Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S...
Transcript of Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S...
![Page 1: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/1.jpg)
Dissecting Idiosyncratic Income Risk
E Halvorsen H Holter S Ozkan K StoreslettenStatistics Norway U of Oslo U of Toronto U of Oslo
XIII REDg Workshop, CEMFINovember 23, 2018
![Page 2: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/2.jpg)
Nonlinear/Non-Gaussian Earnings Dynamics
• Sparked interest in income dynamics: Guvenen et al. (2014, 2015),Arellano et al. (2017) etc.
• Non-Gaussian features of income shocks
• Left skewness and excess kurtosis
S. Ozkan (Toronto) Introduction 1 / 37
![Page 3: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/3.jpg)
Nonlinear/Non-Gaussian Earnings Dynamics
-2 -1 0 1 2
yt+1 − yt
0
0.5
1
1.5
2
2.5
3
3.5Den
sity
US DataN (0, 0.51)
Std. Dev. = 0.51
Skewness = -1.07
Kurtosis = 14.93
S. Ozkan (Toronto) Introduction 1 / 37
![Page 4: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/4.jpg)
Nonlinear/Non-Gaussian Earnings Dynamics
• Sparked interest in income dynamics: Guvenen et al. (2014, 2015),Arellano et al. (2017) etc.
• Non-Gaussian features of income shocks
• Left skewness and excess kurtosis
• Asymmetric/nonlinear mean reversion. Persistence differ by:
• Current shocks change persistence of past ones (Arellano et al.(2017)).
• positive vs negative changes; low vs high income workers
S. Ozkan (Toronto) Introduction 1 / 37
![Page 5: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/5.jpg)
Nonlinear/Non-Gaussian Earnings Dynamics
• Sparked interest in income dynamics: Guvenen et al. (2014, 2015),Arellano et al. (2017) etc.
• Non-Gaussian features of income shocks
• Left skewness and excess kurtosis
• Asymmetric/nonlinear mean reversion. Persistence differ by:
• Current shocks change persistence of past ones (Arellano et al.(2017)).
• positive vs negative changes; low vs high income workers
• New findings make their way into the quantitative macro literature:
• McKay (2014), Kaplan et al. (2016), Golosov et al. (2016), De Nardiet al. (2016) and Wang et al. (2016)
S. Ozkan (Toronto) Introduction 1 / 37
![Page 6: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/6.jpg)
Open Questions: Drivers of Non-Gaussian Features
Focus has been on annual individual labor income dynamics.
• What’s driving skewness/kurtosis of earnings growth?Wages vs Hours?
• Does hourly wage dynamics exhibit non-Gaussian/ nonlinear features?
• What are the real-life events lead to large swings in earnings?
S. Ozkan (Toronto) Introduction 2 / 37
![Page 7: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/7.jpg)
Open Questions: Insurance Against Tail Shocks
How much insurance against large earnings losses/gains from spouseand government?
• Do non-Gaussian features extend to household (husband+wife)earnings? After public insurance?
• For some questions nature of household income risk—before and aftertax—is key.
• How about consumption growth distribution?
S. Ozkan (Toronto) Introduction 3 / 37
![Page 8: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/8.jpg)
What Do We Do?
Use the Norwegian registry data to study above questions.
1. Show that patterns for annual earnings risk are remarkably similar tothe US.
2. Study the role of wages vs hours in non-Gaussian properties ofearnings changes.
• Decompose earnings changes into hours and hourly wage growth.• Do wage and hours growth display non-Gaussian features?
3. Document the insurance against tail shocks of earnings throughspouse’s income and public insurance.
• Distribution of after tax-after transfer household income growth.
S. Ozkan (Toronto) Introduction 4 / 37
![Page 9: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/9.jpg)
Today
1. Data and Empirical Methodology
2. Earnings Growth Distribution
3. Changes in Hours vs Wages
4. Household Income Dynamics
5. Conclusion
![Page 10: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/10.jpg)
Data and Empirical Methodology
Norwegian Registry Data
![Page 11: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/11.jpg)
Norwegian Registry Data
• Administrative data covering the whole Norwegian population.
• Derived from a combination of administrative registers such asannual tax records and employment register
• High quality because
• Third-party reported: employers, banks, brokers, etc.
• No attrition (unless someone emigrates).
• Family identifiers from the population register.
• includes cohabitant couples.
S. Ozkan (Toronto) Data and Empirical Methodology 5 / 37
![Page 12: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/12.jpg)
Norwegian Registry Data: Base Sample
• Panel data between 1998 and 2014.
• Income data goes back to 1993 but not hours.
• Today we focus on males.
• We do the same analysis with women.
• We use ~20M year/individual observations in our analysis
• Labor Earnings for wage and salary workers including bonuses andother remunerations.
• Business income for self-employed workers: no hours data.
• Deflate all values with the 2000 CPI.
S. Ozkan (Toronto) Data and Empirical Methodology 6 / 37
![Page 13: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/13.jpg)
Data and Empirical Methodology
Methodology
![Page 14: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/14.jpg)
Sample Selection
• Revolving panel of 25-60 year olds between 2003-2014.
• In year t select individuals participating in the labor market:
• Y is > Y min
s in t − 1 and for 2 more years between t − 2 and t − 5.
• Y mint is 5% of median earnings; approximately one quarter of
full-time work at the half of the minimum wage.
• For every worker, compute recent wage earnings between t − 1 and
t − 5, Yit−1 ≡ (1
5)∑t−1
s=t−5
(Y i
sdhi,s
).
• Y is : Total earnings in year s.
• dhi,s : Average earnings in age hi,s.
S. Ozkan (Toronto) Data and Empirical Methodology 7 / 37
![Page 15: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/15.jpg)
A Graphical Construct
• Divide the population into 3 age groups in t − 1: 25–34, 35–44,...,45–54.
• Within each age group rank individuals according to Y t−1 into 10 REdeciles.
• Within each age group, against each quantile of Y t−1 on the x-axis:• plot conditional distribution F
(yt+k − yt |Y t−1
)on the y-axis,
k = 1,5.
S. Ozkan (Toronto) Data and Empirical Methodology 8 / 37
![Page 16: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/16.jpg)
Earnings Growth Distribution
Norway vs US
![Page 17: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/17.jpg)
Histogram of yt+k − yt
1-Year Growth
-2 -1 0 1 2
yt+1 − yt, Log Earnings Income Change
0
1
2
3
4
Density
Earnings Year: 2007
Normal Dist.
St. Dev. = 0.40
Skewness = -1.28
Kurtosis = 14.67
5-Year Growth
-2 -1 0 1 2
yt+5 − yt, Log Earnings Income Change
0
0.5
1
1.5
2
2.5
Density
Earnings Year: 2007
Normal Dist.
St. Dev. = 0.50
Skewness = -0.99
Kurtosis = 11.27
• Peaky center, narrow shoulders, long tails⇒ Excess kurtosis.
• Left tail longer than right tail⇒ Left (Negative) Skewness.
S. Ozkan (Toronto) Earnings Growth Distribution 9 / 37
![Page 18: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/18.jpg)
Standard Deviation of yt+5 − yt
Norway Data
2 4 6 8 10
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
25-35
36-45
46-55
US Data
20 40 60 80 100
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
25-34
35-44
45-54
• Changes are smaller in Norway.
• RE and age variation are very similar in both countries.
S. Ozkan (Toronto) Earnings Growth Distribution 10 / 37
![Page 19: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/19.jpg)
Skewness of yt+5 − yt
Norway Data
2 4 6 8 10
-2.5
-2
-1.5
-1
-0.5
0
0.5
25-35
36-45
46-55
US Data
20 40 60 80 100
-2.5
-2
-1.5
-1
-0.5
0
0.5
25-34
35-44
45-54
• In both economies, distributions are similarly left skewed.
• Left skewness increases by RE and age in a similar fashion.
S. Ozkan (Toronto) Earnings Growth Distribution 11 / 37
![Page 20: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/20.jpg)
Kurtosis of yt+5 − yt
Norway Data
2 4 6 8 10
2
6
10
14
18
22
26
25-35
36-45
46-55
US Data
20 40 60 80 100
2
6
10
14
18
22
26
25-34
35-44
45-54
• 5-year earnings distribution exhibits higher excess kurtosis in Norway.
• Excess kurtosis follows hump-shaped pattern over RE in both.
S. Ozkan (Toronto) Earnings Growth Distribution 12 / 37
![Page 21: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/21.jpg)
Changes in Hours vs Wages
Hours Data
![Page 22: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/22.jpg)
Employment Register-Administrative Data
• Hours data reported by employers between 2003 and 2014
• On contractual working hours per week
• Only for wage and salary workers w/ ≥ 4 hours/week contracts
• No self-employed workers or freelancers
• Cover 75% of population between 25 and 60.
• Measurement error in Employment register:
• fail to report employment spells correctly or update hour changes,
• overtime hours are not included,
• employers with irregular employments are more prone.
S. Ozkan (Toronto) Changes in Hours vs Wages 13 / 37
![Page 23: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/23.jpg)
Employment Register-Administrative Data
• Hours data reported by employers between 2003 and 2014
• On contractual working hours per week
• Only for wage and salary workers w/ ≥ 4 hours/week contracts
• No self-employed workers or freelancers
• Cover 75% of population between 25 and 60.
• Measurement error in Employment register:
• fail to report employment spells correctly or update hour changes,
• overtime hours are not included,
• employers with irregular employments are more prone.
S. Ozkan (Toronto) Changes in Hours vs Wages 13 / 37
![Page 24: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/24.jpg)
Employment Survey-AKU
• Survey data on employment and work hours
• Measure of actual hours worked last week.
• Better quality than register data.
• 3200 working-age individuals interviewed 8 quarters in a row.
• Compute annual hours as hannual =∑4
t=1 13× hactual
• Link observations to register data (data on annual wages, registerhours, etc.)
S. Ozkan (Toronto) Changes in Hours vs Wages 14 / 37
![Page 25: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/25.jpg)
Imputation of Hours in the Register Data
• Impute a better hours measure in administrative register data.
• Link observations to register data (data on annual wages, registerhours, etc.)
• Regress the AKU hours on observables in the register data, e.g.,register hours, sick days, unemp. spells, parental leave, etc.
• Estimate an hour measure in register data using the estimates of thisregression.
S. Ozkan (Toronto) Changes in Hours vs Wages 15 / 37
![Page 26: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/26.jpg)
Imputation of Hours in the Register Data
-0.6 -0.4 -0.2 0 0.2 0.4-0.6
-0.4
-0.2
0
0.2
0.4
Actual Annual Hours AKU
Imputed Annual Hours
Employment Register Annual Hours
• Hours changes in register data are smaller than those from AKU.
• Imputation is doing fairly a good job in replicating the AKU measure.
S. Ozkan (Toronto) Changes in Hours vs Wages 16 / 37
![Page 27: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/27.jpg)
Changes in Hours vs Wages
Earnings Growth: Hours vs Wage
![Page 28: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/28.jpg)
Annual Hours vs Hourly Wage
• Decompose changes in earnings to hourly wage or hours components.
• Group workers w.r.t. annual wage growth between t − 1 and t , ∆et ,1
into 20 equally sized bins.
• On top of conditioning on age in t − 1 (young vs prime age) andpast 5-year income (RE) deciles Y
it−1.
• e.g., a group of prime age men with median past income whoexperience 25 log points decline in earnings between t − 1 and t .
• How much hourly wage and hours growth each group experience?
S. Ozkan (Toronto) Changes in Hours vs Wages 17 / 37
![Page 29: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/29.jpg)
Hours vs Wage: Median RE
Median RE Decile
-0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Annual Earnings
Hourly Wage
Annual Hours
• Large earnings swings: hours and wage growth are equally important.
• Smaller earnings changes: wage growth is more important.
S. Ozkan (Toronto) Changes in Hours vs Wages 18 / 37
![Page 30: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/30.jpg)
Hours vs Wage: Bottom vs Top RE
Bottom RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Annual Earnings
Hourly Wage
Annual Hours
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Annual Earnings
Hourly Wage
Annual Hours
• For bottom RE group hours growth plays a more important role.
• For higher RE groups wage changes are main drivers of earningsgrowth. More RE Groups Average of Log Growth
S. Ozkan (Toronto) Changes in Hours vs Wages 19 / 37
![Page 31: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/31.jpg)
Changes in Hours vs Wages
Dynamics of Hours and Wage Growth
![Page 32: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/32.jpg)
Asymmetric Mean Reversion: Dynamics of Earnings
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Bottom RE Decile
Median RE Decile
Top RE Decile
• For bottom (and median) RE group negative changes are transitoryand positive changes are very persistent.
• The opposite is true for top RE group.
S. Ozkan (Toronto) Changes in Hours vs Wages 20 / 37
![Page 33: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/33.jpg)
Asymmetric Mean Reversion:Hours vs Wages
Earnings Growth
-0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Bottom RE Decile
Median RE Decile
Top RE Decile
Wage Growth
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
Bottom RE Decile
Median RE Decile
Top RE Decile
Hours Change
-0.15 -0.1 -0.05 0 0.05 0.1-0.15
-0.1
-0.05
0
0.05
0.1
Bottom RE Decile
Median RE Decile
Top RE Decile
• Wage changes are very persistent (except for top RE).
• Declines in hours are transitory and increases persistent (not so muchfor top RE).
S. Ozkan (Toronto) Changes in Hours vs Wages 21 / 37
![Page 34: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/34.jpg)
Changes in Hours vs Wages
Distribution of Hours vs Wage Growth
![Page 35: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/35.jpg)
Distribution of Hours vs Wage Growth
• Does hourly wage and annual hours growth distribution exhibitnon-Gaussian/nonlinear features?
• Plot their distributions and higher-order moments.
• How much of the left skewness and excess kurtosis of annual earningsgrowth are driven by changes in hourly wages vs hours?
• Decompose skewness and kurtosis of earnings into hours andwage components.
S. Ozkan (Toronto) Changes in Hours vs Wages 22 / 37
![Page 36: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/36.jpg)
Decomposing Higher-Order Moments
et+k − et︸ ︷︷ ︸∆et,k
= wt+k − wt︸ ︷︷ ︸∆wt,k
+ ht+k − ht︸ ︷︷ ︸∆ht,k
• ∆et ,k : log annual wage growth between t and t + k• ∆wt ,k : log hourly wage growth between t and t + k• ∆ht ,k : log annual hours growth between t and t + k
Skewness Decomposition
s∆et,k =
(σ∆wt,k
σ∆et,k
)3
× s∆wt,k +
(σ∆ht,k
σ∆et,k
)3
× s∆ht,k + co-s∆wt,k ,∆ht,k
Kurtosis Decomposition
κ∆et,k =
(σ∆wt,k
σ∆et,k
)4
× κ∆wt,k +
(σ∆ht,k
σ∆et,k
)4
× κ∆ht,k + co-κ∆wt,k ,∆ht,k
S. Ozkan (Toronto) Changes in Hours vs Wages 23 / 37
![Page 37: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/37.jpg)
Variance of yt+5 − yt for Prime Age Male
Variance
2 4 6 8 10
0.05
0.1
0.15
0.2
0.25
Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 100
0.05
0.1
0.15
0.2
0.25Annual Earnings
Hourly Wage
Annual Hours
Co-Variance
• Hourly wage is more volatile than hours especially above the median.
• Similar to the PSID (Heathcote et al. (2014)). 1-Year Growth Variance
5-Year Growth Variance-Gaussian
S. Ozkan (Toronto) Changes in Hours vs Wages 24 / 37
![Page 38: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/38.jpg)
Skewness of yt+5 − yt for Prime Age Male
Skewness
2 4 6 8 10
-2
-1.5
-1
-0.5
Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
Annual Earnings
Hourly Wage
Annual Hours
Co-Skewness
• Both hours and wage growth are left skewed.
• Wage growth and more importantly co-skewness are driving the leftskewness of earnings growth. 1-Year Growth Skewness 5-Year Growth Skewness-Gaussian
S. Ozkan (Toronto) Changes in Hours vs Wages 25 / 37
![Page 39: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/39.jpg)
Kurtosis of yt+5 − yt for Prime Age Male
Kurtosis
2 4 6 8 10
10
15
20
25
30
35
40
Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10
2
4
6
8
10
12
14
16
18
20 Annual Earnings
Hourly Wage
Annual Hours
Co-Kurtosis
• Wage and hours growth are both leptokurtic (especially hours growth).
• Excess kurtosis due to hourly wage dominates the hours.1-Year Growth Kurtosis 5-Year Growth Kurtosis-Gaussian
S. Ozkan (Toronto) Changes in Hours vs Wages 26 / 37
![Page 40: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/40.jpg)
Changes in Hours vs Wages
Earnings Swings and Important Life CycleEvents
![Page 41: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/41.jpg)
Earnings Swings and Important Life Cycle Events
Event—in/out1-Year Earnings Loss 1-Year Earnings Growth
> 0.5 [0.5, 0.25) [0.25, 0.0) [0.0, 0.25) [0.25, 0.5) ≥ 0.5
Unemployment 8% 7% 2% 2% 7% 8%
Sickness 23% 21% 8% 8% 19% 20%
Part time 15% 13% 6% 8% 19% 23%
Parental leave 6% 9% 5% 6% 7% 5%
Firm change 19% 19% 10% 11% 21% 23%
S. Ozkan (Toronto) Changes in Hours vs Wages 27 / 37
![Page 42: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/42.jpg)
Household Income Dynamics
Insurance Against Tail Shocks
![Page 43: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/43.jpg)
Insurance Against Tail Shocks
• How much insurance against tail shocks from
• Spousal income & Government tax and transfers
• After 1 year? After 5 years?
• Capital income includes positive interests, dividends and realizedcapital gains and losses.
• excludes unrealized capital gains.
• Tax and transfers include UI, DI, SS pension, sickness benefits, paidmaternity leave, money received on government activity program.
• No in kind transfers: health care, daycare subsidies, schools, etc.
• Imputed consumption using the budget constraint of the household.
S. Ozkan (Toronto) Household Income Dynamics 28 / 37
![Page 44: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/44.jpg)
Spousal Insurance
Bottom RE Decile
-1 -0.5 0 0.5 1 1.5
-1
-0.5
0
0.5
1
1.5
Spouse Earnings
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Spouse Earnings
• No change in behavior of spouse or her earnings (not showing the SEincome).
S. Ozkan (Toronto) Household Income Dynamics 29 / 37
![Page 45: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/45.jpg)
Public Insurance
Bottom RE Decile
-1 -0.5 0 0.5 1 1.5
-1
-0.5
0
0.5
1
1.5
HH Labor Income
HH Disposable Income
Consumption
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
HH Labor Income
HH Disposable Income
Consumption
• Strong second earner effect (more so for low RE).
• Public insurance is much more helpful with tail shocks for low income.
• High RE can rely on self insurance for consumption.S. Ozkan (Toronto) Household Income Dynamics 30 / 37
![Page 46: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/46.jpg)
Public Insurance, 1 Year Later
Bottom RE Decile
-1 -0.5 0 0.5 1 1.5
-1
-0.5
0
0.5
1
1.5
HH Labor Income
HH Disposable Income
Consumption
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
HH Labor Income
HH Disposable Income
Consumption
• After 1 year bottom RE see a very small decline in consumption (larger for topRE).
• Earnings losses are more persistent for top RE.
S. Ozkan (Toronto) Household Income Dynamics 31 / 37
![Page 47: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/47.jpg)
Public Insurance, 5 Years Later
Bottom RE Decile
-1 -0.5 0 0.5 1 1.5
-1
-0.5
0
0.5
1
1.5
HH Labor Income
HH Disposable Income
Consumption
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
HH Labor Income
HH Disposable Income
Consumption
• Earnings changes are more persistent for top RE (especially negativechanges).
• After 5 years, top RE still hasn’t recovered the losses between t and t − 1.
S. Ozkan (Toronto) Household Income Dynamics 32 / 37
![Page 48: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/48.jpg)
Household Income Dynamics
Distribution of Household Income Growth
![Page 49: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/49.jpg)
Nature of Idiosyncratic Income Risk
• Do non-Gaussian features of annual earnings growth distributionextend to
• household (husband+wife) earnings?
• After tax/after transfer disposable household income?
• For some questions nature of household income risk—before and aftertax—is key.
• Plot their distributions and higher-order moments.
S. Ozkan (Toronto) Household Income Dynamics 33 / 37
![Page 50: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/50.jpg)
Variance of 5-Year Income Growth, yt+5 − yt
2 4 6 8 10
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45Ind. Labor
HH Labor
HH Disposable
• HH labor is less volatile than individual labor.
• Taxes and transfers reduces variance substantially.
S. Ozkan (Toronto) Household Income Dynamics 34 / 37
![Page 51: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/51.jpg)
Skewness of 5-Year Income Growth, yt+5 − yt
2 4 6 8 10
-2.2
-2
-1.8
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
Ind. Labor
HH Labor
HH Disposable
• Spousal income reduces negative skewness due to second earner effect(similar for the US, Pruitt and Turner (2018)).
• Public insurance reduces left tail further.
S. Ozkan (Toronto) Household Income Dynamics 35 / 37
![Page 52: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/52.jpg)
Kurtosis of 5-Year Income Growth, yt+5 − yt
2 4 6 8 10
8
10
12
14
16
18
20
22
Ind. Labor
HH Labor
HH Disposable
• HH labor and disposable income are still substantially leptokurtic, less so
than individual earnings growth though.
S. Ozkan (Toronto) Household Income Dynamics 36 / 37
![Page 53: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/53.jpg)
Summary of Findings
1. Excess kurtosis and left skewness of earnings growth in Norway areremarkably similar to the US.
2. Large earnings swings: wage&hours but more so for higher RE
• Wages are important for skewness and kurtosis of earningsgrowth for higher RE
3. Spousal inc. reduces the variance and skewness of earnings growth
• mainly through second earner effect (no behavior change)
4. Taxes/transfers provide insurance against tail shocks—more for lowRE groups
5. Even the disposable income growth exhibits long tails, less so thenearnings.
S. Ozkan (Toronto) Conclusion 37 / 37
![Page 54: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/54.jpg)
Summary of Findings
1. Excess kurtosis and left skewness of earnings growth in Norway areremarkably similar to the US.
2. Large earnings swings: wage&hours but more so for higher RE
• Wages are important for skewness and kurtosis of earningsgrowth for higher RE
3. Spousal inc. reduces the variance and skewness of earnings growth
• mainly through second earner effect (no behavior change)
4. Taxes/transfers provide insurance against tail shocks—more for lowRE groups
5. Even the disposable income growth exhibits long tails, less so thenearnings.
S. Ozkan (Toronto) Conclusion 37 / 37
![Page 55: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/55.jpg)
Summary of Findings
1. Excess kurtosis and left skewness of earnings growth in Norway areremarkably similar to the US.
2. Large earnings swings: wage&hours but more so for higher RE
• Wages are important for skewness and kurtosis of earningsgrowth for higher RE
3. Spousal inc. reduces the variance and skewness of earnings growth
• mainly through second earner effect (no behavior change)
4. Taxes/transfers provide insurance against tail shocks—more for lowRE groups
5. Even the disposable income growth exhibits long tails, less so thenearnings.
S. Ozkan (Toronto) Conclusion 37 / 37
![Page 56: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/56.jpg)
Summary of Findings
1. Excess kurtosis and left skewness of earnings growth in Norway areremarkably similar to the US.
2. Large earnings swings: wage&hours but more so for higher RE
• Wages are important for skewness and kurtosis of earningsgrowth for higher RE
3. Spousal inc. reduces the variance and skewness of earnings growth
• mainly through second earner effect (no behavior change)
4. Taxes/transfers provide insurance against tail shocks—more for lowRE groups
5. Even the disposable income growth exhibits long tails, less so thenearnings.
S. Ozkan (Toronto) Conclusion 37 / 37
![Page 57: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/57.jpg)
Summary of Findings
1. Excess kurtosis and left skewness of earnings growth in Norway areremarkably similar to the US.
2. Large earnings swings: wage&hours but more so for higher RE
• Wages are important for skewness and kurtosis of earningsgrowth for higher RE
3. Spousal inc. reduces the variance and skewness of earnings growth
• mainly through second earner effect (no behavior change)
4. Taxes/transfers provide insurance against tail shocks—more for lowRE groups
5. Even the disposable income growth exhibits long tails, less so thenearnings.
S. Ozkan (Toronto) Conclusion 37 / 37
![Page 58: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/58.jpg)
September Sample
• Norway also surveys employers in September for more precisemeasure of hours.
• An input to the annual wage negotiations between the unions andthe employers,
• For a sample of the largest private sector employers within eachindustry,
• covering approximately 52% of firms and 70% of employeesworking in the private sector.
• Sample: Workers employed in a big firm two Septembers in a row andthe past three Septs in the last 5 years.
• no extensive margin in September
• Hourly wage=regular monthly wage/regular monthly hours
• thus overtime and bonuses are not included Go back!
![Page 59: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/59.jpg)
Variance of yt+1 − yt for Prime Age Male
September Sample
2 4 6 8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
WAGE
HOURLY WAGE
HOURS
Employment Register
2 4 6 8 10
0.05
0.1
0.15
0.2
0.25
0.3
0.35Annual Wage
Hourly Wage
Annual Hours
• Variance of annual and hourly wages are much smaller in Septembersample than Employment register.
• No overtime and bonuses are not included.
![Page 60: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/60.jpg)
Skewness of yt+1 − yt for Prime Age Male
September Sample
2 4 6 8 10
-2.5
-2
-1.5
-1
-0.5
0
0.5 WAGE
HOURLY WAGE
HOURS
Employment Register
2 4 6 8 10
-1.5
-1
-0.5
0
Annual Wage
Hourly Wage
Annual Hours
• Both samples display left skewness for hours growth and not so muchfor hourly wage growth.
![Page 61: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/61.jpg)
Kurtosis of yt+1 − yt for Prime Age Male
September Sample
2 4 6 8 10
20
30
40
50
60
70
80
90
100
WAGE
HOURLY WAGE
HOURS
Employment Register
2 4 6 8 10
20
30
40
50
60
Annual Wage
Hourly Wage
Annual Hours
• September sample displays more excess kurtosis for both hours andhourly wage growth then employment register.
![Page 62: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/62.jpg)
Higher Order Moments in September Sample
Variance
2 4 6 8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
WAGE
HOURLY WAGE
HOURS
Skewness
2 4 6 8 10
-2.5
-2
-1.5
-1
-0.5
0
0.5 WAGE
HOURLY WAGE
HOURS
Kurtosis
2 4 6 8 10
20
30
40
50
60
70
80
90
100
WAGE
HOURLY WAGE
HOURS
• Variance of annual and hourly wages are much smaller in Septembersample than Employment register.
• Both samples display left skewness and excess kurtosis for both hoursand hourly wage growth.
![Page 63: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/63.jpg)
Hourly Wage vs Annual Hours: 3rd and 7th RE Groups
3rd RE Decile
-0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Annual Earnings
Hourly Wage
Annual Hours
7th RE Decile
-0.6 -0.4 -0.2 0 0.2 0.4-0.6
-0.4
-0.2
0
0.2
0.4
Annual Earnings
Hourly Wage
Annual Hours
• Hourly wage growth plays a more important role for higher RE groups.Go back!
![Page 64: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/64.jpg)
Hourly Wage vs Annual Hours: Average of Log Change
1st RE Decile
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
Annual Earnings
Hourly Wage
Annual Hours
5th RE Decile
-0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Annual Earnings
Hourly Wage
Annual Hours
10th RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Annual Earnings
Hourly Wage
Annual Hours
• Hourly wage growth plays a more important role lower RE groups.
• Wage changes drive earnings growth top RE groups. Go back!
![Page 65: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/65.jpg)
Hourly Wage vs Annual Hours
1-Year Moment Decomposition
![Page 66: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/66.jpg)
Variance of yt+1 − yt for Prime Age Male
Variance
2 4 6 8 10
0.02
0.04
0.06
0.08
0.1
0.12Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10
0
0.02
0.04
0.06
0.08
0.1
0.12Annual Earnings
Hourly Wage
Annual Hours
Co-Variance
• Hourly wage is more volatile than hours above the median.
• Similar to the PSID (Heathcote et al. (2014)). 5-Year Growth Variance
![Page 67: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/67.jpg)
Skewness of yt+1 − yt for Prime Age Male
Skewness
2 4 6 8 10
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4 Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3 Annual Earnings
Hourly Wage
Annual Hours
Co-Skewness
• Hours and wagre growth are left skewed.
• For top earners wage growth is key for left skewness of earningsgrowth. 5-Year Growth Skewness
![Page 68: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/68.jpg)
Kurtosis of yt+1 − yt for Prime Age Male
Kurtosis
2 4 6 8 10
15
20
25
30
35
40
45
50
Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10
5
10
15
20
25
30
35
40
45
Annual Earnings
Hourly Wage
Annual Hours
Co-Kurtosis
• Hourly wage and hours growth are both leptokurtic similar to earningsgrowth.
• Excess kurtosis due to hourly wage dominates the hours above themedian. 5-Year Growth Kurtosis
![Page 69: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/69.jpg)
Hourly Wage vs Annual Hours
Gaussian Decomposition: 5-Year Growth
![Page 70: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/70.jpg)
Skewness of yt+5 − yt for Prime Age Male
Skewness
2 4 6 8 10
-2
-1.5
-1
-0.5
0 Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 10-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
Annual Earnings
Gaussian Hourly Wage
Gaussian Annual Hours
• Hours and wagre growth are left skewed.
• For top earners wage growth is key for left skewness of earningsgrowth. 5-Year Growth Skewness
![Page 71: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/71.jpg)
Kurtosis of yt+5 − yt for Prime Age Male
Kurtosis
2 4 6 8 10
10
15
20
25
30
35
Annual Earnings
Hourly Wage
Annual Hours
Decomposition
2 4 6 8 105
10
15
20
Annual Earnings
Gaussian Hourly Wage
Gaussian Annual Hours
• Hourly wage and hours growth are both leptokurtic similar to earningsgrowth.
• Excess kurtosis due to hourly wage dominates the hours above themedian. 5-Year Growth Kurtosis
![Page 72: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/72.jpg)
Hourly Wage vs Annual Hours
Spousal Insurance After 5 Years
![Page 73: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/73.jpg)
Spousal Insurance, 5 Years Later
Bottom RE Decile
-1 -0.5 0 0.5 1 1.5-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2Spouse Annual Earnings
Spouse Hours
Top RE Decile
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2Spouse Annual Earnings
Spouse Hours
• Almost no change in behavior of spouse.
• For higher RE groups husband and wife’s earnings changes arepositively correlated.
![Page 74: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/74.jpg)
References i
ARELLANO, M., BLUNDELL, R. and BONHOMME, S. (2017). Earningsand consumption dynamics: a nonlinear panel data framework.Econometrica, 85 (3), 693–734.
DE NARDI, M., FELLA, G. and PARDO, G. P. (2016). The Implications ofRicher Earnings Dynamics for Consumption, Wealth, and Welfare.Working paper 21917, National Bureau of Economic Research.
GOLOSOV, M., TROSHKIN, M. and TSYVINSKI, A. (2016). Redistributionand social insurance. American Economic Review, 106 (2), 359–386.
GUVENEN, F., KARAHAN, F., OZKAN, S. and SONG, J. (2015). What DoData on Millions of U.S. Workers Say About Labor Income Risk?Working Paper 20913, National Bureau of Economic Research.
![Page 75: Dissecting Idiosyncratic Income Risk · Dissecting Idiosyncratic Income Risk E Halvorsen H Holter S Ozkan K Storesletten Statistics Norway U of Oslo U of Toronto U of Oslo XIII REDg](https://reader034.fdocuments.in/reader034/viewer/2022050503/5f9527f636caaa6c8f6f6f82/html5/thumbnails/75.jpg)
References ii
—, OZKAN, S. and SONG, J. (2014). The Nature of CountercyclicalIncome Risk. Journal of Political Economy, 122 (3), 621–660.
HEATHCOTE, J., STORESLETTEN, K. and VIOLANTE, G. L. (2014).Consumption and labor supply with partial insurance: An analyticalframework. American Economic Review.
KAPLAN, G., MOLL, B. and VIOLANTE, G. L. (2016). Monetary PolicyAccording to HANK. Working paper, Princeton University.
MCKAY, A. (2014). Time-Varying Idiosyncratic Risk and AggregateConsumption Dynamics. Working paper, Boston University.
WANG, C., WANG, N. and YANG, J. (2016). Optimal consumption andsavings with stochastic income and recursive utility. Journal ofEconomic Theory.