---
schema_version: "library-reader-v1"
source_type: "original_book"
title: "Time for Things Labor, Leisure, and the Rise of Mass Consumption"
author: null
edition: null
publisher: null
year: null
isbn: "9780674250529"
library_document_id: "10f0f318-a2eb-4ea9-8798-d4fb0304f8c3"
source_sha256: "4c8caa621e01543b71e54582edee40922b11dc5ec9e1694e11f876435c81dd6a"
pdf_physical_page_count: 355
pdf_physical_page_index_base: 0
range_start: 70
range_count: 10
html_url: "https://workshop.closeclaw.top/api/ai/read/10f0f318-a2eb-4ea9-8798-d4fb0304f8c3?start=70&limit=10"
markdown_url: "https://workshop.closeclaw.top/api/ai/read/10f0f318-a2eb-4ea9-8798-d4fb0304f8c3.md?start=70&limit=10"
previous_url: "https://workshop.closeclaw.top/api/ai/read/10f0f318-a2eb-4ea9-8798-d4fb0304f8c3.md?start=60&limit=10"
next_url: "https://workshop.closeclaw.top/api/ai/read/10f0f318-a2eb-4ea9-8798-d4fb0304f8c3.md?start=80&limit=10"
---

# Time for Things Labor, Leisure, and the Rise of Mass Consumption

> 页码规则：PDF_PHYSICAL_PAGE_INDEX 是从 0 开始的权威文件物理页索引；PRINTED_PAGE 是页面中识别出的印刷页码，可能为空，二者不得混用。

## PDF_PHYSICAL_PAGE_INDEX=70 | PRINTED_PAGE=62

In a very diferent vein, the failure of work time to diminish could be seen as a consequence of the widespread influence of what social psychologists would describe as an unpopular norm. One of the factors identified by Hochschild, and also by Schor, as responsible for long workdays is the valorization of work efort and the corresponding implicit stigma attached to choosing to work less. Employees feel they must put in more time on the job for fear of being labeled a “shirker” or as “work-shy.” Arguably this is an example of an unpopular norm, upheld not because the majority agrees with it but, more subtly, because each individual thinks that everyone else positively supports the norm. William Macy and his collaborators, in their discussion of this type of norm, illustrate the idea by way of the story of the emperor’s new clothes (Centola, Willer, and Macy 2005; Willer, Kuwabara, and Macy 2009). Each member of the crowd sees that the emperor has no clothes but, because everyone else is behaving as if the emperor is clad in fine garments, opts to behave as if what everyone else seems to be seeing is true so as to fit in with the crowd. For a community to abandon an unpopular norm, it must overcome the challenge of coordinating the communication of true preferences—or, more precisely, of “dis-coordinating” behavioral signals that misrepresent true preferences and normative commitments. If working long hours is a result of an unpopular norm of this sort, then the solution would be for every member of a community to reveal her true preferences at more or less the same time. But concerted collective action of that sort is dificult to coordinate, so once entrenched, unpopular norms can be hard to dislodge. However, while the unpopular norm explanation might give a good account of the dynamic that keeps the suboptimal work-hour norm in place once it has been established, it does not explain how or why the norm became entrenched in the first place. Nor does it account for why the organized collective efort to shift to an alternative norm—a shorter working week—that once animated large parts of American society ebbed significantly during the twentieth century.31

62

Time for Things

## PDF_PHYSICAL_PAGE_INDEX=71 | PRINTED_PAGE=63

One final possible explanation that should be considered for the failure of work hours to decrease is that changes in life expectancy, and in the organization of the life course, reduced the overall proportion of lives typically spent in full-time work. Increasing life expectancy, in conjunction with the greater possibility for retirement and more time spent in full-time education, meant that as the twentieth century progressed, people typically entered the workforce later and lived for more years after the end of their working lives. Since the proportion of life spent in full-time work progressively contracted over the course of the twentieth century, workers might not have worried so much about the failure of the working week to decrease. Indeed, they might have become increasingly concerned about funding their retirements, and an obvious response would be to maximize income during their working years in order to save as much as possible.

Life expectancy increased substantially over the course of the twentieth century in the United States. In 1900, life expectancy for white males was forty-eight years and fifty-one for women. By the end of the century, it had increased to seventy-five for men and eighty for women. An increase in life expectancy of that magnitude might not have much efect on attitudes to hours at work if working life continued to the grave. However, the twentieth century saw the emergence of retirement as a new phase of life—one made possible by increased productivity. The advent of retirement reduced the number of years spent at work as a proportion of years lived. The combined efect of retirement and increases in life span led, between 1850 and 1990, to a sixfold increase in the average amount of time lived after the end of working life (Costa 1998b). Given the increased likelihood of life after the working years, it would perhaps make sense for workers to favor higher wages over more free time in order to build savings to fund retirement.32

However, while an increase in prospective years of life after retirement probably had some influence on attitudes to wages and work time, it is questionable whether the efect was big enough to explain the quite dramatic shift toward higher wage income, at the expense of more free time, during the working years. As discussed earlier in this chapter, in the United States at least, savings rates have been consistently low over the course of the century. This suggests that pushing for higher income, as opposed to a shorter working week, for the most part was motivated by a preference for immediate consumption, rather than by the need to save for retirement. Moreover, there is, to state the obvious, no guarantee whatsoever that a given worker will survive until retirement or for many years after retirement. And, arguably, the extent to which retirement can be enjoyed is limited by overall health, which tends to decline with age. It is not unreasonable, therefore, to suppose that wage earners would discount the value of postretirement time compared to free time when fit and healthy.33 Economists of a psychological bent, such as Tibor Scitovsky (1992), could respond by arguing that people have a bias in favor of minimizing discomfort over maximizing enjoyment and that this influences the rate at which they discount the well-being of their future selves. Since with age comes greater exposure to discomfort, it arguably makes sense for people to adopt a strategy of forgoing enjoyment when young so that discomfort will be minimized in the future. But even if comfort is the overriding desideratum, there is little reason to think that people will act to secure the comfort of their temporally distant selves rather than maximizing their comfort in the present. As we have seen, the experimental findings of behavioral psychologists indicate that people are, to a significant extent, hyperbolic discounters, attaching much more importance to wellbeing in the moment than to that in even the medium-term future. This suggests that the prudent model of the economic agent is not an accurate one (Ainslie 2001). One could, in a philosophical spirit, even follow Derek Parfit (1984) in questioning whether there are any strictly rational grounds at all for supposing a continuity of identity between present and distantly future selves. The supposition that people have rational, self-interested reasons to arrange their lives to take account of the welfare of their far future selves, Parfit argues, rests on a questionable metaphysics of identity, according to which there is an essence that connects the identity of the person across the life course. Why should people sacrifice current satisfaction for future selves given that continuity of identity rests on metaphysically shaky foundations?34

Empirical Pattern in the United States

63

## PDF_PHYSICAL_PAGE_INDEX=72 | PRINTED_PAGE=64

The most powerful objection to the theory that income is preferred to free time because of a concern to fund retirement is, however, that one of its crucial empirical entailments—the prediction that workers will save at high rates when at the height of their earning potential in order to finance their postretirement lives—fails to hold true. For the preference for income over more free time during the working years to be convincingly explained as preparation for retirement, we would have to demonstrate that the life-cycle hypothesis, developed by the economist Franco Modigliani (1970) (and Milton Friedman’s [1957] related permanent income hypothesis), is supported by the empirical evidence. The life-cycle hypothesis posits that people aim to generate a constant amount of utility, through their consumption, over the entire course of their lives. According to the theory, because the marginal utility of consumption decreases, individuals, rather than spending all the money they have at any given point in time, will seek to distribute their spending over time such that it maximizes utility over their entire lives, with the result that consumption is “smoothed” across the life course. The theory predicts that people will go into debt when young and poor, save when afluent and middle aged (in order to finance their retirement), and “dissave” when old, spending down their assets. So it goes in theory, but the empirical support for the life-cycle hypothesis is not compelling.35 Economist Angus Deaton (1992), in an overview of the empirical literature on the topic, concludes that the evidence, on balance, does not support the life-cycle hypothesis. According to Deaton, “it is certainly the case that the household data typically show no clear evidence of asset decumulation among the old.” Moreover, he adds, “there are other problems earlier in the life-cycle, and although there is often some hump saving in late middle age, there is also typically a very close articulation of consumption and income over the whole of life. Survey data also show that many households possess little or no wealth, certainly not in financial assets, and many individuals have little beyond social security after they retire.” Deaton concludes that “the fall in the saving rate in the US is not attributable to a redistribution of aggregate income towards the old, but comes from an increase in expenditures, particularly expenditures on durable goods, across all cohorts, particularly prime-age cohorts” (p. 217 [emphasis added]).36 Deaton’s overall assessment accords with that of other economists. Paul Courant, Edward Gamlich, and John Laitner (1986, pp. 278–79), for example, note that “for all its elegance and rationality the life cycle hypothesis has not tested out well.” Moreover, if Modigliani’s life-cycle hypothesis was correct then, as the length of time lived after retirement increased (as it did consistently during the twentieth century), there should have been a commensurate increase in the rate of saving. Yet household savings in the United States did not increase over the course of the twentieth century, despite rising income, increasing life expectancy, and the growing prospect of life after work (Maddison 1992).

64

Time for Things

## PDF_PHYSICAL_PAGE_INDEX=73 | PRINTED_PAGE=65

Empirical Pattern in the United States

65

## PDF_PHYSICAL_PAGE_INDEX=74 | PRINTED_PAGE=66

In light of the failure of the life-cycle and permanent-income hypotheses to hold empirically, as well as evidence that people tend to discount the future hyperbolically, the argument that increases in life expectancy, in conjunction with the advent of retirement, account for the stagnation of the length of the working week is less than convincing.

At the other end of the life course from retirement, there was an increase, during the twentieth century, in the number of years spent in full-time education prior to entering the work force. In 1910, only 10 percent of the population completed high school. By 1940, that number had leapt to over 50 percent (Goldin 1998, 1999). High school graduation rates continued to climb until around 1960, after which they hovered at around 75 percent, declining slightly through to the end of the century. Most of the increase in the rate of graduation then took place between 1910 and 1960. Enrollment in institutions of higher education also steadily increased, from around 3 percent of the population in the first decade of the century, to 56.6 percent in 1995.37 If increasing productivity was in part used to delay the age at which work begins, then, although it might appear as if working time stayed constant, in fact it was decreasing, as some proportion of increasing output was used to increase the amount of time spent in education. Over the course of the century, individuals entered the workforce progressively later, and this reduced the total number of hours worked over a lifetime.

This trend would go some way to explaining the pattern of work time if full-time education can be characterized as free time. However, that seems questionable, given that school is compulsory (at least until the later teen years) and that to a large extent the telos of schooling is work. From the point of view of the individual, time spent in full-time education might therefore seem more like a backward extension of working life than a mirror image of retirement at the other end of the life course. In the nineteenth century, workers typically began their working lives as apprentices, training “on-thejob.” It makes some sense to think of the time young people spend in fulltime education as somewhat analogous to an apprenticeship, as modern education is to a large extent about training for work. It would certainly be odd to categorize increasing time in education in response to the demands of the labor market as leisure time. This is not to suggest that becoming educated is not, for many individuals, an end in itself. But the main reason for the increasing time spent in full-time education prior to entering the workforce was an increase in the credentials required for work. In that sense, time spent in education should be thought of as continuous with working time rather than as a form of free time.

66

Time for Things

## PDF_PHYSICAL_PAGE_INDEX=75 | PRINTED_PAGE=67

4. Spending and Saving

As we have seen, the history of the United States over the past century or so shows that increases in productivity have predominantly been harnessed to increase output, with some portion captured by labor, through increasing real wages. Up until the middle part of the century, increasing productivity was accompanied by a decline in the length of the working week. However, once the eight-hour day had become the norm, work time began to plateau, even as in the postwar period productivity and real hourly wages continued to rise. But what was done with this increasing wage income? It is quite hard to get a clear picture of past consumption practices, but some general trends are apparent, some of which are prima facie quite puzzling. Aggregate levels of consumer spending, as opposed to saving, are consistently high, especially in the second half of the century. Empirical analysis has shown that, not just in the United States but across capitalist economies, consumption increases proportionately with income (Thaler 1990). This goes against the view that increasing income should lead to a higher proportion of income saved (the increasing marginal propensity to save).

Bureau of Labor Statistics data show that between 1900 and 1950, average household income and consumer spending track one another almost exactly. Spending begins to lag somewhat behind income between 1950 and 2000, but it does so to a surprisingly small extent. Lizabeth Cohen (2003) notes that private consumption remained steady as a proportion of GDP, accounting for at least two-thirds of the national product. Spending in the latter half of the century absorbs a roughly constant (albeit inconsistently so—sometimes it is higher and sometimes lower) rather than a declining proportion of income (National Bureau of Labor Statistics Report 2006). In fact, with the exception of the Depression of the 1930s, levels of consumer spending remained remarkably high across the course of the century, tending to increase at least in proportion to rising income. In 1993, Americans spent over 90 percent of their incomes on consumption, and this is not untypical for the postwar period (Lebergott 1996, p.  61). In line with this pattern, the average household savings rate in the United States was fairly stable for most of the twentieth century, varying only between 6.5 percent and 8.5 percent (with the exception of a spike in savings between 1973 and 1975), before dropping to historic lows in the 2000s, when it averaged 3.5 percent (Maddison 1992).38

Empirical Pattern in the United States

67

## PDF_PHYSICAL_PAGE_INDEX=76 | PRINTED_PAGE=68

The failure of the rate of savings to increase as a proportion of income as real wages increased runs, as mentioned, in some tension with economic theory.39 J. M. Keynes (1936) long ago posited as a “law of consumption” that as real income increases, spending will also increase, but not in proportion to increasing income, with savings making up the diference between rising real wages and consumer spending. Keynes, like many other economists, thought that the marginal propensity to consume decreases as income increases: “It is obvious that a higher absolute level of income will tend, as a rule, to widen the gap between income and consumption” and therefore “as a rule .  .  . a greater proportion of income [will be] saved as real income increases” (1936, p. 97).40 The reasoning here, which is fairly intuitive, is, once more, that the marginal benefits of consumption decline as income goes up, and so, all things being equal, with higher incomes we should expect a shift from consumer spending to saving.41 Although Keynes does not spell out the point, presumably part of his thinking was that, as real income rises, the benefit of having savings (as a kind of insurance against misfortune) increases relative to that of immediate consumption.

But as we have seen, Keynes’s assumption about the marginal propensity to consume (which simply echoed the received wisdom of his discipline) is not borne out by the pattern of consumption in the United States in the twentieth century. Had increasing income been used to increase savings, then the failure of work hours to fall substantially as hourly wages rose would perhaps make more sense. Household savings ofer some protection against unpredictable events, such as illness or unemployment. And as discussed in the previous section, funneling more income into savings as earnings increase might be expected against the background of the quite substantial increase in life expectancy over the course of the century, in conjunction with the emergence of the institution of retirement. Yet household saving in the United States was oddly unresponsive to the combination of increasing real wages during the working years and increased need for savings to fund retirement.

Moreover, not only did household savings rates not increase with rising real income, but the use of credit, the antithesis of savings, expanded over the course of the century. Martha Olney shows how, following World War I, the practice of installment buying spread rapidly across the economy. The beginning of the collapse of inhibitions against borrowing can be traced to the

68

Time for Things

## PDF_PHYSICAL_PAGE_INDEX=77 | PRINTED_PAGE=69

1920s. Consumer credit as a percentage of income doubled between 1920 and 1930, from 4.5 percent to 9.5 percent, and the use of credit of various sorts continued to increase for the remainder of the century (Olney 1991). In 1933, according to Columbia sociologist Robert S. Lynd (1933, p. 862), installment purchases amounted to between 12 percent and 15 percent of total retail sales, including 60 percent of furniture sales, 50 percent of electrical household goods, 75 percent of radios sold, and 60 percent of automobile sales. Robert and Helen Lynd found ethnographic evidence of a new attitude to credit in their study of the economic life of “Middletown” (Muncie, Indiana): “People don’t think anything nowadays of borrowing sums of money they’d never have thought of borrowing in the old days,” Lynd was told by one of his informants (Lynd and Lynd 1929). Even during the economic downturn of the 1930s, the Lynds found that the demand for credit continued unabated. They were informed by “a discerning Middletown woman” that “most of the families I know are after the same things today that they were after before the depression, and they’ll get them the same way—on credit” (Lynd and Lynd 1935, p.  203). The expansion of credit continued into the post–World War II period. Consumer credit grew elevenfold between 1945 and 1960, while credit through installment payment plans increased nineteenfold. Moreover, a new form of credit emerged in the late 1940s with the appearance of the first consumer credit cards.42

One of the reasons why credit took of was the failure of savings to increase in line with wages. Despite increasing productivity and wages, the static savings rate meant that for many consumers it was not possible to pay for bigticket items without borrowing money, and so, from the later nineteenth century onward, new institutions developed to provide credit. Of course, this raises the question of why savings did not increase to fund spending on the consumer durables that were becoming increasingly available. One possible answer is that credit is a kind of saving after the fact of a purchase, such that, rather than saving first to fund a purchase, the purchase is made on credit, then money is “saved” to make payments on the debt incurred. Avner Ofer (2006) has suggested that the appeal of this strategy is that it is psychologically easier to make payments on credit owed than it is to save, since failure to do the former results in very visible financial penalties.43 Yet taking on a commitment to “save” via credit payments has definite downsides. The obvious one is the financial risk entailed by making use of credit. While wage earners accounted for only 35 percent of personal bankruptcies in 1935, by 1958 this number had increased to 85 percent, according to the National Retail Credit Association—an increase directly resulting from the expansion of consumer credit (L. Cohen 2003, p. 124). Moreover, given the rate at which new, and supposedly superior, products become available under capitalism, unless the demand for an item is quite urgent, the consumer would arguably do better by saving up for expensive items rather than using credit. By saving instead of using credit, consumers would likely have more choice when they finally get to the point of making a purchase, because, in all probability, new products would have appeared. Of course, the drawback of saving, as opposed to using credit, is that it entails forgoing immediate consumption. But it is not self-evident that avoiding the frustration of delayed gratification outweighs the benefits of saving, especially given the substantial interest typically paid with credit or installment payments.

Empirical Pattern in the United States

69

## PDF_PHYSICAL_PAGE_INDEX=78 | PRINTED_PAGE=70

Another explanation for the rise of credit is that it allowed rates of spending to be maintained as wages fluctuate. Credit can be used to keep spending in line with current standards, despite decreases in income. But credit can also be used in lieu of rising wages to support increasing consumption. The latter use is likely when expectations become fixed not just about current standards of consumption but also about the rate of increase of consumption. For example, the explosion of consumer credit after 1980 was partly in response to a demand for spending to continue to increase even as real wage levels stagnated—people tried to maintain the postwar pattern of continuously increasing consumption even when their purchasing power was frozen. The increasing use of credit since 1980 thus to some degree indicates a general demand for consumption to increase over time, regardless of what’s happening to real wages. This suggests that the widespread expectation, established in the early postwar era, that material living standards should continuously improve, had become firmly entrenched.

But if the increasing use of credit is explained by fixed expectations about ever-increasing consumption, the question arises of why those expectations should have become so inflexible. Why should consumers, faced with stagnating real income, have chosen to use credit to maintain increasing spending levels rather than keeping their spending in line with their incomes? Why did people expose themselves to the risks associated with accumulating unnecessary debt (some debt is of course unavoidable, especially for people living paycheck to paycheck)? The answer is clearly complex, but one possible explanation that fits with the theory developed in this book is that spending is seen as just compensation received in exchange for work. If the rate of increase of real wages falls, but the perceived labor input of waged workers does not, then wage earners might well feel entitled to keep constant the rate at which their consumption increases, even if that means resorting to credit.

70

Time for Things

## PDF_PHYSICAL_PAGE_INDEX=79 | PRINTED_PAGE=71

So, spending increased roughly in proportion to increasing real wages. But what did people spend their wages on? To a significant extent, people spent their disposable income on durable commodities. Based on Department of Commerce data, between 1925 and 1998 the value in 1998 dollars of consumer durables in the US economy increased from \$32 billion to \$2,418.8 billion (this excludes the value of fixed residential assets, which will be discussed later in this chapter)—a startling seventy-five-fold increase. Using a chain-quantity index measure—which better controls for variability in the purchasing power of the dollar and so gives a more reliable estimate of changes in volume of goods—output of consumer durables still increased more than twenty-onefold between 1925 and 1998 (see table in Appendix 1). Between 1925 and 1998 the US population increased by 2.35 times (from 115,000,000 to 270,000,000) so output of consumer durables per capita increased thirty-two-fold, measured in constant dollars, and ninefold according to the chain-quantity index measure. Most of these goods were destined for the domestic market. According to Avner Ofer’s (2006, p. 178) analysis, over the postwar period in the United States, spending on durable consumer goods grew by 1.2 times as much as income, while consumption as a whole grew slightly less than did income (with income elasticity of 0.94). Economic historian Martha Olney (1991) shows that the shift in the economy toward consumer durables can be traced to the 1920s, which witnessed a “consumer durables revolution.” By this Olney means a structural shift in consumer tastes towards consumer durables. Comparing the periods 1898–1916 and 1922–1929, Olney finds that spending on consumer durables went up and savings declined. Spending on durables doubled, from 3.7 percent of disposable income before the First World War to 7.2 percent afterward, while savings nearly halved, from 6.4 percent to 3.8 percent. As Olney (1990, p. 49) notes: “Such a sharp decline in personal saving rate is astounding, particularly since the 1920s were rather prosperous years and we usually expect savings rates to climb, not fall, during periods of prosperity.” As we have seen, a low saving rate, in conjunction with ongoing spending on consumer durables, was to be the pattern for much of the rest of the century.44 It continued, and indeed intensified, in the postwar era. In 1957, for example, a Federal Reserve study of patterns of credit and consumption drew attention to the shift from consumers buying services from companies that owned durable assets, to consumers acquiring durables to provide services (Garon 2012, p. 319).

Empirical Pattern in the United States

71

下一段：[继续读取](https://workshop.closeclaw.top/api/ai/read/10f0f318-a2eb-4ea9-8798-d4fb0304f8c3.md?start=80&limit=10)