What’s the buzz about Smart Beta?

By: Guy Fletcher, Head of Research & Client Solutions at Sanlam Investments

Every fund manager has a style – it’s how they position themselves along with their people, process and philosophy, says Guy Fletcher, Head of Research & Client Solutions at Sanlam Investments.


In the beginning, we only had passive vs active, and most active managers used fundamental analysis to identify where expected outperformance would eventuate. Some used technical analysis (observing patterns in charts) but this is now almost exclusively the purview of traders. Interestingly, it is also an area in which Artificial Intelligence may be able to play a successful role (we’ll unpack this in another chapter).


Over time, managers gave greater insights into their decision-making and, with the rise of consultants and multi-managers, became more associated with styles. Initial style separation was in the form of value (buying companies that appear undervalued on a fundamental or net asset value basis) and growth (buying companies that are likely to grow earnings faster than the average of the market). Interestingly, over half of all asset managers in South Africa state that they conform to some form of value investing, very often quoting Warren Buffet (arguably the world’s best known and most successful investor) as their influence. What is less well known is that Mr Buffett has been quoted as saying that there is no theoretical difference between value and growth investing!


Style differentiation has grown to the extent that it can be determined quantitatively. Essentially, the amount of data that we have at our disposal, plus our computing power, has allowed analysis to remove bias and to identify company attributes that are sustainable and deliver excess performance – put simplistically, a risk premium. These attributes are known as factors, and the quants are able to create composites that represent these outperforming factors, now known colloquially as smart beta.


The major (but not only) attributes that are packaged into smart beta composites are:


1)      Value – consisting of stocks that, relative to market averages, have

  1. a)lower price : earnings or price : book ratios
  2. b)higher dividend yields
  3. c)higher “margins-of-safety” (discounts to estimated worth)

2)      Growth – stocks that focus on capital appreciation

  1. a)high expected compound revenue / earnings growth
  2. b)high return on equity
  3. c)low pay-outs (company reinvests in itself)

3)      Momentum – stocks that demonstrate

  1. a)positive share price changes over a variety of periods
  2. b)positive forecast earnings revision relative to previous estimates

4)      Quality – companies that demonstrate identifiable fundamental criteria including

  1. a)consistent earnings growth with strong profitability
  2. b)high free cash flows
  3. c)competitive advantage in an expanding industry
  4. d)reliable product portfolio
  5. e)financial strength

5)      Size – smaller companies do better, on average, than their larger counterparts due to the fact that:

  1. a)it is theoretically easier for a smaller rather than a larger company to grow at a high, sustainable rate.
  2. b)the same opportunity in a smaller stock has greater potential impact than in a larger stock

6)      Low volatility – an avoidance strategy

  1. a)minimize downside risk, favour performance stability
  2. b)higher dividend yields to indicate strong balance sheets


Alternative techniques such as fundamental index construction (championed by Research Affiliates) focus on a re-weighting exercise as opposed to a style adjustment. Most indices worldwide are constructed using market capitalisation (or its derivative, free-float market capitalisation) as the sole determinant of weight. The great benefit of this technique is that, in the absence of any exogenous changes, the weight of the constituent will change in accordance with share price movements, thereby minimizing any expensive rebalancing. The downside is, the more “expensive” shares will have a greater weight pro rata than a “cheaper” share.


Fundamental indexation attempts to resolve this by weighting constituents by the size of a company’s line item (such as earnings or revenue) and ignoring its price (and thus market capitalisation). In truth, this is an adjusted value approach, and can be placed within the strategy shown above.


So, which should we choose?


The problem with this question is that it presupposes that one is better than the others. In truth, all factors are to some degree cyclical, and none of them outperform all of the time. However, each has demonstrably delivered excess performance through a full business cycle and over considerable periods. Thus the question should be posed as “which COMBINATION should one choose?”

Let’s look at some charts:


Source : Sanlam Investments, Satrix & Inet, 2017


On the face of it, the average person would choose the momentum strategy since it has delivered the highest return since the inception of our period.


We can represent this differently, by observing the shape of returns relative to the indicated benchmark, namely SWIX. To smooth out “noise” we look at this on a rolling 12-month basis:

Source : Sanlam Investments, Satrix & Inet, 2017


There are a couple of interesting observations from this:


1)      The strategies are often negatively correlated i.e. they deliver excess performance at different times and in contrast with each other

2)      All are reasonably volatile within this relative context – value has the highest extremes!

3)      The range of positive / negative relative performance can be significant

The best aspect of negative correlation is the value of diversification. Let’s build a portfolio with a naïve one-third in each outperforming strategy (i.e. excluding equal weighted) and, once again, look at it within the context of rolling relative returns:

Source : Sanlam Investments – Client Solutions, 2017


The strategy represented above has outperformed the SWIX by an average of 308 basis points p.a. over all rolling 36-month periods since December 2003. Actual experience indicates that the cost of implementation is of the order of 50 basis points p.a. – in practical terms, the client would have received an average excess of over 2.5% above the benchmark, after costs!


The outcomes can be similarly tabulated as follows:

Style Period : Dec-03 to Jun-17 % of rolling 12m periods > index % of rolling 36m periods > index % of rolling 60m periods > index
Ann. Return Ann. STDEV
Benchmark (SWIX) 16.9% 14.1%      
Strategy 20.1% 14.0% 70.2% 99.2% 100.0%
Value 18.5% 14.1% 51.7% 59.8% 80.6%
Momentum 21.0% 14.8% 80.1% 88.2% 100.0%
Quality 20.2% 15.6% 66.2% 78.7% 89.3%
Size (Equally weighted) 15.4% 14.5% 40.4% 35.4% 44.7%


What is immediately apparent is how the introduction of a composite strategy has elevated the probability of delivering excess returns without increasing risk. Interestingly, of the three chosen strategies, value has lowest probability of delivering consistent excess returns. And, despite this, it remains the chosen strategy of over half of active managers!


Let’s look at a scatter plot over the past five years:

Source : Alexander Forbes, Sanlam Investments, 2017

The period above has demonstrated a period of underperformance by the “quality” factors, modest outperformance by the “value” factors and significant outperformance by the “momentum” factors; the composite strategy has delivered excess performance at an annualised risk lower than all three independent factors.




Smart beta is another essential arrow in the portfolio construction quiver. However, it needs to be utilised with due caution for the following reasons:

1)      Individual smart beta strategies can be highly volatile; as such, strategies should generally be considered collectively in order to benefit from their low / inverse correlations

2)      Smart beta composites tend to operate in a lower risk spectrum than the typical active single managers; however, they will exist in a similar space to multi-managers

3)      To maximise diversification, strategies should be created using consistent processes to prevent the introduction of idiosyncratic risks