Estimating path dependence in clean energy transitions

About

Coal is known to be the most climate-damaging fossil fuel, and as a consequence, it is widely recognized that the global economy must permanently transition away from using coal in order to address climate change - but how can a sustained energy transition away from coal be induced? Permanent interventions like a tax on coal would help to facilitate such a transition, but permanent policy interventions may be unrealistic when governments have difficulty committing to long-term policies. Indeed, the history of climate policies is filled with examples of policy revisions, reversals, and withdrawals. 

The theoretical alternative is that a large but temporary intervention that lowers coal use can permanently overcome coal’s abundant supply. Under such circumstances, a sustained long-term transition away from coal could be achieved even after the intervention is lifted. Economies with this feature are broadly characterized as having strong path dependence in energy transitions, path dependence being the dependence of economic outcomes on the path of previous outcomes, rather than simply on current conditions. Whether such dynamics actually govern energy transitions, however, remains an open empirical question. We conducted one of the first empirical studies of long-run energy transitions, specifically examining the transitional dynamics of the U.S. electricity sector over the 20th century.

Approach

We combined modern and historical power plant records to construct a new dataset of county-level, fuel-specific electricity capital for the U.S. midwest across the 20th century to enable an analysis of long-run energy transitions capturing changes in the fuel composition of electricity capital. Using an event study design with county-by-decade panel data, we examined the historical shock induced by the introduction of mechanized mining in the early 20th century. Mechanized mining allowed extraction over previously inaccessible coal held in deep underground deposits, altering coal transport distances and thus the spatial distribution of delivered coal prices. 

We then developed a model of structural change for the electricity sector at the county level featuring increasing returns to scale and local productivity effects to estimate path dependence and the long-run elasticity of substitution between coal and other fuels. We then calibrated the model to draw lessons for future U.S. policies by simulating future electricity sector carbon emissions for the average U.S. county following temporary relative coal price shocks of varying magnitude and duration.

Key Findings

For a better than 50% chance of achieving a permanent switch away from coal, and thus weakly declining carbon emissions, a temporary shock equal in magnitude to recent high relative coal prices (e.g., due to natural gas hydraulic fracturing) must last at least five decades. Alternatively, if the shock can only last one decade, it must then be six times higher than that of recent prices to trigger sustained fuel switching. Altogether, these simulations conclude that in the absence of climate policy, recent economic conditions are insufficient for sustaining a permanent U.S. energy transition away from coal.