0005KPMGIFRSPracticeIssuesConsolidationModel2012

1、Simplifying the model for fund managers

1.1Control requires a link between power and returns

Our previous publication on IFRS 10 Consolidated Financial Statements1 discusses the new consolidation model in its entirety, taking readers through the standard step by step. In this publication we focus in more detail on the key issue for fund managers: should a fund manager consolidate its managed funds?

The diagram below is a reminder of the three elements in the control model. With its delegated power, and usually with an incentive fee paying a share of returns or with a co-investment, the fund management business model will usually meet the first two elements. So it is the third element, the link between power and returns, that is usually key for fund managers. On the assumption that the fund manager has power and exposure to variability, consolidation is then required only when the fund manager has the ability to use its power to affect its returns.

The linkage test determines whether power is deemed to be used:

for oneself, in which case the fund manager is a principal and will consolidate the fund; or

for others, in which case the fund manager is an agent and will not consolidate the fund.

1.2Not all indicators of linkage are equal

IFRS 10 provides a number of tests and indicators to assess whether linkage is present. However, it appears that the analysis for fund managers will come down to a combined assessment of just two key indicators: kick-out rights and aggregate economic interests. Appendix 1 addresses the other tests and indicators, and why we do not consider them to be practical focus areas.

Indicator	Aggregate economic interest	Kick-out rights
Meaning of the indicator	The aggregate economic interest is made up of remuneration plus other interests.The fund manager’s remuneration includes items such as management fees (e.g. 1 percent per annum of total net asset value) and performance fees (e.g. 20 percent of returns when a certain percentage performance is achieved).The fund manager’s other interests represent items such as any equity investment.The key attribute of aggregate economic interest appears to be the fund manager’s variability at the expected level of fund performance. In other words, if the fund does a little better or a little worse than expected, what percentage of that variation does the fund manager gain or suffer?	Kick-out rights represent the power held by another party to remove the fund manager.The key attribute of the strength of kick-out rights is the number of investors who need to act together to exercise such a right.
Relevance of the indicator	The greater the variability of a fund manager’s interest in a fund, then: ●● the greater the weighting placed on its aggregate economic interest; and ●● the higher the likelihood that the fund manager is a principal.	The smaller the number of investors who must act together, then: ●● the greater the weighting placed on their kick-out rights; and ●● the lower the likelihood that the fund manager is a principal.

1.3There is a trade-off between kick-out rights and aggregate economic interest

The indicators are required to be considered together. So, the stronger the kick-out rights, the more aggregate economic interest can be accepted while still being an agent. Conversely, the weaker the kick-out rights, the less aggregate economic interest can be accepted while still being an agent.

The chart below provides a way of visualising a general scheme for the result of combining different strengths of each indicator:

• ●● In the green zone, the combination of strong kick-out rights and low aggregate economic interest suggests that the fund manager is an agent.

• ●● In the red zone, the combination of weak kick-out rights and high aggregate economic interest suggests that the fund manager is a principal.

• ●● In the marginal zone, the combination does not give a clear outcome.

For those cases that fall into the marginal zone, the fund manager will need to consider certain other aspects of these indicators to determine if it is an agent or a principal. These aspects are discussed later in this publication. The question of where the central, marginal zone starts and finishes is not clear; there are no bright lines.

2、Aggregate economic interest – remuneration plus other interests

Look at the variability in expected returns

The fund manager needs to identify a key measure of aggregate economic interest. It appears that this measure is the variability of expected returns, measured at performance levels at which a performance fee kicks in.

2.1 What is in the aggregate?

The first key indicator is ‘remuneration and other interests in aggregate’. (We refer to this as the ‘aggregate economic interest’). It appears from the words quoted that this amount is simply the sum of remuneration and other interests.

2.2Variability around expected returns – the leading measure of this aggregate

IFRS 10 asks for an evaluation of “the magnitude of, and variability associated with, its economic interests ... relative to the total variability of returns of the investee. This evaluation is made primarily on the basis of returns expected from the activities of the investee but shall not ignore the decision maker’s maximum exposure to variability of returns” (emphasis added).

Of these measures, it appears that the variability associated with the expected level of returns (‘expected variability’) is the primary measure of aggregate economic interest. This is because, first of all, the standard is clear that the assessment is based primarily on the fund manager’s expected returns. Second, we believe that variability (at expected return level) is the leading measure, for the following reasons.

● It is a key element of the definition of returns in IFRS 10; the standard focuses on returns as being those that have the potential to vary as a result of the investee’s performance.

● It is identified as a headline factor to assess when performing the principal-vs-agent assessment – e.g. the decision maker’s exposure to variability of returns.

● All of IFRS 10’s examples, bar one, refer exclusively to variability; magnitude, however, is mentioned in only one example. Furthermore, from the facts given in the examples, variability is the only measure that could be calculated (see 2.3).

The guidance also requires consideration of magnitude and the maximum. We believe, as explained below and further in 4.2, that these measures only become important in analysing marginal cases. For the calculation of magnitude, see 4.2.2.

IFRS 10要求对“经济权益的量级和相关变动性...相对于投资对象的总回报变动性”进行评估，主要基于预期从投资对象的活动中获得的回报，但不应忽视决策者的最大回报变动性（强调添加）。

在这些度量标准中，似乎与预期回报水平相关的变动性（“预期变动性”）是衡量总体经济权益的主要度量标准。这是因为，首先，标准明确规定评估主要基于基金经理的预期回报。其次，我们认为变动性（在预期回报水平上）是主导的度量标准，原因如下：

• 这是IFRS 10中回报定义的关键要素；该标准侧重于回报作为有潜在因投资对象表现而发生变化的潜力。
• 它被确定为在进行主体与代理评估时要考虑的关键因素，例如，决策者对回报变动的敞口。
• IFRS 10的所有示例，除一个之外，都专门提到变动性；而量级仅在一个示例中提到。此外，从示例中给出的事实来看，变动性是唯一可计算的度量标准（请参见2.3）。

指南还要求考虑量级和最大值。我们认为，如下文所述并在4.2中进一步解释，这些度量标准只在分析边际情况时才变得重要。有关量级的计算，请参见4.2.2。

2.3How is variability measured?

IFRS 10 requires an assessment of the ”variability associated with its economic interests ... relative to the total variability of returns of the investee”. This is asking: how much, relatively, does the fund manager’s total income vary as fund performance varies?

IFRS 10要求对“与其经济权益相关的变动性...与投资对象总回报变动性相对比”进行评估。这是在问：相对于投资对象的回报变动，基金经理的总收入相对变化有多大？

From this, it appears that variability is the marginal increase/decrease in the fund manager’s returns relative to a marginal increase/decrease in the fund’s returns. Put simply, if the fund’s performance, starting from a particular level, increases or decreases, how much of that increase or decrease does the fund manager gain or suffer; or how geared to changes in fund performance are changes in the fund manager’s income? We believe that this can be calculated relatively simply, as shown by the following example.

因此，变动性是指基金经理的回报在边际基础上相对于基金的回报的边际增加/减少。简单地说，如果从特定水平开始，基金的表现增加或减少，那么基金经理的收益会增加或减少多少；或者说，基金经理的收入对于基金表现的变化有多大杠杆？我们认为这可以相对简单地计算，如下面的示例所示。

Example 1 – calculation of variable interest

Suppose that an investment manager has:

• ●● a 1% management fee calculated on net asset value (NAV); plus

• ●● a performance fee paying 20% of additional profits after management fees once an 8% (profits after management fee) hurdle is reached; plus

• ●● a 10% investment. The variability at any level of return at which a performance fee is due can be calculated as follows2.

In other words, for every 1 more or less of fund performance, the fund manager receives 0.287 more or less. Variability is therefore 28.7%. The calculation can be done in a short-cut way as follows: 1% + 20% x 99% + 10% x (80% x 99%) = 28.7%.

This might also be expressed mathematically as:∆(M + P + U) Where,

∆/∆ = rate of change with respect to ∆R M= income from management fee

U = income (including capital appreciation) from units held

R = returns from the activities of the fund – i.e. before fees

P = income from performance fee

2.4The expected return level at which to measure variability

As noted in 2.2, it appears that variability is measured at the expected level of fund returns. The question is: what is this level? After all, there may be different rates of remuneration at different levels of performance – usually, there is a hurdle return above which a performance fee is paid – and the fund manager’s marginal income (variability) will differ accordingly.

In answer, it appears that the expected performance level should be at a level that includes a performance fee. There are several reasons for this.

●● Performance fees are set with the intention of being achievable. Therefore, the expected range of performance cannot be below the level at which a performance fee kicks in. To calculate the aggregate economic interest at performance levels below the performance fee would, in effect, disregard the performance fee altogether. That would not seem appropriate, particularly when considered in combination with the bullet below – i.e. that the performance fee incentivises the manager to uses its power to affect its own returns.

● The incentive effect of the performance fee is important. IFRS 10 asks whether the fund manager has the ability to use its power to affect its own returns from the fund. The fee incentivises the fund manager to behave so as to obtain this return. For this reason, we believe that it is the incentive effect of the fee, not whether it is achieved, that is the more important feature.

●● IFRS 10’s examples ignore the hurdle. These examples note that the performance fee is paid after a hurdle rate, but the rate is unspecified. The existence of the hurdle, in the sense of a barrier to payment of the performance fee, is seemingly disregarded. That is also consistent with the preceding bullet points. On that basis, we do not believe that the standard and its examples intend that performance fees be routinely disregarded.

Focussing on the level at which performance fees arise is also a practical approach. The calculation is simple and is not sensitive to the precise level of performance above the hurdle, assuming that thefee has only one rate/hurdle. Once the fee has become payable, the fund manager’s share of further changes in the fund performance is always the same. If there is more than one performance fee level, then the decision, as to which performance fee band to build into the quantitative measure, will need to be more precise (i.e. determine which hurdle level is expected and build that into the initial calculation). The existence of higher bands may still need to be taken into account later as a qualitative feature of the aggregate economic interest – see 4.2.

If a performance fee was not expected to be reached in all periods, then this might be factored into the overall assessment, qualitatively, as noted in 4.2.

正如2.2中所指出的，变动性似乎是在预期的基金回报水平上测量的。问题是：这个水平是多少？毕竟，在不同的绩效水平可能有不同的报酬率 - 通常情况下，绩效费用的支付有一定的最低回报率 - 而基金经理的边际收入（变动性）将相应不同。

答案是，预期的绩效水平应包括绩效费用。有几个原因支持这一点。

● 绩效费用旨在能够实现。因此，预期的绩效范围不能低于绩效费用启动的水平。在绩效费用下的绩效水平计算的总体经济权益实际上将忽略绩效费用。尤其在考虑下面的情况时，这似乎不合适 - 即绩效费用激励基金经理利用其权力来影响其自身回报。

● 绩效费用的激励效应非常重要。IFRS 10要求基金经理是否有能力利用其权力来影响基金的回报。费用激励基金经理采取行动以获得此回报。因此，我们认为更重要的是费用的激励效应，而不是是否已经实现。

●● IFRS 10的示例忽略了障碍。这些示例指出，绩效费用是在障碍率之后支付的，但没有明确指定费率。从这个意义上说，似乎忽略了绩效费用的存在。这也与前面的几点一致。在这种基础上，我们认为标准及其示例不打算经常忽视绩效费。

关注绩效费用产生的水平也是一种实际方法。计算简单，不会对绩效费用上的确切水平敏感，假设费用只有一个费率/障碍。一旦费用已支付，基金经理在进一步更改基金表现方面的份额始终相同。如果有多个绩效费水平，那么需要更精确地确定建立在初始计算中的哪个绩效费水平（即确定预期的障碍水平，并将其建立在初始计算中）。更高的带宽存在可能仍然需要在总体经济权益的定性特征中考虑 - 见4.2。

如果预计在所有期间都不会达到绩效费用，那么这可能会作为总体评估的定性特征，在4.2中已经提到。

3、Kick-out rights

3.1The sliding scale of strength

Number of investors required to act together is key

It appears that the strength of kick-out rights may be assessed by categorising without-cause kick-out rights into several classes of strength. The number of investors that need to act together would be the main factor in determining those categories.

While the numbers of investors is most important, barriers to exercising the rights also need to be considered when determining the strength of a kick-out right.

3.1The sliding scale of strength

The second key indicator is kick-out rights, which are assessed on a sliding scale of their strength. We expect that the most significant factor in determining their strength will typically be the number of investors that would need to agree to exercise the rights. In other words, consider how many investors would need to act together.

The standard is clear that the greater the number of parties, the less weighting is placed on their kick- out rights. (At the other end of the scale, single party, without-cause kick-out rights always result in the manager being classified as agent. See Appendix 1.)

It is also clear from the standard that a complete lack (or zero weighting) of a kick-out right is not fatal to being an agent. IFRS 10 gives examples of such cases, some of which are agents and some principals. We explore these examples in Section 4.

It also appears that rights exercisable by an independent governing body (such as the board of directors) are considered at the board level, and are therefore implied to be very strong. This is because a board would generally comprise only a few members; once the investors’ rights have been transferred to the board members, it is, we believe, the number of board members that matters. For example, there may be hundreds of investors in a fund but only a very few directors; if those directors have the ability to remove the fund manager without-cause, then the kick-out right strength would be assessed at the board level. Naturally, the board needs to be independent, and there should be no other barriers to exercise that would weaken the force of that right (see 3.3).

Aside from a few private equity funds with very small numbers of investors, the lowest number needing to act together will usually be the members of a board of certain funds. For example, listed investment trusts usually have independent boards with the power to remove and replace the manager. As noted above, a board right is very strong.

Typically, in the retail fund sector there could be thousands of investors needing to act together, which would result in kick-out rights being considered very weak. In the private equity sector, funds with dozens or perhaps a few hundred investors would fill the ranges in the middle.

3.2With-cause rights have zero weight

The standard clearly indicates that kick-out rights should be without-cause if they are to be considered in the analysis; otherwise they are given zero weight. This is apparent because:

●● the rule about the single party removal rights requires them to be without-cause (see Appendix 1); and

● the standard gives two examples in which a number of investors have removal rights that are exercisable with cause; in both cases, it concludes that the rights are protective only and they are given no weight in the analysis.

3.3Barriers to exercising the kick-out right

As a general rule, IFRS 10 requires one to look only at substantive rights. This is confirmed specifically in relation to kick-out rights.

The standard provides a number of factors to consider in relation to this. The funds sector will recognise these factors as the types of matters considered in the past – for example:

●● financial penalties; ●● conditions narrowly limiting the timing of exercise; ●● absence of a mechanism allowing exercise; and ● inability to obtain the information necessary for exercise.

However, as with kick-out rights generally, it appears that these are not simple yes/no tests. So while some barriers might be so significant as to accord zero weight to the kick-out rights, others may simply weaken the sliding-scale strength of a kick-out right. If there are no barriers, then there is clearly no effect at all on the strength.

Insight – Occasions for casting votes

It appears that the lack of an annual general meeting (AGM) may be seen as weakening the investors’ practical ability to exercise kick-out rights, unless there is another mechanism in place.

Private equity investment funds might have no AGM, but usually have instead alternative mechanisms that give investors the practical ability to exercise kick-out rights. For example:

• ●● there may be quarterly calls and annual meetings at which the investors are provided with information;

• ●● an investor will usually necessarily know who the other investors are – e.g. they will be named in the partnership documents available to all investors; and

• ●● there could be a right to call an extraordinary meeting – often requiring just five percent investment – to vote on removal.The ability to call such a meeting and exercise a right may even be stronger than awaiting an AGM in order to exercise a right.

As a result, the lack of an annual voting opportunity does not necessarily require zero weight to be accorded to kick-out rights.

Insight – Notice periods**

Notice periods can also represent barriers to exercising kick-out rights. It appears that a notice period covering the whole of the life of a limited life fund (or penalties equivalent to total fees for the life of the fund) would result in zero weighting of the kick-out rights; however, other cases may not be so clear cut.

For example, a typical arrangement in the private equity sector is to give six months’ notice plus payment of a further six months’ fees – in effect a combined one-year notice/fees period. However, at the outset the fund life will be five or seven years. It appears that the effect on the strength of the kick- out right is material, but not so great that the kick-out right has no weight.

4、Combining the two key indicators

Matrix approach to assessment

It may be appropriate to use a matrix, grouping together categories of kick-out right strength. For each category, the fund manager could identify a level of aggregate economic interest for which the entity is clearly an agent and another for which it is clearly a principal. A marginal zone would then appear between these points.

If a combined assessment of the key indicators mapped into the marginal zone, then other features of aggregate economic interest and kick-out rights would be taken into account.The other features to consider include expected magnitude, a second tier of performance fee, the proportion of investors required to vote together and nature of variability compared with that of other investors.

4.1The challenge is finding appropriate combinations

Stronger kick-out rights favour a conclusion of agent, and a larger aggregate economic interest favours a conclusion of principal. The basic relationship between these two indicators is easily understood. The challenge is in determining how to apply the relationship to specific cases and conclude whether the fund manager is principal or agent.

The standard (as well as the Effect Analysis3) provides examples of combinations of kick-out rights and aggregate economic interest, as shown below. From these examples, we can see some fixed points about certain combinations and derive a general approach.

Although these examples do not provide any bright lines, they do narrow the areas in which significant judgement is required to operationalise the model. By presenting them in the format below, the concept of a marginal zone between agent and principal becomes clearer. We do not suggest that such a chart can be used as a quantitative tool, but merely that it illustrates the basic idea.

When no kick-out rights exist, then, deriving figures from IFRS 10’s examples, the changeover from agent to principal occurs between approximately 22 and 37 percent (the marginal zone between Examples 14A and 14B).

When stronger kick-out rights exist, a higher variability might still support an agent outcome. For instance, Example 14C has without-cause, board-level, kick-out rights (a very strong case) and can withstand 37 percent aggregate economic interest, while still being deemed an agent. In Example 14B, which has the same variability but no kick-out rights, the manager is deemed a principal.

These examples could be used as a guide to develop a matrix. For each category of kick-out rights strength, zones of variability of aggregate economic interest would be identified, to correspond with clear agent, clear principal and marginal cases. Such a matrix could be presented as follows.

The cells in this matrix have been left intentionally blank, aside from the suggestion of the no-kick-out right zones based on Examples 14A and 14B. As there are no bright lines, it appears that fund managers would need to judge and justify their own assessments of zones.

当不存在罢免权时，根据《国际财务报告准则第10号》的示例，从代理到负责人的变化发生在总体经济权益的约22%到37%之间（在示例14A和14B之间的边缘区域）。

当存在更强的罢免权时，更高的变动性仍可能支持代理的结论。例如，示例14C具有无因罢免的董事级罢免权（非常强大的情况），并可以承受37%的总体经济权益，仍被视为代理。而示例14B具有相同的变动性，但没有罢免权，基金经理被认为是负责人。

这些示例可以用作制定矩阵的指南。对于每种罢免权强度类别，将确定总体经济权益的变动性区域，以对应明确的代理、明确的负责人和边缘情况。这样的矩阵可以呈现如下。

4.2Analysing marginal cases

If a case falls within or along the edges of the marginal zone identified above, then a reasoned judgement as to whether the fund manager is a principal or an agent would be required. It appears that other features of the two key indicators would be considered at this point – for example:

• ●● the expected magnitude (see 4.2.2);

• ●● the likelihood of the expected performance level applied not being reached in all periods (see 4.2.1);

● a performance fee having a second tier applying above the expected performance level, which would increase maximum variability; while the standard asks for the assessment to be made primarily on the basis of variability at the expected return, it requires that the maximum also be considered;

●● the proportion of investors required to vote together to use a kick-out right – i.e. majority or super majority, with the latter being more difficult to achieve; and

●● any difference in the nature of the variability as compared with that of other investors, if this was not enough on its own to result in an outcome of principal (see Appendix 1).

When it is concluded that a fund manager is the principal as a result of seeding a fund (high co- investment), the interaction with IFRS 5 Non-current Assets Held for Sale and Discontinued Operations would need to be considered. That is outside the scope of this publication.

4.2.1Non-cumulative performance fees

In some fund structures, the performance fee is cumulative over the whole life of the fund – e.g. in private equity. In other cases, the fee is determined and paid based on annual performance with no clawback in subsequent years if cumulative performance falls.

It appears that the possibility of an annual, non-cumulative fee not being due may occasionally be a qualitative factor that would be taken into account in deciding marginal cases. If a fee has been designed to be due only occasionally, then this may, in marginal cases, weight the scales towards being an agent. This is because it cannot be factored in quantitatively without converting the variability measure into a figure that does not represent the actual marginal rate at which the fund manager benefits/suffers as performance changes. For example, a probability weighting across several years would become a blended figure.

Such a blended figure would inevitably incorporate cases of a zero performance fee, effectively partially disregarding the fee. It appears difficult to view such an approach as consistent with IFRS 10. The standard is asking for the size of the fee to be assessed in order to decide whether the fund manager has the ability to use its power to affect its own returns from the fund. The fee exists to incentivise the fund manager to behave in a certain way in order to achieve the fee. Therefore, we believe that it is the incentive effect of the fee, not whether it is achieved, that is the more important feature; this conclusion is supported by the fact that IFRS 10’s examples seem unaffected by questions of the possibility of not meeting the hurdle (see 2.4). Accordingly, a blended figure seems difficult to view as consistent with IFRS 10.

Reverting to qualitative matters, it should not be overlooked that performance fees may have a somewhat different nature from the other investors’ interests (see Appendix 1), adding some weight in favour of principal. For example, the existence of the hurdle makes the fund manager’s return at that point more sensitive to changes in fund returns than are those of the investors.

4.2.2Considering the fund manager’s expected magnitude

One of the other features of the aggregate economic interest referred to in IFRS 10 is its magnitude. IFRS 10 asks for an evaluation of the “magnitude of ... its economic interest ... relative to the total variability of returns of the investee. This evaluation is made primarily on the basis of returns expected from the activities of the investee”.

Magnitude is a feature of the aggregate economic interest that can help decide a marginal case. This is because it appears that IFRS 10 requires both variability and magnitude to be sufficiently high to conclude that the manager is a principal. This is because the standard notes that, “the greater the magnitude of, and variability .... the more likely the decision maker is a principal” (emphasis added). So if magnitude is low, then it may tip the scales for a marginal case to be an agent, and vice versa.

Turning to the calculation of the expected magnitude, it appears that this is the fund manager’s total return as a percentage of the fund’s total return measured at the expected performance level (see 2.4). This is because it has to be different from variability, yet still be a measure of fund manager returns relative to fund returns. In other words, while variability is the marginal share of return, magnitude is the absolute share of return.

If magnitude were calculated at low performance levels, then very high magnitude figures would result. Continuing with the above example, if fund performance were 1 percent, then the fund manager would take 100% of it through the management fee, leaving nothing for investors.

如果在低绩效水平上计算量级，那么很高的量级数据将会产生。继续上面的例子，如果基金的绩效为1%，那么基金经理将通过管理费获得100%的绩效，没有留给投资者。

Fortunately, IFRS 10 asks for magnitude to be assessed at expected levels of return (the same as the expected performance level at which variability is measured, see 2.4); it appears that these are at much higher levels, at which such extreme measures do not usually arise. In fact, as returns increase, the magnitude will tend to pull towards being to the same figure as variability. This is because the distortion of the average, by management fees (fixed component) as a percentage of low level performance, is gradually less and less significant as performance increases. In its place, the marginal aggregate income rate increasingly dominates the calculation.

幸运的是，IFRS 10要求在预期回报水平上评估量级（与在其中测量变异性的预期绩效水平相同，见2.4）。在这些水平上，通常不会出现如此极端的情况。事实上，随着回报的增加，量级往往会趋向于与变异性相同。这是因为通过管理费（固定部分）作为低水平绩效的百分比，对平均值的扭曲逐渐变得不那么重要，随着绩效的提高，这种扭曲会变得越来越不重要。在这种情况下，边际聚合收入率越来越主导计算。

It is also worth noting that this relationship between variability and magnitude means that, in practice, it appears that magnitude does not need to be considered outside the marginal zone even though, we believe, IFRS 10 requires both variability and magnitude to be sufficiently high to reach a conclusion of principal. Beneath the marginal zone, this is because if variability is sufficiently low, then it does not matter what the magnitude is. Above it, this is because if variability is sufficiently high, then magnitude calculated at the expected performance level will usually also be sufficiently high (i.e. magnitude moves towards variability as returns increase).

值得注意的是，变异性和量级之间的这种关系意味着实际上，在边缘区域之外，似乎不需要考虑量级，尽管我们认为IFRS 10要求变异性和量级都足够高，才能得出主体的结论。在边缘区域之下，这是因为如果变异性足够低，那么量级是多少就不重要。在边缘区域之上，这是因为如果变异性足够高，那么通常在预期绩效水平上计算的量级也通常足够高（即随着回报的增加，量级趋向于变异性）。

5、Conclusion

This publication identifies what we think at this stage are the two key indicators that fund managers will need to consider in assessing whether they are a principal or an agent under IFRS 10. It also provides some comment as to how these indicators could be operationalised.

The following table summarises the key steps in operationalising the assessment. Overall the analysis highlights that while significant judgement cannot be avoided, nevertheless an initial assessment can be focussed on a more simplified set of indicators than might at first be apparent from the standard’s complexity. This will leave a more manageably sized population at the margins that requires significant judgement.

Measure aggregate economic interest	The fund manager needs to identify a key measure of aggregate economic interest. It appears that this measure is the variability of expected returns, measured at performance levels at which a performance fee kicks in.
Assess strength of kick-out rights	It appears that the strength of kick-out rights may be assessed by categorising without-cause kick-out rights into several classes of strength. The number of investors that need to act together would be the main factor in determining those categories. While the numbers of investors is most important, barriers to exercising the rights also need to be considered when determining the strength of a kick-out right.
Assess key indicators in combination	It may be appropriate to use a matrix, grouping together categories of kick-out right strength. For each category, the fund manager could identify a level of aggregate economic interest for which the entity is clearly an agent and another for which it is clearly a principal. A marginal zone would then appear between these points.
Significant judgement still required at the margins	If a combined assessment of the key indicators mapped into the marginal zone, then other features of aggregate economic interest and kick-out rights would be taken into account.The other features to consider include expected magnitude, a second tier of performance fee, the proportion of investors required to vote together and nature of variability compared with that of other investors.

Appendix 1: Why some tests/indicators are not key in practice

The main part of this publication discusses the indicators of the model that we think will be most relevant to the funds sector. This appendix addresses the other components of the model and why we do not consider them to be practical focus areas.

As shown in the diagram below, the model has only two determinative tests. One is positive, relating to kick-out rights; the other is negative, relating to off-market remuneration. If neither of these tests has resolved the case, then the fund manager considers four indicators in making its assessment: scope of authority; kick-out rights; remuneration; and other interests. Remuneration and other interests held (the ‘aggregate economic interest’), should also be considered together. These effectively become a fifth indicator.

BC126 《征求意见稿第 10号》 建议基于以下假设识别代理关系:

(1) 主要责任人与代理人均力图实现双方的最大利益。因此，主要责任人很可 能引入额外措施以确保代理人不违背主要责任人自身的利益。例如，主要责任人可能 有权利有条件或无条件罢免代理人。

(2) 主要责任人没有动机支付代理人多于与其服务相称的报酬。因此，与所提 供服务不相称的报酬可能表明该决策者并不是代理人。

The two determinative tests, shown in green above, are clearly important. However, as explained below it appears that they will not often be of practical assistance. Of the main group of indicators, those shown in red above are the key ones addressed in the main part of this publication. As explained below, the remaining, shown in grey, appear to be of limited practical relevance.

The positive determinative test – single party kick-out rights: rarely met

A decision maker is an agent without further analysis when:

a single party holds removal rights;

1、those rights are substantive – i.e. they give their holder the practical ability to prevent the decision maker from directing the activities of the investee; and

2、they are exercisable without-cause.

It is clear that this determinative test is met only when a single party holds removal rights. If two parties are required to agree to exercise removal rights, then “those rights are not, in isolation, conclusive”. A single party, however, does not necessarily mean a single entity. It appears that removal rights held by several entities under common control would be regarded as held by a single party.

The standard indicates that rights exercisable by a board of directors or equivalent body are not equivalent to rights held by a single party. However, rights held by an independent governing body are implied to be very strong, as discussed in 3.1.

In our experience, it is rare in practice for a single party to hold substantive, without-cause removal or similar rights.

The negative determinative test – at-market remuneration: rarely failed

For the decision-maker to be an agent, its remuneration needs:

1、to be commensurate with the services provided; and

2、 to include only terms, conditions or amounts customarily present in arrangements for similar services and level of skill, negotiated on an arm’s length basis.

If there is sufficient investment from independent investors, then it appears that the remuneration is generally at market. This is because arm’s length investors should have transacted at arm’s length terms. In order to rely on such an analysis, the outside investors have to hold a material interest. They also need to be genuine third parties: their interest should not be linked to and compensated by, or given as compensation for, some other transaction.

This approach means that there is no need to benchmark fee rates against peers. In any event, it could be difficult for some funds to find comparable funds – for example, due to their specific structure/ strategy.

**Note also that the standard’s focus is on whether the remuneration was negotiated at an arm’s length basis. So if the fees were fixed in such a way some years ago, then it does not matter that market norms for new funds have now changed.

Scope of decision-making authority: not a distinguishing feature

When assessing the scope of its decision-making authority, the manager considers:

1、 the activities that are permitted according to the decision-making agreement(s) and specified by law; and

2、 its level of discretion.

If a case is being assessed for linkage, then it must be the case that one has already concluded that the fund manager has power. That being so, it appears that this indicator will always give the same result for a fund manager. That is, it will generally always be assessed that the fund manager has all of the discretion within the designed activities of fund. Therefore, the fund manager will inevitably have power over the decisions that significantly affect investment returns (the relevant activities).

After all, the whole purpose of investing in a fund is to delegate decision-making, over matters that significantly affect the investment returns, to an expert fund manager who, it is expected, will generate superior returns or, for trackers, reliably close-tracked returns at the lowest cost. Therefore, the fund manager will inevitably have power over the decisions that significantly affect investment returns (the relevant activities); this is regardless of whether the fund’s range of activities is broad or narrow. Indeed, if it did not have power over the relevant activities, then the principal-vs-agent linkage analysis would not be necessary, because without power the IFRS 10 definition of control cannot be met.

We do not think that this necessarily means that the fund manager is a principal. If it did, then all fund managers would be principals. The purpose of the linkage test is to decide whether power is deemed to be used for oneself (in which case the fund manager is a principal) or for others (in which case the fund manager is an agent). As a result, it appears that this indicator is not a distinguishing feature in the principal-vs-agent analysis. In effect, it may be left out of the analysis.

In November 2011 the FASB5 issued an exposure draft on principal-vs-agent evaluation. It is directionally consistent with the conclusions in IFRS 10. It is interesting to note that its exposure draft does not retain ‘scope of decision-making authority’ as a factor to be evaluated in assessing whether a decision maker is a principal or an agent. It appears that being a decision maker in the first place has been looked upon – or this is the inference – as having the scope of decision-making authority, making that indicator redundant as a distinguishing feature.

Remuneration: always counted, but no need to count it on its own

The standard requires the fund manager to assess its level of remuneration. The larger the fund manager’s at-market remuneration, the more likely it is that the fund manager is a principal.Two observations of principle arise from this.

The standard requires the quantum of remuneration to be tested twice: once for remuneration alone, and a second time as part of the aggregate economic interest. Therefore, it appears that in practice the test need only be done once, at the aggregate economic interest level. This is because if the level of remuneration alone resulted in a principal conclusion, then the same conclusion would be reached when assessed in aggregate. Conversely, if the assessment did not result in a principal assessment on a stand-alone basis, then the remuneration would still be required to be tested as part of a larger aggregate.

Other interests: always counted, but no need to count them on their own

If a fund manager holds other interests in the fund – e.g. a co-investment – then this may be an indication of its being a principal. However, this is not a yes/no test. It depends on the quantum of the interest, although no particular lower limit for consolidation is set. The standard asks the fund manager to assess its exposure to variability of returns from its other interests.

The standard requires the quantum of other interests to be tested twice: once alone, and a second time as part of the aggregate economic interest. Once again, in practice it appears that the test need only be done once at the aggregate economic interest level, for similar reasons as stated above in respectof remuneration.

Strong divergence of fund manager’s and investors’ interests: rare

If the aggregate economic interest of the manager is different from that of other investors, then this indicates that the fund manager is a principal. In such a case, the fund manager is incentivised to use its power for itself and not for other investors, because the incentive scheme does not align the fund manager’s and the investors’ interests.

The standard is clear that the fact of at-market remuneration is not sufficient on its own to conclude that the fund manager is an agent. This may be one of most unwelcome points for the investment funds sector. However, that is what IFRS 10 requires.

It follows that at-market remuneration can be too large for the fund manager to be an agent. This is because remuneration of the fund manager, which varies depending on the fund’s performance, is a form of return; the more significant this remuneration, the more likely it is that the fund manager is primarily using its decision-making authority over the fund for its own benefit and not for the benefits of others.

IFRS 10 gives an example from outside of the funds sector: a multi-seller conduit where the decision- maker has a subordinated residual interest and provides credit enhancement. It concludes that the decision-maker is a principal by reference solely to the different nature of this interest, i.e. regardless of all other factors. We expect that such an arrangement is unlikely to be common in the funds sector.

The sector may occasionally, however, have some fee structures that may not be closely aligned with the investors’ interests – for example, if performance fees from one year cannot be clawed back if cumulative performance to the end of the next year falls but the investors cannot realise any gain until the end of the fund life. An example of more significant divergence from investors’ interests might be a geared property fund. If management fees are based on the gross value of the properties, with the manager able to gear up the fund, then there is clearly a difference between the variability to which the geared investor is exposed and that to which the gross-basis management fee is exposed. Such structures were seen before the crisis but are not favoured by investors nowadays.

It appears that, if the variability to which the manager is exposed is strongly different from that to which other investors are exposed – e.g. the geared fees cases – then this might be a sufficiently strong indicator to conclude that the manager is a principal. This is based on IFRS 10’s example (noted above), where the conclusion is for principal on the strength of a different interest alone.

If, on the contrary, the variability to which the manager is exposed is only a little different from that to which other investors are exposed, then it appears that this is taken into account after combining aggregate economic interest and kick-out rights (see 4.2) – i.e. it might be enough to tip a marginal case to being a principal.

原文链接