Modelling the economic impact of US tariffs on Scotland and the rest of the world: technical paper

Methodology

In its simplified formulation, the gravity model predicts international trade flows based on economic sizes and distance between two countries (in analogy to Newtonian gravity). Modern gravity models can take account of many other variables that can predict trade, including sharing a common language and having a free trade agreement.

In this analysis, we use the Scottish Government’s in-house gravity model, based on the GEGravity package^[3] to carry out general equilibrium simulation of the desired scenarios.

Scenario definition

In order to turn the scenarios laid out in the previous section into a form suitable for simulation, we require product-level tariff and trade data to use as a baseline. We obtain this from the World Integrated Trade Solution (WITS) from the World Bank.^[4] Our model contains 15 broad goods sectors, so to construct model scenarios, we aggregate from the product-level tariff data and scenario definitions using a combination of nomenclature concordances from WITS and the detailed documentation of the individual nomenclatures. This aggregation process involves taking averages of the tariffs applied at the product level, weighted by the trade flows between each country, for all products within each of the 15 broad goods sectors.

Simulation dataset

The gravity simulation requires international bilateral trade data. For this, we use the Scottish International Bilateral Trade Dataset for Simulation (SIBTD-S), which was created by OCEA analysts by using a range of trade and production statistics for Scotland and the UK to split UK trade flows in the International Trade and Production Database for Simulation (ITPD-S), created by the United States International Trade Commission.^[5]

SIBTD-S contains 44 countries (including Scotland and rest-of-UK), 25 sectors (15 of which are goods), and covers the years 2002 to 2018. For this analysis, we use the most recent year, 2018.

Trade elasticity

The simulation requires input assumptions for the trade elasticity, which quantifies the direct change in trade value as a result of a change in trade costs. We use Egger et al (2018).^[6]

Simulation method

The GEGravity package uses a one sector Armington-CES trade model to solve for general equilibrium effects of changes in trade costs. The model considers a single factor of production – labour – and treats trade imbalances as an additive component of national expenditure.

We use GEGravity to run several different general equilibrium simulations: one for each of the 15 goods sectors in the model and one for the sum of all sectors. This allows us to obtain some degree of sectoral information, but it has the limitation that all sectors are treated separately, meaning that our model cannot capture “downstream” effects on one sector caused by direct effects on another sector. This is discussed further in the Limitations section.

Limitations

The continuously evolving picture of global tariffs presents a significant challenge to modelling. Gravity modelling is inherently long-term; the results we report here are based on the assumption that the tariff scenarios modelled will be in place for many years, giving the world economy time to adjust to the new situation. The model results describe the “new equilibrium”, and do not capture any short-term details of how that equilibrium is achieved.

We also do not attempt to model other potential changes in the domestic or global economy that could take place, unrelated to the tariffs. Hence, the model does not provide a forecast or prediction of how the economy will look in the future – it only compares two different versions of the future: one with and one without the new tariff scenarios.

The tariffs implemented by the US and retaliating countries have changed several times in recent months and this uncertainty is likely to continue. This uncertainty can, in itself, have an economic impact, as businesses and investors may hesitate to make some decisions in the uncertain environment. The model does not capture the direct or indirect effects of this uncertainty, nor changes in investment patterns that may result from the tariff situation.

Finally, as discussed above, the model treats each sector separately, not taking into account supply chain effects. For example, in the real economy, a steel tariff is likely to increase costs of production for various parts of the manufacturing sector which rely on steel as an input, but this model would not be able to capture such an effect.

These limitations mean that some care is required in interpreting the model results. Nevertheless, the results provide an indication of where the main risks and opportunities in terms of international trade flows are likely to be.