Skip to main content

The FRED® Blog

Income differences across countries: Is the gap closing?

Why are some countries rich and others poor? This question has troubled economists for a long time. Probably the most well-known economic model to study this question is the Solow growth model, which predicts that income per capita of countries that share similar savings rates, capital depreciation, and population growth (among other characteristics) should converge over time to similar values. This convergence hypothesis means that, holding all else equal, countries with low income per capita should grow fast and eventually catch up with countries with high income per capita. FRED data can help us evaluate whether this convergence hypothesis holds over time.

We take the U.S. as our “benchmark” high-income country and analyze whether other countries are catching up with it. The graph above shows real GDP per capita in low- and middle-income countries as a ratio to real GDP per capita in the U.S. For example, if Brazil’s ratio moves closer to 1, that means Brazil’s GDP per capita is converging to U.S. GDP per capita. We see some clear evidence of convergence in three countries: South Korea (solid blue line), China (solid red line), and India (solid green line). GDP per capita in Brazil (dotted blue line) moved toward convergence until 1980, then started decreasing and leveled out. The dashed lines at the bottom, representing Mozambique and Kenya, do not seem to converge at all.

The graph below shows GDP per capita in high-income countries as a ratio with GDP per capita in the U.S.: Japan (dashed purple line) and France (solid blue line) show some stronger signs of convergence, moving from about 20% and 55% of U.S. GDP per capita, respectively, to about 75% of U.S. GDP per capita between 1950 and 2010. The United Kingdom (red line) also shows some signs of convergence to the U.S., while Germany (dotted blue line) and Canada (green line) are mostly flat over the sample period.

Overall, the evidence on Solow convergence is mixed. Some assumptions important for the theory may not hold, and the lack of convergence could be related to differences in savings rates, depreciation, and population growth across some of the countries we analyze here.

How this graph was created: In the FRED search bar, search for “purchasing power parity converted GDP per capita (Chain Series) for China.” Add this series to a graph and search for the same series for other desired countries within the “Add Data Series” tab underneath the graph. After adding all the low- and middle-income countries, go back to the “Add Data Series” tab and modify the existing series by adding GDP per capita for the U.S. Next, for each series, under the “Create your own data transformation” option, type “a/b” in the formula box to create the ratio. You can turn off the title and axis titles on the “Graph Settings” tab to keep the graph clean. Individual series color and line styles can be changed on the “Edit Data Series.”

Suggested by Maximiliano Dvorkin and Hannah Shell.

View on FRED, series used in this post: RGDPC2CNA625NUPN, RGDPCHBRA625NUPN, RGDPCHCAA625NUPN, RGDPCHDEA625NUPN, RGDPCHFRA625NUPN, RGDPCHGBA625NUPN, RGDPCHINA625NUPN, RGDPCHJPA625NUPN, RGDPCHKEA625NUPN, RGDPCHKRA625NUPN, RGDPCHMZA625NUPN, RGDPCHUSA625NUPN

The declining wage component in GDP

The graph above shows the share of GDP from the wages and salaries of employees, which has clearly been on a downward trend over several decades. This post isn’t about the reasons behind this decline, which would require analysis of (i) supplements to wages and salaries such as pensions and other benefits and (ii) proprietors’ income, which is earned by independent workers and business owners that compensates for labor and capital. What we are interested in is whether the decline has bottomed out.

Indeed, the share has been increasing for about two years now. Is this evidence enough to declare the trend has reversed? Well, that call is difficult. If you play with the graph by changing dates—for example, by ending the data in the year 2000 or 1987—you’d find a pretty similar situation in which the decline appears to have reversed. Yet, the share has continued to decline.

But is this time different? Visit this blog in a couple of years and we may have the answer.

How this graph was created: Search for “compensation of employees” and the series used in the graph should be among the first options. Note that a share of it in national income is also among the top options, but it has less current data. Once you have the graph for the series, add a series to the first line, not as a separate line. Then create a data transformation by applying the formula a/b.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: GDP, WASCUR

FRED remembers Bob Rasche

Rasche

We recently celebrated 25 years of FRED, and several posts have looked back at the origins and the originators of the St. Louis Fed’s data services. One person we haven’t mentioned yet in this blog is Bob Rasche. He was the St. Louis Fed’s director of research from January 1999 until 2009, when he was promoted to executive vice president and senior policy advisor. He retired in June 2011.

Bob, who was one of the most vocal and dedicated advocates of delivering high-quality data and information to the public, passed away Thursday, June 2. He is survived by his wife, Dottie, and children, Jeanette and Karl.

One of Bob’s legacies is that he expanded and enhanced FRED at a key moment in its history. He made a compelling case to the Bank’s president and its senior leaders, as well as leaders around the Federal Reserve System, that this mission of public service should continue and thrive. During Bob’s tenure, FRED grew in both size (from under 1,000 series to over 33,000) and recognition. FRED’s global presence and notoriety helped promote other efforts. In fact, Bob envisioned and nurtured a realm of St. Louis Fed data services, including FRED’s sibling sites, GeoFRED and ALFRED, and FRASER, the historical digital archive. He was sincere and energetic, and his process was legendary: He would arrive at work early in the morning, passionately describe the ideas that had occurred to him overnight, and a new project was born.

Bob hired and guided many of the people who are now integral to the development and success of the St. Louis Fed’s data services. The FRED and FRASER teams continue their mission with respect and gratitude for Bob Rasche’s leadership. It was an honor to work for him. It is inspiring to remember him.

A gauge on the minimum wage

FRED recently added more data about the federal minimum wage—specifically, who’s working for it. The top graph shows how many people are paid the federal minimum wage or less. A few points to keep in mind: Workers who receive tips can be paid less than the minimum, and states or localities can impose a minimum wage that’s higher than the federal minimum. So, the long-term decline we see here may reflect many things: states increasing their minimum wage, fewer workers in jobs that earn tips, or movements in the wage distribution. More-specific series and graphing options in FRED can provide some insight.

The middle graph shows the same numbers as above, but divided into education levels. We see that the share of the least-educated workers who receive the minimum wage is actually decreasing, while the share of the most-educated workers is increasing.

Why? The bottom graph shows absolute numbers instead of proportions, which may shed some light: The absolute numbers for the two lower levels of education are decreasing, but the absolute number for those with at least a bachelor’s degree is stable. The release table offers even more series, including one that shows those with a master’s or professional degree who are earning the minimum wage.

How these graphs were created: For the first graph, search for “federal minimum wage prevailing” and click on the series you want, which will be among the first choices. For the second graph, go to the release tables that show minimum wage workers by education, select the series you want, and click on “Add to Graph.” Start the sample in 2003, as some series are missing earlier data. Change the graph type to “Area” and set the “Stacking” setting to “Percent.” For the third graph, use the second graph but change the “Stacking” setting to “None.”

Suggested by Christian Zimmermann

View on FRED, series used in this post: BDAHC2, LHSDC2, SCADC2, T16OC2


Subscribe to the FRED newsletter


Follow us

Twitter logo Google Plus logo Facebook logo YouTube logo LinkedIn logo
Back to Top