Monday, April 20, 2015

Buy and Forget

For those people who don't want to actively manage their portfolios, I usually recommend VT and ACWV to ppl as passive buy-and-hold core holdings for the stock allocations of their portfolios. These are global stock ETFs, which means that they hold basically all the stocks that are publicly trade-able. The main difference is their weighting methodology.

VT is market-cap weighted. That means the stocks held in it are weighted by the size of the company itself. This is the return of the (dollar-weighted) average investor. It is mathematically impossible to always and consistently beat this benchmark (otherwise, you eventually become the market).

ACWV has a higher allocation to low-risk or low-volatility stocks (here we use risk and volatility interchangeably). The correct way to evaluate the worthiness of an investment is to look at return divided by risk (you should NEVER look at solely the return with no regard for the risk). If an investment B has two-thirds the return of investment A but also two-thirds the risk, then you should be ambivalent between the two investments. However, if investment C has two-thirds the return but one-half the risk of investment A, then you should strictly prefer investment C to investment A.

The theory behind ACWV is that ppl are biased towards buying high risk / high reward stocks and structurally overpay for them. When you bid up the price of an asset, you reduce its potential return. Thus, high risk / high reward stocks have their return reduced below what efficient markets would justify.

Why does this bias exist? There are several theories.
  1. The paradigm described above (our analysis with investments A, B, C) is only true if you have access to risk-free leverage. This is obviously not true in the real world. Due to this leverage constraint, investors who have higher risk tolerance, rather than borrowing 50% to buy a 150% allocation of stocks will instead just buy 100% allocation of 1.5x more volatile stocks.
  2. Another theory is that investors solely look at potential return and do not properly account for potential risk. In this scenario, investors will prefer an investment with 1.5x the return but 2x the risk.
  3. Finally, there is theory that investors have behavioral preference-for-gambling and will pay extra for lottery-like binary outcomes. Regardless of which theory is correct, the empirical data confirms that the low-vol anomaly exists.
Thus I think ACWV should outperform VT on a risk-adjusted basis over the long run. Investors should be careful to adjust their asset allocation to take into account the lower risk of ACWV (e.g. rather than a 60-40 stock-bond allocation, a 70-30 or even 80-20 stock-bond allocation might be more appropriate when using ACWV for the stock allocation).

No comments:

Post a Comment