Blog

Nearly 40% of empirical papers in top economics journals use log specifications, and over a third of those face the log-of-zero problem. The most common fix — adding an arbitrary constant — is theoretically unjustified and can produce misleading results. iOLS provides a rigorous alternative that is as easy to implement as OLS.

Most energy transition plans assume a predictable future. Our simulations for Germany show that this blind spot leads to underinvestment, more blackouts, and higher costs — with the burden falling hardest on those least able to cope. The right environmental policies can fix this, even within a deterministic planning framework.

Using data on 340,000 children born in France, we show that the health effects of air pollution are concentrated among the most vulnerable 10% of infants — those born into lower-income families with poorer health at birth. Targeting policies based on vulnerability rather than exposure alone could be far more effective.

The cdid R package implements chained difference-in-differences estimators, providing applied researchers with a robust and flexible approach to causal inference under staggered treatment adoption.

A electricity supplier in Alberta deliberately timed ‘emergency’ outages of its power plants to drive up spot prices and boost its profits. Our study shows this strategy cost the province up to $600 million in a single year.

Notre étude révèle les coûts cachés des choix de transport à Montréal. La mobilité active génère des bénéfices sociaux nets, tandis que l’automobile impose des charges collectives considérables — avec des économies potentielles de près de 2 milliards de dollars par an grâce aux reports modaux.

Our study reveals the hidden costs of transportation choices in Montreal, showing that active mobility generates net social benefits while automobile use imposes substantial collective burdens — with potential savings of nearly $2 billion annually from modal shifts.