Opinion of the Advocate-General
Mr President,
Members of the Court,
1 . In these proceedings the Commission seeks the annulment of Council Regulation ( EEC, Euratom, ECSC ) No 3619 of 26 November 1986 ( 1 ) correcting the weightings applicable in Denmark, Germany, Greece, France, Ireland, Italy, the Netherlands and the United Kingdom of the remuneration and pensions of officials and other servants of the European Communities, in line with the rise in the cost of living recorded between 1 January 1976 and 31 December 1980 .
2 . The importance of the judgment to be given in these proceedings may readily be appreciated by reference to the number of actions brought by officials against the institutions for misapplication of the weightings during the period in question, which have been suspended by order of the Court pending the outcome of the present action . ( 2 )
3 . In the following analysis I shall first outline the legal background to this action, then I shall summarize the history of the dispute and finally I shall consider the merits of the arguments put forward by the parties .
I - The legal background
4 . The principle of equivalence of purchasing power for all officials and other employees of the Communities, regardless of their places of employment and the currencies in which they are paid, has as its legal basis Articles 63, 64 and 65 of the Staff Regulations of Officials, which are applicable to temporary and auxiliary staff by virtue of Articles 20 and 64 of the Conditions of Employment of Other Servants, and has been repeatedly upheld in the decisions of this Court . ( 3 )
5 . The first paragraph of Article 63 provides that officials' remuneration is to be expressed in Belgian francs and paid in the currency of the country in which they perform their duties .
6 . In order to ensure equivalent purchasing power, Article 64 provides that officials' remuneration is to be weighted at a rate above, below or equal to 100%, depending on living conditions in the various places of employment . The weightings are determined by the Council, on a proposal from the Commission, and the weighting for officials employed in Belgium and Luxembourg is to be 100 %.
7 . The weightings for the other countries are reviewed and adjusted periodically according to changes in the cost of living in the various Member States .
8 . Since 1981 the detailed arrangements for that adjustment and the review of the remuneration of officials and other servants have been governed by the Council Decision ( 81/1061/Euratom, ECSC, EEC ) of 15 December 1981 (" the 1981 Decision ") amending the method of adjusting the remuneration of officials and other servants of the Communities, ( 4 ) which had applied since 26 June 1976 .
9 . For the purposes of the annual review of remuneration ( Article 65 ( 1 ) ), provision is made for an annual adjustment of the weightings to be carried out by the Council on a proposal from the Commission, based on data provided by the national statistical departments and taking into account a joint index showing price increases, prepared by the Statistical Office of the European Communities ( SOEC ) - 1981 Decision, Annex, Point II.1 and the last indent of Point II.4 ( c ).
10 . Also, in the event of substantial changes in the cost of living, intermediate adjustments are to be made ( Article 65 ( 2 ) ), which are to be taken into account in the annual adjustments .
11 . In addition, the principle of periodical reviews of the weightings was adopted in 1976, to determine whether they correctly reflect the equivalence of purchasing power of remuneration in the various places of employment .
12 . The 1981 Decision provided that the review was to take place every five years : the SOEC will then verify, in agreement with the statistical departments of the Member States, whether the ratios between weightings accurately reflect purchasing power equivalences between salaries paid to staff serving in the capitals of the Member States and in other places of employment, when it appears that there is a danger of considerable distortion ( 1981 Decision, Annex, Point II, 1.1, second and third paragraphs ). On the basis of the results of that verification, the Commission is to submit a proposal to the Council for a five-yearly adjustment of the weightings .
13 . Under the conditions laid down in the said Point II, 1.1 of the Annex to the 1981 Decision and on the basis of Articles 64 and 82 of the Staff Regulations and Articles 20 and 64 of the Conditions of Employment of Other Servants, on 26 November 1986 the Council adopted the contested regulation, correcting the weightings applicable to the various Member States . ( 5 )
II - The background to the dispute
14 . The process which has given rise to the present proceedings commenced in 1981 and involved four important stages, from which the reasons for the differences between the parties are apparent : the first proposal from the Commission; the withdrawal of that proposal, in view of the objections made within the Council; the Commission' s second proposal; and the adoption by the Council of the contested regulation .
15 . Let us recall the circumstances surrounding the Commission' s initial proposal . With a view to the review of the weightings which was to take place in January 1981, the Commission arranged for surveys to be carried out in 1980 and 1981 in order to determine the consumer habits of European officials . According to the Commission, the surveys were carried out in accordance both with the method for reviewing remuneration approved in 1976 and with the new method adopted in 1981 .
16 . The surveys were carried out by the national statistical departments in collaboration with the SOEC and took account, for all the factors concerned except accommodation, of the prices being charged in the capital cities .
17 . In the case of accommodation, instead of the rents paid by European officials in capital cities, account was taken of the average rents paid in each Member State by the population as a whole, as shown in their national accounts . That method had already been adopted for previous reviews .
18 . However, in the Commission' s view, the outright application of that method would have produced incorrect results : whilst it was found that the results reasonably reflected the situation when the weightings were calculated without taking into account the accommodation item or where the building price-index for each country as a whole was considered, that was not the case when the figures for rents taken from national accounts were included .
19 . Having regard to that contradiction and notwithstanding the doubts which those calculations prompted, the Commission decided that it could not leave out the accommodation item, in view of its importance as part of officials' outgoings ( around 20 %).
20 . In order to minimize the consequences of that anomaly, in its proposal of 17 July 1984 - COM(64 ) 274 - the Commission, using the same values as before, suggested the application of a threshold of 2.5%, within which the variations recorded in purchasing-power equivalences would be ignored, and the weightings would be altered only where there was an upward or downward change in excess of that percentage .
21 . When the Commission' s proposal was discussed by the Council Working Party on Staff Regulations on 8 and 9 October 1984, two delegations raised doubts as to the compatibility of the proposal with the Staff Regulations and with the methods for reviewing remuneration adopted in 1976 and 1981 .
22 . The Council' s legal department was consulted and expressed the view that the application of a threshold was contrary to Article 64 of the Staff Regulations, since officials were entitled to rely on account being taken of any change, even if minimal, in living conditions disclosed by the periodical review .
23 . On the basis of that opinion the Working Party on Staff Regulations reached the conclusion ( at its meeting of 19 November 1984 ) that the introduction of a threshold was unacceptable .
24 . The Commission accepted that view, withdrew its proposal and endeavoured to find more reliable data in respect of accommodation which would obviate the need for the proposed threshold .
25 . Accordingly, on 23 December 1985 the Commission put forward a second proposal . In its preparation, instead of using the national averages for house rents, the Commission arranged for a survey to be carried out as to the rents paid in the capital cities for a number of typical kinds of accommodation . For that survey, estate agencies were approached by the national statistical departments in late 1984 and early 1985 . The results of the survey were processed by the SOEC, which extrapolated them to 1 January 1981; the Commission was then able to conclude that, on that occasion, there was a satisfactory parallelism between the weightings thus determined and those obtained without account being taken of the accommodation item or with the substitution of building costs for rent levels . Consequently, it proposed new weightings, specifying 1 January 1981 as the date on which they were to take effect .
26 . On 26 November 1986 the Council adopted the regulation now being challenged, but made far-reaching amendments to the Commission' s proposals . The changes were in two areas : firstly, the method used for calculating the accommodation item was not that used for the Commission' s most recent proposal but the method used for the previous one, based on national average house rents obtained from national accounting data; secondly, the date on which the review was to take effect was fixed as 1 July 1986 and not 1 January 1981 .
27 . In numerical terms, the difference between the Commission' s proposal and the Council regulation regarding adjustment of the weightings is as follows :
Commission proposal Council Regulation
Denmark + 12.6 + 13.2
Germany + 0.5 + 4.9
Greece - 3.0 - 3.3
France + 12.1 + 1
Ireland + 18.1 + 10.1
Italy + 2.1 - 7.8
Netherlands + 0.1 - 1.4
United Kingdom + 16.8 + 3.0
Varèse + 2.1 - 7.8
28 . It should be noted that, whereas the Commission proposal would have involved a growth in expenditure of ECU 2.8 million for a six-year budget implementation period, the contested Council regulation led, for the same period, to a reduction of ECU 1.9 million .
III - Analysis of the parties' arguments
29 . The Commission maintains that in the two areas in which the regulation adopted departed from its proposal, it infringed both Article 64 of the Staff Regulations and the 1981 Council Decision . On the one hand, in the Commission' s view the values used in calculating the accommodation item as part of the cost of living do not reflect the real costs involved, and they do not enable the levels of remuneration required to ensure the equivalence of officials' purchasing power to be determined with sufficient precision; on the other hand, the Commission also considers that the five-yearly review of the weightings should take effect as from the end of the previous period .
30 . I shall deal separately with each of those aspects, comparing the parties' arguments .
1 . Calculation of the accommodation item
31 . The Council' s use of the national average rents paid by the population as a whole and not the rents paid in the capital cities for standard types of accommodation by European officials is challenged by the applicant on three counts : firstly, that approach is incompatible with the method adopted by the 1981 Decision, which requires account to be taken of the situation obtaining in the capital cities; secondly, it does not take account, with sufficient precision, of the specific situation of the officials concerned; and finally, the grounds on which the Council relied in rejecting the data prepared by the SOEC to determine rent levels are invalid .
1.1 . The obligation to take account of the position in the capital cities of the Member States
32 . According to the Commission, the application of the method adopted in 1981 requires reference to capital cities since the purpose of the five-yearly review is to ensure that the geographical weightings are correctly determined : those weightings were established with effect from 1968 and 1969 by reference to prices in capital cities, and were only determined for Ispra following the judgments of the Court of 15 December 1982 ( which is why Ispra was included in the Commission' s proposal ). In those circumstances, the Council regulation had to take account of the situation in the capital cities and obviously it could not do so by using overall averages obtained from the national accounts .
33 . The Commission also maintains that the 1981 Decision was one of a set of exceptional measures involving sacrifices on the part of officials ( in particular the introduction of an exceptional crisis levy, provided for in the new Article 66A of the Staff Regulations ) and for that reason there must be strict observance of the resultant balance and the rules on which it is based, in particular the provision of the Council decision which requires the capital cities to be taken as the reference point for determining the weightings ( Point II, 1.1, second paragraph ).
34 . The Council puts forward a number of arguments in reply to those points . Firstly, it points out that the parameter "national average rents" had already been used in previous reviews ( most recently in 1976 ) and was used in the Commission' s first proposal for 1981, submitted in July 1984 . Between that date and December 1985 ( the date of the second proposal ) no objective factor had emerged to justify a change in the method used . Secondly, the Commission had never submitted to it a proposal based on average rents in the capital cities, which would have made it possible to achieve more accurate results without changing the method, but had chosen to change the method by recourse to a survey based on information from estate agencies . As regards the set of exceptional measures, the Council observes that the "crisis levy" is not taken into account in relation to the other elements of the annual review and still less for the purposes of the five-yearly reviews .
35 . In its reply, the Commission draws attention to a number of points .
36 . Firstly, it did not make any change to the method, but merely adopted the approach followed in calculating the other cost-of-living items ( direct collection of data in respect of 800 products and services in the capital cities ).
37 . Secondly, the abandonment of the "traditional" factors in the calculation for which the Council criticized it was irrelevant from the legal standpoint : its recourse to data from the national accounts was an exception which could not be regarded as establishing a "tradition" and the change in the Commission' s practice regarding accommodation had become necessary for objective reasons, relating to the SOEC' s finding that that method did not produce reliable results .
38 . Finally, the requirement under the 1981 Decision that the five-yearly review should relate to the capital cities, and also the guarantee that that method would be adhered to for 10 years, and the official recognition of that fact by publication in the Official Journal constituted the quid pro quo secured by officials in return for the imposition of the "crisis levy ".
39 . For its part, the Council ultimately conceded, in its rejoinder, that it did not challenge the principle of surveys in the capital cities but objected to the approach followed in carrying them out, which it considered inappropriate; in particular, it considered that the surface area of typical accommodation was not equivalent in all the capital cities and that in certain capitals European officials resided further from the centre or in smaller houses, for which reason the proper course would have been to approach them for the purposes of the survey .
40 . Let us consider those contentions .
41 . It seems to me to be clear that both Article 64 of the Staff Regulations and the second paragraph of Point II, 1.1 of the 1981 Decision indicate that the cost-of-living factors should express, as accurately as possible, the situation prevailing in the various places to which officials are assigned .
42 . Firstly, Article 64 requires officials' remuneration to be subjected to a weighting "depending on living conditions in the various places of employment ".
43 . Secondly, the 1981 Decision makes it clear that the central feature of the five-yearly review of the weightings is to be an examination of "purchasing-power equivalences between salaries paid to staff serving in the capitals of the Member States ".
44 . It is true that there are officials who work in the capitals but live outside them and indeed there are others whose duties are performed elsewhere than in the capitals ( as in the case of Ispra ).
45 . For that reason, the third paragraph of Point II, 1.1 of the Annex to the 1981 Decision provides that where objective factors suggest that there is a danger of considerable distortion in relation to the data recorded in the capital a check will be carried out for other places of employment .
46 . That check, which is necessary - as made clear in the two judgments of 15 December 1982 - to establish weightings different from those applied in the capital cities, must also give rise, where appropriate, to a review of the weightings, in particular the five-yearly review .
47 . In any event, the objective pursued by the Staff Regulations is to ensure that all the officials enjoy the same purchasing power whatever their place of employment, in accordance with the principle of equality of treatment, as the Court has held with respect to Article 65 ( 2 ). ( 6 )
48 . That is also why, in relation to Article 64, the Court pointed out that what was to be determined was the "exact places where the duties of a sufficiently large number of officials and other employees of the Communities are performed" ( capital cities or otherwise ). ( 7 )
49 . The provisions adopted by the Council fixing the ways and means of adjusting the weightings, as provided for in the Staff Regulations, may not disregard those principles and objectives .
50 . That is the essential background against which must be interpreted the 1981 Decision, by which the Council bound itself regarding the basis for regulating the exercise of the discretionary power conferred on it by the Staff Regulations . ( 8 )
51 . In those circumstances it must be recognized that recourse to national averages as a means of determining the factors on which is to depend the equivalence of officials' purchasing power at their various places of employment does not enable the prescribed objectives to be achieved, in so far as a distorted view of the actual position is given as a result of the calculation methods used .
52 . Only if it should prove impossible - or at least extremely difficult - to obtain the necessary information regarding each place of employment might it possibly be acceptable to use more approximate values, including averages . But I shall consider this matter in due course .
53 . In the mean time it is appropriate to deal with another question raised by the Council' s arguments . Could the Commission validly present a proposal which entailed a change in the method which it had itself used previously ( in its first proposal and in the earlier periodical adjustments ) in order to calculate the accommodation item, and in fact the method on which the contested regulation was ultimately based?
54 . In my opinion, having identified the inadequacies of the results arrived at by the first method, the Commission was under a duty to use the statistical method which proved most accurate .
55 . When presenting the first proposal, the Commission acknowledged the difficulties inherent in the method adopted to calculate the accommodation component of the cost of living and therefore proposed the introduction of a threshold to trigger adjustments . Clarification is provided in that connection by the explanatory memorandum of 24 March 1986 relating to the proposal of 23 December 1985, which refers to the serious defects of the method .
56 . Similarly, the rejection of that proposal by the Council Working Party on Staff Regulations was based on an opinion from its legal department, to the effect that it was "clear that the method employed cannot be regarded as entirely reliable, in view of the margin of error for which the Commission seeks to allow by introducing a threshold ".
57 . Since at the same time - a fact not overlooked in that opinion - the adoption of a 2.5% threshold represented an arbitrary choice not based on specific information, the Commission had no alternative but to re-set its sights and change the method used, in a manner which ran counter to the concern ( also, quite properly, emphasized in the opinion ) to use a reliable method, in order to obtain results which reflected the real situation as closely as possible, so that the new weightings would be truly representative .
58 . The foregoing considerations in themselves point to the illegality of the contested regulation, by virtue of its being based on national average house rents, and it is therefore unnecessary to analyse the question of the alleged interdependence between the review of the weightings and the introduction of the crisis levy .
1.2 . The obligation to take account of the specific situation of officials
59 . The text of Article 64, taken in isolation, does not appear to require account to be taken of the specific living conditions of officials or of their typical consumer habits in order to establish the weightings for their remuneration .
60 . That provision merely refers to the application of different weightings "depending on living conditions in the various places of employment ".
61 . That does not seem necessarily to indicate that it is illegal to use average values for the rents paid by the population as a whole in the various places where officials live, in capital cities in particular .
62 . However, we have just seen that - as made clear in previous decisions of the Court - the purpose of determining and adjusting the weightings is to ensure the equivalence of the purchasing power of the remuneration received by officials, whatever their place of employment .
63 . That was what the 1981 Decision quite properly clarified, in so far as it established that the five-yearly review of weightings is designed to ensure that the ratios between them "accurately reflect purchasing-power equivalences between salaries paid to staff serving in the capitals of the Member States" and, therefore, to staff serving elsewhere than in the capitals .
64 . In order to attain that objective, it does not seem inappropriate to take account of the expenses for which such remuneration is used and therefore, as nearly as possible, of the ordinary consumer habits prevailing among European officials as a whole; against that background it thus appears justified to take account, as was done in the Commission' s proposal, of the type of accommodation occupied by those officials .
65 . In their family budgets, moreover, accommodation accounts for a significant percentage ( around 20%, according to the Commission ) and is therefore an important factor in any comparison of purchasing power .
66 . It is true that, as the Commission stated, the calculation method used for determining the weightings was based on what is known as "the Fischer method", which presupposes that all officials are likely to have the same standard of living as a result of adapting themselves to the way of life available to them in the various places of employment .
67 . In its defence, the Council took that fact as a basis for criticizing the Commission' s position, asserting that the type of accommodation occupied by officials is an improper reference point for the calculation in so far as accommodation requirements differ in different regions ( for example, in a Mediterranean country as opposed to one in northern Europe ), and officials serving in each locality must adapt their habits - including those relating to accommodation - to the way of life obtaining in each region .
68 . The Council' s objection does not appear well founded .
69 . As the Commission explained in its reply, the Fischer method provides intermediate results, representing a compromise between two possible extreme choices : as I understand it, the range of choices extends from the absolute generalization of a way of life characteristic of a particular place to consideration of all the features peculiar to the way of life in each place of residence . ( 9 ) In fact, for a calculation of equivalences of purchasing power to be possible, it is necessary for the cost-of-living components ( in this case, housing ) to be based on identical or comparable products or services, having sufficiently general characteristics to be found in all the places being compared . Accordingly it must be acknowledged - as a basis of comparison - that a particular standard of living can be maintained in the various places where officials are employed, and for that purpose the accommodation must of course possess the characteristics dictated by the specific features of each region . It is on the basis of the costs determined in each region for accommodation capable of providing the same kind of comfort that purchasing power can then be compared .
70 . Moreover, the terms in which the SOEC made inquiries of the national statistical departments described the type of accommodation in question with sufficient accuracy, in a manner which seems consonant with the objectives of the requisite comparison of the purchasing power of remuneration at the various places of employment .
71 . In any event - whatever the pertinence of the Council' s criticisms regarding the method used by the Commission - the national averages used as a basis for the contested regulation could not give a true picture of officials' circumstances .
1.3 . Analysis of the Council' s reasons for rejecting the Commission' s proposal
72 . The second-from-last recital in the preamble to the contested regulation states that "the results of the survey do not appear to be acceptable since it failed, in particular, to deal with a truly representative sample of accommodation; ... furthermore, the survey should have been conducted pursuant to the second paragraph of Point II, 1.1 of the Annex to Decision 81/1061/Euratom, ECSC, EEC, in agreement with the national statistical departments ".
1.3.1 . The unrepresentative nature of the sample
73 . As regards the first aspect, the Commission pointed out in its application firstly that, in view of the objections made within the Council ( 10 ) to the method which it had used at the outset, it had no alternative but to determine the level of rents by reference to the types of accommodation occupied by officials, giving preference in its choice to the criterion of comparability; then it stated that, as already emphasised, that type of accommodation was described in detail in the SOEC' s survey; that after the elimination of the factors giving rise to a degree of uncertainty the Council could not reject the comparisons proposed by the Commission, except on the ground of manifest error and no such error was proved, since the criticisms made of the method are general in scope, and capable of embracing any kind of sampling; and finally that the documents submitted at successive meetings of the Working Party on Staff Regulations demonstrated the reliability of the results obtained by the SOEC, which agreed with those arrived at by total exclusion of the accommodation item or inclusion of the building costs, by contrast with the anomalies resulting from the use of the data obtained from national accounts ( as occurred in the case of London, in respect of which the latter method gave discrepancies of the order of 10 percentage points ).
74 . In support of its allegation that the sample was unrepresentative, the Council first stated that, by contrast with the position in other capitals, the cost of accommodation in Brussels and Luxembourg was determined by reference to the rents actually paid by the officials employed there .
75 . Moreover, the surveys based on information from estate agencies in the other capitals related only, the Council contends, to the types of rented accommodation located in city centres . Those surveys did not take account, therefore, either of the basis of occupation of the accommodation ( owners or tenants ) or of its true situation ( only a small percentage of officials live in city centres, in particular in Brussels ) or of the costs actually borne by officials . In the Council' s view, the situation in London confirms those allegations .
76 . As regards the consistency of the results obtained by the various methods of dealing with the accommodation item, ( 11 ) the Council, although conceding that, for the United Kingdom, Ireland and Italy, those results confirm the Commission' s arguments, contends that that does not occur in the other cases . Thus, in three Member States ( Denmark, the Netherlands and Greece ) the values in the regulation are close to those in the Commission proposal; in three ( Germany, France and Greece ) the values in the regulation are close to those obtained without the accommodation item being included . In other words, in the Council' s view, in the case of five States out of eight, an analysis of the figures does not confirm the Commission' s allegations that the results of the method adopted in the regulation are inconsistent or that the results of the method advocated by the Commission are reliable .
77 . The explanations given by the Commission in its reply appear, however, sufficient to rebut most of the charges made by the Council against the method on which the Commission' s proposal was based .
78 . In the first place, the Commission states that, for the purposes of the five-yearly review, the level of rents paid in Brussels was also the subject of surveys centred on estate agencies; surveys using information from the owners themselves were indeed carried out, but their purpose was to calculate a common index enabling the equivalences of purchasing power to be brought up to date .
79 . In the second place, as explained by the Commission and confirmed by the letters sent by the SOEC to the national authorities containing the terms of the survey to be based on information from estate agencies ( see annexes 1 and 2 to the Commission' s reply to the written questions put to it by the Court ), that survey did not necessarily relate to dwellings situated in city centres but rather to "middle/upper-class residential areas", which may involve the peripheral areas just as well as areas in the centre .
80 . The Commission also states that the calculations made by the SOEC took account of the basis of occupation of the accommodation - owner-occupation or tenancies - by using an "owners weighting" and similarly a "tenants weighting" and recourse to the calculation of notional rents, by allocating the same parity to lettings and to purchases, a method commonly used in calculations for the purposes of national accounts .
81 . Finally, in response to the Council' s comments on the consistency of the results, the Commission accuses the defendant of confusing the method of calculating the accommodation item with the method by which the reliability of the results thereby obtained is tested . The applicant points out that it never stated that, in a given country, the values compared - prices excluding accommodation, prices including accommodation ( in capitals or in general ) and construction costs - never coincided or always coincided . Naturally, the freer the market and the more limited the public interventions or sudden increases in population density, the greater will be the tendency for the different values to be closer to each other; by contrast, restrictions on building, public interventions and sudden increases in population density could lead to divergent trends in the figures in question .
82 . Accordingly, when the reliability of the figures is verified the variations must be kept within reasonable limits, intervention being justified only if those limits are exceeded to the point of contradicting the most basic information available from ordinary experience .
83 . According to the Commission, that is what happened when the results taken from national accounts for each country as a whole were used . Two sets of data - prices excluding accommodation and building price indexes - were in line with each other whilst at the same time, in certain cases, there was a significant divergence from the prices including accommodation, distorting the other sets of data .
84 . As the object of the exercise was to check whether the ratios between the weightings correctly reflected purchasing power equivalences, the Commission focused its attention on those cases in which the cost of accommodation was not in line with the cost of living in general . In particular, it was found that in four member States the indexes which included rent costs were significantly lower than those which did not include them or which used the buildings costs, a result that conflicted with the data based on actual experience relating to the relative development of prices in the various sectors .
85 . In preparing its second proposal, the Commission concluded that the cost components which had previously given rise to problems had become reliable, and the remaining data had not become any less reliable . In the case of London, for example, the Commission emphasizes that the results show that whilst the "agency rent" is not "abnormally high" the cost of "accommodation according to the national accounts" is, without doubt, "abnormally low ".
86 . The Commission' s explanations seem to me to be convincing, in so far as :
( a ) the calculation method based on national accounts led to distorted and unreliable results . In that context, the Council' s statement that the Commission had introduced the threshold of 2.5% in order to avoid a fall in the weightings arrived at by the use of that method seems to be absolutely without foundation, whereas, by contrast, the Commission' s statements as to the reliability of the method appear reasonable;
( b ) it cannot be said that the method used as a basis for the Commission' s second proposal is vitiated by all the technical errors attributed to it by the Council, since the surveys were organized in what appears to be a serious and well-considered manner . In any event, the Council' s criticisms are not supported by any hard evidence to show decisively that the surveys are not sound;
( c ) the method followed by the Commission in preparing its second proposal - although not the ideal method - gave results which, notwithstanding any objections which its application might provoke, better reflected the real position than the method based on the overall figures taken from national accounts .
87 . It must also be pointed out that the arguments put forward by the parties touch upon the delicate problem of the relative powers of the SOEC, the Commission and the Council, a matter with which I shall deal in due course .
88 . Before doing so, I shall consider the Council' s second reason for rejecting the Commission' s proposal .
1.3.2 . The absence of agreement between the SOEC and the national statistical departments
89 . The 1981 Decision ( Annex, II, 1.1, second paragraph ) provides - as does Article 65 ( 1 ) of the Staff Regulations with respect to the preparation of a joint index - that the five-yearly verification is to be carried out "in agreement with the statistical departments of the Member States ".
90 . However, according to the preamble to the contested regulation, the Council considers that condition not to have been complied with by the Commission in preparing its second proposal .
91 . In fact, as the Commission itself recognizes, one of the delegations making up the working party of experts which assists the SOEC in its work on the cost of living ( the British delegation, to be precise ) did not accept the method proposed or the results obtained using it .
92 . Nevertheless, for the reasons which the Commission set out in its application, the SOEC considered itself bound to recommend that the requisite agreement should be dispensed with and proposed that the Commission should adopt the values arrived at as a result of its survey - which the Commission did in its second proposal .
93 . In its defence, the Council went so far as to contend that neither the method nor the results of the survey entrusted to the SOEC by the Commission were ever submitted to the competent working party of statistical experts .
94 . That allegation is not, however, confirmed by the information before the Court, which shows that details of the survey and its results were submitted for consideration by the Working Party on Price Statistics .
95 . Is the objection expressed by one of the delegations on that occasion, therefore, sufficient to undermine the proposal which took no account of it?
96 . In that regard, the Commission maintains that its proposal was prepared in accordance with the customary procedure of collaboration with the statistical departments of the Member States, that the data were compiled through the intermediary of those departments ( an expert from each national statistical department is even said to have accompanied the SOEC representatives at all times during the survey ) and that the lack of agreement of one member of the Working Party on Price Statistics regarding one aspect of the survey cannot block a procedure intended to give effect to a provision of the Staff Regulations .
97 . Accordingly, having before it a proposal from the Commission, the Council could not fail to take account of the SOEC' s conclusions on which that proposal was based ( although there was an objection from one of the delegations to the Working Party on Price Statistics ) and was not entitled to replace the figures arrived at using an accurate method by figures which were manifestly inappropriate .
98 . Specifically, the discussion touches once more upon the relative powers of the two institutions and the limits of the Council' s discretionary power .
99 . How should this question be resolved?
1.4 . Conclusions concerning the first part of the application
100 . The technical nature of the problems arising in these proceedings and also a degree of opacity in the arguments exchanged by the parties rendered necessary, for reasons of clarity, the preceding digression concerning the various aspects of the applicant' s first submission .
101 . In fact, the applicant and the defendant took care to compare their own versions of the other party' s arguments, making it more difficult to clarify their positions and indeed to determine the facts .
102 . In addition, the essential points of law raised by the application appear at times inadequately identified and cloaked with imprecisions which hamper analysis .
103 . My earlier examination has however enabled me to make a number of partial judgments capable of serving as a basis for the solutions which it is incumbent on me to propose :
( a ) in the terms of Article 64 of the Staff Regulations and by virtue of previous decisions of the Court - and also of the Council' s 1981 Decision, which established the procedures for applying that provision - compliance with the principle of equivalent purchasing power for the remuneration of Community officials calls for determination, with the greatest possible precision, of the cost-of-living components which specifically affect the situation of those officials in their various places of employment;
( b ) the average overall figures for house rents obtained from the national accounts are influenced by factors which have nothing to do with the specific localities to which officials are assigned, and therefore recourse to those figures does not enable an important component of the cost of living of Community officials to be determined with sufficient precision to ensure the equivalence of the purchasing power of their remuneration;
( c ) in submitting its second proposal for review of the weightings, the Commission sought to correct the result at which it had arrived, with respect to one of the cost-of-living components, using a method which, being rather unreliable, was immediately challenged within the Council; the new method followed by the SOEC, the results of which were accepted by the Commission, represents, notwithstanding any problems of a technical nature, a better reflection of the real position than is available from average rents derived from national accounts, and therefore it should not be rejected out of hand as inappropriate or unreliable .
104 . It will of course be said that the ideal method, the one which would enable the aims of the five-yearly review to be achieved in full, would have been a survey based on inquiries addressed to the officials themselves, so that their actual accommodation expenses could be determined .
105 . I doubt whether that is the case .
106 . It is true that the fact that a direct survey was carried out for officials residing in Brussels, Luxembourg and Varèse - where they are more numerous - would appear to indicate that the same approach would be feasible in the other capital cities .
107 . However, the Commission explained that that method would not be statistically meaningful, since very few officials live in those capitals and thus no sufficiently representative sample exists .
108 . That argument is credible, since it would be improper if the adjustment factors to be applied to the weightings were heavily influenced by individual decisions or situations which - in a relatively restricted environment - varied significantly from the most usual standard of officials' accommodation .
109 . Accordingly, from that point of view, the recourse to "typical accommodation" might seem to be a method less susceptible to variances . Moreover, it appears to be the method used for their officials by other international organizations ( the OECD and the World Bank ) which are well known for the competence of their statistical departments and the accuracy of their economic surveys .
110 . Furthermore, there is another factor which would certainly have rendered it more difficult to make direct enquiries of officials when it became necessary to prepare a second Commission proposal : I refer to the time which had already elapsed since the end of the reference period ( four years ), as a consequence of which the results of a survey on that basis, and indeed the possibility of carrying it out, would have been more uncertain .
111 . In any event, if that method were used it would certainly be necessary to take the technical measures necessary to neutralize the effects of the variances to which I referred . One of the instruments which could be used to correct the data obtained from a direct survey would then, in all probability, be a survey involving estate agencies ...
112 . Admittedly, it cannot be stated with certainty that the method adopted by the Commission is wholly beyond criticism or, still less, that it is an ideal method . And, in particular, there is no denying that it was challenged by United Kingdom statistical experts .
113 . For that reason, in this highly technical area, cooperation between the SOEC and the national statistical departments is clearly essential in order to avoid stumbling blocks and to arrive at the most accurate answers .
114 . To that end, pursuant to Point II, 1.1 of the Annex to the 1981 Decision, both the determination of the joint indexes ( as required by Article 65 ( 1 ) of the Staff Regulations ) and the verification of the correctness of the weightings call for agreement with the statistical departments of the Member States . It is to be noted that it is not merely a question of the SOEC being required to hear the views of those departments or to work on the basis of information received from them - there must be agreement with them .
115 . In the present case I must confess that the Commission' s conduct causes me to hesitate as to the conclusions to be drawn .
116 . It seems to me that the best meaning to attribute to the obligation to act "in agreement with" the national statistical departments is that the SOEC must "collaborate" or "cooperate" with them .
117 . However, as we have seen, the Commission has provided explanations and documents concerning the manner in which it carried out the survey with the help of the national statistical departments . It does not appear that any of them refused to cooperate or objected to the method advocated, which is succinctly described in SOEC' s letters . And it does not seem to me that the fact that contact was established with SOEC' s usual correspondents in the various national departments is indicative of any irregularity in the procedure or changes the nature of the action taken by them .
118 . Is it also necessary for the method to be discussed with those departments in advance? Or for it to be approved by the Working Party on Price Statistics which supports the SOEC?
119 . What is the importance to be attributed to that working party in the process of verification and review of the weightings?
120 . In so far as it comprises national statistics experts, will its agreement - beforehand or afterwards - be necessary, in the light of the second paragraph of Point II, 1.1 of the Annex to the 1981 Decision?
121 . And what will be the consequences of an objection by one of the national experts in the working party?
122 . The reply to those questions is helpful, although not absolutely essential to the decision to be given in the present dispute .
123 . Let me start by pointing out that, within the framework of the division of powers between the various bodies and institutions involved, it is incumbent upon the SOEC, whenever necessary, to provide a technical basis on which the Commission is to prepare its proposals and the Council take its decisions . In an eminently scientific and technical field, its work is, of course, of fundamental importance .
124 . With respect to the determination or review of weightings, the 1981 Decision is peremptory : it is for the SOEC to verify (" the SOEC shall verify ..."), in agreement with the statistical departments of the Member States, whether the ratios between weightings accurately reflect purchasing-power equivalences ( second paragraph of Point II, 1.1 ). It should be noted that its only responsibility is to verify the accuracy of the weightings, not to adjust them .
125 . In those circumstances, I consider that, since the Commission based its proposal on the work done by the SOEC, the Council, which must take the final decision to adjust the weightings whenever it proves necessary to do so, may override the technical data provided by the SOEC only in the event of a manifest error or procedural irregularity .
126 . That will be the case, for example, where the figures supplied are clearly inaccurate, or where the Council is able to set against them other more credible figures obtained from other bodies of equal standing or, again, if the verification of the ratios between the weightings has not been carried out "in agreement" with the national departments .
127 . It must be stated that, in the case under review, it cannot be concluded without reservation that any of those defects are present .
128 . On the one hand, it must be borne in mind that the results of the survey based on data from estate agents were subjected to tests of accuracy which the Council was not able to challenge convincingly .
129 . Moreover, not only were no better figures produced but it has also been established that the methods adopted are used by other important international organizations with respect to their officials .
130 . And, as has been seen, the SOEC collaborated directly with the national statistical departments .
131 . It is true that, in explaining the basis of that collaboration between the SOEC and the national statistical departments, the Commission mentioned the Working Party on Price Statistics as one of those which ( like the Working Party on Article 65 of the Staff Regulations ) participates in that collaboration . It appears that the Working Party on Price Statistics is even made up largely of representatives of the national statistical departments .
132 . If that is the case, it must be concluded that the intervention of the working party in the work involved in verifying the weightings must start with the preparatory measures for such work and the definition of the methods to be applied . It is doubtful whether that happened in the present case .
133 . It also appears from the information before the Court that, within the working party, there were objections to the results produced after the survey was carried out . However, it appears that only one national expert ( from the United Kingdom ) was involved, and his objection related only to the survey for London .
134 . But neither with regard to this point ( the position of the Working Party on Statistics ) nor with regard to the intervention of the national statistical departments do the proceedings disclose any solidly reliable information making it possible to form a comprehensive view of the sequence of events and to evaluate them in detail .
135 . It must also be said that, in view of the consultative nature of the functions of the Working Party on Price Statistics, an objection by one of its members ought not to be able to obstruct the decision-making process of the SOEC and the Commission, unless there are specific circumstances which, if not taken into account, would lead to an obviously incorrect result .
136 . The main issue in this action is not, however, the method of preparation of the Commission proposal rejected by the Council but rather the final regulation adopted by the Council and attacked by the Commission .
137 . It must be stated that, even if the conclusion were inescapable that the SOEC failed to comply with the procedure laid down in the 1981 Decision and therefore that the Commission' s fresh proposal was properly rejected, that would not mean that the Council would be entitled purely and simply to revive the Commission' s initial proposal ( excluding the 2.5% threshold ) which was rejected as being based on dubious information .
138 . On the contrary, compliance with the Staff Regulations requires not only that a solution should be adopted which takes account of the real equivalence of purchasing power of officials' remuneration but also that that solution should be based on the collaboration of the various statistical departments concerned and on results which the SOEC itself does not regard as distorted .
139 . And, if the United Kingdom experts' objections relate, as the Commission maintains, only to the results for London, then it would be feasible, using the information available and, perhaps, taking the requisite further steps, to make the adjustments needed to secure unanimous approval .
140 . In view of all the foregoing, it must inevitably be concluded that, on the basis of the applicant' s first submission, the contested regulation must be declared void .
141 . It would appear that, on the basis of the information available and by making the necessary marginal amendments in collaboration with the national statistical departments, the SOEC could, without encountering any significant obstacles, fulfil its task of supporting a new Commission proposal complying with the Court' s judgment .
142 . What is now to be said of the second submission?
2 . The retroactive effect of the weightings
143 . When presenting its second proposal for review of the weightings, the Commission proposed that it should take effect as from 1 January 1981, the end of the five-year period which had elapsed since the previous review of the weightings ( 1 January 1976 ).
144 . The Council did not, however, accept the proposed date as the starting point for the new weightings to take effect, claiming that "it (( was )) not now possible to determine with sufficient accuracy the circumstances obtaining as at 1 January 1981 ".
145 . Instead of that date the Council then chose "the first suitable date after the submission of the amended proposal, in this case 1 July 1986 ".
146 . The Commission, quite rightly, alleges that decision to be illegal, on the following grounds :
( a ) the purpose of reviewing the weightings, as is apparent from Article 64 of the Staff Regulations, is to monitor and ensure the purchasing-power equivalences of officials' remuneration, for which reason the adjustment must take effect from the time at which the change in purchasing power was determined ( 1 January 1981 );
( b ) according to the Court' s judgment in Roumengous ( paragraphs 25 and 28 ) and in other cases, ( 12 ) it is clear that, under Article 65 ( 2 ) of the Staff Regulations, the Council has no discretion as to whether or not the measures adjusting the weightings are to be retroactive where there is a substantial change in the cost of living . In such a case, the Court stated, once it is established that there has been a substantial increase in the cost of living, the Council must draw the appropriate conclusions .
In compliance with the judgment of the Court, the Council adopted Regulation No 3681/83 of 19 December 1983 ( 13 ) fixing the new weightings with retroactive effect to the time when the substantial rise in the cost of living occurred ( in Rome and Varèse ).
The Commission considers that the rule laid down by the Court for the application of Article 65 ( 2 ) applies a fortiori to Article 64 which, moreover, was the legal basis for Regulation No 3087/78 of 21 December 1978 ( 14 ) which was challenged in Roumengous, and for the amending regulation of 19 December 1983;
( c ) the Commission also challenges the two reasons for which, according to the Council, it was impossible to determine, with sufficient accuracy, the situation obtaining at 1 January 1981 .
Firstly, it is inappropriate to refer to the "dates on which the original and the amended proposals were submitted", since all that is important is whether or not the body entrusted with calculation of the price levels was in a position to carry out its task, which it did by means of interpolation, a common statistical technique, on the basis of appropriate national price indexes . The one-and-a-half-year period between the delivery by the SOEC of the results of the survey and the submission of the Commission' s first proposal was due to hesitations about inclusion of the accommodation item and to technical discussions with staff .
Secondly, it makes no sense to invoke "the difficulties which have arisen in regard to the exact calculation of the rent item" since the rejection of its first proposal left the Commission no alternative but to determine the cost of living as exactly as possible as at the reference date;
( d ) finally, the Commission considers the Council' s choice of 1 July 1986 to be at odds with the very terms of the regulation, in which the Council admits that "a certain drift has been recorded in the purchasing-power parities between officials" and in which mention is made of the dates on which the weightings were reviewed ( 1980, 1984 and 1985 ) before 1 July 1986 .
In addition, the Council' s choice (" the first suitable date after the submission of the amended proposal ") is regarded by the Commission as arbitrary and meaningless .
147 . It seems to me too that the Council' s position is untenable .
148 . The issue here is strict compliance by that institution with the requirements of Article 64 of the Staff Regulations and the 1981 Decision .
149 . There can be no doubt that the principle laid down in Roumengous can be transposed, mutatis mutandis, to the matter of the periodical ( five-yearly ) review of the weightings .
150 . That follows from recognition of the principle of purchasing-power equivalences, as upheld in that case and embodied in the 1981 Decision .
151 . The Staff Regulations and the decision which laid down the appropriate method of implementation require "the ratios between weightings acccurately to reflect purchasing-power equivalences between salaries paid to staff" in the various places of employment .
152 . Periodical verification of the weightings will bring about their adjustment whenever it indicates that changes in the cost of living are altering the ratios between purchasing power and detracting from the equivalence thereof .
153 . Accordingly, the conclusion is inescapable that the adjustment must be made with effect from the time to which the review of the purchasing-power parities related, when it was decided that the existing weightings no longer accurately reflected them .
154 . Since, as a rule, the verification is carried out ( or completed ) ex post facto, the adjustment will normally be "retroactive" in so far as it will take effect as from a time earlier than the date on which the analysis was carried out and the decision to make adjustments was taken as a result .
155 . That follows from the legal requirement itself and not from any idea of legal certainty or legitimate expectations, whether or not based on a commitment given by the administration .
156 . If that course were not followed, weightings which had long before ( five years ) been found not strictly to reflect purchasing-power equivalences would be left uncorrected for a considerable period .
157 . It should also be noted that application of the new weightings with effect only from 1 July 1986 would mean that no less than 10 years had passed since the previous review .
158 . It is true that the lapse of time since the reference period makes the calculations more delicate . But it does not seem to me that that argument can be used to oppose the retroactive application of the weightings .
159 . It must be borne in mind that the only factor in dispute is the "accommodation item" ( and, it would appear, as far as the technical experts are concerned, only - or basically - with respect to London ), since all the other cost-of-living items were calculated, without being challenged, as at 1 January 1981, on the basis of surveys carried out in 1980 and 1981 .
160 . The information concerning rents was initially obtained from national accounts ( first Commission proposal ) and then ( second proposal ) determined by the surveys carried out in 1984 and 1985, the results of which were, by interpolation, related to 1 January 1985 .
161 . It is thus neither just nor proper to use the lapse of time as a pretext for avoiding retroactive effect when values are available for January 1981 ( most of them having been obtained at that time ).
162 . In my opinion that statement applies both to the contested regulation ( based, as regards the cost of accommodation, on figures from national accounts ) and to the second Commission proposal, rejected by the Council ( which was based, as regards accommodation, on information obtained from estate agencies ). In the latter case, it was possible, by interpolation, to relate the figures to 1 January 1981 . That is a technique which is used universally; admittedly, it does not guarantee 100% certainty as to the results obtained, but that applies to all techniques used in statistics . The Council itself, moreover, appears not to have been able to avoid using it since, in the proposal from its own General Secretariat on which the contested regulation was based, it is stated that the new weightings were calculated by extrapolating to 1 July 1986 the data from the national accounts used for the verification of 1 January 1981 .
163 . Moreover, the point made by the Council concerning the financial cost of the operation is not a valid argument against retroactive application . It is not a consideration on which the Council may rely in order to oppose fulfilment of the requirements of the Staff Regulations and satisfaction of the legitimate expectations of officials arising therefrom .
164 . It now remains to determine the exact date to be adopted .
165 . The Commission refers to the five-yearly frequency provided for in the 1981 Decision in advocating that the date should be 1 January 1981 .
166 . That does not seem to me, however, necessarily to follow .
167 . In fact, the 1976 Decision referred only to a "periodical verification" without giving further details .
168 . The 1981 Decision, which upheld the principle of a five-yearly review, laid down ( in Article 1 ) that the new method for which it provided was to apply for 10 years as from 1 July 1981 .
169 . Since the decision took effect as from 15 December 1981 ( Article 2 ), it must be regarded as having become applicable to the process of reviewing weightings which was under way at that time .
170 . In any event, the inference must be that the review had to be carried out not every five years as from 1 January 1981 but only as from 1 July 1981 .
171 . Nevertheless, although it was not compulsory to relate the five-yearly frequency to 1 January 1981, it was already possible - even under the 1976 method - to base the "periodical review" of the weightings on that date . The adoption of that date will avoid the problem of having to adjust the basic data with reference to some other date, the only implication being that it will be from that date that the further periods of five years will thereafter be calculated .
3 . Conclusion
172 . In view of all the foregoing, I consider that the Commission' s action should be upheld and consequently that Council Regulation No 3619/86 of 26 November 1986 correcting the weightings applicable to the remuneration and pensions of officials and other servants of the European Communities working elsewhere than Brussels and Luxembourg must be declared void .
173 . To avoid any discontinuity in the system of remuneration, I consider it appropriate to accede to the Commission' s request that the Court should :
declare, pursuant to the second paragraph of Article 174 of the Treaty and as it did in its judgment of 5 June 1973 in Case 81/72, ( 15 ) that, with respect to the application of Article 65 of the Staff Regulations, the annulled regulation is to continue in effect until the Council adopts the new regulation .
174 . Since neither of the parties has asked for costs, each should bear its own .
( * ) Translated from the Portuguese .
( 1 ) OJ 29.11.1986, L 336, p . 1 .
( 2 ) Case 194/86 Rydalm v Commission; Case 327/86 Herkenrat and Others v Commission; Case 328/86 Brazzelli and Others; Case 163/87 Nowak v Commission; Case 295/87 Bertolo and Others v Commission; Case 300/87 Baumgaertner and Others v Commission; Joined Cases 312 to 317/87 Nonon, Andreason, Binns, Swinnock, Price, Kingston and Others v Commission, Parliament, Court of Justice and Court of Auditors; Joined Cases 332 and 336/87 Da Po, Actis-Dato and Others v Commission .
( 3 ) See judgment of 15 December 1982 in Case 158/79 Roumengous and Others (( 1982 )) ECR 4379, paragraph 9; see also the judgment of 15 January 1985 in Joined Cases 532, 534, 543, 567, 600, 618, and 660/79 Jan Amesz and Others ECR 55; See also Case 737 Dino Battaglia and Others (( 1985 )) ECR 55, at p . 71 .
( 4 ) OJ 31.12.1981, L 386, p . 6 .
( 5 ) It is true that the preamble to the 1981 Decision only mentions Article 65 and that Point II, 1 . of the decision is preceded by a reference to "Annual review of remuneration ( Article 65 ( 1 ) of the Staff Regulations )". However, although falling outside the scope of the annual review, Point II, 1.1 governs the five-yearly verification of weightings and the first paragraph thereof refers to Article 64 of the Staff Regulations .
( 6 ) Judgment of 6 October 1982 in Case 59/81 Commission v Council (( 1982 ) ECR 3329, at p . 3358, paragraph 33; Judgment in Case 158/89 Roumengous v Commission (( 1982 )) ECR 4379, at p . 4402, paragraph 28 .
( 7 ) Roumengous, supra, p . 4401, paragraph 23 .
( 8 ) See judgment of 26 June 1975 in Case 70/74 Commission v Council (( 1975 )) ECR 795, at pp . 807 and 808, paragraphs 20 to 23 .
( 9 ) At this point the Commission' s submissions are not very explicit and appear to have been curtailed .
( 10 ) "Not without justification", to use the Commission' s own words .
( 11 ) Information from agencies ( Commission proposal ), exclusion of the accommodation item, consideration of building costs and accommodation costs calculated by reference to general averages based on national accounts ( Council regulation ).
( 12 ) See also the judgment of 6 October1982 in Case 59/81 Commission v Council, cited above .
( 13 ) OJ 28.12.1983, L 368 .
( 14 ) OJ 29.12.1978, L 369, p . 10 .
( 15 ) Commission v Council (( 1973 )) ECR 575, at p . 587 .
Full & Egal Universal Law Academy