A common view is that fiscal policy operates not only through direct demand effects, but also by boosting business optimism and therefore investment. This column uses novel data on Italian firms and public procurement contracts to test this hypothesis. It finds that a rise in government purchases makes firms overoptimistic about their future demand. This sentiment channel overturns the standard theoretical result that public spending crowds out investment and therefore leads to larger government spending multipliers. However, financial frictions can limit this amplification: when access to credit is constrained, the sentiment boost has little effect on investment.
How does government spending stimulate economic activity? A view common in policy debates and in the media is that fiscal policy operates not only through direct demand effects, but also by boosting business optimism and hence investment. This idea dates back to Keynes, who, at the height of the Great Depression, famously argued that government expenditures can move beliefs just as much as fundamentals (Keynes and Henderson 1929).
Yet, incorporating this idea in standard macroeconomic models has proven challenging. As noted by Mankiw (2009) and Cochrane (2009), the lack of empirical evidence on how government spending affects private sector beliefs has made it difficult to discipline models that feature such mechanisms. Hence, the literature has largely resorted to models with rational expectations to understand the effects of fiscal stimulus. In this context, there is no well-defined notion of ‘optimism’ and one cannot quantify the extent to which fiscal policy operates by shifting firm beliefs.
In a recent paper (Bellifemine et al. 2025), we bring new data to address this challenge. We show that a rise in government purchases makes firms overoptimistic about their future demand: their expected sales increase by more than actual sales. By stimulating private investment, this sentiment channel overturns the standard theoretical result that public spending crowds out investment and therefore leads to larger government spending multipliers. However, financial frictions can limit this amplification: when access to credit is constrained, the sentiment boost has little effect on investment.
Testing for a sentiment channel of fiscal policy poses an empirical challenge, as it requires simultaneously observing firms’ expectations, actual outcomes, and exposure to government spending — three variables rarely available together. To address this, we assemble a novel dataset linking survey data on Italian firms’ sales forecasts from the Bank of Italy, administrative income statement records, and the universe of public procurement contracts.
Our findings reveal that government spending moves business optimism. Exploiting a 2015 reform that gave some Italian municipalities unexpected fiscal space, we find that firms receiving more procurement contracts become overoptimistic about their future sales — forecasts systematically exceed realisations (Figure 1). The effect is large: a 10% boost to revenues from government contracts led firms’ forecasts to overshoot actual sales by 9% in the following year.
What drives this overoptimism? To better understand the mechanism, we examine how the public spending shock affects firms’ forecasts of different sources of revenue streams. We find that the optimism induced by public procurement is pervasive: firms also become more optimistic about export sales, despite these being entirely unrelated to domestic procurement. This suggests that managers extrapolate positive news about demand from the public sector to unrelated domains. This is consistent with evidence of cross-domain extrapolation recently documented in other economic settings.
Figure 1 Firms become overoptimistic following a positive government spending shock
Does this boost in optimism matter for the aggregate transmission of government spending? To answer this, we build a heterogeneous-firm New Keynesian model disciplined by our empirical results. In the standard model with rational expectations, government spending raises demand but also increases interest rates, which discourages private investment. As a result, the fiscal multiplier is modest: without the sentiment channel, our model predicts that one euro of public spending increases output by around 50 cents.
When we allow expectations to respond to government spending as they do in the data, the results change substantially. Because firms become more optimistic about future demand, they increase investment even as interest rates rise. Rather than being crowded out, private investment is then crowded in (Figure 2). This nearly doubles the government spending multiplier, from 0.5 to 1 — a value much more in line with empirical estimates.
This sentiment-driven crowding-in depends critically on financial conditions. While higher optimism encourages firms to revise upward their expectations of future demand, acting on these beliefs often requires access to external finance. Firms with strong balance sheets or easier credit access are better positioned to increase investment in response to optimism. In contrast, financially constrained firms are unable to scale up investment following a rise in optimism.
We test this mechanism directly in our panel data by comparing firms with different credit scores. The sensitivity of investment to optimism is significantly stronger among firms that are less financially constrained, confirming that financial frictions shape how expectations translate into real activity.
This interaction between sentiment and financial conditions has important implications for fiscal policy. It implies that the effectiveness of government spending is state-dependent: when credit conditions are favourable, fiscal policy generates strong sentiment effects and crowds in private investment, leading to large multipliers. But when credit markets are impaired — as during financial crises — optimism alone is not enough, and the multiplier is much smaller. As a result, our model predicts that the fiscal stimulus was roughly one-third less effective during the 2008–2012 financial crisis than during COVID, when credit conditions were less tight.
Figure 2 The effect of the sentiment channel on the transmission of fiscal policy
Our findings offer new evidence for an old idea: that fiscal policy works partly by boosting business sentiment. Government spending makes firms optimistic about future demand, and this optimism stimulates investment. The effect is large enough to overturn the standard prediction that public spending crowds out private investment, and to double the fiscal multiplier.
These results have practical implications for fiscal policy design. First, they suggest that the composition of a stimulus package may matter for its effectiveness. Procurement contracts — which provide direct, observable demand to firms — may generate stronger sentiment effects than transfers to households, although this remains a question that deserves further work.
Second, and perhaps more importantly, our state-dependency results carry a cautionary message. Fiscal stimulus is significantly less potent during financial crises — precisely when policymakers need it most. This has important implications for the policy mix aimed at stabilising the economy in a recession. If policymakers rely on fiscal spending to lift an economy out of a financial crisis, they may be disappointed unless they simultaneously repair the financial system.
Our results also raise new questions for fiscal policy design. If government spending works partly through sentiment, does the visibility of spending matter? Are infrastructure projects more effective than transfers precisely because they are more salient to firms? Should policymakers communicate differently about fiscal stimulus to amplify sentiment effects? These questions deserve further attention as we refine our understanding of how fiscal policy works.
Source: cepr.org
Macroeconomic forecasters have long struggled to accurately pinpoint the timing of future recessions. In response,…
Global services trade – as the future of global trade – is expanding rapidly (Baldwin…
Governments often provide capital and other direct funding to banks and nonfinancial firms during economic…
The European Commission's Autumn 2025 forecast points to a subdued but positive growth outlook for…
The debate over whether US members of Congress should be allowed to trade individual stocks…
Emerging markets have long been highly exposed to swings in global risk sentiment. During periods…