Beyond the Unemployment Rate: A Structural Labor Market Indicator

Beyond the Headline Rate: Why the Fed is Looking at a New Structural Labor Market Indicator

If you want to know how the U.S. economy is performing, the first number you likely check is the unemployment rate. For decades, it has served as the ultimate economic thermometer—the single headline dial that tells central bankers, market participants, and the public whether the labor market is hot, cold, or just right.

But what happens when the dials on the dashboard start spinning in opposite directions?

Consider these historical puzzles:

• December 2015: The unemployment rate had fallen to a healthy 5.0%, traditionally signaling an economy near full employment. Yet, labor force participation was down nearly two percentage points, and wage growth hovered stubbornly below 3.0%, suggesting substantial hidden slack.
  
  
• 2021–2022: Unemployment rapidly returned to its pre-pandemic low of roughly 4.0%, while job vacancies surged to twice the number of unemployed workers. At the same time, however, average weekly hours worked sat below their historical trend.
  
  
• 2025–2026: Unemployment hovered around 4.0% while job openings (vacancies) fell sharply from their post-pandemic highs. Were labor markets genuinely loosening, or did the earlier vacancy surge reflect temporary, post-pandemic distortions rather than actual tightness?
  
  

These are not just technical academic debates; they represent a fundamental identification and aggregation problem for central bankers. When labor market indicators disagree, the cost of misreading the underlying slack is exceptionally high for monetary policy.

A groundbreaking Federal Reserve working paper by Isabel Cairó, Hess Chung, Francesco Ferrante, Cristina Fuentes-Albero, Camilo Morales-Jiménez, and Damjan Pfajfar (2026) titled "Beyond the Unemployment Rate: A Structural Labor Market Indicator" introduces a powerful solution: the Structural Labor Market Indicator (SLMI).

The Problem with Simple Dashboards

Faced with conflicting signals, the conventional central bank response has been to monitor a dashboard of individual indicators. While checking multiple metrics is intuitive, it presents three severe limitations:

1. Aggregation: How do you mathematically combine the extensive margin (employment and participation), the intensive margin (hours worked), vacancies, and real wages into a single, cohesive signal?
   
   
2. Context-Blindness: Should labor market assessments be based on labor data alone, or should other macroeconomic factors be integrated? For instance, a strong GDP expansion accompanied by stable, subdued inflation suggests there is still room to grow (slack), whereas sluggish growth paired with surging inflation signals that the labor supply is binding (tightness).
   
   
3. Shock Identification: A spike in unemployment driven by a sudden collapse in aggregate demand requires a completely different policy response (monetary easing) than unemployment driven by a structural, supply-side disruption (such as labor supply or matching efficiency shifts). A simple spreadsheet or statistical factor model cannot tell these driving forces apart.
   
   

Purely statistical approaches, such as applying Principal Component Analysis (PCA) directly to raw, filtered labor data (like the Kansas City Fed's LMCI), fail because they cannot distinguish supply from demand, cannot integrate non-labor macro data, and cannot support structural counterfactuals.

Enter the SLMI: A Structural Masterpiece

To solve these puzzles, the authors develop a Structural Labor Market Indicator (SLMI). Rather than relying on purely statistical filters, they ground the SLMI in a state-of-the-art, medium-scale New Keynesian Dynamic Stochastic General Equilibrium (DSGE) model.

This model features:

• Search and Matching Frictions: Modeling the costly process of firms posting vacancies and training workers.
  
  
• Endogenous Labor Supply: Allowing household members to choose whether to participate in the labor force (extensive margin) and how many hours to work (intensive margin).
  
  
• Rigidities: Incorporating nominal price and wage rigidities (staggered Calvo pricing and Nash wage bargaining).
  
  

Raw Macro Data (14 Observables)
  │── GDP, Consumption, Investment Growth
  │── Inflation (GDP Deflator, Core PCE)
  │── Wages (Compensation per hour, Hourly earnings)
  │── Interest Rates, Corporate Credit Spreads
  │── Labor (Unemployment, Participation, Hours, Vacancies)
  ▼
Estimated DSGE Model (Kalman Smoother)
  │
  ▼  (Compares Actual Outcomes vs. Flexible-Price Counterfactual)
Nine Labor Market Gaps Derived
  │── Employment, Unemployment (level & rate), LFPR, 
  │── Total Hours, Average Workweek, Vacancies, Tightness, Wages
  ▼
Principal Component Analysis (PCA)
  │
  ▼  (Extracts dominant, single-factor cyclical component)
Structural Labor Market Indicator (SLMI)

The model is conditioned on 14 macroeconomic and labor market observables over the 1987Q1–2025Q4 period. It extracts the "latent" or unobservable state of the economy by running a Kalman smoother to compute model-implied gaps—defined as the cyclical deviations of actual observed variables from their hypothetical flexible-price-and-wage equilibrium counterparts.

This flexible equilibrium retains all the real structural frictions of the economy (like search and matching costs) but strips away the artificial distortions caused by sticky prices, sticky wages, and markup shocks. It represents the "natural" rate of the economy toward which the market would gravitate if all prices and wages adjusted instantly.

Decoupling and Updating the Policy Engine

The authors introduce three crucial modifications to standard DSGE frameworks to make this model realistic for modern policymaking:

1. Wealth-in-Utility: By adding a wealth term (representing bond holdings) directly into household preferences, they break the standard theoretical link that mechanically ties the steady-state real interest rate to consumption growth. This allows them to calibrate a realistic steady-state real interest rate (steady-state federal funds rate of 3.0%) alongside a long-run GDP growth rate of 2.0%, fitting the actual data.
   
   
2. Demographically Anchored Baselines: The model's steady state is calibrated to match modern fundamentals: a steady-state unemployment rate of 4.3% (matching CBO natural rate estimates) and a labor force participation rate (LFPR) of 61.3%, reflecting the structural decline in participation driven by population aging.
   
   
3. Time-Varying Monetary Policy: Standard models assume a fixed monetary policy rule, which can cause systematic shifts in how the Fed reacts to the economy over decades to be mistakenly flagged as "policy shocks." The authors integrate time-varying Taylor rule coefficients estimated directly from the Federal Open Market Committee's (FOMC) Summary of Economic Projections (SEP) since 2013. This captures how the Fed's response to inflation rose to ~3.5 in the years before COVID-19 before gradually declining to ~1.0 by 2025.
   
   

How the SLMI is Constructed

The SLMI is built by taking the model-implied gaps for nine key labor market variables:

1. Employment Level
   
   
2. Unemployment Level
   
   
3. Unemployment Rate
   
   
4. Labor Force Participation Rate (LFPR)
   
   
5. Total Hours Worked
   
   
6. Average Workweek Hours
   
   
7. Vacancies
   
   
8. Labor Market Tightness (vacancies per job seeker)
   
   
9. Real Wages
   
   

To aggregate these, the authors standardize each gap and apply Principal Component Analysis (PCA). Because the underlying gaps are disciplined by a cohesive structural model, the first principal component is incredibly powerful: it accounts for the vast majority of the total variance and is the only component with an eigenvalue greater than one.

Importantly, the authors scale the final SLMI so that zero corresponds to neutral labor market slack. A positive value represents a tight labor market (overutilized resources), while a negative value represents slack (underutilized resources).

The Weights of the SLMI

The influence of each individual gap on the synthetic indicator is determined by its mathematical PCA loading:

Labor Market Gap	Symbol	Weight in SLMI
Labor Force Participation Rate	LFPR	0.67
Unemployment (Level)	U	-0.32
Average Workweek	ww	0.27
Employment	E	0.23
Unemployment Rate	u	-0.20
Total Hours	H	0.17
Real Wage	w	0.12
Vacancies	v	0.07
Labor Market Tightness	$\theta$	0.01

(Note: Standard signs apply; higher unemployment level and rate pull the SLMI downward, representing slack, while higher participation, hours, and vacancies pull it upward).

Major Insights for Central Bankers

1. GDP and Inflation Carry Crucial Labor Signals

One of the paper's most profound findings is that you cannot fully understand the labor market by looking at labor market data alone.

By comparing the full SLMI with an alternative version constructed using only labor market variables, the authors prove that non-labor macroeconomic data accounts for nearly 19% of the variance in our labor slack assessment.

   [ Strong GDP Growth + Stable Inflation ] ──► Signals *more slack* than labor data alone suggests
   [ Moderate GDP Growth + Rising Inflation ] ──► Signals *less slack* (supply constraints binding)

For instance, during the pandemic recovery in 2020–2021, the unemployment rate was falling rapidly, leading labor-market-only indicators to suggest the economy was tight. However, the full SLMI incorporating weak GDP growth correctly identified persistent resource underutilization (hidden slack). Conversely, during 2021–2023, elevated price pressures (PCE and CPI inflation) validated that the labor market was indeed exceptionally tight, even as certain specific labor margins appeared ambiguous.

2. An "Early Warning" System at Business Cycle Turning Points

The SLMI exhibits distinct cyclical behavior that sets it apart from alternative, purely empirical indicators (like the Kansas City Fed's LMCI or statistical filters):

• Earlier Warnings of Downturns: The SLMI consistently turns negative earlier at the onset of recessions—most notably warning of deteriorating conditions ahead of the 2001 and 2008 downturns.
  
  
• Gradual, Realistic Recoveries: Conversely, the SLMI recovers more gradually during economic expansions. This gradual pace reflects the model's structural friction dynamics, showing how labor market slack dissipates slowly as workers gradually re-enter the labor force and adjust their hours.
  
  

3. Aggregate Demand and Investment Efficiency Rule the Cycle

What actually drives structural shifts in the labor market over the long run? The model's historical shock decomposition reveals two main culprits:

• Risk Premium Shocks (Aggregate Demand): These shocks explain the bulk of labor market variation over the 1987–2026 period, serving as the dominant driver behind the severe slack of the 2008–2009 Global Financial Crisis and the COVID-19 recession.
  
  
• Investment Efficiency Shocks (Supply-Side): These shocks play a heavy role in structural downturns, contributing significantly to the 2001 recession and the notoriously slow, "jobless" recovery of the 2010s.
  
  

Interestingly, monetary policy shocks play only a modest role historically and become almost negligible after 2013. This is a testament to the model's time-varying Taylor rule, which successfully captures the Fed's systematic policy shifts rather than misidentifying them as random shocks.

Operating in the Real World: Uncertainty and Real-Time Limits

Central bankers must make decisions in real time without the benefit of future hindsight. The authors stress two critical considerations for using the SLMI in practice:

Parameter Uncertainty Peaks in Recoveries

Using Monte Carlo simulations over 100,000 parameter draws, the paper establishes 68% credible intervals around the SLMI.

The results reveal that uncertainty about slack is highly asymmetric, with a downward skew (leaning toward greater tightness). Crucially, model uncertainty peaks during economic recoveries. For example, during the protracted recovery of 2010–2015, the 68% credible interval actively straddled zero. This mathematical range perfectly mirrors the intense, real-world policy debate that took place at the time regarding how much genuine labor slack was left in the aftermath of the Great Recession.

Real-Time Performance: The One-Sided Filter

The historical SLMI is calculated using a two-sided Kalman smoother, which relies on past, present, and future data. To test its real-world feasibility, the authors construct a one-sided Kalman filter version of the indicator that only uses information available up to that exact moment.

The results are highly encouraging: the real-time (one-sided) indicator is extremely highly correlated with its smoothed counterpart. While it has exhibited a slight tendency to overestimate tightness since 2015, it fully preserves the vital early-warning properties of the indicator. This makes the SLMI a highly practical, reliable, and actionable tool for real-time policymaking.

Conclusion: A New Compass for Central Banking

The Structural Labor Market Indicator represents a major leap forward for monetary policy analysis. By moving beyond the one-dimensional headline unemployment rate and avoiding the limitations of purely statistical summaries, the SLMI provides a theoretically rigorous, macro-integrated, and empirically validated map of economic slack.

As central bankers continue to navigate complex post-pandemic structural shifts, having a structural compass like the SLMI ensures that monetary policy is guided not by isolated, conflicting dials, but by a cohesive, multi-dimensional story of the U.S. labor market.

Cairo, I., Chung, H., Ferrante, F., Fuentes-Albero, C., Morales-Jiménez, C., & Pfajfar, D. (2026). "Beyond the Unemployment Rate: A Structural Labor Market Indicator," Federal Reserve Board Finance and Economics Discussion Series 2026-058.