Missing but not Forgotten: Modelling New Developments in the Global Economy Using Data with Missing Values

The global economy is changing. The 1990s and 2000s witnessed an acceleration in trade and financial integration between countries. Now globalisation is slowing, giving way to geopolitical tensions and economic fragmentation.

Developing and emerging economies (DEEs) play an increasingly important role. To respond swiftly to sudden shifts in these global economic conditions, policymakers in central banks and international institutions require timely forecasts informed by different types of data. However, time series data on global indicators, DEEs or from non-traditional sources are "short", suffering from missing values due to later start dates and/or release delays. Currently, such data is routinely dropped from analysis.

Building on a nascent literature and our research for the Economic Statistics Centre of Excellence (ESCoE), funded by the Office for National Statistics, this topical studentship will further develop Bayesian econometric methods for incorporating short data with missing values into forecasting models. The existing literature and our ESCoE research point towards two unresolved issues.

First, existing studies consider missing data issues from a domestic perspective. However, jointly modelling developments in multiple countries can improve forecast accuracy and allow us analyse interdependencies between economies. Second, current research deploys a two-step approach, first "filling in" missing values using statistical techniques and then estimating the forecasting model. This leads to "generated regressors", reducing forecast accuracy.

This project therefore has two broad objectives. First, we will develop a new multi-country forecasting model and efficient estimation algorithm which adopt a one-step approach. Second, using our new model, we will evaluate the relative importance of global, regional and domestic information in improving forecasts.

This will allow us to analyse cross-country linkages and investigate which countries are most vulnerable to crises and changes in different global conditions (e.g. supply chain pressure shocks, a slowdown in China's economy or changes in US monetary policy).