Portfolio Management

An Investment Strategy based on  Recent Sector Performance

11/24/2014

Introduction

 This report will evaluate a sector rotation investment strategy by back testing using a risk adjusted performance evaluation and other fundamental analysis.  To test the strategy, exchange traded funds (ETFs) offered by Vanguard will be used.  Vanguards ETFs were deemed preferable because the company does not charge trading fees for its own ETFs if you use a Vanguard brokerage account.  The elimination of such trading fees can make a very significant difference in portfolio performance over time given that most brokerage firms charge at least $4.50 per trade.  Vanguard brokerage services charges $7.00 per trade, for non-Vanguard ETFs, stocks, or mutual funds.  However, Vanguard’s offerings include the main market index, sector, and international ETF types and offer the lowest expense ratios in the industry.  Therefore, it is possible to build a portfolio mainly consisting of Vanguard offerings suitable for the majority of individual investors.  Leaving the free Vanguard advertisement behind, the following paragraphs will discuss the actual strategy to be evaluated.  The strategy is not very difficult or time consuming to implement.  Most importantly, the question this report hopes to answer is if implementing such a strategy adds alpha to a portfolio.

            Vanguard offers ETFs that cover most of the main industry sectors that make up the U.S. economy.  This portfolio strategy involves purchasing those sectors that underperformed by the most and selling those industries that outperformed the most in the prior quarter.  The idea is that an investor would be purchasing the underperformers at relatively low valuations while selling those industries that may be overvalued because of their recent run up.  This approach goes along with several well known investment ideas, such as the concept that prices tend to revert to the mean over time.  There is a related strategy for stocks in which investors purchase the worst performing stocks in the Dow Jones Industrial average of the prior year.  The expectation based on some historical data is that these stocks will outperform the overall Dow Index in the following year.  As is often the case in the markets, investors tend to be overly exuberant on in good times, and then overly pessimistic when the market is declining. 

Another concept that that was also considered is purchasing those industries that have the lowest price to earnings multiple (PE ratio) and sell those with the highest PE ratios.   This strategy is more difficult to implement in practice and more difficult to back test.  For one, different industries historically trade at vastly different PE ratios.  For example, technology stocks always trade at higher valuations than utilities because of the higher growth forecasts associated with the technology industry.  Technology stocks are often fast growing innovative companies with room for expansion while utilities largely operate in a mature and regulated marketplace.  Therefore, using PE ratios alone would lead an investor to only purchase certain industries missing out on opportunities such as mispriced technology stocks.

Data

The table below provides the various Vanguard sector ETFs offered that will be used in this analyses.

Name		Symbol	Expense  ratio	As of 11/24/2014		Average annual total returns as of 09/30/2014*
				SEC yield	YTD returns	1 year	3 year	5 year	10 year	Since inception
1	Consumer Discretionary ETF	VCR	0.14%	1.24%	7.12%	10.77%	26.43%	21.29%	9.59%	8.62%
										(01/26/2004)
2	Consumer Staples ETF	VDC	0.14%	2.38%	14.47%	15.91%	18.42%	15.67%	11.02%	10.30%
										(01/26/2004)
3	Energy ETF	VDE	0.14%	2.19%	-0.17%	11.76%	16.99%	12.19%	11.47%	11.71%
										(09/23/2004)
4	Financials ETF	VFH	0.19%	2.09%	11.92%	15.94%	25.68%	11.29%	1.50%	1.38%
										(01/26/2004)
5	Health Care ETF	VHT	0.14%	1.19%	24.81%	26.96%	29.21%	20.03%	10.99%	9.60%
										(01/26/2004)
6	Industrials ETF	VIS	0.14%	1.68%	9.11%	14.70%	25.43%	17.16%	8.96%	9.12%
										(09/23/2004)
7	Information Technology ETF	VGT	0.14%	1.25%	18.16%	25.39%	22.05%	15.83%	9.91%	7.25%
										(01/26/2004)
8	Materials ETF	VAW	0.14%	1.84%	8.91%	17.86%	21.90%	13.71%	9.34%	9.75%
										(01/26/2004)
9	Telecommunication Services ETF	VOX	0.14%	3.17%	6.41%	12.14%	16.88%	13.35%	8.68%	8.65%
										(09/23/2004)
10	Utilities ETF	VPU	0.14%	3.22%	21.11%	15.73%	12.63%	12.30%	9.38%	9.55%
										(01/26/2004)

                                                                     Table 1 Vanguard Sector Specific ETFs

Due to the inception dates provided in Table 1, the analyses that follows will be based on data spanning from October 2004 through October 2014 as all of the funds have data for this period.  Additionally the performance of this investment strategy will be compared to the performance that could have been achieved by simply investing in Vanguard’s Total Stock Market ETF (VTI).  This ETF tracks the performance of the entire U.S. stock market using market capitalization weightings.      

Name		Symbol	Expense ratio	As of 11/24/2014		Average annual total returns as of 09/30/2014*
				SEC yield	YTD returns	1 year	3 year	5 year	10 year	Since inception
1	Total Stock Market ETF	VTI	0.05%	1.82%	12.75%	17.75%	23.06%	15.84%	8.62%	6.03%
										(05/24/2001)

                                            Table 2 Vanguard Total Stock Market ETF (VTI) Metrics

The portfolio will start off with the weightings in each sector as defined by market capitalization of the sector.  An important rule is that the portfolio will not deviate from these weightings by more than 30%.  This is done to avoid over allocation to a given sector which would add considerable risk to the portfolio.  On the other hand, it will also prevent the portfolio from deviating from the returns of the market to some extent.

	Total Stock Market ETF
	as of 10/31/2014
Basic Materials	2.90%
Consumer Goods	9.80%
Consumer Services	13.00%
Financials	18.80%
Health Care	13.20%
Industrials	12.70%
Oil & Gas	8.30%
Technology	15.80%
Telecommunications	2.20%
Utilities	3.30%

                                                        Table 3 Equity Sector Weightings

Every quarter, 30% of each of the top three sectors will be sold and that money will be evenly divided to purchase the worst three performing sectors for the quarter.  Again, note that the sector weightings in the portfolio will not be allowed to fall outside of the allowable range provided in Table 4.  If all three of the worst performers have reached their maximum allowable range, then the remainder of the money will be invested in the fourth worst performer of the previous quarter and so on. 

This paper will also evaluate if this strategy would have worked better if this asset reallocation were performed semiannually rather than quarterly.  While the reallocation frequency will be change, the evaluation will still start off with the sector weightings representing the actual market which means that the portfolio is initially recreating VTI.  This paper will assume a starting portfolio value of $100,000.

Name		Symbol	Initial Weighting	Initial Amount	Allowed Range (max/min weightings)
1	Consumer Discretionary ETF	VCR	11.40%	$    11,400.00	7.98%	14.82%
2	Consumer Staples ETF	VDC	11.40%	$    11,400.00	7.98%	14.82%
3	Energy ETF	VDE	8.30%	$      8,300.00	5.81%	10.79%
4	Financials ETF	VFH	18.80%	$    18,800.00	13.16%	24.44%
5	Health Care ETF	VHT	13.20%	$    13,200.00	9.24%	17.16%
6	Industrials ETF	VIS	12.70%	$    12,700.00	8.89%	16.51%
7	Information Technology ETF	VGT	15.80%	$    15,800.00	11.06%	20.54%
8	Materials ETF	VAW	2.90%	$      2,900.00	2.03%	3.77%
9	Telecommunication Services ETF	VOX	2.20%	$      2,200.00	1.54%	2.86%
10	Utilities ETF	VPU	3.30%	$      3,300.00	2.31%	4.29%
Total			100%	$  100,000.00

                                                           Table 4 Portfolio Holdings

            The sector weightings for Consumer Services and Consumer Goods from Table 3 were equally divided among the Consumer Discretionary and Consumer Staples ETFs.  Certain sectors will outperform in a given quarter while others will underperform.  The question is if this outperformance will be captured following this investment strategy.  Sometimes sectors will deviate considerably from the overall market.  For example, currently the energy sector is experiencing a significant decline in market value of about 16% driven by a drop in the price of oil.  This is a result of excess supply on the market due to increased production in the U.S. from a technique known as fracking on shale formations.  Additionally, the biggest oil producer, Saudia Arabia, is not decreasing production due to the lower prices as the country does wants to maintain market share.   At the same time the energy sector has gone down considerably the health care sector boasts an impressive 17% return.  Various sector performances compared to VTI for the last six months are seen in the Figure 1 below.  Sectors that performed in line with VTI during the period were removed from the graph to make the graph readable.

                          Figure 1 - Sector Performance for the last 6 Months

 

The correlation matrix in Table 5 shows that all of the sectors ETFs, with the exception of the Utility ETF (VPU), are fairly highly correlated with the overall market (VTI).  This is due to the highly regulated nature of the utility industry.  

	VTI	VCR	VDC	VDE	VFH	VHT	VIS	VGT	VAW	VOX	VPU
VTI	1.00
VCR	0.93	1.00
VDC	0.82	0.75	1.00
VDE	0.72	0.54	0.50	1.00
VFH	0.89	0.86	0.74	0.49	1.00
VHT	0.81	0.70	0.76	0.46	0.72	1.00
VIS	0.95	0.91	0.77	0.65	0.87	0.72	1.00
VGT	0.91	0.86	0.67	0.63	0.73	0.68	0.84	1.00
VAW	0.91	0.83	0.67	0.77	0.75	0.66	0.89	0.84	1.00
VOX	0.78	0.71	0.68	0.53	0.63	0.62	0.72	0.74	0.70	1.00
VPU	0.58	0.43	0.61	0.52	0.41	0.56	0.49	0.46	0.48	0.61	1.00

                                              Table 5 Correlation Matrix

Results

            The results of implementing this strategy starting in the fourth quarter of 2004 through the third quarter of 2014 are pictured in the Figure 2.  This graph compares the outcome of investing $100,000 at the beginning of the period using the strategy described versus simply buying and holding VTI.

                                                                    Figure 2 - Strategy compared to VTI

            The test portfolio ends the 10 year period with $241,126.78, while simply investing in VTI leaves the investor with $228,739.27 at the end of the period.  The investor in the test portfolio had a 141.13% return over ten years or 14.11% average annual return.  VTI provided a 128.74% annual return over 10 years or a 12.87% average annual return.  The investor that implemented the sector rotation strategy had an additional $12, 387.51 at the end of the period or made an additional 12.39% return.

            If this strategy of buying the underperformers and selling the outperformers was implemented performing the reallocations semiannually rather than quarterly the results would not be as good.  This is based on data from the time period going from the second quarter of 2005 through the third quarter of 2014.  The semiannual reallocation portfolio resulted in an ROI of 115.68% while quarterly reallocation resulted in a 120.99% return over the time period.  Both of these portfolios outperformed simply buying and holding VTI which resulted in a 112.33% return.  An investor that started with $100,000 and implemented the strategy discussed in this paper and rebalanced the portfolio semiannually would have about $8660 more compared to a buy and hold investor in VTI. 

                      Figure 3 - Sector Performance Strategy VS VTI

            Some key metrics for the portfolio using quarterly rebalancing are presented in Table 5.  VTI was used as the benchmark when calculating these metrics rather than the S&P 500 since this is what this paper is using for comparison.  This is important since VTI has slightly better performance than the S&P 500 which may be attributed to it also holding small cap stocks.

Column1	Portfolio	VTI
Annualized Sharpe Ratio	0.5449	0.5175
Alpha	0.0013	0.00
Beta	0.9947
R-Squared	0.9894

                                        Table 6 Key Metrics with VTI as the Benchmark

            The sector rotation strategy results in a slightly higher Sharpe ratio indicating slightly higher returns for a given unit of risk compared to VTI.  None of these metrics deviate from VTI greatly.  This makes sense given that the constraints on the portfolio prevent a deviation of greater than 30% for the weightings of any sector.  Therefore, the R-squared value indicates that 98.94% of the portfolio’s performance is explained by the price movements of VTI or the overall market.  Had the constraints allowed a greater deviation in sector weightings compared to the actual market, the results would have showed less correlation.  It is unclear if this would have been beneficial to the portfolio as this would have likely added risk due to possible over exposure to a given sector.  The low alpha number indicates a slight outperformance of the portfolio on a risk adjusted basis.  A beta of nearly one indicates that the portfolio’s value will move with the market as is seen in Figure 2.

It should be noted that every one of the 10 major sectors evaluated was included in the list of worst and best performers for the quarter multiple times throughout the period evaluated.  The only sector that was included in either the best three or worst three sectors for more than four consecutive quarters was financials (VFH).  This included the time period of the financial crisis.  VFH was among the worst three performers from the first quarter of 2007 through the second quarter of 2008.  Financials managed to avoid the list of worst three performers in the third quarter of 2008, but quickly returned to the list for the fourth quarter and first quarter of 2008 and 2009, respectively.  During this extended period of underperformance for the financials the maximum weighting of 24.4% (Table 4), prevented an over allocation to the sector.  However, given that financials are the largest sector in the U.S. economy, the portfolio was not spared during the financial crisis as seen in Figure 2.  The overall market value of the financial sector is still about 25% lower than in April of 2007.  

Conclusion

            Based on the results presented, it appears there may be some value in implementing a sector rotation strategy similar to the one described in this report.  This strategy allows an investor to benefit from the swings in various sectors.  This outperformance by some industries over given periods is not captured to the same extent by simply owning the market.  This strategy also goes along with a common market theory dealing with reversion to the mean.  During certain periods a hot sector may be driven to lofty valuations.  Similarly, market participants often tend to be overly pessimistic on the prospects of a sector that is simply experiencing a rough patch or a cyclical downturn.

            This strategy also somewhat limits risk by enforcing bounds on what percentage of the portfolio can be put into a specific sector.  In this way the investor dos not deviate too much from the overall market.  This may be important in the case of a given sector experiencing a long term decline which would result in the portfolio lacking diversification if the weighting limits were not enforced.  This is because the portfolio would dangerously overweight the sector after a sustained underperformance lasting multiple quarters.

            One negative aspect of implementing this strategy is that the sector ETFs have higher expense ratios than VTI.  This is because the assets under management are much larger for the broad market ETF compared to the sector specific ETFs.  The sector ETF expense ratios are almost three times that of VTI (0.14% vs. 0.05%).  Another factor to consider is the capital gains taxes that will be incurred due to more frequent selling.  A buy and hold investor would incur no taxes over a given year if they do not sell while the investor implementing this strategy would pay the short term capital gains rate on some of their gains (unless the funds are in a retirement account). 

One way to alleviate the tax concern is to simply buy given sectors after underperformance as described, but leave out the selling part of the strategy.  Most investors that are still earning money will continue to contribute to their portfolios over their working lives.  Since they would be investing at various intervals, the investments could be made by buying the underperforming sectors over the previous quarter while considering the weighting boundaries.  Selling the outperformers would not be required in this case although it is unclear if this strategy would perform as well but could be back tested in a separate study.

            A follow on topic that could be evaluated is implementing a strategy similar to this in other areas of the market.  For example, a similar strategy could be implemented for international markets.  The countries with outperforming markets would be sold to buy the underperforming countries over a given time period.  An investor can gain exposure to most of the developed, emerging, and frontier markets through ETFs.  However, the issue of trading fees would diminish the chances of outperformance given that these ETFs are offered by various other firms.

            In summary, the rotation strategy described in this paper may allow some investors to achieve slightly better results with their portfolio.  It may also be refined as different time periods may be looked at among other considerations.  Similarly, the idea of basing purchasing decisions on recent sector performance as described while omitting the selling part of the strategy is of significant interest.  This would eliminated the tax issues and simplify the process for investors still adding money to their portfolios.  The impacts on performance could be evaluated in a similar manner.