Probabilistic approach to assessing macroeconomic uncertainties

Lead Research Organisation: University of Leicester
Department Name: Economics

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description Within the project, we have analysed the measurements and components of macroeconomic uncertainty. We have found that it is possible to discover and quantify footprints which economic policy might be leaving on the uncertainty. Our main findings are: (1): We have developed a new statistical distribution, named by us the weighted skew normal distribution which parameters can tell whether the economic policy is effective in reducing macroeconomic uncertainty. Using relatively simple data (collection of forecast errors), we have used this distribution for removing policy effects from the term structure of inflation forecast and apply it for 38 countries. (2): We have proposed a test which can be used for detecting such footprints of economic policy in time series of economic data. We have applied this test to the series of 10-year government bond yields for 37 countries, positively verifying the hypothesis of policy influence on the uncertainty for more than 50% of them. (3): We have developed a concept of bivariate and conditional uncertainties, which is useful for analysing integrated markets and economies. We have shown that applying measures of conditional uncertainty can lead to reinterpretation of findings of some policy-related inflation probabilistic forecasts, in particular of computing the probabilities of inflation being within some pre-defined bands around inflation target. We have illustrated this by the analysis of inflation uncertainty in Canada as being conditional on that in the U.S. (4): We have identified conditions where overzealous economic policy, resulting in frequent interventions, can be counterproductive and lead to an increase rather than decrease in forecast uncertainty. Consequently, we were able to explain the different patterns in inflation forecasts made by the central banks' forecasters and independent ones.
We also have some negative findings. We have found out that some statistical distributions previously regarded as promising for modelling uncertainty, notably the tempered stable distribution, are not quite fit for this purpose, as their estimation is often awkward and the parameters do not have clear economic interpretation. This negative finding prompted us to abandon the path of investigating tempered stable distributions and lead to formulation of the skew normal distribution and, consequently, findings described above.
We are recently collaborating directly with central banks of Estonia and Poland regarding practical implementations of our results for improvement of current forecasting practices. We are also planning to approach the European Central Bank, Bank of England and Bank of France with propositions of collaboration in the incorporation of our methods into their forecasting practices.
Main research topics which we intend to investigate further in the future are: (1) Further inquiry into the development of concept applications of multivariate uncertainty and (2) Generalisation of our findings for the case of more challenging economic data, such as pooling data of mixed frequencies, and amalgamating time series and panel (experts) forecasts. We continue leading the network of researchers working on macroeconomic uncertianty by organizing special sessions at conferences and co-organizing biannual workshops, in collaboration with the American University, Washington D.C..
Exploitation Route Our findings can be of an immediate use by various macroeconomic bodies, most notably such which base their decisions on outcomes of the forecasting process. Our findings help in overcoming one of the most notorious problems in macroeconomic forecasting that the decisions based on such forecasts might change the realisations of the forecasted phenomenon itself. We suggest the way of forecasting as if the policy decision was not made. Hence, it can be of a particular interest to central banks and fiscal policy bodies. We are currently in collaboration with two European central banks: Bank of Estonia and National Bank of Poland regarding the implementation of our methods. We are also approaching other central banks. We are making inquiry regarding implementation (Bank of France, European Central Bank), inviting researchers and collaborators from these banks to our workshops and special conference sessions (Bank of Canada, Bank of Norway) and collaborating with advisors and experts (Bank of England). We have also presented a research at the Central Bank of Kazakhstan. More substantive talks will take place if some of our papers are accepted to a journal with a significant academic reputation.
Sectors Creative Economy
URL http://pramu.ac.uk
 
Description The ultimate long-term aims of the project were (1) to improve the forward-looking inquiry into macroeconomics by identification possible causes of uncertainty and (2) improving the quality of forecasting. The main directions of possible implementation of the results are to (1) use the probabilistic forecasting methods developed within the project in macroeconomic forecasting practice and (2), apply other methods developed, of testing whether a macroeconomic policy is successful in reducing macroeconomic uncertainty. Hence, the primary recipients of our results are bodies conducting economic policy, monetary and fiscal, which are central banks and other institutions undertaking decisions on the base of macroeconomic projections, and also macroeconomic consultancy firms. We are actively promoting implementations of our methods to such bodies. In particular, Svetlana Makarova's spent three months in 2015 in the Central Bank of Estonia working on such implementation. Her visit resulted in a series of research seminars givren at the Bank and publication of working paper, scheduled for 2016, in which methods developed within the project are applied for the evaluation of monetary policy in the Euro area. There is also an undertaking aiming at the implementation of the results at the National Bank of Poland (with two presentations in 2014 and 2015 by Svetlana Makarova), and the Bank of England, with a presentation by Seohyun Lee in September 2015. These are the most advanced results or our efforts to implement findings of the project in practice. We are also in discussion with researchers from the European Central Bank and Bank of Canada, albeit the progress here is slower. Interest of the practitioners in the results of the project can be illustrated by the dynamics of their intended participation in project-related workshop on modelling uncertainty, first organized at the University College London in May 2014, the second in May 2016 (the latter event is co-financed by the Bank of France), ad third will be organized in 2018 at the Renmin University in Beijing. While in 2014 we had 8 submissions from banks and policy-oriented organisations, in 2016 number of such submissions have risen to 37 (among others, from the Federal Reserve Board, Bank of Canada, Bank of England, Bank of Finland, Bank of France, Bank of Italy, European Central Bank, International Monetary Fund, National Bank of Poland, National Bank of Slovakia and some private consultancy firms). We are currently in discussion with Central Banks of Kazakhstan and Poland regarding the adaptation of the methods developed. In September 2016 we have conducted a seminar at the National Bank of Kazakhstan in Almaty explaining our methods. Also in September 2016 we have presented our results at a workshop at CITE, Stanford University and involved in preliminary discussions with the representatives of US Federal Reserve Board regarding the implementation. Further plans aimed at preparation and dissemination of software which can be used by the end users for conducting probabilistic forecasting and evaluating policy effects.
First Year Of Impact 2014
Sector Creative Economy
Impact Types Economic
 
Description CDAM, Minsk 
Organisation Belarusian State University
Country Belarus 
Sector Academic/University 
PI Contribution Wojciech Charemza was acting as the invited speaker and member of the Organizing Committee of the 10th International Conference 'Computed Data Analysis and Modelling (CDAM) , in Minsk, Belarus, 2013. There is a further collaboration ongoing related to the organization of the next CDAM conference in 2016.
Collaborator Contribution The conference creates an important forum for meeting with Russian and Belarussian applied mathematicians and statisticians, who often publish results relevant for the project locally and in Russian.
Impact Invited paper presented at the conference CDAM 2013: Charemza, W. C. Díaz and S. Makarova: 'Too many skew normal distributions: the practitioner's perspective',
Start Year 2006
 
Description CFE 
Organisation Computational and Financial Econometrics (CFEnetwork)
Country Global 
Sector Charity/Non Profit 
PI Contribution Within the project 7 organized (invited) sessions have been set up at 5 consecutive CFE annual conferences: 2011, London; 2012, Oviedo; 2013, London (two sessions); 2014, Pisa; 2015, London (forthcoming)
Collaborator Contribution The organized (invited) sessions enabled inviting specific scholars working on similar or complementary topics. This in turn provided forum for substantive discussion and exchange of ideas, leading to a progress within the project.
Impact CFE 2010, London. Session: Heavy-tailed time series S. Makarova, W. Charemza, P. Jelonek, Macroeconomic applications of skewed tempered stable distributions CFE 2011, London. Session: Modelling with heavy tails: computational issues S. Makarova, W. Charemza, C. Francq, J. Zakoian, Heavy tailed time series: estimation and numerical issues for dependent observations CFE 2012, Oviedo. Session: Modelling with heavy tails: applications P. Jelonek, Generating tempered stable random variates from mixture representation W. Charemza, Z. Fan, S. Makarova, Y. Wang, Simulated minimum distance estimation in macroeconomics CFE 2013, London. Session: Probabilistic forecasting: density forecasts and disagreement W. Charemza, C. Diaz Vela, S. Makarova, Two-dimensional fan charts and uncertainties CFE 2014, Pisa . Session: Modelling with heavy tails: applications S. Makarova, W. Charemza, C. Diaz, Ex-post inflation forecast uncertainty and skew normal distributions: 'back to the future' approach
Start Year 2010
 
Description CREST 
Organisation Centre for Research in Economics and Statistics (CREST)
Country France 
Sector Academic/University 
PI Contribution CREST is our partner within the joint ESRC/ORA award. We provide macroeconomic and applied econometrics expertize in modelling macroeconomic uncertainty. We also conducted applied analysis and solve numerical problems using high powered parallel computer at the University of Leicester.
Collaborator Contribution They provide theoretical expertize and advice related to solving problems of modelling macroeconomic uncertainty. In particular they contributed by researching problems of estimation and testing of weakly dependent dynamic time series, applied by us in empirical modeling.
Impact Paper 'Testing for policy effects in ARMA-GARCH model' (provisional title), by W. Charemza, C. Francq, S. Makarova and J-M. Zakoïan is at final stages of preparation and will be submitted to journal by the end of February. Papers produced by our CREST collaborators and assigned to the ANR collaborative project have been added to the output. In such case, papers developed solely within the ANR project are, wherever relevant, marked by the following statement: 'Produced within the collaborative ANR project, as the part of joint ESRC/ORA project'. We asked the Researchfish the question about the need of reporting the output of our collaborators within this joint award, but we have not received a reply.
 
Description Eesti Pank 
Organisation Bank of Estonia
Country Estonia 
Sector Public 
PI Contribution The project will help to improve the forecasting system at the Central Bank of Estonia (Eesti Pank), by adding features developed within the project.
Collaborator Contribution Central Bank of Estonia has secured accommodation and subsistence in Tallinn for three months (November 2014-February 2015) for Svetlana Makarova, co-investigator in the project.
Impact There is no output so far. Joint paper with a member of staff of the Central Bank of Estonia is expected in 2015.
Start Year 2014
 
Description NBP 
Organisation National Bank Poland
Country Poland 
Sector Public 
PI Contribution We are offering help in implementation by the Bank the forecasting methods developed within the project.
Collaborator Contribution The national Bank of Poland has invited, at their expense, Wojciech Charemza to give a seminar in September 2013 and Svetlana Makarova, co-investigator within the project, to participate in a workshop in November 2014, which might lead to the implementation.
Impact None so far. It is possible that a joint paper with members of staff of the National Bank of Poland will be produced.
Start Year 2013
 
Description UCL, Economics 
Organisation University College London
Country United Kingdom 
Sector Academic/University 
PI Contribution Workshop describing current developments of the theory and practice of modelling of macroeconomic uncertainty was organized jointly with the Department of Economics, UCL (with another contributor being UCL School of Slavonic and East European Studies), which benefited staff and postgraduate/PhD students of the UCL Department of Economics.
Collaborator Contribution UCL department of Economics provided funds for travel and accommodation of some participants, and also provided secretarial and administrative assistance to the workshop.
Impact To avoid duplication, the seminar output is listed under the entry for the School of Slavonic and East European Studies.
Start Year 2013
 
Description UCL-SSEES 
Organisation University College London
Department School of Slavonic and East European Studies (SSEES)
Country United Kingdom 
Sector Academic/University 
PI Contribution Workshop describing current developments of the theory and practice of modelling of macroeconomic uncertainty was organized in collaboration with, and at the premises of the School of Slavonic and East European Studies, which benefited their staff and postgraduate/PhD students. Another collaborator was the Department of Economics, UCL.
Collaborator Contribution School of Slavonic and East European Studies contributed to organization and financing of the project workshop 'Uncertainty and Economic Forecasting', which took place at the School premises. Their contribution consisted of financing travel and accommodation of some participants, contribution to catering, providing office space for the workshop.
Impact Papers presented at the workshop: Aastveit, K.A. (Norges Bank): Economic uncertainty and the effectiveness of monetary policy Baker, S. (Stanford University): Measuring economic policy uncertainty Bidder, R. (Federal Reserve Bank of San Francisco): Robust stress testing Conrad, C. (Heidelberg University): Cross sectional evidence on the relation between monetary policy, macroeconomic conditions and low-frequency inflation uncertainty Galvão, A. (Warwick Business School): Measuring macroeconomic uncertainty: US inflation and output growth Giacomini, R. (University College London, Cemmap and CEPR): Forecasting with judgment Glass, K. (University of Hamburg): Did the real-time information content of Eurostat's macroeconomic data worsen? Istrefi, K. (Goethe University Frankfurt): Economic policy uncertainty and inflation expectations Krüger, F. (Heidelberg Institute for Theoretical Studies): Combining survey and Bayesian VAR forecasts of US macro variables: evidence from entropic tilting Ludvigson, S. (New York University and NBER): Measuring uncertainty Makarova, S. (University College London): Term structure of inflation forecast uncertainties and skew normal distributions Müller, U.K. (Princeton University): Measuring uncertainty about long-run predictions Orlik, A. (Federal Reserve Board): Understanding uncertainty shocks and the role of Black Swans Rich, R. (Federal Reserve Bank of New York): The measurement and behaviour of uncertainty: evidence from the ECB survey of professional forecasters Sekkel, R. (Bank of Canada): Macroeconomic uncertainty through the lens of professional forecasters Sheng, X.: (American University): Measuring Uncertainty of a Combined Forecast Tamoni, A. (London School of Economics): The scale of predictability
Start Year 2010