Another View On 'Diversification Fallacies, Part 1: Asset Allocation'

by: Daniel Moser

It has been too long since I have posted anything on Seeking Alpha. Recently, I had the pleasure of reading a post from Dane Bowler titled "Diversification Fallacies, Part 1: Asset Allocation". What a topic! The title alone jumped out at me as a must read, largely because I think asset allocation is one of the most interesting aspects of portfolio/investment management. Between the title and the first paragraph I was hooked. He provided a great setup but once I passed the first paragraph I found little content on which I could agree, leaving me no other choice but to refute some of his arguments in the hopes of having a spirited discussion. Let's get to it!

According to Mr. Bowler, "diversification is a very important concept in investment[s] as it is one of the primary tools for reducing risk." As he shifts gears and extends his argument into asset allocation he argues that 1.) "Asset allocation is a diversification method endorsed by highly reputable universities", 2.) That "asset allocation has become so prevalent that few even question its validity" and 3.) "[Asset allocation] is an arbitrary constraint". His third point is where the disagreement begins.

Asset allocation is not a set of arbitrary constraints. At some investment institutions, an asset allocation policy might be a set of misguided, misleading, naive or outright bad constraints-but they are not really arbitrary. You could make a case that over some period of time, if a portfolio's asset allocation policy is not revised with new data/assumptions on expected correlations, returns, risks, etc. the policy could become counterproductive towards maintaining a desirable level of diversification and potentially poor portfolio performance - but that is about as far as claims can go.

The next section of Mr. Bowler's article explored 3 different challenges to the effectiveness of asset allocation: theoretical, by size, and by industry.

The theoretical argument attempted was made through a simple analogy but, in reality, the analogy turned out to be rather nonsensical-largely because it totally neglects how the modern world has come to think about "asset classes" i.e. financial instruments that deliver a variety of cash flow/return profiles with varying degrees of statistical characteristics including risks, returns, and correlations among one another. The underlying assumption in the analogy used is that color is the only deterministic factor of a particular baskets ability to hold eggs as opposed to the materials used to construct the basket creating a certain level of "structural integrity" i.e. paper, plastic, aluminum, or steel. Much like the analogy, the underlying assumption the author makes about real life asset allocation/diversification is that the underlying cash flows/return profiles among asset classes have no bearing on the level of diversification inherent in a portfolio-which happens to be the real underlying fallacy in the discussion.

Next, Mr. Bowler shifts to a real life example of asset allocation: asset allocation across market capitalization. He makes the following case, "many will suggest various weightings between small, mid, and large cap stocks. The only way doing so would increase diversification of a portfolio would be if each size had correlated risks." This statement does a pretty outstanding job of totally missing what drives a portfolio's diversification … which is less correlation among assets - not more. If market capitalization had no impact whatsoever on a stock's return profile, market capitalization would be entirely irrelevant in an asset allocation strategy. So why not simply ask, does market capitalization matter? Consider the following anecdotal question: during the peak of the credit crises in 2007, did large companies experience as much difficulty obtaining credit as small companies? No! Small companies had a much harder time securing new financing, rolling existing financing, etc. and for far longer than larger companies. Why? Well, small companies are frequently perceived as riskier. This is manifested into the underlying data through an inverse relationship between a company's size and their stock's volatility (on an asset class level) as well as (we hope) a modestly higher level of return over time to compensate investors for bearing this added risk in their portfolio. This underlying manifestation in the data is what leads portfolio managers to the conclusion that they can improve a portfolio's diversification by varying a portfolio's weightings by small, mid, and large market capitalizations. Companies of varying market capitalization classes are not "perfectly" correlated which means there are diversification opportunities available.

The third argument posed: asset allocation by sector. This argument, again, misses the underlying premise of risk management through asset allocation: not all industries share the same risk characteristics. Some industries are inherently more prone to ebbs and flows, bull and bear markets, ups and downs. Compare and contrast industrial metal mining with the utilities industry. These industries are substantially different from one another in terms of the underlying business and that is typically manifested in the underlying statistical data. Hint: one industry exhibits practically 3 times the level of volatility than the other. Moreover, it is these differences that create diversification opportunities. Again, if every industry were perfectly correlated there would be no opportunities to diversify, but it just so happens that empirically, industries do vary from one another.

Let's take a moment now to step back and flip this analysis on its head. What would a portfolio manager have to do to make asset allocation and diversification irrelevant? He or She would have to pick winners a VAST majority of the time if not ALL the time-which is virtually guaranteed impossible unless the portfolio manager uncovers a hidden pool of assets that ONLY increase in value. With this in mind, take a look at the authors next argument on the merits of asset allocations in which he examines the cost/benefit analysis of asset allocation ...

"Suppose the fundamentals are lining up perfectly for a certain sector yet its market prices have not yet risen to meet the new information. An investor may reasonably feel this sector has a higher expected return than other sectors, yet the constraints put in place by his asset allocation strategy prevent commitment of more than 10% to this sector. He must also commit 10% to the sector he believes has the lowest expected return to meet that sector's constraint. In such a scenario, the use of said asset allocation strategy would substantially reduce the expected returns of the portfolio while only providing heuristic, fundamentally flawed, diversification."

Let's "suppose" the fundamentals are lining up perfectly for a certain sector and yet the market price for this particular sector have not baked the new fundamentals into the cake so to speak. Next, let's suppose -- just for the sake of argument -- we are wrong! Compare and contrast the relative impact of being long and wrong with 100% of your assets in a single sector vs. 10% in 10 sectors. Surely we can agree there will be a pretty big difference in these two scenarios. Diversification is about risk reduction not return maximization. The core of the author's argument against asset allocation is really on a particular instance in which an asset allocation policy mandates an equal weighting (by % of underlying assets) in each sector-which is by no means a default requirement for an asset allocation strategy. Perhaps the more compelling argument along the authors line of thinking is that an equal sector weighting asset allocation policy is sub-optimal. But that does not in any way prove asset allocation is irrelevant or counter-productive towards achieving good performance and good diversification on a portfolio level.

Mr. Bowler goes on to provide a "better" alternative to asset allocation. But before I can even consider his alternative, I am left asking myself … better than what asset allocation strategy-an asset allocation strategy based on equal dollar value, market cap, growth vs. value, sector, asset class, risk? There is really no substantive comparison portfolio to evaluate the "handcrafted fundamental diversification" approach against. We, as readers, are literally left with no idea what the watered down asset allocation portfolio even looks like but nonetheless lets evaluate the proposal. He proposes a portfolio comprised of 5 REITS specifically: American Realty Capital Properties (ARCP), Weyerhaeuser (NYSE:WY), Northstar Realty Finance Corp. (NRF), RAIT Financial Trust (NYSE:RAS), and Omega Healthcare Investors (NYSE:OHI). He goes on to suggests that a portfolio constructed with only these 5 securities can "demonstrate substantial diversification". He then provides a hypothesis of how each of these 5 REITS will perform with respect to two potential risk factors: inflation and economic risk.

After reading this portion of the analysis, I am left with no better understanding of what a diversified portfolio looks like. But, I do feel like I just played a game of "Am I Diversified" from "Mad Money". Literally, isn't this the same game? I am wondering if Jim Cramer would bless a 100% REIT portfolio as sufficiently diversified. I guess I will never know. Mr. Bowler feels justified in arguing that this portfolio is diversified as he asserts that "diversification is not immunity to risk but rather the asynchrony of it". However, without any sort of data demonstrating that the underlying "asynchronous" business risks inherent in these 5 REITS are manifested in "asynchronous" returns (read: a diversified portfolio), I am left wondering … am I diversified?

In my opinion a 100% REIT portfolio does not constitute anything remotely approaching "asset allocation" level diversification-not at least in the way I have come to think about portfolio diversification. The only thing I can suggest is that he has constructed a 100% REIT portfolio that diversifies some portion of the unsystematic business risk between the 5 REITS he believes are winners. In fairness, they might be winners. But that implies there is a case to be made for overlooking broad level asset allocation in order to overweight (or go ALL IN) on this 5 REIT portfolio as a superior alternative to a diversified portfolio.

Mr. Bowler argues that diversification waters down expected returns. I would rephrase that by saying that bad diversification waters down expected returns. "Bad" diversification could mean being long a portfolio 100% allocated to REITS when income oriented assets fall out of favor vs. some other asset class. "Bad" could also mean allocating capital to an array of assets that are strongly correlated to one another providing few, if any, offsetting characteristics to the portfolio.

In conclusion, Mr. Bowler introduced a good topic for discussion. However, diversification is not about maximizing the return on a single trade (i.e. long REITS). Diversification is about reducing the risk that a single trade takes you out of the game. Just remember, you have to earn a 100% return to make up for a 50% loss just to get back to even. There are many asset allocation strategies that have not been explored that merit a good bit of research and discussion before we can rule Asset Allocation is an inferior means to diversify a portfolio.

Thanks for reading.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.