Dear friends,
Grantham, Mayo, van Otterloo (GMO) just released their latest 5-7 asset class forecasts. By way of background, these forecasts are released monthly, they "ruthlessly normalize" variables such as profit margins and p/e ratios, and they've been pretty accurate in the past.
They project the "real" (i.e., after inflation) returns for twelve types of assets, from U.S. large cap stocks to timberland. They assume 2.2% inflation over 15 years. (Grantham, personally, suspects it will be far higher.) Here are the projections, from highest to lowest return:
- Timber: 6.5% (give or take 5.5%)
- Emerging market stocks: 6.1% (+/- 10.5%)
- International large cap stocks: 4.9% (+/- 6.5)
- International small cap stocks: 4.4% (+/- 7.0)
- US high quality stocks: 4.2% (+/- 6.0)
- Emerging market debt: 1.9% (+/- 8.5)
- Cash: 0.1% (+/- 1.5)
- U.S. large cap stocks: 0.0% (+/- 6.5)
- U.S. small cap stocks: -0.5% (+/- 7.0)
- US total bond market: -1.4% (+/- 4.0)
- International bonds, currency hedged: -1.6% (+/- 4.0)
- Inflation-linked bonds: -2.7% (+/- 1.5)
Given GMO's analysis:
- emerging markets stock look to be the most profitable asset class
- emerging market securities look to be the most volatile asset classes
only TIPs have no prospect of positive real returns - their best case is a real loss of 1.2% per year
For what it's worth,
David
Comments
So, there's a bit of a mixed message here. While TIPS will do poorly if those assumptions play out, they are ultimately linked to inflation. And if inflation turns out to be "far higher" (as Grantham suspects), than the return on TIPS will be far higher as well.
It might be akin to stocks: if optimistic assumptions are already "baked in" to a stock's price (a forward P/E of 36 because investors are betting on huge and sustained earnings growth in the years ahead), then merely getting the expected good news leaves you flat and anything less than consistently good news leads to painful returns.
I might be wrong about the logic but it's the best I've got.
For what it's worth,
David
Thank you for the alert to the recent GMO forecast.
GMO, essentially Jeremy Grantham, has been making monthly forecasts of this type for many years. During this extended period they have revised their format over time just a little. For example, their original work on this matter projected 10-year returns instead of the current 7-year forecast. Grantham and GMO do great work that benefits all investors.
The firm merits a well-done acknowledgement both for the relative accuracy of their predictions and the courage to publish them in an open forum. No one has access to a perfect crystal ball: GMO is not excepted, and my version seems more cloudy than most. I certainly praise GMO for the courage and boldness to offer monthly updates. The firm exposes itself to testable predictions.
But given the longevity of the prediction (now a 7 year proposition), their risk is substantially attenuated. In the investment world, long horizon projections are an easier task. Historical data sets backstop this assertion. Over the long haul speculation plunges tend to cancel one another, and returns are mostly dependent upon fundamental market factors. Grantham’s analyses rests on the more historic and fundamental characteristics of the investment universe.
John Bogle recognized the advantages of such an approach in his 1999 classic book “Common Sense on Mutual Funds”. In that volume Bogle examines returns over decade long periods. He relates returns to fundamental components (corporate profits and dividend payments) and to a speculative element. Over a 10-year timeframe, the speculative component approaches zero because of both over-exuberance and over-caution among the investing classes.
One natural question concerning the GMO forecast is just how good it is? In general, the reviews have been exemplary on the positive side. Have out-of-sample tests been rigorously applied to challenge the accuracy of the predictions? Surely that is needed.
I have identified two references that, at least partially, performed that necessary task. One is from CXO Advisory Group and the other from a college professor. Here are Links to the two referenced documents:
http://www.cxoadvisory.com/3124/individual-gurus/gmos-stunningly-accurate-forecast/
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1701947
In general, they report favorably on the GMO work. However, some specific reservations were expressed. In particular, please take note of the cautionary comments made by the CXO Advisory Group. Statistics is a helpful investing tool, but must be used with great care.
Along the lines of interpreting statistical results with considerable care, I direct your attention to the third of your conclusions. You claim that “only TIPs have no prospect of positive real returns - their best case is a real loss of 1.2% per year”. Not so.
The results are presented in a statistical manner. I presume the TIPs estimates represent a most likely average outcome with a one standard deviation value of minus 1.5 %. Again, assuming that a normal distribution was postulated, that translates into a probability of a 3.6 % positive real return. That’s surely not very attractive, but it is somewhat more affirmative than your quick interpretation. I recognize that this is a picayune objection, but a precise rendering of statistical analysis goes with the statistical territory.
Once again thanks for the summary, but especially thank you for the website.
Best Regards.
On TIPS, negative 2.7% is labeled as "Annual Real Return over 7 Years." The +/- 1.5% is the "estimated range of 7-year annual returns." I read that as: "we expect TIPS to lose something between 1.2% - 4.2% per year for the next decade." If the loss of 1.2% represents an outlier to the right side of the bell curve - that is, accounting for a variety of factors, that's the most optimistic outcome GMO will project and it might or might not represent a one- or two-deviation boundary (they don't say) - I'm not sure how you derive a 3.6% positive real return.
Curious, as ever,
David
> I'm not sure how you derive a 3.6% positive real return.
my thoughts exactly. Maybe a tail at 99.9999999%
Thanks for reading my post; thank you again for taking time to respond.
My analysis assumes that a Normal distribution of annual real returns was postulated. That means that the future expected average real return per GMO projections for the Inflation linked Bond category is -2.7 % with a one standard deviation of 1.5 %.
Since a Normal Bell curve distribution was postulated, GMO anticipates that a negative return of -1.2% to -4.2% will be generated about 68 % of the annual data periods; that’s the area under a Bell curve from minus one sigma to plus one sigma. It does not include all possible return outcomes nor outliers.
By looking at a cumulative Normal distribution table (available in any statistical handbook), a zero or positive return at the 1.8 sigma level (that’s 2.7/1.5), is achieved 3.6 % of the time. The 3.6 % estimate is the likelihood of all positive real returns, not a guesstimate of an exact annual return.
That is the value I cited in my original posting. It is NOT any expected return; it is the probability that a positive return of any magnitude will be realized given the GMO analysis.
Accipiter, a plus 3.6 real return is hugely unlikely. Again, given the GMO forecasts, that high a return would represent a plus 4.2 standard deviation (sigma) outcome. From the cumulative Normal probability table, one would expect that to happen only 0.00133 % of the time. As a recovered Mid-Atlantic state resident, I recommend you “forgetaboutit”.
At that extreme Sigma (standard deviation) value, a Normal distribution is likely not a reliable modeling of the data or any worthwhile prediction. This is the “fat-tails” issue with most investment modeling. Avoid it whenever possible.
This exchange further demonstrates the need to read and interpret statistical analysis with extra care. The devil is indeed in the details.
I hope this clarifies my submittal.
Best Wishes.