holland.
Yaari, ME. (1977). ‘Endogenous Changes in Taste: A Philosophical Discussion’, Erkenntnis, 2, 157–96.
(1) Medical economists employ the concept of the quality of life, and something like it is an essential ingredient of the cost–benefit analysis of a medical intervention which prolongs the lives of gravely ill patients. One could not substitute the term standard of living for the term quality of life. They plainly mean different things.
(2) The idea is inspired by modern physics in which the results of its application have been most impressive.
(4) This is equivalent to saying that the index is defined up to a linear transformation, as the values attaching to the worst and best outcomes are arbitrary. Note that the postulate that there exist worst and best outcomes already violates unrestricted domain.
(5) Strictly, measured preference relative to given best and worst outcomes.
(8) Note, however, that Rawls himself is not a utilitarian and would not accept the idea that expected utilities could be attached to agents in his original position.
(10) To say that the idea has meaning is not to claim that it must be true. It could be held that the lift from the abject poverty of Zaire to the less appalling deprivation of Ghana is a much more important step than the rise to the hugely greater Swiss standard.
(11) Traditional cartography is assumed. Today maps are prepared from satellite photographs, and this alters the point completely.
(15) Peleg and Yaari (1973) show that an optimal policy may not exist if the present self chooses subject to the constraint that a future self will not deviate. I think this point is fascinating but rather technical as it stands. My reason is that a policy may be constructed which cannot be improved by more than a small number, as small as desired.
(19) We can define the super‐consumption set to be the mathematical sum of the various individual consumption sets. However this does not reduce life‐styles to a case of the classical theory for two reasons. The super‐consumption set is not necessarily convex, and preferences and prices vary across it in a systematic manner.
(22) For a broad review of the human consequences of the British Industrial Revolution, see Hobsbawm, 1969, esp. ch. 4 and 5.
(23) Although more common in Africa, where the country never loses its attraction.
(24) Hobsbawm, 1969: 61–2.
(25) This way of expressing things is illuminating and is often employed. Thus a reporter might accompany film of a certain country with the commentary: these motor cars would cost the workers making them two years' earnings. The price of a car has been given in terms of years of income, as 2 units.
(26) There are many important points which this simple account ignores. Fisher and Shell (1972), for example, argue that with an intertemporal standard of living comparison the end‐weighted Paasche cost of living index number is to be preferred on the grounds that it is more appropriate when new goods are introduced.