From microeconomic favouritism to macroeconomic populism

Gilles Saint-Paul 07 March 2019



Why would people support macroeconomic policies that are more likely to lead to sovereign or balance-of-payments crises? Dornbusch and Edwards (1991) noted that policies that they labelled 'macroeconomic populism' continually recur, particularly in Latin America, even though they typically end in crisis and painful internal and external adjustment.

In my recent work with Andrea Vindigni and Davide Ticchi (Saint-Paul et al. 2017), we argued that it would be rational for some social groups to support those policies. This is because the adjustment mechanism was biased in their favour for institutional or cultural reasons, and so they bore a lower fraction of the costs of the crisis – a feature we called 'favouritism'. When the painful fiscal adjustment came, some social groups had better access to publicly provided goods, often rationed, than others.

There are many reasons for this, including differences in political influence in bargaining over the burden of adjustment, the structure of taxation, outright discrimination, differences in access to social networks, and corruption. 

In our work we are agnostic with respect to these root causes. Our key assumption is that people are entitled to some reference level of publicly provided goods (healthcare, education, daycare, and so on) and that they are rationed in accessing their entitlement in a fiscal crisis. This process makes some people 'more equal than others', and we show that the likely winners have an incentive to engineer the crisis by supporting macroeconomic populism.

In a recent paper on how 'microeconomic favouritism' becomes 'macroeconomic populism' (Saint-Paul 2018), I provide some additional insights, as well as evidence about the mechanism. 

I consider an election between a populist who favours some social groups when allocating public goods, and a technocrat who sticks to anonymity and will strive to balance the budget so that as many citizens as possible can access their entitlement of publicly provided goods. In other words, people can now choose the degree of favouritism they get from the state, with the populist being, by assumption, associated with greater favouritism than the technocrat.

If the support for the populist is greater, there is a greater likelihood of a fiscal crisis. The reason is simple: the resource costs of rationing access to public goods have to be incurred in a crisis, regardless of whether the country is run by a technocrat or a populist. This makes a populist relatively more attractive during crises. 

In particular, the analysis predicts that populists are more likely to conquer power if:

  • there is a greater the likelihood of default,
  • the macroeconomic environment is more depressed,
  • there is a higher inherited level of public debt, or 
  • the state has a lower fiscal capacity. 

These mechanisms are illustrated by the history of French pension reform.

The French pension reform saga

In 1981, voters elected the left-wing administration of President François Mitterrand, with disproportionate support from civil servants (Rouban 1999). Mitterrand had promised to reduce the retirement age from 65 to 60, despite projections that this would be financially unsustainable in the long run.

Workers are treated differently by the French pension system depending on their industry, occupation, or type of labour contract. Therefore, these different régimes have different financing needs, and so restoring fiscal balance involve may imply one cross-subsidises another. 

These inequities reflect the bargaining strengths of different social groups. Civil servants are better organised than private sector employees, and are likely to be favoured at times of fiscal consolidation of the pension system. This makes them more likely to favour unsustainable increases in the generosity of the system, which helps explain their disproportionate support for Mitterrand in 1981.

This is confirmed by the first fiscal consolidation that followed Mitterrand’s election, the 1993 pension reform implemented under Prime Minister Edouard Balladur. While not formally overturning the change in retirement age, the duration of contributions rose from 37.5 years to 40 years for private sector employees only. The Balladur reform also toughened conditions for retirement for private sector employees in other dimensions, in particular the indexation of the pension. This made it less likely that private sector employees would retire at 60, and so gave public employees more favourable terms.

The reform had curbed expenditures on the general régime, while expenditures on the régime for civil servants continued to grow at a higher rate. The right-wing Balladur administration did not reform public sector pensions because of the superior ability of public employees' unions to organise. 

Consistent with my theory, it was rational for public sector employees to support Mitterrand in the 1981 election, despite overwhelming evidence that the reduction in the retirement age was fiscally unsustainable. Public sector voters correctly expected that any subsequent adjustment would fall predominantly on private sector employees, who then effectively subsidised the superior entitlement granted to civil servants.

Statistical evidence on favouritism and fiscal indiscipline

The Institutional Profiles Database (IPD), a panel of country surveys on institutional quality, provides statistical evidence for this theory. I estimated the effect of equality of treatment by the state on macro fiscal performance measures. I controlled for other indicators of institutional quality that are likely to affect aggregate fiscal discipline: fiscal capacity, conflict, and corruption. 

The data show that better equality of treatment reduces the debt-to-GDP ratio. The point estimates suggest that, all else equal, countries that have the highest degree of equality of treatment have a debt-to-GDP ratio that is 10 to 15 percentage points lower than countries with the lowest level of equality of treatment. Higher fiscal capacity has a positive effect on debt, suggesting it makes it easier for governments to borrow. 

Corruption, which may be interpreted as another measure of unequal treatment, also appears to raise debt levels. The conclusion is the same if budget deficits are used as the dependent variable instead of public debt.

My model predicts that the decisive voter will tend to vote for higher public expenditure, the greater the degree of favouritism. This is also validated by the data. The magnitude of the effect is 8-10 points of GDP between countries with the most and least favouritism.

So does favouritism make fiscal crises more likely to occur? An estimate of the likelihood of sovereign default based on the CRAG database (Beers and Mavalwalla 2017) is not so convincing. It provides mild support for a direct effect of favouritism on the likelihood of fiscal crises, but the effect is not statistically significant. The outstanding level of public debt, however, has (unsurprisingly) a strong positive effect on default, and so favouritism makes crises more likely through the indirect channel of raising public debt. 

When do populists conquer power?

To measure whether adverse fiscal and macroeconomic conditions help populist governments, I used the Interamerican Development Bank's Database of Political Indicators, and defined a populist platform as either nationalist, rural, regional, or religious. 

We might expect adverse macroeconomic and fiscal conditions to raise the likelihood of a populist government. Empirical results support this, although not entirely. Macro variables were not significant, though fiscal variables were and have the predicted effect on the likelihood of a populist coming to power. As expected, debt has a positive significant effect on populism, while the budget surplus has a negative significant effect.

Finally, a higher level of spending reduces the likelihood of populism. This is not surprising –controlling for net lending, greater spending means greater revenues, hence a greater fiscal capacity, which has a negative predicted effect on the likelihood of populism. 


So inequality of treatment is associated with higher levels of public debt, public spending, and public deficits, consistent with the favouritism theory of fiscal indiscipline. Furthermore, the greater the fiscal adjustment required, the more likely it is that voters elect a populist who is more likely to discriminate between groups. 


Beers, D and J Mavalwalla (2017), "Database of Sovereign Defaults", Bank of Canada Technical Report 101.  

Dornbusch, R and S Edwards (eds) (1991), The Macroeconomics of Populism in Latin America, University of Chicago Press.

Rouban, L (1999), "Les Attitudes politiques des fonctionnaires: vingt ans d'évolution", Les cahiers du CEVIPOF 24.  

Saint-Paul, G, D Ticchi and A Vindigni (2017), "Engineering Crises: Favoritism and Strategic Fiscal Indiscipline", CEPR Discussion Paper 12291.

Saint-Paul, G (2018), “From Microeconomic Favoritism to Macroeconomic Populism”, CEPR Discussion Paper 13434.



Topics:  Macroeconomic policy Politics and economics

Tags:  populism, pension reform, democracy, crises, favouritism

Professor, ENS-PSL, Paris School of Economics and NYUAD


CEPR Policy Research