During the past 20 years, the world economy suffered two major crises – the global financial crisis of 2008-2009 (GFC hereafter) and the pandemic crisis of 2019-2020 (COVID-19 hereafter). The drivers of the two crises are fundamentally different since one was a financial crisis while the other was a public health crisis. The GFC, also known as the Great Recession, was ultimately the result of lax financial supervision, which failed to keep up with questionable financial engineering that sought to maximise yield in a low-interest rate environment. In contrast, COVID-19 generated a pandemic recession (Diebold 2020). Fatás et al. (2020) noted that past recessions have left permanent scars on long-term growth, known as hysteresis, and that proper policies should be put in place to minimise the long-term effects.
A common denominator between the two crises is that both impacted the entire world rather than just one region or one group of countries. In a recent paper (Aizenman et al. 2025), we analyse the patterns of recessions and recoveries of 101 advanced and developing economies, identifying the turning points of recessions and expansions between 1990 and 2022, and perform cross-country analysis of domestic and external drivers of economic recovery. In addition to the standard independent variables, we include institutional development, political stability, the extent of democracy, and trade restrictions indexes, and explore their roles in explaining recessions and recovery patterns.
Two distinct models of economic recessions can be identified. The first, a Hamiltonian recession, is derived from the pioneering work of James Hamilton (Hamilton 1989) and foresees recessions that prevent economies from returning to their pre-crisis growth trajectory (Cerra and Saxena 2008). This type of recession typically leads to a permanent reduction in an economy’s productive capacity and income level. The second model of recession, conceptualised in modern economic discourse by Milton Friedman (Friedman 1964, 1993), assumes dynamics known as a Friedman-like recession akin to the response of a stretched guitar string. The further the economy is pushed downward, the more forcefully it rebounds.
Productive capacity remains largely intact and the economy does not suffer a permanent loss of income. The supply side remains resilient, in contrast to the Hamiltonian scenario. Countercyclical monetary and fiscal policies may yield very different results in the two models.
To identify economic recessions and recovery, we use the Bry–Boschan algorithm. It automates the cycle-dating procedure in line with the NBER tradition (Bry and Boschan 1971). Using the Bry–Boschan algorithm, we identify 419 recessions in our sample of 101 countries over the period 1990-2022. We found that 59 recoveries occurred in 2009 (i.e. the GFC) and 94 occurred in 2020 (i.e. the COVID-19 crisis). Notably, the number of recessions during the COVID-19 crisis is twice as high as during the GFC, illustrating the significant impact of the pandemic. Although many emerging market economies (EMEs) experienced financial crises in the late 1990s and early 2000s, the number of recessions was not as high, suggesting that the crises in emerging market economies were regionally contained.
Figures 1 and 2 illustrate the different recovery patterns between industrial economies in Figure 1 and EMEs (Figure 2) in the context of comparing the recovery from the GFC and COVID-19 crises.
In Figure 1, the GFC seems to have had a longer lasting impact on this group, with recovery mostly sluggish. For instance, Switzerland and Canada only managed to reach their pre-crisis real output level in the first quarter of 2010. Meanwhile, the peripheral euro-area countries were subsequently hit hard by the euro crisis. In contrast, the impact of COVID-19 was much bigger than that of the GFC, with Japan and Spain suffering real GDP losses of 20%. However, the recovery was also much faster and stronger than during the GFC – the downturn lasted less than two quarters in most cases. For instance, the US and Switzerland managed to recover their pre-crisis real output in the second quarter of 2020. Once again, the recovery was more sluggish for peripheral euro area countries.
Figure 1 Comparing two recoveries: The GFC versus COVID-19 in industrialised economies
In Figure 2, we compare the two recoveries for a selective group of EMEs. In the left panels, we observe that recoveries after the GFC were faster and stronger in EMEs than in IDCs. Since GFC primarily affected the financial systems and real economies of financially well-developed advanced economies. In the left panel of Figure 2, we observe that the post-COVID-19 recovery pattern was similar for the broader group of EMEs and IDCs.
Figure 2 Comparing two recoveries: The GFC versus COVID-19 in emerging market economies
Regression results
We investigate the determinants of the variables related to recessions and recovery. Candidate variables include macroeconomic and institutional variables that are observed at the annual frequency. In addition to the macroeconomic variables identified as important in Eichengreen et al. (2024), we include institutional variables, based on the principal components (PCs) of political risk ratings in the ICRG database. The first PC is legal development, based on the ratings for bureaucratic quality, anti-corruption measures, and the respect of ‘law and order’. The second PC is a political stability index based on ICRG ratings for government stability, the lack of military in politics, and the lack of external and internal conflicts, religious tensions, and ethnic tensions. We also include aggregate trade restrictions (Estefania-Flores et al. 2024). With the return of trade tensions and restrictions at the global level since 2018 (Bown and Kolb 2022), we expect that trade restrictions may influence the extent of the recoveries after the most recent recession episodes.
First, we use panel logit models to estimate the probability of an economy entering a recession.
We find that higher government debt level or budget deficit, excessive credit creation, fuel importers, and greater exchange rate market pressures would lead to a higher probability of a recession.
The estimation results of the depth of recession suggest that tighter trade restrictions are associated with shallower recessions during the GFC and among industrialised countries. One possible explanation is that trade restrictions may help mitigate the impacts of external shocks and help stabilise the economy.
Given the heterogeneity of our sample economies, we obtain insightful results when we apply a panel logit estimation augmented with interaction terms. Worsening fiscal space (i.e. a higher budget deficit or government debt), being a major fuel importer, and excessive bank credit availability are associated with a higher likelihood of recession, while a higher level of political stability can help reduce the probability of a recession. When the level of political stability is sufficiently high, holding a higher amount of international reserves as a percentage of GDP reduces the probability of recession.
Higher levels of international reserve holdings reduce the probability of a recession, but only for low levels of trade restrictions (i.e. freer trade). This result echoes the finding of Aizenman et al. (2023) on the complementarity between the holding of international reserves and capital account restrictions in the context of terms-of-trade shocks. The buffer effect of international reserves is only observed when the economy is sufficiently open to trade. When the level of trade restriction is too high, the holding of international reserves is no longer associated with a reduction of the probability of a recession. When trade restrictions are too high, the buffer effect of macroeconomic variables disappears.
Next, we examine whether Hamilton’s model or Friedman’s model better depicts the recovery path in the aftermath of a recession. The results suggest that in a stable political environment, recessions during which GDP decreases by an additional 1% induce a stronger output recovery of around 0.9% after four quarters, and the length of the recession has no significant effects on the extent of the recovery four quarters later. When the number of trade restrictions is very low, recessions during which GDP decreases by an additional 1% induce a stronger output recovery of around 0.8% after four quarters, and the length of the recession has no significant effects on the extent of the recovery four quarters later.
This can be explained by the buffer effect of international reserves holding explored in Aizenman and Riera-Crichton (2008). For the whole sample, we find that deeper recessions are followed by stronger recoveries, in line with Friedman’s ‘plucking model’ of the business cycle. However, the impact becomes weaker if institutional development is limited and trade restrictions are tight. We show that recessions with political instability or trade tensions differ sharply from those without, which is highly relevant to the current global climate of heightened trade tensions and geopolitical uncertainty.
Summary
We analyse a large sample of industrialised and emerging countries between 1990 and 2022, a period of unprecedented trade and financial globalisation. We perform an in-depth analysis of the drivers of different patterns of recessions and recoveries, with a focus on the impact of political stability and institutional development. In addition, we empirically explore the role of trade restrictions in economic recovery. We also empirically test the validity of Friedman’s plucking model of the business cycle (Friedman, 1964, 1993). Notably, we provide global empirical evidence that Friedman’s plucking model is less relevant in describing an economy’s recovery path in the presence of political instability, weak institutions, and extensive trade restrictions.
Relative to industrial economies, EMEs tend to have weaker institutions and more restrictive trade barriers. Our empirical findings suggest that when policymakers seek to mitigate global shocks through countercyclical monetary or fiscal policy, these are more effective when the economy benefits from more political stability and fewer trade restrictions.
References
Aizenman, J, S H Ho, L D T Huynh, J Saadaoui and G S Uddin (2023), “Real exchange rate and international reserves in the era of financial integration”, VoxEU.org, 6 November.
Aizenman, J, H Ito, D Park, J Saadaoui and G S Uddin (2025), “Global Shocks, Institutional Development, and Trade Restrictions: What Can We Learn from Crises and Recoveries Between 1990 and 2022?”, NBER Working Paper No 33757.
Aizenman, J and D Riera-Crichton (2008), “Real exchange rate and international reserves in an era of growing financial and trade integration”, The Review of Economics and Statistics 90(4): 812-815.
Bown, C P and M Kolb (2022), “Trump’s trade war timeline: An up-to-date guide”, PIIE blog.
Bry, G and C Boschan (1971), “Programmed selection of cyclical turning points”, in Cyclical analysis of time series: Selected procedures and computer programs, NBER.
Cerra, V, A Fatás and S C Saxena (2020), “The persistence of a COVID-induced global recession”, VoxEU.org, 14 May.
Cerra, V and S C Saxena (2008), “Growth dynamics: the myth of economic recovery”, American Economic Review 98(1): 439-457.
Diebold, F X (2020), “Real-time real economic activity: Exiting the great recession and entering the pandemic recession”, NBER Working Paper No. 27482.
Dominguez, K M (2010), “International reserves and underdeveloped capital markets”, in NBER International Seminar on Macroeconomics, Vol. 6, No. 1, pp. 193-221.
Eichengreen, B, D Park and K Shin (2024), “Economic resilience: Why some countries recover more robustly than others from shocks”, Economic Modelling 136, 106748.
Estefania-Flores, J, D Furceri, S A Hannan, J D Ostry and A K Rose (2024), “A measurement of aggregate trade restrictions and their economic effects”, The World Bank Economic Review, lhae033.
Fatás, A (2019), “The 2020 (US) recession”, VoxEU.org, 14 March.
Fatás, A (2021), “The short-lived high-pressure economy”, VoxEU.org, 27 October.
Friedman, M (1964), “Monetary studies of the National Bureau”, The National Bureau Enters Its 45th Year 44: 7-25.
Friedman, M (1993), “The ‘plucking model’ of business fluctuations revisited”, Economic Inquiry 31(2): 171-177.
Hamilton, J D (1989), “A new approach to the economic analysis of nonstationary time series and the business cycle”, Econometrica: Journal of the Econometric Society, 357-384.
Kohlscheen, E, R Moessner and D M Rees (2024), “The shape of business cycles: A cross‐country analysis of Friedman’s plucking theory”, Kyklos 77(2): 351-370.