We measure sector and country contributions to the returns of the constituents of the MSCI World Index. We find that over the past 25 years, companies in developed markets have actually had a stronger relationship with their home country than sector. This domestic dominance does not hold for all countries and sectors, though, with energy and information technology stocks representing notable exceptions.
Reduced trade barriers and the rise of multinational conglomerates have accompanied an increase in the integration of global financial markets. But if companies are becoming more globalised, do their share prices still follow region-specific trends? Or are returns instead more influenced by what is happening in a company’s sector? Understanding issues such as these are important when constructing equity portfolios.
To answer these questions, we attribute stock returns to sector and country “factors”. We use the constituents of the MSCI World Index from January 1990 to November 2015, with the underlying sector and country indices representing our factors. As an example, to measure country and sector contributions to the returns of Google, we use the MSCI USA Index and the MSCI Information Technology Index, respectively.
For results that are straightforward to interpret, we regress out the MSCI World’s total returns to calculate market-neutral returns, which we adjust to the same volatility. A linear, multivariate regression model is then fitted using the relevant country and sector factors for each stock. These regressions are performed at the end of each month, over the previous 36-month period, using daily returns .
If the factors are uncorrelated, then each regression coefficient is equal to the correlation between stock and factor. The square of a regression coefficient then gives the fraction of the stock’s volatility that can be explained by the volatility of the factor. It is worth noting that this type of analysis does not attempt to indicate causation or measure contributions to a stock’s long-term performance; instead, it describes how a stock’s return co-moves with the sector and country returns.
We split our analysis into four sections. First, we introduce stock-level results for two companies. Next, we average the results of companies within sectors and countries. And, finally, we measure sector and country contributions for all the constituents of the MSCI World Index.
We begin by looking at the contributions of country and sector factors to the returns of two companies, both within the energy sector. The results for Exxon Mobil and Showa Shell Sekiyu K.K. are shown in Figure 1, with sector coefficients in red and country coefficients in blue. The standard errors of the regression coefficients are indicated by the shaded regions .
Figure 1: Sector and country contributions towards the returns of two energy stocks
Exxon Mobil is one of the largest energy companies globally, using resources from all over the world to supply energy to many countries. As we would expect from a large multinational corporation, its location had less bearing on how the stock performed than whether the global energy sector has done well.
In contrast to Exxon Mobil, Showa Shell Sekiyu K.K. is more domestically focused. It sells products internationally through Royal Dutch Shell’s global network, but most of its revenues come from Japan. Figure 1 shows that the company has historically seen a stronger relationship with the Japanese market than the energy sector. The results of these two companies highlight the extent of differences which occur, even within the same sector.
We now turn our attention to average results across sectors. In Figure 2, we show the market-capitalisation-weighted average contributions of country and sector factors to the returns of stocks within each sector. The shaded error represents the standard error of these averages . The sectors are ordered by the relative strength of the sector contribution.
Figure 2: Contributions to the returns of stocks within all 10 GICS sectors.
Over the past 25 years, energy stocks have had a stronger relationship with their sector than their home country. This makes sense as companies within this sector have generally become more reliant on global resources such as oil, and have expanded to provide energy to many different countries. Similarly, information technology companies have had an increasingly global presence and have become less correlated to their domestic markets.
In contrast, financial stocks continue to co-move with their domestic economies. This is perhaps expected as banks, which are the largest constituent of the financials sector, are heavily influenced by government policy. Unlike some sectors, the financials sector encompasses a wide range of industries from real estate, through to insurance. As a result, the financials sector index may have a less explanatory power of individual company returns than, say, energy or information technology.
Similar to financial companies, industrials and consumer discretionary companies have had strong relationships with their home country. The remaining sectors, in general, have become less correlated to their domestic market since the early 1990s. Over recent years, these sectors have mostly received equal contributions from both the country and sector factors.
In Figure 3, we compare the country and sector regression coefficients for stocks within the US, Japan and Germany, which are three of the largest economies in the index. Our results show that US stocks have become less related to the domestic market over time and more related to their sector returns.
Figure 3: contributions of sector and country to the returns of stocks within three countries
Our results for the US contrast with our findings for Japan. In the late 1980s and early 1990s, Japan experienced a real estate-driven boom and bust. Figure 3 shows that the sector contribution was relatively large during this time, but gradually weakened over the next 10 years, leaving Japanese companies with a stronger relationship to their home country ever since. This relative dominance of the domestic marketplace can be seen in other Pacific countries as well.
The results for European stocks are mixed. In the region’s largest economies ‒ Germany, for example ‒ the contribution of the country factor has decreased over time and the sector contribution has risen, eventually resulting in similar levels between 2000 and 2010. However, more recently, a stronger relationship between stock and country has developed.
Finally, in Figure 4, we average the sector and country contributions by market-capitalisation weighting (solid lines) and equally weighting(dashed lines) each constituent within the MSCI World Index .
Figure 4: Average contributions of sector and country to the returns of stocks globally
When equally weighting, we find that the relationship between stock and country has weakened, but still remains stronger than with a company’s sector. This is also true when using market capitalisation weighting, but to a lesser extent.
During the recent crashes, we see that the sector contribution to stock returns rose. This is understandable given that the booms and busts were mainly driven by individual sectors – technology stocks in the early 2000s and financial stocks in the mid-2000s.
Using the methodology presented here, we find that companies within developed markets have generally had a stronger relationship with their country than their sector. This is not always the case, though. The level of contributions for different sectors and countries can vary to a large extent. In particular, we found that information technology and energy companies have been less correlated with their domestic markets and more with the performance of their sector globally.
 In order to diminish any effects from the different closing times of exchanges, we use rolling five-day returns for all response and explanatory variables. USD returns are used; local currency returns do not materially change the results.
 The standard error is determined by block bootstrapping the returns.
 These errors were calculated by bootstrapping the stock-level regression coefficients contributing to the average.
 By equally weighting the constituents, we reduce the influence of the US on our overall results. Currently, the US constitutes around 60% of the index by weight. When equal weighting, this reduces to close to 30%.
MSCI makes no express or implied warranties or representations and shall have no liability whatsoever with respect to any MSCI data contained herein. The MSCI data may not be further redistributed or used as a basis for other indexes or any securities or financial products. This report is not approved, endorsed, reviewed or produced by MSCI. None of the MSCI data is intended to constitute investment advice or a recommendation to make (or refrain from making) any kind of investment decision and may not be relied on as such.