Returns-chasing investors value top-performing funds and prioritise them for their investment allocations. But what does ‘top-performing’ actually mean? The naïve answer is that this is self-evident – the funds with the highest rolling returns over a given period – but this concept is quite misleading.
Because investors are all targeting a different future point at which their accumulated funds will be converted into cash and redeemed from the fund, commentating on which fund is ‘ahead’ at a given month, quarter or year-end is of limited relevance, unless that date coincides with an investor closing out their investment. At that point, the investor could assess whether they had chosen wisely or otherwise.
Ranking investment funds as if they were realising their accumulated returns at a specific point in time is no more than a ‘convenient fiction’. It satisfies a human need for estimation, but is of almost no practical use and can frequently lead to mistaken assumptions about both the risk undertaken in achieving a particular reported investment return and likely future performance.
Here are some crucial facts about reported/ranked investment performance, and six important warnings about reading too much into investment performance ranking lists.
1. Investment timeframes
Performance surveys include multiple timeframes but, in our view, only the three and five year+ timeframes contain any meaningful performance information. Periods of two years or less are problematic for several reasons, including insufficient exposure to several stages of the economic cycle, significant elements of market noise, luck and style, and base effect distortions due to unit pricing timing or intra-year cost changes. Valuation factors and fee levels are also issues. Multi-manager portfolios are inherently built for their diversification benefits, so their relative advantages will almost certainly never be visible over less than three years.
Periods as long as 10 years are also problematic. While over long time frames markets should have passed through all potential phases and any statistical anomalies resolved, there are still a few issues such as the small pool of New Zealand fund managers possessing a 10-year track record. Changes within that decade in terms of investment style and staff can also have a huge impact on the prospective outlook for a fund. In addition, the fact that 10 years is the minimum intended holding period for growth and aggressive funds makes it unlikely that impatient investors will wait for the full recommended term before making a decision to dis-invest – particularly when market volatility surges as in the GFC.
Morningstar’s New Zealand funds survey runs its returns comparison only out to seven years, and beyond that (for some funds) is listed a ‘since inception’ time horizon. This reflects a lack of continuous, consistent performance history from a sufficiently large number of investment managers. Ten year returns therefore still remain outside an appropriate performance period from a New Zealand investor’s point of view.
Allowing instead for a five year returns history as the ideal, it is striking how widely spread are measured returns:
The dispersion of returns shown above appears to suggest an equally wide range of skill levels on the part of managers, but it mainly reflects the asset mix within these five broad fund categories. Both the asset mix variation and the impact of returns sourced from actively managing the fund contribute to the investment outcome. The predominant share of the raw return over any given period derives from general market movements for the assets within the fund. This would not create a problem, were it not for the fact that returns differences are regularly masked behind the fairly crude categorisation schemas used in ranking summaries.
2. Holdings variations in funds lumped together for convenience
There is a wide range of growth assets within the conservative, balanced and growth fund types. In periods when growth assets perform well, returns for some funds are likely to be higher just because of larger holdings of growth assets. But if growth assets then perform badly, returns are likely to be lower. So when comparing the returns of two funds within a fund type, it is important to take into account what each fund holds.
Variations in the definition of fund category are large enough to allow asset class holding variations to swamp skill assessments within the respective categories. For instance, a balanced fund manager who is only allowed a maximum to 50% in the growth asset classes will have a hard time beating the gross returns level of a manager with a cap of 70% ̶ regardless of the first manager’s ability to add value to the benchmarks in their fund.
3. Category migration and cloning
The periodic phenomenon of category migration is not just theoretical, and distorts comparisons over time. For instance, 2017 saw several New Zealand funds that were previously classified as ‘aggressive’ in the Morningstar survey migrate to the ‘growth’ category. Survivorship bias can also be an issue – the closing or market exit of funds that contributed track records in the past but have subsequently disappeared without due statistical adjustment.
It may be that the survey producer is willing to move funds from one category into another (and back again) depending on the proportion of their assets held in a given risk category. However, the question arises with these migratory funds: do they take their track records along with them when they ‘migrate’ between categories? Or is their longer term positioning within the new category adjusted for the fact that earlier that fund had a different asset mix?
An associated problem with the main New Zealand performance surveys is that they have allowed for several instances of what we call ‘cloning’. This takes place when a fund manager splits a near-identical strategy into multiple vehicles and contributes a clutch of funds to the surveys whose performance may only be a few basis points apart. That has the effect of ‘crowding the quartiles’ or causing other managers’ funds to be pushed into a different performance position, due to variants of what is, in effect, only one underlying strategy.
The chart below shows the uptrend in the total funds appearing in the Morningstar New Zealand survey, which includes new entrants, but also the impact of significant fund ‘cloning’ by some providers.
4. Return (not risk-adjusted return) is what makes the league tables
Investors do not sufficiently account for the volatility which funds have undergone in earning historical returns, and are likely to confuse risk-taking with skill. This can lead to an over-allocation of capital to lucky past winners, and to excessive risk-taking by fund managers to attract inflows. The corollary may be an incentive to maximise risky asset exposure during good times, on the assumption that in a major bear market or crash scenario everyone will be hard-hit ̶ so prior prudence will not be adequately rewarded. Such an outcome, however, is rarely the case.
At present, ten years after the GFC, because of the strong multi-year performance of the New Zealand equity market, which has taken it to extremely highly valued levels by historical standards, there is a crowded trade getting underway where some managers are concentrating their fund exposures to the domestic market. However, history indicates that this is inevitably a risky undertaking.
5. International equities’ dominance and variable hedge ratios
Note should be also made of the impact of international equities on overall portfolio returns. International equities tends to be either the largest or second-largest allocation in funds with a composition classification of ‘balanced’ and above, so have a huge impact on funds’ comparative returns. This asset class’s returns profile is notable in New Zealand for usually being only partially hedged into New Zealand dollars (NZD).
At times wide differences can show up in the short term between hedged and unhedged asset class returns accruing from international equities. A given manager’s exposure to these returns streams in their funds could well be differentiated from their manager peers not by the returns in the underlying equity markets, so much as by their varying hedging levels. Over time, currency effects tend to equalise, but this longer-term resolution of differences between hedged and unhedged returns will not be visible short-term.
6. The problem of persistence – concluding thoughts
International research over many years has established that, in all probability, an investment fund is doing well if it outperforms its benchmark for any multi-year period. The majority of funds in most markets do not do so – mainly for statistical reasons founded in the randomness in market returns, and the cumulative impact of transaction costs. It is even more remarkable if such a fund outperforms its benchmark on an after-fees or after-costs basis.
Nevertheless, there is a set of funds that do beat their benchmarks, their index-fund competitors and even other active funds at times. The best an active asset manager can do is identify funds which can fairly be called ‘high-quality’ and that offer a greater comfort that decent returns will be achieved with acceptable regularity, when compared with lower-quality fund structures. Yet naïve results tabulators persist in presenting unadjusted raw returns as a guide for investors. We believe the complicating factors need to be recognised, and advise against using ranking lists.
This blog post has been prepared to provide general information and does not constitute financial advice in accordance with the Financial Markets Conduct Act 2013. An individual investor should, before making any investment decisions, consider the information available in the relevant Product Disclosure Statement and seek professional advice. While every care has been taken in the preparation of this document, AMP Capital Investors (New Zealand) Limited and the AMP Group (together, 'AMP') make no guarantee that the information supplied is accurate, complete or timely and do not make any warranties or representations in respect of results gained from its use. The information is not intended to infer that current or past returns are indicative of future returns. The views expressed are those of the author and do not necessarily reflect those of AMP. These views are subject to change depending on market conditions and other factors.