Today is a big day for the EU Savings Tax Directive. I discussion paper on its reform is being discussed.

The paper is welcome. But it does not go far enough to tackle the weaknesses in the current arrangements. The key issues it raises are:

1. Whether the scope of the EU STD should be extended from natural people (real human beings) to include legal persons (companies, trusts and foundations) that those real people won and which can be used by them to escape the obligations of the EU STD.

2. Changing the definition of a paying agent so that some recipients of taxable income become paying agents in their own right to avoid some of the current problems of not being able to identify the beneficiaries of income paid. These might include discretionary trusts and some legal entities such as limited liability partnerships which are 'tax transparent'. The paying agent is the person who has to actually operate the EU STD and decide whether to withhold tax or not.

3. Whether the scope of the income covered by the EU STD should be extended to include interest wrapped up into other products e.g. UCITS, pension products and life assurance. It's worth noting that:

a. dividends and capital gains are specifically excluded from consideration, noting that these are best dealt with by 'information exchange'.
b. The possibility of a 'substance over form' test is suggested to determine what might and might not be covered: a move towards a targeted anti-avoidance principle.

This is fine, but some of the conclusions are weak. In particular it is assumed that the operation of the paying agent arrangement imposes a considerable burden upon the paying agent. Since these paying agents are banks or their representatives there is at present no evidence that this is causing them undue economic hardship: other things may be.

There is also some considerable doubt expressed about how the EU STD may be extended to legal entities within the EU, but some apparent belief that this would be easier to do with regard to entities beyond the EU (presumably therefore those in tax havens such as the Channel Islands which are covered by the EU STD but which are not EU members). There is considerable risk in this: the concept of the level playing field is important, and the need to prevent market distortion is noted throughout the document. To create an obligation on non-EU locations covered by the STD not imposed on EU members would be dangerous and might induce the withdrawal of these locations from the scheme altogether. That would not be beneficial.

What worries me is the tone of the document. It assumes that:

1. Since the onus of proof is on the paying agent they have reason to resist change;
2. A source of income is excluded unless included;
3. A tax transparent entity is not taxable as a paying agent until made so.

On (2) and (3) there are welcome signs that a) on (2) a substance over form test could be used, and a committee might be established to monitor innovation to ensure that market developments are tackled and b) on (3) there is again the idea that the list need not be fixed and will be kept under review.

If this is the case then I think we should welcome these approaches, whilst encouraging open and automatic exchange of information on capital gains and dividends and then look at the paying agent problem as the focus for attention. Here I suggest:

1. To make the EU STD work the logic has to change. By 2011 withholding rates will be 35%. The risk of major market distortion arising is considerable unless this applies to all entities through which an individual might manage their affairs. This means that the current logic of information exchange by default in most countries has to be changed to withholding by default, and that this should apply to legal entities (companies, trusts, foundations and tax transparent vehicles) that can be shown to have a link with an EU resident unless:
a. The recipient can be proven by way of supply of taxpayer information to reside in the state in which payment arises;
b. The person / identity provides taxpayer sufficient information in a required form to ensure that effective information exchange takes place;
c. The person / entity proves that they have no reasonable likelihood of connection with an EU resident person. There's an obvious weakness here of who will certify this, and that may need to be the bank into which the funds will be deposited who has a duty to know their client rather than the fund holder themselves.
2. Funds then not positively opted out in this way will be subject to withholding. If the beneficial owner of the funds cannot then be allocated to a state withholding will take place and apportionment should be on the ratio of national ownership of identified deposits within the state in which payment arose, with the retaining state keeping 25%, as now.

This first of all does not prejudice the havens, who are prejudiced now (to some degree). Second, it ensures that if 35% is likely to be deducted at source from all entities the incentive to opt into the tax declaration system will be high. Third, the burden on the paying agent is massively reduced. They need only to know they have been supplied in good faith with taxpayer identification data to ensure that they can pay gross. It would be reasonable to assume that methods for cost effective positive checking of this data should be capable of being established (and I think that is a reasonable expectation) so that once this has been done the burden on the paying agent is limited.

This then creates what the EU wants: a level playing field and reduced obligation on paying agents plus a massive disincentive to evade with a considerable incentive to comply. There is no market distortion outside the EU: payments will still be made gross there so long as certifiable evidence of identity is secured - and that is a legal obligation already imposed on banks.

I urge the EU therefore to be bold: change is possible here and the potential gains are enormous. Tax evasion has to be stopped.