Federal Reserve Economic Data: Your trusted data source since 1991

The FRED® Blog

Vehicle miles traveled and transportation carbon emissions

Vehicle miles traveled (VMT) reflects the total distance traveled by cars and trucks during a given period. Prior to 2000, VMT in the US grew proportionally with real GDP, shown by the blue and red lines, respectively, in the FRED graph above. But this relationship appears to have weakened after 2000.

VMT statistics from the US Federal Highway Administration begin in 1970, and we give that year an index value of 100. Between then and 2005, VMT-per-capita also grew (shown by the green line), rising from about 450 monthly miles per person in 1970 to a high of about 840 monthly miles per person during 2003-2006. Since then, VMT-per-capita has stayed relatively stable, apart from a sharp decline during the COVID-19 pandemic. Over the same period, total VMT growth has declined, falling below real GDP growth.

So, how have vehicle emissions changed during this evolution in VMT? The purple line shows the growth of carbon emissions from the transportation sector relative to levels in 1970—again, with a starting index value of 100. Together, cars and trucks are responsible for the vast majority of carbon emission from the transportation sector—81% in 2021 (EPA, 2023a). While the share of electric vehicles is growing, most of the fleet is still combustion: For a conventional vehicle, emissions increase proportionally with driving volume. As a result, the growth in VMT been a main driver of the growth in transportation emissions since 1970.

However, transportation sector emissions have grown much more slowly than VMT growth alone might suggest, at a rate just slightly below that of VMT-per-capita. This has occurred, in part, because vehicle efficiency has improved. With technological innovations, conventional cars produced today tend to require less fuel and emit less carbon dioxide than cars of comparable weight and power produced five decades ago.

Average emissions efficiency of new cars produced is one way to measure changes in fleet efficiency over time: The average carbon dioxide (CO2) emissions per mile for a new car was almost 700 grams in 1975, the first year data on average new car efficiency is available. By 2021, average CO2 emissions per mile for a new car had fallen by about half, to 347 grams per mile. Most of those efficiency gains came during 1975-1985, when average new car emissions dropped sharply to just over 400 grams per mile, catalyzed in part by the consumer response to high gas prices. Emissions per mile declines are primarily due to these technological improvements in fuel efficiency.

Over the same period, the car market’s shift toward SUVs and trucks—which use more fuel per mile on average than smaller vehicle types—acted as a countervailing force, limiting the extent to which vehicle emissions fell (EPA, 2022).

To the extent electric cars become more common, vehicle emissions and VMT will continue to uncouple; and further increases in VMT may be less closely associated with transportation sector emissions (EPA, 2022).

Recent emission standards set by the EPA—including those set in 2023—focus specifically on the levels of CO2 emitted from vehicles (EPA, 2023b). Because there’s a lag between when standards for new car emissions are imposed and when those newer cars begin to dominate the fleet of vehicles overall, these newer standards do not yet show up on the FRED graph. But they may eventually lead to reductions in transportation emissions in the future.

How this graph was created: In FRED, search for and select the “Vehicle Miles Traveled” series. Use the “Edit Graph” panel to change the frequency of the series from monthly to annual using averaging as the aggregation method. With the “Add Line” tab, search for and select “Real Gross Domestic Product.” Use the “Add Line” tab again to search for and select “Vehicle Miles Traveled,” change the frequency from monthly to annual using averaging as the aggregation method. Within the same tab, search for and add the series “Population” and apply formula a/b. Use the “Add Line” tab again to search for and select “Transportation Carbon Dioxide Emissions, All Fuels for the United States.” Under the units option, select “Index” with date 1970-01-01 and click “Copy to all.” Use the date selector above your graph to choose 1970-01-01 as the start date.

Suggested by Marie Hogan and Mike Owyang.

How common is “moonlighting”?  

Moonlighting—simultaneously holding multiple jobs—is fascinating from a macro labor economics perspective: It adds jobs to the economy without increasing the actual level of employment. Fortunately, FRED provides data to help us understand how prevalent this group of workers is in the US economy.

The blue line in the FRED graph above plots the fraction of employed individuals simultaneously holding more than one job for each month. It shows that moonlighters make up a sizable fraction of the employed. From 1995 to the present, about 5-6% of employed persons have held multiple jobs in any given month. The proportion of moonlighters tends to drop at the onset of recessions and recovers afterward, with the starkest decrease occurring during the COVID-19 recession.

How common is moonlighting among men vs. women? The proportion of employed women who hold multiple jobs (green line) is now noticeably higher than the proportion of employed men who do (red line). This divergence became particularly clear after the 2001 recession, with differences of up to 1 percentage point. Since the COVID-19 recession, the proportion of women holding multiple jobs has exceeded its pre-pandemic level, while the proportion of men holding multiple jobs is similar to its pre-pandemic level.

How this graph was created: In FRED, search for “Multiple Job Holders as a Percent of Employed.” Click on “1 other format” (below the first search result) and choose “Monthly, Percent, Not Seasonally Adjusted” to be consistent with percentages by gender. From this graph, use the “Edit Graph” panel in the top right corner to select the “Add Line” tab. Search for and select the unemployment rate for men, i.e., “Multiple Job Holders as a Percent of Employed, Men.” Repeat for women.

Suggested by Serdar Birinci and Ngân Trân.

How binding is the federal minimum wage?

The real value of $7.25 per hour and the declining number of workers who make it

The federal minimum wage hasn’t changed since 2009. But in those 15 years, the real value of $7.25 per hour has changed considerably due to inflation, especially in the past few years.

The FRED graph above uses data from the Bureau of Labor Statistics to show the federal minimum wage adjusted for inflation using the consumer price index, or CPI.

Keep in mind that the CPI data here correspond to actual prices in 1982-84. Thus, the real federal minimum wage (the blue line) is measured relative to the prices of that period. But it is the visual here that really matters: The real federal minimum wage has never been as low as it is now, at least within the period shown in this graph. (If you move the date tracker below the graph, you’ll see that it was lower in 1947-49.)

Of course, the federal minimum wage is only impactful if the state-level minimum wage is not set higher. And most states do set a higher minimum wage. This FRED map identifies each state’s approach to the minimum wage, including some that do not set one at all.

Now, the lower a minimum wage is, the less it’s going to matter in raising overall wages. And the fewer employees who earn that minimum wage, the less it will affect the US economy. The line in red shows the proportion of US employees who earn the minimum wage or less. (There are exceptions for certain professions.) The red and blue lines do indeed follow a very similar downward path.

How this graph was created: Search FRED for “minimum wage.” Use the “Edit Graph” panel to add the series “CPI”; change the frequency to annual, taking averages; and apply formula a/b*100. Use the “Add Line” tab to search for and select the “paid at or below minimum wage” series. Use the “Format” tab to move the y-axis for one of the series to the right. Start the sample period in 1979.

Suggested by Christian Zimmermann.

Subscribe to the FRED newsletter

Follow us

Back to Top