The 2014 issue of the Credit Suisse Global Investment Returns Yearbook, has an interesting section on emerging markets https://publications.credit-suisse.com/ ... B5D14A7818 Dimson, Marsh and Staunton create an emerging market index back to 1900 (using GDP per capita as the annual sort metric to classify whether a country is developed or emerging). Here are some extracts/key findings:
On construction of the index:
- For the 23 countries in the Dimson, Marsh and Staunton (DMS) database in 1900, seven would have been deemed emerging markets [using the GDP per capita criteria] – China, Finland, Japan, Portugal, Russia, South Africa, and Spain. Three are still emerging today – 114 years later – namely China, Russia, South Africa.
Countries beyond the DMS database were also added when data became available. Other markets included were Brazil and India since 1955; Korea and Hong Kong since 1963 (the latter moved to developed in 1977); Singapore since 1966 (until it moved to developed in 1980); Malaysia since 1970; Argentina, Chile, Greece, Mexico, Thailand, and Zimbabwe since 1978; Jordan since 1978; Colombia, Pakistan, Philippines and Taiwan since 1985; and Turkey since 1987, and the MSCI Index was used since 1988.
- Outperformed developed markets 1900 to 1917
Hit by October 1917 Revolution in Russia when investors in Russian stocks lost everything
Returned less than developed markets from early 1920s to 1929 peak, but also declined by less during Great Depression.
Matched developed market returns from mid 1930s to mid-1940s
Collapsed in 1945-1949. Largest contributor was Japan equities that lost almost 98% of their market value in US dollar terms. Markets in China closed in 1949 following the communist victory, and where investors in Chinese equities effectively lost everything. Other markets such as Spain and South Africa also performed very poorly in the immediate aftermath of World War II. [Just to note the DMS index only included 6 countries over this period – China, Japan, Portugal, Russia, South Africa, Spain; Finland moved to developed in 1932).
From 1950 to 2013 emerging markets staged a long fight back, albeit with periodic setbacks. Highest frequency of crises was in 1980s and 1990s – on average, in these two decades, emerging countries averaged more than 3 times more crises per country than developed countries.
From 2000-2010 the annualized returns on MSCI EM index was 10.9% versus just 1.3% for developed markets.
Emerging markets underperformed developed markets over 1900-2013 period, but outperformed since 1950
- Annualized return (%)
Emerging markets equities = 7.4%
Developed market equities = 8.3%
Emerging markets equities = 12.5%
Developed markets equities = 10.8%
Since 1950 emerging markets outperformed but since 1900 they underperformed developed markets (the risks that showed up in 5 of the 114 years (1945-1949) where significant enough to lead to underperformance over the full 114 year period.
- Correlations of returns with developed market have increases from around 0.55 (correlation coefficient) in 1980s-1990s to about 0.8 in 2000s, and average volatility of EM country returns has declined.
- With the data they have, 2000-2013, value outperformed growth in EM by an average of 4.1%, which was larger than the 3.1% for developed markets (small caps underperformed in EM relative to developed markets). Just to note Jason Hsu's EM database which starts in 1995 also shows a large value premium in EM
From 1976-2013, quintiles sort of EM countries by dividend yield, annually rebalanced, showed the quintiles with highest yields provided an annualized return of 31%, while those with the lowest start-year yields gave an annualized return of 10%. The highest yielding countries outperformed the lowest yielders by 19% per year.
- EM are not guaranteed to outperform developed markets
EM returns had larger ‘negative tails’, and when these showed up as in 1945-1949 they had significant negative effects
Diversification matters, with only half a dozen countries in an index, bad performance of a few countries can dramatically effect aggregate performance (perhaps a lesson for the frontier market indexes)
With higher correlations with developed markets and declining volatility, EM diversification benefits have likely declined
- If standard deviations and correlations of EM and developed market equities is similar to those since 1990, then (at least by my estimate), EM returns would need to be 2% higher than developed markets to justify inclusion in a portfolio on a mean-variance basis – and only up to about a 10% allocation. Swensen in his 2009 edition of the Pioneering Portfolio Management), expects a 2% premium (as was the case since 1950). I presume his analysis of expected capital market returns, volatilities and correlations informed his suggestion to investors for a 7-14% EM allocation within equities. This is similar to what I also get for an 'optimal' allocation (on a mean-variance efficiency basis) using a 2% expected EM premium, with similar volatility and correlations to the last two decades. If we expected the volatility to be lower, and correlations to be higher, but with a continued 2% premium then the ‘optimal’ allocation would be higher, and if we expected the premium to be lower, but correlations and volatility to remain the same then the optimal allocation would be lower. Bernstein suggests (in his Rational expectations book) a 2% expected EM premium over US stocks, but a higher expected return for Non-US developed market equities than EM equities - suggesting a less than 2% premium on EM above developed market equities in aggregate.
Some have argued that EM stocks don’t add much beyond developed market small value. And if the latter have an expected premium of 2% over a DM market portfolio, then the argument has some validity on a mean-variance basis. However, if EM value stocks are used instead of an EM market portfolio, and if EM value stocks have a similar 2% premium over developed market value stocks, then it still makes sense, on a mean-variance basis, to include EM stocks.
This implies that adding an EM market allocation to a small cap and value tilted developed market portfolio may not add much to ‘expected’ mean-variance efficiency, while adding an EM value allocation would.
- Annualized return (%)
I will stick with my 13% equity allocation to EM (no plan to change that), but seems prudent to capture some of my targeted portfolio value exposure in EM.