When it comes to the multi-asset income offerings on the market today, the investment universe is often more important than a robust investment process.
We see below that investment universes vary considerably from one ETF to the next.
There is not a single exposure that is shared across all ten ETFs that we considered. The most common exposures are U.S. and international equities (most often with a bent towards dividend stocks), U.S. REITs, preferreds, MLPs, and high yield bonds.
Convertible bonds, TIPS, bank loans, and mortgage-backed/asset-backed securities only make an appearance in one ETF each. We find this surprising since each of these exposures offer unique structural characteristics that make them very useful diversifiers.
But apparently the term “multi-asset” does not always mean “diversified”.
Unsurprisingly, large differences in each ETF’s investment universe create similar disparities in risk and return metrics. For the year-ended April 2016:
- Return ranged from -17.6% (GYLD) to +1.5% (FDIV)
- Annualized volatility ranged from 8.2% (IYLD) to 22.8% (GYLD)
- Maximum drawdown ranged from 10.8% (INKM) to 37.8% (GYLD)
- Correlation to the SPRD S&P 500 ETF (ticker: SPY) ranged from 0.51 (DVHI) to 0.94 (INKM)
- Correlation to the iShares Barclays Aggregate Bond ETF (AGG) ranged from 0.11 (DVHI) to 0.48 (IYLD)
Reminder #3: Beware static approaches.
Investors looking for a one-stop, managed solution to increase overall portfolio yield should be wary of multi-asset income ETFs with a static allocation approach.
The optimal path towards achieving a given income objective will vary considerably over time.
Take MLPs as an example. At their peak in August 2014, MLPs were yielding 4.3%, the same yield offered by long-term investment grade corporate bonds. Fast-forward to February 2016 and the yield on MLPs had skyrocketed to 11.0% while corporate yields only increased modestly to 4.3%.
When the risk/reward trade-off for a given asset class changes this dramatically, it certainly is worth reconsidering the sizing of that position. MLPs at a 4% yield are a totally different animal than MLPs at an 11% yield. Static processes, without the ability to adjust position sizes, do not afford themselves the ability to take advantage of evolving yield and risk dynamics.