Following the methodology used by the Chicago Fed, I construct the National Activity Index (NAI) for Finland. The NAI is a monthly indicator designed to measure the state of the economy. It is constructed as the lead principal component of 28 monthly economic activity series.

The Federal Reserve Bank of Chicago describes their National Activity Index for the United States as follows: “The [Chicago Fed National Activity Index] CFNAI corresponds to the index of economic activity developed by James Stock of Harvard University and Mark Watson of Princeton University in “Forecasting inflation,” an article published in the Journal of Monetary Economics in 1999. The idea behind their approach is that there is some factor common to all of the various inflation indicators, and it is this common factor, or index, that is useful for predicting inflation. Subsequent research by former Chicago Fed President Charles Evans found that the CFNAI is also a useful gauge of the business cycle in the United States.”

Here I aim to construct a similar measure for Finland. Due to data availability, the exercise is limited to 28 readily available time series between June 1997 and October 2024.

Data

The time series used in this exercise follow in principle those used in the CFNAI. These include industrial and retail turnover and production, building, unemployment, wages and confidence indicators. All data are collected from StatFin with the exception of the unemployment rate (ECB Data Portal) and confidence indicators (Eurostat). Whenever seasonal adjusted data is appropriate but not available, I have manually conducted an X-13 adjustment using the seasonal R-package. The series are transformed into differences or log-differences. In order to conduct principal component analysis, each series is centered around zero and normalized to have unit standard deviation.

Series	Transformation	Manual seasonal adjustment
Building and dwelling production, building permits, residential buildings, m3	log-diff	yes
Building and dwelling production, building permits, non-residential buildings, m3	log-diff	yes
Building and dwelling production, building starts, residential buildings, m3	log-diff	yes
Building and dwelling production, building starts, non-residential buildings, m3	log-diff	yes
Building and dwelling production, building completions, residential buildings, m3	log-diff	yes
Building and dwelling production, building completions, non-residential buildings, m3	log-diff	yes
Bankruptcies instigated, number of enterprises	diff	yes
Bankruptcies instigated, number of employees	diff	yes
Turnover of trade (2021=100), wholesale and retail trade*	log-diff	no
Turnover of trade (2021=100), wholesale and retail trade and repair of motor vehicles and motorcycles*	log-diff	no
Turnover of trade (2021=100), wholesale trade, except motor vehicles and motorcycles*	log-diff	no
Turnover of trade (2021=100), retail trade, except motor vehicles and motorcycles*	log-diff	no
Volume index of industrial output (2021=100), total industries	log-diff	no
Volume index of industrial output (2021=100), mining and quarrying	log-diff	no
Volume index of industrial output (2021=100), investment goods	log-diff	no
Volume index of industrial output (2021=100), durable goods	log-diff	no
Volume index of industrial output (2021=100), other consumer goods	log-diff	no
Wage and salary indices by industry monthly (2021=100), whole economy*	log-diff	no
Index of turnover in industry monthly (2021=100), export turnover, capital goods*	log-diff	no
Index of turnover in industry monthly (2021=100), export turnover, durable consumer goods*	log-diff	no
Index of turnover in industry monthly (2021=100), export turnover, intermediate goods*	log-diff	no
Index of turnover in industry monthly (2021=100), export turnover, non-durable consumer goods*	log-diff	no
Unemployment rate	diff	no
Construction confidence indicator	diff	no
Industrial confidence indicator	diff	no
Retail confidence indicator	diff	no
Consumer confidence indicator	diff	no
Services confidence indicator	diff	no

NAI for Finland

I conduct principal component analysis to extract the common drivers of the data. As we can note below, the lead principal component catches about 11% of the total variance of the data set.

	PC1	PC2	PC3
Standard Deviation	1,770	1,541	1,473
Proportion of variance	0,112	0,085	0,077
Cumulative proportion	0,112	0,197	0,274

Below we can see the lead principal component plotted as the stationary series and as its cumulative value for illustrative purposes. The vertical blue lines are placed at September 2008 (Lehman Brothers bankruptcy), April 2020 (lockdowns) and February 2022 (Russias full-scale invasion of Ukraine).

Data and R codes