Why do households repay their debt during COVID 19 crisis? well-being and financial implications.

Lead Research Organisation: Birkbeck, University of London
Department Name: Management

Abstract

Bank of England (Money and Credit, June 2020) reports that households repaid £7.4 bn of consumer credit in April 2020, the largest net repayment on record. The lockdowns due to COVID 19 could have initially shifted household behaviour towards debt repayments as households opt for prudent management of their finances, while recent data also show an increase in households savings. Alas, this might not persist as recent reports shows that British households could increase their debt exposure by £6bn due to the pandemic as they fall behind on credit card payments, council tax and utility bills (Guardian, 9 Jun 20). This could be also the case in other countries, i.e. US (FT, 23 Jun 20).

The project explores whether higher household debt repayment would persist or it could be of transitory nature also in light of the severe adverse economic conditions due to the pandemic. Any findings on household debt repayments patterns would impact upon total indebtedness of the economy and upon financial stability.
So it is of importance to understand what happens with household debt repayments in real time.

In some detail, the project examines the impact of the pandemic and government interventions on household debt repayments and on household financial resilience, measured as months during which households can pay for subsistence consumption and debt with liquid assets in case of income loss. Lockdowns by reducing household spending and employment stimulus packages (i.e. furloughing) have affected debt repayments. Monetary expansion and debt repayment moratoriums have reduced household debt burdens. In addition, household specific characteristics, such as ethnicity, age, gender, assets, health, employment could interact with the above.

The project uses a plethora of data to analyse household debt repayments and financial resilience also in light that supportive government interventions are going to be faced out while also lockdowns vary. The project employs an innovative methodology of a unique panel Vector Autoregressive model that nests neural networks. This methodology allows to study the responses in UK household debt repayments, financial resilience and well-being to shocks in COVID 19 and government interventions while examines the interconnectedness between household financial resilience, household well-being and financial stability. In addition, the project provides evidence, for comparison and to capture variability in government interventions and data, across countries: USA, Canada, Japan, France and Germany.

The project's empirical findings would inform policy making decisions in real time while its forecasts have medium term horizon in line with the pandemic trends.
 
Description If, indeed, increases in household debt repayment of the first six months of the Covid-19
lockdowns were to last, it could have caused a structural change in the financial industry.
However, our results show that there is no strong persistence in household debt repayments
and other household financial data.
These results are supported by the recent data showing that households' debt repayments have
fallen behind in 2021. The prolonged uncertainty over the pandemic and the associated
restrictions has adversely affected household finances and in particular household debt
repayments. However, as the lockdowns are lifted it appears that the Covid-19 shocks are
diminishing and household financial data converge to the levels prior to the pandemic albeit
with some lags.
A common pattern emerges from the above graphical analysis. Households' debt repayment
and household credit increased and decreased respectively in the first six months of the
pandemic and during the first lockdown, but ever since there are underlying dynamics that
dominate while there is little persistence in either debt repayments or credit.

In another strand of our research, we show that non-pharmaceutical interventions impact on the initial exponential growth of the infected population and the final exponential decay of the infected population. We employ a Bayesian dynamic model to test whether there is a Bayesian learning, a random walk pattern or other type of learning with evolving epidemiological data over time. Using a sample of UK country-specific data and also for 168 countries and 51,083 country-date observations, we estimate the model with time-varying parameters in a dynamic panel vector autoregressive model. Although learning does not seem to be taking place, and despite the absence of evidence of governments' learning from the past, most policy measures appear to assert some effect on the parameters of the number of susceptible people, the number of infected, and the number of recovered persons. We also provide estimates of time-varying parameters that can be used widely, and we develop novel testing procedures for testing for Bayesian learning.

We continue our research to study household debt and household wellbeing.
Exploitation Route All the above evidence of our research on Covid is in the public domain. The research is ongoing and by early in May we would have concluding remarks and what can be taken forward and put to use by others.
Sectors Communities and Social Services/Policy,Financial Services, and Management Consultancy,Healthcare,Leisure Activities, including Sports, Recreation and Tourism,Pharmaceuticals and Medical Biotechnology
 
Description Our empirical methodology and evidence are attracting the attention of policymakers as our research is openly available and is downloaded by government officials. We have also been contacted by policymakers in the Bank of England. Over the coming months as we finalise our findings we expect that we will attract more attention. However, at this stage and given our research is ongoing it is early to measure impact.
First Year Of Impact 2022
Sector Financial Services, and Management Consultancy,Leisure Activities, including Sports, Recreation and Tourism,Pharmaceuticals and Medical Biotechnology
Impact Types Economic,Policy & public services
 
Description School Impact Grant: The response of household debt to Covid-19
Amount £2,433 (GBP)
Organisation Birkbeck, University of London 
Sector Academic/University
Country United Kingdom
Start 09/2022 
End 05/2023
 
Title Coronavirus (COVID-19) latest insights A live roundup of the latest data and trends about the coronavirus (COVID-19) pandemic from the ONS and other sources. 
Description A live roundup of the latest data and trends about the coronavirus (COVID-19) pandemic from the ONS and other sources. 
Type Of Material Database/Collection of data 
Year Produced 2022 
Provided To Others? Yes  
Impact The data set provides up-to-date information about Coronavirus (COVID-19) latest insights. It provides a live roundup of the latest data and trends about the coronavirus (COVID-19) pandemic from the ONS and other sources. And provide all to date recent data on Covid-19 on various topics such as: Comparisons Infections Hospitals Deaths Vaccines Antibodies Well-being Lifestyle 
URL https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/articl...
 
Title Money and Credit, Bank of Enlgand 
Description These monthly statistics on the amount of, and interest rates on, borrowing and deposits by households and businesses are used by the Bank's policy committees to understand economic trends and developments in the banking system. 
Type Of Material Data handling & control 
Year Produced 2020 
Provided To Others? Yes  
Impact Based on Bank of England report (Money and Credit, June 2020) households repaid £7.4 billion of consumer credit, on net, in April 2020 the largest net repayment on record. 
 
Title OECD Quarterly National Accounts, 1947-2021 
Description The OECD's quarterly national accounts (QNA) dataset presents data collected from all the OECD member countries and some other major economies on the basis of a standardised questionnaire. It contains a wide selection of generally seasonally adjusted quarterly series most widely used for economic analysis. It contains a wide selection of generally seasonally adjusted quarterly series most widely used for economic analysis from 1947 or whenever available:- GDP expenditure and output approaches (current prices and volume estimates);- GDP income approach (current prices);- Gross fixed capital formation (current prices and volume estimates) broken down separately by type of asset or product and by institutional sector;- Disposable income and Real disposable income components;- Saving and net lending (current prices);- Population and Employment (in persons);- Employment by industry (in persons and hours worked);- Compensation of employees (current prices);- Household final consumption expenditure by durability (current prices and volume estimates).Please note that OECD reference year changed from 2010 to 2015 on Tuesday 3rd of December, 2019. These data were first provided by the UK Data Service in October 2006. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
Impact OECD country specific data on household finances like household debt, household disposal income 
URL https://beta.ukdataservice.ac.uk/datacatalogue/doi/?id=5540#3
 
Title The Oxford COVID-19 Government Response Tracker (OxCGRT) 
Description The Oxford COVID-19 Government Response Tracker (OxCGRT) systematically collects information on several different common policy responses that governments have taken to respond to the pandemic on 18 indicators such as school closures and travel restrictions. It now has data from more than 180 countries. The data is also used to inform a Risk of Openness Index which aims to help countries understand if it is safe to 'open up' or whether they should 'close down' in their fight to tackle the coronavirus. 
Type Of Material Data handling & control 
Year Produced 2020 
Provided To Others? Yes  
Impact Since the outbreak of the COVID-19 pandemic in the United Kingdom (UK) in early 2020, the four nations of England, Scotland, Wales and Northern Ireland have responded with a wide range of measures to break the chain of infection and manage the broader impacts of the disease. Due to the devolved powers afforded to the governments of Scotland, Wales and Northern Ireland, all four nations of the UK have used their autonomy to implement and ease restrictions. While economic support and public health measures have been similar across the four UK nations, the different governments have implemented different closure and containment policies since May 2020. The Oxford COVID-19 Government Response Tracker provides a systematic way to measure and compare government responses to COVID-19 across the four nations from 1 January 2020 to the present, and will be updated continuously going forward. The tracker combines individual indicators into a series of novel indices that aggregate various measures of government responses. These can be used to describe variation in government responses, explore whether the government response affects the rate of infection, and identify correlates of more or less intense responses. KEY POINTS All four nations of the UK implemented similar lockdown-style restrictions in mid March to combat the increasing cases and spread of COVID-19, but diverged in their approach to easing restrictions. Scotland has been the most consistent with their restrictions and stringency throughout the COVID-19 pandemic in the UK thus far. England eased restrictions earlier than the other three nations and maintained the lowest restrictions until early October. All four nations of the UK saw increased COVID-19 cases and deaths following the return of students to schools and universities in September. With the increase in cases across the UK in September and October, all four nations have imposed stricter measures to reduce and contain the spread of COVID-19. Where the four nations have varied most in their policies regarding COVID-19 restrictions is regarding domestic movement: Scotland was the last to fully drop a 5 mile domestic travel restriction on 7 July compared to 13 May in England. 
 
Title UK Data Service: Understanding Society: COVID-19 Study, 2020-2021 
Description The COVID-19 survey From April 2020 participants from our main Understanding Society sample were asked to complete a short web-survey. This survey covered the changing impact of the pandemic on the welfare of UK individuals, families and wider communities. Participants completed a regular survey, which included core content designed to track changes, alongside rotating content adapted each wave as the coronavirus situation developed. There was a telephone version of the survey in some waves for participants unable to take part online. The COVID-19 study was funded by the Economic and Social Research Council and the Health Foundation. Fieldwork for the web surveys was carried out by Ipsos MORI; fieldwork for the phone surveys by Kantar. Both agencies have conducted the COVID-19 youth surveys. If you would like to receive updates from Understanding Society on the COVID-19 survey please join our mailing list. If you are a researcher or analyst and have a question about the COVID-19 study you can contact the COVID survey team by email: covid@understandingsociety.ac.uk. If you are a participant of the study and have a question please contact the participant liaison team. COVID-19 datasets Data from the COVID-19 survey is available from the UK Data Service. If you have a question about using the COVID-19 dataset please contact our user support team. You can find the Understanding Society COVID-19 dataset, Study Number 8644 here. Special Licence data Special Licence datasets are available: SN 8663 - Understanding Society: COVID-19 Study, 2020: Special Licence Access, Census 2011 Lower Layer Super Output Areas SN 8664 - Understanding Society: COVID-19 Study, 2020: Special Licence Access, Local Authority District SN 8730 - Understanding Society: COVID-19 Study, 2020: Special Licence Access, School codes Questionnaire content The COVID-19 survey has the core modules: Household composition Coronavirus illness Long-term health conditions management GHQ Loneliness Employment 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
Impact This data provides survey data for individuals and households in UK 
 
Title World Economic Outlook (WEO) (April 2021 Edition) 
Description The World Economic Outlook (WEO) database contains selected macroeconomic data series from the statistical appendix of the World Economic Outlook report, which presents the IMF staff's analysis and projections of economic developments at the global level, in major country groups and in many individual countries. 
Type Of Material Database/Collection of data 
Year Produced 2021 
Provided To Others? Yes  
Impact The data set provides country specific data for a global sample so as to investigate the main research question of the project using a global panel VAR. 
URL https://beta.ukdataservice.ac.uk/datacatalogue/doi/?id=5761
 
Description Professor Mike Tsionas 
Organisation Lancaster University
Country United Kingdom 
Sector Academic/University 
PI Contribution Mike G. Tsionas is Professor of Economics at the Lancaster University Management School. He is a Fellow of the Journal of Econometrics, a Distinguished Author of the Journal of Applied Econometrics, and an Associate Editor of Empirical Economics, Journal of Productivity Analysis, Economic Modelling, Journal of Mathematics, and the Journal of Banking and Finance in the past. He has authored 160 articles and is an expert of Bayesian econometric theory.
Collaborator Contribution Mike provides the programming tools to compute the empirical evidence of the research and develop new novel estimation methods.
Impact Why Do Households Repay Their Debt during COVID-19 Crisis? A Panel Var Using Neural Networks 41 Pages Posted: 23 Feb 2022 E. C. Mamatzakis Birkbeck College, University of London Mike Tsionas Lancaster University Steven Ongena University of Zurich Date Written: February 21, 2022 Abstract Based on a report from the Bank of England (Money and Credit in April 2020) households have been repaying loans from banks while consumer credit has been dramatically fallen. Households repaid £7.4 billion of consumer credit, on the net, in April 2020, the largest net repayment since the series began. Higher payments towards household debt would enhance both household financial sustainability and financial resilience. Household debt repayments would also have implications for the financial industry and financial stability. In this paper, we investigate whether higher household debt repayments have been caused by COVID-19 and the subsequent lockdowns in the UK and whether they will persist over time. We also perform a forecasting exercise. Keywords: COVID-19, household debt, ANN, panel VAR, MIDAS Suggested Citation: Mamatzakis, E. C. and Tsionas, Efthymios G. and Ongena, Steven, Why Do Households Repay Their Debt during COVID-19 Crisis? A Panel Var Using Neural Networks (February 21, 2022). Available at SSRN: https://ssrn.com/abstract=4039952 or http://dx.doi.org/10.2139/ssrn.4039952 Bayesian Policy Learning Modeling of COVID-19 Interventions: the Impact on Household Debt Repayment in UK and Internationally 37 Pages Posted: 27 Jul 2021 Last revised: 22 Feb 2022 E. C. Mamatzakis Birkbeck College, University of London Steven Ongena University of Zurich Mike Tsionas Lancaster University Date Written: July 17, 2021 Abstract The rapid spread of COVID-19 across the globe primed a variety of non-pharmaceutical interventions. Given these interventions, we examine the impact of those interventions on the initial exponential growth of the infected population and the final exponential decay of the infected population. We employ a Bayesian dynamic model to test whether there is a Bayesian learning, a random walk pattern or other type of learning with evolving epidemiological data over time. Using a sample of UK country specific data and also for 168 countries and 51,083 country-date observations, we estimate the model with time-varying parameters in a dynamic panel vector autoregressive model. Although learning does not seem to be taking place, and despite the absence of evidence of governments' learning from the past, most policy measures appear to assert some effect on the parameters of the number of susceptible people, the number of infected, and the number of recovered persons. We also provide estimates of time-varying parameters that can be used widely, and we develop novel testing procedures for testing for Bayesian learning. Keywords: Bayesian learning, COVID-19, UK household debt repayment, interventions JEL Classification: G20, IOO, CO1, C11 Suggested Citation: Mamatzakis, E. C. and Ongena, Steven and Tsionas, Efthymios G., Bayesian Policy Learning Modeling of COVID-19 Interventions: the Impact on Household Debt Repayment in UK and Internationally (July 17, 2021). Available at SSRN: https://ssrn.com/abstract=3888559 or http://dx.doi.org/10.2139/ssrn.3888559
Start Year 2020
 
Description Professor Steven Ongena, University of Zurich Main navigation, Department of Banking and Finance 
Organisation University of Zurich
Country Switzerland 
Sector Academic/University 
PI Contribution Steven Ongena is a professor of banking in the Department of Banking and Finance at the University of Zurich, a senior chair at the Swiss Finance Institute, a research professor at KU Leuven, a research professor at the Norwegian University of Science and Technology NTNU Business School, and a research fellow in financial economics of CEPR. He is also a research professor at the Deutsche Bundesbank and a regular research visitor at the European Central Bank. Before moving to Zurich, he taught at CentER-Tilburg University and BI Norwegian Business School and studied at the Universities of Oregon (PhD), Alberta (MA) and KU Leuven (MBA, Hir). He is publishing in economics, finance, law and management journals, including in the American Economic Journal: Macroeconomics, American Economic Review, Econometrica, Journal of Finance, Journal of Financial Economics, Journal of Political Economy, Management Science, Review of Finance, and Review of Financial Studies, among other journals. He co-authored, with Hans Degryse and Moshe Kim, the graduate textbook Microeconometrics of Banking: Methods, Applications and Results published by Oxford University Press. He is currently a co-editor of Economic Inquiry, the International Journal of Central Banking, the International Review of Finance and the Journal of Financial Services Research, and an associate editor of the Journal of Financial Stability, Economic Notes, the Asian Review of Financial Research, and the Journal of Financial Management, Markets and Institutions. In the past he has served as a co-editor for the Review of Finance and as an associate editor for the Journal of Finance, the Journal of Financial Intermediation, the Journal of Financial Services Research, the European Economic Review and the Journal of Banking and Finance, among other journals. He is a fellow of CFS and serves on the scientific advisory board of EBES, FINEST, GOLCER, the Halle Institute for Economic Research, IBEFA, the Research Data and Service Centre of the Deutsche Bundesbank and SAFE. In 2017 he received an ERC Advanced Grant lending, in 2012 an NYU-Fordham-RPI Rising Star in Finance Award and in 2009 a Wim Duisenberg Research Fellowship of the European Central Bank.
Collaborator Contribution The outcomes of our research of the implications of household debt repayments on household income and wellbeing will inform MaPS. In addition, we have been in conduct with the Bank of England as Steven Ongena has been recently visiting the Bank and has ongoing research engagement while Emmanuel Mamatzakis is in conduct with Jonathan Haskel (Bank of England). Policymakers will be interested to hear about the findings of our research in relation to household financial behaviour during the crisis, household financial resilience, and financial stability. In addition, there is a growing public interest in household debt repayments as evidenced by the extensive news coverage on household finances during the pandemic and the popularity of TV programs linked to household debt. The project in a systematic manner would provide research feedback to economic policymakers and briefs to newspapers and TV news programmes. All members of the research team have substantial experience and track record of achieving a strong impact on economic policy and in particular on issues related to finance. In addition, newspapers and TV programs have frequently featured members of the research team. It is also worth noting that the research team has been involved in policymaking for over three decades.
Impact Mamatzakis, E. C. and Ongena, Steven and Tsionas, Efthymios G., Bayesian Policy Learning Modeling of COVID-19 Interventions: the Impact on Household Debt Repayment in UK and Internationally (July 17, 2021). Available at SSRN: https://ssrn.com/abstract=3888559 or http://dx.doi.org/10.2139/ssrn.3888559 Abstract The rapid spread of COVID-19 across the globe primed a variety of non-pharmaceutical interventions. Given these interventions, we examine the impact of those interventions on the initial exponential growth of the infected population and the final exponential decay of the infected population. We employ a Bayesian dynamic model to test whether there is a Bayesian learning, a random walk pattern or other type of learning with evolving epidemiological data over time. Using a sample of UK country specific data and also for 168 countries and 51,083 country-date observations, we estimate the model with time-varying parameters in a dynamic panel vector autoregressive model. Although learning does not seem to be taking place, and despite the absence of evidence of governments' learning from the past, most policy measures appear to assert some effect on the parameters of the number of susceptible people, the number of infected, and the number of recovered persons. We also provide estimates of time-varying parameters that can be used widely, and we develop novel testing procedures for testing for Bayesian learning. Keywords: Bayesian learning, COVID-19, UK household debt repayment, interventions JEL Classification: G20, IOO, CO1, C11 Mamatzakis, E. C. and Tsionas, Efthymios G. and Ongena, Steven, Why Do Households Repay Their Debt during COVID-19 Crisis? A Panel Var Using Neural Networks (February 21, 2022). Available at SSRN: https://ssrn.com/abstract=4039952 or http://dx.doi.org/10.2139/ssrn.4039952 Abstract Based on a report from the Bank of England (Money and Credit in April 2020) households have been repaying loans from banks while consumer credit has been dramatically fallen. Households repaid £7.4 billion of consumer credit, on the net, in April 2020, the largest net repayment since the series began. Higher payments towards household debt would enhance both household financial sustainability and financial resilience. Household debt repayments would also have implications for the financial industry and financial stability. In this paper, we investigate whether higher household debt repayments have been caused by COVID-19 and the subsequent lockdowns in the UK and whether they will persist over time. We also perform a forecasting exercise. Keywords: COVID-19, household debt, ANN, panel VAR, MIDAS
Start Year 2020
 
Title A Global panel VAR of ANN 
Description We fit all information/ data of various frequencies into a vector autoregression (VAR) with µ a vector of constant terms, B a matrix containing unknown coefficients, and S the unknown covariance matrix. Within this VAR household debt repayments, household financial resilience and financial stability are prominent to examine their underlying responses to shocks. Within the VAR our focus is on households with their needs for assets and liquidity as well as their resilience and recovery in the aftermath of an extreme event/shock such as Covid19. To this end, we relate economic and financial-oriented information to address how extreme events/shocks are reshaping the paradigm of household debt repayments and household financial resilience while controlling for household characteristics. Note that we have variables with different frequencies (Covid 19 data are daily, household debt repayments monthly) necessitating an application of Mixed Data Sampling (MIDAS) or alternatives. From the VAR we estimate how exposure coefficients d vary with COVID19 infections to derive their impact on household repayments and household financial resilience. This model treats COVID19 as a forcing variable; not as an exogenous shock since VARs are more appropriate for normal times but not when an extreme persistent shock takes place as in the case with COVID19. To estimate the model's parameters, we employ neural networks. It is well known that neural networks can approximate well any functional form to arbitrary accuracy. Household specific characteristics like ethnicity, well-being would also form part of the neural network equations so that generalized impulse response in the functions can be computed separately. Our model allows deriving forecasts to inform policymaking under different scenarios of the main variables, i.e., COVID19 infections. The steps for our paper are as follows: first, we perform historical simulations to examine model fit; second, given different scenarios for COVID19 as well as different government interventions we estimate the impact on the household debt repayments and household financial resilience in the VAR using generalized impulse response functions; the focus is on how lockdown, lifting lockdown and government financial assistance would affect household debt repayment over the remaining 2020, into 2021 and 2022; we, therefore, provide simulations for future paths of household debt payments and household financial resilience based on different scenarios that would also control for new health developments such as test and trace applications, drag and vaccine discovery. Lastly, we employ generalized response functions and provide forecasts of the main variables of our modeling (household debt repayments, household financial resilience and financial stability) as well as their interactions with government interventions and feedback loops. This is of particular importance given the risk of further waves of the pandemic. Our results are useful for policymakers as they provide evidence of how government intervention household behaviour is key to overcoming the crisis. 
Type Of Technology New/Improved Technique/Technology 
Year Produced 2021 
Impact We fit all information/ data of various frequencies into a vector autoregression (VAR) with µ a vector of constant terms, B a matrix containing unknown coefficients, and S the unknown covariance matrix. Within this VAR household debt repayments, household financial resilience and financial stability are prominent to examine their underlying responses to shocks. Within the VAR our focus is on households with their needs for assets and liquidity as well as their resilience and recovery in the aftermath of an extreme event/shock such as Covid19. 
URL http://dx.doi.org/10.2139/ssrn.4039952
 
Title The Response of Household Debt to COVID-19 Using a Neural Networks VAR in OECD 
Description This technique investigates responses of household debt to COVID-19-related data like confirmed cases and confirmed deaths within a panel VAR framework for OECD countries. We also employ a plethora of non-pharmaceutical and pharmaceutical interventions as shocks. In terms of methodology, we opt for a global panel VAR (GVAR) methodology that nests underlying country VARs. In addition, as linear factor models may be unable to capture the variability in the data, we use an artificial neural network (ANN) method. The number of factors, as well as the number of intermediate layers, are determined using the marginal likelihood criterion and we estimate the GVAR with MCMC techniques. Results reveal that household debt positively responds to COVID-19 infections and mortality as well as lockdowns, though this response is valid in the short term. However, vaccinations and testing appear to negatively affect household debt. Lockdown measures such as stay-at-home advice, and closing schools, all have a positive impact on household debt in GVAR, though of transitory nature. 
Type Of Technology New/Improved Technique/Technology 
Year Produced 2022 
Impact In GVARs it is typical that various VARs are connected through some observed variables like an export / import index converted to lie between zero and one. Here, we connect the different VARs through several common dynamic factors (an ×1 vector, ). is the unknown number of components in the ANN. The number of factors as well as the number of intermediate layers () are determined using the marginal likelihood criterion (Diccio et al., 1997). We use the MCMC technique in Appendix A.1. We use 150,000 iterations, the first of which are discarded to mitigate possible start-up effects. For each country, the VAR can be estimated on an equation-by-equation basis resulting in a substantial reduction in computation time. The dynamic factors in (6) are computed beforehand to simplify computations so, for all equations of the GVAR as well as different countries, equations can be estimated on an equation-by-equation basis. We use a flat prior for the coefficients in (7). For the diagonal elements of () we assume that they are normally distributed with mean 1 and standard deviation 0.2. The non-diagonal elements have a normal prior with mean zero and standard deviation 0.2. For the elements of ?() and F() we assume that they have a standard normal distribution. 
URL http://dx.doi.org/10.2139/ssrn.4087551
 
Title The rapid spread of COVID-19 across the globe primed a variety of non-pharmaceutical interventions (NPIs). Given these NPIs, whether the SIR parameters followed a Bayesian learning, a random walk pattern or other type of learning. 
Description It is a Bayesian method of estimation. We use a recent advance in sequential Monte Carlo methods known as the particle Gibbs (PG) sampler. see technical appendix at: https://eprints.bbk.ac.uk/id/eprint/45203/1/paper_COVID%2015July21.pdf 
Type Of Technology New/Improved Technique/Technology 
Year Produced 2021 
Open Source License? Yes  
Impact We use a recent advance in sequential Monte Carlo methods known as the particle Gibbs (PG) sampler, see Andrieu et al. (2010). The algorithm allows us to draw paths of the state variables in large blocks. Particle filtering is a simulation-based algorithm that sequentially approximates continuous, marginal distributions using discrete distributions. This is performed by using a set of support points called ''particles'' and probability masses; see (D. Creal, 2012) for a review. The PG sampler draws a single path of the latent or state variables from this discrete approximation. As the number of particles M goes to infinity, the PG sampler draws from the exact full conditional distribution. As mentioned in (Creal and Tsay, 2015): "The PG sampler is a standard Gibbs sampler but defined on an extended probability space that includes all the random variables that are generated by a particle filter. Implementation of the PG sampler is different than a standard particle filter due to the ''conditional'' resampling algorithm used in the last step. Specifically, for draws from the particle filter to be a valid Markov transition kernel on the extended probability space, Andrieu et al. (2010) note that there must be a positive probability of sampling the existing path of the state variables that were drawn at the previous iteration. The pre-existing path must survive the resampling steps of the particle filter. The conditional resampling step within the algorithm forces this path to be resampled at least once. We use the conditional multinomial resampling algorithm from Andrieu et al. (2010), although other resampling algorithms exist, see Chopin and Singh (2015)" (page 339). 
URL https://eprints.bbk.ac.uk/id/eprint/45203/1/paper_COVID%2015July21.pdf
 
Description On line Workshop. 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact On-line day workshop in the Department of Management in July 2021, Birkbeck College on the impact of Covid-19 on household finances. During the workshop, the principal investigator presented empirical evidence of whether we observe Bayesian learning during Covid-19 using a variety of non-pharmaceutical
interventions. There was constructive dialogue and various feedback was given to the research team
As a result of the workshop, an eprint was generated at: https://eprints.bbk.ac.uk/id/eprint/45203/
Year(s) Of Engagement Activity 2021
URL https://eprints.bbk.ac.uk/id/eprint/45203/1/paper_COVID%2015July21.pdf
 
Description One Day Workshop on Covid-19 and household finance, Birkbeck College, Clore Building, 31st May 2023 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact The workshop will provide a platform to disseminate the findings of my research entitled 'Why do households repay their debt during COVID-19 crisis? Well-being and financial implications, the study examines the impacts of COVID-19 policy interventions on household debt repayments and financial resilience.
The study finds a plethora of results that have been already published in academic journals and provides also evidence of the persistence of both household debt repayments and debt exposure. Recent statistics show that British households have increased their debt exposure by £6 billion as they fall behind on credit card payments, council tax, and utility bills. The study delivers evidence and projections that will inform policy-making decisions in real-time and enhance impact.
Year(s) Of Engagement Activity 2023