Two main methods have customarily been employed in attributing the investment return to non-resident general insurers.
(UK Technical provisions/World wide technical provisions) x (World wide investment return)
([UK technical provisions + solvency margin + ‘comfort
margin’]/World wide assets) x (World wide investment return)
These methods may not be fully compatible with the separate
enterprise principle in Article 7(2) of the OECD Model. Arguably
Method 1, in apportioning worldwide yield in the ratio of UK
technical provisions to the world technical provisions attributes
income on shareholders’ funds that would not be available to
the establishment were it a separate entity. However, if the
activities of the permanent establishment and of the rest of the
enterprise are homogeneous, Method 1 may give a reasonable
arm’s length result, assuming the national regulatory
requirements involved are similar. Otherwise adaptation may be
required.
Method 2 arguably seeks to confer some attributes of a
separate company with its own share capital and reserves upon a
permanent establishment. The hypothesis in Article 7(2) does not
say that the branch is deemed to be a separate company (and
therefore must have a share capital). But an enterprise looked at
on a stand alone basis would in general need capital in some form.
A partnership or Lloyd’s syndicate needs to be capitalised
although not in the form of share capital.
Whatever the justification, by reference to
General Reinsurance v Tomlinson, or the
OECD Model Commentary in the light of these arguments, the actual
circumstances of the branch activities must always be considered.
Any total transfer of cash-flow, as it arrives, to Head Office
would for example clearly be a ’dealing’ with the
enterprise of which the branch is a permanent establishment, and
not one that would exist at arm’s length. It is perhaps less
obvious that an initial under funding of a branch by its Head
Office is. The existence of the first situation clearly justifies
the attribution of investment yield to the branch above what is
shown in the branch books or regulatory return.
Attributed investment yield is not ’notional’
investment income. It is real, but has not been attributed by the
company to its branch. On this analysis it might seem impossible to
attribute to a UK branch more than the total investment return of
the company as a whole, but where there are investment gains and
losses this is a possibility. If the entity of which the UK branch
forms part has had insufficient investment return attributed to it
as a result of non- arm’s length dealings with other members
of a group Article 9 of the OECD Model (Associated Enterprises) and
possibly ICTA88/SCH28AA may be in point (see INTM430000+).
Methods 1 and 2 should now be seen as simplified
applications of the approaches discussed in the OECD Report on the
Attribution of Profits to Permanent Establishments. They may give a
reasonable result in some circumstances. See
GIM10210+.
