We use cookies on our website to give you the best online experience. Please know that if you continue to browse on our site, you agree to this use. You can always block or disable cookies using your browser settings. To find out more, please review our privacy policy.

COVID-19 RESOURCES AND INFORMATION: See the Atlanta Fed's list of publications, information, and resources; listen to our Pandemic Response webinar series.


Policy Hub: Macroblog provides concise commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues for a broad audience.

Authors for Policy Hub: Macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.

Comment Standards:
Comments are moderated and will not appear until the moderator has approved them.

Please submit appropriate comments. Inappropriate comments include content that is abusive, harassing, or threatening; obscene, vulgar, or profane; an attack of a personal nature; or overtly political.

In addition, no off-topic remarks or spam is permitted.

July 15, 2019

Making Analysis of the Current Population Survey Easier

Speaking from experience, research projects often require many grueling hours of deciphering obtuse data dictionaries, recoding variable definitions to be consistent, and checking for data errors. Inevitably, you miss something, and you can only hope that it does not change your results when it's time to publish the results. It would be far less difficult if data sets came prebuilt with time-consistent variable definitions and a guidebook that makes the data relatively easy to use. Not only would research projects be more efficient, but also the research would be easier to replicate and extend.

To this end, we have worked closely with our friends at the Kansas City Fed's Center for the Advancement of Data and Research in Economics (CADRE) to produce what we call a harmonized variable and longitudinally matched (HVLM) data set. This particular data set uses the basic monthly Current Population Survey (CPS) data published by the U.S. Census Bureau and the Bureau of Labor Statistics. The HVLM data set underlies products such as the Atlanta Fed's Wage Growth Tracker and the various tools on the Atlanta Fed's Labor Force Participation Dynamics web page.

You may be wondering how this data set is different from the basic monthly CPS data available at IPUMS. Like the IPUMS-CPS data, the HVLM-CPS data set uses consistent variable names and includes identifiers for longitudinally linking individuals and households over time. Unlike the IPUMS-CPS data, the HVLM-CPS also has time-consistent variable definitions. For example, the top-coded values for the age variable in the IPUMS-CPS is not the same in all years, whereas the HVLM-CPS age variable is consistently coded by using the most restrictive age top-code. As another example, the number of race categories is not the same in every year in the IPUMS-CPS (having increased from 3 to 26), while the race variable in the HLVM-CPS data set is consistently coded by using the original three race categories. Applying these types of restrictions means that the resulting data set can be more readily used to make comparisons over time.

The screenshot below shows how accessible the HVLM-CPS data are. For a visual of each variable over time, click on Charts at the top to see a PDF file of time-series charts. Code Book is an Excel file containing the details of how each variable has been coded. You can see in the screenshot how each variable ends with two numbers. These two numbers correspond to the first year that variable is available. For example, mlr76 is coded with consistent values (1 = employed, 2 = unemployed and 3 = not in labor force) from 1976 until today. The Data File is a Stata (.dta) format file with variable labels already attached. For users wishing to use the panel structure of the CPS survey, lags of many variables are provided on the data set already—for example, mlr76_tm12 is an individual's labor force status from 12 months ago).

screenshot of CPS survey data

Clicking on the c icon under Code Book opens a screen with the values of the corresponding variable. The screenshot shows lfdetail94 and nlfdetail94 as examples. The first variable, lfdetail94, contains a large amount of detail on those engaged in the labor market, while nlfdetail94 contains detailed categories for those not engaged in the labor market.

screenshot of CPS survey data

screenshot of CPS survey data

The HVLM-CPS data set is freely available to download and is updated within hours of when the CPS microdata are published, thanks to sophistical coding techniques and the fast processors at the Kansas City Fed. To access the data, go to the CADRE page (using Chrome or Firefox). At the top right, select Sign in, then Google Login. Then, under schema, select Harmonized Variable and Longitudinally Matched [Atlanta Federal Reserve] (1976–Present).

June 1, 2018

Part-Time Workers Are Less Likely to Get a Pay Raise

A recent FEDS Notes article summarized some interesting findings from the Board of Governors' 2017 Survey of Household Economics and Decisionmaking. One set of responses that caught my eye explored the connection between part-time employment and pay raises. The report estimates that about 70 percent of people working part-time did not get a pay increase over the past year (their pay stayed the same or went down). In contrast, only about 40 percent of full-time workers had no increase in pay.

This pattern is broadly consistent with what we see in the Atlanta Fed's Wage Growth Tracker data. As the following chart indicates, the population of part-time workers (who were also employed a year earlier) is generally less likely to get an increase in the hourly rate of pay than their full-time counterparts. Median wage growth for part-time workers has been lower than for full-time workers since 1998.

Wage Growth Tracker

This wage growth premium for full-time work is partly accounted for by the fact that the typical part-time and full-time worker are different along several dimensions. For example, a part-time worker is more likely to have a relatively low-skilled job, and wage growth tends to be lower for workers in low-skilled jobs.

As the chart shows, the wage growth gap widened considerably in the wake of the Great Recession. The share of workers who are in part-time jobs because of slack business conditions increased across industries and occupation skill levels, and median part-time wage growth ground to a halt.

While part-time wage growth has improved since then, the wage growth gap is still larger than it used to be. This larger gap appears to be attributable to a rise in the share of part-time employment in low-skilled jobs since the recession. In particular, relative to 2007, the share of part-time workers in the Wage Growth Tracker data in low-skilled jobs has increased by about 3 percentage points, whereas the share of full-time workers in low-skilled jobs has remained essentially unchanged. Note that what is happening here is that more part-time jobs are low skilled than before, and not the other way around. Low-skilled jobs are about as likely to be part-time now as they were before the recession.

How does this shift affect an assessment of the overall tightness of today's labor market? Looking at the chart, the answer is probably “not much.” As measured by the Wage Growth Tracker, median wage growth for both full-time and part-time workers has not been accelerating recently. If the labor market were very tight, then this is not what we would expect to see. The modest rise in average hourly earnings in the June 1 labor report for May 2018 to 2.7 percent year over year, even as the unemployment rate declined to an 18-year low, seems consistent with that view.  A reading on the Wage Growth Tracker for May should be available in about a week.

April 18, 2018

Hitting a Cyclical High: The Wage Growth Premium from Changing Jobs

The Atlanta Fed's Wage Growth Tracker rose 3.3 percent in March. While this increase is up from 2.9 percent in February, the 12-month average remained at 3.2 percent, a bit lower than the 3.5 percent average we observed a year earlier. The absence of upward momentum in the overall Tracker may be a signal that the labor market still has some head room, as suggested by participants at the last Federal Open market Committee (FOMC) meeting, who noted this in the meeting:

Regarding wage growth at the national level, several participants noted a modest increase, but most still described the pace of wage gains as moderate; a few participants cited this fact as suggesting that there was room for the labor market to strengthen somewhat further.

Although wages haven't been rising faster for the median individual, they have been for those who switch jobs. This distinction is important because the wage growth of job-switchers tends to be a better cyclical indicator than overall wage growth. In particular, the median wage growth of people who change industry or occupation tends to rise more rapidly as the labor market tightens. To illustrate, the orange line in the following chart shows the median 12-month wage growth for workers in the Wage Growth Tracker data who change industry (across manufacturing, construction, retail, etc.), and the green line depicts the wage growth of those who remained in the same industry.

As the chart indicates, changing industry when unemployment is high tends to result in a wage growth penalty relative to those who remain employed in the same industry. But when the unemployment rate is low, voluntary quits rise and workers who change industries tend to experience higher wage growth than those who stay.

Currently, the wage growth premium associated with switching employment to a different industry is around 1.5 percentage points and growing. For those who are tempted to infer that the softness in the Wage Growth Tracker might signal an impending labor market slowdown, the wage growth performance for those changing jobs suggests the opposite: the labor market is continuing to gradually tighten.

February 28, 2018

Weighting the Wage Growth Tracker

The Atlanta Fed's Wage Growth Tracker (WGT) has shown its usefulness as an indicator of labor market conditions, producing a better-fitting Phillips curve than other measures of wage growth. So we were understandably surprised to see the WGT decline from 3.5 percent in 2016 to 3.2 percent in 2017, even as the unemployment rate moved lower from 4.9 to 4.4 percent.

This unexpected disconnect between the WGT and the unemployment rate naturally led us to wonder if it was a consequence of the way the WGT is constructed. Essentially, the WGT is the median of an unweighted sample of individual wage growth observations. This sample is quite large, but it does not perfectly represent the population of wage and salary earners.

Importantly, the WGT sample has too few young workers, because young workers are much more likely to be in and out of employment and hence less likely to have a wage observation in both the current and prior years. To examine the effect of this underrepresentation, we recomputed median wage growth after weighting the WGT sample to be consistent with the distribution of demographic and job characteristics of the workforce in each year. It turns out that this adjustment is important when the labor market is tight.

During periods of low unemployment, young people who stay employed tend to experience larger proportionate wage bumps than older workers. In 2017, for example, the weighted median is 40 basis points higher than the unweighted version. However, both the unweighted version (the gray line in the chart below) and the weighted version of the WGT (the blue line) declined by a similar amount from 2016 to 2017. The decline in the weighted median is also statistically significant (the p-value for the test is 0.07, indicating that the observed difference is unlikely to be due to chance).

Another issue that could affect comparisons of wage growth over time is the changing demographic characteristics of the workforce. In particular, we know that workers' wage growth tends to slow as they approach retirement age, and the fraction of older workers has increased markedly in recent years. To examine this trend, we re-computed the weighted median, but fixed the demographic and job characteristics of the workforce so they would look as they did in 1997.

Our 1997-fixed version shows that median wage growth in recent years would be a bit higher if not for the aging of the workforce (the dashed orange line in the chart below). Moreover, this demographic shift appears to explain some of the slowing in median wage growth from 2016 to 2017. Whereas the 1997-fixed median also slows over the year, the difference is not statistically significant (a test of the null hypothesis of no change in the 1997-fixed weighted median between 2016 and 2017 yielded a p-value of 0.38).

Long story short, our analysis suggests that median wage growth of the population of wage and salary earners is currently higher than the WGT would indicate, reflecting the strong wage gains young workers experience in a tight labor market. Moreover, the increasing share of older workers is acting to restrain median wage growth. Although the decline in median wage growth from 2016 to 2017 appears to be partly the result of the aging workforce, there still may be more to it than just that, and so we will continue to monitor the WGT and related measures closely in 2018 for signs of a pickup. We also want to note that with the release of the February wage data in mid-March, we will make a monthly version of the weighted WGT available.