RELATED ARTICLE "A Better Way: Productivity and Reorganization in the American Economy," Annual Report 2003

Free Enterprise
Measures of Productivity

Julia Kedrova looks at measures and limitations of overall business sector and industry productivity.

America’s 139 million workers churn out $11 trillion a year in goods and services, making them among the world's most productive. Most of the time, we look at one number as a gauge of that productivity—output per hour of work. It’s a key economic measure because societies grow wealthier and attain higher living standards only by becoming more productive. Firms compete and increase productivity by improving business processes, creating technological innovations, investing in useful business assets, educating employees, and trading for goods and services. Our economy becomes more productive through reorganizing capital goods, labor skills and new technology and trading with other countries. While most of us recognize productivity's importance, we're fuzzy on what the statistics capture and what they miss.

The Bureau of Labor Statistics (BLS)[1] reports the most widely cited number—labor productivity in the business sector. This index shows changes in the relative amount of output to hours of labor input. Output data originate from the gross domestic product (GDP) adjusted for inflation. Labor data come from labor hours adjusted for actual hours worked, not hours paid. The BLS reports business sector productivity quarterly and annually. Even though the index relates output to one input—labor time—it does not measure the specific contribution of labor, capital or any other factor of production. Rather, labor productivity reflects the joint effect of a number of interrelated influences such as changes in technology and capital investment, utilization of capacity, methods of production, skill and effort of the workforce, and managerial expertise.

While the new productivity reports catch our attention, we need to keep in mind the limitations of this measure. Since we don’t measure productivity directly, the numbers are only as good as the GDP and labor market data that underlie them. The key issues include:

• Our measure of inflation is overestimated because it doesn’t fully capture improvements in the quality of goods and services.[2] Overestimated inflation results in underestimated real GDP, yielding undervalued productivity numbers.
• Near economic turning points, statistics miss the pace of changes in output and employment. The data surveys struggle to account for business closings during economic slowdowns and new business creation during economic expansions. As a result, quarterly productivity numbers may reflect business-cycle changes rather than lasting trends.
• Revisions in GDP and employment introduce volatility into quarterly productivity, but it still shows the general trend. The annual number averages the quarters, so it’s less volatile.

The chart below illustrates that labor productivity tripled over the last 53 years. Recently, the economy experienced a productivity surge. From 1995 to 2003, productivity grew 3.2 percent a year on average, far above the historical trend of 2.3 percent.

The productivity index reflects the performance of the economy as a whole. The BLS also calculates annual productivity numbers for about half the industries in the business sector. For example, manufacturing, mining, wholesale and retail trade, and transportation have detailed productivity statistics. Due to lack of reliable information, the reports do not cover some of the more interesting industries such as health care, legal, financial, insurance and real estate services.

The BLS calculates industry labor productivity as output per hour, but it uses data other than GDP and hours worked. Labor input comes from the number of hours paid. In industries with clearly defined output, such as raw commodities, the physical quantities produced count as output. For instance, barrels of oil and cubic feet of gas serve as output data in calculating the oil and gas extraction productivity, which increased 44 percent from 1987 to 2001.

However, most industries produce non-uniform products, making output calculations complex. For example, the BLS reports that output per hour in the semiconductor industry rose an impressive 1,304 percent from 1987 to 2001. During this period, the variety, quality and prices of semiconductors changed substantially. Moreover, semiconductor manufacturing standards differ among firms. To account for such intricate details, the BLS has a method to normalize output. It arranges similar goods and services into detailed product groups with comparable characteristics. For each group, sales are divided by prices adjusted with the corresponding price deflator. Aggregated results become the output measure similar to physical quantity measure of output. The BLS adjusts this measure for changes in inventories, resale, intra-industry transfers, and changes in industry structure.

As with the overall business sector productivity, the industry numbers depend on availability and accuracy of underlying data. The BLS releases preliminary industry productivity numbers after two years (manufacturing after three years) and revises them a year later. When the BLS and other reporting agencies[3] modify survey samples and industries change classifications, a mismatch of output to corresponding labor input may occur. Changes in quality of products introduce another potential for errors, since they affect adjustments in sales in the output calculation. Revisions and lack of timely data result in imprecise measure of annual industry productivity. In fact, the faster the industry evolves and the less standardized its output, the more inaccurate its productivity statistics.

Therefore, we should use historical productivity numbers to study industry progress. Although the industries’ index values don’t relate to each other because they measure different events, we can compare productivity growth rates of different industries. Similarly, we can compare industries across countries. The table below shows the leading U.S. industries in productivity growth.

In conclusion, the limitations of productivity measures don’t cancel the message of increasing efficiency. Though some industries advance faster than others, their long-term productivity growth contributes to the general improvement in our living standards. In the meantime, the BLS reviews its methodology for better ways to measure productivity and plans to include more industries in the future productivity measures.

About EYI | Global Economy | U.S. Economy | Regional Economy | Free Enterprise | Money & Banking | Technology
Federal Reserve Bank of Dallas | FRB Dallas Publications