📈 economics

Effective Demand, Inflation, and the Persistence of Unemployment: The Limits of Full-Employment Policy

This paper argues that while government expenditure positively influences economic growth, its ability to secure full employment is severely limited by the weak predictive power of policy instruments on output and the finding that unemployment is driven primarily by the output gap rather than inflation, implying that achieving full employment requires demand to consistently outpace productive capacity rather than merely sustaining growth.

Original authors: houssam boughabi

Published 2026-10-27

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: houssam boughabi

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of economics, there is a long-standing hope that governments can simply turn a dial to fix the job market. The idea, rooted in the work of John Maynard Keynes, is that when people stop spending money and businesses stop hiring, the state can step in with public projects or lower taxes to stir up demand. If the government spends enough, the logic goes, factories will run at full capacity, and everyone who wants a job will find one. This concept relies on a simple chain of cause and effect: more spending leads to more production, and more production leads to more workers. For decades, policymakers have operated under the assumption that if they can just keep the economy growing, full employment will follow naturally.

However, this chain of logic has a hidden weak link. Just because an economy is growing does not mean it is growing fast enough to catch up with the number of people looking for work, nor does it guarantee that the money people earn will actually buy what they need. Prices can rise faster than wages, eating away at the value of a paycheck, or the economy's ability to produce goods might expand at the same speed as the demand for them, leaving the jobless rate unchanged. The question facing modern economists is not just whether the government can make the economy bigger, but whether it can make it big enough to employ everyone without causing prices to spiral out of control.

A recent study by Houssam Boughabi, an economist at the Institut National de Statistique et d'Economie Appliquée in Morocco, tackles this problem by looking at the long-term history of the United States economy. The researcher built a model to trace how government spending, interest rates, and taxes actually influence the growth of the economy over time, and then tested whether that growth translates into lower unemployment. The study examined quarterly data spanning more than seventy years, from 1955 to 2026, a period that includes every major boom and bust in modern American history. The goal was to see if the old promise of demand management still holds true when you account for the messy reality of inflation and the changing speed of the economy's productive capacity.

The findings paint a picture that is far more cautious than the optimistic view of the past. The study confirms that government spending does have a positive connection to future economic growth. When the government increases its spending relative to the size of the economy, the economy tends to grow a bit faster in the following quarters. However, the power of this lever is surprisingly small. The three main policy tools—spending, interest rates, and taxes—explain only about 3 percent of the fluctuations in economic growth. The vast majority of what drives the economy forward is simply its own momentum, a tendency to keep moving in the direction it was already going. Furthermore, the connection between these tools and growth has weakened significantly since the mid-1970s. What worked in the past to stimulate the economy does not seem to work as reliably today.

Perhaps the most critical discovery concerns the relationship between a growing economy and the number of people with jobs. The study found that the sheer size of the economy does not determine the unemployment rate. A country can have a massive economy with high output and still have high unemployment. Instead, what matters is the gap between how much the economy is actually producing and how much it is capable of producing. If the economy is growing but is still operating below its full potential, unemployment remains high. Only when the growth of demand outpaces the growth of the economy's capacity to produce goods and services does unemployment begin to fall. This means that simply sustaining growth is not enough; the growth must be faster than the rate at which the economy's potential expands.

The role of inflation in this equation is also more complex than a simple trade-off. The study looked for a stable relationship where higher inflation would consistently lead to lower unemployment, a concept known as the Phillips curve. The data shows that this relationship was present in the early decades of the study but has largely disappeared since the 1970s. In the modern era, there is no reliable statistical link between rising prices and falling jobless rates. This suggests that trying to lower unemployment by accepting higher inflation is no longer a dependable strategy. The study also noted that unemployment has a stubborn quality; once it rises due to a recession, it tends to stay high for a long time, even after the economy begins to recover. This persistence means that the damage done by a downturn is not quickly undone by a return to normal growth.

When the researcher used the model to project what would happen if the government took specific actions today, the results were modest. The simulation suggested that if the government were to increase its spending by one percentage point of the total economy for a year, it would lower the unemployment rate by only about 0.15 percentage points two years later. While this is a positive effect, it is small. Lowering interest rates or cutting taxes had even smaller impacts. The study concludes that while active government policy can nudge the economy in the right direction, it cannot single-handedly secure full employment. The path to full employment requires demand to grow faster than the economy's ability to produce, a condition that is difficult to achieve and maintain given the small and unstable influence of current policy tools.

Ultimately, the research suggests that the dream of a simple policy fix for unemployment is elusive. The economy is not a machine where turning a few knobs guarantees a full workforce. Instead, it is a system where the relationship between money, jobs, and prices is fragile and constantly shifting. The study does not say that government action is useless, but it does warn that the tools available are blunt and their effects are limited. To truly achieve full employment, policymakers must understand that growth alone is not the answer; the growth must be of a specific kind that outstrips the economy's capacity, and even then, the margin for error is narrow. The limits of full-employment policy are not just about inflation, but about the fundamental difficulty of making demand grow fast enough to absorb the entire workforce in a changing economic landscape.

Technical Summary: Effective Demand, Inflation, and the Limits of Full-Employment Policy

Problem Statement
The paper investigates whether active Keynesian demand management can secure and maintain full employment when the cumulative trajectory of output and accompanying price dynamics are explicitly accounted for. While the principle of effective demand posits that employment is determined by aggregate demand rather than automatic market adjustment, the paper questions whether a policy stance that reliably raises output necessarily translates into sustained full employment. The central problem is that an expansion in nominal output may fail to reduce unemployment if it is matched by the growth of productive capacity or if inflation erodes real purchasing power. The study seeks to disentangle the cumulative expansion of activity from price dynamics to determine the analytical limits of full-employment policy.

Methodology
The author develops a reduced-form theoretical framework and confronts it with quarterly United States data from 1955Q1 to 2026Q2. The methodology proceeds in three stages:

Theoretical Framework:
- Output Dynamics: Expected output growth is modeled as a function of three policy instruments: public expenditure ( $G_t$ ), the interest rate ( $r_t$ ), and the tax rate ( $\tau_t$ ). The growth equation is specified as $^g_t = \delta G_t + \beta r_t + \gamma \tau_t$ .
- Unemployment and Inflation: Unemployment is modeled as a joint function of inflation ( $\pi_t$ ) and an output measure ( $Y_t$ ). The framework distinguishes between the level of output and the output gap (relative to productive capacity). Inflation evolves through proportional growth factors ( $f_t$ ).
- Cumulative Expression: By iterating the growth and inflation equations, the author derives a horizon-dependent expression for expected unemployment (Eq. 9). This expression separates the contribution of cumulative output expansion from that of inflation, allowing for an analysis of long-run prospects.
Econometric Strategy:
- The study employs a "specification as written" approach, estimating equations exactly as derived, and an "adjusted specification" only where unit-root, cointegration, and stability diagnostics necessitate deviation.
- Data Properties: The analysis utilizes Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), DF-GLS, and KPSS tests to determine integration orders. It finds that real GDP, government expenditure, and interest rates are $I(1)$ , while unemployment and the output gap are $I(0)$ .
- Estimation: Ordinary Least Squares (OLS) with Newey-West heteroskedasticity- and autocorrelation-consistent (HAC) standard errors is used. Parameter stability is tested using Andrews' sup-Wald (QLR) test, CUSUM tests, and rolling window estimates.
- Forecasting: A pseudo out-of-sample evaluation (1990Q1–2026Q1) compares the predictive content of the policy-conditional framework against benchmarks like the random walk and direct autoregressions. Diebold-Mariano tests assess forecast accuracy.
Conditional Projections:
- Using the estimated parameters, the paper generates conditional projections for unemployment from 2026Q2 to 2028Q2 under various deterministic policy paths (e.g., raising government expenditure by 1% of GDP).

Key Contributions

Analytical Distinction: The paper provides a compact framework that analytically separates cumulative output expansion from inflation dynamics. It demonstrates that full employment cannot be inferred from cumulative output growth alone; the outcome depends on whether policy-induced growth outpaces the growth of productive capacity.
Empirical Validation of Spurious Relationships: The study empirically demonstrates that regressing the unemployment rate on the level of output yields a spurious relationship. The valid empirical relationship exists only between unemployment and the output gap (output relative to potential), confirming that employment depends on capacity utilization rather than the absolute level of activity.
Policy Instrument Effectiveness: The paper quantifies the limited explanatory power of standard policy instruments. It finds that while public expenditure is positively associated with subsequent growth, the three instruments (expenditure, interest rate, tax rate) collectively explain only about 3% of the variance in quarterly output growth.

Key Results

Policy and Growth: Government expenditure relative to GDP is positively associated with subsequent output growth ( $\delta > 0$ ), but this association has weakened significantly since the mid-1970s. The interest rate coefficient is negative and significant only in pre-1975 samples. The tax rate shows no significant association. The policy-induced component of growth is small and does not improve upon forecasts based solely on the persistence of past output.
Unemployment and Output: The relationship between unemployment and the level of output is spurious. When the output gap is used, the relationship is robust: a one percentage point increase in the output gap is associated with a 0.62 percentage point decrease in the unemployment rate. This coefficient has strengthened over time, moving from approximately -0.38 in the 1960s to -1.16 in recent windows.
Inflation and Unemployment: The association between inflation and unemployment is small and unstable. Before 1971, a significant negative trade-off (Phillips curve) existed; post-1971, the coefficient is small and statistically insignificant. Consequently, inflation dynamics contribute minimally to unemployment projections in the model.
Forecasting: Conditioning output forecasts on policy instruments does not improve accuracy compared to using the history of output alone. The policy-conditional framework adds no predictive content beyond the persistence of output.
Conditional Projections: Under a baseline scenario, policy-induced growth (2.36% annually) slightly exceeds potential growth (2.13% annually), leading to a gradual decline in projected unemployment. A counterfactual scenario where government expenditure is raised by one percentage point of GDP for four quarters lowers the unemployment rate two years ahead by approximately 0.15 percentage points.

Significance and Claims
The paper claims that the limits of full-employment policy are not solely defined by the inflation-unemployment trade-off but are fundamentally constrained by the relationship between demand, productive capacity, and the instability of policy transmission mechanisms.

Limits of Demand Management: The results suggest that a demand-management strategy reduces unemployment only to the extent that it raises the growth of output above the growth of productive capacity. Sustained growth at the rate of potential output leaves the unemployment rate unchanged.
Instability and Endogeneity: The effectiveness of policy instruments is found to be small, unstable over time, and difficult to distinguish from the inherent persistence of output. The paper emphasizes that the estimated coefficients reflect conditional associations within a reduced-form framework rather than causal effects, noting that endogenous policy responses (counter-cyclical actions) likely weaken the observed reduced-form relationships.
Purchasing Power Constraint: While the statistical link between inflation and unemployment is weak in the data, the framework highlights the theoretical constraint of purchasing power: if prices rise faster than nominal incomes, real gains from expansion are eroded. The paper concludes that achieving full employment requires demand to outpace capacity, a condition that current policy instruments struggle to secure reliably.

More like this