Deep Trade Agreements and UK Industries of the Future

Research Proposal Abstract All modern industries rely, to some degree, on international trade
and trade agreements with partnering nations. Reduced trade barriers can have many
economic benefits, from increased welfare and national income, advances in innovation,
reduced costs from economies of scale, and increased competition. In 2017, the UK
government's industrial strategy white paper emphasised the "industries of the future"
(IOTF), such as artificial intelligence, big data, clean energy, and self-driving vehicles, i.e.,
industries having key strategic value to the UK economy. Negotiating tailored trade deals to
best nurture these industries is critical for post-Brexit international trade. Modern trade
agreements contain many provisions besides tariff reductions in areas as diverse as services
trade, competition policy, trade-related investment measures, or public procurement. Existing
research has struggled with overfitting and severe multicollinearity problems when
estimating the effects of these provisions on trade flows. Recent research has tried to move
beyond estimating the overall impact of Preferential Trade Agreements (PTAs) and
establishing the relative importance of individual trade agreement provisions in determining
an agreement's overall impact (e.g., Kohl, Brakman, and Garretsen, 2016, Mulabdic, Osnago,
and Ruta, 2017, Dhingra, Freeman, and Mavroeidi, 2018). Most provisions jointly appear in
several trade deals, thus creating a high degree of collinearity among them. Mattoo et al.
(2017) use the number of provisions in a deal to measure the depth, Dhingra et al. (2018)
instead of group provisions in relatively small groups of bundles. Both approaches to
overcome collinearity come at the cost of losing important information about the contribution
of specific provisions. I will be using the techniques of Breinlich, Corradi, Rocha, Ruta, Santos
Silva and Zylkin (2021). They instead use the complete set of individual provisions and
machine learning techniques, such as LASSO (Least Absolute Shrinkage and Selection
Operator), to choose the subset that affects bilateral trades adaptively. The main goal of my
research is to use data disaggregated by industries (e.g., Chen and Novy 2011) to study and
identify the effects of provisions on bilateral trade flows in specific industries, with an
emphasis on the IOTF. This paper will build on recent developments in the machine learning
and variable selection literature to propose novel data-driven methods for selecting the most
important provisions and quantifying their impact on trade flows in these specific sectors. I
aim to contribute to the international trade literature by merging the big data/machine
learning approach of Breinlich et al. (2021) to the sectorial disaggregated approach of Chen
and Novy (2011). My research will also contribute to estimating the gravity equation with
four-layer panel data, i.e., in the presence of export and import countries, industries, and
time. To consider the role of time and country-dependent time effect would not be trivial. I
aim at devising computationally feasible estimators, extending, e.g., Larch et al. (2019) to
industry disaggregated data. I will apply these methods to a comprehensive data set recently
made available by the World Bank of PTA provisions (Mattoo, Rocha and Ruta, 2020). This
data set provides evidence on the content of trade agreements both at the extensive margin
(number of policy areas covered) and the intensive margin (commitments within a policy
area). Importantly, my dissertation will have an impact outside of academia. I believe that
identifying the provisions that affect exports in a given industry may advise a government
about which agreements/provisions it should subscribe to in negotiations.