Research

Policy-makers of some fossil fuel-endowed countries wish to know if a given fossil fuel supply project is consistent with the global carbon budget that would prevent a 2 °C temperature rise. But while some studies have identified fossil fuel reserves that are inconsistent with the 2 °C carbon budget, they have not shown the effect on fossil fuel production costs and market prices. Focusing on oil, we develop an oil pricing and climate test model to which we apply future carbon prices and oil consumption from several global energy-economy-emissions models that simulate the energy supply and demand effects of the 2 °C carbon budget. Our oil price model includes key oil market attributes, notably upper and lower market share boundaries for different oil producer categories, such as OPEC. Using the distribution of the global model results as an indicator of uncertainty about future carbon prices and oil demand, we estimate the probability that a new investment of a given oil source category would be economically viable under the 2 °C carbon budget. In our case study of Canada’s oil sands, we find a less than 5% probability that oil sands investments, and therefore new oil pipelines, would be economically viable over the next three decades under the 2 °C carbon budget. Our sensitivity analysis finds that if OPEC agreed to reduce its market share to 30% by 2045, a significant reduction from its steady 40–45% of the past 25 years, then the probability of viable oil sands expansion rises to 30%.

Who buys imports? This question is fundamental to understanding which households are most exposed to trade shocks. By augmenting U.S. grocery purchase data to include the origin countries of both products and households, we provide the first direct evidence that immigrants exhibit substantially stronger preferences for imports than natives. Immigrants therefore disproportionately gain from trade, and higher import prices reduce consumer welfare by 42% more for immigrants than for natives. We estimate a quantitative model of trade to show that immigrants also reduce trade costs, thereby generating gains from trade for natives due to increased import supply. These gains accrue primarily to urban and high-income households. In the aggregate, however, immigrants generate nearly three times as much import expenditure via their preferences than by reducing trade costs.

This paper evaluates the extent to which additional Canadian pipeline infrastructure might mitigate the cost of proposed U.S. tariffs on crude oil exports. We generate three key findings via a highly disaggregated computable general equilibrium model of the North American crude petroleum industry. (1) Under the existing pipeline network, a 10% tariff applied by the U.S. generates an annual export revenue loss of $10.6B for Canadian crude producers. (2) If previously canceled east-west pipelines (Energy East and Northern Gateway) were operating at full capacity, tariff costs would be reduced by a third and the tariff elasticity of substitution would effectively triple. (3) To fully insulate producers from U.S. tariffs, pipeline capacity to tidewater would have to expand by a factor of 4.5. Tariff pass-through to U.S. refineries is 20 percentage points higher when canceled pipelines are operational versus baseline, but this benefit is masked when measured using average prices: producers endogenously concentrate exports of low-price bitumen in tariff-affected export markets and substitute high-price lighter grades of crude to non-tariff destinations. We therefore urge caution in interpreting ex-post empirical estimates of pass-through in this setting: such estimates may exhibit downward bias due to composition effects. 

This paper links product country-of-origin data to US household purchase records and estimates a model of import demand with flexible consumer heterogeneity and attribute-based import substitution patterns. Counterfactual price shocks applied to all imported varieties reveal concentrated costs in urban US counties, while China-specific price shocks are strongly anti-rural. I provide evidence that differences in retail market composition drive these disparate outcomes by acting both in addition to, and independently of, household income. I provide evidence that import substitutability is greater than under typical Armington assumptions and justify the modeling choice of this paper by providing evidence that variety attributes – such as quality and markups – are highly separable between the location of production and the firm which designed that variety.