Decoupling or Increased Correlation Across the Globe
1 like183 views
The document analyzes the correlation among major asset classes, sectors, and countries while investigating the hypothesis of decoupling in global business cycles. It concludes that while correlation has generally increased since the 2008 financial crisis, stock pickers still benefit from diversification at the individual stock level. Key findings include that developed and emerging markets are more synchronized now, with diversification benefits diminished across asset classes and sectors.
Decoupling or Increased Correlation Across the Globe
- 1. Presented by: Andrew Stotz, CFA 25 May 2016 Decoupling or increased correlation across the globe?
- 2. 25 May 2016 2 Ask a research question Summarize previous research Formulate a hypothesis to test Select a relevant data set, remove errors and outliers Formulate a methodology to test the hypothesis Present and analyze the results Advise how to apply to improve investment decisions Our academic-style research format Source: A. Stotz Investment Research Hypothesis Method Results Question Review Action Data
- 3. 25 May 2016 3 Hypothesis Method Results Question Review Action Data Correlation between three major asset classes Source: A. Stotz Investment Research What benefit can you get from diversifying across: • The three major asset classes? • Emerging and developed markets? • The ten major sectors? • Countries? • Stocks? Does it add more or less diversification benefit now compared to the past?
- 4. 25 May 2016 4 Hypothesis Method Results Question Review Action Data Investigated the degree of global cyclical interdependence during 1960-2005 Three groups: Industrial countries, emerging markets, and other developing economies Concluded that business cycles within the groups had converged, but became less synchronized between industrial economies and emerging markets Mixed Kose, Otrok, and Prasad (2008): Mixed conclusion Sources: A. Stotz Investment Research, Kose, M.A., Otrok, C. and Prasad, E.S., 2008. Global business cycles: convergence or decoupling? (No. w14292). National Bureau of Economic Research.
- 5. 25 May 2016 5 Hypothesis Method Results Question Review Action Data Investigated links between banks in developed countries and emerging market bond, stock and credit markets during the 2008 global financial crisis Conclusion: No decoupling at time of crisis. Liquidity and bank solvency measures had become more correlated with emerging market bond, stock and credit markets during that crisis period Highly correlated Frank and Hesse (2009): High correlation between EM and DM assets Sources: A. Stotz Investment Research, Frank, N. and Hesse, H., 2009. Financial spillovers to emerging markets during the global financial crisis (No. 9-104). International Monetary Fund.
- 6. 25 May 2016 6 Hypothesis Method Results Question Review Action Data Looks at deviations from GDP trends rather than actual GDP growth rates on data from 1980-2008 Rejected decoupling in the years before the global financial crisis, and concluded that business cycle interdependence had become stronger Highly correlated Wälti (2012): Rejects decoupling Sources: A. Stotz Investment Research, Wälti, S., 2012. The myth of decoupling. Applied Economics, 44(26), pp.3407-3419.
- 7. 25 May 2016 7 Hypothesis Method Results Question Review Action Data Tested decoupling in two ways in 1993-2010: Real decoupling – Sensitivity to business cycles and other economic factors such as GDP, and Financial decoupling – Sensitivity of emerging market assets (credit, equity and FX returns) relative to a global market portfolio Concluded that business cycles in emerging markets have gradually decoupled from developed markets Rejected financial decoupling as they could see high and increased ‘comovement’ in EM assets Highly correlated Yeyati and Williams (2012): Rejects financial decoupling Sources: A. Stotz Investment Research, Yeyati, E.L. and Williams, T., 2012. Emerging economies in the 2000s: Real decoupling and financial recoupling. Journal of International Money and Finance, 31(8), pp.2102-2126.
- 8. 25 May 2016 8 Hypothesis Method Results Question Review Action Data Looked at US (developed) and Poland (emerging) GDP growth rates’ dependency on the underlying long-term GDP trend Compared deviations from the long-term GDP trend Conclusion: No evidence to support decoupling, the Polish economy was well synchronized with that of the US Highly correlated Wyrobek and Stańczyk (2013): High correlation between EM and DM Sources: A. Stotz Investment Research, Wyrobek, J. and Stańczyk, Z., 2013. Decoupling hypothesis and the financial crisis. European Financial Systems 2013, p.362.
- 9. 25 May 2016 9 Hypothesis Method Results Question Review Action Data There are diversification benefits (relatively low correlation) to gain from investing in different asset classes, regions, sectors, countries and stocks Correlation will have increased over time in all the above cases Hypothesis Source: A. Stotz Investment Research
- 10. 25 May 2016 10 Hypothesis Method Results Question Review Action Data We collected price data for MSCI indices from December 1994 to September 2015 Indices for bonds, commodities, and world equity Equity indices for developed markets, emerging markets, global sectors, and countries For company level correlations we used a data set of about 27,000 companies listed anywhere in the world beginning in March 2002 and ending in May 2015 MSCI indices and a universe of 27,000 international stocks Source: A. Stotz Investment Research
- 11. 25 May 2016 11 Hypothesis Method Results Question Review Action Data We calculated the returns of the indices and the stocks Then calculated the pairwise correlations between asset classes, regions, sectors, and companies Looked at the average correlation within each grouping For asset classes, regions and sectors we looked at a 10 year rolling correlation For companies we calculated the 3 year rolling correlation as the data set spanned a shorter time period Comparing average rolling correlations Source: A. Stotz Investment Research
- 12. 25 May 2016 12 Hypothesis Method Results Question Review Action Data Correlation: Increased correlation after the global financial crisis Sources: A. Stotz Investment Research, Thomson Reuters 3 asset class average pairwise correlation increased to 45% from 14% Diversifying across asset classes still offers benefits, but less so Correlation was low during stock market rise (peaked in Oct 2007) Switched to very high (and rising) by Feb 09 equity trough 14 45 - 10 20 30 40 50 60 70 80 90 100 Mar-05 Sep-07 Mar-10 Sep-12 Mar-15 Bond, Equity, Commod.10yr-rolling correlation (%)
- 13. 25 May 2016 13 Hypothesis Method Results Question Review Action Data Equity and bonds are no longer uncorrelated Sources: A. Stotz Investment Research, Thomson Reuters Correlation between global equity and global gov’t bonds has increased to 35% from 3% Equity and bonds correlation shifted up Correlation between EM equity and commodities has doubled to 62% from 34% No decoupling 3 35 34 62 (10) - 10 20 30 40 50 60 70 80 90 100 Dec-05 May-08 Oct-10 Mar-13 Aug-15 Global equity vs Global gov't bonds EM equity vs Commodities 10yr-rolling correlation (%)
- 14. 25 May 2016 14 Hypothesis Method Results Question Review Action Data Correlation: Increased correlation after the global financial crisis Sources: A. Stotz Investment Research, Thomson Reuters Correlation between EM and DM equity markets has increased slightly to 87% from 77% Implies increased financial integration and fewer benefits from diversification between EM and DM But they were always highly correlated at this aggregate level No decoupling 77 87 14 45 - 10 20 30 40 50 60 70 80 90 100 Mar-05 Sep-07 Mar-10 Sep-12 Mar-15 Developed vs emerging Bond, Equity, Commod. 10yr-rolling correlation (%)
- 15. 25 May 2016 15 Hypothesis Method Results Question Review Action Data Correlation across sectors has increased massively Sources: A. Stotz Investment Research, Thomson Reuters The average correlation among the 10 major global sectors has risen to 79% from 53% Diversification benefits across sectors have been drastically reduced since the crisis 77 87 14 45 53 79 - 10 20 30 40 50 60 70 80 90 100 Mar-05 Sep-07 Mar-10 Sep-12 Mar-15 Developed vs emerging Bond, Equity, Commod. Sectors 10yr-rolling correlation (%)
- 16. 25 May 2016 16 Hypothesis Method Results Question Review Action Data Correlation: Country-level diversification still offers benefits Sources: A. Stotz Investment Research, Thomson Reuters The average correlation across countries has increased from 36% to 55% At the country level, it is easier to benefit from diversification, even though correlation has increased No decoupling 77 87 36 55 14 45 53 79 - 10 20 30 40 50 60 70 80 90 100 Mar-05 Sep-07 Mar-10 Sep-12 Mar-15 Developed vs emerging Countries Bond, Equity, Commod. Sectors 10yr-rolling correlation (%)
- 17. 25 May 2016 17 Hypothesis Method Results Question Review Action Data Good news: Stock pickers can still diversify Sources: A. Stotz Investment Research, Thomson Reuters The least amount of correlation exists between stocks Though correlation among stocks tripled during the financial crisis, it has been decoupling since 2012 This means that a stock picker can still construct a highly diversified portfolio Stocks have decoupled 77 87 36 55 14 45 8 4 53 79 - 10 20 30 40 50 60 70 80 90 100 Mar-05 Sep-07 Mar-10 Sep-12 Mar-15 Developed vs emerging Countries Bond, Equity, Commod. Stocks (3yr-rolling) Sectors 10yr-rolling correlation (%)
- 18. 25 May 2016 18 Hypothesis Method Results Question Review Action Data Grouping Begin Recent Level Change over period Asset classes 14 45 Moderate Very large increase DM vs EM 77 87 High Small increase Sectors 53 70 High Large increase Countries 35 55 Moderate Very large increase Stocks 9 7 Low Small decrease Equity market price correlation summary Source: A. Stotz Investment Research
- 19. 25 May 2016 19 Hypothesis Method Results Question Review Action Data Correlation has increased over time, except for stocks Expect portfolios of indices’ to be highly correlated Construct portfolios of stocks to benefit from decoupling Stock pickers can still diversify Source: A. Stotz Investment Research
- 20. 25 May 2016 20 Know it first! Get access to additional research & articles Sign up for free membership and newsletter Go to: becomeabetterinvestor.net/join
- 21. 25 May 2016 21 References Frank, N. and Hesse, H., 2009. Financial spillovers to emerging markets during the global financial crisis (No. 9-104). International Monetary Fund. Kose, M.A., Otrok, C. and Prasad, E.S., 2008. Global business cycles: convergence or decoupling? (No. w14292). National Bureau of Economic Research. Wyrobek, J. and Stańczyk, Z., 2013. Decoupling hypothesis and the financial crisis. European Financial Systems 2013, p.362. Wälti, S., 2012. The myth of decoupling. Applied Economics, 44(26), pp.3407- 3419. Yeyati, E.L. and Williams, T., 2012. Emerging economies in the 2000s: Real decoupling and financial recoupling. Journal of International Money and Finance, 31(8), pp.2102-2126.
Editor's Notes
- #13: E:\Dropbox (LifeFORCE)\Astotz\Reports\company level correl (2)
- #14: E:\Dropbox (LifeFORCE)\Astotz\World correlation2_AW sheet “pair correl”
- #15: E:\Dropbox (LifeFORCE)\Astotz\Reports\company level correl (2)
- #16: E:\Dropbox (LifeFORCE)\Astotz\Reports\company level correl (2)
- #17: E:\Dropbox (LifeFORCE)\Astotz\Reports\company level correl (2)
- #18: E:\Dropbox (LifeFORCE)\Astotz\Reports\company level correl (2)