Here I am going to do a "back-of-the-envelope-analysis" of its fund performance, using target date fund 2040 from Fidelity and Vanguard as case studies. The result confirms my prior belief, retirement product with purely passive index tracking, as exemplified by Vanguard, is superior to "active" asset allocation, as represented by Fidelity.
Disclaimer: The article is not an investment advice and the opinion expressed here does not represents my personal opinion.
I. The Repackaging Game
It is an open industry secret that each power house constructs their retirement finance product by repackaging the funds bred within the institution. For example, Vanguard retirement funds use other vanguard funds as basis while Fidelity retirement funds are exclusively composed of Fidelity funds.
But how? One can do it in two ways, cheap or expensive.
The cheap way is just to pool funds that represent the market in the most general way without worrying too much about the fine tuning. Because there is not much work involved once the style funds are nailed down, the institution cannot charge too high a fee in a competitive market.
The expensive way is to pool funds based on a proprietary view point. Managers try to represent asset styles of the market on a finer grade and even impose their own investment view on the portfolio. Because all the human work that goes into the sausage making process, the seller feel entitled to charge for a higher expense ratio.
That is all mumble jumble. What does that mean in English exactly? :Let me illustrate the two styles with Vanguard and Fidelity 2040 target date Fund.
Vanguard is the text-book case of the cheap repackaging. It only consists of 3 funds, which tracks three major asset style: US equity, International Equity and US Bond market.
| Fund Composition: Vanguard 2040 |
| Vanguard Total Stock Market Index Fund Investor Shares |
| Vanguard Total International Stock Index Fund Investor Shares |
| Vanguard Total Bond Market II Index Fund Investor Shares |
In contrast, Fidelity has much finer asset style classification. It has 16 funds in total. Not only it has specific exposure to commodity and real estate, it also has very subjective asset allocation strategy, such as "disciplined equity fund" and "small cap opportunity fund". There appears to be a lot of sleepless nights involved to carve out the investment allocation, or at least so the investor should hope.
| Fund Composition: Fidelity 2040 |
| Fidelity Series AllSector Equity Fund |
| Fidelity Series Large Cap Value Fund |
| Fidelity Growth Company Fund |
| Fidelity Disciplined Equity Fund |
| Fidelity Series 100 Index Fund |
| Fidelity Blue Chip Growth Fund |
| Fidelity Series Small Cap Opportunities Fund |
| Fidelity Small Cap Growth Fund |
| Fidelity Small Cap Value Fund |
| Fidelity Series Real Estate Equity Fund |
| Fidelity Series Commodity Strategy Fund |
| Fidelity Series Investment Grade Bond Fund |
| Fidelity Series High Income Fund |
| Fidelity Series Floating Rate High Income Fund |
| Fidelity Series Emerging Markets Debt Fund |
| Fidelity Series Real Estate Income Fund |
II. Methodology: Synthetic Portfolio Replication
How can we compare the performance of two totally different funds?
The naive view is to compare the expense ratio. Vanguard charges 19 bps a year while Fidelity charges 78 bps a year. So Vanguard 2040 is a better choice than Fidelity 2040.
Wrong. Because Fidelity folks will argue that their portfolio offers better risk-return reward than its index tracking vanguard peers.
So why not just compare their historical performance? If one plotted on Google finance, it is indeed the case that Vanguard 2040 is doing better than Fidelity 2040 since 2006 (LINK). The Vanguard fund over perform Fidelity fund by 11%, while the expense ratio only accounts for 3.5% of the difference.
Fidelity folks still have excuses for their sorry performance. They chooses a more risky market portfolio, and they just have a tough ride thanks to the Global Financial Crisis.[1]
To account for this argument, one could run a Fama-French three factor model and see if the alpha comes out different for the two funds. But here, I will introduce a more practical version of test:
Can one replicate the fund performance using a series of ETF that serves as proxy for the asset style?
If one can perfectly replicate the fund performance with a cheaper ETF portfolio, it is hard to argue that high expense has any merit. Of course, it is next to impossible to find an exact replicate.
So the natural question is that is the alpha over the synthetic portfolio justifies the high expense ratio?
For a more detailed version of the method, please refer to Razor Hedge LLC's website. Here I will use a water down version of it. Leveraging Morningstar's public info, I will use SPY (US equity), EFA (international equity) and AGG (US bond) to replicate the fund.
To be more specific, I recorded the alpha of the each composite over the style benchmark reported by morning star. I chose the longest history possible and annualized them according to the available time length. Then I aggregate the alpha according to the composite weight of the original portfolio. It is the alpha of the true fund over a synthetic portfolio replacing style benchmark with aforementioned ETFs. There will be tracking error of the ETFs but I choose to ignore them in this excise. The fund allocation is not constant over time, and I also ignore them for convenience. As such, the simulation provided here is by no means 100% accurate, but I would argue it is accurate enough to make qualitative judgement call.
Since I know the expense ratio of three ETFs, I can also estimate the "real return" had one actually constructed the portfolio. In fact, one could read the quarterly release of the fund and rebalance accordingly. So this is not entirely an ivory tower thought experiment.
III. Result
I offer the full excel working sheet for download HERE
Here is the big picture:
| expense | synthetic expense | fund alpha | |
| Fidelity | 0.78% | 0.16% | 0.05% |
| Vanguard | 0.19% | 0.17% | 0.07% |
If we generate the synthetic portfolio for the two fund, neither fund outperform the synthetic portfolio by a wide margin. For Vanguard, it is expected, because it is engineered to track the index. However, for Fidelity, it is quite embarrassing: with all the allegedly effort, they did not even beat the dumb synthetic portfolio.
But wait, Fidelity is charging an expense that is 3 times higher than that of vanguard.
I rest my case.
ENDNOTE
[1] In a perfect efficient market, this is non-sense. There will be one market portfolio if we can agree on what the market is. Since neither fund can leverage, their risk characteristics should be exactly the same. Well, in reality, people do not agree on what market looks like (part of the thesis for active management) and market may not be perfect efficient, so please forgive my theoretical sloppiness here.
No comments:
Post a Comment