What do you think of oil and gas partnerships? Where do they fit in your model portfolios?
Let me start with the second question. Oil and gas limited partnerships are often divided into three categories: exploratory drilling (at the high risk/high return end of the spectrum), development drilling (to increase production from a field with proven reserves), midstream (lower risk businesses that process hydrocarbons, e.g., to remove natural gas liquids from natural gas) and occasionally pipelines (at the lower risk/lower return end of the spectrum). As you can see, the first two are effectively investments in real assets, similar to investments in timber, gold or commodities. As a result, we would include them under our allocation to commodities. In contrast, the latter two are investments in businesses that provide a service related to bringing commodities to market - one would not normally expect their revenues to fluctuate with commodity prices as much as those partnerships that are directly involved with the production and sale of the commodities themselves. Given this, the latter two types of limited partnership seem closer to equities in their essential nature. In terms of your first question, let me draw a distinction between three alternatives. The first is an investment in a public equity index that tracks performance of relatively large companies that are primarily engaged in oil and gas exploration - for example, ETFs like IEO or XOP. The second is an investment in the publicly traded stock of a single small E&P company that is not included in the ETF. And the third is an investment in a limited partnership or LLC unit issued by an organization engaged in either exploratory or development drilling. What distinguishes these three investments?
In terms of risk, the ETFs provide some diversification benefits that reduce net exposure to single company risk, leaving a mix of exposure to the E&P sector and the overall equity market (e.g., these ETFs have recently had a roughly a .60 correlation with the S&P 500). Interestingly, over the past two years, IEO has outperformed the broad equity market when hydrocarbon prices were rising (but not risen as high as investments that more closely track energy prices, like JJE, which tracks the GSCI energy sub-index). On the other hand, it has outperformed JJE and similar commodity investments when energy prices were falling. In contrast, with both the small E&P stock and the E&P partnership, you are relatively more exposed to company-specific risk (e.g., lease location and cost, drilling execution, etc.), with arguably less exposure to the overall equity market. In terms of functional differences between the public corporation and private partnership form of organization, the former may provide superior disclosure, but less effective control. In terms of the costs associated with each form, the public corporation undoubtedly has higher reporting costs, while the partnership may (and I emphasize may) have higher potential employee compensation costs if the firm is successful (and therefore may also be able to attract higher quality employees who reduce execution risk). E&P partnerships and LLCs may also have better access to capital than the smaller public E&P companies, for which excessive dilution is often a problem (remember, many penny stocks were originally resource plays).
Turning to the return side of the equation, because of the lower company specific risk profile, ETFs like IEO and XOP should deliver returns in line with the relative riskiness of E&P as a function or sector - in other words, better than the overall market, but not as high as the potential returns on a single E&P company or partnership, where potential returns should also reflect compensation for much higher company specific and small company risk. Finally, with respect to midstream and pipeline partnerships, some of these are traded publicly, in the form of master limited partnerships, or MLPs. Their performance has historically been tracked by the Alerian MLP Index. Since this past summer, an ETN issued by JPMorgan Chase tracks this index, and trades under the ticker AMJ (of course, this ETN also requires you to hold JPMorgan Chase credit risk, which may give you second thoughts).
I'm sure you get asked this question a lot (given what you do), but I'm still curious about how you answer it: What's the best asset allocation today?
Unfortunately, there is no simple answer to that question. Compliance officer-types might say that the answer depends on the answers you give on a survey that attempts to measure your "risk tolerance" and/or "risk capacity." The problem we have with this approach is that risk tolerance is much more situation specific than compliance officers like to admit. To use a glib example: Under normal circumstances, I would think it insane to run into a burning building - it would be well outside my risk tolerance, given my highest priority goal at the time - say, staying alive. But if my child was inside, running inside would be well within my risk tolerance, given the change in my highest priority goal - saving my child. In general, while research seems to indicate that people have degrees of risk aversion that are relatively stable over time, this is very different from their willingness to take risk, which can change dramatically under different circumstances. Consider three different situations. In the first, a couple has a clear plan to accumulate a million (choose your currency) by the time they retire in thirty years time. This plan includes an annual savings target and a clear asset allocation and rebalancing plan. One day, a letter arrives saying a distant uncle has left them half a million. This could affect their accumulation plan in many different ways, including shortening their accumulation period, raising their accumulation target, reducing their saving contribution, and/or making their asset allocation more conservative. Now consider another situation. Same couple, same initial circumstances, but the value of their portfolio is reduced by half in the 2007 - 2008 crash. Once again, how does this affect their plan? Do they save more? Extend their target retirement date? Plan to accumulate less, and possibly reduce their expected post -retirement standard of living? Adopt a riskier asset allocation? Again, I don't know. And I'd guess that they don't either, until they have worked through, thought about, and discussed the alternatives (hopefully with a good adviser). Finally, consider a third person: the manager of a global macro hedge fund, whose performance is measured on a yearly basis. At the end of December, he asks your opinion on the best asset allocation. Presumably, he is interested in the allocation that will maximize his or her fund's returns over the next 12 months. But maybe this isn't the case. Maybe his brother is a sales manager who is trying to plan for his retirement asked him the same question. The point is this: until you understand all the facts and circumstances facing the person who is asking it, there is no simple answer to the question, "what is the best asset allocation today?" The glib, honest answer is, "it depends."
Is it possible to over-diversify a portfolio?
Yes, it is. And unfortunately, it happens all too frequently. Consider the person who thinks they have a well-diversified portfolio because they have a well diversified portfolio because 25% is in a corporate bond index, 25% is in a large cap value stock index, 25% is in a small cap growth stock index, and 25% is in an international stock index. Diversification works when a portfolio is exposed to a wide range of return generating factors and processes that have relatively little relationship with each other (e.g., the growth of trees, changes in real risk free interest rates, changes in inflation expectations, and growth in investors' share of corporate profits). Given this basic principle, how well diversified is the portfolio described above? I would say it is minimally diversified. To be sure, the returns on these four investments won't be perfectly correlated; however, I would also guess that a more sophisticated analytical technique (e.g., principal components analysis) would show that just one or two factors account for well over fifty percent in the variation of returns on this portfolio (e.g., such a factor might have a powerful affect on corporate cash flow, before its division between debt and equity holders). So in this sense, a portfolio can be overdiversified, or, perhaps, provide a false sense of the potential size of the likely diversification benefit. A different way of approaching this question is to measure the risk reduction benefit of adding a new asset class to a portfolio. Even when the asset classes in question have low levels of correlation with each other, the incremental diversification benefit usually declines as additional asset classes are added. So at some point, the incremental cost of adding another asset class might exceed its benefit. However, there is no hard and fast rule in this regard; it is also the case that new asset classes that become available to investors can still add substantial benefits, even if a portfolio already contains many other asset classes. In recent years, examples of this have included commodities, timber, real return bonds, and volatility.
Mohamed El-Erian from PIMCO generally recommends a heavier weighting than you do for emerging markets and foreign equities. Any thoughts?
We have two responses. The first is contained in our answer to the question above on what is the best asset allocation. If the PIMCO portfolio in question is based on a different set of investor goals and constraints than the ones we assume for a given model portfolio, I could easily explain the difference. More broadly, however, we admit to being somewhat cautious when it comes to investing in emerging markets. Undoubtedly, this has something to do with the number of years we have spent working in them, and our experience with the challenges they present for investors. These include relatively weak institutional structures compared to more developed markets (e.g., contract, property, and shareholder protection law and the effectiveness of the judicial system); the consequent preference for organizing as private companies, often in the context of extensive family groupings; quite extensive corruption that can sap free cash flow; often times weak education systems that limit the supply of skilled labor that is critical to growth; in many cases, either weak political parties and/or a history of populist and potentially destabilizing political uprising when economic conditions deteriorate; and low levels of domestic savings and consequent reliance of foreign debt which heightens the possibility of value destruction due to inflation, debt, and exchange rate crises. We don't disagree with the obvious economic logic pointing to potentially higher returns to investment in emerging markets that are long labor and resources, but short capital. However, we also have a very healthy respect for the challenges that must be overcome to realize this potential. And having seen those challenges triumph over potential gains on more than one occasion over the past thirty years (many more, in fact), we are cautious about allocating substantial amounts of a portfolio to emerging market investments.
In your timber article last month, I think you overlooked two key points. On the demand side, you have missed the very important woody biomass market (being driven by renewable fuels, global warming and energy self sufficiency issues) and its evolving impact on stumpage prices. Second, historical timber prices have risen at a real rate of about 2%. The assumption you use is too low.
Point taken about potential biomass related revenues (e.g., cellulosic ethanol). However, we chose to err on the side of conservatism here, as we did with the treatment of revenues from CO2 sequestration. On the real price change point, we accept that 1% to 2% real has been a North American TIMO rule of thumb. However, in trying to construct a valuation framework for timber as a broad global asset class, we chose to use the IMF's data for the historical evolution of global softwood and hardwood prices. Clearly, it is possible to take issue with their methodology, from weighting to quality adjustments; to cite just one recent example, new technologies that make softwood more durable should expand the range of its potential uses, and thus add to demand and upward price pressure. Again, we took the most conservative approach. The bottom line remains the same: any attempt to value timber as a global asset class is unavoidably "noisy" and uncertain; however, even a conservative approach leads to the conclusion that timber is likely undervalued over a medium term time horizon.
What do you think of Daniel Rohr's recent articles on timber on Morningstar ("If a Tree Falls in the Forest, Does It Generate and Adequate Return?")
We found them very interesting, and would strongly recommend them to our readers. As we are fond of saying, in complex adaptive systems like the financial markets, nobody has a monopoly on truth, and forecast accuracy is improved by combining analyses that are based on different methodologies. This is a perfect example of that. In this case, we agree on a number of key points, including the limitations of the NCREIF Index and the way discount rate arbitrage between industrial sellers and institutional buyers of timberland has likely driven up timberland prices and returns in the past. We also commend Rohr for his very detailed analysis of physical supply and demand in the North American forest products industry. And we agree with Rohr's investment conclusion: "We are by no means suggesting that timberland has no role to play in institutional portfolios. In our view, owing to timberland's unique characteristics [e.g., the fact that timber growth is uncorrelated with the return generating process on other asset classes], it certainly does. Rather, we would simply argue that in making asset allocation choices, decision makers are ill served by relying on timberland's historical returns." To the extent we differ, it is probably on the issue of what future returns are likely to be. At the margin, we probably see more potential upside in timber (from its CO2 sequestration benefits) than Rohr; however, in the interest of conservatism, our valuation model does not include them. Hence, like Rohr, our basic model forecasts relatively modest future total real returns from an investment in timberland, given today's dividend yields on our two proxies for this asset class, Plum Creek (PCL) and Rayonier (RYN). However, when it comes to valuation, the future returns an asset class is expected to supply is only half the equation. The other half is the rate of return an investor should demand to hold the risks inherent in that asset class (with an additional illiquidity premium if the investment is in a partnership rather than a publicly traded timber REIT like PCL or RYN). In this regard, the uncorrelated aspect of part of timber's return generating process argues for quite a low required return - we currently use a three percent risk premium over real return bonds, but there is an argument that even that may be too high. After taking both the supply of and demand for returns from timberland into account, we still believe that liquid timber REITs are likely undervalued today.
However, if we were to challenge our own position on this, we might argue that we should also add some type of company specific risk premium to reflect that fact that our proxy for timber as an asset class is based on only two companies. In our most recent analysis, this "breakeven" company specific risk premium would be 1.37%, on top of the 3.00% risk premium we already use for timber as an asset class, or 4.37% in total. If we use 3.50% as an overall long-term risk premium for the public equity market as a whole, and use the average of PCL and RYN's reported betas (1.15) to adjust it upward, we get a required risk premium of 4.03%. So at worst, using this alternative approach, it could be argued that PCL and RYN are at worst approximately fully valued, but not overvalued. Bottom line: we agree with Rohr that there remains a strong case for including timberland as a risk reducing asset class in a portfolio.