Economic policy uncertainty: an index to measure it

Summary:

· The Economic Policy Uncertainty Index (EPU) measures the level of uncertainty surrounding economic policy through press coverage, fiscal measures, and discrepancies in economists’ forecasts.

· The EPU provides an original approach for estimating levels of uncertainty and comparing countries and major economic and geopolitical events.

· There are limitations to the construction of the index.

· The selection of media outlets used tends to skew the measurement of uncertainty…

· … but it could be improved, in particular by expanding and modifying the selection of newspapers.

For JM Keynes, the concept of uncertainty is central to economic analysis. In his major work,  » The General Theory of Employment, Interest and Money, » he argues that uncertainty causes economic agents to be cautious, investing and hiring less than they should. In economics, this « uncertainty » can be defined as a phenomenon of mistrust and lack of visibility in economic activity in the short and/or medium and long term.

The purpose of this article is to present the Economic Policy Uncertainty (EPU) indicator, which is an extension of initial work to quantify the uncertainty generated by economic policy-making. We will first present the methodology used to construct the indicator and its value in macroeconomic analysis. We will then highlight its limitations before concluding with potential avenues for improving the EPU indicator.

1. Definition and methodology

The EPU indicator, developed in 2012 by N. Bloom, Scott R. Baker, and Steven J. Davis, is an original and innovative tool that could become a benchmark for measuring uncertainty related to economic policy. The EPU covers 13 of the major economies, plus a^14th for Europe.

Not all indicators are constructed in the same way. Given the nationality of its creators, the methodology was logically established for the United States, and the US EPU is therefore the only one to use the following three elements:

The^first element: The News Index

This uses a fixed selection of newspapers to quantify economic policy uncertainty. It corresponds to the number of articles published per month in the same newspaper in which a combination of the keywords « uncertain » and « uncertainty » appear with « economy » and « economic, » as well as one or more political terms relating to political uncertainty. Each monthly series from newspapers is normalized so that they have a standard deviation of 1 before 2011, and an average is calculated for the different newspapers for each month in order to obtain EPU indicators at the European and national levels. Finally, these EPU indicators before 2011 are normalized so that their average is equal to 100. It should be noted that EPU Europe will use newspapers[2] from Germany, France, Italy, Spain, and the United Kingdom.

The second element: Tax provisions

This component, which only applies to the United States, uses reports from the Congressional Budget Office and establishes a list of temporary provisions in the US federal tax code that are set to expire within the next ten years. This component of the EPU is obtained by calculating the dollar amount of provisions set to expire for each year. This component quantifies the degree of uncertainty regarding changes in tax provisions. European indicators do not use this component.

To calculate this, we look at the Federal Reserve Bank of Philadelphia’s Survey of Professional Forecasters; the standard deviation between respondents’ forecasts[4] on CPI variables (inflation forecasts), federal spending, and state and local government spending, thus measuring the uncertainty in economists’ forecasts regarding macroeconomic variables and public spending.

To obtain the weighted overall index, each element must be normalized by its own standard deviation prior to January 2012. The average value of each element is then calculated using 1/2 for the News Index and 1/6 for the other three measures (the fiscal policy indicator and the CPI and budget deficit forecast spreads).

Since 2014, the European indices no longer include this^third component, which used forecast deviations from Consensus Economics[5]. European EPU indices are now only indicators because they use only the first element, meaning that EPU Europe, France, Germany, Italy, and Spain = News_Index. This difference can also be explained by the absence of a European equivalent to a « list of temporary tax provisions. » This element cannot therefore be taken into account for European indicators.

2. The value of such an index

The EPU is useful because it provides an initial measure of uncertainty, which can be compared between countries (or regions) and between different major economic and geopolitical events.

The News Index provides an original measure based on press coverage, showing the attention and media coverage given to an event or political announcement. The^second element helps to better visualize the pace and scale of a country’s tax measures. Companies and investors need fiscal visibility to plan their investments. It is therefore an important component of the EPU index. As for forecast deviations, they allow us to observe the ability of financial markets to read and predict the economic situation. Significant deviations would reflect their difficulty in anticipating economic variations, which could be due to a context of uncertainty.

The EPU would help to better visualize the levels and shocks of uncertainty associated with economic policy-making and thus help to interpret the economic situation.

3. An indicator with limitations

We can identify three major limitations. The first is that the volume of articles published by the media is not an accurate reflection of the concerns of socio-economic and/or political actors. The number of articles is more indicative of whether economic news is generating a lot of reaction from investors, businesses, and the media, and therefore measures the media importance of a news item or economic policy announcement. It is also regrettable that the sample only includes two newspapers per country for the European indicators.

The second concerns taxation. European EPUs do not take it into account in the calculation of the index, yet taxation is at the heart of economic policy. Only_{the UPUUSA measures} future tax provisions in concrete terms. Their pace and scope are therefore known, giving greater visibility to economic actors, who can better plan their activities.

The third concerns the specific characteristics of each country, which tend to influence their national indicator in a specific way. These parameters are difficult to quantify, but we can cite the country’s culture, the organization of political life, and the way the press deals with economic and political issues. One identifiable bias is the « political slant » of a newspaper; a right-wing newspaper, for example, will tend to be more critical of a left-wing government.

4. Areas for improvement

The News Index offers a new approach based on the occurrence of articles dealing with economic uncertainty in the press. While this approach has some limitations, there are several avenues for improvement in measuring economic uncertainty.

One avenue would be to modify the sample of newspapers to better reflect a country’s overall uncertainty. Newspapers could be selected based on circulation, i.e., those that are most widely read. Alternatively, a selection of « neutral » newspapers, i.e., those with no political bias, could be made, which would better reflect general opinion. This would be a broader selection of newspapers, including newspapers from across the political spectrum, which could mitigate the bias. Another possibility would be to combine the news index with other variables, such as changes in borrowing rates, stock market prices, exchange rates, or reports from rating agencies, which are closely followed by economic actors.

Another solution would be to use social media. Considering that uncertainty and confidence can change very quickly and that the print media is becoming less and less influential, the use of social media may be justified. The instantaneous nature of reactions and the circulation of information should have a much greater impact on uncertainty than the print media. In addition, a significant number of revelations, but also potentially harmful rumors, first appear on social media. A tweet is likely to circulate much more quickly and be read by a larger number of people than a press article on the economic situation.

The third area for improvement is to recreate a document on tax provisions in Europe (^second element). One could imagine a European « congressional budget office » establishing a list and amounts of tax provisions for the eurozone and for each country. This would be beneficial for centralizing and quantifying savings measures and tax policy.

Another option would be to reintegrate the uncertainty calculated from forecast deviations using the ECB’s « survey of professional forecasters SPF. » Adding new variables in addition to the CPI and budget deficit, such as manufacturing output, could strengthen the measurement of the indicator.

Conclusion

The EPU indicator offers an innovative approach based on the occurrence of articles dealing with economic uncertainty in the press, tax provisions, and financial market forecast deviations. It has limitations, particularly in terms of its construction but also in terms of the sample of newspapers used. Increasing the number of newspapers or using social networks or online media are potential avenues for improvement. Nevertheless, the question of whether uncertainty is a cause or a consequence of the economic downturn remains unresolved. In a future article, we will look at the specific case of the indicator for France and how other aggregates could be used to visualize a country’s economic uncertainty.

References:

Scott R. Baker, N. Bloom, and Steven J. Davis (2015): « Measuring Economic Policy Uncertainty » NBER Working Papers Series

Bloom Nicholas (2009): « The impact of uncertainty shocks » in Econometrica, vol. 77, No. 3, pp. 623-685, The Econometric Society