Here Comes the Rain Again: Weather and the Intertemporal Substitution of Leisure. Marie Connolly Princeton University

Size: px

Start display at page:

Download "Here Comes the Rain Again: Weather and the Intertemporal Substitution of Leisure. Marie Connolly Princeton University"

Marcia Webster
5 years ago
Views:

1 Here Comes the Rain Again: Weather and the Intertemporal Substitution of Leisure Marie Connolly Princeton University ATUS Early Results Conference, December 8-9, 2005

2 The weather Influence on leisure activities Good weather: desire to enjoy it Example: first nice days of the spring Bad weather: more work? Depends on the flexibility of work schedules Weather and stock returns

3 Literature Lucas and Rapping (1969): model of intertemporal substitution of labor In practice: little supporting evidence (Card 1994, Ham and Reilly 2002) Farber s critique (2005): the standard neoclassical model assumes that workers can set their hours in response to changes in the wage, or can select a wage-hours combination from a dense joint distribution of jobs

4 Recent advances Look at labor supply for jobs in which workers are free to set their own hours: Taxi drivers (Camerer et al. 1997, Farber 2004 and 2005) Bike messengers (Ferh and Goette 2002) Stadium vendors (Oettinger 1999)

5 Purpose of this paper Use time-use data combined with weather data Analyze the impact of weather shocks (which are exogenous) on the marginal adjustment of labor supply Horizon: very short-run Time-use categories: work, leisure, home production

6 Outline of talk Introduction Model Data Results Conclusion

7 Preview of the results Effect of a rainy day (0.1 inches of rain or more in 24 hours) Men: 13 minutes go from leisure to work (45 minutes in the Sunbelt) Men: Rain yesterday reduces time at work today by 25 minutes Women: Rain yesterday and today: 33 minutes more in leisure

8 Model (1) Weather (s) influences utility derived from leisure, observed when planning the day Worker chooses labor supply (N) and time in leisure (L) today and tomorrow (t and t+1) Constraints: total time (T=N+L), and contract with employer (sort of budget constraint) Adjustments only on the margin

9 Model (2) ( ) 1 1,, max t t t t L L L s L s U t t C r N w N w t t t t Subject to the budget constraint: And the time constraints: And E(s)= = + = t t t t N L T N L T

10 Model (3) Solution: U U L L W W t+ 1 s s t+ 1 ( 1 r) t t t = + t+ 1 If s t >s t+1 (today is a nicer day than tomorrow), then the worker will want to increase L t and decrease L t+1. w/s weather-adjusted wage

11 Data (1) American Time Use Survey 2003 and 2004 Activities: work, leisure, and home production; total number of minutes in a day Match with latest CPS (8 th MIS) to get geographical information (county code or MSA/PMSA code), lose ~1/3 of sample

12 Data (2) National Climatic Data Center Over 8,000 individual weather stations Daily summary: info on precipitation, snowfall, maximum and minimum temperature (from which mean temperature is calculated) Daily normals: normal precipitation and temperature

13 What is a Nice Day? How to get a value for s? Would like to know the form of the function f: S = f (TMEAN, normtmean, PRCP, normprcp, SNOW, month, region) Strategy used: impact of rain Impact of temperature not clear Tobits of time in each time-use category on weather variables (rainy day dummies)

14 Mark Twain Hypothesis «Everybody talks about the weather, but nobody does anything about it.» -Mark Twain Twain Hypothesis (H0): The rain dummies are not (jointly) significant Test of the Twain hypothesis: F-test of the rain dummies

16 Table 2 Summary statistics (continued) 8.3 hours 4.7 hours 3.5 hours

23 Table 9 (continued)

24 Results Effect of a rainy day (0.1 inches of rain or more in 24 hours) Men: 13 minutes go from leisure to work (45 minutes in the Sunbelt: due to low persistence of the weather?) Men: Rain yesterday reduces time at work today by 25 minutes Women: Rain yesterday and today: 33 minutes more in leisure

27 Conclusion Workers respond to weather shocks Intertemporal substitution of labor supply and leisure Marginal adjustments Flexibility Twain Hypothesis rejected People DO something about it!

28 Future research Add persistence of the weather to the model (should explain larger coefficients for the Sunbelt) Better understand the impact of weather on mood and well-being Data on what is a nice day (s), or estimate the f function Data on cloud cover Direct impact of bad weather? Are workers who lose their workday compensated in any way? Flexibility and well-being: higher productivity?

1 Static (one period) model

1 Static (one period) model The problem: max U(C; L; X); s.t. C = Y + w(t L) and L T: The Lagrangian: L = U(C; L; X) (C + wl M) (L T ); where M = Y + wt The FOCs: U C (C; L; X) = and U L (C; L; X) w +