Can we better predict recessions? And can we bend those same tools to meaningfully speak to underlying economic well-being, before those tools break?

Working people connect the economy, and because labor is so central to all aspects of production, many popular recession indicators rely mostly on employment-based measures. But with so many measures of labor utilization, which metric is the most predictive of recessions, and can we use disaggregated data in recession indicators to say something even deeper about labor market inequities?

Utilizing labor utilization measures within a Modified Sahm Rule reveals more comprehensive measures are more predictive. Utilizing disaggregated unemployment data, by education, race, gender, etc., compared to the standard subtrahend of U-3, creating a Relative Sahm Rule , can be used to show underlying inequities within the economy. Allowing for diagnosis beyond merely saying that labor market conditions are simply different.

The Modified Sahm Rule consists of two components: the minuend, \( \lambda_{i,t} \), and the subtrahend, \( \gamma_{j,t} \). The Modified Sahm Rule is defined as:

where:

The recession indicator \( R_t \) is triggered when the Sahm Rule exceeds a threshold \( \alpha \):

Note the following:

• \( k \) is typically set to 3 months, but the function can be smoothed over any \( k \)-month average. Generally, subcategories of unemployment with increased volatility require more smoothing.
• \( \alpha \) is generally set to 0.5, but this parameter can be adjusted to establish upper and lower bounds, such as 0.3 and 0.8, as discussed in Michaillat and Saez (2024).
• Indices \( i \) and \( j \) allow for comparing different unemployment measures (e.g., \( i \neq j \), such as comparing Hispanic women unemployment or the unemployment rate in Cocke County, Tennessee against the baseline \( U_3 \) as the subtrahend, creating a Relative Sahm Rule). This formulation can also be extended to include job vacancy or any labor utilization measure.

Most of the popular recession indicators now rely on employment data, generally unemployment data, even more specifically the U-3 measure of unemployment. But what are the benefits of using the most standard measure of unemployment? And is there a better metric?

The Modified Sahm Rule utilizing different measures of labor utilization highlights the fact that more comprehensive measures, like the U-4 to U-6 measure, are more responsive to changes in the underlying economy. This becomes intuitive when we understand recessions as effecting the demand for labor, and the more comprehensive measures of labor utilization are simply better-suited at capturing more sesnitive changes in the labor market.

The Modified Sahm Rule, using U-6 measure of labor utilization, is the most responsive indicator. Providing well-calibrated early warning signs of any concerning trends in the labor market. The Modified Sahm Rule, as opposed to the Relative Sahm Rule, behaves similarly to the Traditional Sahm Rule. Whereas the Relative Sahm Rule is designed to highlight relative differences across the economy, with more exacting diagnosis. The Modified Sahm Rule is designed to simply create faster indicators.

The Modified Sahm Rule can also be used with unemployment data disaggregated by educational attainment. The Modified Sahm Rule, using unemployment data disaggregated by educational attainment, is the most responsive indicator. It provides well-calibrated early warning signs of emerging trends in the labor market. Unlike the Relative Sahm Rule, which is designed to highlight disparities across different education levels and provide more nuanced diagnoses, the Modified Sahm Rule focuses on delivering faster and broader indicators. This makes it particularly useful for identifying shifts in labor utilization that might otherwise go unnoticed in aggregate data.