US headline and core inflation have increased sharply in June and July 2021, reaching the highest levels in more than a decade. Transitory factors and base effects have played a major role in recent US inflation developments. Moreover, a significant part of the observed strong pick-up in US inflation has been driven by a small number of items associated with the rebound in consumer spending following the re-opening of the economy (Figure 1), namely, developments in energy prices, demand for travel, prices for used cars (a close substitute to new cars), global supply disruptions – particularly chip shortages disrupting car production – and shipping delays. Many of these factors are expected to be largely temporary and their impact on inflation should wane amidst a gradual resolution of supply bottlenecks and stabilising demand.
Figure 1 Drivers of consumer price inflation (month-on-month percentage changes and contribution)
Source: Bureau of Economic Analysis and ECB calculations.
Risks to inflation being more persistent
However, rising US inflation has highlighted concern about risks to the inflation outlook prompted by the rapid rebound of the US economy in the aftermath of the COVID-19 recession, the large US fiscal stimulus and a pick-up in inflation expectations. The large scale of the fiscal response of the US administration to the COVID-19 crisis could elicit more persistent demand pressures and, thereby, compound the inflationary pressures. The Biden administration passed in March 2021 a recovery package, the American Rescue Plan (ARP), of around 8.8% of 2020 US GDP ($1.84 trillion), and, subsequently, two additional packages were proposed: the American Jobs Plan (AJP) to scale up infrastructure investment, and the American Families Plan (AFP) that focuses on welfare spending and tax credits to advance social objectives.1 As a result, US economic activity is expected to exceed potential output already in the second half of 2021 and record a positive output gap over the next years. Whether – and mostly how – this will affect inflation is at the epicentre of an ongoing policy debate. While some are warning of the surge of inflation (Blanchard 2021, Summers 2021), others are less worried and suggest that there are low chances to face an inflationary spiral similar to what observed in the 1960s (Ball 2021, Gopinath 2021).
Furthermore, short-term inflation expectations have increased at shorter time horizons, although longer-term inflation expectations appear better anchored (Figure 2). The rise in financial market-based inflation compensation measures may also reflect risk premia dynamics rather than genuine increases in inflation expectations. However, one could not preclude that a more persistent rise in inflationary pressures could feed into more persistent increases in inflation expectations with the risk of generating feedback loops as experienced in the 1960s/1970s.
Figure 2 Inflation expectations (percentage points)
Source: US authorities, Haver and ECB staff calculations
Quantifying the upside risks to US inflation
A quantification of these risks to the US inflation outlook is attempted through the lens of the ECB Global model (Georgiadis et al. 2021). The policy simulation scenarios are motivated by three questions: What if the Phillips curve slope is higher than commonly assumed? What if fiscal multipliers are larger than expected? And what if inflation expectations rise and eventually become de-anchored?
Under our baseline scenario, the fiscal stimulus implies a relatively moderate increase in CPI inflation, peaking in 2022 and raising by around 0.3 percentage points in 2022–23 (Figure 3A). Our baseline estimates imply a fiscal multiplier of about 0.8, while the assumed Phillips curve slope is around 0.1,2 which is broadly in line with recent academic estimates for the US economy (e.g. Bianchi et al. 2021). Moreover, we assume that inflation expectations remain anchored and that monetary policy is constrained to the median response – as indicated by the June Fed dot plot – which implies that the Federal Reserve will not raise interest rates before 2023.
In scenario I, we depart from our baseline by assuming a different Phillips curve slope coefficient. Indeed, the empirical literature reveals high uncertainty on the size of the Phillips curve slope. Some studies, for example, find higher slope estimates depending on model specifications and on the way inflation expectations are treated (Coibion et al. 2018). Therefore, our first alternative scenario assumes a higher slope coefficient of 0.3. Such a steeper Phillips curve slope would raise the inflation impact from the fiscal stimulus by 0.8 percentage points at its peak (Figure 3.A) compared to the flat Phillips curve baseline.
Scenario II mainly relates to the transmission of the fiscal shock to economic activity. The baseline scenario is based on an average fiscal multiplier of 0.8, but several reasons could argue in favour of a higher fiscal multiplier. First, US households may release part of their excess savings accumulated during the pandemic period (Attinasi et al. 2021), implying a larger impact of the fiscal stimulus. Second, the AJP is specifically designed to scale up infrastructure investment, and some studies suggest that multipliers for investment projects are larger than multipliers for government consumption. Third, there is evidence that multipliers are higher in recessions (Auerbach and Gorodnichenko 2012).3 To account for those uncertainties, we also consider a larger fiscal multiplier of 1 compared to 0.8 in the baseline scenario, which would raise GDP by an additional 0.9% and inflation by 0.1 percentage points in 2023 (Figure 3.B).
Source: ECB-Global and ECB staff calculations.
Scenario III assesses the impact from rising inflation expectation and from a potential de-anchoring of expectations from the Fed’s target. In choosing the magnitude of the shock, we look at changes in the cross-sectional forecast distribution of long-term inflation forecasts from the Survey of Professional Forecasters and construct two scenarios: in the moderate increase scenario, long-term inflation expectations increase by 0.2 percentage points; in the high increase scenario, inflation expectations increase by 0.4 percentage points.4 In addition, we consider also an exceptional de-anchoring scenario in which agents adjust their inflation expectations as they observe higher inflation outcomes. In all the scenarios, higher inflation expectations lead to higher inflation outcomes. The strongest impact is observed in the ‘de-anchoring scenario’, which results in an increase in inflation of around 1 percentage point above baseline by 2023 (Figure 3.C).
Overall, the unprecedented large fiscal stimulus could shift US inflation upward, but model simulations suggest inflation risks are moderate, although more persistent inflationary pressures could emerge if, in an unlikely scenario, all the risk factors materialise simultaneously. As shown in Figure 4, a steeper slope of the Phillips curve could raise US inflation by an additional 0.7 percentage points in 2023. Adding in a higher fiscal multiplier could raise inflation by another 0.3 percentage points. Finally, in a ‘worst’ upside scenario with inflation expectations getting de-anchored and being strongly influenced by past inflation (in addition to a steeper Phillips curve and higher fiscal multipliers), inflation could increase by 1 percentage point and persistently remain above the Fed target. In such a scenario, however, it would be likely that the Fed would withdraw monetary accommodation earlier than currently suggested by the ‘dot plot’, thus attenuating inflationary pressures.
Figure 4 Additional model-based effects on CPI inflation 2023 (percentage points)
Source: ECB-Global and ECB staff calculations.
Notes: The chart shows the effects of the fiscal stimulus on CPI inflation in 2023 under different assumptions. In the baseline scenario the fiscal packages contribute to an increase of 0.3 percentage points in inflation, while a steeper Phillips curve would imply an extra 0.7 percentage points. A higher multiplier with a flat Phillips curve increases inflation by 0.1 percentage points with regards to the first scenario, with the steeper Phillips curve case that adds an extra 0.3 percentage points compared to all the previous. Same reasoning for the inflation expectation shock scenarios.
These simulations are only illustrative and subject to high uncertainty. Uncertainty stems from a variety of sources. The size, composition, and pace of implementation of further fiscal stimulus is still the subject of debate. In addition, a faster-than-expected unwinding of the excess saving accumulated by households could substantially increase inflationary pressures.
Attinasi, M G, A Bobasu and A S Manu (2021a), “Winding down of the excess savings and its implications for outlook”, European Central Bank.
Auerbach, A J and Y Gorodnichenko (2012), “Measuring the output responses to fiscal policy”, American Economic Journal: Economic Policy 4(2): 1–27.
Ball, L, G Gopinath, D Leig P Mitra and A Spilimbergo (2021), “US Inflation: Set for Takeoff?”, VoxEU. org, 7 May.
Bianchi, F, J D Fisher, L Melosi et al. (2021), “Some inflation scenarios for the American Rescue Plan Act of 2021”, Chicago Fed Letter 453: 1–8.
Blanchard, O (2021), “In defense of concerns over the $1.9 trillion relief plan”, Realtime Economic Issues Watch 18, Peterson Institute for International Economics.
Blanchard, O (2016), “The Phillips Curve: Back to the’60s?”, American Economic Review 106(5): 31–34.
Coenen, G, P Karadi, S Schmidt and A Warne (2018), “The New Area-Wide Model II: an extended version of the ECB’s micro-founded model for forecasting and policy analysis with a Financial sector”, ECB Working Paper 2200.
Coibion, O, Y Gorodnichenko and R Kamdar (2018), “The formation of expectations, inflation, and the Phillips Curve”, Journal of Economic Literature 56(4): 1447–91.
Georgiadis, G, S Hildebrand, M Ricci, B Schumann and B Van Roye (2021), “ECB-Global 2.0: A global macroeconomic model with dominant-currency pricing, tariffs and trade diversion”, ECB Working Paper 2530.
Gopinath, G (2021), “Structural Factors and Central Bank Credibility Limit Inflation Risks”, IMFBlog, 29 February.
Powell, J (2021), “Virtual Hearing - Oversight of the Treasury Department’s and Federal Reserve’s Pandemic Response”, 23 March.
Summers, L (2021). “The Biden stimulus is admirably ambitious. But it brings some big risks, too”, Washington Post, 4 February.
Yellen, J L (2015), “Inflation dynamics and monetary policy: A speech at the Philip Gamble Memorial Lecture”, University of Massachusetts, 24 September.
1 Although there is some uncertainty on the final configuration of the fiscal packages under negotiation, the expectation is that one will cover infrastructure and another one social safety net aspects and green investment, for amounts that could be comparable to the initial proposals for AJP and AFP. For that reason, the information on those are used in our scenario analysis.
2 In ECB-Global, the true Phillips curve slope is a parameter measuring the change of inflation to changes in real marginal costs. To facilitate comparison with empirical estimates and public discussion on the issue, in the analysis above we refer to the ex-post elasticity of inflation to GDP changes as the 'Phillips curve slope'. In the reminder of the analysis higher slope/elasticities are obtained by modifying this parameter.
3 For this recession in particular, households might still spend some of the excess savings they accumulated during the COVID crisis once the pandemic is fully under control.
4 This is calibrated on the difference between median expected long-term inflation in 2020Q4 SPF and the most recent 2021Q2 survey.
5 To study the effects of a de-anchoring of inflation expectations ECB-Global has been modified along the lines of the NAWM II (Coenen et al. 2018). In this set-up, private-sector agent’s perception of the central bank’s inflation target is allowed to differ from the central bank long-run target thus allowing for a possible drift of longer-term inflation expectations. More precisely, the perceived inflation target is determined by an adaptive scheme according to which the current value of perceived inflation is determined by a weighted average of past inflation and the value of the perceived inflation objective in the previous period.