The Hidden Slack Problem: Inflation Above Target and a Labor Market Weaker Than It Looks

The Hidden Slack Problem: Inflation Above Target and a Labor Market Weaker Than It Looks

The chart presents a policy configuration that is more difficult than either a conventional soft landing or a conventional recession. PCE inflation began around 1%–2% before the pandemic, surged toward 7% in 2022, fell to roughly 2.2%, and then turned upward again toward 3.5% in 2026. That last move matters more than the earlier disinflation. It implies that the final mile back to the Federal Reserve’s 2% target did not merely stall; inflation momentum reappeared after markets and households had begun to expect normalization. At the same time, headline unemployment, after its pandemic spike and its decline to about 3.5% in 2022–23, now sits near 4.2%, close to a familiar 4% full-employment benchmark. Read only through those two series, the economy looks overheated enough to justify restraint but healthy enough to tolerate it.

The third series changes that interpretation. A slack-adjusted unemployment measure that broadly followed headline unemployment before the pandemic fell to just above 1% in 2021–22, then rose to approximately 5.5%. Its widening gap from the 4.2% headline rate suggests that standard unemployment is understating unused labor capacity. The United States may therefore be experiencing an uncomfortable combination: inflation well above target and a labor market whose effective weakness is greater than the most quoted statistic indicates. This is not the simple demand-driven inflation of an economy operating uniformly beyond capacity. Nor is it the clean disinflation of a labor market cooling in an orderly way. It is a mixed regime in which price pressure can persist even while employment opportunities deteriorate beneath the surface.

That distinction is decisive for policy and markets. If inflation reflects excessive aggregate demand, restrictive rates can cool spending, reduce vacancies, and return prices to target with manageable employment costs. If inflation increasingly reflects supply constraints, sector-specific wage pressure, housing persistence, energy costs, tariffs, insurance, or unresponsive productivity, then weaker labor demand may not buy much disinflation. Monetary restraint can continue to damage employment while the inflation rate remains sticky. The chart is therefore not simply a snapshot of two mandates. It is evidence that the mapping from labor slack to inflation may have shifted, and that policy based on the headline unemployment rate alone risks diagnosing the wrong economy.

Three Curves, One Policy Conflict

The first curve describes the inflation cycle. Moving from roughly 1%–2% to nearly 7% was the combined product of reopening demand, supply-chain disruption, fiscal transfers, commodity shocks, housing pressure, and extraordinary monetary accommodation. The subsequent decline toward 2.2% demonstrated that much of the shock was reversible. Goods supply normalized, energy effects faded, inventories rebuilt, and policy tightened. Yet the rebound to about 3.5% shows that the remaining inflation is not simply the dying tail of the pandemic. A renewed rise after a large decline is more troubling than a steady but slow descent because it can alter expectations, wage bargaining, pricing conventions, and the policy rate required to restore credibility.

The second curve, headline unemployment, tells a comparatively benign story. The pandemic spike was temporary, the recovery to around 3.5% was exceptionally rapid, and the current 4.2% level is only modestly above the lows. Against a 4% full-employment benchmark, the labor market appears close to balance. In the textbook dual-mandate framework, this would permit the central bank to concentrate on inflation without fearing a large departure from maximum employment. A 4.2% unemployment rate has historically been consistent with respectable growth, ongoing job creation, and limited recession risk.

The third curve disputes that comfort. At about 5.5%, the slack-adjusted measure is 1.3 percentage points above headline unemployment. More important than the exact gap is its direction and history. Before the pandemic the measures moved together; in 2021–22 the adjusted measure fell to just above 1%, reflecting an extraordinarily tight labor environment; since then it has risen substantially faster. A widening wedge means that the composition and intensity of labor utilization have changed. Workers may technically remain employed while receiving fewer hours, accepting lower-quality work, leaving the labor force, enduring longer job searches after displacement, or finding that vacancies are less credible than headline counts imply.

What Headline Unemployment Cannot See

The official unemployment rate is carefully constructed, but it answers a narrow question. A person is unemployed if they lack a job, are available to work, and have actively searched recently. That definition excludes several forms of economic slack. A discouraged worker who stops searching is outside the labor force, not unemployed. A part-time worker who wants full-time hours is employed. A recent graduate delaying entry, a caregiver unable to find workable hours, or a displaced worker taking a much lower-productivity job can all disappear from the headline measure’s signal about labor underutilization.

Hours are particularly important because firms often adjust labor input at the intensive margin before cutting headcount. They trim overtime, shorten shifts, rely less on temporary staff, or leave positions unfilled. Payroll status remains intact, so unemployment barely moves, but household income and aggregate labor demand soften. This is rational behavior under uncertainty: preserving trained workers avoids rehiring costs if demand rebounds. It also means a cooling cycle can advance considerably before layoffs make it obvious. An adjusted measure that incorporates hours, participation, job-finding rates, or underemployment may therefore lead the headline rate.

Labor-force participation creates a second blind spot. The unemployment rate can remain low because people find jobs, but also because people cease searching. Aging, health constraints, caregiving, immigration dynamics, geographic mismatch, and skill obsolescence all affect the denominator. None should be interpreted mechanically as cyclical weakness, yet each changes how much productive labor is actually available and desired. The widening gap in the chart invites a decomposition: how much comes from involuntary part-time work, weaker participation, longer durations, slower hiring, or declining hours? Even without that full decomposition, the gap warns investors against treating 4.2% as a complete summary of household labor security.

The Beveridge Curve and Declining Matching Efficiency

The Beveridge curve relates unemployment to job vacancies. In a normal expansion, vacancies rise and unemployment falls; in a contraction, vacancies fall and unemployment rises. Pandemic reopening pushed the relationship outward because employers posted extraordinary numbers of jobs while matching was impaired by health concerns, sectoral reallocation, location constraints, and changing worker preferences. The subsequent decline in vacancies without a major rise in unemployment was celebrated as evidence of a soft landing: excess labor demand could disappear without mass layoffs.

But vacancy counts are not equivalent to effective job offers. Listings may be duplicated, kept open for future needs, posted without urgency, or aimed at applicants whose skills and location do not match available workers. When financing costs rise and uncertainty increases, a firm can preserve the option to hire by maintaining a vacancy while delaying commitment. Search-and-matching models associated with Diamond, Mortensen, and Pissarides emphasize that employment depends not just on vacancies and job seekers but on matching efficiency. If efficiency falls, the same vacancy rate produces fewer hires and longer searches.

This helps explain why adjusted slack can rise while headline unemployment stays near full employment. A worker may move between short jobs, accept fewer hours, or search longer without immediately appearing in a dramatic unemployment statistic. Firms may complain about labor scarcity in specialized roles while broad pools of workers experience weak bargaining power. The aggregate labor market becomes segmented. A central bank observing high vacancies or low headline unemployment might infer excess demand, while the lived market is already loosening. The policy danger is overtightening in response to indicators whose relationship to actual wage pressure has degraded.

A Phillips Curve with a Moving Intercept

The Phillips curve is often summarized as an inverse relationship between unemployment and inflation. In modern formulations, inflation depends on expected inflation, economic slack, and supply shocks. A simple representation is π = πᵉ − κ(u − u*) + s, where π is inflation, πᵉ expected inflation, u unemployment, u* the equilibrium unemployment rate, κ the sensitivity to slack, and s supply-side pressure. The chart’s puzzle arises because u appears close to u* while π is about 3.5%, far above a 2% target. If adjusted unemployment near 5.5% better captures slack, the puzzle becomes sharper: inflation is high despite material underutilization.

There are several possible explanations. The supply term may be positive because of energy, trade restrictions, housing, insurance, healthcare, or other bottlenecks. Expected inflation may have drifted upward after repeated price shocks. The equilibrium unemployment rate may be changing because matching efficiency and sectoral composition changed. Or κ may have become smaller, meaning additional unemployment produces less disinflation than policymakers expect. These mechanisms are not mutually exclusive. Together they describe a Phillips curve whose intercept and slope are unstable.

This instability changes the cost of policy. If κ is low and s is high, raising unemployment yields only a modest reduction in inflation. The sacrifice ratio—the cumulative output or employment loss required to lower inflation—rises. Conversely, if inflation expectations remain anchored and supply pressure fades, current slack may generate delayed disinflation without much additional tightening. The rational response is not to abandon the Phillips curve but to treat it as conditional. Policymakers must identify which component is moving rather than assuming that a low headline unemployment rate proves excessive demand.

Inflation Composition Matters More Than the Headline

PCE inflation is the Federal Reserve’s preferred measure because it covers a broad consumption basket, adjusts weights as spending patterns change, and differs methodologically from CPI. Yet one aggregate rate can arise from very different processes. Goods inflation driven by supply chains responds differently to policy than shelter inflation driven by slow-moving leases. Market services respond differently from administered healthcare prices. Energy can reverse quickly; insurance can reset with long lags. A 3.5% rate dominated by cyclically sensitive services is a different policy problem from 3.5% dominated by lagged housing and supply-constrained essentials.

The rebound from 2.2% toward 3.5% therefore demands attribution. If wage-sensitive services are accelerating broadly, labor demand may still be stronger than adjusted slack suggests. If housing, energy, tariffs, or insurance lead the move, interest rates may have weak near-term traction and significant collateral costs. Higher rates can even worsen selected measured costs through housing supply, financing, and insurance channels. Monetary policy ultimately restrains nominal demand, but the path from restraint to each price category is uneven and sometimes perverse in the short run.

Investors should separate breadth, persistence, and impulse. Breadth asks how many categories are rising rapidly. Persistence asks whether those increases repeat. Impulse asks whether the marginal change is accelerating or decelerating. A rebound to 3.5% with broad acceleration is much more concerning than a temporary energy-driven spike. Median and trimmed-mean measures, rent on new leases, wage trackers, productivity, import prices, and inflation expectations help distinguish these cases. The chart supplies the warning; composition determines the appropriate response.

Why the Dual Mandate Is Asymmetric Here

The Federal Reserve seeks price stability and maximum employment, but the two goals do not have equal measurement structures. Inflation is observed through several price indexes with revisions and known lags. Maximum employment is not a single number. It depends on participation, demographics, hours, wages, job quality, vacancies, and matching. A policy framework that treats 4% unemployment as a precise full-employment point can therefore create false confidence, especially when an adjusted measure signals 5.5% slack.

The risks are also asymmetric. If the Fed eases too early and the rebound in inflation becomes embedded, restoring credibility may require more severe tightening later. If it stays restrictive too long while hidden slack is already rising, employment deterioration can become nonlinear. Hiring slows before layoffs surge; once firms begin broad cuts, weak income feeds weak spending, delinquencies, tighter credit, and further job losses. The Sahm-rule intuition—that relatively small increases in unemployment can identify recession dynamics—reflects this nonlinearity, though no single rule should dominate policy.

Optimal policy under uncertainty therefore resembles risk management rather than point forecasting. The Fed should ask not only whether 3.5% inflation is too high and 4.2% unemployment is low, but how reliable each measure is and how costly each error would be. Robust-control logic argues for policies that perform tolerably across competing models. That may mean avoiding both aggressive easing and reflexive additional tightening, while using communication to preserve inflation credibility and watching high-frequency labor indicators for evidence that hidden slack is becoming outright contraction.

The Neutral Rate and the Real Restraint Problem

The policy stance depends on the real interest rate, approximately the nominal policy rate minus expected inflation. But the neutral real rate, r*, is unobservable and can move with productivity, demographics, fiscal policy, global saving, and risk preferences. If r* has risen, a seemingly high nominal rate may be less restrictive than past comparisons suggest. If hidden labor slack indicates demand is already weak, however, the effective stance may be tighter than aggregate spending data imply.

Distribution strengthens this ambiguity. High-income households with appreciating assets and fixed-rate mortgages may be relatively insulated from higher rates. Lower-income households, renters, small businesses, and firms needing refinancing feel them directly. Aggregate consumption can remain firm because wealthy households spend, even while marginal borrowers face stress and labor opportunities deteriorate. Monetary transmission becomes uneven: the sectors most sensitive to rates contract, while less sensitive sectors sustain inflation. This can produce both 3.5% inflation and 5.5% adjusted slack without contradiction.

Financial conditions also mediate the stance. Rising equity prices, narrow credit spreads, and abundant liquidity can offset policy restraint for large corporations and affluent households. Regional banks, commercial real estate, small firms, and consumers using revolving credit may simultaneously face tight conditions. The relevant question is not whether policy is restrictive on average, but where it is restrictive and whether those locations overlap with the sources of inflation. If they do not, additional hikes can deepen labor weakness while leaving stubborn price categories comparatively untouched.

Market Pricing: The Bad Mix for Both Bonds and Equities

For government bonds, the chart creates two opposing forces. Above-target inflation and a renewed upswing argue for higher expected policy rates and possibly a higher term premium. Hidden labor slack argues for future easing and lower real growth. The yield curve becomes a battleground between inflation persistence and recession insurance. Front-end yields respond to the near-term reaction function; long yields also reflect inflation uncertainty, fiscal supply, and term premium. Duration can rally on weak labor data yet sell off if investors believe the Fed cannot ease safely.

Equities face a similarly uncomfortable mix. Persistent inflation can compress valuation multiples through higher discount rates. Labor weakness can reduce revenue growth, particularly for consumer-sensitive firms. Yet slower wage pressure may support margins, and nominal growth can sustain reported earnings. The cross-section matters more than the index. Firms with pricing power, low refinancing needs, recurring demand, and productivity gains are better positioned than leveraged businesses dependent on discretionary consumption. Small caps may benefit from eventual easing but are often more exposed to credit and domestic labor weakness.

Credit markets can appear calm late into such a regime because defaults lag and nominal revenue cushions debt service. But hidden slack eventually matters through household cash flow, delinquencies, and corporate earnings. Investors should watch lower-quality spreads, bank lending standards, consumer ABS performance, and refinancing calendars rather than relying on broad investment-grade spreads alone. A regime with inflation at 3.5% and effective slack near 5.5% can generate dispersion before it generates an index-level crisis.

Portfolio Construction Under Stagflationary Tail Risk

The chart does not prove stagflation, but it increases the probability of a mild stagflationary state: inflation above target alongside weakening labor utilization. Traditional stock-bond diversification is less reliable in that state because inflation shocks can hurt both equity multiples and nominal duration. Portfolio construction should therefore distinguish recession hedges from inflation hedges. Long government bonds protect best when weak labor demand pulls inflation expectations down. They protect less when supply pressure keeps inflation elevated.

Inflation-linked bonds can protect realized purchasing power, though they remain exposed to changes in real yields. Short-duration government paper offers liquidity and carry with less duration risk. Commodities and energy exposure can hedge supply shocks but are volatile and sensitive to the precise source of inflation. Quality equities with strong balance sheets and pricing power can provide partial resilience. Explicit options can add convexity, though their cost must be budgeted rather than treated as free insurance. No single asset hedges every branch of the distribution.

A practical allocation process uses scenario weights. In a clean disinflation scenario, inflation returns from 3.5% toward 2%, adjusted slack stabilizes, and duration performs well. In a renewed inflation scenario, both nominal bonds and long-duration equities face pressure. In a labor-break scenario, the 5.5% adjusted measure leads a rapid rise in headline unemployment and recession hedges pay. In a productivity scenario, output remains firm, labor reallocates, and inflation falls without a large downturn. Portfolio robustness comes from surviving all four, not from maximizing exposure to the most comfortable forecast.

Indicators That Can Resolve the Contradiction

The first group of indicators measures labor quantity: payroll growth, average weekly hours, temporary-help employment, job-finding rates, continuing claims, layoffs, and involuntary part-time work. Hours and temporary employment often turn before headline unemployment. A sustained deterioration would validate the adjusted-slack warning. Stabilization would suggest that the wedge reflects structural or compositional changes rather than an imminent cyclical break.

The second group measures labor price and productivity: wage growth, compensation per hour, unit labor costs, quit rates, and output per hour. Wage inflation without productivity growth can sustain services inflation. The same wage growth paired with strong productivity is less threatening because unit costs remain contained. Quits reveal bargaining confidence, while job-to-job transitions reveal whether workers can still move toward better matches. These measures help determine whether 4.2% unemployment is genuinely tight or merely optically low.

The third group measures inflation propagation: trimmed-mean PCE, market-based services, rent on new leases, import prices, producer margins, energy, and survey or market expectations. If breadth narrows and expectations remain anchored, the rise toward 3.5% may fade as slack works with a lag. If breadth expands and expectations drift, the Fed faces a genuine credibility problem. The key is the joint distribution. Labor weakness alone does not guarantee disinflation, and inflation alone does not prove broad overheating.

Policy Should Target the Mechanism, Not the Symbol

Interest rates are a broad instrument. They work mainly by changing financing conditions, asset prices, spending, and ultimately labor demand. They cannot produce apartments quickly, expand energy supply, repair trade networks, lower insurance losses, or instantly improve matching efficiency. When inflation is driven by those mechanisms, complementary policies matter. Housing supply, labor mobility, childcare, immigration administration, competition, infrastructure, and energy resilience can reduce inflationary bottlenecks without requiring additional unemployment.

This does not absolve monetary policy. A central bank that accommodates every supply shock can allow temporary inflation to become persistent through expectations and contracts. The distinction is between preventing second-round propagation and pretending that demand destruction repairs supply. Credible monetary policy can anchor the nominal environment while fiscal and structural policy address real constraints. Coordination need not mean sacrificing independence; it means recognizing comparative advantage among policy tools.

Communication should also reflect uncertainty. Declaring victory because inflation touched 2.2% would have been premature, but treating 4.2% unemployment as proof of unlimited room to tighten would be equally simplistic. The public can understand a conditional reaction function: policy will remain sufficiently restrictive to return inflation to 2%, but officials will respond if broad labor deterioration shows that restraint is exceeding what price stability requires. Credibility comes from coherent contingency plans, not from false precision.

The Investment Conclusion: Respect the Wedge

Before reaching the portfolio conclusion, it is useful to consider how the chart can be misread in real time. Data revisions are one danger. Employment surveys can diverge, population controls can alter historical comparisons, and seasonal adjustment is unusually difficult after structural breaks. PCE components are revised as source data improve. An adjusted slack index necessarily embeds modeling choices about which workers, hours, or transitions deserve weight. None of this makes the signals useless. It means conviction should rise when independent measures agree and fall when a conclusion depends on a single estimated series. A robust reading treats the 5.5% measure as a warning to investigate, not an oracle that replaces every other labor statistic.

Another danger is confusing level with change. A 4.2% unemployment rate may look low historically, but its rise from the 3.5% region can matter more than its absolute level. Labor markets often display momentum because hiring and separations reinforce one another. Similarly, 3.5% inflation is far below the 7% peak, but its rise from roughly 2.2% changes the direction of the policy problem. Markets price marginal change and reaction functions, not only long-run comparisons. Both curves are moving in an unfavorable direction: inflation is reaccelerating while labor utilization is weakening. That joint deterioration is the chart’s central macro signal.

The household consequences also deserve emphasis. Aggregate slack is not evenly distributed. Younger workers, recent entrants, lower-wage service employees, temporary workers, and people without specialized credentials typically absorb cooling first. Their marginal propensity to consume is high, so lost hours can affect demand quickly. Meanwhile, asset-owning households may benefit from high nominal yields and resilient portfolios. This bifurcation can keep aggregate consumption stronger than labor stress would imply, delaying the macro signal while deepening distributional strain. Eventually, however, weaker income growth among liquidity-constrained households appears in revolving credit, auto delinquencies, rent stress, and discretionary spending.

For corporations, the same wedge produces conflicting planning signals. Nominal revenue may remain healthy because prices are rising, while real volume and labor utilization weaken. Managers can initially protect margins by limiting hiring and hours rather than cutting posted prices. That behavior raises measured slack but allows inflation to persist, especially in concentrated industries with pricing power. Over time, the choice becomes harder: pass costs through and risk volume, or absorb them and lose margin. Earnings dispersion therefore widens well before aggregate profits collapse. Analysts should decompose revenue into price and volume, and labor cost into wages, hours, and headcount, rather than taking nominal growth at face value.

International linkages add another layer. A strong dollar can suppress import prices and tighten global financial conditions, helping domestic disinflation while weakening foreign demand. A weaker dollar can support U.S. manufacturing and earnings translation but raise imported inflation. Commodity shocks can lift PCE even as domestic labor slack rises. Foreign central banks may face different inflation-growth combinations, creating divergent rate paths and currency volatility. The domestic chart is therefore not a closed system. Part of the apparent failure of unemployment to control inflation can originate in global supply, exchange rates, and geopolitical risk rather than U.S. excess demand.

Finally, policy lags make restraint cumulative. Rate increases affect floating-rate borrowers quickly but fixed-rate mortgages and corporate debt only at refinancing. A labor market can therefore weaken gradually as successive cohorts refinance, even without new hikes. At the same time, inflation can remain sticky because shelter and contract prices reset slowly. The visible combination of 3.5% inflation and 5.5% adjusted slack may be a transitional overlap between lagged price persistence and lagged demand destruction. If so, further tightening would be especially risky: it would add future weakness to a pipeline already carrying substantial restraint. If instead inflation expectations are becoming unanchored, waiting would be costly. This is why forward-looking evidence matters more than mechanical rules.

The chart’s most important information is not any single level. It is the wedge between a familiar headline and a broader measure of slack. PCE inflation near 3.5% is clearly above the Fed’s 2% target. Headline unemployment near 4.2% is close to the 4% benchmark and looks benign. Adjusted slack around 5.5% says that the labor market beneath the headline may already be materially weaker. Those facts can coexist when inflation is supply-sensitive, matching efficiency deteriorates, monetary transmission is uneven, and aggregate statistics conceal distribution.

For the Fed, the wedge argues for patience, model humility, and close attention to hours, hiring, participation, and inflation composition. For investors, it argues against a one-dimensional “higher inflation means stronger growth” trade or a one-dimensional “weaker labor means immediate cuts” trade. Rates can stay restrictive longer than labor fundamentals justify if inflation credibility is at risk. Labor can weaken further than headline unemployment suggests before policy relief arrives. That is precisely the environment in which volatility and cross-sectional dispersion rise.

The final lesson is that full employment is not a line on a chart and price stability is not secured by one favorable print. Both are latent economic states inferred from imperfect data. When different measures diverge, the divergence is itself a signal. Here it says that the economy may be paying a larger employment cost for disinflation even as inflation stops improving. A disciplined policy and portfolio framework must respect both sides of that message: inflation is too high, and labor slack may already be too high as well.