US inflation and growth risk through expert eyes

Chart 1: Time-varying probability of a stagflationary episode in the US

Is inflation driven primarily by demand or by supply? And are price pressures transitory or persistent? These questions are central to policy decisions, shaping whether central banks tighten or hold back, whether governments stimulate or consolidate, and how policymakers balance price stability against growth. The Covid-19 shock made the answers especially elusive. Unlike most recent recessions, the pandemic triggered large and simultaneous disruptions to both demand and supply: lockdowns suppressed consumption even as unprecedented fiscal stimulus supported household incomes, global supply chains broke down, labour markets tightened and uncertainty surged. Subsequent geopolitical tensions have further complicated the outlook, keeping these questions at the forefront of the policy debate. In a recent paper (Mouabbi, Renne and Tschopp, 2025), we exploit surveys of professional forecasters to develop a framework that jointly evaluates risks to inflation and growth. Our methodology disentangles demand and supply drivers, and tracks evolving concerns about stagflationary episodes.

A new lens on inflation and growth risk

Surveys of Professional Forecasters are regular questionnaires that collect predictions from economists and other experts about future economic conditions, including inflation and GDP growth. These surveys are known to have a good forecasting performance relative to standard time-series models. In Mouabbi, Renne and Tschopp, 2025, we show that professional forecasters’ responses, by virtue of their forward-looking nature, can be used to jointly assess the risks to inflation and output growth. The probabilistic responses of these surveys are especially useful in periods of high uncertainty, when large shocks hit the economy, as they capture the full distribution of expected future inflation and growth. They give policymakers insight into the perceived likelihood of extreme outcomes, so-called tail risks, without the need for such fears to materialise, making them a rich and relevant source of information for our analysis.

Using a dynamic factor model that allows uncertainty and asymmetry to change over time, we study the joint behaviour of inflation and real activity in the United States. The model combines realised data and survey-based expectations, capturing the macroeconomic risk perceptions of informed agents. This approach makes it possible to distinguish between demand- and supply-driven forces and to track how risks evolve across the full distribution of future inflation and GDP growth. 

We identify demand and supply factors using simple economic principles: demand shocks tend to move inflation and real activity in the same direction, while supply shocks move them in opposite directions. We further refine this identification by allowing changes in the shape of the forecast distributions to inform the nature of the shocks. For example, rising downside risks to growth alongside upside risks to inflation point to supply-side stress – an especially useful signal when demand and supply shocks occur at the same time. Finally, the model decomposes inflation and growth into persistent trends and transitory cyclical components driven by the estimated latent factors, providing a clearer picture of the forces shaping prices and output.

The shifting forces behind US inflation and growth

Using quarterly US data from 1981Q3 to 2024Q1, spanning the Volcker disinflation, the Great Recession and the Covid-19 pandemic, we show that the forces driving the economy have shifted markedly over time. Before the Great Recession, and again since the pandemic, the output gap reflects a combination of supply and demand factors. By contrast, between 2008 and 2020, it was driven largely by demand. The period 2022-24 points to a sharp rise in the cyclical component of prices, mainly the result of negative supply shocks. 

This supports the view that part of the post-pandemic inflation surge stems from temporary supply disruptions. At the same time, the underlying inflation trend has steepened, signalling the presence of more persistent forces, mostly linked to strong demand. Taken together, the post-pandemic inflation episode appears to reflect a mix of temporary supply pressures and sustained demand strength (Chart 2).

Turning to macroeconomic risk, we find that inflation uncertainty and asymmetry were especially pronounced in the 1980s and during the Great Recession. For GDP growth, these features stand out in the 1980s and again in the period following the Covid-19 shock.

Chart 2: Decomposition of GDP growth and inflation into supply and demand factors
 

The return of stagflation risk in the US

We examine the relationship between inflation and economic activity through the lens of professional forecasters. The model shows that this relationship changes over time, sometimes even switching signs. Prior to the early 1990s, the correlation between inflation and growth is mildly negative or close to zero, suggesting that supply and demand forces largely offset each other. Since the Great Recession, demand forces have played a dominant role, reflected in a stronger and positive link between inflation and economic activity. This correlation peaks during the Covid-19 crisis. However, since the pandemic (and up to the end of our sample in 2024Q1), the relationship has weakened sharply, pointing to a growing role of supply-side factors in shaping inflation and growth. Moreover, we find that the inflation-growth correlation is negatively related to nominal term premiums. This aligns with structural models on the term structure of interest rates that show that nominal term premiums decrease as demand shocks become the primary driver of economic fluctuations.

Finally, we estimate the probability of a stagflationary episode from the model-implied joint distribution of future inflation and GDP growth, over short- to medium- term horizons (Chart 1). In 2022, these probabilities rose sharply, reaching levels not seen in three decades. While they remain relatively low – around one quarter of their early-1990s peak – the increase is striking after 30 years in which stagflation risk was close to zero.