Sensitivity to other adjustments and the effect of capital non-smoothing
As said before, according to EC’s official methodology, smoothing the series of Working Hours or Participation Ratio is carried out via Hodrick-Prescott (HP) Filtering. These series affect the estimated potential output path as described in equation (6.4). The associated “end-point bias” problem is surpassed by extending the series forward with forecasted values from estimated autoregressive models. The order of the autoregressive model and the length of the extension are
Figure 6.8 Estimates of the Non-Accelerated Wage Rate of Unemployment (NAWRU) for Greece under alternative anchors for the level of Structural Unemployment (SUR) in 2026: (1965-2018). NAWRU (EC Anchored) is calculated using the default anchor used by European Commission (SUR = 12.05%)
both chosen based on well-defined statistical criteria. The sensitivity of the final smoothed series to the choice of each of the above parameters was assessed via a battery of estimations and extensive filtering. Results appear rather insensitive to the choice of the parameters and remain available upon request. Small short-run variation of the unsmoothed series of the Greek economy, constitutes the basic reason for their insensitivity to HP filtering.
A comment on the role of capital in EC’s methodology should be made here; as said, EC’s methodology conducts no smoothing on the series of capital stock in the economy. From equation (6.2), this implies that at times capital stock falls more rapidly than the economy’s Gross Domestic Product, TFP increases. This makes the potential output path behave less procyclically, which is reasonable to assume since capital outflows are easier to revert than disinvestment. This counter-cyclical effect could be further enhanced by using a smoothed capital series in the potential output equation (6.4). This could prevent rapid falls in potential output and subsequent narrowing of output gaps even though no actual growth takes place in the economy.
The effect of rapid capital reduction is of concern for Greece, which has experienced a huge fall in capital stock as a result of the vast disinvestment (Christodoulakis and Axioglou 2017) coupled with large capital flights abroad since the outburst of the economic crisis. The Greek economy has roughly lost 100 billion euros in the capital since 2010, corresponding to 4.5 percentage points loss in terms of potential output, according to equation (6.4). This reconfirms the evidence depicted in Figure 6.1, that the shrinking potential output explains part of the narrowing of the Greek economy’s output gap.