Browsed by
Tag: Factors

Factor Investing 5 – RA Symposium #2

Factor Investing 5 – RA Symposium #2

Today’s post is the fifth in a series on factor investing, otherwise known as smart beta. We’re going to take another look at a recent symposium hosted by Research Associates in California.

RA symposium

A few months ago we took our first look at the Research Affiliates annual advisor symposium.

  • Research Affiliates is run by Rob Arnott, and focuses on fundamental indexing, a version of smart beta or factor investing.

Last time out we looked at a couple of slide decks from the symposium.

  • Today we have another three to get through.
Three mistakes

Three Mistakes

Juhani Linnainmaa pointed out three problems with factor investing:

  1. Returns aren’t as high as people expect
    • There’s probably an element of data mining involved in the discovery of factors
    • Factors become crowded over time, driving up valuations
    • Implementation and trading costs are often neglected
  2. Drawdowns can be worse than expected
  3. The diversification benefits can sometimes vanish as correlations increase (for example, during crashes).

Factor performance

The next two tables show the long run performance of a wide range of factors (1963 to 2018).

Factor performance 2


This chart shows the impact of market cycles, which we will look at in more detail in a later presentation today:

Factor performance over cycles

Here’s the real problem – performance for the first 10 years out of sample (after publications) is less than half of that for the last 10 years of in-sample returns:

Before and after publication

That said, 3.2% pa is very much still worth having.

Recent performance

Unfortunately, recent performance (2003 to 2018) is even worse, at less than 2% pa.

Recent performance 2

It’s easier to see in the charts.

  • First, the long-run performance:

long run performance chart

Next, the chart for 2003 to 2018:

Recent performance chart

Performance is mediocre (similar to the market) until the 2008 crash, and pretty terrible afterwards.

Here’s what happened to Momentum distribution

Those are pretty extreme monthly returns, and here is the effect of them:

Momentum crash

The next table shows how abnormal recent returns have been for all the factors:

Not normal returns

It seem that all the factors have “excess kurtosis” (fat tails) and factors like Factor portfolio crash

Here’s the same chart we saw earlier for Multi-signal approach

Multi signal approach

Jim Masturzo looked at the competing objectives of investors:

Objectives

The standard solution today is an MVO (mean variance optimisation) portfolio with a tactical overlay.

  • This approach is called TAA (Tactical Asset Allocation) today, but I still cling to the terminology of my youth and refer to it as Core / Satellite.

Recipe

So far, not so factor oriented. But the next couple of slides are useful:

Bond returns

Ten-year bond returns are strongly linked to starting yield.

Equity returns

Whilst ten-year equity returns are strongly correlated to the starting earnings yield (inverse of CAPE).

Tactical strategies

The four “tactical strategies” (I’m pretty sure that’s not a thing) are:

  1. Carry (yield)
  2. Value
  3. Reversal, and
  4. Tactical returns table

    The tactical portfolio has done pretty well, with a sharpe ratio of 0.8.

    • Note the pretty high turnover, however.

    Tactical returns

    Here’s how it all fits together:

    Total return

    Market cycles

    Market cycles

    Michele Mazzoleni looked at the impact of market cycles on factors.

    Factor strategies

    Here’s a table of the strategies under consideration.

    Hedging during drawdowns

    This chart shows what a great hedge those strategies can provide during drawdowns.

    • It uses data from the five worst S&P 500 drawdowns between 1989 and 2018.

    Not a perfect hedge

    Which is not to say that factor strategy performance doesn’t vary between up and down months for the market.

    Market cycles from MAs

    Michele uses 1 month and 12 month returns to divide the market into four states:

    1. Bull (up after up)
    2. Rebound (up after down)
    3. Correction (down after up)
    4. Bear (down after down)

    Macro cycles

    Here’s how those phases relate to macro conditions.

    Factors and cycles

    And here’s how the factor strategies work in each of those phases (1971 to 2018).

    Up month performance

    Note that the factor strategies underperform during up months for the market, so it’s important not to chase returns during the good times.

    Great moderation

    Michele points out that the Great Moderation is real.

    • We’ve had fewer recessions and lower aggregate volatility since the 1990s.

    Cycle changes

    Bull phases have become more frequent, and bear markets have become deeper.

    Hedging unchanged

    But the hedging properties of the factor strategies have not changed very much.

    Conclusions

    That’s it for today, and for our detour to the RA Symposium.

    • It wasn’t all about factors, but a lot of it was.

    The key takeaways for me are:

    1. Factors work less well over time, but they still work.
    2. All factors are cyclical, and a multi-factor strategy works best.
    3. Drawdowns can be worse than expected, and all factors can crash together.
    4. Valuations still matter, and lots of things are expensive right now.
    5. Top Dogs are dangerous to own for the long term.
    6. Carry (yield), value,

      Mike Rawson

      Mike is the owner of 7 Circles, and a private investor living in London. He has been managing his own money for 35 years, with some success.

Article credit to: https://the7circles.uk/factor-investing-5-ra-symposium-2/