Value-free economics?

A recent volume by Vivian Walsh and Hilary Putnam,  The End of Value-Free Economics, brings to a fine point a line of argument that has been brewing for fifteen years: is the logical positivist insistence on separating "fact-based" science from "value-based" ethics any longer a tenable one? Most particularly, are there now compelling reasons for declaring that mainstream economics needs to recognize that the distinction is wholly untenable? Is the zeal for insisting on "positive" economics now unsupportable? Should economists at last recognize that Lionel Robbins' strong exclusion of normative language from the science of economics both unjustified and unwise?  Walsh and Putnam argue that the answer to each of these questions is definitive: the strict dichotomy between fact and value in economics can no longer be supported.

The issue of facts and values has a number of sources within the empiricist tradition.  There is Hume's view that we can't derive "ought" from "is"; or in other words, that moral judgments are logically independent from empirical beliefs.  There is the positivists' criterion of significance, according to which the meaning of an utterance reduces to the empirical experiences that would demonstrate its truth or falsity.  (The two propositions together imply that moral sentences are meaningless or "non-cognitive", since the first proposition concedes that no empirical experience can demonstrate the truth or falsity of a normative statement.) And there is the positivists' idea that science is exclusively concerned with "facts"; but the first two propositions consign moral statements to the category of "value" rather than "fact", so science cannot contain normative vocabulary.  Another source was internal to debates within neoclassical economics itself: Lionel Robbins's arguments against interpersonal comparisons of utilities, based on the idea that making such comparisons unavoidably involves taking an evaluative stance towards the individuals in question.

The key idea advanced in The End of Value-Free Economics is that none of these philosophical ideas have survived the critique of positivism that was offered within philosophy of science and philosophy of language over the past fifty years.  The attempt to draw a sharp line between "fact" and "value" turns out to be impossible.  And this is equally so in economics.

Consider an example.  The concept of Pareto efficiency is defined in value-neutral terms: a distribution is Pareto-efficient if there is no other distribution that improves some individuals without harming at least one individual.  The concept of distributive justice is not value-neutral; it invokes the idea that some distributions are better because they are more fair or more just than others.  The positive economist holds that the latter set of distinctions are legitimate to make -- in some other arena.  But within economics, the language of justice and equity has no place.  The economist, according to this view, can work out the technical characteristics of various economic arrangements; but it is up to the political process or the policy decision-maker to arrive at a governing set of normative standards.  Walsh and Putnam (as well as Amartya Sen) dispute this view on logical grounds; and this leaves the discipline free to have a rational and reasoned discussion of the pros and cons of various principles of distributive justice.

Raising the issue of value-neutrality for economics is a frontal assault on the uncritical positivism that neoclassical economics incorporated from the 1930s and forward. But it is also an attack on something else--the no-longer acceptable idea that economists can only tell us how things are, not how they should be. Is famine worse than food sufficiency? Is literacy better than illiteracy? Is good health an improvement in wellbeing? If we take the view that "positive economics" cannot contain normative judgments, then none of these questions could be answered by an economist. "It depends on what you value." What Walsh, Putnam, Sen, and other contributors to this volume want to say is that this response is idiotic, and there is no basis in logic, science, or methodology that would support it. Of course economics, and economists, can find that starvation is a bad thing. Instead, they maintain that the best philosophy of language and philosophy of science supports the idea that value concepts and descriptive concepts are intermingled or "entangled", and that we can offer good reasons and evidence for evaluating claims involving both.

Why, some readers will ask, has Hilary Putnam become a central figure in this emerging debate? Putnam is known as a technically astute philosopher of mathematics, logic, and physics, and a philosopher of language; he is known for a sometimes wavering adherence to several versions of scientific realism; and he has made contributions of the greatest importance to each of these fields. But how did he come to get deeply immersed in the issue of the role of values in economics?

Vivian Walsh is one important part of the answer. Walsh undertook a series of articles in the 1980s and 1990s that were critical of the logical positivist assumptions that have lingered within the methodology of neoclassical economics. He took encouragement from the writings of Amartya Sen on welfare economics that confidently dismissed these positivist assumptions -- for example, the idea that science could not incorporate values or that statements about values were meaningless. (Lionel Robbins is offered as a particularly clear advocate of these views.) And Putnam worked up his reactions to these ideas into a novel book in 2002, The Collapse of the Fact/Value Dichotomy and Other Essays.

A key construct in the collaborative thinking that Putnam and Walsh have done together is the idea of the "second phase of classical theory." (Harvey Gram discusses this construction in detail in his contribution.) Walsh introduces the idea and Putnam follows up in his essay. What this refers to is the fact that classical political economy, as expressed by Smith and Ricardo, underwent a major intellectual revival in the 1960s when thinkers like Piero Sraffa proposed reappropriating some of their key analytical ideas. Sraffa's Production of Commodities by Means of Commodities : Prelude to a Critique of Economic Theory was a key product of this rethinking. The rethinking itself came about because of an uneasiness about the premises of neoclassical economics, and it stayed close to the core logical ideas. The first revival focused on Ricardo, but the second phase, Walsh argues, has given a much more nuanced interpretation of Smith himself.  Walsh finds that this reconsideration has been led by Amartya Sen and is more wide-ranging. Here is why Walsh thinks this reconsideration of Smith is important:

This is because Smith embedded a remarkable understanding of the core concepts of a political economy whose implications for moral philosophy he understood and explored.  The Smith texts as a whole offer a rich tapestry, interweaving threads of classical analysis, moral philosophy, jurisprudence, and history. (7)

And here is how Putnam summarizes Sen's contribution to this reconsideration of classical political economy:

If we are to understand Sen's place in history, the reintroduction of ethical concerns and concepts into economic discourse must not be thought of as an abandonment of "classical economics"; rather it is a reintroduction of something that was everywhere present in the writings of Adam Smith, and that went hand-inhand with Smith's technical analyses. This is something that Sen himself stresses. (quoted by Walsh, 29)

Amartya Sen has argued throughout his career for the robust possibility of reasoning about value issues -- in economics and elsewhere. (A very early place where Sen takes up this topic is in "The Nature and Classes of Prescriptive Judgements"; link.) Much of what Sen brings to this debate within economics, according to Walsh and Putnam, is found in his capabilities theory as a foundation for a theory of welfare or wellbeing. This theory is based on the idea of human functionings; and there is a plain intermingling of factual and evaluative ideas associated with this notion.  We need to know what human beings can and want to do, before we can say how well off they are. And this means bringing in orienting human values at the foundations. Putnam draws attention to Martha Nussbaum's list of core human capabilities. Anyone reading these descriptions would agree that they presuppose human values. And Nussbaum (as well as Sen and Putnam) believes that we can rationally discuss and evaluate these. But if welfare economics is to incorporate a substantive notion of human wellbeing, then it plainly cannot be maintained that it is "value-free".

Another important locus for Sen's reintroduction of ethical concepts into economics is his critique of the narrow conception of individual economic rationality.  As Sen puts the point in "Rational Fools" (link),

A person thus described may be "rational"in the limited sense of revealing no inconsistencies in his choice behavior, but if he has no use for these distinctions between quite different concepts, he must be a bit of a fool. The purely economic man is indeed close to being a social moron. Economic theory has been much preoccupied with this rational fool decked in the glory of his one all-purpose preference ordering. To make room for the different concepts related to his behavior we need a more elaborate structure. (336)

Sen introduces the idea of "commitments" directly into the concept of economic rationality.  Individuals choose among preference rankings based on their commitments -- to each other, to political ideas, to groups with whom they have decided to affiliate.  And this brings normative ideas directly into economic decision-making -- and therefore into the domain of economics.

Walsh and Putnam insist on a point that seems very important to me as well: it is the dichotomy between facts and values, or between positive and normative analysis, that they reject. They do not reject the idea that there are facts and there are values. But they believe in important respects these categories are intertwined and inseparable. They argue for "entanglement" and "rich description." They believe that it is fully possible and acceptable to engage in rational debates over the best theory of justice, or human nature, or human freedom; and to do so within economics as well as outside of economics.  And they believe that science can handle its goals without this sharp dichotomy.

The politics color wheel

The problem of mapping or classifying people's political attitudes is more complicated than it looks. Placing people on a spectrum from left to right is convenient but over-simple. It assumes that there is a single dimension of political difference, ranging from conservative to liberal, and that everyone can be placed somewhere along that spectrum. But social and political attitudes aren't single-dimensional. So we'd like to have a way of mapping the space of attitudes that captures the separate dimensions that go into political attitudes.

Most of the schemes above take a step forward by hypothesizing two dimensions reflecting attitudes towards the involvement of the state in the economy and in individual social and moral life. An economic conservative opposes state involvement in the economy. A social liberal opposes state involvement in individual choices and lifestyle, while a social conservative favors intervention in individual behavior when it comes to certain "social" issues. A libertarian opposes both forms of state intervention and an authoritarian supports both.

This kind of classification is an improvement, but it doesn't fully satisfy. It doesn't really capture the substantive moral ideas that might underlie these attitudes, the political philosophy that might underlie political preferences.

A more substantive effort at classification might take the form of a cluster of values shared by various groups. Conservatives favor maintaining traditional values (religion and morality), maintaining an existing system of wealth, power, and prestige (market and social inequality), and limiting the use of state power to these functions. Progressives favor greater social equality and opportunity; they favor the free expression of individual life choices; and they favor using the authority of the state to increase the equality and freedom of ordinary people. Conservatives defend existing inequalities and oppose redistribution. Progressives condemn certain kinds of inequality and justify redistributive measures to redress these inequalities.

Yet another approach is to consider a diagnostic set of issues and classify the individual according to where he/she falls on these issues. For example:

• Guaranteeing a woman's right to choose
• Restricting the sale of assault rifles
• Carbon tax
• Inheritance tax
• Immigration reform
• Affirmative action

We might imagine grouping the population into a sort of Venn diagram.  Suppose that each circle represents individuals who include the specific issue within his/her top five issues.  Overlaps indicate the circumstance of individuals who share the issues of each overlapping circle.  So region A represents individuals who rank civil liberties, social justice, and environmental issues in their top five; region B represents individuals who rank markets, guns, right to life, and religious fundamentalism in their top five priorities.  On this diagram, there is no one who is a militia activist and a social justice advocate.

Several questions arise here: are there correlations across individuals with respect to these issues? Is there a set of philosophical principles that underlie the choices that individuals make within connected clusters?  And do the clusters correspond to identifiable political groups?

Ideal types, values, and selectivity

image: cover of Max Weber, The History of Commercial Partnerships in the Middle Ages

I've never really understood why the exposition of one of Max Weber's most important methodological ideas, his theory of ideal types, occurs in the context of an essay that is primarily about the role of values in the social sciences.  This is his essay, "'Objectivity' in Social Science and Social Policy" in The Methodology of The Social Sciences.  The central thrust of the essay is to attempt to spell out the ways that objectivity and truth relate to partisanship and values within the course of social science research and teaching.  And Weber draws a sharp distinction between what can be known and demonstrated (empirical facts and causal relationships) and what can be internally tested for consistency but not demonstrated (fundamental value commitments).  Here is how Weber puts the point early in the essay:

What has been the meaning of the value-judgments found in the pages of the Archiv regarding legislative and administrative measures, or practical recommendations for such measures? What are the standards governing these judgments?  What is the validity of the value-judgments which are uttered by the critic, for instance, or on which a writer recommending a policy founds his arguments for that policy?  In what sense, if the criterion of scientific knowledge is to be found in the "objective" validity of its results, has he remained within the sphere of scientific discussion? We will first present our own attitude on this question in order later to deal with the broader one: in what sense are there in general "objectively valid truths" in those disciplines concerned with social and cultural phenomena? (50-51)

The theory of ideal types is an important contribution to the specification of the nature of concepts in the human and historical sciences.  But why is this subject particularly relevant in the context of a discussion of values and social policy?

One reason for the conjoining of ideal types and values is the unavoidable fact of selectivity in the social sciences.  Weber makes the point repeatedly in this essay that there is an infinite depth to social phenomena -- even to a single phenomenon or event -- and therefore it is necessary to select a finite representation of the object of study if we want to approach a problem scientifically.  But how do we select a specific aspect of a phenomenon for study?  We do so on the basis of a judgment of what aspects are important -- and this is a value judgment, either directly or indirectly.  In a very specific sense, our interests (material and intellectual) guide the formation of our social-science research projects.  "The quality of an event as a 'social-economic' event is not something which it possesses 'objectively.' It is rather conditioned by the orientation of our cognitive interest, as it arises from the specific cultural significance which we attribute to the particular event in a given case" (64).  And a few pages later: "Social economic problems do not exist everywhere that an economic event plays a role as cause or effect -- since problems arise only where the significance of those factors is problematical and can be precisely determined only through the application of the methods of social-economics" (66).

Or in other words: what constitutes an economic situation as a "problem" is the fact that it has consequences that intersect with things we care about -- or our fundamental scheme of values.  So the subject matter of "social-economics" is doubly dependent on our interests: our cognitive interests in how things work, and our practical interests in how to promote "good" outcomes and avoid "bad" outcomes.

This is where the issue moves into connection with the theory of ideal types.  Weber makes it clear that he regards the formulation of a scientific research topic as being generated by a set of interests -- cognitive and practical.  And this is where the idea of an ideal type is relevant.  Because an ideal type is a selective, one-sided representation of an aspect of social life.  Here is the foundational description that Weber offers in conjunction with the idea of a "commodity-market":

This conceptual pattern brings together certain relationships and events of historical life into a complex, which is conceived as an internally consistent system.  Substantively, this construct is itself like a utopia which has been arrived at by the analytical accentuation of certain elements of reality.  Its relationship to the empirical data consists solely in the fact that where market-conditioned relationships of the type referred to by the abstract construct are discovered or suspected to exist in reality to some extent, we can make the characteristic features of this relationship pragmatically clear and understandable by reference to an ideal-type.  It is not a description of reality but it aims to give unambiguous means of expression to such a description.  It is thus the "idea" of the historically given modern society, based on an exchange economy, which is developed for us by quite the same logical principles as are used in constructing the idea of the medieval "city economy" as a "genetic" concept. ... An ideal type is formed by the one-sided accentuation of one or more points of view and by the synthesis of a great many diffuse, discrete, more or less present and occasionally absent concrete individual phenomena, which are arranged according to those one-sidedly emphasized viewpoints into a unified analytical construct. In its conceptual purity, this mental construct cannot be found empirically anywhere in reality. (90)

This specification of the concept of an ideal type links back to Weber's discussion of the value of specialization in the social sciences: "the justification of the one-sided analysis of cultural reality from specific 'points of view' -- in our case with respect to its economic conditioning -- emerges purely as a technical expedient from the fact that training in the observation of the effects of qualitatively similar categories of causes and the repeated utilization of the same scheme of concepts and hypotheses offers all the advantages of the division of labor" (71).  In other words, the specialization of research methods and training that is implied by the establishment of disciplines in the social sciences is justified pragmatically rather than epistemically -- it is the practical advantage in research productivity rather than the nature of the social world that justifies the establishment of disciplines.

There is no absolutely "objective" scientific analysis of culture -- or put perhaps more narrowly but certainly not essentially differently for our purposes -- of "social phenomena" independent of special and "one-sided" viewpoints according to which -- expressly or tacitly, consciously or unconsciously -- they are selected, analyzed and organized for expository purposes. (72)

This locates the role and justification of the "ideal type" very precisely.  An ideal type -- of a market economy, of a university, or of a banking institution -- creates a selective model of social organization that can then be explored analytically and empirically by specialists.  The ideal type is not intended to be a general representation of a category of phenomena; but rather as a heuristic model that permits exploration and extension of some of the characteristics of the concrete social institutions and behaviors that it partially represents.

Throughout this discussion is woven the methodological debate between advocates of nomothetic and idiographic interpretations of the social sciences (link).  Weber indicates his own understanding of the crucial importance of the individuality and particularity of historical phenomena -- without abandoning the viability of discovering limited generalizations and regularities.

Laws are important and valuable in the exact natural sciences, in the measure that those sciences are universally valid.  For the knowledge of historical phenomena in their concreteness, the most general laws, because they are most devoid of content are also the least valuable.  The more comprehensive the validity -- or scope -- of a term, the more it leads us away from the richness of reality since in order to include the common elements of the largest possible number of phenomena, it must necessarily be as abstract as possible and hence devoid of content.  In the cultural sciences, the knowledge of the universal or general is never valuable in itself. (80)

Here again, the historically detailed ideal type is a better basis for historical and social analysis.

Here is one additional qualification that Weber offers that is very important to his understanding of social-scientific knowledge:

We have designated as "cultural sciences" those disciplines which analyze the phenomena of life in terms of their cultural significance.  The significance of a configuration of cultural phenomena and the basis of this significance cannot however be derived and rendered intelligible by a system of analytical laws, however perfect it may be, since the significance of cultural events presupposes a value-orientation towards these events.  The concept of culture is a value-concept. (76)

This makes a final important connection between the two themes of the essay, objectivity and values.  Values come into the social science at the stage of defining and examining social-economic-cultural problems; they come into the choice of subject matter by establishing the framework in terms of which a phenomenon is a "problem"; and they must be invoked in our interpretations of the phenomena when we attribute cultural meanings to the participants.  The construct of the ideal type provides conceptual resources for each of these zones of intersection between social-science inquiry and the schemes of values that we humans endorse.