Businesses are always striving to minimize the cost of producing a given level of output. When designing a factory, for example, the inputs would include predicted labor and capital costs, energy costs, costs of intermediate products such as sheet steel and components purchased from vendors, and other costs such as local taxes and transportation. The drive to minimize costs ensures continuing pressure to increase productivity by applying new technology, negotiating lower prices with suppliers, and looking for ways to increase labor productivity.
But labor is not homogenous, so companies measure outputs of different types of labor separately using appropriate goals for each. In a car factory one could use the number of cars a painter finishes in a day as an appropriate metric for judging the productivity of painters. Amazon is famous for their employee metrics and the incentives, positive and negative, they use to encourage productivity.
While firms measure productivity to minimize costs, countries collect aggregate productivity data as one measure of how the economy is performing. Rather than directly dividing output (cars say) by labor hours, output is usually measured in dollars. In the United States, the Bureau of Labor Statistics publishes data on labor productivity at the sector and major industry level. To compare productivity numbers over time, output must be adjusted for inflation. If the price of cars has doubled in 20 years, $2 billion dollars’ worth of cars today are the same as $1 billion dollars’ worth of cars 20 years ago, and so the productivity numbers for the automotive industry in terms of dollar output of cars divided by hours of labor must be adjusted accordingly (in this case, halved). The best-known inflation rate index used to adjust current (aka nominal) dollars is the Consumer Price Index, or CPI, which is based on the cost of a “basket” of goods and services that a consumer might buy. If the same basket of goods and services costs 5% more this year than last, then inflation is said to be running at 5% and any dollar figures must be reduced by 5% to compare with the prior year. Using the CPI, and other inflation indexes such as the Producer Price Index, one can compare productivity over the years in output dollars adjusted for inflation divided by labor hours. The actual calculation of an inflation index is quite tricky since what was in the “basket” 20 years ago may not correspond well with what is available today. We’ll consider some of these details in Part II when we look at the data.
The aggregate measure of productivity we’ve just discussed lumps all labor together and measures output per hour of labor. When we looked at how a company measures productivity, it was quite different, output is determined by the combined efforts of different types of labor employing different types of machinery (or more generally capital) and raw or intermediate materials such as energy.
Countries often also produce a measure of productivity that includes other factors of production, not just labor. We mentioned this measure briefly before, it’s called total factor productivity, also referred to as multifactor productivity.
Total factor productivity (TFP) is computed by looking at the percent change in output of an industry or sector over the course of a year and comparing that to the percent changes in inputs including labor and capital and intermediate goods. Each input is weighed by how much of it is used (in dollars). So, for example, if an industry uses 60 million dollars’ worth of labor, 30 million dollars’ worth of capital, and 10 million dollars of materials (total = 100 million dollars), then the weights for each factor would be 0.6, 0.3, and 0.1 which adds to one. If the output of the industry increases by 10% but labor only goes up 5%, capital by 7% and materials by 10%, even without applying the weights to each input we can see that output went up more than the inputs. The extra output cannot be “explained” by the increases in labor, capital, and materials, and is thus said to be due to an increase in total factor productivity. This increase in productivity can be quantified: it is simply the percent change in output divided by the combined weighted change in inputs. If the output grows as fast as the inputs, then they both grow at the same rate and the ratio of output increase to input increase is one. If output grows faster than the inputs, productivity increases and TFP is greater than one.
Why would one want to compute total factor productivity? Simply, it is a better measure of “real” productivity growth. To take an example, let’s suppose that output in an industry grows ten percent but the amount of labor doesn’t change at all. Then calculated labor productivity also rises by ten percent (ten percent more output from the same amount of labor). If the amount of capital used also stayed the same in dollar terms, then total factor productivity also rises ten percent because we are now producing ten percent more output with the same inputs (neither labor nor capital used has gone up, but we’re getting ten percent more output). In this case we can safely say that productivity in the industry has gone up, probably because of improved technology. But suppose that the amount of capital used has gone up twenty percent while labor stayed the same. Then calculated total factor productivity does not go up ten percent, and the increase in labor productivity is likely due to the addition of more capital, not just technology improvements. Total factor productivity statistics make it possible to see what is happening more clearly.
The total aggregate output of a country is referred to as it’s “gross domestic product” or GDP[1]. GDP per capita indicates how much output the country produces in dollars per person, while GDP per worker is a measure of average labor productivity. Both can be used to compare the relative productivity and income of different countries and can also give us an idea of how actual incomes of groups of workers fare against the average.
Gross domestic product includes spending for consumption or investment by consumers, government, and businesses but doesn’t include intermediate goods because their cost is built into final sales. GDP also includes sales overseas but purchases from overseas have to be subtracted from total sales since they weren’t produced here. Since all these sales are income to someone in the country, GDP balances neatly to national income which is why “gross national product”, and “national income” can be used interchangeably and are in fact calculated separately as a cross check.
In the USA the GDP per worker in 2019 was $131,047. For comparison, GDP per worker in China was $31,416, in Germany $103,013, and in India $19,270. These numbers correspond to average productivity, or output, per worker[2].
We will make extensive use of GDP per capita and per worker numbers when comparing the economies of countries in the sections on world productivity and income distribution.
[1] GDP doesn’t include the cost of replacing machines and other capital as it wears out. GDP net of depreciation is called Net Domestic Product (NDP) and is a better measure of output and income that could be consumed. GDP in the USA was $21.4 billion in 2019 while NDP was $18 billion.
[2] “World Bank GDP (current US$).” n.d. Accessed December 15, 2021. https://data.worldbank.org/indicator/NY.GDP.MKTP.CD.