Thesis
Essays in macroeconomic interdependence, business cycles and nowcasting in a multi-country context
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2024
- Thesis identifier
- T16821
- Person Identifier (Local)
- 201984442
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- This thesis employs a multi-country approach and builds upon the existing literature on the Bayesian Panel Vector Autoregressions (PVARs) as its foundation for analysing empirical macroeconomic interdependence, business cycles synchronisation and economic forecasting. The contribution is provided in three essays. The first essay (Chapter 2) examines macroeconomic interdependency of main macroeconomic variables in terms of dynamic, static, and cross-sectional homogeneity features by using a PVAR model. In order to accurately measure these features, a stochastic search specification selection (S 4 ) prior algorithm is employed to investigate their interdependencies within the G-7 countries. The results indicate that while cross-sectional homogeneity is of little significance among the G-7, dynamic and static interdependencies are of great importance. In brief, the S 4 algorithm is beneficial for classifying each type of the panel structure of macro-financial interlinkages. This essay also compares the inflation forecasting performance of the S 4 algorithm with the original factor shrinkage prior of Canova and Ciccarelli (2009) and finds that the PVARs with the S 4 algorithm give a better point forecasting performance, particularly in the short-term forecast horizons. Regarding the density forecasts, the PVARs with the S 4 prior outperform the PVARs with the factor shrinkage prior for all the G-7 in the short-term horizons, whereas in the long-term horizons, although the PVARs with the factor shrinkage prior give an improved performance, they still only forecast better for two of the seven countries, namely Canada and Japan. The second essay (Chapter 3) investigates the economic interdependencies between the ASEAN+3 and the US as well as between the ASEAN+3 members themselves through the lens of business cycle synchronisation, by using a Bayesian panel Markov-switching VAR approach (The PMS-VAR model). The main reason for investigating this phenomenon is that the increasing level of regional economic integration of the ASEAN+3 has led to a discussion over the past decade about whether or not the ASEAN+3 is decoupling from the US economy. The results provide evidence that the business cycles of the ASEAN+3 economies are much more synchronised with each other than any of them are with the US economy, especially for real economic variables. However, for financial variables, the results indicate that after the US subprime crisis of 2008 the synchronisations of the ASEAN+3 and the US have become more substantial, particularly of their stock price indices and exchange rates. The third essay (Chapter 4) studies recent literature on nowcasting. Upon study, there is a substantial gap to be found regarding investigation into whether or not multi-country nowcasting models can give predictive gains, no doubt due to the historical issue of over-parameterisation, and this thesis meets the challenge of filling that gap. These models are helpful when considering the role of interdependence among a particular group of economies and have potential to help in the assessment of nowcasts of several different GDPs. Therefore, this chapter focuses mainly on comparing nowcasting performance between multi-country models - large Bayesian VARs, Panel VARs and a multi-country dynamic factor model, and individual-country models - MF-BVARs, MF-DFM, with mixed-frequency approaches, applied to the four largest European economies during both normal periods and the Covid-19 pandemic. The results show that country-specific models outperform the other models when it comes to nowcasts for almost all countries, especially the pandemic period.
- Advisor / supervisor
- Koop, Gary
- Darby, Julia
- Resource Type
- DOI
- Funder
关系
项目
| 缩略图 | 标题 | 上传日期 | 公开度 | 行动 |
|---|---|---|---|---|
|
|
PDF of thesis T16821 | 2024-03-07 | 公开 | 下载 |