Manager Research
all members of our manager research team have over ten years of investment experience.
The Orbis fee structure has been widely debated over time as investors have weighed up what appears to be an expensive fee, with what has in practice been a significant level of value add, totaling almost 6%[1] additional return per year for 30 years. Where alternative funds have charged significantly less, they have also delivered less for the most part, and have also not demonstrated the same continuity in terms of process, culture and business stability. While Orbis’ investment process is not designed to deliver a steady outcome, the business which underpins their capability has been exceptionally well managed which has allowed them to take the contrarian positions they need, such as avoiding tech shares during the Dot Com bubble in 1999/2000. The chart below shows their performance cycle of adding value since 1990, which has averaged 4% net of fees since inception, but is currently negative after fees.
Source: Orbis; FE Analytics; Fundhouse
Through various cycles they have been able to add value given the flexibility inherent in their approach. They have performed well when markets are contracting or expanding, unlike a more typical value-biased manager which is more cycle dependent. The combination of a strong contrarian approach, with high levels of research insight have delivered the results over time. Over the past ten years, as the GFC has driven certain asset prices higher, Orbis has lagged given their tendency to avoid overpaying. We would expect a better relative outcome looking forward.
During our last Orbis review in 2019 we noted that there had been some significant changes made within the team and process which is quite unusual given their propensity to avoid making structural or organizational changes. This after a period of relative weak performance (the first time in 30 years they had failed to add value over the market over a 10-year rolling horizon).
The main changes were the shortening of the portfolio manager list which in part drove a significantly higher concentration of shares compared to the past few years. Historically their holdings had averaged around 80 shares, however by 2014 this had drifted up to over 140 holdings. Today this number sits just below 60 shares, indicating a much more concentrated and higher conviction portfolio. With the smaller management team, the focus is purely on idea generation and competition for a place in the portfolio – the strongest ideas win, and this is at the heart of their performance culture.
In short, we retain a Tier 1 rating for the Orbis Global Equity process and expect them to continue to add value over time.
The question now is, at what cost?
Fee changes being made in May 2020
The nature of performance related fees is quite complex, so we have attempted to simplify the assessment below by stepping through the various elements of the fee. Fees are also complicated by administration charges, and in this case there is significant impact due to embedded administration costs.
To outline the changes, at a high level:
| Item | Details | Current | New |
| Base Fee (to Orbis) | Fixed % fee irrespective of returns, charged on fund total assets, paid to Orbis | 1.00% | 0.80% |
| Unit Trust Administration Fee[2] | Fixed % fee irrespective of returns, charged on fund total assets. | 0.50% | 0.30% |
| Performance Fee | Variable % fee, charged if Orbis adds value (or refunded, if they lose value) | 12.5% of value add (or value lost) | 25% of value add (or value lost) |
| Measurement Period for Performance Fee | Horizon over which the performance fee is measured and charged | 3 years | Weekly |
|
Fund Costs (audit, bank, trustee, legal, regulatory, etc) |
Operating costs to maintain the fund operations and governance. | Average +- 0.06% per year of fund assets. | |
|
Fee Minimum |
0.50% | No Minimum | |
|
Fee Maximum |
2.50% | No Maximum | |
|
Measurement Benchmark |
For the performance fee calculation | FTSE World (all developed markets) | MSCI World (all developed markets) |
| 1. | Absolute fee level and the expected proportion of client return paid to the manager. |
|
| a. | At 0.80% this would be near the top end of a typical global equity fund of this nature (0.60% to 0.80% for equivalent classes with an average of 0.75%), but not outside of the bounds of what is a reasonable fee, excluding any value for money/merit argument (this should be in the performance fee). So, while it’s a bit high for the base, its not exceptional. It does need to be contrasted with pure passive at +-15bp, so there is a ‘fee’ of 65bp paid for optionality on value add by Orbis. One could argue that this 0.65% premium should already compensate Orbis for the potential value add they may bring. This base fee level as a result is relatively high considering there are additional performance fees, and would need to go hand in hand with a strong active process to warrant any level of consideration. |
|
| b. | The 25% share rate appears relatively high on the surface, but given you need to earn back the base fee plus administration cost of 1.1% (0.80% + 0.30%) before Orbis accrues a performance fee, it’s actually lower than it appears. If you take the expected ranges of value add plus the base fee as a % of the total expected return, you get to a range of 10-15% of the total return paid to Orbis for the Global Equity fund. With the performance fee on top of the ‘option’ fee above, it does get quite expensive. However, given there is a new “Refundable Reserve Fee”[3] pool, this enables the sharing rate to be fully symmetrical whereby underperformance is refunded to investors out of this pool. In the case where the pool is exhausted and refunds can’t be made, the refund is accrued until such time as it is earned back. So, provided investors remain invested, it is equal sharing of out- and underperformance. |
|
| c. | For value add of zero before fees (as they are experiencing currently), the total fee is 0.94% (including the 0.30% administration charge, ie 0.64% for fund management) calculated as 0.30% (administration) plus 0.80% (fund management) plus 25% sharing rate multiplied by (0% value add less the 1.1% base fees)[4] . This is probably acceptable for most investors given they are investing with the hope of value add. |
|
| 2. | Administration costs: we have quite a big issue here: there is a 0.30% charge irrespective of whether you are a small individual investor (e.g. direct to Orbis) or whether you have a single bulk account (e.g. a large fund administration platform). This makes it very expensive, particularly platforms other than Allan Gray. You effectively pay 0.80% for fund management and 0.70%[5] for administration as an investor through a third-party platform, before performance fees). In addition, most managers have this administration charge embedded in the fund manager fee (the 0.80%), not charged separately. It does seem to be unjustifiably high, particularly where the fund is accessed through a third-party platform. |
|
| 3. | Embedded costs (fund costs, bank charges, regulatory, etc) |
|
| a. |
Current costs at 0.06% are reasonable and we don’t expect these to change materially.
|
|
| 4. | Where performance fees are applicable, we test: |
|
| a. |
Symmetry. Does the client pay for performance fees, and receive an equivalent discount when the manager underperforms? Yes, in this instance they do, provided they stay invested to be able to recover any underperformance they may have experienced. This is a strong feature of the performance fee, and one we often see skewed to the fund manager in other cases. The table below shows expected fee levels for value add: |
|
|
|
||
| b. |
Performance Fee Measurement Benchmark.This is currently an issue for us, albeit relatively minor. Orbis invests with an unconstrained mandate, investing across both Developed and Emerging Markets (DM and EM). They have now changed the benchmark to MSCI World from FTSE World, which are both Developed Market benchmarks. With significant holdings across Emerging Markets we don’t follow the logic of choosing this as a comparative, given the potential misfit over time. This is also more important going forward than in the past, given the prevalence of China in the All Countries (DM and EM) index and its increasing prominence and exposure. An additional smaller change is the move from a gross of tax index, to the net of withholding tax index which typically lags the gross of tax index by 0.5% per year. However, this is in line with what investors receive in the fund and is common practice globally. |
|
| c. |
Time horizon.While the weekly calculation is good for ‘pay for what you get’, it also has the potential to be significantly more volatile. Clients will experience a much more volatile fee ‘ride’ than they have had in the past. This is a function of the improved symmetry and lower base fee as a proportion of the value add. Our experience here is that the manager will get questioned when fees are high, but will not get reciprocation when fees are low, leading to potential client dissatisfaction. This is not a real issue however, provided investors stay invested, the symmetry of the fee works in clients’ favour. |
|
Overall Conclusion
The new Orbis fee was not driven by a need to lower fees; rather it was driven by two intentions: to make the fee fairer to clients when value add was lower or negative, and to make the fee more reactive to relative performance. This does result in a lower fee though, on average. Should Orbis deliver the same 6% per annum total value add as they have done since 1990, the expected fee to investors will be largely the same, just earned in a way which balances the interests of Orbis and the investor more fairly along the way. Should Orbis deliver zero value add over time, the investor will pay a fee not too dissimilar to what they will pay for the average active fund in the market.
Given this range of outcomes, the Orbis fee will remain a divisive topic, however we can take these changes as improvements to the original structure. As investors we then need to consider the Orbis investment proposition relative to their expected value add and allocate accordingly. We do believe there is a role for this in investors’ portfolios, where the ultimate payoff is determined by Orbis’ ability to add value, which we rate highly, as well as the set of alternatives, of which there are few.
Implications for local Allan Gray Funds
- Local funds will access the 1.1% fee class (which includes the 0.30% unit trust administration fee) with the performance fee added over and above. Through all third party platforms (ie not via Allan Gray) there will then be the usual platform administration charge in addition to the 0.30%, which makes the fee structure relatively expensive on these platforms.
- There are some minor differences in terms of how the new Orbis fees are charged within Allan Gray Balanced and Stable funds:
- The measurement period will reduce from three years to two years to align with the Allan Gray fee structure, rather than the weekly Orbis period.
- There will be a fee cap and a floor remaining as per the current structure. This also helps align fees where Allan Gray allocates poorly (value negative) but Orbis is performing well (and there is an underlying performance fee charged).
- The local feeder funds (Allan Gray Orbis Global Equity Feeder and Allan Gray Orbis Global Fund of Funds) will adopt the new 1.1% class as well. Note that it is likely the Fund of Funds will transition to a Feeder fund in time (to Global Balanced, launched in 2012 and currently rated Tier 2.)
[1] Before fees are charged
[2] For direct and 3rd party platforms, this 0.30% is charged within the fund fee. Through the Allan Gray platform, the 0.30% is charged as the administration fee. To then access the fund on 3rd party platforms, their additional administration fee is levied.
[3] This RRF pool is funded through performance fees earned by Orbis, but as they only pay out a third of fees earned per year, this leaves the balance of two thirds available to refund investors for future underperformance.
[4] There is a Dividend Withholding Tax Adjustment, and the calculation is Geometric
[5] We have assumed an additional 0.40% administration fee charged by the 3rd party platform in this example
Orbis Fee Changes
31 Jan 2020