gwp 2015
DOI: 10.24149/gwp256
|View full text |Cite
|
Sign up to set email alerts
|

To Bi, or not to Bi? Differences in Spillover Estimates from Bilateral and Multilateral Multi-country Models

Abstract: Asymptotic analysis and Monte Carlo simulations show that spillover estimates obtained from widely-used bilateral (such as two-country VAR) models are significantly less accurate than those obtained from multilateral (such as global VAR) models. In particular, the accuracy of spillover estimates obtained from bilateral models depends on two aspects of economies' integration with the rest of the world. First, accuracy worsens as direct bilateral transmission channels become less important, for example when the … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
6
0

Year Published

2016
2016
2021
2021

Publication Types

Select...
7
1

Relationship

2
6

Authors

Journals

citations
Cited by 11 publications
(6 citation statements)
references
References 29 publications
0
6
0
Order By: Relevance
“…In this regard, we use the FAVAR as a multicountry framework, assuming that developments of economic variables share strong similarities which can be approximated by common factors. Our empirical approach is also motivated and encouraged by the results of Georgiadis (). His results show that multicountry models such as the FAVAR (and GVAR, global vector autoregressive) are, in contrast to bilateral VARs, more appropriate to model global and regional shocks as well as cross‐country spillovers, because these models are capable of modeling higher‐order spillovers.…”
Section: Data and Empirical Approachmentioning
confidence: 99%
“…In this regard, we use the FAVAR as a multicountry framework, assuming that developments of economic variables share strong similarities which can be approximated by common factors. Our empirical approach is also motivated and encouraged by the results of Georgiadis (). His results show that multicountry models such as the FAVAR (and GVAR, global vector autoregressive) are, in contrast to bilateral VARs, more appropriate to model global and regional shocks as well as cross‐country spillovers, because these models are capable of modeling higher‐order spillovers.…”
Section: Data and Empirical Approachmentioning
confidence: 99%
“…Although we find that our results are robust to different specifications and consistent with existing studies, identification of spillover effects from the bilateral perspective may result in possible practical consequences. As argued in Georgiadis (2015) spillover estimates obtained from bilateral models are in general inconsistent asymptotically and less accurate than those obtained from a multilateral model in finite samples due to omitted variable bias and failure to account for higher-order transmission channels. In this respect, our results from the bilateral perspective may not indicate all possible cross-country interrelations in the globalized world.…”
Section: Figurementioning
confidence: 98%
“…15 Brute force alternatives for carrying out counterfactual analysis in VAR models are to set to zero autoregressive parameters after or before estimation (see, for example, Ramey, 1993;Vicondoa, 2019;Degasperi et al, 2020;Dees and Galesi, 2021). However, setting to zero VAR coefficients before estimation implies a mis-specified empirical model and induces biased estimates; in general, the bias is not informative about the strength of the channel that is being shut down (Georgiadis, 2017). In turn, setting to zero VAR coefficients after estimation may be understood similarly as the MRE approach in the sense that it reflects some counterfactual VAR model.…”
Section: Conceptual Considerationsmentioning
confidence: 99%