El riesgo como paradigma referencial en el estudio de los instrumentos de valoración y gestión del riesgo
CUADRÍCULA
1.2. La tesis gubernamental
Table 14 presents the estimates of the cyclical safety margin for each Member State (column 2) which were prepared by the Commission for the first set of stability and convergence programmes. They were obtained by multiplying the budgetary sensitivity to the cycle (in col- umn 1) with an output gap estimate which encapsulates the size and frequency of cyclical fluctuations in output for each Member State. Naturally, the higher the sensitiv- ity of the budget to the cycle and the higher the volatility of the economy, the higher the estimated safety margin is. Estimates of the budgetary sensitivity to the cycle are arrived at by measuring the impact of a rise/fall in GDP on public expenditures and revenues. For the euro area, it has been estimated that a 1 % fall in GDP relative to trend will increase the deficit by around 0.5 percentage points of GDP (see Chapter 1 in Part VI). As to the largest negative output gap which a Member State is likely to encounter, the Commission calculations took the mid-point of two worst output gaps of the following estimates: (a) the largest negative output gap recorded in each Member State between 1960 and 1997; (b) the unweighted average of the largest negative output gaps in EU Member States over the period 1960–97, which is estimated to be to 4 % of trend GDP; and (c) the average volatility of the output gap in each Member State, as measured by two times its standard deviation (1).
The difference between the 3 % reference value and the estimated cyclical safety margin is the so-called country’s ‘minimal benchmark’. Overall, these estimates of the cyclical safety margin show that Belgium, Denmark,
51 (1) When output gaps are normally distributed, around 95 % of the
observations fall within the range of two times the standard devia- tion around the mean. Thus, only 2.5 % of the observations fall outside this range in the case of negative output gaps.
Spain, Ireland, Luxembourg, the Netherlands, Portugal and the United Kingdom should aim for a structural deficit of between 0 % and 1 % of GDP. Germany, Greece, France, Italy and Austria could aim for a deficit target even slightly above 1 % of GDP. Sweden and Finland would have to aim for a surplus as their budgets have a high sensitivity to the cycle and their economies have in the past shown a high degree of volatility. These minimal benchmarks are largely in line with those computed in other studies (see Box 8).
As mentioned above, these calculations were originally performed in 1998–99. Two relevant developments have taken place since the estimates were first made.
• The introduction of ESA95 has changed the historical GDP time path (see Chapter 2 in Part VI). In general, GDP levels are higher, but the changes smoother. This does not have a major impact on the estimated maximum negative output gaps, with the exception of Greece where such a variable worsens somewhat. • The cyclical sensitivity of the budget has been recal- culated recently on the basis of new provisional OECD estimates (OECD, 1999a). The new budget sensitivities are shown in the third column of Table 14.
While the average euro area sensitivity does not change appreciably, there are some noticeable differ- ences at country level. For instance, a substantial fall occurs for Austria, Spain, Portugal and the UK, and, to a minor extent, France, Italy, Ireland and Sweden. This reduction may reflect the reforms in the past decades which trimmed the generosity of the welfare state and lowered the progressivity of tax systems (Part IV). The only countries where the budget elas- ticity appears to have increased significantly are Denmark and the Netherlands.
New estimates of the cyclical safety margins are pre- sented in the last column of Table 14 (1). The cyclical
safety margin remains broadly unchanged in eight coun-
52
P u b l i c f i n a n c e s i n E M U - 2 0 0 0
Table 14
Estimates of the cyclical safety margins
Former estimates New estimates
Budgetary Cyclical safety Budgetary Cyclical safety sensitivity margin sensitivity margin
1997 (% GDP) 2000 (% GDP) B 0.60 2.0 0.65 2.2 D 0.50 1.9 0.50 1.9 E 0.65 2.6 0.40 1.6 F 0.55 1.5 0.45 1.4 IRL 0.55 2.1 0.40 1.8 I 0.50 1.8 0.40 1.4 L 0.60 3.0 0.60 2.9 NL 0.75 2.9 0.85 2.9 A 0.50 1.7 0.30 1.0 P 0.50 2.4 0.30 1.5 FIN 0.65 4.3 0.65 3.4 EUR-11 0.55 2.0 0.50 1.8 DK 0.70 2.3 0.85 3.2 EL 0.40 1.6 0.35 1.6 S 0.90 3.8 0.80 2.6 UK 0.70 2.9 0.45 1.9 EU-15 0.60 2.2 0.50 1.9
Source: Commission services.
(1) Given the provisional character of the new budgetary elasticities,
these estimates are to be considered highly tentative at this stage. In addition to elements above, the estimates also take into account the fact that some of the largest negative output gaps in the origi- nal estimates have occurred in periods of severe recession, which, according to the SGP rules, could have triggered the application of the exceptionality clause allowing countries to exceed the 3 % limit (section 2.2). Adjusting the calculation for events where real GDP growth was lower than – 0.75 % has a non-negligible effect only in Finland and Sweden. This is of course not surprising given the economic downturn these countries experienced in 1992–93.
53 (1) For a summary of the arguments, see European Commission (1998).
P a r t I I I
T h e S t a b i l i t y a n d G r o w t h P a c t o n e y e a r o n
Using a similar approach to that presented above, IMF (1998) and OECD (1997a) find that a structural deficit in the range of 0.5 % to 1.5 % of GDP and below 1.5 % of GDP, respectively, would be enough to allow the auto- matic stabilisers to operate without breaching the 3 % of GDP deficit threshold even in periods of pronounced cyclical slowdown. Similar conclusions were obtained by applying more sophisticated methodologies. Dalsgaard and de Serres (1999), in the context of an estimated struc- tural VAR (vector autoregressive model), show that for a majority of EU countries a structural deficit between 1 % and 1.5 % of GDP would help to avoid breaching the 3 % of GDP threshold with a 90 % certainty over a three-year horizon. If governments aimed for a structural position between zero and 1 % of GDP, the confidence horizon was extended to between five and seven years.
Stochastic simulations on the NiGEM model of the National Institute for Economic and Social Research show
that, for the five large EU economies, the probability of breaching the 3 % of GDP threshold is still high at present for France and Italy, but drops significantly by 2002 (see Barrell et al., 1999). In a later paper, also based on stochastic simulation of the NiGEM model, Dury and Pina (2000) indicate very low probabilities of collision with the sanctions procedures of the Pact, a result that is robust to variation in assumptions about the monetary strategy pur- sued by the ECB. Similarly, Barrell and Pina (2000)’s esti- mates of the automatic stabilisers by applying methods of stochastic simulation are generally lower than normally assumed. Their study confirms that, if the countries adhere to the budgetary targets laid down in their stability and convergence programmes, the full working of built-in stabilisers and the respect of the 3 % deficit ceiling are expected to be compatible.
Box 8: Estimates of the cyclical safety margin: other studies
tries (Belgium, Germany, Greece, France, Ireland, Italy, Luxembourg and the Netherlands), while it goes up in Denmark. In all other countries, the cyclical safety mar- gin is reduced somewhat reflecting a lower budgetary sensitivity and/or smoother cyclical fluctuations. These estimates have to be treated with caution. In par- ticular, factors influencing the cyclical fluctuations may change in the new EMU framework. A number of argu- ments point towards a higher synchronisation of business cycles in the future, thereby raising the stabilisation role of the single monetary policy (1). However, it is too soon
to verify whether the pattern of cyclical fluctuations will change in EMU. If evidence of a changing cyclical pattern emerges in EMU, these estimates will need to be revised.