In a previous post I mused: "Why use a monthly leading index when you have a weekly one to hand?" Forget that - how about a daily economic index? Martin D.D. Evans of Georgetown University has a new article in the latest International Journal of Central Banking which proposes just such a thing.
His paper, Where Are We Now? Real-Time Estimates of the Macroeconomy, describes a method for calculating daily realtime estimates of the current state of the US economy:
The estimates are computed from data on scheduled U.S. macroeconomic announcements using an econometric model that allows for variable reporting lags, temporal aggregation, and other complications in the data. The model can be applied to find real-time estimates of GDP, inflation, unemployment, or any other macroeconomic variable of interest.
In this paper, I focus on the problem of estimating the current level of and growth rate in GDP. I construct daily real-time estimates of GDP that incorporate public information known on the day in question. The real-time estimates produced by the model are uniquely suited to studying how perceived developments in the macroeconomy are linked to asset prices over a wide range of frequencies. The estimates also provide, for the first time, daily time series that can be used in practical policy decisions.
Evans argues that you can model GDP growth "as the quarterly aggregate of an unobserved daily process for real economy-wide activity."
The model also specifies the relationship between GDP, data releases on GDP growth, and data releases on a set of other macroeconomic variables in a manner that accommodates the complex timing of releases. In particular, I incorporate the variable reporting lags that exist between the end of each data collection period (i.e., the end of a month or quarter) and the release day for each variable. This is only possible because the model tracks the evolution of the economy on a daily basis.
His model captures quarterly GDP quite well, though it is more accurate to consider it as a continually-updated estimate rather than a daily one. He concludes:
In this paper, I have presented a method for estimating the current state of the economy on a continual basis using the flow of information from a wide range of macroeconomic data releases. These real-time estimates were computed from an econometric model that allows for variable reporting lags, temporal aggregation, and other complications that characterize the daily flow of macroeconomic information. The model can be applied to find real-time estimates of GDP, inflation, unemployment, or any other macroeconomic variable of interest.
Evans notes that "these findings give but a flavor of the possible uses for realtime estimates". I quite agree. I look forward to longer time series, and modelling of other economic data. It surely can't be long before one of the investment banks starts crunching out their own proprietary version, either.