ANNEXES
Restricted
Annex 1. GDP regression results
For 2018 a further extension of the official AutumnForecast was needed. The approach taken by Unit B3 was to extend the output gap series (ygap) with one year and add this to the potential output (YPOT) estimated in the framework of the longer time horizon forecasts (t+5 and t+10) to become an estimate of expected GDP (Y) and GDP growth in 2018[1]. And thus GDP in 2018 is:
where is known from the Autumn2015 Final Forecast and is calculated using an AR(1) or AR(2) extension of the output gap:
The second lag on the output gap (AR2) was only added if the coefficient was significant. When insignificant we first tried to add an MA(1) term. Only when the ARMA(2,1) did not result in significant coefficients we returned to an AR(1) model. The specifications are shown in table 1. Since we did not find any ARMA specification which would lead to significant coefficients for Malta, the EPC's mechanical rule was applied.
GDP growth in 2018 is then:
Table 1 ARIMA specifications per country (values for beta, significance level and R²)
*: sign. with p<0.1**: sign. with p<0.05
Annex 2 :Investment regression results
Up until 2017 the information from the Autumn 2015 Final Forecast from the Commission has been used to calculate growth in investment (measured as gross fixed capital accumulation) for the different member states, as well as averages for EU28 and the EuroArea. The approach from the winter forecast is simply extended to 2018 to forecast investment. The ratio of investment (IQ) to potential output (YPOT) is regressed on past behaviour (AR4) and a constant. For some countries a time trend is added.
To calculate the value for2018, we forecast this regression. Since information is already available on potential output for 2018 in the Autumnforecast, we can then calculate the investment for 2018: and the growth in investment in 2018is . The specifications are shown in table 1.
Table 1 -ARIMA specifications per country on IYPOT(values for beta, significance level & R²)
IYPOT / constant / beta 1 / beta2 / beta3 / beta4 / time / R²be / 0.04 / ** / 0.82 / ** / 0.70
de / 0.06 / ** / 1.06 / ** / -0.32 / ** / 0.66
dk / 0.04 / ** / 1.25 / ** / -0.46 / ** / 0.77
el / 0.02 / 1.10 / ** / -0.26 / 0.27 / -0.20 / * / 0.85
es / 0.03 / ** / 1.58 / ** / -0.70 / ** / 0.92
fr / 0.06 / ** / 1.15 / ** / -0.43 / ** / 0.75
ie / 0.04 / ** / 1.30 / ** / -0.45 / ** / 0.83
it / 0.06 / ** / 1.30 / ** / -0.71 / ** / 0.41 / -0.27 / * / 0.00 / 0.82
lu / 0.08 / ** / 0.55 / ** / 0.31
nl / 0.06 / ** / 1.18 / ** / -0.46 / ** / 0.71
at / 0.05 / ** / 0.78 / ** / 0.02 / 0.69
pt / 0.04 / ** / 1.07 / ** / -0.29 / * / 0.72
fi / 0.09 / ** / 1.15 / ** / -0.57 / ** / 0.75
se / 0.08 / ** / 1.17 / ** / -0.57 / ** / 0.17 / -0.11 / 0.00 / 0.66
uk / 0.03 / * / 1.30 / ** / -0.74 / ** / 0.29 / * / 0.77
cz / 0.12 / ** / 0.88 / ** / -0.31 / 0.45
ee / 0.08 / ** / 1.22 / ** / -0.87 / ** / 0.34 / 0.62
hu / 0.05 / 1.15 / ** / -0.43 / 0.02 / 0.66
lv / 0.09 / ** / 1.06 / ** / -0.45 / ** / 0.59
lt / 0.09 / * / 1.02 / ** / -0.66 / ** / 0.17 / 0.48
pl / 0.06 / * / 1.40 / ** / -0.80 / ** / 0.11 / 0.77
sk / 0.07 / ** / 0.70 / ** / 0.60
si / 0.04 / 1.30 / ** / -0.54 / 0.09 / 0.77
cy / 0.02 / 1.36 / ** / -0.47 / * / 0.86
mt / 0.16 / ** / 0.25 / -0.01 / -0.10 / -0.10
bg / 0.04 / ** / 1.09 / ** / -0.27 / -0.02 / 0.79
ro / 0.06 / * / 0.74 / ** / 0.54
hr / 0.05 / 1.19 / ** / -0.43 / 0.02 / 0.71
*: sign. with p<0.1
**: sign. with p<0.05
1
Annex 3 :Inflation regression results
Annex 4 : Consumption regression results
1
Annex 5: Output growth scenarios under alternative methodologies (ARMA process versus mechanical rule)
BE/ AT
DE
/ DK
EL
/ ES
FI
/ FR
IE
/ IT
LU
/ NL
PT
/ SE
UK
/ BG
CY
/ CZ
EE
/ HU
LT
/ LV
MT
/ PL
RO
/ SI
SK
/ HR
EA12
/ EU29
1
Annex 6: Unemployment rate forecasts under alternative methodologies (Kalman filter forecast of the unemployment gap versus mechanical rule)
Annex 7: Comparison with the IMF WEO 2015
Output growth (percentage change in GDP in constant prices)
Unemployment rates, in %
Inflation (end of the year prices; percentage change)
Investment growth in %
1
[1] The EU average output gap is calculated by dividing the sum of GDP over the countries by the sum of potential output over the countries (). The exchange rates are taken from the Ameco storage of the Autumn forecast.