The Changing (Dis-)Utility of Work Greg Kaplan, University of Chicago and NBER Sam Schulhofer-Wohl, Federal Reserve Bank of Chicago The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Chicago or the Federal Reserve System.
What is work? Standard macro model: All work is the same. Decide how much to work by trading off wage against disutility of losing leisure time. Adam Smith (1776): The wages of labour vary with the ease or hardship, the cleanliness or dirtiness, the honourableness or dishonourableness of the employment....a journeyman taylor earns less than a journeyman weaver. His work is much easier. A journeyman weaver earns less than a journeyman smith. His work is not always easier, but it is much cleanlier.... Honour makes a great part of the reward of all honourable professions.... Disgrace has the contrary effect.
The shifting distribution of occupations in the U.S. Managerial Professional Specialty Technicians & Related Support Sales Admin Support & Clerical Service Farming/Forestry/Fishing Precision Production Repair Construction & Extractive Operators/Assemblers/Inspectors Transportation & Material Moving men women Source: Census and American Community Survey..2.15.1.05 0.05.1.15.2.25
What we do Question: How has changing occupation distribution changed the aggregate utility/disutility of work? Key findings: Since, work in the U.S. has gotten better on all dimensions for women. But mentally less pleasant for men. Changes concentrated at lower education levels. Scope: We measure only non-pecuniary aspects of market work. Not wages/compensating differentials or non-market work.
Related literature Sectoral change, job polarization, inequality we add a non-wage dimension. Deaths of despair value of a job is more than the wage. Changing value of leisure (video games). Gig economy and flexible jobs. Women s LFP and non-market work. Why do people in rich countries work at all (Keynes)?
Outline Empirical method. Baseline results. Robustness check with alternative empirical strategy.
Empirical method
Strategy Occupation scores: Use recent survey data to score each occupation on 6 dimensions happiness, sadness, pain, tiredness, stress, meaningfulness. Aggregate changes over time: Compute average of occupation scores across workers, weighting by distribution of occupations in the year of interest. Example: Aggregate stress of work in is average of occupations present-day stress scores, weighted by distribution of occupations. Key assumption: Feelings about an occupation today are the same as feelings about that occupation in the past.
Occupation codes Need to measure distribution of occupations in uniform manner from to present. Use OCC1990 occupation code from IPUMS. Maps census occupation codes in other years to 1990 years. Aggregate to 12 broad categories from 389 in data. Some detailed occupations too small to estimate feelings precisely. Changing detail in original census codes means the 389 detailed occupations aren t coded uniformly over time. Some details: Exclude military. Measure occupation distributions in decennial census ( 2000) and 2011- American Community Survey.
The shifting distribution of occupations in the U.S. Managerial Professional Specialty Technicians & Related Support Sales Admin Support & Clerical Service Farming/Forestry/Fishing Precision Production Repair Construction & Extractive Operators/Assemblers/Inspectors Transportation & Material Moving men women Source: Census and American Community Survey..2.15.1.05 0.05.1.15.2.25
Measuring feelings about occupations in the American Time Use Survey The survey: ATUS asks respondents how they spent each minute of the previous day. Feelings: In 2010, 2012 and 2013, ATUS also asked respondents to describe their feelings for three randomly chosen activities during the day. From 0 to 6, where a 0 means you were not stressed at all and a 6 means you were very stressed, how stressed did you feel during this time? We use only responses for activity of working on main job (don t observe occupation for other jobs).
Feelings data 0.8 0.6 0.8 0.6 0.8 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 <0 0 1 2 3 4 5 6 >6 Happiness scale <0 0 1 2 3 4 5 6 >6 Sadness scale <0 0 1 2 3 4 5 6 >6 Stress scale 0.8 0.8 0.8 0.6 0.6 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 <0 0 1 2 3 4 5 6 >6 Tiredness scale <0 0 1 2 3 4 5 6 >6 Pain scale <0 0 1 2 3 4 5 6 >6 Meaningfulness scale raw data residuals Source: American Time Use Survey observations where activity is main job. N=6,061.
Work vs. the rest of life 0.8 0.8 0.8 0.6 0.6 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 0 1 2 3 4 5 6 Happiness scale 0 1 2 3 4 5 6 Sadness scale 0 1 2 3 4 5 6 Stress scale 0.8 0.8 0.8 0.6 0.6 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 0 1 2 3 4 5 6 Tiredness scale 0 1 2 3 4 5 6 Pain scale 0 1 2 3 4 5 6 Meaningfulness scale work on main job other activities Source: American Time Use Survey. N=6,061 (main job), 56,105 (other activities).
Difference in feelings: main job minus other activities 0.8 0.8 0.8 0.6 0.6 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 6 4 2 0 2 4 6 Happiness scale 6 4 2 0 2 4 6 Sadness scale 6 4 2 0 2 4 6 Stress scale 0.8 0.8 0.8 0.6 0.6 0.6 fraction 0.4 fraction 0.4 fraction 0.4 0.2 0.2 0.2 6 4 2 0 2 4 6 Fatigue scale 6 4 2 0 2 4 6 Pain scale 6 4 2 0 2 4 6 Meaningfulness scale Source: American Time Use Survey. N=5,151 respondents.
Correlation in feelings: main job vs. other activities other activities 6 5 4 3 2 1 0 Happiness scale 0 1 2 3 4 5 6 main job other activities 6 5 4 3 2 1 0 Sadness scale 0 1 2 3 4 5 6 main job other activities 6 5 4 3 2 1 0 Stress scale 0 1 2 3 4 5 6 main job other activities 6 5 4 3 2 1 0 Fatigue scale 0 1 2 3 4 5 6 main job other activities 6 5 4 3 2 1 0 Pain scale 0 1 2 3 4 5 6 main job other activities 6 5 4 3 2 1 0 Meaningfulness scale 0 1 2 3 4 5 6 main job fraction of respondents <.015.015.03.03.045.045.06.05.075.075.09.09.105.105.12 >=.12 Source: American Time Use Survey. N=5,151 respondents.
Adjusting for demographic differences across jobs Regression model: z i = θ 0 + a α a + e γ e + r β r + o δ o + ɛ i z i : respondent i s report of a particular feeling (e.g., stress). α a, γ e, β r : dummies for age, education, race. δ o : occupation dummy. When we disaggregate results by demographics, we run regression separately by demographic group. Adjusted mean feelings for occupation o: z o = ˆθ 0 + a ˆα a s a + e ˆγ e s e + r ˆβ r s r + ˆδ o s a, s e, s r : sample shares at each age/education/race.
Exogeneity assumption Regression assumes error ɛ i is uncorrelated with occupation. Effect of occupation on feelings is causal. If we see high stress levels among managers, we re assuming that management is inherently stressful... not that inherently stressed-out people, who would report high stress in any job, happen to be managers. Robustness check will use fixed effects to (imperfectly) investigate this issue.
Baseline results
Changes in aggregate feelings at work, -present 4.00 happy 0.80 sad 2.45 stress 3.95 0.75 2.40 3.90 0.70 2.35 3.85 0.65 2.30 3.80 0.60 2.25 2.55 tired 1.05 pain 4.50 meaning 2.50 1.00 4.45 2.45 0.95 4.40 2.40 0.90 4.35 2.35 0.85 4.30
Changes in aggregate feelings at work by sex happy 4.00 3.95 3.90 3.85 3.80 tired 2.75 2.65 2.55 2.45 2.35 2.25 sad 0.85 0.80 0.75 0.70 0.65 pain 1.20 1.10 1.00 0.90 0.80 stress 2.65 2.55 2.45 2.35 2.25 2.15 meaning 4.50 4.45 4.40 4.35 4.30 4.25 women men
Sources of the change (happiness) 5.0 women Repair 5.0 men occupation happy score 4.5 4.0 3.5 Transportation Professional Farm Service Managerial Admin/Clerical Sales occupation happy score Construction/Extractive Precision Production Operator/Assembler Technician 4.5 4.0 3.5 Operator/Assembler Construction/Extractive Precision Production Sales Farm Service Managerial Repair Professional Transportation Admin/Clerical Technician 3.0.15.1.05 0.05.1 change in share of workers, to 3.0.1.05 0.05.1 change in share of workers, to Area of circle is proportional to occupation s share of workers in.
Sources of the change (pain) 3.0 women 3.0 men occupation pain score 2.5 2.0 1.5 1.0 0.5 Operator/Assembler Precision Production Farm Transportation Service Technician Construction/Extractive Sales Professional Admin/Clerical Managerial occupation pain score 2.5 2.0 1.5 1.0 Farm Transportation Precision Production 0.5 Construction/Extractive Repair Service Admin/Clerical Operator/Assembler Technician Sales Managerial Professional.15.1.05 0.05.1 change in share of workers, to.1.05 0.05.1 change in share of workers, to Area of circle is proportional to occupation s share of workers in.
Sources of the change (meaningfulness) occupation meaning score 6.0 5.5 5.0 4.5 4.0 3.5 Admin/Clerical women Farm Construction/Extractive Repair Service Professional Sales Technician Transportation Precision Production Operator/Assembler Managerial.15.1.05 0.05.1 change in share of workers, to occupation meaning score 6.0 5.5 5.0 4.5 4.0 3.5 Farm Construction/Extractive Professional Operator/Assembler Precision Production Sales Technician Managerial Repair Service Transportation men Admin/Clerical.1.05 0.05.1 change in share of workers, to Area of circle is proportional to occupation s share of workers in.
Sources of the change (stress) occupation stress score 3.0 2.5 2.0 1.5 1.0 0.5 women Precision ProductionTechnician Operator/Assembler Professional Admin/Clerical Sales Managerial Construction/Extractive Farm Transportation Repair Service.15.1.05 0.05.1 change in share of workers, to occupation stress score 3.0 Managerial Professional Precision Production Technician Sales Repair Service Admin/Clerical Operator/Assembler Transportation Construction/Extractive Farm 2.5 2.0 1.5 1.0 0.5 men.1.05 0.05.1 change in share of workers, to Area of circle is proportional to occupation s share of workers in.
Changes in women s feelings at work by education 4.3 happy 2.9 stress 4.8 meaning 4.2 2.8 2.7 4.7 4.1 2.6 4.6 2.5 4.0 2.4 4.5 3.9 2.3 2.2 4.4 3.8 2.1 4.3 high school some college bachelor s
Changes in men s feelings at work by education 4.2 happy 2.6 stress 4.7 meaning 4.1 2.5 4.6 4.0 2.4 4.5 3.9 2.3 4.4 4.3 3.8 2.2 4.2 3.7 2.1 4.1 3.6 2.0 4.0 high school some college bachelor s
Changes in women s feelings at work by race (education HS) 4.7 happy 2.4 stress 4.7 meaning 4.6 2.3 4.6 4.5 2.2 4.5 4.4 4.3 2.1 4.4 4.2 2.0 4.3 4.1 1.9 4.2 4.0 1.8 4.1 white black
Changes in men s feelings at work by race (education HS) 4.3 happy 2.1 stress 4.9 meaning 2.0 4.2 1.9 4.8 4.1 1.8 1.7 4.7 4.0 1.6 4.6 1.5 3.9 1.4 4.5 white black
Robustness check: fixed effects
Does our regression estimate a causal relationship? The concern: Maybe high-happiness jobs look that way only because inherently happy people happen to choose those jobs. Partial solution: fixed effects model See how a given person s feelings change as he or she moves between work and non-work activities. Takes out any permanent differences in feelings for that person. Drawback: spillovers Won t solve the problem if work affects how you feel at home. Positive spillovers (pleasant job makes you happy at home) FE underestimates causal effect of job. Negative spillovers (happy to come home from unpleasant job) FE overestimates causal effect.
Fixed effects estimation Regression model: z ij = η i + w ij δ o + (1 w ij ) ζ n o n ( + (1 w ij ) α a + a e γ e + r β r ) + ɛ ij z ij : i s report of a particular feeling during activity j {1, 2, 3}. η i : respondent fixed effect. w ij : = 1 if activity j is work on main job. ζ n : dummies for various non-work activities. Demographics can affect difference in feelings between work, home. Fixed-effects-adjusted mean feelings for occupation o: z FE o = η + ˆδ FE o
Changes in aggregate feelings at work for women happy 4.20 4.10 4.00 3.90 3.80 tired 2.75 2.65 2.55 2.45 2.35 2.25 sad 0.85 0.80 0.75 0.70 0.65 pain 1.20 1.10 1.00 0.90 0.80 stress 2.65 2.55 2.45 2.35 2.25 2.15 meaning 4.50 4.45 4.40 4.35 4.30 4.25 baseline fixed effects
Changes in aggregate feelings at work for men happy 4.00 3.95 3.90 3.85 3.80 tired 2.75 2.65 2.55 2.45 2.35 2.25 sad 0.85 0.80 0.75 0.70 0.65 pain 1.20 1.10 1.00 0.90 0.80 stress 2.65 2.55 2.45 2.35 2.25 2.15 meaning 4.50 4.45 4.40 4.35 4.30 4.25 baseline fixed effects
Why are FE results sometimes different from baseline? Mechanics: An occupation will affect the trend only to the extent that its share of the work force changed over time. How does difference between baseline and FE scores for each occupation relate to change in occupation s share of work force?
Relationship between FE and baseline scores, women fixed effects score 5.0 4.5 4.0 3.5 3.0 11 happy 7 3.0 3.5 4.0 4.5 5.0 baseline score fixed effects score 1.5 1.0 0.5 0.5 1.0 sad 1.0 0.5 0.5 1.0 1.5 baseline score fixed effects score 3.0 2.0 1.0 stress 7 1.0 2.0 3.0 baseline score fixed effects score 4.0 3.0 2.0 1.0 tired 11 1.0 2.0 3.0 4.0 baseline score fixed effects score 2.0 1.0 1.0 pain 11 2.0 2.0 1.0 1.0 2.0 baseline score fixed effects score 6.0 5.0 4.0 3.0 11 meaning 3.0 4.0 5.0 6.0 baseline score Area of circle proportional to absolute value of change in occupation s share of workers from to. Codes 7 = farming, 11 = operators/assemblers.
Relationship between FE and baseline scores, men fixed effects score 4.2 4.0 3.8 3.6 3.4 happy 11 3.4 3.6 3.8 4.0 4.2 baseline score fixed effects score 1.0 0.8 0.6 0.4 3 4 sad 5 0.4 0.6 0.8 1.0 baseline score fixed effects score 2.8 2.6 2.4 2.2 2.0 1.8 7 stress 8 1.8 2.0 2.2 2.4 2.6 2.8 baseline score fixed effects score 2.8 2.6 2.4 3 2.2 2.0 11 tired 2.0 2.2 2.4 2.6 2.8 baseline score fixed effects score 1.4 1.2 1.0 0.8 0.6 pain 12 0.6 0.8 1.0 1.2 1.4 baseline score fixed effects score 4.8 4.6 4.4 4.2 4.0 3.8 meaning 6 8 7 3.8 4.0 4.2 4.4 4.6 4.8 baseline score Area of circle proportional to absolute value of change in occupation s share of workers from to. Codes 3 = technicians, 4 = sales, 5 = admin/clerical, 6 = service, 7 = farming, 8 = precision production, 11 = operators/assemblers, 12 = transportation.
Changes in women s feelings at work by education, fixed-effects-adjusted 4.4 happy 2.9 stress 4.7 meaning 4.3 4.2 2.8 2.7 2.6 4.6 4.1 2.5 4.5 4.0 3.9 2.4 2.3 2.2 4.4 3.8 2.1 4.3 high school some college bachelor s
Changes in men s feelings at work by education, fixed-effects-adjusted 4.1 happy 2.6 stress 4.6 meaning 4.0 2.5 4.5 3.9 3.8 3.7 2.4 2.3 2.2 2.1 4.4 4.3 4.2 4.1 4.0 3.6 2.0 3.9 high school some college bachelor s
Conclusions Since, work has become: Less physically taxing for everyone. Happier and more meaningful for women. Less happy, less meaningful, and more stressful for men. Changes appear larger for people with less education. What s next on our agenda: How can these changes help us understand evolution of employment, labor force participation, etc.? What about non-market work?
The Changing (Dis-)Utility of Work Greg Kaplan, University of Chicago and NBER Sam Schulhofer-Wohl, Federal Reserve Bank of Chicago The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Chicago or the Federal Reserve System.