Why This Counts: How much time do Americans spend working and engaging in other activities outside of work?

The U.S. Bureau of Labor Statistics is best known for our monthly job and inflation reports, but we also publish data on many other topics. My blog series, “Why This Counts,” explains why we conduct our surveys and how people can use BLS data at work and home. I hope this series will help bring BLS data to life—to transform a set of abstract statistics into a snapshot of life in America.

“Time is money,” said Benjamin Franklin in Advice to a Young Tradesman, Written by an Old One. One of the most precious resources each of us has is our time. Indeed, how we use our time can tell you a lot about our priorities, opportunities, and constraints.

Today I am happy to announce our annual news release and 2014 data for the American Time Use Survey (ATUS)—and that we’ve made it easier to compare this year’s results to previous years.

Unlike other BLS surveys that track employment, wages, and prices, the ATUS tracks a less conventional economic resource that we never have enough of: time. The survey compiles data on how much time Americans spend doing paid work, unpaid household work (such as taking care of children or doing household chores), and all the other activities that compose a typical day. With this information, economists, sociologists, and other researchers can examine the different time choices and tradeoffs that people make every day.

ATUS data are behind many of the facts you often see in the news, such as:

  • At what times of day do workers with different occupations work? (25% of workers in protective service occupations, such as firefighters and police, are at work at midnight.)
  • How many hours of sleep do people get at different ages? (For teenagers, 9.8 hours!)
  • How many hours per day do employed parents spend caring for their children? (See chart below.)


How do we get this key information? The ATUS is the only nationally representative, continuous survey that provides estimates of how, where, and with whom Americans spend their time. In 2014, more than 11,500 respondents completed the survey. The results represent the national population by gender, age, race and ethnicity, and other characteristics. We select the ATUS sample from households that have already responded to another BLS survey, the Current Population Survey. This enriches ATUS data because we already know the demographic characteristics and household composition and don’t have to ask for that information again. Instead, most of the ATUS interview involves asking respondents to report a time-use “diary” for the previous day.

In service to the rest of us, ATUS respondents voluntarily and carefully report every activity they did for the 24-hour period beginning at 4 a.m. on the day before the interview and ending at 4 a.m. on the day of the interview. So, dear reader, each of us owes the ATUS participants a huge thank you for their public service!

Economists from the U.S. Census Bureau and BLS put these responses together in a way that prevents anyone from identifying individual respondents. BLS then makes the data available to the public in the form of public-use data files, tables, charts, and news releases.

People often ask us how these estimates have changed over time. It’s now much easier for you to examine trends in time use because we have added ATUS historical data to the Labstat database on our BLS website. You can view thousands of ATUS time series going back to 2003, the first year of data collection. These data have always been available through the ATUS news release archive and tables. We are pleased to present them now in a more accessible and convenient format.

There are two big benefits to having ATUS data available from the Labstat database. First, all the historical data are now available in one place. You no longer have to combine data from different Excel or PDF documents to create a time series. Second, all estimates in Labstat have been created using the same statistical weighting method, which has been consistent since 2006 but changed slightly each year before that. This allows you to directly compare estimates over time.

We’ve selected a few data series that we think will be of particular interest. You can view these “Top Picks” at http://data.bls.gov/cgi-bin/surveymost?tu

Some examples are:

  • The percent of employed people who did some or all of their work from home on days they worked
  • The average hours per day mothers and fathers spent providing childcare, by age of their youngest child
  • The average hours per day spent in leisure, including watching TV, and socializing, on weekdays and weekend days

Many more data series are available through the Historical Tables extraction tool. These tables allow you to select one or many estimates from ATUS tables, specify a time range, and retrieve the historical data. I hope you will explore this new tool.

For help working with ATUS data and for more information on the survey, please visit the ATUS Labstat Tips webpage and the ATUS homepage.

Visualizing BLS Data to Improve Understanding

If a picture is worth a thousand words, what’s the value of a striking, cool chart or map of some BLS data? At the U.S. Bureau of Labor Statistics, we’re always thinking of better ways to help our users understand the information we produce. The global economy is complex, and the statistics to explain the economy can be complex too.

Data visualizations are one tool we use to present our data more clearly. What are data visualizations? They are any method of presenting numerical information visually—most commonly through charts and maps. Good data visualizations can improve understanding for all types of audiences, from students of all ages to experts with advanced degrees in economics, statistics, or other fields.

In recent years we’ve done more to include data visualizations in nearly all our publications. We have designed two of our publications to showcase data visualizations. One is The Economics Daily—or TED, as we call it. We publish a new edition of TED every business day, and we’ve done that since 1998. Each edition of TED typically includes a chart or map, sometimes two, with a few words to explain the data in the visualization.

Another publication geared toward data visualizations is Spotlight on Statistics. Spotlight tells a longer, more detailed story about a topic through a series of visualizations presented in a slideshow format. As with TED, Spotlight includes brief written analysis to explain more about the data.

Even our publications that feature mostly written analysis often include visualizations to tell a more complete story. Our flagship research journal, the Monthly Labor Review, has evolved a lot over its 100 years of publication to serve readers better; that evolution includes more and better data visualizations. Beyond the Numbers and BLS Reports often include visualizations as well.

We take pride in crafting our words carefully, but good data visualizations can complement the words. For example, during and after the Great Recession, the monthly Employment Situation news release has discussed the historically high levels of long-term unemployment. The number of long-term unemployed—those jobless 27 weeks or longer—has remained high years after the recession ended in June 2009. It’s one thing to read about long-term unemployment, but a good chart can tell the story even more clearly. long-term-unemployment

For an even broader perspective, we have a Spotlight on Statistics that examines long-term unemployment more fully.

Not only have we presented more data visualizations in recent years, but our visualizations also have gotten more sophisticated. A basic image can present information effectively. Take this simple map that shows the proportion of each state’s population age 16 or older that had a job in 2014. state-employment-population ratios

Now check out the interactive version of this map that we published in the March 9, 2015, edition of TED. When you hover over each state, more information pops up to show the state’s employment–population ratio in 2014 and how much it changed from 2013. When you hover over the items in the map legend, the states in each category light up more brightly to help you see the states with similar employment–population ratios. When you click on each state, you go to a webpage that provides even more information about the state’s labor market. Interactive features in our charts and maps give you the power to choose what information you want to see.

If you like the interactive features in our charts and maps, I think you’ll love the animation in some of our visualizations. Animation adds a time dimension to our data to let you see how measures change. For a great example of animation, see a TED we published last year that shows state unemployment rates before, during, and after the Great Recession.

The BLS website will feature even more data visualizations soon. Watch this space to learn more about them.

We share many of our data visualizations on Twitter, so follow us @BLS_gov. You also can sign up to receive email alerts for TED, Spotlight on Statistics, and our other publications.

And if you have created a great visualization of BLS data, please share it with us and the readers of this blog!

Why This Counts: Employer Costs for Wages and Benefits

The U.S. Bureau of Labor Statistics releases a wide range of economic statistics and analysis almost every day. Families, employers, policymakers, and others use these statistics to make better decisions. BLS data inform Americans and people around the world about employment, prices, compensation, productivity, worker safety, and a host of other topics. Today I am using my blog series, “Why This Counts,” to shed some light on the quarterly Employment Cost Index, a measure of changes in employer costs for wages and benefits.

The Employment Cost Index—the ECI—began in the mid-1970s by tracking changes in employer costs for wages in the private sector. Over time, the program has expanded to include the cost of about two dozen benefits and currently includes both private industry and state and local government workers.

We track a sample of employers and occupations selected to represent the U.S. economy. We return to these employers quarterly and ask them to update their information on the cost of wages and benefits. By tracking the same occupations over time, the ECI is free from changes in wage and benefit costs due to shifts in employment from one industry or occupation to another.

The ECI provides quarterly rates of change. The data are for March, June, September, and December, and we release the numbers a month later. Today we released the March 2015 data. I’m always impressed at how quickly we can collect and report on dozens of pieces of information from thousands of employers each quarter. We’re able to do this because of the excellent cooperation we have with employers; my thanks to those who provide data for all our programs. We’re also able to release the data quickly thanks to our regional office staff who work with employers and our national office staff who turn the raw data into useful numbers and publish the results.

As we reported today, total compensation costs (wages and benefits together) rose 0.7 percent for the quarter and 2.6 percent over the past year. These data are for all workers in private industry plus state and local government—a group we call civilian workers. Wages for civilian workers grew 0.7 percent over the quarter and 2.6 percent over the year; benefit costs rose 0.6 percent over the quarter and 2.7 percent over the year.

The cost of many benefits tracks closely with employee wages. For example, as wages increase, an employer’s cost to provide vacation benefits increases about the same rate. This is also true for 401(k) retirement plans; the employer cost for these plans—usually a match of employee contributions—is a percent of the employee’s pay and increases as pay increases.

In contrast, the cost of some benefits moves independent of wages. Health insurance is one notable example. For the 12 months ending in March, private employer costs for health insurance rose 2.5 percent. As the chart shows, the change in health insurance costs can drive the change in overall benefit costs.


The ECI provides evidence of the slow growth in wages since the 2007–2009 recession. From December 2002 to December 2008, private industry wages grew at an average rate of 2.9 percent per year. From December 2008 to December 2014, private industry wages grew at an average rate of 1.8 percent per year.

The ECI can provide an early warning of inflationary pressure in the economy. Rising compensation costs may lead firms to raise prices for their goods and services. In a preview of this week’s economic calendar, one media outlet said about the ECI, “It’s seen by some as perhaps the best gauge of whether labor costs are on the rise. This is of utmost importance to the market, since if labor costs are accelerating, it could spur the Fed to make monetary policy tighter.” In addition, employers look to the ECI when setting their budgets for pay raises; employers and workers look to the ECI during salary discussions or contract negotiations.

We also use the data on wages and benefits to estimate the cost in dollars and cents per hour worked that employers spend. We call this series Employer Costs for Employee Compensation. This series provides information on the share of compensation dollars that go toward wages versus benefits, and toward individual benefits. In December 2014, about 69 percent of the compensation dollar of private sector employers went toward wages, with the rest going toward benefits. The most expensive individual benefit for private industry employers was health insurance, costing employers $2.39 per hour worked in December 2014.

You can visit the Employment Cost Trends page of our website to get current and historical data and analysis and learn more about how we collect the data.

An Important Improvement in the Consumer Price Index

The Consumer Price Index (CPI) has measured price change in the U.S. economy for more than 100 years.

The index measures the average change over time in the prices paid by urban consumers for a market basket of consumer goods and services. It is one of the most widely used measures of inflation. Policymakers, government leaders, business executives, and others use the index as a guide in making and evaluating economic decisions.

As you might imagine, producing the CPI is a large and complex monthly undertaking that is subject to constant, careful review. We occasionally make minor changes to our process that are narrow in scope. However, we do not introduce any major changes without undertaking years of rigorous research and testing and informing stakeholders at each step of the process.

To underscore that point, we launched the first major improvement to the existing system in more than 25 years in February!

The new CPI estimation system changes are pretty technical. In short, the state-of-the-art processes we implemented give us key new flexibilities and efficiencies in how we calculate and measure price changes in the economy. It’s important to note the methodology underlying production of the CPI has not changed and the new system is not designed to produce a higher or lower estimate of price change.

As part of the redesign, we also eliminated paper completely in all review steps of producing the CPI. That’s not only good for the environment but also improves automation and accuracy in our work processes.

I’ve challenged my staff to get the best we can for the nation’s data dollar and to continually adapt our programs to meet the challenges of a changing economy. The redesigned CPI estimation system is a huge step toward those goals and provides us with the opportunity for more research and faster innovations in the future.

Measuring inflation is complicated but vitally important to support public and private decision making. Launching this estimation system was a huge undertaking, and I applaud the staff here at BLS who worked incredibly hard to make it a success.

Why This Counts: Tracking Labor Market Experience over a Lifetime

The U.S. Bureau of Labor Statistics is best known for our monthly job and inflation reports. We also publish data on many other topics, ranging from how Americans spend their time and money to workplace injuries and the growth of entrepreneurship. My blog series, “Why This Counts,” explains why we conduct our surveys and how people can use the data at work and home. I hope this series will take the mystery out of our data and make our work come to life for both new and advanced users.

Today I want to tell you about a fascinating group of surveys called the National Longitudinal Surveys. It’s especially timely to talk about these surveys for two reasons: 1) we published a news release this week with the latest data from one of the surveys, and 2) the program is one of the important legacies of former BLS Commissioner Janet Norwood, who passed away recently.

The National Longitudinal Surveys stand out because they are designed to answer key long-term questions about people’s paths through life. Most of our measures about the labor market and economy focus on current conditions. What’s the national unemployment rate? How rapidly is employment growing in California or North Dakota or Georgia? How many job openings are there in manufacturing? What are the trends in consumer prices for food, energy, clothing, and shelter? It’s important to have up-to-date answers for these and other economic questions. But some questions take longer to answer—years or even decades.

Some long-term questions we care about include: How many jobs do people hold over their lifetimes? How do earnings grow at different stages of workers’ careers? The surveys designed to answer these and other long-term questions are called “longitudinal” surveys. What’s that mean?

A longitudinal survey asks questions about the same people at different points in their lives. Longitudinal surveys are useful for studying changes that occur over long periods. These surveys are also useful for examining cause-and-effect relationships. For example, how do events that happened when a person was in high school affect labor market success as an adult? This week we published a new report that looks at the experiences of baby boomers from age 18 to age 48.

The survey follows a set of people born in the latter years of the post-World War II baby boom, 1957 to 1964, and living in the United States when the survey began in 1979. To answer my earlier questions—using just-released data—these baby boomers held an average of 11.7 jobs from age 18 to age 48. Their inflation-adjusted hourly earnings grew the most during their late teens and early twenties, and earnings generally grew faster for college graduates than for people with less education.

Real wage growth-final

The survey doesn’t just ask about labor market activity. It also asks about education, training, health, marriages and other relationships, children, use of government programs, juvenile crimes and arrests, drug and alcohol use, and much more. Why do we ask about these topics, some of which are pretty sensitive? In short, we’re trying to understand all the things that affect or are affected by labor market activity. That covers nearly every part of our lives.

Before this survey of baby boomers began in 1979, four other longitudinal surveys began in the 1960s of earlier generations. BLS began another survey in 1997 that represents people born in the years 1980 to 1984 and living in the United States at the start of the survey. We only still conduct the surveys of the two more recent generations, but we have learned so much from all the surveys. These surveys are some of the most analyzed in the social sciences. Researchers in economics, sociology, psychology, education, and health sciences have used the surveys to examine a broad range of topics. Here are just a few examples of what researchers have learned from the surveys:

  • Nobel Prize winner James Heckman and his colleagues found that noncognitive skills, such as motivation and perseverance, are as important to future labor market success as are skills such as reading and math.
  • People who obtain a GED or other exam-certified high school equivalent have labor market outcomes that are similar to those of high school dropouts, rather than to people who earn a regular high school diploma.
  • The labor market effects of a 4-year college degree are similar for those who start at a 4-year college and those who transfer from a 2-year college to a 4-year college.
  • Obesity is not only a health concern, but a labor market concern. Workers pay a price for obesity with lower wages and employment, and this price is higher for women than men.
  • Low birth weight is a better predictor than cognitive test scores of whether people either work or attend school at ages 24 to 27. Birth weight also is a better predictor of adult wages.

You can find information about thousands of other research studies in the National Longitudinal Surveys bibliography.

Although we learn a lot each time we update our monthly and quarterly data on employment, compensation, prices, and productivity, there is so much we could not learn without these longitudinal surveys.

This is all possible thanks to Janet Norwood—and to the people who have agreed to participate in the surveys for so long—so that we can understand people’s paths over time!