Assignment No.

1 Paper
Review
Cities and the growth of wages among young
workers: Evidence from the NLSY

Paper Review Cities and the growth of

wages among young workers:
Evidence from the NLSY
Main idea
The workers are more productive in cities. It is found that the period of time in which workers
typically see their wages grow the most (i.e. the first 10 years of a career) is also a period of
frequent job changes.
Similarly, one could argue that larger between-job wage changes in cities may emerge from
faster learning if workers continue to learn as they make the transition from one employment
position to the next.
Nevertheless, faster within-job wage growth can be viewed as a necessary outcome of any theory
which rests upon a learning mechanism.

Data Source
Author used the geocoded version of the National Longitudinal Survey of Youth 1979 Cohort
(NLSY79) which provided a weekly labor force history for a sample of more than 12,000 men
and women who were between the ages of 14 and 21 as of December 31, 1978. Several workers
were found in more than one job at a time.
Author then limited the sample to the 3003 male respondents from the cross sectional part of the
survey. He did this to avoid labor force participation issues.
The final sample he took included 1258 male workers who held a total of 5150 jobs between
1978 and 1994. These 1258 males are then classified according to local labor markets.
3 sources are used to describe these local markets;
The Census Bureaus Population Estimates Program (set reports estimates of total
resident population for each county in the US for each year between 1978 and 1994.)
The USA Counties 1998 on CD-ROM (US Bureau of the Census) ( information on
county-level land area)
County Business Patterns (CBP) files for the years 1978 to 1994. (Data on total
employment in each county for industries at the four-digit (SIC) level)

Methodology Used
The author the above 1258 male workers data to calculate their personal characteristics
(experience and marital status), overall wage growth, the within-components (using their 5150
jobs) and between-job wage growth (using 3888 job changes).

Local market characteristics are given by the averages for each of the 385 local markets
identified in the sample. Mean is calculated and minimum and maximum values are used.

Results
2 criterions were used to calculate within and between job growths.
Wage growth is decomposed into the sum of two parts:
1. Associated with the growth of wages on (or within) particular jobs
2. Due to job changes.
The data reveal a set of initial and final (log) wages {wj,istart, wj,iend} as well as starting and
stopping times {tj,istart, tj,iend} for jobs j = 1, 2, . . . , Ji for each worker i.
The wage growth that workers experience within the jobs they hold (W Gi ) amounts to roughly
2.6 percent per year, while that due to job changes (B Gi ) is approximately 1.9 percent per year.
Between the two sets of results, the estimated associations between the three market size
variables and wage growth are somewhat larger in the between-job regressions

Conclusion
Workers in large urban areas tend to earn more, on average, than similar workers who live in
smaller places as evident from the paper.
Significance of the estimates vary depending on the particular sample of workers used. Wage
growth is achieved through job changes rather than from growth on-the-job.
Yet, workers in large markets do not tend to see greater wage gains experienced on-the-job than
workers in small markets.
Interpreting faster within-job wage growth as a necessary implication of a learning mechanism
and faster between-job wage growth as a necessary implication of a matching mechanism, these
findings provide greater support for the latter explanation for agglomeration economies