by
Gikas A. Hardouvelis
www.hardouvelis.gr
The 10-year old Greek economic depression is unprecedented. Since 2007, not only did real GDP decline by over 26%, but beginning in early 2015 economic activity has stagnated as well. The original fall in output was the outcome of a partly necessary ultra-restrictive fiscal policy, which brought the primary general government deficit from 10% of GDP down to a zero balance in 2014. Yet the subsequent prolonged stagnation is not just due to bad economics but primarily bad politics. In 2015, the moment the economy had just begun taking off, when FDI was flowing into the country and economic sentiment was rising a new and inexperienced coalition government took office. It was led by left wing and right wing populists who aggressively confronted Greece’s lenders, destroyed credibility, starved the country and its banks of liquidity, brought capital controls and aborted the recovery. The country muddles through ever since in what is Greek crisis Phase II. In 2017, positive growth came back but was the lowest in Europe.
Economists have not yet provided a full quantitative account of the economic forces behind the Greek depression. Most scientific quantitative attempts rely on theoretical Dynamic Stochastic General Equilibrium models and ad hoc calibration exercizes, all of which are riddled with unrealistic assumptions about optimal economic behavior. The role of uncertainty is missing from that analysis. Yet uncertainty, economic or political, seems prima facie to be a key variable behind the drop in output and its subsequent stagnation. Ιt is worth studying in detail.
Together with co-authors from the University of Piraeus and the London School of Economics, I have constructed indices of economic policy uncertainty (EPU) and political uncertainty (POLU).i The methodology is based on textual analysis. We have swiped through more than half a million articles published in four Greek newspapers from January 1999 to December 2017 and recorded the frequency of key words related to the previous specific concepts of uncertainty. We also constructed sub-indices of EPU, related to monetary policy uncertainty (EPUM), fiscal policy uncertainty (EPUF), currency or Grexit uncertainty (EPUC), pension policy uncertainty (EPUP), and banking uncertainty (EPUB). We further partitioned the fiscal uncertainty into debt uncertainty (EPUD) and tax uncertainty (EPUT).ii
During the last twenty years EPU fluctuated a lot and rose significantly during major international events, such as the September 2001 terrorist attack, the 2003 Iraq war, or the 2008 Lehman Brothers episode. Naturally, it captures major Greek events as well, such as the fiscal crisis of late 2009 – early 2010, the Papandreou plans for a referendum in October 2011, the May 2012 and January 2015 elections, or the June 2015 announcement of a referendum.
The Greek EPU index is positively correlated with a corresponding - and pre-existing - global EPU index. The correlation went up during the international financial crisis, apparently because the crisis affected Greece as well. It declined in the period 2010-2014, during Greek Crisis Phase I, as uncertainty was driven mainly by local events. Then, since early 2015, during Greek crisis Phase II it has collapsed to zero, as Greece is completely sidelined from the international scenery, being absorbed by its own domestic problems.
Greek EPU fluctuations provide information over and above global EPU fluctuations, which are able to explain both the direction and the magnitude of the change in key domestic macroeconomic variables. An increase in EPU is followed by a decline in industrial production, a decline in employment, a decline in GDP, a decline in investment, a decline in economic sentiment, a decline in the stock market and a rise in bond yields. All these directional changes are consistent with economic theory. The magnitudes of the responses also appear reasonable. Average EPU increased by 22% from the pre-crisis part of the sample to the crisis part. This size increase can explain about 2/3 of the cumulative drop in industrial production, employment or economic sentiment and the full contraction in household deposits.
Political uncertainty is less strong in explaining the fluctuations in economic variables. It does, however, perform better in the explanation of the 10-year bond spread over Germany. And it explains quite well household deposit behavior.
Three EPU sub-indices seem to dominate the fluctuations of EPU and all three are related to different aspects of the Greek crisis: EPUD, which captures the debt crisis, EPUB, which captures the banking crisis, and EPUC, which captures uncertainty about Grexit. These three are able to explain not only the depth of the Greek crisis but its unusual length as well, as they contribute significantly to the variability of the macroeconomic variables during the crisis period.
Overall, the negative relation between economic policy uncertainty and the economy is strong and robust. An open question is the direction of causality: Is it primarily policy uncertainty that drives the negative effects on the economy or the opposite? That is, could it be instead that a decline in economic activity raises uncertainty? Or perhaps do both negative effects operate simultaneously, feeding on each other and boosting the observed negative association? Future research might tell us the answer.
i See G. Hardouvelis, G. Karalas, D. Karanastasis, and P. Samartzis, “Economic Policy Uncertainty, Political Uncertainty and the Greek Economic Crisis,” April 3, 20018, SSRN Working Paper # 3155172.
ii Those indices will soon be freely available to the research public through the site www.hardouvelis.gr.
