Presenters:
Nancy Folbre, University of Massachusetts-Amherst
Robert Pollak, Washington University in St. Louis
Barbara Fraumeni, Northeastern University
Frank Stafford, University of Michigan

Michael Horrigan opened the session.

Nancy Folbre first gave a brief history of the collection of time-use data. She cited a 1878 letter from the Association for the Advancement of Women written to Congress protesting that the Census Bureau did not pay attention to nor measure non-market work. Housework was not considered "productive;" yet sexual prostitution was, and prostitutes, unlike housewives, were classed as workers. Initially the BLS, headed by Carroll Wright, did measure housework.

Folbre gave three reasons why it is important to collect information regarding time use:

1. Economics: the inability to measure non-market hours have biased statistics including the SNA on economic growth.
2. Sociology: time stress and our time crunch resonates with numerous welfare issues and quality of life concerns which need to be studied and credited.
3. Developmental psychology: how does time use affect child, family, and community outcomes?

The language of human capital needs to be expanded. There are some outcomes for families that do not affect economic growth but are still important. Folbre suggested thinking in terms of the production of human capabilities. She cited Sen's sense of capabilities, and Human Resource Accounting (see Jan W. van Tongeren's paper, "Human Resource Accounting (HRA) for Integrated Socio-Economic Analysis").

What kinds of synergies emerge from looking at these disciplines? Three arguments arise when combining these disciplinary approaches:

1. We should look more at quality, not just quantity of time. How time is organized and structured is important. For example, the number of hours an adult spends on child care may not be as important as adult availability or how adults spend the time.
2. Using an opportunity cost value has big disadvantages. For example, it measures an hour of leisure as being equal to an hour of child-care. An emphasis on capabilities will intensify the dissatisfaction with the opportunities cost (OC) approach. Should an hour of leisure be valued the same as an hour with a child or an hour preparing dinner? This deflects attention from the process of household production.
3. What are other inputs in addition to input time? We need to think more systematically about what inputs other than time we should be surveying and how they affect outputs like child development. We need to relate time to "child development" and "educational attainment." We need multi-use surveys and to add time-use modules to existing surveys, and vice versa.

In essence, how are we going to score the game? A danger is that such a multi-purpose approach is considered over ambitious, but Folbre thinks we could try to do everything at once.

Robert A. Pollak asked what can we say about the determinants of time use. Does tax policy affect time use, and if so, how? What can we say about the consequences of time use, especially with regard to children? Since we cannot do controlled experiments, how do we identify the effects of time use? How do we deal with sample selection bias?

There are numerous index and measurement issues. Index number theory is the touchstone in telling us what we are supposed to measure. But how do you measure national output? What exactly are we trying to measure? How can we measure it?

We need theory about what we are trying to do with time-use. We need to consider both market goods and services and non-market outputs. Measurement of non-market outputs gives you the ability to talk about technical progress. We need a measure and a concept of household technology and of joint production. We also have to look at both market and non-market inputs including time and household capital. We need other measures of productive inputs, not just capital.

Should we focus on non-market work or on non-market time? Does it matter if the time now spent in market work was previously devoted to non-work or to leisure? Is an increase in output really an increase in welfare if leisure is decreased? We need to think in a broad theoretical way how to pick up on issues dealing with leisure as well as issues of non-market work.

Barbara Fraumeni: "Human Capital, Intangible and Tangible Capital: Market and Non-market Activities"

Fraumeni described an all-encompassing economic model which includes the Education, Health, Market-all other, and Non-market sectors. She is focusing on investment activities that take place in the non-market sector because the non-market sector has a tremendous impact on the market sector.

She looks at non-human and human capital with an emphasis on human capital. There is much human capital formation going on in the non-market sector.

The Additional Scope of Output — What is in the economic model?

1. Investment in Education
2. Investment in Knowledge
   Investment in R&D (research and development)
   Investment in Other Knowledge (e.g., education)
3. Borrowed R&D and Knowledge Capital
4. Investment in OJT (on-the-job training)
5. Investment in Health
6. Investment in Children
   Births
   Raising Children
7. Household Production and Leisure Consumption
8. Other maintenance consumption

General Principles:

1. Human capital output imputations from lifetime income (numbers will look huge)
2. Value of imputations on the factor outlay side from the value of output minus the value of other inputs
   Exceptions:
   a. Borrowed R&D and other knowledge
   b. Other Knowledge
3. Loss of experience and slower career advancement as well as the value of parent time imputed income

Productivity is estimated incorrectly if you do not include non-market with market activity.

Frank Stafford described the importance of time-use data. It will help to highlight that investment is taking forms other than as savings. When looking at what is contributing to economic growth, we need to think of information technology also. How are shifts in demand for labor affecting household bargaining? Are we better off now than we were 25 years ago?

Skill Biased Technical Change

A paper by B. Wolfe showed that there has been a substitution of knowledge workers for other workers in nearly every industry. Skilled workers have extended their capacities to perform tasks previously done by others.

What do we mean? Juster & Stafford mean skill-biased technical change, or rather, progress. There are two types, skill-intensive and skill-extensive bias. This is skill-extensive bias. It may take place in the non-market as well as the market sector. It is part of the simple general equilibrium, the "big picture." It is not in a traditional human capital format. There is heterogeneous skills and/or heterogeneous consumer output.

How do we model this? Look at a General Equilibrium Model with heterogeneous human capital. Consider two groups of workers with skill-extensive change. One high-skilled group of workers with information technology now has greater productivity than they did before. They do not reduce their wage even with an influx of new workers, and this lowers the price of goods produced by less-skilled workers which drives down the wage.

Suggestions about what kinds of data we need:

Very comprehensive data
Contextual data
Spousal pairs and other family members
Multiple time diaries, over time to reduce bias
Non-diary measures
Comparable definitions across countries
Panel data

Group Discussion and Comments

Cathleen Zick, University of Utah, addressed to Folbre: There is little variation in the technology measures we have. It works out to be constant. What would we be measuring it in terms of home technology?

Timothy Smeeding, Syracuse University: He thinks computers (Internet access) are the differentiating factor between families.

Zick: We can use computers outside the home and still uniform across families. She agrees that it is important, but it is more difficult.

Folbre: Agrees in a cross-section, but across time and across families and countries, it will be different. We also need to look at household-relevant technologies—market substitutes, availability and costs.

Clyde Tucker, BLS: Suspects change is over time. Directed to Pollak: Does value of leisure change in relation to the substitutes you have for time, i.e., whether you do market work or non-market work?

Pollak: Responds that he does not know. He has not thought out all the complexities. Heterogeneity is very important and difficult to handle. For example, a surgeon painting his/her own house vs. somebody else doing it. Comment on the technology issue: Cannot look at technology if we do not have an independent measure of output. The difference will be in human capital, how we turn inputs into outputs.

Tucker: Leisure is one thing you cannot have others do for you.

Pollak: True. Think about the whole spectrum. Usually start with dichotomy. We need to broaden our thinking, but we should not focus solely on the notion of goods and services produced outside the market. We need to look at leisure too. An increase in output is not necessarily an increase in well-being if leisure has decreased.

Robert Michael, University of Chicago: Where are we in terms of concept and measurement? Opportunity cost of time is a correct concept but is not measured well, i.e., the wage rate is a bad measure. It should be measured some other way. We do not want to throw out the concept of opportunity costs because the measures are bad. What is the distinction between leisure and non-market time? Leisure is production of another kind. There is much leisure that you cannot buy.

Folbre: She does not want to throw away idea of the opportunity cost concept, but the problem with the OC concept is that it is a utility measure. She wants to look at the actual value of transferable goods and services to others because of some action. We are more interested in welfare of family/community than in utility of individuals. What an activity produces is of interest. A most fascinating result from the original Juster and Stafford study is that people value work more than leisure.

Pollak: A possible distinction between leisure and other non-market time is that some uses of time produce some output that we can measure directly. When we talk about a measure of output, it is important to ask, what is the output? Is output the painted house or the painted house plus the satisfaction derived from painting the house yourself? It's easy to run the models as if people do not care how they spend their time, but this is not true.

Jan W. van Tongeren, United Nations: Commented on the concept of capital: We need to broaden the concept from purely material capital (tangible) to intangible capital. The link between capital and production should be expanded. We can see capital as the store of wealth, e.g., art. This concept of capital as a store of wealth is also relevant for human capital which can be used for leisure too, e.g., happiness, which is not necessarily linked to production (of income). Some human capital is used to produce, but not all of it, as we are not solely producers. It is hard to describe human capital in only physical terms. We need to unlink capital from production.

Fraumeni: She is not sure how to use "other knowledge. " The model is tied to an income-type concept, but there is something more going on.

Jacqueline Eccles, University of Michigan: First, what role does social capital as an output play in economic models? Second, psychologists have spent a lot of time thinking about how to value "leisure," including personal pleasure. There is much talk about individual identities and social identities. These are aspects of both the person and their membership within a group. When thinking about investments in families as units, consider what function the investment plays in these units. Consider the aspects of a person and the aspects of groups. For example, father may paint his house to be role model for his kids, showing that there is "attainment value" from the activity. Activities convey messages about values and who we are.

Folbre: Good point. How do people talk about time use and convey time use? Some work is being done on "social capital"—community time, but the use of the word "capital" may not be a good idea here. We need to think about qualitative dimensions. Some forms of social capital are good, but some are bad (the Junior League vs. gangs). Usually more (capital) is better, but this is not necessarily a good thing.

Katharine Abraham: She is struck by the discrepancy between how much time people say they work and how much they actually work. This might be changing social expectations. Study has had people indicate what activities are work and what are leisure—hard to quantify because one person's work is another's leisure.

Andrew Harvey, Saint Mary's University: His work has been focused on measuring outputs. The same activity can be coded the same or differently when done at different times by the same person. The study asked people to indicate which activities are work and which are leisure. This is hard to quantify because one person's work is another's leisure. Leisure may be an output.

Smeeding: The concept of joint production (getting two things done at once) is very important. There are negatives as well as positives, e.g., caring for someone.

Sarah Fenstermaker, University of California at Santa Barbara: We need to appreciate the gulf between the conceptual frameworks and empirical capabilities. "Irrational" behavior exists within households. Families have been devoted to a consistent rationing of household labor. The surgeon who appears irrational by painting the house may just be sending a message about how a parent should spend their weekend. If you are invested in a production of identity, work and leisure become the occasion for jointly produced outputs of complicated social identities. Also there is joint production of gender in households as well as of goods and services, and a production of identity, of inequality.

Pollak: The notion of attainment value is important for discussing determinants of time use. What does attainment value have to do with measuring output? It probably does matter, but how?

Eccles: What some of us call outputs, others call inputs.

Pollak: We need a theoretical framework. It could give us answers to what counts and what does not count.

Folbre: What we are engaged in, here, is a production of identity, a process of challenging the income accounts approach from economists.

J. Steven Landefeld, Bureau of Economic Analysis: National income accountants have never claimed that GDP is a measure of welfare. GDP is not a welfare measure; perhaps we need additional measures. The accountants have no problem including all these different things when measuring GDP, it's just that they are very hard to measure.

Folbre: This is not a critique of GDP per se, but acknowledgment of the tendency to measure success by national income accounts. We are criticizing the interpretation and use of the measure. It's a matter of changing the cultural idea of winners and losers.

Fraumeni: In the model, we are trying to construct a measure of GDP that includes market and non-market production. We are not trying to measure welfare either.

Sandra Hofferth, University of Michigan: The question is, who is receiving return from non-market work? The parent? The child? Society? Another individual? Oneself? Who benefits? There are different levels of accounting.

Fraumeni: That's true. She has not dealt with the levels yet.

Luisella Goldschmidt-Clermont, France: What about the time-use survey itself? What have we learned in the past century on what we can get from a time-use survey?

Folbre: It is important to have a broad theoretical discussion and a large-scale survey that allows us to contextualize time-use better. The implicit agreement is that there is a need for time-use surveys. We want to develop a survey that can get at necessary information. "Time use" is forcing us to question and rethink basic economic theoretical concepts.

Miron Straff, National Academy of Sciences: Is there an impediment to developing measures? Can we not separate the measurement of time from the experience of time? How can we get a measurement of time that is rational? There are different values of time for different people.

Folbre: The same objections and rules apply to money. Some people get much more pleasure from $10 than others.

Straff: No, they get more pleasure from $10 than $5. We need to get at the measurement.

Folbre: She wants to collect data in such a way that persons from other disciplines can use it too. The interdisciplinary approach is helpful and separates the division of labor.

Horrigan: There will come a day when we need to make fundamental decisions about what data we will collect and for what purpose the data will be collected. This will come down to methodology of collection. There is a gulf between theory and measurement. The data collected may not be data to inform theories.

Tucker: We need to talk to people to get a real idea.

Robin Douthitt, University of Wisconsin-Madison: She has good standards on how we measure time use. We need to be able to get at context. The measurement differences between consumption, production and leisure are only going to worsen as we look at increasing technology and more people working at home. We need to look closely at how people perceive the use of time. How will physical separation, i.e., working at home, impact families? Will it be a benefit?

Thomas Juster, University of Michigan: Every activity produces essentially two products, an output and satisfaction, i.e., process benefits. A finished product may also give satisfaction to others. The surgeon gets satisfaction from painting the house, but the painted house gives satisfaction to the family and to people driving by as well. Work and play are social activities, whereas housecleaning is a solitary activity. We want to get a measure of all outputs, e.g., the painting and the painted house. We want to get satisfaction measures to see why people want to do the things they do, but we can probably only get at the average, we cannot get this "at the margin," for modeling. An objective is to produce a database that can be used for modeling behavior.

Session ended at scheduled time.

back to top