Portfolio design

Building optimal emerging market equity portfolios

Ben Ho| Jennifer Delaney |May 22, 2019

As asset allocators evaluate their emerging market (EM) equity exposures, we explore ways to combine different investment styles to build efficient EM equity portfolios.

  • Building upon BlackRock’s Index, Factor, & Alpha Framework for designing more efficient portfolios, we utilized optimization models that seek to maximize the information ratio (IR) of the portfolio (highest active return per unit of active risk).
  • Our findings suggest that Index, Factor, and Active strategies that have differentiated and uncorrelated sources of active return can be combined to build more efficient emerging market portfolios.
  • This view is consistent across allocations that are constrained to long-only exposures, as well as those that have more flexibility to relax the shorting constraint (i.e., 130-30 or absolute return-oriented strategies).

Market overview

Emerging market (EM) equities, as represented by the MSCI Emerging Markets Index, have had a strong start to 2019, rallying 10% in the first quarter1 with strong flows into both active and passive products.2 Relative to other risk assets, BlackRock finds EM equities compelling from a risk-return perspective3 and we find many asset owners reviewing their allocations to the asset class.

The client discussion is not limited to how much to allocate to the asset class; and extends to portfolio construction, including how to how best to build an emerging markets portfolio. This consideration is acute for emerging markets, as the dynamics of a changing landscape, coupled with under-representation in many portfolios and a marked increase in the number of available investment vehicles over the last decade4, has led to more options for asset allocators.

Mean variance optimization framework*

To explore what an optimal allocation within EM equities might look like, we utilized a mean variance optimization (MVO) approach that seeks to design portfolios with maximized information ratios (IR)5.

Through setting maximum IR as the objective, we can construct efficient portfolios as the optimization creates a frontier of maximum return portfolios at different levels of active risk. Our work provides options for portfolios that are constrained to long only strategies and those that include long-short and absolute return strategies. See Chart 1 below.

Inputs for each of the MVO models include6:

  1. A mix of systematic and fundamental strategies across long-only, 130/30, and absolute return
  2. Historical active risk (tracking error) for each strategy to MSCI EM Index
  3. Imputed strategy, hypothetical, excess returns over MSCI EM Index, partly based upon historical information ratios for each strategy
  4. Historical correlation of excess returns between strategies
  5. Constraint on strategy type (i.e. long-only or 130-30 and Absolute Return-oriented strategies)

We ran two optimizations: #1 (long only EM), which combines Index, Factors, and long-only EM active equity strategies and #2 (leverage constrained), which also allows for a 30% total allocation to 130-30 and absolute return strategies. The efficient frontiers from these optimizations are displayed in Chart 1.

Chart 1: Mean variance optimizations for emerging market equity portfolios

Mean variance optimizations

Source: BlackRock as of 31 December 2018. It is not possible to invest directly in an index. Please refer to the Appendix for further details on the methodology being used. *All figures shown throughout this paper are on a gross of fee basis. If performance was shown on a net of fee basis, returns would be lower and results of the optimization may vary dramatically. Please refer to the GIPS composite slides for net of fee performance for each of the underlying strategies shown. **This information is supplemental to the GIPS composite disclosure slides located at the end of this paper.

Combines index, factors and active long-only EM strategies

For our first optimization, we combine long-only index, factor and active strategies that provide broad-based exposure to global emerging market equities with the following results:

  • The optimal portfolio at each of the 2%, 3% and 4% active risk levels (see Charts 2-4 for reference) suggests allocating to multifactor, systematic and fundamental active strategies. This implies there are portfolio benefits to combining the three strategies, supported by the low correlation of their excess returns7.
  • Similarly, the low correlations of excess returns across the multifactor, systematic and fundamental active strategies limits the weighting to an index strategy. This implies that at a given level of active risk, allocating to non-indexed strategies in EM could deliver a better outcome than an index only portfolio8. (Table 1)
  • The model’s inclusion of the Multifactor strategy at each of the active risk levels supports our view that a dedicated factor allocation can complement a lineup of active managers in emerging markets.
  • While the allocation to a multifactor strategy remains relatively constant at different levels of active risk, the mix of the fundamental and systematic strategies is more sensitive to active risk tolerance. For allocators with a higher active risk budget, a larger allocation to the fundamental strategy is suggested. For those with a lower active rusk budget, however, a tilt towards the systematic strategy provides potential for excess return with lower active risk. See charts 2, 3, and 4 for reference.
Table 1: Hypothetical portfolio outcomes

*All figures shown throughout this paper are on a gross of fee basis. If performance was shown on a net of fee basis, returns would be lower and results of the optimization may vary dramatically. Please refer to the GIPS composite slides for net of fee performance for each of the underlying strategies shown. **This information is supplemental to the GIPS composite disclosure slides located at the end of this paper. +As compared against the MSCI Emerging Markets Index. It is not possible to invest directly in an index.

 

Chart 2: Hypothetical long only portfolio assuming 2% active risk

Pie chart showing hypothetical long only portfolio assuming 2% active risk

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

 

Chart 3: Hypothetical long only portfolio assuming 3% active risk

Pie chart showing hypothetical long only portfolio assuming 3% active risk

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

 

Chart 4: Hypothetical long only portfolio assuming 4% active risk

Pie chart showing hypothetical long only portfolio assuming 4% active risk

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

Expands upon the first option

In our second optimization, we extend this framework and include 130-30 and absolute return strategies. We note that while active risk is often higher for 130-30 strategies, the structure of the strategy can allow a manager to more effectively execute on its insights in a long only construct while maintaining a beta of one to the market. Absolute return strategies, meanwhile, provide the opportunity for

significantly higher alpha, alongside even higher levels active risk. For allocators who are willing to focus on a higher Sharpe ratio and rely less on active risk to the market cap index as a constraint, a larger investment in absolute return strategies may be warranted.

  • Adding 130-30 & absolute return strategies increases the slope of the efficient frontier, implying that a higher information ratio portfolio can be achieved by including these strategies in combination with long-only strategies (Table 2)
  • The information ratio begins to decrease at around 2% active risk, showcasing that adding 130-30 & Absolute Return strategies, even with a constraint, does not necessarily force an allocator to take significant active risk if the 130-30 & Absolute Return strategies are sufficiently diversifying to other strategies in the portfolio.
  • More active risk needs to be taken before absolute return strategies are incorporated into the portfolio. Allocations to these strategies are only included at an active risk of 3% or above.
Table 2: Hypothetical portfolio outcomes

*All figures shown throughout this paper are on a gross of fee basis. If performance was shown on a net of fee basis, returns would be lower and results of the optimization may vary dramatically. Please refer to the GIPS composite slides for net of fee performance for each of the underlying strategies shown. **This information is supplemental to the GIPS composite disclosure slides located at the end of this paper. +As compared against the MSCI Emerging Markets Index. It is not possible to invest directly in an index.

 

Chart 5: Hypothetical leverage constrained portfolio assuming 2% active risk

Pie chart for hypothetical leverage constrained portfolio assuming 2% active risk.

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

 

Chart 6: Hypothetical leverage constrained portfolio assuming 3% active risk

Pie chart for hypothetical leverage constrained portfolio assuming 3% active risk.

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

 

Chart 7: Hypothetical leverage constrained portfolio assuming 4% active risk

Pie chart for hypothetical leverage constrained portfolio assuming 4% active risk.

Source: BlackRock, as of 31 December, 2018. This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The above hypothetical portfolios were constructed using the historical tracking error from the respective BlackRock composite to illustrate optimal allocations based on investor risk tolerance. No investor has achieved any of the portfolio outcomes shown.

Conclusion: findings, limitations and future Work

Our work shows how Index, Factor, and Active strategies that have differentiated and uncorrelated sources of active return can be combined to build an efficient emerging markets portfolio. The mix of these strategies depends on the active risk tolerance of the allocator.

For allocators with higher active risk tolerance, a higher allocation to fundamental or leveraged strategies could be warranted. For those with a lower tolerance, however, a tilt towards systematic strategies provides potential for excess return with lower risk.

We also show that 130-30 and Absolute Return strategies can also be additive without the need to take much more active risk and can be an additional tool in building diversified emerging market equity portfolios.

We note that we did not address a common question when building an emerging markets equity portfolio, as clients increasingly ask whether or not a regional approach could be superior to a global approach. While we believe a regional approach remains an option for allocators it comes with a number of potential challenges and costs. Challenges include researching and allocating to regional managers in appropriate size, as well as the added complexity of balancing the investment between regions over time. We theorize that one solution may be to build a global emerging markets portfolio with satellite exposures focused on regions or countries that are either underrepresented in the benchmarks or have high potential for excess return. We look forward to exploring this topic further.

Jennifer Delaney
Director, Product Strategist, Global Emerging Markets
Read more
Ben Ho
Associate, Client Insight Unit
Read more

 

 

1 MSCI, Bloomberg, as of March 29th 2019

2 UBS as of 03/29/2019

3 BlackRock Capital Market Assumptions as of 31 December 2018 (https://www.blackrock.com/institutions/enus/insights/charts/capital-market-assumptions)

4 Additional information on the assumptions behind each of the above MVO data inputs can be found in the appendix.

5 Information Ratio is defined as the amount of excess return per unit of active risk (also defined as tracking error) to that manager’s respective benchmark. It is used to determine a manager’s ability to efficiently source excess return.

6 Additional information on the assumptions behind each of the above MVO data inputs can be found in the appendix of the PDF.

7 See page 6 of PDF for correlations of excess returns

8 We note that we have utilized alpha strategies that have generated alpha historically and we assume will do in future. If strategies that failed to deliver alpha were used then we would expect the index component to be more heavily featured.