Increasing Inequality in Long‐Term Earnings: A Tale of Educational Upgrading and Changing Employment Patterns

• Published date: 01 September 2022
• Author: Matthias Seckler
• Date: 01 September 2022
• DOI: http://doi.org/10.1111/roiw.12511

© 2021 The Authors. Review of Income and Wealth published by John Wiley & Sons Ltd on behalf of

International Association for Research in Income and Wealth.

617

INCREASING INEQUALITY IN LONG- TERM EARNINGS: A TALE

OF EDUCATIONAL UPGRADING AND CHANGING EMPLOYMENT

PATTERNS

by Matthias seckler*

University of Tübingen

This paper provides a detailed decomposition analysis of rising long- ter m earnings inequality among

West German men born between the years 1955 and 1974 based on high- quality administrative data.

Educational upgrading is identified as a leading factor behind increasing inequality in the upper part

of the long- term earnings distribution. The study also reveals a substantial shift from full- to part- time

employment and shows this to be an important factor in explaining rising inequality in the lower part.

This effect seems to be quantitatively more important than the increasing incidence of non- employment

for the studied cohorts. Overall, increasing inequality in long- term earnings can primarily be attrib-

uted to an increasing inequality in average yearly earnings during times of employment as opposed to

changes in the total years of employment. The analysis also reveals similarities with the development in

the US by documenting a stagnation in long- ter m earnings among the cohorts studied.

JEL Codes: C14, D31, D33

Keywords: income inequality, long- term earnings inequality, RIF decomposition

1. introduction

Growing wage and earnings inequality around the world has caused an

increasing interest in the topic among both policymakers and academics. The latter

have so far mainly focused on the increase in cross- sectional inequality over time as

documented in a vast literature (see Acemoglu and Autor 2011, for a general over-

view, and Dustmann etal., 2009, for the German case). Surprisingly, relatively little

is known about how this increasing cross- sectional earnings inequality has affected

the evolution of individual long- term and lifetime earnings across different birth

cohorts. From a purely cross- sectional perspective, which usually compares earn-

ings distributions at different points in time, cohort differences are usually non-

distinguishable from life cycle trends. For example, when comparing the German

*Note: I would like to thank the two anonymous referees, the editor Conchita D’Ambrosio, as well as

Martin Biewen, Bernd Fitzenberger, Uta Schönberg, and participants at the IAAE 2019 Annual Conference,

the 32nd Annual Conference of the European Society for Population Economics (ESPE), and the 2018

Summer School of the DFG priority program 1764 for their helpful comments and discussions. The data

basis of this paper is the weakly anonymous Sample of Integrated Labour Market Biographies (SIAB)

1975– 2014. The data were accessed on- site at the Research Data Centre (FDZ) of the Federal Employment

Agency (BA) at the Institute for Employment Research (IAB) and via remote data access. Financial sup-

port through DFG priority program 1764 and DFG project BI767/3- 1 is gratefully acknowledged.

*Correspondence to: Matthias Seckler, School of Business and Economics, University of

Tübingen, Mohlstr. 36, 72074 Tübingen, Germany (matthias.seckler@uni-tuebingen.de).

Review of Income and Wealth

Series 68, Number 3, September 2022

DOI: 10.1111/roiw.12511

This is an open access article under the terms of the Creative Commons Attribution License, which

permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Review of Income and Wealth, Series 68, Number 3, September 2022

618

© 2021 The Authors. Review of Income and Wealth published by John Wiley & Sons Ltd on behalf of

International Association for Research in Income and Wealth.

earnings distribution of the early 1990s with the one two decades later, it remains

unclear to what extent the standard of living of later cohorts differs from that of

their predecessors. This is a consequence of the fact that observable differences in

cross- sectional earnings are the result of individuals being observed at different

points of their career. Moreover, studying long- term earnings from a cohort per-

spective is likely to be more informative with regard to an individual’s or cohort’s

standard of living, which is ultimately determined by lifetime earnings rather than

by earnings at a certain point in time.

Recent studies by Bönke etal. (2015a) and Guevenen etal. (2017) document a

dramatic increase in long- term and lifetime earnings inequality for both Germany

and the US among men in later birth cohorts. Though being an ongoing debate,

the previous literature has identified different channels underlying the increase

in cross- sectional inequality, most prominently skill- biased technological change

(SBTC), demographical and institutional factors, as well as internationalization

and changes in individual employment biographies. A more comprehensive discus-

sion of these channels is provided in Section 2 of this paper. At the same time, it

remains unclear to what extent these factors are also responsible for the increasing

inequality in long- term and lifetime earnings. This paper intends to shed light on this

blind spot by disentangling the increasing inequality in long- term earnings using

high- quality administrative employment data for Germany. Methodologically, the

paper uses state- of- the- art recentered influence function (RIF) decomposition

techniques based on unconditional quantile regressions introduced in the seminal

contributions by Firpo etal. (2009, 2018).

This paper makes the following contributions to the literature. First, to the

best of my knowledge, this is the first study providing a detailed decomposition

analysis aimed at explaining the rising inequality in long- term earnings. It reveals

that much of the rising inequality at the top of the distribution is associated with

educational upgrading, whereas changing employment patterns are associated with

rising inequality at the bottom. Second, the results confirm previous findings by

Bönke etal. (2015a) who documented a sharp rise in long- term and lifetime earn-

ings inequality in Germany based on a different database. Going a step further, this

paper also shows that German men born between the years 1955 and 1974 did not

only face a higher level of inequality, but also equally suffered from a stagnation in

long- term earnings over a major part of their career. This paper also provides first

evidence that these trends tend to accelerate for the youngest cohorts.

The rest of this paper is structured as follows: Section 2 summarizes the

related literature. Sections 3 and 4 describe the data and the econometric method.

Section 5 presents the decomposition results. Section 6 concludes with a discussion

of the major findings.

2. related literature

This section provides an overview on the most relevant literature for this

paper. Most importantly, the study directly adds to the literature on the evolution

of individual long- ter m and lifetime earnings inequality. Using data for the US,

an important contribution by Bowlus and Robin (2004) finds that inequality in

Review of Income and Wealth, Series 68, Number 3, September 2022

619

© 2021 The Authors. Review of Income and Wealth published by John Wiley & Sons Ltd on behalf of

International Association for Research in Income and Wealth.

cross- sectional and lifetime earnings appears to follow a similar pattern over time.

Moreover, they show that the level of inequality in lifetime earnings is substan-

tially lower than inequality in cross- sectional earnings because of earnings mobil-

ity among young workers. However, changes in earnings mobility are not identified

as an important factor in explaining the rising dispersion in lifetime earnings. As

the study builds on a relatively short panel, the used measures of lifetime earn-

ings are simulated based on estimates for different parameters (job destruction/

re- employment rates, promotion/demotion rates). Kopczuk etal. (2010) provide

evidence for increasing inequality in male long- term earnings, especially for the

US baby- boomers born after 1945. This trend is found in all stages of the career,

with the level of inequality being generally higher in later episodes of the working

life. In a more recent contribution, Guevenen etal. (2017) document both a sub-

stantial decline in median lifetime earnings of the US men born between the years

1942 and 1958 (after observing gains in earlier cohorts) and a long- run trend of

increasing inequality within male cohorts. The authors conclude that the observed

changes are mostly because of differences in early career earnings across cohorts.

Importantly, they show that later cohorts suffered from earnings losses at young

age that were not compensated by a higher future earnings growth. In fact, the

study finds that women realized substantial gains in lifetime earnings (starting

from a very low level) across the study period which, however, only partly offset the

losses suffered by men.

In a seminal contribution for Germany, Bönke etal. (2015a) documented a dra-

matic increase in long- ter m and lifetime earnings inequality based on an Insurance

Account Sample (Versicherungskontenstichprobe) containing West German men

born between the years 1935 and 1969. The authors resort to the concept of up-

to- age X earnings (UAX) as a measure for individual long- term earnings, which

is defined as the present value of all earnings before reaching a certain age. By

imputing earnings for periods of un- and non- employment, they show that parts

of the increasing dispersion in lifetime earnings at the bottom can be explained by

differential unemployment patterns. Moreover, they establish two other results that

are important for the subsequent analysis. First, they show that earnings mobility,

which is high at the beginning of the working life, mostly vanishes after age 40.

Second, they conclude that the evolution of inequality in lifetime earnings most

likely reflects the development in long- term earnings up to age 40. Following this

argument, the subsequent analysis focuses on earnings up to age 40, which does

not only offer important insights into changes in individual long- term earnings for

a major part of the career, but can likely be generalized to inequality in lifetime

earnings (see Bönke etal., 2015a, p. 186). Another advantage of this approach is

to obtain new evidence on very recent cohorts. In a further contribution, Bönke

et al. (2015b) provide evidence for an increase in the transitory component for

younger workers in the 1970s and a related increase in short- term earnings risk.

This paper intends to directly add to these previous findings by trying to pin down

the aforementioned increase in lifetime earnings across cohorts to different explan-

atory factors.

In this aspect, the present study connects to a vast literature trying to explain

the well- documented increase in cross- sectional inequality during the past decades

as described by various authors (see, for the German case, Dustmann etal., 2009;