Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

Evaluation Of Some Low Volatility Strategies In Canada After Half A Decade Of Existence

Summary

Since inception in Canada Low Volatility investing has produced outstanding results

There is a likelihood that the strategy concentrates on few sectors, and investors might end up being overexposed to these sectors

The history is short and the persistent outperformance appears uncertain in the future

Investors can build a simple portfolio similar to the traditional “60/40 BUY and Hold”that incorporates some Low Volatility ETF rather than being significantly exposed to it.

Low volatility factor investing has been documented in empirical finance more than four decades ago. The whole idea behind it is that stocks that exhibits lower fluctuations (or in another words safer stocks) outperform their riskier peers on a risk-adjusted basis. The idea gained popularity after the last great recession a decade ago as correlation between asset classes increased considerably compared to the pre-crisis era; thus investors started to turn to strategies that can provide significant downside protection. Also known in the industry as one of the smart beta strategies, its construction can be done through various risk measures, including market risk (Beta), total volatility (Variance), residual volatility or the minimum variance portfolio. In this article I will evaluate the performance of that strategy in Canada through few ETFs with more than five years of history. To crunch the numbers I will use R (a free open source software package and language for statistical analysis) rather than Excel. R has more advanced statistics and data sciences capabilities and is faster for this type of analysis that can take countless hours to complete in Excel. I will make the code available so that people who are interested can replicate the results for other global markets.

The table below presents a quick profile of the three leading low volatility ETFs in Canada. In addition the table shows a proxy for the entire Canadian equity market (XIC) as well as a proxy for the entire Canadian bond market (XBB).

 

As of July 2017

One quick observation is that low volatility ETFs are pricier than their plain vanilla counterparts (3 to 4 times more expensive) looking at their MERs. This is not a surprise as they are actively managed and require more work. As a consequence, one would expect them to outperform their passive peers. To test that view, I plotted the cumulative performance of the three standalone low volatility ETFs (ZLB, XMV,TLV), and a simple “buy and hold” portfolio with one rebalance each calendar year of 60% XIC (the market), 40%XBB (the bond). I also added XIC as the benchmark. TLV was issued in 2012 and is the most recent fund; thus the 5 year data starts from TLV issue date for better comparison purposes. As the following graph shows, in the last five years all low volatility funds have outstanding cumulative returns versus the index (XIC) or versus a simple buy and hold portfolio.

The more robust results on a risk adjusted basis in the following table corroborates the previous findings as all of low volatility strategies posted higher Sharpe ratios than both the benchmark and the 60/40 portfolio over the last five years. 

One of the primary expectation investors have about low volatility vehicles is protection of their assets during bear markets. The last five calendar years saw the Canadian market in bear territory only in 2015 with -8.3% return and it would be insightful to assess how these ETFs behave during that year. The following chart shows the performance of the ETFs during the 2015 bear market:

Table 2 provides further details on a risk adjusted basis during 2015 bear market: 

As one can see all low volatility ETFs posted better performance and particularly proved resilient during the bear market in 2015, which is one of their primary goals. Clearly ZLB appears as the best low volatility ETF in Canada over the last five years on the basis of its relative outstanding Sharpe ratio in both the bull and the bear markets; but the purpose of this study is not to investigate the rationale behind a single issuer’s outperformance; rather I am investigating low volatility as whole, its current status and how investors can utilize it to improve the performance of their investments. Most previous studies concluded that the success of low volatility strategies is a function of their sector bets; in other words these strategies tend to be heavily skewed towards few sectors that are characterized as stable such as Staples, Discretionary, Financials, Utilities and REITs. In Canada particularly, Financials, Utilities and REITs have currently the highest weight in the S&P/TSX Composite Low Volatility Index; thus I will consider these three sectors as the least volatile. Then I wanted to test whether that is the case for the ETFs by dissecting their sector exposure over the last five years. To achieve that, I normally need the sector breakdowns of the ETFs in the last five years. The funds profiles posted on the issuers web sites provide these data every quarter but most issuers do not keep a history of the fund profiles going back five years. Thus, due to the unavailability of some of the data, I only use the exposure according to their latest quarterly profile. Figure 1 shows that all low volatility ETFs have a higher exposure to Financials versus the other sectors. The cumulative weigh of Financials, utilities and REITs is between 47% and 69% which appears significant given that about half or more than half of the asset can potentially be invested in these sectors. As a result their performance will greatly be influenced by the three sectors.

               Low Volatility ETFs Sector Exposure as of July 2017

Due to the higher relative weight of low volatility vehicles towards some sectors, there have been concerns with regards to their continued outperformance in the future. Some prominent analysts like Rob Arnott at Research Affiliate (How Can “Smart Beta” Go Horribly Wrong?, February 2016) think that the valuation of some smart beta strategies including low volatility have become very rich, thus should underperform and even “crash” in the future. Others like Nicolas Rabener (Smart Beta vs Factors Returns, in Factor Research, August 2017) think that true factor returns is isolated through the construction of zero investment portfolios. In the case of Low Volatility ETF, that means going long the low volatility underlying securities and shorting the higher volatility ones to really isolate that factor; which is not the case in the majority of ETFs. Thus most Low Volatile Strategies mimic the factor but fail to provide the true return of that factor. Some issuers attempt to circumvent the sector exposure issue by capping the sector weight. So Investors need to scrutinize the sector exposure when picking low volatility ETFs.

Passive investors might be tempted to invest all their assets into low volatility vehicles or substitute them to the traditional 60/40 portfolio with regards to the former’s better risk/return profile as well as its relative lower correlation to the market (Table 3).

However it is crucial to remind investors that low volatility investing should not be equate to low risk investing. During periods of economic or financial stress, the correlation of these low volatility ETFs can dramatically increase and expose investors to serious drawdowns. Thus the basic principle of constructing well diversified portfolio with a range of asset classes that counterbalance each other, like equities and bonds remain important. In Canada XIC, XBB and some other ETFS have zero trading cost through few Brokers like Scotia Itrade or Virtual Brokers. There is an opportunity for Investors to build an alternative portfolio resembling a 60/40 with minimal rebalancing trading cost. That portfolio will be 40/30/30, meaning 40% will be invested in XIC (the equity index), 30% XBB (the bond index) and 30% ZLB (the low volatility). The purpose of such portfolio is to reduce the allocation to ZLB due to its higher MER and to prevent the potential sector concentration risk that we previously outlined regarding some low volatility vehicles; that portfolio will also keep the overall cost very low because in Canada XIC & XBB have zero trading cost (via some brokers) and very low MERs. Finally that portfolio will look like a traditional 60/40 while presenting a better risk/return characteristics as shown in chart3. 

On a risk adjusted basis in the following table 3, it appears obvious that such portfolio can mimic the risk of a traditional 60/40 as the two standard deviations are statistically similar and at the same time that portfolio provides a better return potential. It also helps mitigate some of the fears about low volatility as its allocation is limited to 30%.

 

Conclusion and takeaways

This brief evaluation of low volatility in Canada confirms the expectations that these vehicles are supposed to meet in theory. But the data covers only five years which is a very short period of time to make this study exhaustive. A comprehensive study should cover decades with various economic cycles. Nevertheless as these vehicle have been around in Canada for only few years, the findings can be read as a preliminary attempt to gauge the performance of these products. In a next article I will try to find whether this holds as well in other global markets in US or Europe, then where data is available I will attempt to decompose the returns of low volatility strategies against Fama & French’s Three Factors to determine the possible source of structural and persistent outperformance.


Appendix: R code to replicate the cumulative performance and risk adjusted returns

library(quantmod)

library(xts)

library(PerformanceAnalytics)

start<-"2012-08-03"

end<-"2017-09-08"

secs<-c("XIC.TO","XBB.TO","ZLB.TO","XMV.TO","TLV.TO")

getSymbols(secs, src = "yahoo", auto.assign = T, from = start,to=end)

prices <- list()

for(i in 1:length(secs)) {

prices[[i]] <- Ad(get(secs[i]))

}

prices <- do.call(cbind, prices)

colnames(prices) <- gsub("\\.[A-z]*", "", colnames(prices))

head(prices,2)

zeroprices= apply(prices, 1, function(row) all(row !=0 ))

prices<-prices[zeroprices,]

pricesreturns <- na.omit(Return.calculate(prices))

head(pricesreturns,2)

### XIC, port 60% XIC & 40% XBB vs ZLB, XMV,TLV

w.60.40<-c(.6,.4)

XIC.XBB<-pricesreturns[,c("XIC","XBB")]

port.60.40.returns<-Return.portfolio(XIC.XBB,rebalance_on="years",weights=w.60.40,wealth.index=TRUE,verbose=TRUE)

XIC.ZLB.XMV.TLV.port.60.40<-cbind(pricesreturns[,c("XIC","ZLB","XMV","TLV")],port.60.40.returns$returns)

colnames(XIC.ZLB.XMV.TLV.port.60.40)[5]<-"60%XIC/40%XBB"

charts.PerformanceSummary(XIC.ZLB.XMV.TLV.port.60.40,main="Chart1: Cumulative performance of the Low Vol ETFs vs the BM & vs the 60/40 portfolio",geometric=FALSE)

table.AnnualizedReturns(XIC.ZLB.XMV.TLV.port.60.40)

     #2015 Bear market

charts.PerformanceSummary(XIC.ZLB.XMV.TLV.port.60.40["2015"],main="Chart2: Performance of the Low Vol ETFs vs the BM & vs the 60/40 portfolio during 2015 bear market",geometric=FALSE)

table.AnnualizedReturns(XIC.ZLB.XMV.TLV.port.60.40["2015"])

#### 60%XIC/40%XBB vs 40% XIC 30% XBB 30% ZLB

w.40.30.30<-c(0.4,0.3,0.3)

XIC.XBB.ZLB<-pricesreturns[,c("XIC","XBB","ZLB")]

port.40.30.30.returns<-Return.portfolio(XIC.XBB.ZLB,rebalance_on="years",weights=w.40.30.30,wealth.index=TRUE,verbose=TRUE)

port.60.40.vs.40.30.30.returns<-cbind(port.60.40.returns$returns,port.40.30.30.returns$returns)

#head(port.60.40.vs.40.30.30.returns,3)

colnames(port.60.40.vs.40.30.30.returns)<-c("port.60%XIC/40%XBB","port.40%XIC/30%XBB/30%ZLB")

charts.PerformanceSummary(port.60.40.vs.40.30.30.returns,main="Chart3: Cumulative performance of the 60%XIC/40%XBB portfolio vs 40%XIC/30%XBB/30%ZLB",geometric=FALSE)

table.AnnualizedReturns(port.60.40.vs.40.30.30.returns)

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

Additional disclosure: I do hold few other ETFs not mentioned in the article that are manufactured by some of these companies.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it.
The portfolio is for educational purposes only, and not an actual portfolio. I am not offering investment advice. This is simply a test/illustration. Consult your financial advisor before any investment decision.