Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Unpaid Productive Activity Over the Life Course James N. Morgan We all start life being dependent on others, net recipients of goods and services. Then, after an extended period of productive activity, some of it paid for in money, we once again consume more than we produce, although we may well have saved enough to pay for most or all of that consumption. Much is known about people's paid work, both about their earnings and their hours of work. The rapid increase of paid work by women, even those with small children to care for, has been studied thoroughly. It is clear that as a society we are at some kind of peak in terms of the aggregate amount of productive time available in relation to the population. Even without a resurgence in births, as the baby- boom generations reach retirement age we shall have fewer paid workers and probably fewer aggregate paid work hours available relative to the total population to be fed, clothed, and housed. These factors put unpaid work particularly the possibility of more mutual and self-help among the elderly in our population- in a new perspective. Unfortunately, because there are no mar- kets, no official records, and no taxable income from unpaid pro- ductive effort, our data base on the subject is considerably less than adequate. There are measurement problems in evaluating unpaid productive effort that are far more serious than those James N. Morgan is a research scientist at the Institute for Social Research, and professor of economics at the University of Michigan. 73
74 JAMES N. MORGAN found in assessing paid work, which at least has some readily available data in the form of wages. Particularly after taxes, wages may not measure productivity precisely, but they are com- monly accepted as an approximation, assuming that markets, even for labor, work reasonably well and that taxes are not too unfair or discrepant. But how should we value unpaid work? For those who also work for wages we could argue that they rationally equate their marginal returns in the two kinds of activity, but there are diff;- culties with this approach. Many people, particularly among the young who do the most unpaid work, claim to want more work than they can find; others get higher overtime wages than their regular wage. Also, imputing wages to those not currently work- ing poses serious problems of "selection bias" for which there are only proximate solutions. And many unpaid productive activities have a component of pleasure, or direct consumption. Parents may well spend more time caring for their children than just what is required for investment in their education and rearing (Herman c.anital) and many a home gardener does things that he ~ ~ ~ ~¢ ~ ,, ~ ~ ,, c, or she would not pay others to do. ~ Others have proposed valuing unpaid productive time accord- ing to market wages paid for that kind of activity, even breaking down such activities as housework and child care into compo- nents for the purpose. This method, however, requires very detailed accounts of activities, complete with time-diaries, and so far has not proved practical for large-scale research. Given these circumstances, and particularly the lack of much variation in the reports of amounts saved per hour in relation to paid earnings, this paper focuses on hours devoted to productive activities, which are defined here as activities that produce goods or services that otherwise would have to be paid for. This paper looks at hours reported as spent per year in such activities with- out attempting to solve the problem of potentially different val- ues depending on who is putting in the hours or what the activity is. For example, for work done on the car or house, in which case * We have, in the case of do-it-yourself activities, asked people how much they think they saved by doing that activity. After subtracting the cost of materials and the recreational aspects, we tended to get amounts that appear reasonable-in relation to the hours reported spent-implying modest hourly earnings. Those implied earnings do not vary much with the individual's paid wage, however.
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 75 we asked about both hours and the amount of money saved, both can be reported. In the case of growing and preserving food, we asked only about amounts saved, on the assumption that people could not recall the scattered hours spent on these activities. Among the data presented here on hours spent in various unpaid productive activities are the following: For some activi- ties difficult to attribute to individuals, there are data at the family level: volunteer work for organizations (for example, churches and charities), time helping friends and other individ- uals, and repairing houses or cars. At the individual level there are data on housework, in relation to paid market employment, and on time kept available for paid work but diverted by unem- ployment, strikes, illness, or caring for others who are ill. Here we follow men and women separately, focusing on age differences and only occasionally going beyond age to the demographic changes that affect the use of people's time: marriage, children, home ownership, and length of residence in an area. Most of the data are from a pair of national studies on philan- thropy that were conducted for the Commission on Private Phi- lanthropy and Public Needs (Morgan et al., 19791. Although there are a few other studies on volunteer activity and, of course, many larger samples giving data on paid work hours, this paper deals primarily with data from the two studies referred to. There are also time-budget studies in progress at the Institute of Social Research (ISR) that report time use over a whole year or for an average week (Hill, forthcoming). Preliminary comparisons indi- cate no glaring discrepancies. In comparing individuals or families, a researcher must either use the same short period, expand to a uniform longer period, or use the longer period in the questions. The presumed accuracy of time-diary reports on yesterday is substantially reduced when they must be multiplied by 366, but a more important issue- interpersonal comparisons can be distorted. Because diaries com- pel accounting for 24 hours, however, they can be useful in per- haps constraining the exaggerations one might get in asking about only one or two activities. We must assume that by asking about all of last year, we may get exaggeration from a respon- dent's focusing on a few things, which is offset by understate- ment from memory loss. In the case of housework and child care where we ask about current time per week, we may also get some exaggeration, compounded by the fact that people can do more than one thing at a time.
76 JAMES N. MORGAN PRODUCTIVE ACTIVITIES MEASURED AT THE FAMILY LEVEL Volunteer Time In 1974 we interviewed two national samples: one based on an area probability sample and the other on an Internal Revenue Service sample interviewed by the Census Bureau that asked about gifts of time and money to organizations and to individuals (Morgan et al., 19791. The pooled data indicated that some 6 billion hours a year were given to religious and charitable organizations in 1973, somewhat more than half of it by husbands and single (male and female) heads of households. If this time were valued at some reasonable opportunity-cost wage, it would be roughly equiva- lent to the aggregate money contributions that year. And even though it is mostly education that accounts for differences in volunteer time and income that accounts for differences in the donation of money, the two forms of altruism go together: Those who gave time also gave money. Clearly, then, there is a philan- thropy syndrome. We know much less, however, about the dynamics by which people begin either type of philanthropy (donating time or money), or whether there is some order in which they get into the syndrome. The few questions we asked about parents' altruism had little explanatory power. It was mostly the well-educated, the affluent, and the young who gave time to church or charity, and even among these groups a few gave substantial amounts while many gave little or none, making the aggregate estimates shaky. There was also time spent helping friends, neighbors, or relatives (for example, caring for sick persons not living with the individual surveyed). Some 30 percent of families reported giving some such time, again with a few giving a lot, and a crude estimate of the aggregate for 1973 was 2 billion hours, as compared to the 6 billion hours donated to · ~ organizations. The focus in this paper is on age, particularly in view of the changing age distribution of the population. But because volun- teer time is so strongly affected by education, and education is correlated with age, we can see a purer age relationship by relat- ing each individual's volunteer time to the average for his or her education group (averaging the ratio of husbands and wives for families containing married couples). Table 1 shows that there is
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 77 TABLE 1 Time Given to Philanthropic Organizations, in 1973, in Hours and in Relation to the Average for One's Education Group Ratio of Time to Average for Each Average Hours Individual's Age per Household Education Group 18-34 58 0.58 35-44 132 1.09 45-54 91 0.84 55-64 94 0.84 65-74 63 0.80 75 + 60 0.72 still an age pattern although it is somewhat attenuated. The greatest time in relation to the amount expected for one's educa- tion level is given at the very age when the combination of job and family is putting the most pressure on people's time anyway. When looking at age differences, however, one must always spec- ulate whether they also or merely reflect cohort differences among the generations. There could be a pure decline with age plus a tendency for the younger generation to be less altruistic, and of course there are more one-person households at both ends of the age spectrum. We also know from an earlier study asking about all volunteer work for organizations or relatives that income dominated the explanations, so much so that, surprisingly, the single best pre- dictor was the number of modern appliances in the home (Mor- gan et al., 19661. But the next best predictor was education, fol- lowed by Tong residence in an area. Age never accounted for much in that analysis. In some earlier studies of retirement we asked people not yet retired whether they expected to do more volunteer work when they retired, and we asked the retired whether they were doing more volunteer work than before retire- ment. Of course these are different generations or cohorts and there are undoubtedly memory biases, but substantially more expected to increase their volunteer work than reported having done so (Barfield and Morgan, 19691. A major implication of these findings is that it is not available time that seems to drive philanthropic activity but abilities and purposes. People with children get involved in activities, ranging from church to Little League, that are directed at the socializa- tion of their offspring. Active, visible people are urged to take
78 JAMES N. MORGAN leadership roles, and the more money they give, the more they are asked also to give time. All this is indirect inference, how- ever, as ~ know of no studies of the dynamic process. Tax laws would appear to encourage more time donations, in comparison to money donations, among Tow-income people who do not itemize and cannot therefore get income tax rebates on their charitable contributions of money. But they may also be spending more time earning a living or looking for more paid work. The other implications of the data are that paid work and unpaid work do not appear to be interchangeable and that the reduction in paid work hours with age, even after retirement, does not appear to lead to any substantial increase in volunteer work. Perhaps incentives are more important than free time. With volunteer work we usually assume that the opportunities are there. Emergency Help From many studies we know that there is relatively little money being given to relatives not living together and not much time regularly devoted to such things as caring for grandchil- dren. A recent analysis of a sample close to retirement age found only 1 in 40 persons currently providing any support for parents A. Morgan, 19831. But a major element in interfamily help might be the availability of help in emergencies. The 1980 interviews of the Panel Study of Income Dynamics asked about emergency giving or receiving of time or money to or from friends or rela- tives in the previous five years. A chapter in Volume 10 of Five Thousand American Families gives the details of the findings (Morgan, 19831. A summary will have to suffice here. In the study we asked about whether emergency help was given and to whom or from whom; we also asked about availabil- ity and, in the case of money, whether it was to be repaid and if so whether it would be repaid with interest. The focus was on time rather than money, but from the data it can be said that the giving of money in emergencies was largely from those in their peak earning years to their children. Four out of five persons felt some emergency help would be available to them in an emer- gency, and two out of three said money help was available. But what about actual time help? A constant 30 percent of the individuals surveyed reported
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 79 giving some time in the past five years, at all ages until the most advanced. After 75 the frequency dropped off. A smaller fraction reported getting help, indicating some memory bias but perhaps also a tendency for those in trouble to get help from more than one person. Reports of getting time help fell with age from an early 20 percent to a little over 10 percent of those aged 55 to 64; the figure then rose to slightly over 20 percent for those aged 75 and older. Admittedly these are crude data, based on fuzzy recall and a somewhat ambiguous def~r~ition of giving, but they indicate rela- tively little activity. On the other hand, the vast majority indi- cated help would be available, and this insurance aspect may be more important than the actual time spent. Unpaid Productive Work Around the House The remaining activity that has been measured mostly at the family or household level is home production or do-it-yourself activities: repairing or improving homes or cars, or growing or preserving food. Prior to the Panel Study, similar questions were asked in the 1965 Productive Americans study (Morgan et al., 19661. They covered the year 1964, asking separately and specif~- cally about the following: Percentage Yes 100 Work around the house, such as preparing meals, cleaning, and straightening up 50 Painting, redecorating, or major housecleaning 78 Sewing or mending (women only) 25 Growing own food 24 Canning or freezing 25 Anything else that saved you from having to hire someone else to do it 46 Volunteer work without pay such as work for church or charity or helping relatives 17 Taking courses or lessons (investing in self) (head of family only) The other side of the coin was also considered, namely, whether the family paid to have outside help with these activities: Percentage Yes 14 Things that have to be done around the house, such as preparing meals, cleaning, and making repairs 30 Sending out any of the laundry
80 10 Care of children 80 Eating out (42 percent over a week or more) 14 Painting, redecorating, or spring cleaning 16 Lawn work JAMES N. MORGAN Although substantial fractions of respondents reported buying back some of their own time in these ways, the amounts reported were usually trivially small. When we totaled the five items of home production in the first list above, they came to a total of 206 hours per year (heads and wives together). It was discovered that it was not age that mat- tered but the absence of one or more inhibiting factors, many associated with the family life course. Those who did the most such work were married, lived in single-family structures with large families, had a highly educated family head, lived in a rural area, and had no young children under two years of age. (A systematic search program resulted in such a high-order interac- tion.) When the residuals from that analysis were examined in a second search, which included some potentially endogenous vari- ables like home ownership, only home ownership and a measure of achievement motivation mattered. Because the crude separate quantification by brackets makes quantitative analysis difficult, and because we have more recent data on home production, we turn to the Pane] Study of Income Dynamics, which in 1968 to 1972 and in 1979 asked about these activities in the previous year: (1) hours spent repairing cars, (2) money saved repairing cars, (3) hours spent working on the house, (4) money saved working on the house, and (5) money saved growing or preserving food. Sewing or mending had been part of the ways people said they saved on clothing costs in 196S, but those questions were dropped because few people reported saving more than $200 on clothes in any way. We also asked in 1968 to 1972 whether the respondent was taking any courses or lessons a steady 11 to 12 percent said yes each year. The proportions of individuals that reported working on cars or houses were remarkably stable a little over a third of the house- holds each year except for a substantial increase in reported work on the house for 1971 in relation to the four years before or the report seven years later. Because more recent years have seen increases in both unem
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 81 ployment and inflation to pressure people toward more produc- tive use of their time, we compare 1972 with 1979 estimates of such activities in 1971 and 1978. Able 2 shows modest increases in the proportions that report working on their car or growing food, and a decrease, perhaps toward normalcy, in reports of working on the house. For those who like to compare cohorts we give the 1978 estimates both by age in early 1979 and by age in early 1972. More important, the table shows that it is largely the young who work on cars, the middle-aged who work on their homes, and the older people who grow food. The age patterns are similar when we ask for the average amounts saved in the three activities (Table 3), although the effect is partly automatic because the averages include the zeroes of those who reported no activity. Note, however, that the average amount saved by working on the house increased substantially while the proportion reporting any fell, implying fewer but larger projects. For work on cars and on the house we asked both hours and amounts saved so that we could estimate the implied hourly earnings from the activities. Table 4 shows it by age, with people aged 24 to 45 apparently doing things that required more skill, or valuing their time more highly. MarginaTists will notice that this is not the marginal wage on do-it-yourself activities but the average; it does not allow us to mode} decisions about unpaid versus paid work, even after adjusting for income taxes on the latter. Indeed, if we look at the implied earnings per hour work- ing on cars or houses according to the hourly earnings of the respondent, there is almost no correlation. Unpaid do-it-yourself work was also affected by marital status, whether the wife worked for money, home ownership, and the gender of single household heads. There was no apparent substi- tution of time no negative correlation between paid work and unpaid work. (Indeed, there was a positive correlation that disap- peared when we took account of age, sex, marital status, a work- ing wife, and home ownership.) Nor was there any apparent mar- ginal comparison of earnings. There was, in fact a positive simple correlation between paid wage rates and hours of unpaid work, which again withered in the multivariate analysis. Those not in the labor force at all did less unpaid work, too, perhaps the extra free time being more than offset by a lack of opportunities or skills, or by disabilities.
82 be ¢ - o o by ·_I o so o A o o V o ·_. 5- to ¢ Cal U2 o Cal 5- ~o o V ~ ¢ o ~ v rim oo ~ ¢ ~ Cal Cal ~ ~ CC ~ CO m¢ m¢ ~ ¢ CD U: ~ 0 ~ o Cal Cal Cal ~ ~ Cal CO Cal N Cam Cry CO ~ cO Cal ~ 00 U) Cal ~ O CC U: Cal ~ ~4 ~ Cal U: ~14 CD U: ~ C~ CS) U: C~ ~ ~ ~ C~ ~ CC ~ ¢ C~ ~ U: L~ CO C~ C~ ~ ¢ C~ U: C~ . ~ ~ ~ ~ C~ ~ ¢ U: ~ U: ~ C~ C~ 0;~ ¢ bD ¢ O ~ O CO U: ~ ~ C~ _~ C~ CO y CO I + C~ U: ~ ~ U: ~ ~ - C~ C~ ~ U: CO ~ ¢ .s CQ CO .o .> . - Ct ·_e c--o. Ct
83 Cal ._, US ·_. - To Go of Us so < - o, _' En 5: v be ¢ o ¢ . ~ of a) CQ to, 1 i_ o ._ ~ Cal a) ~ Cal · - ~Q U] En . a) C<5 -~ _ to Cal ~ CO ¢ ~ E" Ct oo "o JO ~ CQ o V So x Go rid "o =¢ == ~ ¢ =¢ ¢ ¢ ¢ ¢ ¢ == ~ ¢ ¢ ~ oo ~ ~ C~ C~ ~ C~ C~ ~ ~ CD o ~ C9 ~g c~ u: r~ 0 ~ ~ ~ c~ c~ ~ r~ 0 0 ~ ~ cg u7 ~ ~ 0 ~ ~ CO ~ CS ~ C~ 69 r~ ~ oo 0 c~ ~ c~ r~ ~ 0 ~ ~ c~ C~ U: ~ C~ c~ ~ ~ c~ c~ ~ r~ L~ ~ ~ ~ ~1 ~ ~ C~ N ~ O7 e~ r~ c~ 0 ~ c~ ~ co c~ C~ ~ ~ O CO ~ C~ ~ c~ ~ C~ ~1 ~C~ 6g ~ ~ ~ ~ O 0 U: ~ ~ CC 69 ~ cg r~ ~ ~ ~ ~ 0 c~ ~ c~ ~ r~ c~ ~g ~ CO ~ ~ C~ ~ ~ ~ 0 ~ oo r~ ~ ~ co U' C~ ~ ~ ~ CO C~ 1 1 1 1 1 1 + _ 0 ~ U: ~ ~ m} U:) - ~ C~ ~ ~ U~ CO ~ ~ cd . - U' C) S" CO . - ._ . ~ .o 3 0 o s. c<5 h_
84 5- =-o ~ r ¢ ~ ·> ~ ~ ED - ~0 Cal Sol CL ~ __ . ... ~ .= ~ ~ ~ ~ ~ ~ ~ ¢ be ~ ~ U: ~ U' U. ~¢ ~ Cal CO ~ U: U) O ~ U: 00 Cal O ~ ~ ~ ~ ~ cat co Rae ~ Cal ~ CS 00 00 ~ ~ O ~ ~ ~ a, ~ cut ~ rim ~ ~ 0 . . . . . . . . co ~ cat ~ ED ~ u: 0 00 ~ ~ Car CO ~ ~ ~ . ~ ~ . . . . rim ~ cat ~ Rae ~ Rae Cal di d4 ~ Cat d4 ~ ~ ~ Cal Cal O CC ~ U: ~ ~ Cal ) It:l en cat - oo c~ c~ o ~ ~ o ~ r~ u~ ~ u~ . . . . . . . . r~ ~ ~ r~ oo u: u: oo oo ~ u: c~ u: . . . . . . . . ~ oO d. ~ 0 u: ~ C~ c~ ~ ~ ~ ~ c~ c9 oo ~ c~ r~ u: oo CD ~ C~ ~ r~ c~ b~ cd - ¢
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 85 Most important of all for our discussion here, these unpaid productive activities drop off with advanced age, in spite of more free time, lower incomes, and greater home ownership. It seems unlikely that this is a cohort effect and that today's young will continue their work activities unabated, although they might be healthier and affected by their unemployment and inflation experiences. We have no information about their accumulation of human capital (skills), an accumulation that might presage more such activity by them than by earlier generations. Indeed, with- out question there is a need to collect such information. PRODUCTIVE ACTIVITIES MEASURED AT THE INDIVIDUAL LEVEL Housework Housework hours for regular cooking, cleaning, and the like are clearly dominated by family composition and, among fami- lies with two adults and not more than one child, by the wife's absorption in paid work. Indeed, relatively few subgroups account for 41 percent of the variance in housework hours. Fig- ure 1 shows the subgroups that do this, though we must remem- ber that a search program found them in a sequential process examining many alternatives at each step; tests and extrapola- tions thus are suspect. Many other analyses indicate that when a wife goes to work, her leisure suffers more than her housework and her husband is unlikely to increase his work contribution around the house appreciably. Simple descriptions of the hours spent in productive activity by gender and age fad] to illuminate the forces associated with gen- der and age that really affect productive effort. A considerable amount of work in labor economics attempts to fit utility maxi- mizing models of choices between market and nonmarket work and between work and leisure. But this work is hampered by the fact that none of our surveys has provided these researchers with information on people's alternative opportunities nor on their marginal wage rates. There is a kind of natural experiment, however, in the changing composition of families if we are will- ing to treat these changes as approximately exogenous in the short run. People do make decisions about who lives with whom
86 JAMES N. MORGAN All 6,620 families in 1981. 1,550 hours | One adult | 737 hours ·1 I One child or none. 647 hours 1,827 cases - Two or more children. 1,585 hours - - One child or none. 1,790 hours - Two or more adults in familY - 2,018 hours , / \ ~ ' 1 Two or more children. 2,546 hours 344 cases L7 ~L7 Wife worked Wife worked less than for pay in half time or 1980,1,000 not at all. hours or more. 1,838 hours 1,386 hours 1,290 cases 958 cases 557 cases ~ 1 Two children. 2,298 hours 954 cases Three or more children. 2,943 hours 690 cases FIGURE 1 The housework hours of families as determined during inter- views conducted in 1981. and about having children and about when the children leave home; but individuals inevitably grow one year older each year and many other family changes are in large part determined by social and legal forces. Able 5 relates the level of or changes in total housework of the family to the number of individuals, or changes in that number, by age and gender (Morgan, forthcoming). In spite of the rather crude method of measuring housework and the fact that differ- ences are subject to large relative errors, all the regression coe~- cients for family composition are highly significant and intui- tively sensible. A baby adds greatly to housework, teenaged sons
87 - Cal Cat o ._ V o - Q en ¢ CQ ¢ U] o o CO o a_ ~ . - o b'0- CtS · _ V ho _ ·_' got ~ so ,~ o In U. by Ct V ho 1 Rae o Do l Rae IV o a, oOo U: >1= . a, Ct v o to ~ ~ ~ ~ o ~ o ~ o CC ~ do d4 cat co C] u ~ ) a) co cat co co Cal ~ Cal Cal ~ U 1 c¢: c~ ~ ~ ~ c,`, oo ° a) co oo ~ CD ~ U 1 C~ ~ C~ 0 U: 00 0 CD 0o C~ U 00 C~ ~ ~ C~ ~ ~ 0 ~ CS oo ~ co ~ c~ ~ co c~ ~ ~ c~ oo 1 ~ ~ ~ C~ 0 CC 0 ~ 0 oo r~ 0 ~ ~ c~ ~ a: cg C~ CC C~ oo C~ C~ oo oo ~ oo ~ CS) C ~ C~ c~ ~ ~ ~ r~ c9 ~ C~ r~ U: C~ ~ ~ ~ ~ C~ CD C~ 0 ~ ~ C~ E ,` E & ~r E x E~ I o E v V v ~ ~ 1Q oo ~ ~ ~ 3 Ct C) . · o z 11
88 JAMES N. MORGAN add less than teenaged daughters, and other adults add most of all. Since we have some families that add or lose a spouse, much of the change in "others" is marriage or divorce. If we include the change in the wife's paid work hours as a pseudoexplanatory variable, dodging the simultaneity of deci- sions about paid work and housework, there would appear to be a small short-run dynamic effect. There is no significant effect for longer lags, but a large significant cross-section or long-term effect does occur, which may well reflect interpersonal differences in preferences or skills rather than individual marginal choices. The effects of a change in family composition on the wife's paid work is much less clear. There are practically no dynamic effects in this case except for changes in "others," which is partially an artifact of the circumstance that an increase or decrease in "oth- ers" often means a spouse and without a spouse there cannot be a wife to work for money. A change in the housework hours has no effect on the wife's paid work hours in the dynamic data, leaving the suspicion that the static negative correlation between house- work and the wife's paid work hours is spurious, resulting from omitted variables. We will turn now to some other data on individuals. In our examination we will identify their own hours of productive work, again focusing on age-gender patterns without going much beyond marital status, and investigate other variables such as home ownership, job seniority, or family size, which might be affecting the results. Lifetime Patterns of Family Situations for Individuals The activities of individuals over their life course are affected by their family situations, particularly whether they live with a spouse and whether there are children under 18 at home. For the 1981 wave of the Panel Study sample individuals, Table 6 shows their distribution within each age group according to family situ- ation. One can estimate the probability for a man or woman of being a household head or wife, and then the probability of living with a spouse, followed by the probability of having children at home. The probability of being neither household head (house- holder) nor wife is of course 100 percent for young children, remains appreciable for the late teens and early twenties, and then only rises again for those 75 or older.
89 8 ·> it; ~2 rid Dot ·_' be ¢ I I I d Ct Sol do CD ¢ U: rid rid CO CO or C9 Car l U. l U: CO Cal Cal 00 rid l O ~1 CO O ~ ~ Cal ~ CO Cal O N ~ an ~ Cal ~Cal ~ of ~ ~ ~ O ~ O ~ Cal Cal O Cal O CO CO ~ ~ ~ ~ O ~ CO ~ ~ 0 ~ co 0 Cal Cal ~ rid ~ ~ Cal 0 0 Cal ~ CC Cal O Cal _ CC Cal ~ up ~ ~ ~ Cal ~ Cal an Cal CO ~ Cal CO CO ~ ~ ~ ~ O ~ ~ ~ C~ tD d~ C~ C~ ~ a co c~ ~ ~ c~ ~ ~ co ~ c~ r" ~ u: ~ u: ~ ~ u~ C~ u: ~ ~ ~ O C~ ~ ~ ~ ~ ~ c~ ~ o ~ ~ r" o o o ~ c~ c~ ~ oo u: ~ ~ ~ ~ o o c~ ~ ~ c~ ~ ~ ~ ~ c~ ~ o ~ O ~, 1 ~ ~ ~ c~ cO ~ o ~ c~ c~ cD ~ c~ o ~ c~ ~ c~ ~ ~ ~ cO o co ~ c~ co ~ r~ ~ c~ ~ r~ r~ ~ ~ oo O O ~ ~ co ~ C~ oO di ~ CO a) co ~ : o - cO oo c~ - o o o o o o o o o o o o o o o o o ~ u: cq ~ ~ s" ~ ~ C ~ ~ ~ ~ ~ co ~ ~ ~ ~ ~ ~ ~
90 JAMES N. MORGAN There are more precise demographic data from the U.S. Census Bureau, but we present these partly to reassure the reader that the Pane} Study sample still reasonably represents the country as a whole. The patterns move smoothly through the life course. Multivariate analyses of housework hours, for example, would take account of family status, age and numbers of children, and age that is a proxy for several of these situational variables. On the other hand, the resulting patterns are strong and systematic by age without getting into the process details. Housework The major unpaid productive activity in most households is housework, which is usually combined with child care. Only in one year of the Panel Study (1976) when we interviewed the wives, too, did we ask separately about child care. It is useful to compare husbands' reports from other years on housework done by their wives and themselves with the 1976 reports by wives on their housework and child care and their husbands' child care. (Even in that year husbands were asked about their own house- work time.* Analysis of the effects of family composition on total family housework hours revealed systematic and meaningful effects of individuals by age and gender, whether estimated from a single cross-section or by using the panel to relate changes in family composition to changes in housework. Table 7 shows the age patterns of housework-child care hours of individuals by gender, containing some of these family- composition effects. Table ~ uses the 1976 interviews in which the wife reported on the husband's child care, but he reported his own housework, perhaps with some duplication. Adding child care explicitly increases the estimates of unpaid work for mar- ried men and women. *Note that housework was elicited by three questions asked of the husband, or two (the second and third) asked of single heads: (1) About how much time does your wife spend on housework in an average week? I mean time spent cooking, cleaning, and doing other work around the house. (2) About how much time do you (head) spend on housework in an average week? I mean time spent cooking, cleaning, and doing other work around the house. (3) Does anyone else here in the household help with the housework? This final question was asked of each respondent: About how much time does he/she spend on housework in an average week?
91 o ¢ A ·_1 ~ - o U2 v) cd ·~. cd s-^ ¢ CQ o 1 o CQ o ~ A) ¢ u) rot ·£ as. _§ £ V) - v ¢ . v - _ A.) Go o ~ o CS oO d4 ~ Cat O Cat ~ Cat ~ ~ Cal CC Cat Cat ~ ~ Go rut 0 CO ~ Cal CO ~ CC ~ ~ rat cat CO ~ Cal ~ o Go ~ U: CS) ~ Cal ~ U: C9 - ^ ~ - ^ ~ 1 0^ ~ c~ ~ ~ c~ ~ c~ ~ ~ oo o c~ ~ co ~ cs' ~ oo ~ r" C~ ~ ~ ~ ~ ~ d4 ~ C~ ~ r~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ c~ - ~ ~ - ~ ~ ~ o ~ ~ ~ ~ ~ ~ ~ ~ c~ cx) u: co ~ ~ c~ ~ ~ ~ ~ cC o o ~ c~ ~ oo o c~ o o ~ ~ ~ ~ ~ - ~ ~ ~ ~ - - - ~ ~ ~ ~ ~ o ~ ~ c~ oo o 0 ~ ~ co r~ co co r~ r~ oo ~ ~ ~ a, ~ a, ~ ~ ca ~ oo ~ ~ ~ ~ ~ ~ c~ co c~ ~ 0 ~ 0 ~ ~ r~ c~ O C~ co ~ co ~ ~ ~ ~ ~ ~ ~ - ^ ec s~ ~s: =- .- t ~ - 5 V ¢ bD ._ ~Q o ~ o ~ ~ CO ~ ~ CO o oo CC ~ o ~ U: C~ ~ CO ~ ~ CO C~ C~ CO CO ~ ~ C~ ~ ~ CO U: o U. ~ o ~ oo CO ~ C~ CC ~ o C~ ~ ~ o ~ ~ oo ~ C~ C~ ~ ~ CO ~ U: ~ ~ ~ U: C~ "o ~ _ _% 1= (53 CS oo ~ C~ ~ o ~ N ~i4 ~ 00 C- C~ ~ ~ /3, o ~ u' ~ ~ ~ ~ ~ C~ ~ bD .- o ~ ~ ~ ~ ~ ~ ~ ~ + ~ ~ a) C~ I I .1 uL ~ ~ o ~- ¢ ~Vl C~ Co ~, u: ca to r~ ~ ¢ U' ·q) :^ CQ C) P. ~ s~ aS ,= 3 ] 11 3=
92 JAMES N. MORGAN TABLE 8 Annual Hours of Housework Plus Child Care, From 1976 Interviews Housework Men Women Age in 1976 Single Married All SingleMarriedaAll 18-24 426 624 554 3852,1111,532 25-34 377 761 695 5272,4612,045 35-44 446 690 668 8052,0591,807 45-54 558 427 436 8621,6201,462 55-61 677 282 308 8981,4261,244 62-64 471 398 405 8991,1681,066 65-69 891 350 406 8291,1941,018 70-74 709 471 501 9621,062109 75 + 690 462 534 893882891 Averages 484 559 548 7411,8991,549 aWives reported for themselves. Questions referred to hours per week at the time of the interview. Total Productive Use of Time ng: We can define productive use of time (or offer of time) as includ · paid employment or self-employment; · commuting time; · strike or unemployment time; · time lost from work by illness of self or others; · housework and child care (including time given in return for child care); · volunteer work for organizations (church, charity) · time spent helping friends and relatives; · unpaid productive work on house, car, food, and clothing (sewing); and ~ taking courses or lessons (investment in human capital). In addition, if people report that they wanted more (or less) work than they could get last year, we might infer that their "supply" of labor was actually greater. (If they were currently unemployed, we did not ask the question, but in this analysis we treat these people as though the answer was yes.) The question was only asked of heads of households, not wives, except in 1976.
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 93 Substitutions of Paid and Unpaid Productive Time Economic models focus on marginal choices, leading to the notion that when paid market work becomes less available, unpaid productive activities might flourish. The response to eco- nomic incentives at the margin, however, may well be over- whelmed by constraints and pressures. And we have already seen that volunteer work is largely done by young overworked people with children and that most older people do not report doing more volunteer work after they retired. Housework hours reported do tend to rise a little, particularly for men in their 60s. Hence we Took at housework, paid work, and the total supply of productive hours, in which we include commuting time and hours lost from work because of strikes, unemployment, or ill- ness of self or others. From the individual's point of view that total is what it takes to earn his standard of living, even though society may only see the benefit of the hours that are paid for plus the housework. The 1981 reports on housework currently available, and reports on the other work times for the prior year, show the expected patterns, with the fraction of the total hours that are paid for declining after age 40 for men and from the beginning for women (Table 9~. Particularly for women, of course, there may well be a historic trend or generation effect added to the age effect as succeeding generations of women have fewer children and work more for money even when there are children at home. Because married women's hours are reported by their hus- bands in most years of the Panel, we also provide some estimates from the 1976 interviews in which wives were also interviewed and separate questions were asked about child care. Here there may still be some bias in favor of husbands, who reported their own housework, perhaps double-counting some of the time their wives credited them for as doing child care. The patterns, how- ever, are much the same (Table 101. Individual Change Admittedly, we have been using the Panel Study of Income Dynamics in an inefficient way as a series of cross-sections pro- viding different information in different years. Life course
94 ¢ · ~ so B Cal o U] o 1 3 o C;l ·- C) ~ F° o ~ o 3 3 ·C~ Cal o he ~ C<, ·Ce a, ~ F 3 3 3 3 . Go ._ be ¢ Go ~ cot cat co O ret ~ ~ ~ ~ ~ Cat ~ ~ ~ Cat CO ~ ~ CD ~ o o ~ ~ o ~ o ~ ~ o US ~ ~ ~ ~ CO Cat CO CC U: Cat ~ ~ ~ ~ Cat CO Cal ~ ~ ~ Cat Go o ~ o CO ~ CO o Do ~ ~ Cal o Cat ~ Cal ep N CO ~ ~ cO N C~ ~ C~ c~ o c~ ~ o ~ oo ~ o cD N It: ~ ~ 0) It: ~4 d. CKl 1 - O C~ C~ C~ ~ 00 cs C~ co e~ ~ O c~ r" ~ ~ ~ ~ ~ c~ c~ ~q 00 S~ O C~ cn C~ O U' cn _ _ O _ CO .~ o - C~ o CQ C) . - o B ._ P4 o ca o _ . '~ s.4 . _ 50 o 3 ~ J:2 s~ 5- o ( ._ ~ ~ s" + 6 aY B 0 ~ ~ c~ ~ ~ co r~ co r~ co c~ ~ 0 c~ u: c~ u: c~ ~ r~ oo oo C~ ~ oo o o ~ ~ ~ ~ C~ C~ CO CO ~ ~ ~ ~ ~ U: U: ~ ~ oo o ~ o ~ ~ ~ CS o CC ~ CC CC ~ UD U: CO U: ~ ~ o oo ~ o ~ CO ~ t~ c~ ~ ~ u: c~ r~ O C~ O C~ _ _ _ _ _ _ _ ~ C~ d. ~ I ~ CO CO 00 ~ U: U: ~ C~ ~ ~ C~ ~ ~ CO C~
95 5- -o. - C: . _, Cal 5~^ a) ¢ ~ ~ 3 CQ . - ~, - C~ ,~,5 ~ to , sol be .= U2 _m ~Ct ~ .~ a, .= On be so ._ ~ Cd PI ~ I, o ~ Ed ~ ~ CQ ._ Ct o C: Cal so ._ ~ Ct A _ CQ ._ Ct o Pa ~ Ct C) so ._ - P. U) U) ._ Ce o be Cal C) 5- ·_ ~ Ct PI ~ CO _ - % rid en up ~ cat ~ ~ 0 cut ~ ~ An cO Ha co At ~ -~ - ~ ~ ~ At At ~ ~ By ~ ~ At ~ U: 0 CO CD 0 0 U) ~ O CC CS ~ CD Cal ~ ~ ~ 0 ~ ~ ~ ~ At At ~ ~ ~ At - ~ Go cC ~ At Go Ha ~ ~ ~ co cc ~ At ~ 'c ~ 0 0 ~ to At ~ ~ to 0 ~ At ~ to rid rid ~ cat 0 ~ CD Cal 0 ~ CO CD CO ~ ~ ~ Cal ~ UP 0 0 0 rid 0 cut rid co ~ rid At ~ ~ to ~ ~ ~ ~ ~ to co lo ~ ~ ca At ~ ~ ca At ~ At ~ ~ ~ ~ Go ~ ~ To ~ ~ O CO 0 ~ 0 CD ~ ~ At lo co ~ ~ co At ~ At 0 ~ ~ 0 rid ~ ~ 0 0 cat ~ ~ 0 Go 0 0 ~ co cat 0 lo ~ ~ 0 0 ~ cs ~ ~ As lo ~ ~ At ~ ~ ~ ~ ~ cO 0 cat cat 0 cc 0 cat ~ ~ cO At ~ At cs O ~ ~ ~ Cat ~ ~ ~ rid ~ cO 0 O ~ ~ ~ 0 ~ ~ CC ~ UP ~ ~ CD ~ CO Cal ~ Cal ~ Cal ~Cal ~ _ ~ ~ ~ CD ~ ~ ~ ~ Cal 0 0 0 rid ~ ~ ~ co 0 CS ~ ~ ~ U: Cal CO U ~ ~ Cat ~ ~ rid 0 can cat 0^ At at ~ cat ~ ~ ~ ~ Cal so CS 0 ~ 0 ~ ~ 0 U: cat ~ rid ~ cat ~ cat co cat ~ ct O O ~ Cat ~ CD ~ ~ Cat -0 _ ^^^^^^__ ^ At; Cal Cal ~ ~ ~ ~ ~ CQ - CC) Cal Cat ~ U) CS CO CC ~ + ='5, 0 ~ I ~ ~ ~ ~ 0 ~ ~ ¢ ~ Cal CO ~ ~ CS CD ~ ~ ¢ U' 3 Cal 3 ._ ~ to BY . o 3 ~ ~ o o sit ·o 8 ~ CQ o a 0 c: co ~ ~ lo, ~ ,~ cot 5- c' 0 ~ 3 _ ~ ._ ~ sol UO ~ o U) ,~ =._ CQ Ct o ~ As == of A en, 5- o 3 ~ sit U) o I= o U) 3 ~ .o ~ lo e.) . ~ CO &, ._ ~ 3 u: _ ~- ~ CQ Cal 3 11
96 JAMES N. MORGAN changes of individuals can in fact be studied and analyzed more carefully. One can even pool these changes for example, by cre- ating a record for each person for each three-year period he or she was in the Panel and was either a household head or wife. Taking such a file we produced 66,326 records, but because we omitted the questions on housework in one year, three years worth of triplets must be discarded in looking at housework, which leaves 42,500 records. Able 11 shows the average changes for men and women in their paid work and housework, and for comparison the average level of hours for the year between the two measur- ing the change. It is clear that paid work increases for only a few years, then falls off at an increasing rate, which is dramatized by our selection of age ranges. Housework, on the other hand, TABLE 11 Mean Two-Year Change in Paid Hours and in Housework, by Age and Gender, Compared With Mid-Year Levels (for 42,500 Records of Sample Household Heads or Wives for Each Available Three-Year Period) Paid Hours Housework Age in Middle-Change FromMiddle-Change From Mid-Year Year LevelYear 1 to Year 3Year LevelYear 1 to Year 3 Men 18-29 2,13510286342 30-39 2,30951,023- 36 40-49 2,223-251,109-50 50-54 2,058-72923-48 55-59 1,925- 123858-21 60-61 1,677-27788433 62 1,557-34290428 63-64 1,350-566999-85 65 792-62592846 66 + 412- 105864-47 Women 18-29 970251,38952 30-39 847661,593-75 40-49 914481,506-81 50-54 897- 31,39897 55-59 837-641,291- 11 60-61 745- 1601,251-43 62 583-2031,330-70 63-64 454- 1881,30517 65 246-2171,320-86 66 + 130-481,251-74
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 97 declines much less sharply and even seems to increase a little around retirement age. It is likely that reports on housework hours are considerably less accurate than the reports on paid work hours because the husband is reporting paid hours and housework hours for both spouses in families containing a mar- ried couple. The mean changes appear to reflect the changes in the means, of course. It is when one wants to study interpersonal differences in these patterns that the Pane! data become advantageous. Our earlier analysis of the effects of change in family composition on housework is an example. Expressed Desires for More or Less Work One reason for looking at the total supply of hours is that a person may have to devote more hours to earning a living than he or she is paid for-that is, the total supply includes unpaid commuting, strike, and sick time. But beyond that there is an additional labor supply we can only call underemployment. We attempt to tap it by asking people whether they could have worked more (or less) last year and if not whether they would have liked more (or less) work. Tables 12 and 13 show a very few at any age reporting wanting less work, even at less pay. More importantly, the tables show substantial proportions, particu- larly among the young, saying they wanted more work than was available. We went back to 1976 because wives were asked about their own desires then, whereas for the 1981 interviews we can only report on husbands and on single male and female house- hold heads. One serendipitous finding is the appearance of a substantial increase between 1976 and 1981 in the proportion of young single women aged 18 to 44 who said they wanted more work. There was a smaller increase among the younger single men. Of course, there was also an increase in unemployment, which hit the young the hardest. Although the question was not asked of the currently unemployed, we assume that they wanted more work. There is a pervasive myth that many people who would rather keep on working are forced to retire, in spite of a variety of studies over many years that show most people retire from paid work as soon as they can afford to, and rarely regret it (Streib, 1971; Parnes, 1970; Morgan, 19801. The elimination of compul
98 be · _' A A a) o CQ CQ o o ·C~ o a) be ·_1 o CQ o U] CO :> o it' ·5- .= U] a) ·50 V) En so C: _ CO . AD ~ Go Cal ~ ~ CS o o o ~ . . . . . . . . . . Cal ~ ~ Cal Cal ~ o o ~ Cal At ~ ~ o Cal ~ ~ o o CO IS) O CO N O O O Hi Cal ~ ~ Cal ~4 lo) O O . . . . . . . . . . Cal ~ ~ Cal ~ ~ ~ O O Cal ~ ~ o ~ ~ ~ . ~ . . . . ~ ~ O CD O ~ o ~ - . co o ~ ~w u: ~ ~ ~ o o ~ o . . . . . . · · · · t-\ co c~ ~ ~ c~ o cO o o o - ' ~ ~ - ^ c~ ~ ~ ~ ~ o o ~ ~ c~ ~ ~ . . . . . . . . . . ~ CC U: ~ O CD ~ C~ ~ C~ ~ ~ cO ~ c~ c~ ~ ~ ~ o ~ ~ o . . . . . . . . . . ~ ~ c~ ~ ~ ~ ~ ~ o o c~ o c': ~ c~ c~ ~ ~c ~- c~ ~ ~ ~ o c~ . . . . . ~ oo ~ ~ ~ ~ ~ c~ c~ ~ a: ~ - `. c~ ~ cq d4 C~ ~ ~ ~ CO CD CO 1 1 1 1 1 1 1 1 + _ a~ u~ ~ ~ ~ c~ ~ 0 ~ - ~ c~ c~ ~ ~ cs~ co r~ r~ ¢ E~ u' o s~ c4-o-` 5- o · - o CQ o · - ct ct ?> Ct CQ ? _ U) ,~ o 3 o o o ~ ~o ? ~ 3 ~ CQ CQ Cd o ~ o 3 CO ~ ?, o = ~R U) C~ U: C~ C: 3 11
¢ ~ - CQ .> - ~ To ·_1 a) 5- o o U2 A 5- c~o4 ·_I CQ en · - 3 o A o A ca 3 co a, o A GR o oo 5- yO U) CQ bC ¢ 99 ~ o ~ ~ ~ Go ~ ~ ~ Cal . . . . . . . . . . CO ~ ~ ~ oo Cal ~ o CO Cal Cal ~ A ~ C~ ~ ~ o o o ~ o ~ . . . . . . . . C~ ~ CC U) ~ o o C9 C~ O 00 CD U: ~ ~ O . . . . . . . .. C~ oo ~ ~ ~ ~ o C~ ~C~ C~ C~ ~ ~ ~ ~ o C~ CO C~ ~ CO C~ o CC ~ o C~ CO ~ C~ ~ ~ ~ o C~ C~ ~ ~ ~C~ C53 CO CC U: ~ O O C~ . . . . . . . d. ~ ~ ~ ~ ~ ~ o ~ o ~ ~ c~ ~ ~ o - c~ c~ c~ c~ r~ oo oo ~ 0 ~ ~ 0 c~ ~ ~ ~ c~ u: c~ c~ ~ ~ ~ co co r~ 1 1 1 1 1 1 1 1 + : ~ u: c~ co ~ u: co co r~ r~ u) CQ cd c~ o c~ ~ - ~s - ¢ ~ s~ o 3 5- o cd 3 Ct o U] U) Ct o s~ C) 1
100 - .> BY ~ ho ~ C: o ~ U2 o ¢ ¢ ~ o o ~ ~ A 4 a) - · - 1 ·s ~ . _ _ ~ A · - Cd _ ~ Cat M · - a) of ~ CO it ho ¢ be =2 o ~ A ho ~ _4 ~ if, U) sat - U] _ ¢ ·5- cd bD . - u] bC ¢ ~ C~ ~ O O CO O ~ (D ~ O co u: 0 r~ c~ ~ ~ ~i CS O ~ CC d4 ~ u: C~ _ O o . 0 ~ o ~ c~ o co ~ u) ~ ~ c~ ~ ~ ~ ^ c~ c~ 0 r~ co c~ 0 ~ ~ ~ O . . . . .. . . . . ~ ~ ~ 0 c~0 c~ ~ ~ ~ ~ : u: u) cc D c~ 0 u: . r~ a) CD O ~ ~ ~ ~ U: O cc} ~4 ~14 ce c~ O C~ 0 ~ d4 0 CD ~ 0 ~ CO 0 0 0 0 c ~ c~ 'c ~o 0 c~ o u: ~ co ~ ~ '~ co ~ . -, o co u: oO ~ C~ CO ~4 ~ C9 CC) C~ C~ CO ~ U: CD ~ CD ~ 1 1 1 1 1 1 1 1 + 0 ~ ~ ~ ~ c~ ~ o ~ ~ ~ c~ ~ u: co co r~ u, cd c) c~ o _ ~ ¢ ~ U) 5- o U) ~Q U) Cd C.) 5- C~ C~ ~Q _ · - ct B CO CO .> 11
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 101 very retirement, even when combined with inflationary pres- sures, appears to have had little effect on when people retire, although providing more flexible hours and appropriate working conditions might produce some changes. Capacity IlIness and Disability It is commonly agreed that not only are people living longer- their health also is better. Substantial changes do occur with aging, however. Using the Pane! Study data, ~ include two tables on illness and disability (even though there are much better data on health and disability). Table 14 shows the percentage of peo- ple, by age and gender and marital status, who report disabilities that limit the type or amount of work they can do (as before, husbands report for their wives as well). The dramatic increase in this measure with age raises questions about the productive potential of an older population if no attention is paid to these physical limitations. The much higher reported incidence of dis- ability among unmarried (single, widowed, divorced, or sepa- rated) people says nothing about causal direction and may be partly the result of a reporting bias (more time to focus on one's own problems). The higher incidence may also be explained at least for women by the direct reports made by single women as compared to the husbands' reports for wives. The apparent ten- dency for more women than men to be disabled in the middle years may also reflect that, with more husbands reporting for both, men may understate their own disabilities in relation to those of their wives. It is the general age pattern that concerns us here. Table 15 shows the expected increase in disability and hospi- taTization with age but a flatter pattern of intensive illness (sick bed days). Some Traps in Comparing Age Groups We have all become familiar with the age-period-cohort para- dox: We are aware that what seems to be an effect of age can be a difference between generations and that a change in the same people over time can be the effects of a historic period (changing levels of inflation and unemployment) rather than an effect of growing older. But there are other problems.
102 on ·_1 - ·_1 o En be Cal a' m .O Ct o a .s ·,- o a o - U2 ~ a ~ V O A, =0 .C) U] ~ _ C) ~ ~ co ~ V O Clq O .= .C~ U: C) Ct by ¢ ~ ~ ~ Cal ~ 00 ~ ~ ~ O Cal ~ ~ ~ Cal CO ~ ~ (3~ Cal co Cal ~ r. ~ ~ cat ~ ~ Co Cal Cal ~ US ~ ~ ~ ~ ~ 00 . . . . . . . . . ~ ~ O ~ ~ Cal Cat ~ ~ . ~ ~ ~ rip ~ 0 ~ ~ ~ cat Cat ~ ~ CD CO ~ N ~ O O Cal O . . . . CD ~ O Cat Cal ~ Cat ~ CO O Cat CO di ~ ~ Cal ~ ~ cat co Do ~ 0 rip CD oO ~ ~ ~ ~4 Cal ~ ~ ~ ~ u:' ~ ~ ~ ~ ~ ~ d4 ~ CC 0 co rip ~ ~ ~ ~ co ~ can . . . . . . . . . . O O O ~ ~ Cal ~ did CO d CO O a) CD CO . . . . . . . . . . U: ~ Cal Cat ~ Cal Co ~ ~ U) ~ ~ CO oO ~ . . . . . . . . . . Do tS~ O ~ ~ ~ ~ ~ O Cal Cat ~ ~ CD ~ ~ ~ ~ ~ ~ ~ ~ ~ be Cal ~ ~ ~ CO Can CD 1 1 1 1 1 1 1 1 + 00 U: ~ U: ~ Cal ~ O U. - ~ Cal CO ~ ~ CO CS ~ ~ ¢
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 103 Measuring things, particularly nonmarket or uncompensated items such as the value of unpaid time or the "free rent" on an owned home, becomes more uncertain and potentially biased with older people. For- example, for various tax reasons and because of inertia and emotional attachments, they tend to live in houses that are too large. But to add an imputed net rental value to their income exaggerates their economic well-being. They cannot eat or pay their medical bills with that imputed rent. If we want to value their housework or their hours of volun- teer work, there is no market "opportunity cost" wage to start from, even apart from the selection biases in using that method with younger people, some of whom work. Nor can we trust averages; inequality in income and wealth increases with age, and an average does not tell how many are below some threshold. More importantly, among younger people, those with low incomes or no assets are not likely to stay that way; the older people in those circumstances are likely to remain so. Comparing the fraction of poor in one year is comparing a set composed mostly of persistently poor older people with a set com- posed mostly of temporarily poor younger people. Asset means tests may be discouraging to younger people, but they can at least hope to build some savings later. Measuring time use may also involve problems of comparabil- ity if there are differences in the intensity of work. With older people housework time tends to spread out and take up available hours while younger people may even be doing several things at the same time. The main issue, however, is not measurement and analysis but seeking ways to allow, encourage, and facilitate increased productive activity among the elderly. What Affects OIcler Persons' Productive Activity? We can model the forces affecting the productive activity of older persons in terms of incentives and disincentives or barriers. Five general propositions apply: -1. Multiple incentives are better than solitary ones, and it is essential that all the barriers and most of the disincentives be removed. Both the incentives and the barriers range from the monetary through the more broadly economic to the psychological-sociological (altruism, social roles, inertia, and
104 JAMES N. MORGAN social obligations), and then to physical conditions (health and the physical environment). 2. Changing people's behavior is difficult at best and becomes increasingly so as people get older. A lifetime of habit and becom- ing accustomed to one set of roles and environments does not predispose an individual to venturesome changes, and yet small one-thing-at-a-time changes are unlikely to provide substantial rewards. 3. Ordinary paid work in the usual marketplace is the least promising activity to expect older people to expand. Care for themselves and others is probably the most likely, with various forms of semimarket activity (growing food, sewing, or repairs) probably somewhere in the middle. 4. Money incentives may well be less important, and affective rewards more important, among older people. Both, however, are required. 5. The disparity among individuals in the capacity to produce, relative to the level of needs, becomes greater with age. Some people stay healthy, accumulate assets, and acquire multiple use- ful skills while others become increasingly dependent. For all our concern with incentives, it is crucial to remember that most dif- ferences in productive activity are the involuntary result of life histories, health, and an individual's particular environment. Attempts to increase economic incentives can result in reward- ing the fortunate and punishing the unfortunate. The terms "unpaid" and "voluntary" are imprecise. We need to distinguish market transactions that is, paid labor from (1) productive effort with expectations of some reciprocal bene- fits, (2) productive effort with direct benefits (e.g., home produc- tion and repairs), and (3) productive effort with no expectation or contract for anything in return (pure altruism). Any of these three can be in an individual or two-person context or through an organization such as a church or volunteer group. There are serious difficulties with the notion of increasing the altruistic unpaid work of older people. If there is no reciprocity or return benefit, the burden is likely to be quite unequal because of the unequal capacities and preferences of older people. And there are limits to reciprocal (bilateral) mutual help arrangements, just as there are to any barter arrangement the necessity for each pair to stay in balance severely restricts activity. (For exam
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 105 pie, one report on an experiment in encouraging shared housing found relatively little permanent success; more people wanted to share their homes than wanted to live in someone else's (Prit- chard, 19831. Multilateral trade is unquestionably more complex, requiring agreements, prices, and money (or some substitute), but it vastly increases trade. It seems likely that compatibility and equitable financial arrangements are much more difficult with very small numbers of people than with larger groups. A larger group or community can have more special skills available, can afford to invest its members' time in working out equitable and efficient arrange- ments, and can achieve economies of scale. It can also allow sub- groups of friends to provide social support to each other without insisting that everyone get along swimmingly with everyone else. When we combine the need for multilateral arrangements with the disparity in abilities and needs among individuals, it becomes apparent that we need a combination of savings accu- mulation and insurance to make the system work. Each individ- ual would contribute to a fund that would then pay for the invol- untary needs of some individuals as a form of insurance but would also accumulate a substantial sum. (Younger individuals would pay more, on the average, than they took out knowing that later they, as an age group, would get more paid out than the fund was taking in from them.) Individuals would then receive compensation for helping one another, directly from the helped person if the need was voluntary but from the insurance fund if it was involuntary (e.g., nursing during illness, etc.~. Implicit in all this is an attempt to separate efficiency from equity considerations, to focus not on changing the amount of redistribution or subsidy but on opening up a wider variety of opportunities and alternatives. There is much to be said for dis- tinguishing economically self-sustaining arrangements from subsidies, bribes, and coercions. Indeed, the usual tax-break methods of encouragement are particularly likely to be inequita- ble among the elderly; those kinds of benefits will probably be limited to those with high tax rates. A good test of any program is whether it is viable on its own or whether some funds are needed, such as a temporary development fund to start an "infant indus- try" or a loan rather than a permanent flow of subsidy. For some kinds of new living arrangements to facilitate efficient self- and
106 JAMES N. MORGAN mutual help among the aged, what appears to be needed are some small development funds to assist the first attempts and export the successful mechanism (not rules, but rules for develop- ing rules). Another important method is the use of nonamortized loans; older people may not always need or want to be accumulat- ing further equity in their last years. There is an important difference between a project that covers its own costs, including interest on loans, and one that is subsidized or one that mimics condominiums for the young intent on saving taxes while build · · . long equities. It seems likely that removing all the barriers to more produc- tive activity among the elderly will do a great deal of good. Because many things can inhibit the process or cause changes in the issues at hand, we need coordinated, multifaceted, compre- hensive programs that solve all the problems. Furthermore, because many of the problems will arise in midstream, what we need are not fixed projects but the creation of new viable institu- tions or communities that with the full participation of the mem- bers solve each problem as it comes along. Indeed, the very solv- ing of those problems may be the most productive thing some of our elderly could be doing. WHERE DOES THIS LEAVE US? Are we then back to our "common pleas" for more research and more data? Yes, indeed, we are but as preparation for field trials of new societal arrangements. Even if the scare statistics about the ratio of-total population to employed workers are misleading and exaggerated, we do face a much larger and older retired population and potential further increases in life expectancy. The very economic security and medical insurance that may well have increased that life expectancy is now an object of contro- versy. It seems likely that paid employment will remain scarce and will require younger, better trained, more flexible workers, even though the demands for low-tech products like nursing care and other services grow because we have no efficient "affordable" ways to provide those services. The greatest potential for produc- five activity among older people may well be in self-help and mutual help and in activities formerly performed largely by vol- unteers. This does not mean that we should expect great
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 107 increases in unpaid activity among older people but perhaps that we should facilitate and provide at least some compensation for many of these activities. A first step would be assessment of the "human capital" of skills and experience among the retired and those soon to be retired. They are unlikely to provide good direct forecasts of how they would respond to new opportunities and incentives, but we can at least infer their capabilities. A second step might well be some field trials (a better term than experiments) facilitating the formation of communities that would work out the physical, economic, and insurance-type arrangements to encourage all kinds of productive activities within a framework allowing maximum choice, freedom, rewards, etc. Take the case of the care of the sick and temporarily disabled, or the partially permanently disabled. Currently, in about half the cases this care seems to take place in extremely expensive hospitals and nursing homes where costs are often out of control. Or sometimes this work is performed by a spouse at considerable physical, emotional, and sometimes financial costs (since the insurance schemes do not cover most of it). Extended families only rarely can be expected to help because they tend to be scattered and have their own children to care for. Natural communities do not develop. Indeed, most older people see their network of social support withering as their friends die, and the isolation in single-famiTy homes does not encourage the develop- ment of new social support networks. We know that a mixture of economic incentives, combined with inertia and emotional attachment, makes staying in the family home attractive unless alternatives appear that offer a solution to several problems at once: · protection against outliving one's savings; · protection against being financially devastated by a medical or other emergency; · protection against inflation; development and maintenance of a social support network; · provision of opportunities for increased productive activity in a variety of ways; and · provision of opportunities for social and economic interac- tion in a community but maintaining privacy when desired and control over some indoor and outdoor space.
108 JAMES N. MORGAN No one-at-a-time changes or experiments to isolate the effects of only one thing will do us much good in finding solutions to these problems. We must take some longer leaps, and in that kind of activity, preliminary research will be of little help. SUMMARY We have shown that there is little reason to expect increased productive activity among the elderly under the present system of living arrangements, tax laws, and expectations, particularly the expectation that some work we do must be performed with little or no compensation. Furthermore, attempting to rely on tax incentives or coercions like changing the retirement age threaten to punish the unfortunate or benefit the fortunate instead of changing behavior. Increasing options and reducing barriers may well be important, particularly if new arrange- ments provide many different benefits and solve many different problems all at the same time. It seems unlikely that piecemeal approaches or rigid experiments that do not have built-in adjust- ment mechanisms will do more than prove what cannot work. Further research might reduce the risks of field trials of compre- hensive new arrangements, but for this to succeed, risks will have to be taken. Finally, we suggest that the design of any new plan or system should emphasize economically viable aTterna- tives, separating subsidies and income redistributions from attempts to improve the efficiency of our social and economic arrangements. REFERENCES Barfield, Richard, and James Morgan. 1969. Early Retirement, The Decision and the Experience. Ann Arbor, Mich.: Institute for Social Research. Hill, Martha S. Forthcoming. "Patterns of Time Use." In Time Use, Goods, and Well Being Edited by F. Thomas Juster and Frank Stafford. Morgan, Leslie A. 1983. "Intergenerational Financial Support: Retirement-Age Males, 1972-1975." The Gerontologist 3(April): 160-166. Morgan, James N. 1980. "Retirement in Prospect and Retrospect." In Five Thousand American Families: Patterns of Economic Progress. Vol. 8. Ann Arbor, Mich.: Institute for Social Research. Morgan, James N. 1983. "The Redistribution of Income by Families and Institutions and Family Help Patterns." In Five Thousand American Families: Patterns of Economic Progress. Vol. 10. Edited by Greg J. Duncan. Ann Arbor, Mich.: Institute for Social Research.
UNPAID PRODUCTIVE ACTIVITY OVER THE LIFE COURSE 109 Morgan, James N. Forthcoming (1985). "Effects of Changing Family Composition on Housework and Food Consumption." Journal of Consumer Research. Morgan, James N. Forthcoming. "The Role of Time in the Measurement of Transfers and Well-Being." In Conference on Social Accounting for Transfer Payments. Edited by Marilyn Moon. Cambridge, Mass.: National Bureau of Economic Research. Morgan, James N., Ismail A. Sirageldin, and Nancy Baerwaldt. 1966. Productive Americans. Ann Arbor, Mich.: Institute for Social Research. Morgan, James N., Richard F. Dye, and Judith H. Hybels. 1979. Results from Two National Surveys of Philanthropic Activity. Research Report Series. Ann Arbor, Mich.: Institute for Social Research. Reprint. Commission on Private Philanthropy and Public Needs. 1977. Research Paper Series. Vol. 1. Washington, D.C.: U.S. Treasury Department. Parnes, Herbert S., et al. 1970. "From the Middle to the Later Years: Longitudinal Studies in the Preretirement and Postretirement Experiences of Men." In The Retirement Experience. Columbus, Ohio: Center for Human Resource Research, Ohio State University. Pritchard, David C. 1983. "The Art of Matchmaking: A Close Study of Shared Housing." The Gerontologist 23 (April 1983): 174-179. Streib, Gordon F., and Clement J. Schneider 1971. Retirement in American Society: Impact and Process. Ithaca, N.Y.: Cornell University Press.
