Distilling the macroeconomic news flow

Alessandro Beber, Michael Brandt, Maurizio Luisi 19 April 2013

a

A

Timely measurement of the state of the economy relies traditionally on low-frequency observations of a few economic aggregates that refer to previous weeks, months, or even quarters. A prominent example is the advance estimate of GDP released quarterly about a month after the end of the quarter.

The low frequency and delayed observation of any such economic aggregate considered in isolation stands in sharp contrast with the rich macroeconomic news flow that market participants observe almost daily. This news flow contains information that agents use to learn about the economy in the absence of private information. In particular, the macroeconomic news literature has identified a large cross-section of dozens of different news releases that have significant and immediate effects on financial markets. Figure 1 shows the typical timeline for the most important macroeconomic releases in the US.

Figure 1. Reporting structure of main US macroeconomic news

The macroeconomic news flow

We obtain data on the dates, release times, and actual released figures for 43 US macroeconomic announcements covering the period from 1997 through 2011, for a total of more than 8,000 announcements over about 3,800 working days. Most macroeconomic indicators are released on different days and at different frequencies, making it difficult to process the flow of information in a systematic and consistent way. Figure 1 shows that actual news releases occur with a variety of different lags with respect to the month they are referring to.

Based on both empirical evidence and economic rationale, we first separate the aggregate economy into two broad dimensions: the nominal inflation-related announcements and news that relates to real growth. Growth data, in turn, come in two flavours: objective realisations of past economic activity and subjective often forward-looking views derived from surveys which we label `’macro sentiment’. Finally, economic activity can be split one last time into information relating to output versus employment. The following diagram summarises the different subset of macroeconomic news:

It is worth reiterating at this point that we do not include any market-based data (such as stock prices, interest rates, credit spreads, or VIX) in our analysis, as they represent already the market's interpretation of the macroeconomic news flow.

A simple methodology

We propose to distill the economic news flow observed by market participants into a set of indicators measuring the distinct aspects of the economy described above: inflation, output, employment, and macroeconomic sentiment. Specifically, within each category subset, we let the data speak for itself by extracting the first principal component of that subset of data.

The key inputs to our methodology are the within news category correlation. There are two issues that need to be addressed in computing these correlation matrices.

  • First, the data is in the form of an unbalanced panel due to some of the series being initiated after the start date of the estimation window;
  • Second, the data is naturally persistent, partly due to autocorrelation of the data in announcement time, partly due to the cross-sectional misalignment of the news in calendar time, and partly due to the forward filling of missing data.

We address the first unbalanced panel issue by using a correlation matrix estimator along the lines of Stambaugh (1997), who shows how to adjust first and second moments estimates for unequal sample lengths. The intuition of his approach is to use the observed data on the longer series, along with a projection of the shorter series on the longer ones estimated when both are observed, to adjust the moments of the shorter time series.

To correct for the persistence in the data, we adopt the same approach used in high-frequency asset prices, where asynchronous and infrequent trading creates a misaligned and locally constant panel of observations (e.g. Ait-Sahalia, Mykland, and Zhang 2005). Specifically, we use average correlation matrix estimates obtained on a sparse backward subsampling of the news series (see Beber, Brandt, and Luisi 2013 for further details).

Our paper contributes to the literature that attempts to measure the state of the economy in a time-series setting based on fundamental economic data (see Banbura et al. 2012 for a survey). Our approach is simple, robust (no numerical optimisation is required), and can effectively handle the large number of announcements that are relevant for tracking the evolution of macroeconomic conditions in real time. At the same time, our methodology deals with data released at different frequencies and with missing observations.

Empirical evidence

Figure 2 shows that the economic activity factor (which combines output and employment information as they are highly correlated) as well as a macroeconomic sentiment factor have sensible dynamics. The greatest dips in both series are well aligned with the ex-post defined National Bureau of Economic Research recession phases. The macroeconomic sentiment factor, obtained from consumer and business confidence releases, is highly correlated with economic activity, but appears to lead fundamentals especially around the most important turning points.

Figure 2. Real-time US economic activity (in blue) and macroeconomic sentiment (in red) factors.

Figure 3 shows that our real-time inflation factor exhibits dynamics that seem only weakly correlated with the growth factor (which combines economic activity and macro sentiment), with much more erratic variation, and has an unclear pattern in expansion versus recession phases.

Figure 3. Real-time US Growth (blue line) and Inflation (red line) factors.

We formally relate a real-time factor of economic growth to the CFNAI index (Chicago Federal Reserve National Activity Index), which is constructed by the Chicago Federal Reserve Board based on Stock and Watson (1989), released on a monthly frequency. The CFNAI is a very interesting comparison for our high-frequency indicator, because it utilises information for a large cross-section. However, the CFNAI indicator remains at the low monthly frequency level and does not deal with information released on different days over the month and with missing observations.

Figure 4. Real-time Growth factor (red line) and CFNAI index (blue line)

Figure 4 plots the CFNAI monthly indicator and our real-time growth factor. The two series are very similar. However, our growth index anticipates the turning points of the CFNAI index and, most importantly, is available at daily frequency in real-time.

We now compare our real-time growth index to the vintage version of the ADS index of Arouba, Diebold and Scotti (2009) at the weekly frequency. The construction of the Arouba, Diebold and Scotti index allows high-frequency measurement, but can only handle a limited number of news releases.
Figure 5 plots the Arouba, Diebold and Scotti index and the growth factor. The two series are again very similar. Our real-time growth index is much smoother and seems to somewhat anticipate the turning points of the Arouba, Diebold and Scotti index.

Figure 5. Real-time growth factor (red line) and ADS index (blue line)

 

Figure 6. Real-time Growth factor (in red), actual GDP release (in blue), Survey of Professional Forecasters GDP forecast (in black)

 

In a related empirical exercise, we find that our real-time growth factor has predictive power for future actual GDP releases and is highly correlated with the quarterly GDP expectations in the Survey of Professional Forecasters (see Figure 6). This is a remarkable feature given that, unlike other methodologies, our approach is not optimised to forecast GDP. The large correlation with the quarterly releases of the Survey of Professional Forecasters offers an intuitive interpretation of our growth factor as the high-frequency daily reading of economist expectations about macroeconomic fundamentals.

In summary, our empirical analysis shows that the output of our real-time approach is more timely and informative than more sophisticated but also more difficult to implement statistical techniques. Intuitively, the potential disadvantage of a simpler modelling framework is more than compensated by the sheer quantity of data our approach can effectively process.

Extensions

Our methodology is well suited to handling the many dimensions of news that could characterise various geographic areas. At the same time, the approach is extremely flexible and allows the construction of a real-time factor tailored on a specific group of countries, even changing over time.

For example, Figure 7 shows the daily real-time growth factor for the Eurozone, which is obtained from 183 different sources of macroeconomic information. The overall factor can be broken down in a ‘core’ and ‘periphery’ real-time growth component, which is constructed using macro news from a (potentially) time-varying group of countries.

Figure 7. Real-time growth factor for the Eurozone (in red), core countries (in orange), periphery countries (in blue)

Figure 8 shows a similar example for emerging markets. Also in this case, the cross-section of news can potentially be very large (for example, for Asia and Pacific we use 311 news items). It might be of interest to construct real-time growth factors for some specific geographic areas (e.g. Latin America), as well as the more overarching factor measuring growth for all emerging markets.

Figure 8. Real-time Growth factors Emerging Markets and Asia Pacific

References

Ait-Sahalia, Yacine, Mykland, Per A, and Zhang, Lan (2005), “How often to sample a continuous-time process in the presence of market microstructure noise”, Review of Financial Studies 18, 351-416.

Aruoba, S B, Diebold, F X and Scotti, C (2009), “Real-Time Measurement of Business Conditions”, Journal of Business and Economic Statistics, 27, 417- 427.

Banbura, Marta, Giannone, Domenico, Modugno, Michele and Lucrezia Reichlin (2012), “Now-casting and the real-time data flow”, CEPR Discussion Paper 9112.

Beber, Alessandro, Michael W Brandt, and Maurizio Luisi (2013), “Distilling the Macroeconomic News Flow”, CEPR Discussion Paper 9360.

Stambaugh, Robert F (1997), “Analyzing investments whose histories differ in length”, Journal of Financial Economics 45, 285-331.

Stock, J H and M W Watson (1989), “New Indexes of Coincident and Leading Economic Indicators”, in O J Blanchard and S Fischer (eds.) NBER Macroeconomics Annual, 352-394.

a

A

Topics:  Macroeconomic policy

Tags:  Information, news

Professor of Finance, Cass Business School

Professor of Economics at Fuqua School of Business, Duke University

Managing Principal, Quantitative Investment Solutions Ltd

Events