Agency Costs on the Investment Performance of Australian Pension Funds Finance Essay

In Australia, the superannuation industry has undergone a rapid expansion since the introduction of a compulsory pension system in 1992. Superannuation assets grew from $245 billion to 942 billion in 10 years (Benson et al., 2010). Investors and consumers now have vast choice in their ability to pick a pension fund, ranging from non-profit funds, such as corporate, public sector and industry superannuation funds, and for-profit retail funds. However, there was a dearth of empirical studies which aimed to clarify any systematic differences in the performance of these respective funds. This is an important policy consideration as even small differences in cost and returns can have large remuneration ramifications in the long term.

Bateman (2001) identified that a 1% change in costs could lead to a 22% change in returns on retirement. As such, it is crucial to analyse potential differences in returns between fund types and ensure retirees maximise their retirement income savings as performance ultimately affects welfare (Clark and Unwin, 2008). Systematically lower net returns will reduce final retirement savings. This can increase reliance upon the public pension system than would have otherwise been required under lower cost and higher return funds. In Coleman et al.’s (2006) article, the authors utilised a large cross-sectional database which encapsulated not-for-profit and for-profit funds, elaborating on prior studies which tended to only consider for-profit providers. The article aimed to gain understanding of the performance of the Australian superannuation industry from 1997 to 2002, evaluating in terms of risk, returns and expenses, and what factors in particular were the determinants of investment performance. The core approach to these questions adopted a fund type comparison to enable the authors to ascertain whether the varying characteristics between not-for-profit and for-profit affected fund net return and performance. Despite the adoption of a ‘prudent person’ approach according to the Superannuation Industry (Supervision) Act 1993, the driving force behind potential disparities was hypothesised to be agency costs which manifest due to variations in the trustee board structure. In not-for-profit funds, consisting of the corporate, public sector and industry funds, equal representation is required from members of the fund and employers. In contrast, for-profit funds are non-representative and trustee board members are often employees of the professional financial institution who provide and manage the fund. As such retail funds are expected to have much higher agency costs which emerge in various ways. Without member representation, having only employees of the company on a trustee board creates conflicts of interest as the company aims to make a profit from providing superannuation services, whilst simultaneously attempting to maximise their customer’s retirement savings.

Unlike not-for-profit funds which have established customer bases originating from industries of work, retail funds need to create networks and entice employees into their fund, often through commission-based financial planners. Commissions involve more cost for consumers, but more earnings for planners than fee-for-service charges. As retail funds are the only fund type that pays commissions, it creates an incentive for financial planners to advise their clients into retail funds, and encourages purchase of other financial products which is driven by commission, not independent and sage advice.

Effective fund choice is further limited by agency costs which generates high exit fees to deter customers transferring their pension to another fund, another result of how retail funds have no established customer base. By dissecting performance through fund type, various features within each fund type could be assessed for any material effect on performance, such as size and economies of scale, returns over time and benefit structure. It is presupposed that for funds with a larger amount of financial assets, any fixed costs involved in administration and management, both of the fund and individual accounts, would be spread out over a greater volume of cash, resulting in lower average costs. Likewise, large funds should be able to negotiate more competitive fees due to the increased influence and bargaining power they can attain due to their size. Using return on assets, volatility on returns and expense ratios as a measure of performance Coleman et al.’s (2006) analysis found that not-for-profit funds outperformed for-profit funds, with not-for-profit funds reporting significantly higher risk-adjusted returns. The difference in returns amount to an estimated 228 to 318 basis points per annum (Coleman et al., 2006). Another significant finding was that for-profit retail funds have significantly higher expense ratios than not-for-profit funds, which helped to drive the difference in net returns between the fund types. The higher expenses for retails funds were associated with marketing and networking costs which aim to establish and expand their customer base. Performance across fund types was also assessed in terms of risk-adjusted performance using the Sharpe ratio and Jensen alpha, both of which are measures of the returns received relative to risk taken. In terms of risk-adjusted performance using the Sharpe ratio, retail funds again had a statistically significantly lower performance, with a ratio 20-30% lower than not-for-profit funds.

The alternative measure of the Jensen alpha also recorded significant differences between not-for-profit and for-profit funds. The greatest disparity occurred between the industry and public sector funds with the retail funds in the multiple index model with market timing. This suggests industry and public sector funds have more active management of their asset allocation to take advantage of fortuitous market movements. The results were concluded to be consistent with the hypothesis that agency costs exist to create differential performance between not-for-profit and for-profit funds (Coleman et al., 2006). The lack of a representative trustee board creates behaviour that fails to minimise expenses, resulting in lower returns for for-profit funds. That fund features such as size and benefit structure were not significant determinants of performance simply allows greater explanatory power to be attributed to agency costs for the stark performance between not-for-profit and for-profit funds. This study expands upon prior studies to provide a quantitative figure on the adverse effect of agency costs upon superannuation funds. Whilst Coleman et al. (2006) published a thoroughly researched articles, there are some methodological issues that emerged. A chronic problem was the underreporting of expenses where external expenses were often deducted from the gross return on assets.

Therefore explicit figures of external management costs were often unavailable yet still need to be taken into account for the expense ratio. To remedy this, external costs were estimated using the average management ratio reported in the annual returns of 25 pension funds during 2001 or 2002, which is at least on par with (Rice and McEwin, 2002) or is a conservative estimate (Clare, 2001) compared to other studies. This would be more accurate if the sample taken was representative of the superannuation industry and not skewed towards particular cost functions of any fund type. However, it homogenises costs between fund types, which is far from the truth according to Coleman et al’s results. Subsidies in corporate funds also posed problems, possibly leading to an underestimated expense cost.

This is difficult to correct for as subsidies may not be used uniformly, providing only a caveat for potentially understated corporate expense figures. The article’s sample time period from 1997 to 2002 may also detrimentally affect the results. Its relevance to the current superannuation industry can be questioned due to the introduction of the Superannuation Choice Act No. 82 in 2005. Future research should determine whether member fund choice has affected performance or fees. Fund behaviour may have altered in order to gain competitive advantage, e.g. increased advertisements to attract and retain members. The survey period of 7 years is sufficient to observe several peaks and troughs, however, whether these trends are an accurate estimation of performance throughout a member’s lifetime is uncertain. In Coleman et al. (2006), the results of the Jensen alpha multiple index model suggested that corporate and public sector funds had more active management which allowed them to reap higher returns throughout the sample period. This contrasts Drew and Stanford (2003), who subscribes to the Efficient Market Hypothesis which states that securities markets are informationally efficient.

Hence, profit is already maximised and active management strategies are futile. A longer time frame of study should be adopted in the future in order to ensure the long term trend is captured, particularly since retirement savings is also a long term goal. Whilst Coleman et al. (2006) delivered a thorough analysis of investment performance by fund type, they acknowledged that they the experiment was not completely controlled. Differing allocation strategies can influence investment growth and this would create bias in the results if fund types were correlated with particular asset allocation and growth trends.

For example, where retail funds tend towards more conservative asset allocation than not-for-profit funds (Ellis et al., 2008). Evidence of the alignment of asset allocation in the data was displayed by the high correlation between fund performance with a known balanced asset allocation. However, Ellis et al. (2008) improve the analysis further by directly controlling for asset allocation through analysing performance by investment option. This controls for the varying risk preferences and financial products utilised by fund members, synchronising fund type analysis. Ellis et al.’s (2008) results showed that expenses and taxes were the only significant determinant of investment performance. It confirms Coleman et al.’s (2006) findings, where net return relative to a benchmark return for retail funds in a standardised default option significantly underperformed the other fund types.

Underperformance and higher costs in retails funds reflect the agency costs associated with for-profit funds as the trustee board struggle between attaining pension fund growth and earning profit. Hence Coleman et al.’s (2006) are evidently robust and exposed the varying performances of pension fund types.

Cite this page