1.

In the present case and one other case, ( 2 ) the Court is dealing with the implementation and interpretation of its case-law by the Member States, in this instance the Kingdom of Sweden. The point at issue is whether a Swedish parent company has the right, on the basis of Article 49 in conjunction with Article 54 TFEU, to deduct losses in an indirectly and wholly owned Spanish subsidiary (a sub-subsidiary) from its profits made in Sweden if the sub-subsidiary has been liquidated and was not able to use all its losses in Spain (that is, to set them off against its own or other profits of the Spanish group).

2.

The Grand Chamber of the Court of Justice ( 3 ) ruled in 2005 that cross-border use of losses within a group is not required in principle by the fundamental freedoms. Losses arising abroad would be forfeited and could not therefore be used by other members of the group in national territory. Provision is to be made for cross-border use of losses only in the case of final losses in accordance with the principle of proportionality.

3.

A number of problems have grown up around this category of ‘final losses’ created by the Court, which has already led to several decisions by the Court ( 4 ) (including two others by the Grand Chamber). However, none of these decisions thus far have been able to clarify definitively the conditions for final losses. ( 5 )

4.

In this regard, the Court has another opportunity — if it still wishes to adhere to the final losses exception ( 6 ) — to refine this category.

5.

The framework for the case in EU law is provided by freedom of establishment of companies or firms under Article 49 in conjunction with Article 54 TFEU.

6.

In Swedish law, what is known as intra-group transfers are used to achieve the balancing of results within a group of companies. The intra-group transfer is set off by the company which makes the transfer and is entered for taxation by the company which receives it. By making an intra-group transfer to an (even indirectly owned) loss-making subsidiary, a parent company can ‘assign’ those losses economically to itself.

7.

The rules on intra-group transfers ( 7 ) do not apply, however, where the subsidiary is not liable for tax in Sweden. In this regard, only the rules on group relief introduced on the basis of the requirements of EU law permit a cross-border use of losses. Under those rules, a Swedish parent company may, on certain conditions, apply group relief to a definitive loss made by a wholly owned foreign subsidiary. ( 8 ) The rules provide that the subsidiary must be established in a State within the EEA and must correspond, inter alia, to a Swedish limited company.

8.

A loss is definitive if it has not been possible and will not be possible for the subsidiary or any other person in the State where the subsidiary is established to use it. It is further necessary that the reason that the loss cannot be used by the subsidiary is not that there is no possibility in law to do so or that this possibility is limited in time.

9.

Under Paragraph 5, the application of group relief is conditional, inter alia, on the subsidiary having been placed in liquidation and the liquidation having been completed (point 1). Further, according to the request for a preliminary ruling, the subsidiary must have been wholly owned by the parent company for all the tax years for the parent company and the subsidiary until the liquidation was completed or have been wholly owned by it since it began operating in any way until the liquidation was completed (point 2). The parent company must also have no associated companies which, on completion of the liquidation, carry on business in the State where the subsidiary is established (point 5).

10.

During the drafting of the legislative proposal, according to the referring court, objections were raised to the provisions being applicable solely to losses in directly and wholly owned subsidiaries. The view was taken, however, that if losses in indirectly owned subsidiaries were also to be included, it would become possible for companies to choose in which State the losses should be used.

11.

The case concerns a preliminary decision by the Skatterättsnämnd (Revenue Law Commission, Sweden). The preliminary decision is based on the following facts:

12.

Holmen AB (‘Holmen’) is the parent company in a group with subsidiaries in several countries, including Spain. The Spanish part of the group — in so far as is relevant here — is structured as follows: Holmen owns all the shares in the subsidiary Holmen Suecia Holding S.L. (HSH). HSH owns all the shares in the two sub-subsidiaries Holmen Paper Madrid S.L. (HPM) and Holmen Paper Iberica S.L. (HPI).

13.

Since 2003, the Spanish companies have formed a tax grouping and are taxed in accordance with the Spanish tax consolidation system. Under that system, the profits and losses of the entities which are members of the grouping can, without restriction, be set off against each other. That is achieved by the grouping drawing up a joint consolidated tax declaration for revenue. Unused losses may, without limit in time, be carried forward and deducted against any profits in future years.

14.

However, in Spain since 2011, only part of the profits made in a certain year may be set off against previous years’ losses. Losses which may not be deducted as a result of that amendment are carried forward, in the same way as other unused losses, to the next year. If the tax grouping is dissolved because the entities in the grouping are liquidated, any outstanding losses are allocated to the companies in which they arose. In the year of liquidation, it is not possible for any entity to use them other than that in which the losses arose.

15.

The Spanish part of the Holmen group made losses. The majority of the losses arose in the sub-subsidiary HPM. Since 2003 that company has accumulated operating losses relating to the activity in Spain totalling approximately EUR 140 million. The profits made by the Spanish part of the group during the period covered by the application (from 2003 inclusive forward) are insignificant. Holmen now intends to liquidate its Spanish activity. The liquidation was begun in 2016 when most of the assets of HPM were sold to an outside buyer.

16.

In the preliminary decision applied for, it must now be determined whether Holmen is entitled to apply group relief to the losses in HPM (the sub-subsidiary of Holmen) when the liquidation is complete. Two alternatives for liquidation are assessed in the preliminary decision applied for.

17.

Under the first alternative, the sub-subsidiary HPI, the sub-subsidiary HPM and the subsidiary HSH are liquidated in the same tax year and in the order stated. Under the other alternative, the subsidiary HSH is absorbed by the former sub-subsidiary HPM in a reverse merger, after which HPM (as a subsidiary) is liquidated.

18.

It is a condition for both alternatives that the group will not carry out any operations during the period when the liquidations are being completed and will not have any company remaining in Spain after the measures have been implemented. It also does not plan to operate there in future.

19.

The Skatterättsnämnd (Revenue Law Commission) found that Holmen is not entitled to apply group relief to the losses of the sub-subsidiary HPM if it were liquidated according to the first alternative. However, in the case of liquidation according to the second alternative, Holmen is entitled to apply group relief in respect of the definitive losses in HPM, then as a subsidiary.

20.

The Skatterättsnämnd (Revenue Law Commission) essentially based the preliminary decision, as regards the first alternative, on the ground that HSH (as the subsidiary) does not have any possibility in law in Spain of using the losses in the sub-subsidiary (HPM). The losses cannot therefore be regarded as definitive for the purposes of the case-law of the Court of Justice. A refusal to grant group relief to the Swedish parent company Holmen in respect of the losses thus cannot be regarded as disproportionate and accordingly does not run counter to EU law.

21.

As regards the second alternative, in essence the Skatterättsnämnd (Revenue Law Commission) based its conclusion on the ground that HPM (then as a directly owned subsidiary) has the possibility in law, on taxation in Spain, of itself making use of at least part of the losses in question. The fact that there is no longer any possibility for Spanish entities other than HPM to make use of the losses after the tax grouping is dissolved does not mean that the losses are definitive. At least part of the losses at issue could thus be regarded as final for the purposes of the Court’s case-law.

22.

Both the Skatteverket (Tax Board) and the applicant Holmen have appealed against the preliminary decision before the Högsta förvaltningsdomstol (Supreme Administrative Court, Sweden).

23.

The Högsta förvaltningsdomstol (Supreme Administrative Court), which is hearing the dispute, has referred the following questions to the Court:

24.

In the proceedings before the Court, Holmen, the Kingdom of Sweden, the Federal Republic of Germany, the Kingdom of the Netherlands, the Republic of Finland and the European Commission submitted written observations on these questions. The Tax Board, Holmen, the Kingdom of Sweden, the Federal Republic of Germany, the Republic of Finland and the European Commission took part in the hearing on 24 October 2018.

25.

By its first question — which concerns the interpretation of the judgment in Marks & Spencer — the referring court wishes to know whether freedom of establishment (Article 49 in conjunction with Article 54 TFEU) obliges Sweden to take into account the losses of a sub-subsidiary established and being liquidated in Spain. This question arises because, on the one hand, the losses could be set off only to a limited extent under Spanish tax law and would now be forfeited in respect of the Spanish sub-subsidiary on account of the liquidation. On the other hand, Sweden could never tax any profits of the sub-subsidiary.

26.

The question is therefore whether in the present case final losses in the sub-subsidiary of Holmen can be assumed to exist. The second to fourth questions also relate to the finality of those losses, such that to a large extent all the questions can be answered together.

27.

It must be examined, first of all, whether non-use of losses of non-resident sub-subsidiaries constitutes a restriction of freedom of establishment.

28.

Freedom of establishment, which Article 49 TFEU grants to European Union nationals, includes, in accordance with Article 54 TFEU, for companies formed in accordance with the law of a Member State and having their registered office, central administration or principal place of business within the European Union, the right to exercise their activity in other Member States through a subsidiary, branch or agency.

29.

It is settled case-law that all measures which prohibit, impede or render less attractive the exercise of the freedom of establishment are restrictions on that freedom. ( 9 )

30.

In order for tax legislation of a Member State to infringe freedom of establishment of companies, it must result in a difference in treatment to the detriment of the companies exercising that freedom; that difference in treatment must relate to objectively comparable situations and must not be justified by an overriding reason in the public interest or proportionate to that objective. ( 10 )

31.

Swedish law permits full loss relief between companies belonging to a group which are taxed in Sweden by way of intra-group transfers. In the case of non-resident group companies whose revenue is not taxed in Sweden, loss relief is dependent on the subsidiary being directly owned by a Swedish parent company. This excludes non-resident sub-subsidiaries (which are not taxed in Sweden) from loss relief. There is therefore a difference in treatment.

32.

This difference in treatment is liable to render less attractive the exercise of freedom of establishment through the creation of sub-subsidiaries in other Member States. It is, however, incompatible with the provisions of the Treaty only if it concerns situations which are objectively comparable.

33.

It should be recalled that, according to the case-law of the Court, the comparability of a cross-border situation with an internal situation must be examined having regard to the objective pursued by the national provisions at issue. ( 11 ) According to the referring court, the reason for the exclusion of foreign sub-subsidiaries is that otherwise the group’s management would have a right to choose where the losses in sub-subsidiary are used. Possible options are, for example, the Member State of the subsidiary or the Member State of the parent company.

34.

Germany considers that there is no comparability in this regard, making reference to the Court’s judgment in Timac Agro Deutschland ( 12 ) and my Opinion in Commission v United Kingdom. ( 13 )

35.

Thus far, with regard to the comparability of domestic and foreign permanent establishments, the Court has focused on whether the Member State concerned also exercises any tax powers over the foreign permanent establishment. It thus expressly ruled ( 14 ) that ‘[i]n the present case, it must be held that, since the Federal Republic of Germany does not exercise any tax powers over the profits of such a permanent establishment, the deduction of its losses no longer being permitted in Germany, the situation of a permanent establishment situated in Austria is not comparable to that of a permanent establishment situated in Germany in relation to measures laid down by the Federal Republic of Germany in order to prevent or mitigate the double taxation of a resident company’s profits.’ This idea could also be applied to sub-subsidiaries resident abroad and not taxed in national territory.

36.

However, the Court has developed a settled case-law concerning the cross-border use of losses between subsidiaries and parent companies where comparability has been implicitly or expressly accepted. ( 15 )

37.

In addition, recently in Bevola the Court again expressly accepted, as regards final losses attributable to a non-resident permanent establishment, the comparability of taxed domestic and untaxed foreign permanent establishments. ( 16 ) This would seem to have to apply a fortiori to taxed domestic and untaxed foreign controlled sub-subsidiaries.

38.

Lastly, the criterion of comparability is vague. Given that all situations are comparable in some respect, if they are not identical, ( 17 ) this test should in any case be abandoned. ( 18 )

39.

Accordingly, comparability must be taken to exist. Differences which exist — here the lack of symmetry between taxation of profits and use of losses ( 19 ) — in the case of a foreign as opposed to a domestic sub-subsidiary are to be taken into consideration only in respect of the justification. There is thus a restriction of freedom of establishment.

40.

A restriction of freedom of establishment may be justified by overriding reasons in the public interest. Justifications can be the preservation of the balanced allocation of the power to impose taxes between Member States and the avoidance of double use of losses (even though they were only taxed once). ( 20 ) In addition, the measure must be appropriate to ensuring the attainment of its objective and not go beyond what is necessary to attain it. ( 21 )

41.

According to the Court’s settled case-law on ‘final’ losses, it is disproportionate if the Member State refuses the parent company use of losses even though the foreign subsidiary has exhausted all possibilities of having the losses taken into account and it is no longer possible for those losses somehow still to be used. This must be demonstrated by the taxable person. ( 22 ) However, it could not be shown by a liquidation following a merger that there was no possibility of taking into account the losses that existed in the subsidiary’s State of residence. ( 23 )

42.

Double use of losses seems to be ruled out in the case at issue. According to the preliminary decision, Holmen has ceased all economic activity and no longer has any realisable economic assets. Only the losses stemming from the previous financial years still remain and the company is to be liquidated. As both the subsidiary and the sub-subsidiary are resident in the same Member State and use of losses in the case of the subsidiary is precluded by Spanish tax law, there is also no risk that the losses in the sub-subsidiary could be used twice in respect of the parent company and the subsidiary.

43.

If, however, there is no risk of double use of losses, this justification does not apply.

44.

As regards the balanced allocation of the power to impose taxes between Member States, it should be pointed out that it is a legitimate objective recognised by the Court, ( 24 ) which may make it necessary to apply to the economic activities of taxable persons established in one of those Member States only the tax rules of that State in respect of both profits and losses. ( 25 )

45.

In the present case, however, it is not possible, on the basis of this justification, to assume the existence of final losses to be used, for three reasons. First, use of the losses in the sub-subsidiary made in Spain over the years would undermine the fiscal autonomy of the Member States (see point 46 et seq.). Second, the condition of losses which are usable in law, but not in fact is not satisfied in this case (see point 57 et seq.). Third, final losses in relation to the parent company are precluded in principle in the context of an indirect participation (that is, in the case of a sub-subsidiary) (see point 73 et seq.).

46.

As the Court has already ruled, the fundamental freedoms cannot have the effect of requiring the Member State of residence of that parent company to grant that company a use of losses for an amount originating solely from the tax system of another Member State, if the first Member State is not to see its fiscal autonomy limited by the exercise of fiscal power of the other Member State. ( 26 )

47.

As the Court has expressly stated, ( 27 )‘losses sustained by a non-resident subsidiary cannot be characterised as definitive, as described in paragraph 55 of the judgment in Marks & Spencer, ( 28 ) by dint of the fact that the Member State in which the subsidiary is resident precludes all possibility of losses being carried forward’. ( 29 ) A Member State would then have to adapt its tax legislation to that of another Member State.

48.

If, according to the Court’s case-law, ( 30 ) losses cannot be characterised as definitive by dint of the fact that the Member State in which the subsidiary is resident precludes all possibility of losses being carried forward, this must also apply to a preclusion of a transfer of losses to a third party. The same holds for the completion of a transfer of losses. In both cases, a future use of losses — once by the taxable person itself, once by a third party — is prevented. The situations are therefore also to be treated in the same way.

49.

As was stated in point 41 et seq. above, before ‘final losses’ are accepted it must therefore be examined, according to the Court’s case-law, whether they could not have been used previously by a transfer to third parties. This can only mean third-party transferees or other companies in the group. If this is not possible, however, because it is precluded by Spanish tax legislation — in this case on account of the liquidation leading to the end of the Spanish group consolidation — this legal preclusion of setting off losses does not mean that there are now final losses.

50.

In any case, the Court has ruled that it is not contrary to the fundamental freedoms if a loss which can be set off transnationally is always to be established as a final loss at the end of the assessment period. ( 31 ) Therefore, any loss which can be carried forward is non-final, at least initially. ( 32 )

51.

Such accumulated (carried forward) losses which are regarded as non-final in one year (because they can be carried forward or setting off the losses was precluded under national law) cannot subsequently become final losses because they cannot be carried forward further on account of the liquidation.

52.

Otherwise, the initially successful activity in Spain would be taxed solely in Spain, while the subsequently loss-making activity would be financed by the tax revenue of the States in which the group parent companies are established. This would run counter to the preservation of an appropriate allocation of the power to impose taxes.

53.

This is particularly clear from the relevant Spanish provision, which since 2011 has taken into account losses carried forward only to a limited extent, even if there are sufficient profits in that year which Spain has also taxed. It is apparent that these losses still exist primarily from the legal restriction on setting off losses imposed by Spain in 2011. The Swedish system of tax law does not have to adapt to this (principle of autonomy).

54.

Similarly, the character of the non-final loss does not change subsequently upon a transfer to the third party (here the group parent company in Spain) merely because the Member State puts an end to the transfer of losses. Rather, it is contrary to the principle of autonomy that, because of the particularities of national law in another Member State, non-final losses in one assessment period become final losses again in a subsequent assessment period.

55.

Along the same lines, the Court considers in Commission v United Kingdom that there can be no subsequent change to finality once absent. ( 33 ) In any case, the statements made in that judgment indicate that at most the loss in the subsidiary made in the last year of liquidation must still be able to be set off (transnationally) somehow, but not the losses accumulated up to then and carried forward under national (here Spanish) law. ( 34 )

56.

As the existence of the losses at issue in this case stems solely from Spanish law, they are not final losses in the sub-subsidiary.

57.

Against this background, almost all the parties to the proceedings distinguish, in assessing the finality of a loss, between losses which cannot be used in law and in fact (final losses).

58.

Losses which cannot be used because they are not legally recognised in the Member State in which they arose or are not usable because of legal restrictions (for example, they cannot be carried forward or back) are not intended to constitute final losses in accordance with the Court’s case-law. Only losses which would be usable in law but cannot be used in fact in future could be regarded as final losses. This is compelling on account of the autonomy of systems of tax law (point 46 et seq.).

59.

It nevertheless seems doubtful whether there can actually be losses which are usable in law, but not in fact. I would like to illustrate this with an example. The only case where a loss remains despite the possibility of carrying forward or back losses without restriction would be the case of an undertaking which is loss-making on the whole and which has never made sufficient profit, even after all economic assets have been sold. In that case, even the loss from the last year could not have any effect (in fact) despite the possibility of carrying back the loss.

60.

However, even in this case there would still be the possibility of transferring those losses to a purchaser with the sale of the undertaking, ( 35 ) provided this is permitted by the Member State of establishment. The purchaser will take into account the value of the existing losses through the purchase price for the undertaking, with the result that the seller thus ‘realises’ those losses.

61.

If the legal order in question permits a transfer of losses to other persons, it is also always possible in fact to use those losses. It may not be particularly successful in a specific case because the purchaser of a loss-making undertaking will not necessarily pay much money for such an undertaking. Nevertheless, this does not affect the usability in fact of the losses.

62.

The definitive nature of the losses in that case is thus also based either on the legal order of the Member State (preclusion of any possibility of transferring losses) or on the decision by the taxable person not to sell the company, but to place it in liquidation. In both cases, however, it is not obvious why non-use of losses in another Member State should be disproportionate. It is also not without reason the Court requires that all possibilities of having the losses taken into account have been exhausted. This includes the losses being transferred to a third party by way of a sale.

63.

Holmen confirmed at the hearing that these scenarios were considered, but it was decided to liquidate. In this respect too, it can therefore be stated that there are no final losses in the case of Holmen.

64.

This is also not precluded by the recent judgment in Bevola. ( 36 ) First, in that case the Court ‘merely’ applied the Marks & Spencer exception to ‘final’ losses of permanent establishments and did not call into question the reservations made above. ( 37 ) In particular, it did not make any more specific comments on when final losses exist.

65.

Second, the arguments raised in that more recent judgment relate primarily ( 38 ) to the ability-to-pay principle tax. This may be understandable in the case of permanent establishments as permanent establishments legally form a dependent part of a taxable person’s undertaking. This line of argument would not hold, however, in the case of subsidiaries and sub-subsidiaries. They are autonomous legal entities which also have an independent financial ability to pay (if this is understood to mean the ability to pay taxes based on their revenue). ( 39 ) The Court — rightly — did not decide that it is necessary for the correct taxation of the parent company’s ability to pay to take into account the losses in the subsidiary.

66.

From the point of view of tax law, group relief constitutes a breach of the ability-to-pay principle because the ability to pay of a number of legal entities is added together. The inclusion of other legal entities cannot therefore be justified in any case by the principle of taxation according to the ability to pay.

67.

On the contrary, it even runs counter to the principle of taxation according to the ability to pay if a Member State takes account of only one side (that is, only revenue or only expenditure). In addition, to my knowledge there is neither a general principle of tax law nor a general principle of EU law to the effect that relief should somehow be granted for all losses at the end of a life cycle of a legal entity. In particular, the ability-to-pay principle does not require losses to be exported to other Member States.

68.

Consequently, in accordance with the judgment in Bevola, there are no deductible final losses which can be exported from Spain to Sweden in this case.

69.

This conclusion based on case-law is also compelling from the point of view of a ‘fair’ internal market, which has been brought back into focus again in the light of the ‘BEPS debate’. ( 40 ) A possibility of setting off final losses transnationally would, specifically in the particular situation at issue, favour above all large groups operating across borders as opposed to smaller undertakings (which do not generally operate across borders). For example, if Holmen knows that all losses incurred from the Spanish business model can ultimately be set off against the profits in Sweden, then, in attempting to position itself in the Spanish market, Holmen can compete very differently from a Spanish competitor that has to assume that its losses will be forfeited if it ceases its commercial activity in Spain. For Holmen the ‘Spanish losses’ would be a much lesser burden than for a domestic competitor without a similar group structure.

70.

Bearing this in mind and consistently applying the Court’s case-law (see point 41 et seq. and the case-law cited therein), the following conclusion is therefore reached: If the use of losses is precluded by law in the State of the sub-subsidiary, there are no final losses. If it is possible for that State to use losses, the taxable person must have exhausted those possibilities. According to the judgment in Marks & Spencer, ( 41 ) this includes realising the losses by transferring them to a third party, which did not occur in this case.

71.

Accordingly, the preclusion by Sweden of the setting off of losses in a sub-subsidiary resident abroad and not taxed in national territory is not disproportionate.

72.

I will turn to the specific answers to the individual questions.

73.

By its first question, the referring court wishes to know whether the losses in a sub-subsidiary resident and being liquidated in Spain are to be regarded as final losses. That is not the case.

74.

The Marks & Spencer exception for final losses does not differentiate, as regards final losses, between subsidiaries and sub-subsidiaries. At first glance, it would thus permit both the group’s parent company and the intermediate subsidiary to set off losses against the final losses in the sub-subsidiary.

75.

The Marks & Spencer judgment also related to an indirect chain of holdings, as is rightly pointed out by Holmen and the Commission. The sub-subsidiary, the parent company (a holding company) and the grandparent company (group parent company) in that case were even resident in three different Member States. This is only apparent from the request for a preliminary ruling and the Opinion of the Advocate General, however. ( 42 ) It is not mentioned in the facts in the judgment, nor is it examined by the Court.

76.

However, it appears excessive to me, contrary to the view taken by the Commission and in agreement with the Netherlands and Sweden, to infer from this that the Court implicitly ruled that the group’s parent company must also be able to use the (final) losses in a sub-subsidiary. The Court did not have to examine this question more closely in that case because it had not been asked about this point.

77.

Above all, however, such an approach would result in a right to choose within a group, which could decide in which Member State a subsidiary or a parent company will use the ‘final’ losses in the sub-subsidiaries.

78.

In particular, if all three companies were resident in different Member States and there were profits to be set off, this right to choose would be important with a view to optimising the group tax rate. As the Member States participating in the proceedings rightly assert, however, such a right to choose cannot exist. It would also jeopardise the preservation of the balanced allocation of the power to impose taxes between Member States. In addition, there is a risk that the losses could be used in more than one Member State.

79.

As it is still possible in principle to set off losses in another Member State for the direct parent company, there is a fundamental precedence for setting off losses against the direct parent company over setting off losses against the indirect parent company (in this case the group’s parent company in Sweden). That precedence also avoids the abovementioned risks of a right of choice for the taxable person and the possibility of double use of losses in three State scenarios.

80.

That precedence applies even where — as here — the sub-subsidiary and the subsidiary are resident in the same Member State. In this case there is no risk of optimisation of the group tax rate by selecting the Member State in which losses are set off, as Holmen rightly argues. Likewise, an increased risk of multiple use of losses is ruled out. Here too, however, the crucial question is not whether the subsidiary and the sub-subsidiary are resident in the same country, but whether final losses exist for the sub-subsidiary in relation to the parent company in the other Member State. As stated above, however, that is not the case.

81.

The answer to the first question is therefore that the losses in an indirectly owned company (a sub-subsidiary) do not in principle constitute final losses in relation to the ‘grandparent company’ (the parent company of the subsidiary).

82.

By the second question, the referring court wishes to know whether a loss which had to be carried over as a result of a restriction on setting off losses is also to be regarded as definitive.

83.

The answer to this question, in accordance with the above statements, is that a loss ‘merely’ carried over is not to be regarded as a ‘final loss’ even if it has not been possible to set off against earlier profits as a result of a restriction on loss relief in the subsidiary’s State.

84.

The existence of that loss depends solely on the organisation of Spanish tax law and cannot compel Sweden to recognise that loss as reducing the tax burden.

85.

By its third and fourth questions, the court wishes to know whether in the assessment of the finality of a loss account must be taken of the fact that, under the rules in the subsidiary’s State, the possibilities for parties other than the party making the loss itself to deduct the loss are restricted and, if so, whether regard must be had to the extent to which the restrictions have in fact led to it not being possible to set off any part of the losses.

86.

The answer to these two questions, on the basis of the above statements, is that in the assessment whether there are final losses in the sub-subsidiary only the restrictions in respect of the sub-subsidiary are significant. If it is possible for the sub-subsidiary to transfer the losses to a third party (a subsidiary for example), final losses in the sub-subsidiary are ruled out. It is therefore immaterial whether third parties were actually able to use the loss in the specific case. This is relevant at most for the question whether for such third parties there are final losses in relation to their parent companies.

87.

On those grounds, I propose that the questions referred by the Högsta förvaltningsdomstol (Supreme Administrative Court, Sweden) be answered as follows: