Working Papers
[Abstract]
Carbon pricing is a key policy instrument to induce a green transition. However, political support for carbon pricing is uncertain and may change over time. I study the green transition in the electricity sector when climate policy switches stochastically between regimes with and without a carbon tax. I find that regime risk slows the transition: relative to a deterministic tax, green electricity capacity builds up more slowly and brown capacity is reduced at a slower pace, resulting in more carbon emissions during the transition. The rate at which the government switches regimes has only a limited effect on median outcomes at the end of the century. But longer expected durations lead to larger forecast intervals for the transition paths, with worse outcomes in bad scenarios. Thus, while fast switching of regimes reduces the effective tax rate over the whole transition horizon, it also mitigates the risk of being stuck in a no-tax regime for a majority of the transition period. Overall, a higher carbon price accelerates the transition more than changes in the expected duration of governments.Political Competition and Climate Policy: A Dynamic Game of Pollution Control (February 2025) [Abstract]
In this paper, I analyze a dynamic game of political pollution control where two policy-motivated parties compete for votes through tax policies. Voters evaluate the direct effects of the tax policies and vote probabilistically. Depending on the voters' time horizon, the ambition of the green party's climate policy is drastically affected. If voters only consider one legislative period, the green party's climate policy is virtually indistinguishable from the brown party's to remain politically competitive. The brown party is barely affected by the political competition and can almost exactly implement its dictatorially preferred policy. Extending the voter's time horizon allows the green party to be more ambitious and forces the brown party to move closer to the median voter's preferred policy. I find that voters would need a time horizon close to ten legislative periods for the steady-state pollution stock to be lower than under random voting. This implies that for shorter time horizons the green party is better served not making the pollution issue a political campaign topic. Increased polarization in party objectives/beliefs decreases societal welfare and increases both the level and the volatility of steady-state pollution.
Learning dynamics in a network of Cournot economies (July 2023) [Abstract]
In this paper, firms in a network of Cournot economies learn about the intercept of the demand curve using past sales history. I decompose the learning process into economically interpretable components and find that learning aggregate quantities happens at a faster rate than individual quantities both within markets and within firms. This speed depends on the network topology and the slope of the demand function. The slowest learning component is the distribution of a correct aggregate amount between markets, which drives the slow convergence of individual quantities. The convergence rate of individual quantities is the same for all sufficiently connected networks and is independent of the slope of the demand function. Increasing the density of a random network has a non-monotonic effect on the convergence speed of aggregate quantities. Convergence speeds first decrease relative to isolated market-firm pairs, increasing again after the graph becomes sufficiently connected.