Imputing monthly values for quarterly time series. An application performed with Swiss business tendency survey data
METADATA ONLY
Loading...
Author / Producer
Date
2022-09-14
Publication Type
Conference Paper
ETH Bibliography
yes
Citations
Altmetric
METADATA ONLY
Data
Rights / License
Abstract
This paper documents an applied investigation into strategies and algorithms to deal with the problem of missing higher frequency data. We refer to Swiss business tendency survey (BTS) data, in particular the KOF manufacturing surveys, which are conducted in both monthly and quarterly frequency. Accordingly, some information is available at quarterly frequency only. There is a wide range of ways to address this problem comprising univariate and multivariate approaches. We resort to different multivariate imputation algorithms and apply them to generate monthly series out of quarterly series from the KOF BTS in the Swiss manufacturing sector and compare the results. Our strategy to compare the suitability of the different approaches is to make sure that we do possess adequate reference series for the model selection stage. To this end, we apply our procedures to series that are monthly, from which we create artificial quarterly data by deleting the same two out of three data points from each quarter. The candidate series for the imputation of the missing (deleted) observations are given by the entire set of time series that are resulting from the monthly KOF manufacturing BTS survey. In this way, we resort to a set of indicators that share the common theme, the Swiss business cycle. With this set of indicators, we conduct the different imputations. On this basis, we then run standard tests of forecasting accuracy by comparing the imputed monthly series to the original monthly series (internal validation). Finally, we take a look at the congruence of the imputed monthly series from the quarterly survey question on firms’ technical capacities with existing monthly data on the Swiss economy (external validation). Due to the massive shock from the Covid-19 pandemic, we restrict the in-sample analysis to data from 1967m2 to 2019m12. This notwithstanding, we also take a look at how well our imputations would have fared in real time during the pandemic in 2020 and 2021
Permanent link
Publication status
published
External links
Editor
Book title
Journal / series
Volume
Pages / Article No.
477
Publisher
Centre for International Research on Economic Tendency Surveys
Event
36th CIRET Conference
Edition / version
Methods
Software
Geographic location
Date collected
Date created
Subject
Frequency transformation; Business tendency surveys; internal/external validation
Organisational unit
02525 - KOF Konjunkturforschungsstelle / KOF Swiss Economic Institute
06331 - KOF FB Konjunkturumfragen / KOF Business Tendency Surveys
06335 - KOF Institutsdienste / KOF Institute Services
06330 - KOF FB Konjunktur / KOF Macroeconomic forecasting
Notes
Conference lecture held on September 14, 2022.