What do you think about the Ivy Portfolio and other forms of Global Tactical Asset Allocation?
My Reply
I don't know the details of the Ivy Portfolio but in general I have a very low opinion of tactical asset allocation, global or otherwise. It's just a fancy way of saying that the fund manager can do whatever he wants. Why would you want to give anybody a free pass to play with your capital like that?
If there was decent evidence that some fund managers are truly skilled and, more importantly, if there was a good way of identifying which ones will be skilled in the future, then I'd be all for it. I'd find my skilled fund manager and let him run the portfolio as he pleases. As long as he delivers alpha that exceeds management fee plus frictional costs of active management it will be worth it.
Unfortunately, there is next to no evidence of skill among fund managers or financial advisors. Plenty outperform, but since there are tens of thousands of them it is statistically expected for thousands to outperform. Let's be (very) optimistic and suppose there really are some that are skilled enough to generate alpha big enough to overcome the management free plus expenses. If these managers were well known, then their funds would either be (a) already closed to new investors or (b) so big that the manager's skill would not be enough to overcome the fact that the size of his trades overwhelms liquidity available in the market, thus moving the price against him. And if these managers were not well known, then (a) how can you be sure that they are skilled -- since they must not have much history or else they would be better known -- and (b) what are the chances of you coming across them ahead of millions of other investors?
In order for you to believe that your manager can provide alpha that exceeds the extra costs you have to assume that:
- A skilled manager exists.
- This skilled manager chooses to work in the relatively unrewarding and unprestigious area of mutual funds that are open to the general public instead of the much more exclusive and lucrative hedge funds.
- This skilled manager is generally unknown to other investors -- probably because he doesn't have much history.
- You have come across this skilled manager and somehow recognized his skill.
Alternatively, it may be just that your manager is:
- Same as vast majority of manager in that he claims to be skilled, is able to string together intelligent-sounding arguments for whatever, delivers before-costs returns that are randomly distributed around the market mean, and costs you around 2% annually in expenses and taxes.
Which do you think is more likely?
A follow-up email from a reader...
ReplyDelete----------------------------------
LTR, Great blog! But you didn't really answer the question on GTAA. The Ivy portfolio essentially takes your assets and puts 20% into each of 5 low cost ETFs in 'different' categories ala an Ivy endowment: US stocks, foreign stocks, total bond market, REIT, and commodities. Buy and hold, rebalance annually. The GTAA method would alternatively look at holding on the trading day of the month, and sell if the price was below it's 40 week SMA and buy [hold] if the price was above the SMA. The portfolio is to be managed by the investor using low cost ETFs in a discount brokerage, preferably in a tax-sheltered account, minimizing expenses and taxes.
Unfortunately there's no more that I can say about the Ivy Portfolio or any other GTAA model... I will not be able to mathematically prove that it will lag the market. Statistics says it has ~50% chance to beat the market before expenses and maybe ~40% chance to do so after expenses any given year. So obviously, yes, it's possible it will succeed.
DeleteBut the onus is not on me to definitively prove that this is a bad idea. The onus is on the fund manager to definitively prove that it is a good idea -- even after expenses. The fact that the fund manager is no doubt an intelligent person with a lot of experience adds zero to his case because everyone in the industry is an intelligent person with a lot of experience and collectively they will all lose to the market because of expenses.
Why is this particular model better than others and why should it be expected to beat the benchmark even after its additional headline and internal trading expenses? That is the question you should be asking. Why 40 day SMA and not 50/200MA crossover? Or lowest CAPE? Or top decile of low P/B and low PEG? Or countries with fastest GDP growth? Or countries with budget and trade surplus?
The onus is on the fund manager who will take your money to prove to you that his model has not only beat an appropriate passive benchmark in the past, has not only done so after considerable management and trading expenses, but will also continue to do so in the future. That should be your criteria for any actively managed investment, Ivy Portfolio or any other.
One reason might be that is has historically hypothetically beat the market for some time. See this paper which has considered many of the questions you asked.
ReplyDeleteAbstract:
The purpose of this paper is to present a simple quantitative method that improves the risk-adjusted returns across various asset classes. A simple moving average timing model is tested since 1900 on the United States equity market before testing since 1973 on other diverse and publicly traded asset class indices, including the Morgan Stanley Capital International EAFE Index (MSCI EAFE), Goldman Sachs Commodity Index (GSCI), National Association of Real Estate Investment Trusts Index (NAREIT), and United States government 10-year Treasury bonds. The approach is then examined in a tactical asset allocation framework where the empirical results are equity-like returns with bond-like volatility and drawdown.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=962461
Thanks for posting the link to the paper. Just from skimming it, my immediate concerns with it are:
Delete1) Why that particular time period? It starts and ends with two massive bear markets for S&P 500, EAFE, REITs, and Commodities (in 2008) -- the kind of steep drops that benefit trend-following strategies. If either of those bear market ends was not include, buy-and-hold would have outperformed for equities.
2) It ignores commissions, spreads, management expenses, and taxes (granted, last one could be ignore for tax-advantaged accounts).
3) Why 10-month SMA? I know the answer -- because it provided best results in backtesting. But *some* period will always provide best results in backtesting. What is the confidence that the 10-month SMA results are statistically meaningful as opposed to just being the one thing that ended up providing best results after exhaustively testing every possible interval?
4) Related to (3), what is the reason to believe that this pattern will continue going forward? Especially now that it's published and known to the world.
5) Commodity and possibly REIT and even EAFE results are not meaningful since those indices were not investable for much/most of the time period -- at least not without huge frictional costs.
6) Comparison to the diversified rebalanced model like the 50/50? S&P500 + 5-year T-Bonds in 50/50 provided 9.3% annual return (say, 9.2% after expenses) with maximum drawdown of 15% for the same time period. Seems to be on par or better than the timing model with a whole lot less effort and unaccounted-for expenses.
1) He's looked at a variety of out of sample time periods and asset classes and the results are fairly consistent. There are some periods of underperformance, but on average there is a risk-adjusted return benefit.
Delete2) He examines the topic of costs. Turnover is fairly low so commissions, spreads are negligible.
3) The results are insensitive to moving average period. He looked at a range from 3 months to 12 months and got broadly similar results. I expect he used 10 month because 200 days SMA is a commonly used technical indicator.
4) Momentum is not a new concept. It's been known for decades and this had not eliminated the anomoly. He speculates that it is due to investor irrationality. Of course, bull and bear markets are products of irrationality.
5) Well, okay. I don't think his results hinge on commodities and REITs.
6) Is that likely to be repeated, given expected returns for bonds at present yields? How would a portfolio of 30 year bonds fared over the same period?
Thanks for following up. Unfortunately there's just no way I will be able to spend time on this paper and simulations over other time periods to confirm to rebut what you said. Without doing so, my attitude towards this method remains very skeptical. As you said, momentum investing has been around at least for a century, but mutual funds that actually implement it have decidedly subpar returns (and maybe I'll try to make a post out of looking at them). I find it unlikely that they are/were doing it "wrong" while Mebane Faber is doing it "right" in this paper. Much more likely in my opinion it was and is a combination of torturing data until it yields a backtested strategy plus lack of appreciation for frictional costs which have been much, much higher in the past than they are now -- see http://www.longtermreturns.com/2012/12/small-cap-premium-or-liquidity-premium.html for example. Strategies work great on paper in backtesting and not so well going forward with real money. Don't forget the management fee which will likely exceed all the frictional costs combined.
DeleteAs to your last observation, I don't expect 50/50 to repeat those returns, but then I don't expect this strategy to repeat either. A portfolio of 30-years would have done amazingly well in 1972-2008 period, by the way. I don't have 30-years data (not even sure 30-years treasuries were issued back then), but 20-years did nearly as well as stocks with much less volatility: http://www.longtermreturns.com/p/historical-investment-returns.html?fromyear=1972&toyear=2008&real=0&log=1&plots=500_Corp_20yr&customplots1=none
And here's a quick overview of performance of momentum funds I know of that are open today: http://www.longtermreturns.com/2013/01/momentum-mutual-funds-review.html
DeleteAnother strike against Mebane Faber's global tactical asset allocation is this ETF managed by him: http://quote.morningstar.com/ETF/chart.aspx?t=GTAA®ion=USA&culture=en-us , trailing every conceivable benchmark
Delete