Structural Break Adjustment of National Accounts and Employment Data Series of Liechtenstein
Published data series may contain structural breaks. This is normal and there are many reasons for this, for example because methodological concepts or guidelines change over time. To make time series more interpretable, the original data series should be adjusted for the breaks (before the statistical analysis (i.e., the historical series should be revised). Also, temporary statistical outliers in economic time series may also require adjustment if they cannot be justified economically but originate from compilation errors.
For econometric analysis, the question of whether underlying data should be adjusted for structural breaks and outliers is a recurring task. In the research project, an updated and consolidated adjustment for structural breaks was made based on earlier, own backward estimates and structural break adjustments of individual Liechtenstein data series. In the process, these various adjustments were also reconciliated to further account for summation conditions and other considerations (e.g. should the totals of the economic sectors sum to the total economy aggregate, or should individual aggregates in the national accounts sum up to GDP).
On the one hand, the adjustments concerned breaks in the national accounts in 2013 (ESA conversion), 2016 (new definition of economic sectors), and 2018. In the course of this process, the following time series were newly adjusted: Liechtenstein's gross domestic product back to 1972, gross national income back to 1995, national income back to 1954, and gross value added by economic sector (industry, financial services, general services, agriculture) back to 1998. On the other hand, employment figures (number of employees and full-time equivalents, total and by economic sector) were adjusted back to a time point between 2000 and 1970, depending on the data series.
The generated time series were integrated into the Applied Economic Analysis database and will be made available to the public via the data pool.
Project duration: 2022
