Measuring a dynamic economy: What should data users expect from the federal statistical system?

Federal economic data and statistics are essential for both public and private sector decisionmakers across the United States. They make it possible to monitor and understand the performance of the economy, craft public policy to effectively address challenges facing households and the nation, and make informed business and financial decisions. Their collective value to the users of these data—from policymakers to businesses to researchers—is immense.

Changing needs, and the need for changing data

At the same time, the needs of data users are evolving. Policymakers and businesses increasingly demand more timely, localized, and detailed information. Economic research continues to identify new relationships and concepts that are important for data to capture, and for statistical series to incorporate and reflect.

Most of all, economic data and statistics require constant innovation to keep pace with a dynamic and changing economy. Factors like the rise of artificial intelligence, gig work, digital assets, and increasingly complex sources of income and wealth can pose challenges for traditional economic data. Consider examples that arise across four key domains of economic data: employment, prices, income, and wealth:

Employment data: Understanding evolving labor markets

Federal employment statistics are among the most widely referenced economic indicators. These data—tracking labor market conditions, measuring job growth, calculating the unemployment rate, observing trends in and the distribution of wages across workers, and so on—are closely followed by policymakers, financial markets, researchers, voters, and the media.

But some important questions about today’s labor market are difficult to answer with existing data.

Consider the rise of artificial intelligence (AI), which some analysts believe may already be disrupting the labor market. Understanding and responding to AI’s labor market highlights key data needs. Which workers are most at risk of losing work? Which skills are becoming more or less in-demand? How are patterns of employment changing across occupations? 

Current data can and do provide valuable evidence on these points, but also run up against limits. For instance, employment data currently provide only limited ability to track occupational changes. Both from the worker perspective, as people move between jobs, and from the broader labor market perspective, as categories of work evolve. Data on these types of dynamics would provide much needed indicators for policymakers seeking to monitor any changes in the occupational structure of the labor market due to AI adoption and use.

The Occupational Employment and Wage Statistics (OEWS) series, for example, which reports data on employment levels and wages by occupation, cannot reliably be used for time series analysis. The employment data that form the basis for headline employment numbers, turnover statistics, and job flows capture industry but not occupational information. Ongoing innovation, such as the enhancement of unemployment insurance wage records to include occupational information, could lead to the development of richer wage and employment statistics by occupation that parallel currently available statistical products by industry. More broadly, as AI changes how people live and work, new data on time use may become increasingly valuable.

Price statistics: Capturing how inflation affects different households

Inflation statistics likewise play a central role in economic policymaking, especially in the midst of an ongoing period of elevated inflation. Policymakers such as the Fed closely monitor and respond to inflation data. And firms, workers, and consumers throughout the economy gauge and react to inflation trends in making purchasing, investment, and saving decisions. 

Yet many users also seek more nuanced information than traditional price indices can provide.

For example, there is a growing recognition that when it comes to feeling the pinch of rising prices, not all households are affected in the same way. Consumers often experience inflation differently depending on where they live, what they buy, and their income level. Housing costs, healthcare expenses, and energy prices can vary across regions and demographic groups.

Recent research has found, for example, that lower-income households have tended to experience relatively higher rates of inflation, and that Black households may face more inflation volatility than white households. While many other statistical series report differences across groups—allowing for tracking income inequality, for instance, or racial disparities in the unemployment rate—inflation series generally do not. Especially in a period of high inflation and concerns about affordability, this may reduce visibility into critical trends in real consumption that might be obscured by unmeasured price variation between groups.

To be sure, available inflation data do provide some window into these differences. The main inflation series, for example, are disaggregated by expenditure category, and report geographic variation by region and for some cities. And there are research efforts such as the research CPI by equivalized income quintiles (R-CPI-I) provides inflation estimates for households at different income levels. There are also examples of unofficial series, such the New York Fed’s Economic Heterogeneity Indicators (EHI), which provides estimates of inflation across demographic groups. More generally, there are a number of efforts taking on the important challenge of improving price measurement, such as through the innovative use of transaction data.

Income statistics: Monitoring income mobility and volatility

Income statistics, such as household median income, measures of income inequality, and poverty rates, provide a key index for both the overall performance of the economy as well as an indicator for how well the economy is meeting the material needs of households.

But users increasingly want to see not just a snapshot of the income distribution, but how the incomes of households evolve over time, and who moves into higher income groups.

These issues have gained salience in part as measures of consumer sentiment have continued to fall in recent years even as real incomes have generally continued to rise and other economic indicators remain healthy. This puzzle has defied easy explanation. But one line of questions considers the role of income trends beyond levels alone, such as how have factors like declining income mobility, or significant income volatility, may play a role in undermining economic security.

Current federal income data and statistics generally have only limited ability to speak directly to issues like income mobility or volatility because for the most part they do not track households over time. Surveys like the National Longitudinal Survey of Youth (NLSY) do, and are often used for research on these questions, but were not designed to produce statistical series. Administrative data provide some of the best available information in federal sources on mobility and volatility, though also have limits for statistical purposes. 

Advances in measurement from research conducted both in academia and within the federal statistical system show directions for innovation. Linked tax data, by capturing rich income information on the near universe of households, and allowing for following units over time, are the most promising basis for improved statistics on mobility and volatility. Research projects such as the Census Bureau’s Mobility, Opportunity, and Volatility Statistics project (MOVS) provide an example of drawing on linked tax data to develop statistics.

Wealth statistics: The largest data gap

Among the major economic indicators, wealth measurement may face the greatest unmet needs.

While income provides a snapshot of economic resources, wealth captures long-term economic security and opportunity. Wealth influences retirement readiness, educational opportunities, entrepreneurship, and resilience to economic shocks.

Yet measuring wealth remains difficult because many assets are not regularly observed, and important forms of wealth continue to evolve.

Current data gaps include:

• Better measurement of wealth inequality and the coming wave of intergenerational wealth transfers—how might we capture the different starting endowments that set up some families and children for economic success, while others start from zero wealth?
• Improved information on business ownership—how might we more consistently measure whether businesses are inherited, cooperatively owned, saddled with debt, or other aspects that indicate how much wealth a business asset could contribute?
• More comprehensive coverage of retirement assets—retirement savings accounts, Social Security, and the homes where people have invested for their retirement years.
• Measurement of emerging asset classes—cryptocurrency and other financial technologies offer new, but potentially less traceable, forms of wealth, so how might we capture these in portfolios?

These limitations make it challenging to fully assess economic opportunity and financial security across the population. As wealth concentration becomes an increasingly important policy issue with the world’s first trillionaire, demand for improved wealth statistics continues to grow.

Meeting needs through investment and innovation

The federal statistical system is one of the nation’s most important public assets. Its data guide decisions that affect economic growth, public policy, financial markets, and household well-being. Yet the economy is evolving faster than many of the systems designed to measure it.

Meeting evolving needs will require sustained investment and innovation. The reward will be a stronger evidence base for policy, better-informed economic decisions, and a clearer understanding of how economic change affects the lives of Americans.