A research paper about my paper. Cool! I haven’t had the chance to read it yet…
—-
Draw your own conclusions on endowment performance last year, fiscal year ending June 30th. Below are facts.
Bonds +7%
60/40 -12.7%
Texas -13%
Penn -15.7%
Columbia -16.1%
MIT -17%
Williams -18%
Amherst -20%
Brown -23%
Yale -24.6%
Stanford -25.9%
US Stocks -25.95%
Cornell -26%
Harvard -27.3%
Foreign Stocks -31%
Buy and Hold -31.1%
Real Estate -41%
Commodities -61%
(Data sources: Global Financial Data, Harvard and Yale Annual Reports)
US Stocks – S&P 500
Foreign Stocks – MSCI EAFE
Bonds – 10 Year US Govt
Commodities – GSCI
REITs – NAREIT
Buy and Hold is an equally-weighted, monthly rebalanced allocation to the above 5 asset classes
Harvard and Yale are announced returns for 2009
60/40 is the traditional 60% stocks, 40% bonds allocation
I quickly read the paper. WTF are they talking about?? I guess I'm not a quant but usually I can dope out just about any research paper for the gist. But this may have lost something in translation. I have no idea if they confirm or deny you system, Meb. can someone tell me what “bootstrapping” means in the context of this paper?
Are they saying the returns on the SP using 10 month MA are within the realm of pure chance?
If so, what would they estimate the odds of the 5 asset classes all showing spurious returns using the same, unoptimzed system?? My stat tells me the odds would be very small.
Help me out here.
see page 2, footnote #1. the #1 ranking is impressive!
basically they're saying that he's onto something but further research is needed and, imho, they're acknowledging that EMH might be suspect.
Does the term “bootstrapping” apply to resampling assumption methods?
Many believe Malkiel's EMH becomes a do-do bird at times.
Some of the words were transposed leading to some confusion, and this quote is perplexing:
“These hypothesis are intriguing and constitute a possible research agenda”.
It should probably read “hypotheses” and I hope I am interpreting it that “possible research agenda” hints at further research and not bias.
Steel always becomes stronger when termpered by fire.
Charts show Meb is consistently left of center which is usually a good thing. LOL.
Bootstrapping is a way of calculating a confidence interval. Like a 95% interval. Basically the paper says that Faber's results are within the interval. That is, the good performance may be due to chance.
The essential reason for this result is (my opinion) that Faber does not have enough data yet to prove his method. And the main reason he needs more data is that his sample has only about 3 bear markets in it. In a way he only has 3 data points.
nick is absolutely correct-sort of. while the data points are few say on the sp 500, if one expands that to the other markets traded the confidence level rises. I'm far more concerned about the popularity of the approach. too much attention will be its doom. fortunately (depending on your perspective) the market actually outperformed the model from post 1973 to 1999 so as (if) the market outperforms the model in the future it will lose its popularity (until of course it blows up again)
the article posted by Meb (I think) by Theodore Wong, Moving Average: Holy Grail or Fairy Tale Part 1 published by Advisor Perspectives seems to reach back to 1871 and confirms the method using a range of lengths.
http://www.advisorperspectives.com/…/Moving_Average-...
rsmlp is correct-sort of. Expanding to other markets does add a bit of confidence but only a little because all the markets are VERY correlated. Regarding the second point – if the model outperformed the market ALWAYS then there would be a risk-free arbitrage opportunity and, sorry, that's not allowed. So the model MUST underperform the market occasionally. So doom prediction gets tricky.
actually inoddy the markets ARE NOT correlated. not sure how you measure correlation but commodities are not correlated to stocks and so on. sure, during a complete meltdown everything that trades pretty much tanks but that's extremely unusual. I think everyone but the complete naive expect the model to USUALLY underperform the sp. the key is that the we also expect the sp to rise over the long run and we're content with similar returns with less drawdown. I will be very surprised if the classic 5 asset 10 mo ma model does not underperfrom a bh of the sp from 2009-2018-and quite happy if it still makes 10% or so.
This is very interesting, but it would be great to see how all of these endowments have done over the last ten years. The worst performers this year could still be the best performers over the long term.
Hrm. I disagree Name.
What I got out of the paper was less that “he's on to something” and more that the Ivy model falls within the expected range for a random distribution of possible market outcomes. It happens to beat the market with historical data, but it could just as easily fall below in different market conditions. There's no intrinsic bias above the efficient frontier in this approach (which is a bummer, since I've just implemented the Ivy Portfolio in my retirement account).
But I'm very curious to hear Mebane's take on it – I could easily be misunderstanding the paper.
However, I freely admit that I did not understand the paper's conclusions regarding volatility and drawdown risk.
http://dshort.com/articles/2009/diversification...
I wonder how *any* strategy could be proven fail safe.
actually inoddy the markets ARE NOT correlated. not sure how you measure correlation but commodities are not correlated to stocks and so on. sure, during a complete meltdown everything that trades pretty much tanks but that's extremely unusual. I think everyone but the complete naive expect the model to USUALLY underperform the sp. the key is that the we also expect the sp to rise over the long run and we're content with similar returns with less drawdown. I will be very surprised if the classic 5 asset 10 mo ma model does not underperfrom a bh of the sp from 2009-2018-and quite happy if it still makes 10% or so.
This is very interesting, but it would be great to see how all of these endowments have done over the last ten years. The worst performers this year could still be the best performers over the long term.
Hrm. I disagree Name.
What I got out of the paper was less that “he's on to something” and more that the Ivy model falls within the expected range for a random distribution of possible market outcomes. It happens to beat the market with historical data, but it could just as easily fall below in different market conditions. There's no intrinsic bias above the efficient frontier in this approach (which is a bummer, since I've just implemented the Ivy Portfolio in my retirement account).
But I'm very curious to hear Mebane's take on it – I could easily be misunderstanding the paper.
However, I freely admit that I did not understand the paper's conclusions regarding volatility and drawdown risk.
http://dshort.com/articles/2009/diversification...
I wonder how *any* strategy could be proven fail safe.