Why would people support policies that are macroeconomically unsound, in that they are more likely to lead to such events as sovereign crises, balance of payments crises, and the like? This may arise if decisive voters are likely to bear a lower fraction of the costs of the crisis, while benefitting from the short-run gains associated with those policies, such as greater public expenditure or lower taxes. I first discuss an illustrative model based on Saint-Paul et al. (2017), based on the assumption that in a crisis, not everybody can access his or her entitlement to publicly provided goods, a feature labelled “favoritism”. If the decisive voter is relatively favored in this rationing process, then people are more likely to finance public expenditure by debt, the greater the degree of favoritism. Furthermore, favoritism and the likelihood of a crisis raises the level of public spending. Next, I consider the choice between electing a “populist” who reneges on anonymity when allocating the public good, even in normal times, and a “technocrat” who sticks to anonymity, and does all it takes to balance the budget. I show that the support for the populist is greater, (i) the greater the likelihood of default, (ii) the more depressed the macroeconomic environment, (iii) the greater the inherited level of public debt and (iv) the lower the state’s fiscal capacity. I then argue that the model helps understanding some episodes in French pension reform. Some occupational groups supported unsustainable reductions in the retirement age because they expected that other workers would bear a higher proportion of the burden of future adjustment. Finally, using a panel of countries, I provide evidence in favor of some of the predictions of the model. As predicted, favoritism raises public debt, budget deficits, and public spending. It also raises the likelihood of a fiscal crisis through its effect on public debt. Furthermore, “populists” are more likely to conquer power, the higher the degree of debt and budget deficits, and the higher the level of government spending – the latter finding being consistent with the model’s prediction on the effect of fiscal capacity.