MONETARY POLICY REPORT | SUMMER 2021 (box 2)
(authors: Tomáš Adam, Ondřej Michálek, Eva Slezáková)
The COVID-19 pandemic presented fundamental challenges for macroeconomic forecasters and monetary policymakers. Its unprecedented nature meant that the econometric models they normally use had difficulty assessing the depth of the downturn and identifying turning points in the cycle in a timely way. Traditional indicators come with a delay. They were also subject to significant data uncertainty during the pandemic. Forecasters internationally and in the Czech Republic therefore started to make greater use of alternative indicators of economic activity. This opened the way for new approaches to monitoring domestic economic activity almost in real time.[1]
The Rushin, a weekly index of economic activity, is one of the new instruments that help us refine our knowledge of GDP in the current quarter.[2] Its basic building block consists of high-frequency indicators (daily or weekly), complemented by routinely used monthly indicators. The index contains information from the most important sectors of the Czech economy. This ensures that it is comprehensive, robust and unbiased. The first step in creating the index was to align the time series (by interpolating the monthly data) to a common weekly frequency and to create 13-week moving averages and subsequently their quarter-on-quarter changes so that the index corresponded to the published quarterly GDP growth.
The resulting index is made up of ten variables, chosen on the basis of the following criteria: economic relevance, correlation with domestic GDP, lags in data availability, and correlation with other indicators. Developments in manufacturing, which are so important for the Czech economy, are captured by high-frequency electricity consumption indicators and the monthly industrial production index. The frequency of internet searches for the term “benefit” (meaning unemployment benefit) from Google Trends monitors the equally important labour market developments. Household consumption, which accounts for about half of GDP, is represented by monthly sales in services and retail. Data on freight transport on Czech and German motorways and the ifo Business Climate Index capture international trade and external demand. The OECD Composite Leading Indicator and the Prague Stock Exchange index add a forward-looking dimension.
The Rushin very accurately captures historical GDP growth (see Chart 1). The periods of turbulent economic change seen during the Global Financial Crisis and the current coronavirus pandemic are particularly worth mentioning. In both these periods, the weekly index of economic activity managed to capture the ongoing downturn of the domestic economy and estimate its depth with a high level of accuracy. Its GDP estimate for the shallow recession that hit the domestic economy in 2011 and 2012 was also good.
Chart 1 – The Rushin estimates mostly match the historical GDP growth
GDP – observed q-o-q changes in %; Rushin – weekly moving q‑o‑q changes in %
The Rushin’s ability to capture ongoing economic output was also tested by conducting an exercise in pseudo real time. An estimate of the index was calculated for each week from 2016 to the present using only the data that were available at the time.[3] They were used to estimate GDP growth in the given quarter at the end of each month (see Chart 2) by applying the “bridge equations” approach, which converts data from monthly to quarterly frequency and has been assessed in the academic literature as being very successful for short-term forecasting.[4] This generated three quarterly GDP growth estimates for each quarter. Chart 2 shows how they compare with the GDP figures published by the CZSO. The volatility and estimates of economic growth were low in the pre-pandemic period, but the situation changed with the start of the pandemic. The Rushin nonetheless dealt with this challenge well, and in most cases, as the dataset expanded during the quarter, the index converged towards the GDP growth estimates published several weeks later by the CZSO. It now ranks among the instruments used to help create short-term GDP forecasts at the CNB.
Chart 2 – The Rushin provides good GDP estimates even during the pandemic
real q-o-q changes in GDP in % in individual quarters; GDP estimates using Rushin at end of each month
The Rushin is named in honour of Alois Rašín, a prominent economist and politician from the period of the establishment of the independent Czechoslovakia. The modified name is a phonetic pun on Mr Rašín’s surname, while also capturing the timeliness (rush-) of the index (-in).
[1] The first estimate of the impacts of the pandemic on the Czech economy, based on electricity consumption data, was published by T. Adam and A. Michl in April 2020 – see https://www.cnb.cz/cs/o_cnb/cnblog/Prvni-odhad-dopadu-pandemie-COVID-19-na-ekonomiku-CR-zverejneno-8.-4.-2020/ (in Czech only).
[2] A preliminary version of the Rushin is introduced in the blog article: https://www.cnb.cz/en/about_cnb/cnblog/The-Rushin-An-Index-of-Czech-Economic-Activity/.
[3] The analysis does not take into account potential changes in the variables entering the index, although in view of the constantly evolving structure of the economy such ongoing potential adjustments are relevant. On the other hand, structural economic changes are only gradual and are of little significance for the period length chosen.
[4] This analysis will form part of a forthcoming CNB working paper – The Rushin Index: A Weekly Indicator of Czech Economic Activity (Adam, T., Michálek, O., Michl, A., Slezáková, E., 2021).