From an investment strategy viewpoint there are two key approaches: active management, which is the art of stock picking and market timing with the objective of generating returns greater than the market (i.e. an alpha); and, passive management, which is the buy-and-hold approach that often follows the performance of a particular index in order to generate average market returns.
Despite many institutional investors being owned by a global parent corporation with broadly defined investment strategies, most Mutual Fund Managers, Investment Advisors, and Pension Funds are typically independent legal entities within a country with their own remit for investment strategies. Moreover, some of these institutional investors manage a number of mutual funds that have individually defined active or passive investment objectives.
Investor Relations professionals take their quoted company's senior management to meet institutional investors in a round of road-shows when the interim and final results are announced, in that the hope that their investment case will result in said institutional investors purchasing their equity. However, the author frequently meets Investor Relations professionals who believe that some institutions and funds that are supposedly actively managed are in fact closet index trackers investing passively and thus it is a waste of their time meeting with them; in addition they believe that as some mutual funds are managed by one person or a team that indexing tracking behaviour could occur across these multiple funds rather than individual ones.
Active managers require expensive data and research to support fundamental or technical analysis techniques and thus charge high fees for their services, but actively managed funds remain popular despite relying on exploitation of market inefficiencies as objective when the strongest arguments are in favour of efficient markets and therefore passive investing. The author is motivated to tackle this topic because if it is possible to demonstrate that a large percentage of funds that claim to be active but are in fact index trackers then their active fund management fees would be unjustified and it would call into question the fund management structure of the entire fund management industry.
This topic ties in with the author's Management Studies, particularly within the Finance & Account track, and builds upon what the author learnt in four modules: Investment Strategies, Quantitative Business Analysis, Business Finance, and Financial Reporting.
The underlying academic theories and constructs are of particular relevance to the author's research are principal-agent problem, loss aversion from prospect theory, correlation and dependence, spearman's rank correlation coefficient and other non-parametric statistics, modern portfolio theory, efficient market hypothesis, and perhaps some elements of the capital asset pricing model.
The scope of the authors intended investigation is to examine a representative sample of the largest equity fund institutional investment firms in the British equity market. Institutions and funds holding long positions in equity will be included, but certain types of institutions and funds will be excluded and thus out of scope due to either technical reasons or lack of transparency of their holdings to external agents.
In this dissertation it is the author's intent to investigate a small number of equity fund management firms that represent the majority of the market capitalization and access the extent of which they are closet index trackers in their portfolios. Specifically, the dissertation aims to answer the following questions:
Which institutions cover 80% of British equity capital?
What proportion of these institutions/funds are declared index trackers?
Does the remainder reflect in whole or in part global benchmark indices such as FTSE, MSCI, Dow Jones, and STOXX?
What proportion of active funds are closet index trackers?
Does fund and institutional consolidation matter when attempting to identify whether entities reflect indices (e.g. parent, institution, or mutual fund level)?
Does the degree of index tracking, if present in active investors, vary depending on market sentiment (e.g. bull, bear or distressed markets)?
Does the degree of index tracking, if present in active investors, vary by asset allocation (e.g. domestic versus international; large, mid, small cap; sector; etc.)?
Does the degree of index tracking, if present in active investors, vary by declared investment strategy (e.g. growth, looking for capital gain; value, seeking undervalued stocks)?
The null hypothesis is that claims of active fund management made by active equity fund institutional investors are justified. The alternative hypothesis is that despite claims of active management the institutional investors are tracking portfolios to a lesser or greater extent across their portfolios.
2. Methodology:
Fred N. Kerlinger said "There's no such thing as qualitative data. Everything is either 1 or 0," (Miles & Huberman, 1994: 40 cited by Masucii, 2006: 7) and that is the case with the research data in this proposed dissertation, there are only numbers and statistics not words nor any non-numeric data; therefore the research methodology is quantitative. Moreover, the author knows clearly, what he is looking for, which is to test a hypothesis using objective analysis of data. Furthermore, many of the aspects of the study can be designed before the data is collected.
First step will be to build the model, this will require extracting the constituents and their index implied weights from the major parent indices (e.g. FTSE, MSCI, STOXX, DJ, etc.) and their sub-indices (e.g. cap sized, sector specific, etc.) from sources and create lookup tables. The next activity will be to setup a model that takes an institutional profile, extracts its holdings for each index and converts them into percentages (the actual weights). The final element of the model will correlate the index implied weights to the actual weights.
Second step will be to validate the procedure of identifying index tracking via by using a positive control. This is achieved by running a known index tracker (such as Legal & General Investment Management), which is the positive control, through the analytical procedure; if the known index tracker (positive control) produces the expected outcome, then we can conclude that the approach is competent in observing the effect. In addition, the author plans to run a number of know index trackers through the steps to not only act as positive controls, but also to identify non-sampling errors (such as measurement errors) and thus define errors bands which can be used to reduce bias caused by systemic error (Wolverton, 2009:373).
Third step will be to identify a representative sample, i.e. one that has a similarity to the larger population, so that the author can account for sampling error and further reduce systemic error biases. The author's approach will likely be stratified random sample, where the population will be divided into sub-groups (strata) based on one or more characteristics, and random samples selected from these strata (Wolverton, 2009:374). The size of the sample will be dependent on how long it takes to analyze each entity versus the time available for analysis, both these times are unknown as of writing this proposal but the author hopes to get a large sample size and cover most of the capital in the markets under examination.
Fourth step will be to extract the sample's data from sources and run it through the model, thus generating correlation analysis for each entity versus the numerous indices.
Fifth step will be to perform various statistical analysis on the output of the fourth step in order test the hypothesis and answer research questions, potentially using multiple regressions to identify linear relationships between multiple categorical variables (e.g. Large Caps are index trackers and Small Caps are not).
3. Feasibility:
The author has access to all the required data through his own business's subscriptions to Thomson Reuters and Factset research services, and also has the appropriate usage license from data-providers.
The author does not plan to examine the entire British publically quoted equity market, but rather the few institutions that hold 80% of the equity Assets under Management (AuM). For example, although American institutional investors hold approximately US$11-trillion of equity AuM, as few as 160 institutions hold 80% of this equity AuM (where 160 is 4.86% of the total 3,293 American institutions); from personal experience the author believes that the UK market is similar to the US one so a similar limit to the top 80% of UK equity AuM reduces number of institutions to analyze from the thousands down to less than a couple of hundred
However, the scope of the study may change as it is determined by the evolutionary research process that is being proposed. For example, in the pilot phase the author will test the model again an institutional entity and against its constituent mutual funds to determine the appropriate level of granularity; moreover, it might be possible to analyse by fund manager
Challenges will include data cleaning, combing data sets (the extracts are not entire sets), if the author needs to run the models at mutual fund level of granularity this will add significantly to the number of data extracts required, and most complex of all mapping the data providers records together; unfortunately this task is required whether one analyses 1 or 100 institutional investors.
4. How Your Work Fits Existing Published Work:
Closet tracking in a term that appears in the press but does not appear to have much coverage in academic literature. Where it does exist the usual indicator for closet index tracking is the 'tracking error' measure, which is the standard deviation of the differences between the return of a portfolio and the return of the benchmark index, where a low tracking error indicates likely index tracking (Barro & Canestrelli, 2008: 47); but this approach relies on knowing which index is being tracked.
There are several well-known studies [add citing] that examine individual mutual fund performance versus benchmark indices; therefore the author's approach of examining multiple funds managed by one individual or group or at an institutional level should supplement the existing field of knowledge. Thomas & Tonks (2001: 322) considered that a number of UK policy documents had suggested that pension funds invest in index funds because "there is little evidence that active fund management can deliver superior investment returns for the consumer (Office of Fair Trading, 1997: 71, para. 420; see also Consumers' Assocation, 1997; Department of Social Security, 1998; Financial Services Agency, 1999)," (ibid). Therefore Thomas & Tonks (2001: 342) investigated the performance of the UK equity portfolios of 2,175 segregated UK pension funds over the period of 1983-97, which was the longest set of UK pension fund data analysed to date; the benefit of such a large dataset was that it enabled Thomas & Tonks to "examine performance over three distinct sub-periods," (ibid). Their approach was to "analyse the shift in the distribution of returns relative to an external benchmark" and to determine whether fund performance, on average, was better or worse in bull markets or bear markets; or if they add value in markets with broad spread of activity versus narrowly focused sectors; and whether small or mid-cap gives better performance that large cap. They used Jensen's technique to measure fund performance and found that most of the pension funds in their sample had an "equity beta close to unity" and that the "coefficient of determination in the regression of fund returns against returns on the market was very high," (Thomas & Tonks, 2001: 342). Thomas & Tonks conclusion based upon these two findings was that it appears that the funds in their sample are 'closet trackers'. The author believes that this article provides an excellent start for the Literature Review and placing the topic in the context of body of research done by others, furthermore it gives some weight to the author's alternative hypothesis that active funds are in fact closet trackers. Further implications for the author's research are firstly in the scope, the UK market alone is adequate; secondly, that similar research questions could be defined, in regard to bull and bear markets, diversity of sectors and sub-sectors, and cap strategy (these last two will be factors of various sub-indices under examination). One area the author has not considered was 'survivorship bias', but does not think this will be factor; rather the recent consolidation and change of ownership of funds and institutions will be a predominant factor to consider, e.g. BlackRock acquiring Barclays Global Investors.
Waring & Siegel (2005) considered eleven 'myths' of Active Management and attempted to debunk them with a brief argument around each. Waring & Siegel (2005) argue in favour of Active Management in all their responses, for example that in the mid- to late-1990s most US-based international equity managers beat their benchmarks by 5% to 7% and achieved this by "underweighting Japan relative to MSCI EAFE index when Japan was experiencing lower returns on the benchmark as a whole," (Waring & Siegel, 2005: 21). Waring & Siegel (2005: 24) briefly discuss 'closet index trackers' and attempt to argue that they are simply "any fund managed without skill," (ibid). Article such as this one will provided a conter-point in the author's literature review, however, one has to be vary careful of the veracity of such articles because the authors have vested interests in promoting their point of view (they are both active fund managers).
Ennis (2005) exclaims that the pursuit of alpha is the zeitgeist of the times, and examines the question that "if the potential payoff from active management has waned since 1960, why is the price of active management at or near it's all-time high?" (Ennis, 2005: 44). Firstly Ennis examines market efficiency and argues that not only have active returns reduced since 1960 but also the equity trading costs as percent of trade value have dramatically decreased over the same period, secondly Ennis (2005: 46) examines the average equity fund expense ratio (equally weighted) over time and the apparent trend in price rises. Jensen (1968, cited by Ennis 2005: 45) consistently demonstrated that, on average, active managers underperform their benchmark indices by an amount approximately equal to their fees; which implies that the are in fact closet index trackers. This article gives the author insight into Jensen's 1968 findings, and a relatively up to date examination of them; the data and approach are not useful for the author's analysis but the content and articles that it cites are useful for the Literature Review.
5. Why You Are Doing This Topic:
Firstly, the author's business supports the Investor Relations departments of publically quoted companies with independent equity capital advice. One of the key services offered is Precision Targeting, helping Investor Relations departments select which institutional investors their executives should meet with during interim and full results road-shows. CEOs tell us that they believe many institutional investors are closet index trackers and it is a waste of their time going to see these people, because no matter what is said to them or how compelling an investment case they will simply track the index. If the author could identify any active institutional investors that are really index trackers this will be extremely valuable information for these executives.
Secondly, closet index trackers are funds / institutions that are supposed to be actively managed, but nevertheless run in a conservative manner that is little different from a index tracker. If investors wish to track indices they are better off investing in true index tracker funds because: the fees will be significantly lower, and as the fund generates the same return it means a greater return for the investor; a real tracker fund will carefully chose a benchmark index and track it highly accurately rather than lazily pick one and sloppily rebalance around it and destroy value - in addition, many active fund managers are measured against benchmark indices and can be tempted to simply track it rather than risk underperforming it. If the author could identify closet index trackers, this would be extremely valuable information for investors who wish to get the best equity returns on their capital.
6. Timing Mileposts
N.B. You must reach stage 8 at least four weeks before your deadline; a month contingency provision is also advisable to allow for slippages. This is to enable your DA to give you sufficient Feedback on your final draft.
You should produce a final Proposal for submission to the SM for approval (as agreed with your DA) in between one and two months from your classroom date. However, please aim for 7 weeks from your start date (class date) at the latest, to ensure time for any necessary revisions and final approval by the 8 week cut-off.
Milestone
Description
Due date
Remarks
1
Stage 1: Area of interest identified
2
Stage 2: Specific topic selected
3
Stage 3: Topic refined to develop dissertation proposal
4
Stage 4: Proposal written and submitted
28 April 2010
Week 7 deadline extended by two weeks.
5
Stage 5: Collection of data and information
6
Stage 6: Analysis and interpretation of collected data/information
7
Stage 7: Writing up
8
Stage 8: Final draft prepared - submission of dissertation
9
Final Deadline - nine months from classroom date.
