Nowcasting industrial production index with high-frequency toll data

Abstract

Nowcasting is a phenomenon that has recently become a widely researched area in the context of official statistics. The main reason is an exploration of new data sources that can be used in the early estimation of economic activity. An electronic system that records passages of vehicles through roads generates such big data of early available information. It assembles high-frequency data, which can be collected relatively shortly afterwards, that holds information related to the transport of goods and commodities through roads. The transportation of goods and commodities, usually carried out by trucks, through roads can provide an early indicator of the economic activity of industries that rely on road transport. The objective of this paper is to propose an index, which measures monthly changes of the number of vehicles on roads. The truck-passage index is able to provide a nowcasting estimate of the economic output of industries that heavily rely on road transport. In the empirical study, we use toll data from the Slovak electronic toll system that captures in-and-out passages of vehicles through monitored sections of highways. The truck-passage index is then compared to a monthly Industrial Production Index (IPI) that aims to validate its accuracy, and, potentially, provide its early nowcasting indicator. Furthermore, we use an Empirical Mode Decomposition (EMD) to extract cyclical components of both truck-passage and industrial production indices that allow for the in-depth analysis of fluctuations of economic time series. Based on the empirical results, we can conclude that the truck-passage index has a high potential of capturing fluctuations in the industrial production output.