ETFs Vs. Mutual Funds: A Case Study - Part I

 |  Includes: CL, IWD, KO, PG, XLP
by: Eli Inkrot

I have previously written about a rationale I developed for comparing a dividend growth strategy of holding individual stocks to that of an ETF. I compared the Select Sector Consumer Staples SPDR (NYSEARCA:XLP), which has many dividend growth companies within it, to its counterparts of recognizable dividend stalwarts, such as Procter & Gamble (NYSE:PG), Coca-Cola (NYSE:KO), and Colgate-Palmolive (NYSE:CL).

Based on the categories of Portfolio allocation and Dividend Yield & Growth, I found that a dividend growth strategy likely carries a slight edge over holding a single ETF. Furthermore, when you add in the categories of Fees and Selective holdings, the dividend growth strategy is recognizably more efficient in the long-run.

But that's not to say that ETFs don't have their place. Oft they are not used singularly, but instead as a specific function in one's portfolio; more generally, ETFs offer a low-cost alternative to a wide range of diversifiable benefits.

So let's say that you want to gain entrance into the Large Value Equity arena. As one is likely well aware, simply buying a portion of 100s of large value companies is going to be quite costly. Luckily there are alternatives. One could go with an ETF, say the iShares Russell 1000 Value Index (NYSEARCA:IWD).

Another route would be to go with a mutual fund, say the JP Morgan Large Cap Value (OLVAX). For those of you wondering why I selected these two options, I simply pulled them out of a hat for illustrative purposes. Much in the same way, there are hundreds of alternatives for whatever style or position gap you are looking to fill.

In developing the case study between the IWD and the OLVAX, I knew that the ETF likely had lower fees. This turned out to be precisely the case, as the IWD has an annual expense ratio around 0.20% and the OLVAX has an annual expense ratio around 1.08%. Other than that, my working assumptions were limited in that I expected that the OLVAX must provide superior risk-adjusted returns in order for someone to consider it as a practical alternative to the benchmark.

It should be noted that my comparisons of the OLVAX to the large value sector were made directly to the Russell 1000 Large Value index - and not the IWD. However, in order to invest in the benchmark, one needs a proxy. For such purposes, I refer to the IWD. Additionally, the data is from 1997 to 2007, as to avoid any funny-business in the last four years. This was an effort to quantify "more normal" times. I do realize that 1999-2001 weren't exactly normal themselves, but the analysis mainly focuses on the "more reasonable years". Let's begin.

First, I created a 3-yr Rolling Window graph of the OLVAX manager's fund performance against that of the applicable style weight benchmarks.

3-yr Rolling Window Graph and Style Weights

Click to enlargeClick to enlarge

Click to enlarge

The 3-year Rolling Window of asset allocation (as seen above) allows for a pictorial view of the managers selective asset allocation over the years. Time is on the horizontal axis, while the style allocation is on the vertical axis. It demonstrates both how closely the manager stuck to his mandated allocation requirement and whether or not there was a significant shift during this time period. It should be made known that this graph captures a backwards looking period of three years such that each date on the horizontal axis demonstrates what happened three years ago up until the specified point.

In this particular case (OLVAX), the fund manager appears to overwhelming stick with his required holdings, namely large cap value stocks. Although other holdings appear to sneak their way into the portfolio, the large cap value dominates the frame. Additionally, it is possible that these discrepancies are a result of definition rather than an inherent disconnect between the specified mandates.

For example, the benchmark might consider large growth stocks to be those with a dividend yield under 2%, for example, while the fund manager might see value here. It would be interesting to see whether or not the fund manager drastically changes their holdings in worsening times as to keep up with their counterparts.

In addition to the Rolling Window graph, we also have the applicable style weights. This Asset Allocation style weights graph is precisely similar to the 3-year rolling window in that it provides information on how often a fund manager adhered to a specified asset class. To create this graph, I used the provided returns of specific indices (large value, large growth, small value, and small growth) to best characterize the style of the JP Morgan Large Value fund.

As you can see, more than 85% of the time, the manager adhered to a selection of assets that tracked the large value index. This is reasonable and definitely demonstrates that the manager was mostly keeping to his requirements. Although to be sure, if you already had a variety of other asset classes and were buying the OLVAX for the added value of large value stocks only, then you might not be so accepting of an allocation outside of 100% in large value.

For example, if investors viewed their portfolio in its entirety and they already had their desired allocations in large growth and small growth, then they would not buy the OLVAX to achieve an even higher allocation in these classes. Once more, it is important to recognize the definition error possibilities here, but also the added value of the manager. For example, if the manager is able to produce consistently outsized risk-adjusted returns, then one might not care so much about what allocation he is using to get there.

The R-squared graph below can provide insight into the influence of the fund manager.

Click to enlarge

This graph was created by running a regression on the OLVAX fund returns as compared to the returns of the Russell 1000 Large value index. It should be noted that while these returns were calculated by first subtracting the risk-free rate, either way yielded the same 94.46% to 5.54% ratio. The overwhelming portion of this graph, the 94.46%, represents the R-squared to the Benchmark. More specifically, this is measure of how well the manager's returns are a resultant of the benchmark. That is, 94.46% of the variance in the JP Morgan Large Value fund can be explained by the Russell 1000 Large Value index. This is a very high explanatory power, but is not necessarily surprising, as many of the holdings in the Russell 1000 Large Value Index are likely to be held in the OLVAX. Furthermore, the Russell Index is a value weighted measure such that the largest holdings have the greatest clout in return movement. However, it should be noted that the precise holding weights are assuredly different.

The 5.54% residual shows how much variance cannot be explained by the benchmark and thus is the resultant of the manager's stock selection and weightings. In choosing between the Russell 1000 Large Value fund and the OLVAX for a return associated with large value, one would not see a significant deviation from the benchmark. Add in the fees and simply buying the benchmark for this given allocation is likely a better option. That is, unless the manager provides consistent, yet small by definition, positive gains over the benchmark time and again. We move on to this precise question with regard to relative performance.

The graph below the R-squared graph (see above) details the manager performance with regard to the benchmark.

This graph represents how well the manager performed against the benchmark, in this case the Russell 1000 Large Value index. The values were calculated by first normalizing the beginning of our time frame. That is, both OLVAX and the Russell index were set to 100. In this way, performance before January 1998 was not captured nor measured. The graph demonstrates how well the JP Morgan Large Value fund performed relative to the benchmark in the given time frame.

Calculations were made by starting with the normalized value and adding or subtracting subsequent returns of both the index and the fund. As you can see, the OLVAX fund performed better than the benchmark in the 2000-2001 range, while relatively underperforming the benchmark from 2003 until the beginning of 2007. Overall one can see that the two allocations followed each other quite closely. This is to be expected given the high R2.

Thus far in comparing the IWD ETF and a large value mutual fund, the OLVAX, we have seen fundamental similarities. The OLVAX appears to stick to its mandate of large value equity holdings as seen in the asset allocation and style weight graphs. Furthermore, the variance in the OLVAX is largely explained by the underlying holdings in the benchmark. When viewing returns we see that the OLVAX outperformed in the early years and underperformed in the latter years. However, thus far we have not mentioned the applicable risk related to each holding. This is precisely what the next part of my article seeks to accomplish.

Disclosure: I am long KO, PG.

Additional disclosure: This report, especially with regard to data analysis, was completed with the instrumental help of Kai van Hooff.