What is the ultimate goal of the model?
Our model produces a “nowcast” of GDP growth, incorporating a wide range of macroeconomic data as it becomes available. With this approach, we aim to read the real-time flow of information and evaluate its effects on current economic conditions. The platform provides a model-based counterpart to the more routine forecasts produced at the bank, which have traditionally been based on expert knowledge.
What is the modeling strategy?
The platform employs Kalman-filtering techniques and a dynamic factor model. The approach has a number of desirable features. It is based on:
- a reliable big data framework that captures in a parsimonious way the salient features of macroeconomic data dynamics;
- a design that digests the data as “news,” mimicking the way markets work.
What are the input data? What has been driving the data selection?
We include all the market-moving indicators—the same data that are also constantly monitored by market participants and commentators.
Why should we trust the model?
Extensive back-testing of the model, research findings, and practical experience have shown that the platform is able to approximate best practices in macroeconomic forecasts. The model produces forecasts that are as accurate as, and strongly correlated with, predictions based on best judgment.
The methodology has been tested for accuracy in many countries, including large developed economies (the euro area, Italy, France, Germany, Spain, the United Kingdom, Japan, and Canada), small open economies (Australia, Ireland, Belgium, New Zealand, the Czech Republic, and Scotland), fast-growing economies (Brazil, Russia, India, China, and South Africa), and developing economies (Mexico, Indonesia, and Argentina).
How should we read the output of the model?
- The model produces forecasts for all variables taking into account their dynamic interactions.
- Since it is a fully specified dynamic model, the platform provides an intuitive reading of the incoming data as “news.”
- The difference between two consecutive forecasts (that is, the forecast revision) is the weighted average of the news during the week.
- News is defined as the difference between released data and model predictions. The weights account for the information content as well as the timeliness of the data releases.
- The contribution of new data to the forecast revision is reported in the charts with colored bars. To make the charts easier to read, we grouped variables in a few broad categories. Detailed information about the contribution of specific variables is available in the accompanying tables.
Banbura, M., D. Giannone, M. Modugno, and L. Reichlin.
2013. “Nowcasting and the Real-Time Data Flow.
” In G. Elliott and
A. Timmermann, eds., Handbook of Economic Forecasting
, Vol. 2. Amsterdam: Elsevier-North Holland.
Bok, B., D. Caratelli, D. Giannone, A. Sbordone, and A. Tambalotti.
2017. “Macroeconomic Nowcasting and Forecasting with Big Data.
” Federal Reserve Bank of New York Staff Reports
, no. 830, November.
Giannone, D., L. Reichlin, and D. Small.
2008. “Nowcasting: The Real-Time Informational Content of Macroeconomic Data.
” Journal of Monetary Economics
55, no.4 (May): 665-76.