How to estimate a vector autoregression after March 2020

Lenza, Michèle; Primiceri, Giorgio E.

doi:10.1002/jae.2895

Cited by 138 publications

(92 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…14 We also test for information deficiency in the VAR following Forni and Gambetti (2014) and find no significant concern. 15 Last, we replicated the results on a sample excluding the COVID-19 pandemic (Lenza and Primiceri, 2020), and we find the same qualitative results, with a somewhat weaker and less persistent response of industrial production.…”

Section: Asymmetric Responses In Stress and Non-stress Periodssupporting

confidence: 59%

“…The non-linear nature of the model partly takes into account the abrupt changes in macroeconomic and financial variables during the pandemic, so our results do not change when we end our sample at the end of 2019 before the start of the pandemic. Thus we are less subject to the critique of Lenza and Primiceri (2020) who argue that VAR models should not include the pandemic in the estimation period.…”

Section: Modelling the Nonlinearitiesmentioning

confidence: 99%

See 1 more Smart Citation

Systemic Financial Stress and Macroeconomic Amplifications in the United Kingdom*

et al. 2021

View full text Add to dashboard Cite

We develop a daily composite index of financial stress for the United Kingdom over 50 years, the UKFSI. The index includes market stress indicators based on their incremental information to capture financial crises. During the COVID-19 crisis, financial stress peaks but remains less severe than during the Global Financial Crisis. The UKFSI is used in a threshold vector autoregression to differentiate the economic dynamics between tranquil and stressful periods. We highlight the importance of nonlinearities that amplify shocks. But we find no evidence of financial shocks contributing to the COVID-19 crisis, possibly reflecting effective policy interventions. I. IntroductionOver the past five decades, the UK financial sector has been buffeted by a few major shocks, which have highlighted the importance of macro-financial linkages. On the one hand, the severe economic and financial meltdown during the global financial crisis (GFC) was amplified within the banking sector. On the other hand, the spread of the COVID-19 pandemic unleashed profound economic and financial stress, but the shock was partly absorbed by a more resilient banking sector and unprecedented policy support.Against this backdrop, we develop a daily systemic financial stress index for the United Kingdom, the UK Financial Stress Index (UKFSI). We demonstrate its importance for assessing the transmission of financial shocks and the amplification of small shocks during stressful periods.The UKFSI is a daily-from 1971 onwards-composite index of systemic financial stress, that is, it tracks periods characterized by higher volatilities, valuation losses and a JEL Classification numbers: C31, C54, G01, G15. *The authors thank Andy Blake, David Aikman, Lee Foulger and colleagues who participated in Bank of England seminars on the subject. The views expressed are those of the authors and do not necessarily reflect those of the Bank of England (its Monetary Policy Committee, Financial Policy Committee or Prudential Regulatory Committee) nor those of the Bank of Canada. All errors remain our own.

show abstract

Section: Asymmetric Responses In Stress and Non-stress Periodssupporting

confidence: 59%

Section: Modelling the Nonlinearitiesmentioning

confidence: 99%

Systemic Financial Stress and Macroeconomic Amplifications in the United Kingdom*

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In particular, the fluctuations in macroeconomic variables are large enough to materially change the estimates of forecasting models, such as our VAR in equation ( 5). Lenza and Primiceri (2021), Schorfheide and Song (2020), and Carriero et al (2021) provide further discussion. To handle the COVID-19 pandemic in our forecasting models, we change the VAR in equation ( 5) to be…”

Section: Results For 2020 and 2021mentioning

confidence: 98%

Advance Layoff Notices and Aggregate Job Loss

Krolikowski

Lunsford

2022

Working Paper (Federal Reserve Bank of Cleveland)

View full text Add to dashboard Cite

We collect data from Worker Adjustment and Retraining Notification (WARN) Act notices and establish their usefulness as an indicator of aggregate job loss. The number of workers affected by WARN notices ("WARN layoffs") leads state-level initial unemployment insurance claims, and changes in the unemployment rate and private employment. WARN layoffs move closely with aggregate layoffs from Mass Layoff Statistics and the Job Openings and Labor Turnover Survey, but are timelier and cover a longer sample. In a vector autoregression, changes in WARN layoffs lead unemployment rate changes and job separations. Finally, they improve pseudo real-time forecasts of the unemployment rate.

show abstract

“…In essence, the extreme data in the first half of 2020 caused estimates of the full sample MF-VAR coefficients to change in a manner which led to poor forecasts. Lenza and Primiceri (2020) propose an alternative VAR-based approach which allows the error covariance matrix to have a mixture distribution. In essence, the pandemic is treated as a large variance shock and pandemic observations are, thus, drastically downweighted in the model estimation.…”

Section: Introductionmentioning

confidence: 99%

“…However, disregarding these recent data is inappropriate for forecasting the future evolution of the economy, because it vastly underestimates uncertainty." Thus, although Schorfheide and Song (2020) and Lenza and Primiceri (2020) adopt very different approaches, they end up with similar advice: discard the pandemic observations when estimating the model.…”

Section: Introductionmentioning

confidence: 99%

Nowcasting in a Pandemic using Non-Parametric Mixed Frequency VARs

Huber¹,

Koop²,

Onorante³

et al. 2020

Preprint

View full text Add to dashboard Cite

This paper develops Bayesian econometric methods for posterior and predictive inference in a non-parametric mixed frequency VAR using additive regression trees. We argue that regression tree models are ideally suited for macroeconomic nowcasting in the face of the extreme observations produced by the pandemic due to their flexibility and ability to model outliers. In a nowcasting application involving four major countries in the European Union, we find substantial improvements in nowcasting performance relative to a linear mixed frequency VAR. A detailed examination of the predictive densities in the first six months of 2020 shows where these improvements are achieved.

show abstract

How to estimate a vector autoregression after March 2020

Cited by 138 publications

References 37 publications

Systemic Financial Stress and Macroeconomic Amplifications in the United Kingdom*

Systemic Financial Stress and Macroeconomic Amplifications in the United Kingdom*

Advance Layoff Notices and Aggregate Job Loss

Nowcasting in a Pandemic using Non-Parametric Mixed Frequency VARs

Contact Info

Product

Resources

About