We use the UK Click-and-Drag Diary Instrument (CaDDI) data (Sullivan et al., 2021b) to examine changes in time use patterns before, during, and after the Covid-19 pandemic.4 This dataset comprises time use diaries collected in the UK during multiple periods: prior to the pandemic in 2016, during the pandemic (May-June 2020, November 2020, January 2021), and post-pandemic when confinement measures eased (Summer 2021). The diaries, structured in 10-minute episodes from 4 am to 4 am the following day, capture primary activities, their locations, and the presence of others. Our analysis aims to explore shifts in time allocation across these periods. Time use diaries, such as those in the CaDDI sample, have become pivotal in understanding daily worker behaviors (Gimenez-Nadal and Molina, 2022). They offer more accurate data compared to surveys reliant on general questions (Harms et al., 2019), providing more precise estimations (Bonke, 2005; Yee-Kan, 2008).
The UK initiated lockdown measures in March 2020 in response to the escalating Covid-19 pandemic. Subsequent lockdowns were implemented to address evolving pandemic situations, with adjustments made to the timeline and severity in accordance with changing circumstances. The initial lockdown announcement occurred on March 23rd, legally enforced from March 26th. This lockdown was extended on April 16th, and by May 10th, guidance suggested a return to workplaces for individuals unable to work from home, while advocating avoiding public transport. Schools and non-essential shops gradually reopened towards the end of June, albeit with ongoing adherence to lockdown measures and social distancing protocols. This period, spanning from the onset of the lockdown to June 2020, aligns with the first data collection period for the CaDDI diaries during the lockdown.
Several measures relaxing lockdown restrictions were implemented in July and August 2020, marking a period of eased restrictions. However, as the pandemic evolved, a second national lockdown commenced from November 5th to December 2nd. This second lockdown coincides with the CaDDI's second data collection phase during Covid-19 lockdowns. Subsequently, due to evolving circumstances, a third national lockdown began on January 6th, 2021. Throughout the spring and summer of 2021, most restrictions were gradually eased or lifted, with the majority of legal limits on social contact removed by July. This phase of relaxed, eased, or lifted measures corresponds to the relaxation period, during which the CaDDI collected time use diaries.
3.1 Sample requirements
We began with the original CaDDI sample, comprising 3,423 individuals. Initially, we excluded 312 individuals due to incomplete data on key variables, resulting in a reduced sample of 3,111 individuals. Following this, we further refined the sample by excluding individuals who reported diary entries on atypical days, yielding a sample size of 2,734 individuals. Given our focus on WFH practices, we narrowed the scope to employed individuals, including both employees and self-employed workers, leading to the exclusion of 848 individuals not employed. Moreover, to capture work-related diary entries specifically, we retained only workers who completed diaries on workdays, culminating in a sample of 1,425 individuals. Additionally, considering the temporal structure of the CaDDI data, we eliminated 215 individuals who recorded diary entries during a brief period between lockdowns in Summer 2020. This finalizes our sample selection process. These restrictions leave a final sample of 1,210 individuals, and 1,808 observations, as some interviewed individuals filled in two diaries during two different workdays.5 234 individuals (294 observations) correspond to the period pre Covid-19; 774 individuals (1,177 observations) correspond to the lockdowns period; finally, 202 individuals (337 observations) correspond to the period of relaxation that followed lockdowns. Furthermore, the 57.36% of the individuals in the sample (i.e., 694 individuals) are male workers, and the remaining 42.64% (516 individuals) are females.
3.2 Variables
We begin by utilizing the diary structure of the CaDDI data to determine the time allocated by interviewed workers to paid work, leisure, and unpaid work activities, based on their primary activity throughout the day. Paid work is delineated in accordance with the CaDDI data guide, encompassing activities categorized as 117 "paid work including at home," 118 "formal education," and 125 "work break”. Leisure activities are identified based on previous research methodologies (e.g., Aguiar and Hurst, 2007; Gimenez-Nadal and Sevilla, 2012), encompassing activities such as dog walking, hobbies, reading, engaging in sports, etc. Lastly, unpaid work comprises activities related to household chores, excluding childcare duties.6
We also utilize the diary structure of the CaDDI data to determine workers who have the capability to work from home (WFH) during the diary day. Specifically, we leverage information regarding the location of each diary episode, distinguishing between episodes at home, the workplace, and other or unspecified locations. Our primary identification of WFH workers is as follows: we categorize a worker as WFH if all their paid work episodes occurred at home, indicating that the individual did not commute to or from work on the diary day and solely worked from home. Additionally, we identify WFH individuals as those who spent at least one hour engaged in paid work activities at home (referred to as "WFH (bis)" throughout the analysis).
The third pivotal variable in our analysis measures the instantaneous enjoyment experienced during daily activities. Within the CaDDI data, information is provided on the instant enjoyment associated with each episode of paid work, leisure, and unpaid work, elicited through the question: “How much did you enjoy this time?”, utilizing a scale from 1 “didn’t enjoy at all” to 7 “enjoyed very much”. This variable serves to define the instant enjoyment linked to each episode, capturing the affective feelings, also known as “experienced instant utility” or “instantaneous well-being,” that individuals experience during specific activities (Kahneman et al., 2004). Consequently, this well-being measure encapsulates the moment-to-moment flow of enjoyment (Kahneman and Krueger, 2006; Kahneman and Deaton, 2010).7
The CaDDI data enables the definition of several variables potentially associated with time allocation and well-being, which we designate as control variables. These variables encompass age (measured in years), highest level of formal education attained, marital status (cohabiting or single), UK citizenship status, family size, number of children, self-employment status, and usual weekly work hours. Additionally, we categorize the day of the week when diaries were completed; residential region (London, Yorkshire & Humberside, East Midlands, East Anglia, South East, South West, West Midlands, North West, Scotland, Wales, Northern Ireland, and North East); social status (upper and middle class, lower middle class, skilled working class, and working class) as a proxy for income or earnings.8 Moreover, we consider worker occupation (semi or unskilled manual, skilled manual, clerical/administrative, supervisory/junior managerial, intermediate managerial/professional/administrative, higher managerial/professional/administrative), and worker self-reported health status (ranging from 1 “very bad” to 5 “very good”).
Finally, we define certain controls at the episode level, which could potentially influence the experienced instantaneous enjoyment of episodes (Gimenez-Nadal et al., 2023). These controls encompass whether episodes are conducted alone, with a spouse, with children, or with others; the location of episodes (at home, workplace, or another/unspecified location); the engagement in secondary activities while performing the main activity; and the hour of the day when activities are undertaken, intended to capture potential fatigue accumulated throughout the day that might impact well-being.
3.3 Descriptive statistics
Table 1 displays the proportion of WFH workers within the sample (both in our baseline and alternative identification). Preceding the Covid-19 period, 16.9% of observations corresponded to workers who exclusively conducted their paid work activities from home, while 20.2% represented workers who spent at least 1 hour working from home.9 These figures surged to 63.6% and 67.0%, respectively, during the lockdown period, and decreased to 44.8% and 50.5% during the relaxation period. These trends align with existing data indicating a substantial increase in teleworking practices during the Covid-19 period (Barrero et al., 2023), mirroring the measures adopted to adapt to the evolving pandemic circumstances.
Table 1
Workers from home and away from home
|
Pre Covid-19
|
Lockdowns
|
Relaxation
|
VARIABLES
|
Mean
|
D.Dev.
|
Mean
|
D.Dev.
|
Mean
|
D.Dev.
|
WFH
|
0.169
|
0.375
|
0.636
|
0.481
|
0.448
|
0.498
|
WFH (bis)
|
0.202
|
0.402
|
0.670
|
0.470
|
0.505
|
0.501
|
N. Observations
|
294
|
1177
|
337
|
Note: The sample (CaDDI) is restricted to employed individuals that worked during the diary day. WFH are defined as workers whose paid work activities were all reported at home. WFH (bis) are defined as workers who spend at least 1 hour in paid work activities at home. |
Table 2 showcases descriptive statistics of the primary variables, namely the time allocated by workers to paid work, leisure, and unpaid work activities. Before the Covid-19 pandemic, average daily durations for WFH female workers were 435.8 minutes in paid work activities, 264.4 minutes in leisure activities, and 69.4 minutes in unpaid work activities. In comparison, average daily durations for male workers were 494.1 minutes in paid work activities, 273.7 minutes in leisure activities, and 24.8 minutes in unpaid work activities. These figures indicate that prior to the Covid-19 outbreak, both female and male WFH individuals spent less time in paid work activities and more time in leisure activities compared to their WAFH counterparts, consistent with findings for the US by Gimenez-Nadal et al. (2020).
Table 2
Descriptive statistics of main variables
|
FEMALES
|
MALES
|
|
WAFH
|
WFH
|
Diff.
|
WAFH
|
WFH
|
Diff.
|
VARIABLES
|
Mean
|
S.Dev.
|
Mean
|
S.Dev.
|
|
Mean
|
S.Dev.
|
Mean
|
S.Dev.
|
|
A. Pre Covid-19
|
|
|
|
|
|
|
|
|
|
|
Paid work time
|
435.805
|
123.221
|
316.751
|
192.974
|
119.054***
|
494.082
|
106.875
|
437.433
|
204.534
|
56.649**
|
Leisure time
|
263.402
|
121.861
|
328.766
|
187.838
|
-65.364**
|
273.674
|
136.942
|
335.216
|
209.522
|
-61.542**
|
Unpaid work time
|
69.390
|
86.911
|
85.386
|
74.427
|
-15.996
|
24.758
|
39.366
|
38.223
|
54.656
|
-13.465
|
N. Observations
|
99
|
26
|
|
149
|
20
|
|
B. Lockdowns
|
|
|
|
|
|
|
|
|
|
|
Paid work time
|
469.786
|
149.002
|
436.935
|
144.607
|
32.851***
|
478.235
|
125.202
|
434.067
|
137.288
|
44.168***
|
Leisure time
|
260.567
|
151.305
|
257.973
|
149.899
|
2.594
|
268.647
|
144.017
|
295.051
|
148.805
|
-26.404**
|
Unpaid work time
|
52.727
|
67.067
|
61.872
|
69.083
|
-9.145
|
34.727
|
53.227
|
51.605
|
78.281
|
-16.878***
|
N. Observations
|
186
|
319
|
|
253
|
419
|
|
C. Relaxation
|
|
|
|
|
|
|
|
|
|
|
Paid work time
|
462.155
|
121.770
|
470.080
|
180.852
|
-7.925
|
488.778
|
146.319
|
474.154
|
131.333
|
14.624
|
Leisure time
|
202.570
|
137.595
|
234.420
|
168.315
|
-31.850
|
289.822
|
198.526
|
251.370
|
171.138
|
38.452
|
Unpaid work time
|
45.612
|
56.479
|
64.817
|
85.939
|
-19.205
|
41.660
|
58.522
|
45.638
|
61.656
|
-3.978
|
N. Observations
|
59
|
50
|
|
129
|
99
|
|
Note: The sample (CaDDI) is restricted to employed individuals that worked during the diary day. All time allocations are defined in minutes per day. WFH are defined as workers whose paid work activities were all reported at home. Difference defined as the average value for WAFH minus the average value for WFH. *** significant at the 1%; ** significant at the 5%; * significant at the 10%. |
However, Table 2 also illustrates a reduction in these disparities during lockdowns, especially among women, and further reductions during the relaxation period. This suggests that the normalization of WFH due to Covid-19 generated a convergence in daily behaviors between WFH and WAFH workers. Specifically, during lockdowns, average durations for WFH female workers were 469.8 minutes in paid work activities, 260.6 minutes in leisure activities, and 52.7 minutes in unpaid work activities. Meanwhile, average durations for male workers were 478.2 minutes in paid work activities, 268.6 minutes in leisure activities, and 34.7 minutes in unpaid work activities. During the relaxation period, average durations for WFH female workers were 462.2 minutes in paid work activities, 202.6 minutes in leisure activities, and 45.6 minutes in unpaid work activities. For male workers during the relaxation period, average durations were 488.8 minutes in paid work activities, 289.8 minutes in leisure activities, and 41.7 minutes in unpaid work activities.
Table 3 presents the average enjoyment ratings for episodes of paid work, leisure, and unpaid work activities at the episode level. Paid work episodes appeared to be more enjoyable for WFH individuals than for WAFH individuals before the Covid-19 pandemic. During lockdowns and the relaxation period, this trend persisted for females, yet a reverse trend was observed for males. Male WFH reported lower levels of enjoyment during paid work compared to WAFH counterparts, with the difference being statistically significant only during lockdowns. In contrast, for leisure activities, Table 3 indicates that WAFH individuals derived greater enjoyment from their leisure episodes than WFH individuals, for both women and men, during lockdowns and the relaxation period. However, before the Covid-19 outbreak, this was solely observed among males, while female WFH and WAFH individuals reported similar enjoyment levels for their leisure activities. Lastly, concerning unpaid work, averages indicate that before the Covid-19 pandemic, female WAFH individuals found more enjoyment in their unpaid work episodes compared to WFH individuals, while the opposite was observed for males. However, during lockdowns and the relaxation period, it was WAFH workers, regardless of gender, who found more enjoyment in their unpaid work episodes.
Table 3
Averages of enjoyment at the episode level
|
PAID WORK EPSODES
|
LEISURE EPISODES
|
UNPAID WORK EPISODES
|
VARIABLES
|
WAFH
|
WFH
|
Diff.
|
WAFH
|
WFH
|
Diff.
|
WAFH
|
WFH
|
Diff.
|
A. Pre Covid-19
|
|
|
|
|
|
|
|
|
|
Women
|
4.081
|
5.446
|
-1.365***
|
5.685
|
5.687
|
-0.002
|
5.557
|
5.162
|
0.395**
|
N. episodes
|
4,350
|
786
|
|
2,564
|
860
|
|
653
|
218
|
|
Men
|
4.399
|
4.818
|
-0.419***
|
5.673
|
5.506
|
0.167***
|
4.935
|
5.092
|
-0.157***
|
N. episodes
|
7,400
|
844
|
|
4,046
|
679
|
|
384
|
78
|
|
B. Lockdowns
|
|
|
|
|
|
|
|
|
|
Women
|
4.453
|
4.582
|
-0.129***
|
5.390
|
5.385
|
0.005***
|
5.093
|
4.964
|
0.129***
|
N. episodes
|
8,864
|
13,785
|
|
4,743
|
8,366
|
|
994
|
1,881
|
|
Men
|
4.925
|
4.786
|
0.139***
|
5.719
|
5.526
|
0.193***
|
5.561
|
4.920
|
0.641***
|
N. episodes
|
12,165
|
18,245
|
|
6,775
|
12,482
|
|
815
|
2,028
|
|
C. Relaxation
|
|
|
|
|
|
|
|
|
|
Women
|
4.560
|
4.863
|
-0.303***
|
5.365
|
4.954
|
0.411***
|
5.147
|
4.660
|
0.487***
|
N. episodes
|
2,728
|
2,317
|
|
1,216
|
1,159
|
|
288
|
364
|
|
Men
|
5.234
|
4.563
|
0.671
|
6.035
|
5.649
|
0.386***
|
5.420
|
5.067
|
0.353**
|
N. episodes
|
6,319
|
4,583
|
|
3,750
|
2,625
|
|
551
|
501
|
|
Note: The sample (CaDDI-episode level) is restricted to episodes of employed individuals that worked during the diary day. Difference defined as the average value for WAFH minus the average value for WFH. *** significant at the 1%; ** significant at the 5%; * significant at the 10%. |