On Estimating Long-Term Returns; Fitness Versus Fatness!

From: aditya rana
Date: Sat, Jan 24, 2015 at 2:12 PM
Subject: On Estimating Long-Term Returns; Fitness Versus Fatness!


The first month of the year is an appropriate time to have a rethink about long-term return expectations on asset portfolios, and the investment advisory firm Research Affiliates is a leading light in this area, with a focus on developing asset allocation strategies which have about $177 billion dedicated to them globally. Their CIO, Chris Brightman, wrote an illuminating note recently on the 10-year return expectations for a broad variety of asset classes which is summarised below:

-Over the last 100 years, the simple 60% (equities) and 40% (bonds) portfolio has returned a very respectable nominal return of 8.4%, with equities returning 10.3% and bonds providing 5.3%. As shown in the table below, over the last 50 and 25 years the portfolio has provided even higher returns, with bonds returns improving and equity returns dropping slightly..

-This remarkable consistency of long-term returns might tempt one to extrapolate them into the future – so it is important to analyse the initial conditions which provided support for such superior returns. As the table below illustrates, both the 100-year and 25-year periods had low P/E ratios and moderate to high dividend yields, while the 50-year period had a high P/E, a low dividend yield and moderate bond yields. Yet, despite this headwind the 60/40 portfolio returned over 9% over the next 50-years.

-However, this long-term performance masks some significant variance in returns over 10-year periods (see graph below)– over the first decade of the 50-year period (’65-’74) the portfolio returned only 2.3% (nominal, minus 2.8% after inflation), with equities returning 1% and bonds 3.7%. While the portfolio subsequently rebounded, the combination of high initial P/Es and low yields provided for inferior returns over a 10-year period.

-The most recent decade (’05-’14), despite having witnessed ultra-low interest rates, a housing bubble, the Global Financial Crisis and the Great Recession, still managed to provide a respectable 7.2% nominal return (5% real). Looking at the real returns of other asset classes over the last decade (see chart below) we can clearly see that while the basic 60/40 portfolio did reasonably well (outperforming 9 out of the 16 asset classes), there were ample other high return opportunities to enhance returns.

-Returns over longer-term horizons like 10-years are relatively more predictable than returns over shorter periods which are essentially random – as the “noise” tends to dissipate over time.

-Looking ahead, using a simple model to determine return expectations based on the starting yields of the particular asset class, provides estimates not very different from those derived from using more complicated approaches. Specifically:

-The returns on sovereign bonds are equally to the starting nominal yield, as changes in the interest rate are offset by changes in the reinvestment rate.

-Credit returns are equal to the starting nominal yield minus the expected credit losses.

-Equity returns are the average of the starting dividend yield and the earnings yield – and are thus higher than the dividend yield to account for reinvestment of retained earnings and lower than earnings yield to account for dilution.

-REIT returns are equal to their starting yield.

-Commodity prices are influenced by inflation and therefore their estimated returns are equal to expected inflation.

-Historical evidence for this simple approach is illustrated by the two chart below – a) plotting the relationship between the nominal returns for various U.S. bond indices and their starting nominal yield, and, b) the relationship between real global equity returns (developed and EM) and the average of the starting dividend and earnings yield. While the relationship between stock returns and starting yields is not as strong as that between bond returns and starting yields, it’s direction is similar(i.e. higher starting yields are associated with higher returns).

-While globalization may continue to push corporate profits to even a higher share of economic output, and QE policies may continue to generate corporate cash-flow in excess of profitable investment opportunities leading to growth in earnings per share through unprecedented stock buybacks, these trends cannot last forever and are likely to mean revert over a 10-year horizon.

-Looking at the expected real returns for the 16 asset classes using the simple and reliable model (see chart below) we can calculate the expected returns for a 60/40 portfolio (1.2%), a equally weighted portfolio cross the 16 asset classes (1.6%) and a typical institutional portfolio (1.8% – see table below),

-While they utilize a more complex model than relying on just the simple initial valuation metrics, the results are similar. There are two important refinements which they make:

-1) Slower global growth than what has been seen historically, arising from the negative productivity impact of an ageing population and the need to deleverage the large amount of debt around the globe.

-2) The impact of valuations changes, which shows the U.S. trading at much higher levels (based on Shiller P/E) relative to history compared to other countries (see chart below). This implies even lower returns for the U.S. than implied by the simple approach, and the reverse for other markets.

-Interesting note which highlights the likelihood of low returns on a diversified portfolio over the next 7 to 10 years. Real returns of 1-2%, plus inflation of 2 to 3%, provides for nominal returns of 3-5% as the base case. If one is able to achieve returns in excess of this it can be attributed to a more concentrated portfolio, skill or plain old luck!

-The announcement of the much anticipated QE program by the ECB (Euro 1.1 trillion of assets to be purchased by September 2016, with the possibility of further expansion), confirms the passing of the “liquidity baton” from the Fed to the ECB. As Gavyn Davies notes, the QE program by the ECB is larger than that of the Fed when scaled to the relative size of their respective economies and bond markets. The important point to make note of is that going forward the Bundesbank no longer has an effective veto at the ECB, and that it has voted, along with the rest of the European central banks, in regarding QE as an effective monetary policy tool. This is a big deal. The focus going forward will be on fiscal and structural reform, and it is likely that the austerity programmes in southern Europe will start being rolled-back, paving the way for relatively higher growth in Europe. This is good for European assets – stocks as well as government bonds and credit markets.

Fatness versus Fitness:

Excerpts from another great piece by David Katz, Director of the Yale University Prevention centre, on the merits of fitness levels versus obesity levels.

Huff Post: 01/16/2015

-A new study about physical activity, obesity, and mortality has been propagating so many headlines and so much commentary, it’s pretty much a given that if you are seeing this, you have seen something like: "inactivity kills more than obesity."

-As is generally true in such cases, few of the people opining about the study seem to have actually read it. I suppose that’s understandable — once you get to particulars about ‘random-effects meta-analysis,’ the typical reader’s eyes glaze over. Still, it does prove helpful to read a study before interpreting it, however inconvenient our culture may find such labour. Here is what the study actually showed.

-First, going from inactive to moderately active was, indeed, associated with marked reductions in premature mortality risk across all levels of body-mass index. That sounds at first like a validation of the idea that being fit trumps being fat. But there’s more.

-In the data tables, almost all of the moderately active to active people had BMIs below 30. The numbers of those both obese but also moderately active to active were very, very low. This recapitulates a finding published before: it is possible to be fit but fat, but few people actually are. Generally, being active is associated with weight control.

-Second, the investigators looked not only at BMI, but also waist circumference — a far better measure not just of body fat excess, but its distribution. Not all overweight and obesity is a threat to health, but excess fat round the middle is. We have long known that not all variations on the theme of more weight correspond to more mortality.

-Just as they compared varying levels of activity across levels of BMI, the authors compared varying measures of waist circumference across levels of exercise. The result was that going from a high to normal waist circumference was associated with almost the identical reduction in mortality risk as going from inactive to moderately active.

-So, really, fitness and fatness both matter. Excess weight from muscle, with a lean waist, is not a risk factor for premature death. Excess body fat distributed in the lower extremities, as often prevails in premenopausal women, is also not a marker of risk. Weight around the middle is, however, and its effect on mortality appears in this large analysis to be much the same as the risk of inactivity.

-If you have excess weight around your middle, and can reduce that whether with more exercise or by eating better, you will likely reduce your risk of mortality (and morbidity) meaningfully. If you are inactive and become moderately active, you will do the same — whatever your weight. Fitness and fatness both matter.

Here’s to keeping an eye on weight and being active!




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: