In the last episode of the EtlaTalks spring season the topic of discussion is nowcasting and how does it differ from forecasting? And what has caused its meaning to grow in economics? Researcher Paolo Fornaro from Etla also estimates the possible developments of nowcasting in the future. The interviewer is the editor-in-chief of TalousTV Jouko Marttila.
In recent years, nowcasting has become a common activity performed by economic institutions and private companies, alongside forecasting. Forecasting, which has been used since the beginning of the macro economics, aims to predict the direction of the economy in the medium and long run (e.g. 6 months or one year in the future). Nowcasting again focuses on receiving better picture of the current state of the economic. The need to get timely predictions, together with the improved availability of data and computation capability, enabled the emergence of nowcasting.
– It takes couple of months or longer for statistical institutes to publish their official estimates of the economy. For an example GDP figures are published 60 days after the end of the reference quarter. That is why economists were interested to know if they could use some statistical method or model that could faster give clearer picture of the current state of the economy. And so nowcasting was born, says Fornaro.
According to Fornaro the emergence of the nowcasting in economic is due to new developed statistical methods that can process lot of data. In traditional statistical models, big amount of data could lead to poorer prediction or the so-called “curse of dimensionality”. Instead, novel statistical techniques have allowed economists to include large number of variables in their models, without deteriorating the quality of the predictions. Digitalization also enabled the possibility to gather and use more data.
Fornaro thinks that nowcasting is not going to replace forecasting but both are still relevant and different both in terms of perspective and use in the economics. Fornaro also sees possible developments in nowcasting.
-Nowcasts are like black boxes: they give us a certain prediction, but we are not entirely sure why we got that prediction. There have been attempts to solve this problem, e.g. by examining the contributions that individual variables have on the nowcast, in what researchers call “news analysis”. However, there is progress to be made in this direction.