I blogged recently about the Canadian Couch Potato DIY service. There are in fact more advanced similar services in the US, which is an indictment of the lack of innovation and competition in the Canadian financial services market place.
Wealthfront, fronted by Burton Malkiel is one which charges 0.25% per annum for what is an advanced automated ETF management service. This is not DIY, but then again neither is the Canadian Couch a truly DIY platform.
In my opinion, the key differentiating benefit of using a disciplined ETF based service as opposed to a non ETF disciplined based service, is cost. There are a couple of things I wanted to point about some of the performance claims of these types of services though:
Number 1 – It states that the service provides, in total from the factors noted, an additional performance of 4.6% a year – this is based on the data used, is an average not a constant, is period dependent, and the performance thereby derived is not an absolute but a relative. Total returns could easily be negative, but not as negative.
Number 2 – tax loss harvesting produces some 1% per annum. Tax loss harvesting is really a deferral of tax, and while tax is deferred it is reinvested and generates return, which of course requires returns to be compounding upwards. This a) assumes that the alternative ignores tax loss selling and b) the average from which the data is drawn is based on the period 2000 to 2011, which one would argue has produced more than the average tax loss selling opportunity. The figures do not appear to include bid/offer spreads and some information on the average historical bid offer spread of their vehicles would be useful. I would assume that bid offer spreads would be wider during periods of heavy index selling, times when tax loss harvesting may be highest.
Number 3 – They claim a 0.5% optimal allocation return benefit relative to the Fidelity allocation. Given the difficulty in obtaining “correct” return, risk and correlation inputs, I would have thought that this additional return is a random return and therefore not one which can be guaranteed. It may also be explained by the allocation structure of the options. I am not so sure why they need to assume annual rebalancing for the Fidelity allocation, given that the Fidelity fund allocation would be rebalancing irrespective.
Number 4 – Here they have drawn from David Swensen’s book whereas it would have been useful to see the annual rebalancing benefit from an explicit analysis of their own portfolios. Rebalancing benefits depends on the size and duration of the relative price movements of various asset classes, the relative width of the rebalancing benchmarks and the point at which you made your original target allocation. Selling more towards the end of an uptrend and buying more towards the end of a downtrend enhances return, whereas buying and selling equidistant would result in a sum zero outcome. This driver is not a function of automatic rebalancing, but rebalancing per se.
When individuals need to draw down on capital, a capital expenditure allocation algorithm needs to be executed: this would allocate rebalancing sales to expenditure rather than lower valuation assets. While it would subvert the rebalancing return (unknown) it would obviate the reinvestment risk of the overweight sale.
Number 5 – The point about increasing the number of asset classes against which you are rebalancing is valid: this increases the opportunity to benefit from from buying relatively low to selling relatively high. But again, actual returns are period specific and dependent on relative price movements and benchmark widths used to trigger a rebalancing. Some periods may produce no such benefits.
In all, the key differentiating benefit from using a disciplined ETF based service as opposed to a non ETF disciplined based service, is cost. The other factors are variable and relative to unknown benchmarks.