/203230179344.jpg?imageView2/2/w/1000/format/jpg/interlace/1/q/85″ data-w=”1080″ data-h=”627″>
GDP growth of different types of countries (left) | Index returns of different types of countries (right)
Figure 1: GDP growth and stock index returns
The GDP data comes from (IMF, 2013), and the forecast time of 2013 GDP is October 2013.
Emerging markets and frontier market index returns: Based on the market capitalization weighted combination of MSCI Emerging Market Index (MXEF) and MSCI Frontier Market Index (MXFM).
However, as many market participants can confirm, correlation does not prove cause and effect. Instead, a careful breakdown of the drivers of historical returns and expected returns can provide more valuable insights. This applies to individual companies like Pear, as well as national or regional market indexes. In fact, when stockbrokers sell Pear’s investment advice to investors, the above five questions that investors may ask are also applicable to investing in emerging or frontier markets.
This article attempts to solve these five problems in the context of stock investment in emerging and frontier markets. The first section discusses confidence in economic growth forecasts. During the two-year forecast period, the forecasts of private sector economists appear to be statistically significant predictors of future GDP growth. Long-term economic forecasts, such as the 5-10 year period considered by many asset allocators, are relatively more difficult to find. One of the data sources is the International Monetary Fund (IMF), which has published long-term GDP forecasts for various countries since April 2008. The evidence from these data shows that long-term economic forecasts are full of more noise, but they are not without some predictive value.
Accurately forecasting GDP growth is only valuable to investors when the growth is transformed into a high-quality risk-adjusted portfolio return. The rest of this article will focus on how GDP growthContributing to the decision-making, and why investing in a market similar to the company’s Pear (high-growth economy) may not be an investment How much value the combination adds. The second section attempts to dispel a common myth that economies with faster GDP growth tend to produce higher returns on equity. The third section describes the source of emerging market stock returns, and shows that most of the excess returns have already taken the growth impact on investors (ie, unexpectedly high growth)< /span> and currency risk are compensated. Section 4 shows that the correlation between the returns of large listed companies in developed and developing markets has increased over the past 20 years. This may be due to the increase in their cross-regional input and output markets (ie international trade). The last section summarizes the enlightenment for investors, and believes that emerging market stocks still belong to a diversified investment portfolio, provided that the expected returns allocated to emerging and frontier market stocks remain at a moderate level.
1. GDP(to some extent) at least in the short term Within is predictable
Accurately forecasting GDP growth rates often proves to be very challenging. An old joke compares an economic forecaster to a bad clock that accurately displays the time twice a day. Perhaps a more appropriate modern joke would say that economic forecasters will only approach twice a day, depending on the standard error of the forecast and possible backtesting in the next few days.
The bad reputation of this dull science in prediction may be unfair. Many people still remember some major mistakes, such as the famous 20th century economist Irving Fischer (Irving Fischer) in 1929, That is, the stock market crash (marking the beginning of the Great Depression) optimistically predicted in the first three days that stock prices and the overall economy will have “permanent highs” “. However, ordinary predictions that reflect the value of consensus tend to receive less attention, but it turns out to be more accurate.(Laster, Bennett and Geoum, 1999) . In fact, the forecasts of integrated economists seem to create a perspective of group wisdom(Denrell and Fang, 2010) .
1. More recent economic forecasts are relatively more accurate
A large number of tests of varying complexity and robustness are used to evaluate the accuracy of macroeconomic forecasts, especially in the more recent forecast range. For the sake of simplicity, this article will focus on the panel regression model with the actual annual GDP in the economic forecast of country i in year t as the dependent variable (real GDP growth). Equation 1 summarizes this simple model.
Formula 1
Real GDP growth it = β0 constant + β1 real GDP growth i,t−1 +β2 real GDP growth i,t−2 +β3 real GDP growth forecast t−2 it + εit
The independent variables include one-year (real GDP growth i,t−1) and two-year (real GDP growth i,t−2) the actual GDP growth rate before and the growth forecast for this year two years ago( Real GDP growth forecast t−2 it). The time series of most markets covers 1991-2013, although the data is not a balanced data set due to the different years when data is initially available for each country. The forecasted data comes from consensus economics, which surveys economists in the private sector and aggregates their forecasts of real GDP in different countries. By aggregating data from a large number of organizations, this approach seems to be more accurate than a single source (such as IMF or OECS) prediction accuracy (Batcheloe,2007).
Table 1 shows the results and made three main findings. First, the actual GDP growth of the previous year is a statistically and economically significant forecast indicator of the actual GDP growth of this year. The 1% increase in GDP in the previous year may herald 0.29%(first column) to 0.44%(third column) real GDP growth. The real GDP growth two years ago may be more statistically significant than economically significant.
Table 1 Actual annual GDP growth vs. expected growth
Second, the consensus forecast of economic growth two years ago also seems to provide a reasonable forecast for future growth rates. The second table shows that the expected GDP growth of 1% indicates a real GDP growth of 1.01%. Even if the previous value of the actual GDP(column 3) is included, the coefficient of the predicted growth rate is still 0.49 which is statistically significant and economically significant percentage point.
Third, even if GDP forecasts have a certain predictive value, they can only explain about one-third of the change in GDP growth rate (column 2 R2 is 0.28, and the third column is 0.39). Part of the reason for the lower R2 is that forecasters tend to better predict trends rather than shocks. For example, the consensus reached in January 2007 was that the US GDP growth rate in 2008 was 3%, which is very close to the long-term average. In fact, the US GDP growth in 2008 was -0.3%. At the beginning of 2008, the forecast of US GDP growth in 2009 was even more inaccurate.(predicted 2.5% vs. actual -2.8%).
As mentioned earlier, there are many ways to test the accuracy of macroeconomic forecasts. If you think that Table 1 proves that the macroeconomic forecasts of private sector economists have effective statistical forecasting capabilities, then this situation is a bit exaggerated . Other studies, such as Laster et al. (1999), solved this problem in more detail. The basic research result is that the overall forecast of GDP growth tends to be accurate, but individual forecasters have shown behavioral deviations. The results in Table 1 are consistent with these overall results. At the very least, Table 1 shows that the GDP predicted by economists is more valuable than a broken clock.
2. Long-term economic forecasts are more challenging
(5-10 years) forecasting economic growth rates in a longer time dimension is more challenging than short-term forecasting. One source of long-term growth forecasts is the “World Economic Outlook” (World Economic Outlook) published every six months by the IMF. Since 2008, the World Economic Outlook includes long-term forecasts of GDP growth rates for various countries.
Figure 2 depicts the 2008 real GDP growth forecast for 2013 (i.e. five-year forecast) and October 2013 for 2013 The relationship between real GDP growth forecasts (that is, three-month forecasts). Assuming that the recent forecast is more accurate, one way to assess the accuracy of long-term forecasts is to draw a 45-degree line through the origin of the chart. The value on this line indicates that the 2008 forecast is the same as the more recent forecast.
Figure 2: Realized Real GDP vs. Projected Real GDP Growth Rate
An obvious revelation of Figure 2 is that forecasters in 2008 are too optimistic. The 2008 forecast for every country is greater than zero, while the 2013 forecast has some negative expected growth rates. Except for three countries, the forecasts of other countries are below the 45-degree line, which is consistent with the global economic recession and weak recovery. The second meaning is that in 2008 and 2013, the growth forecast of (blue triangle) in the developing market surpassed that of the developed market(orange circle) growth forecast.
For many investors, actual growth rates are not as important as relative growth rates. In other words, without knowing the global average growth rate, it may be sufficient to know that developing markets will rank higher on the global growth rate chart than developed markets. The right panel of Figure 2 adjusts the 2008 and 2013 values according to the global average predicted growth rate in 2008 and 2013, respectively. Based on this indicator, the IMF’s forecast seems more accurate. The relative ranking of the growth rate by country is more stable, with values above and below the 45-degree line almost evenly distributed. The data again shows that the prediction that developing markets will grow faster than developed markets is correct.
Although the IMF’s long-term forecast may prove to be accurate in direction, it is not accurate. The root mean square error of the prediction is 2.5. Ordinary least squares regression (not reported) shows that the predicted value in 2008 is a statistically significant predictor of the actual value in 2013. But R2 is only 0.43, indicating that the predicted value is noisy. Based on the results of this data and the accuracy of short-term economic forecasts, economists’ forecasts seem to be more accurate than random guesses, but the uncertainty about the future makes longer-term forecasts look more like a useful guide than precise charts.
2. GDP growth cannot be converted into return on equity
Even if GDP growth is predictable within a reasonable margin of error, knowing the future growth rate is helpful only when the growth is translated into return on equity. There is some evidence to support this hypothesis. Fama and French(1998) studied the global stock market and found that the return on “value” stocks was 5% higher than that of “growth” stocks~ 8%. At the individual stock level, growth may erode profits or make capital use inefficient, thereby destroying value (Koller, Goedhart, and Wessels, 2010). There are similar concepts at the national level. That is, the economic growth of a country may not produce a positive return on equity. For example, some high-growth countries overinvest in infrastructure, resulting in inefficient use of resources. Some countries also allocate a larger proportion of their economic surplus to non-equity investors (such as managers, labor, consumers, or the government).
Figure 3 depicts the country’s gains in 25 developed markets and 24 developing markets since the beginning of this century. (in U.S. dollars) and The ratio of national growth rates. Table 2 reports the regression results represented by the solid line in the graph. A simple linear regression shows a positive but statistically meaningless result (the coefficient value reported in column 1 is 0.65, but the standard error is 0.52) , this means that the relationship between growth and stock returns is unclear. The correlation between real GDP growth in developed markets and stock market returns is greater (coefficient of 1.26), but statistically compared to emerging markets no difference. The average annual growth rate of emerging markets is 6.45 percentage points higher than the average growth rate of developed markets over the same period (standard deviation is 2.06 percentage points).
Table 2: Revenue vs. Real GDP Growth (2000-2013)
Figure 3: Revenue vs. Real Revenue vs. Real GDP Growth (2000-2013)
Choosing a longer time frame or a different data set illustrates a similar story. Statistical tests have not found a robust and positive relationship between real GDP growth and stock returns. Henry and Kannan(2008) studied 19 emerging markets in 30 years( 1976-2005) stock market returns. They not only found that there is a statistically insignificant relationship between economic growth and stock returns, but the sign of this relationship is still negative (that is, actual GDP growth is lower than High countries have lower return on equity). Dimson, Marsch, and Staunton(2002, 2011) and Ritter(2005, 2012 ) ResearchA longer time series (1900-2011), although these studies mainly focus on stock returns in developed markets in the early 20th century. Dimson et al. (2011) and Ritter(2012) report Show that in the entire time period (1900-2011), the post-World War II period (1950-2011) and the “modern” economic era (1970-2011), the relationship between stock return and GDP growth There is a negative correlation. Ritter(2012) also considered the results of 15 emerging markets from 1988 to 2011, and still found a negative correlation.
3. The source of stock returns in emerging markets: growth shocks and currency risks
If it were not for faster real GDP growth, then the excess returns generated by emerging and frontier markets relative to developed market stocks (as shown in Table 2 And as described in the academic literature) there must be another explanation. There are many potential explanations, including liquidity differences and market irrationality. If you do not consider the validity of these views in the short term, then for long-term investors, these two views are not convincing. The increase in wealth in developing markets and the reduction in cross-border transaction costs will logically narrow the gap in liquidity over time (unless there are any regulatory obstacles). According to the behavioral finance and behavioral economics literature, predicting that the market remains predictable and irrational in the long term may not be completely unreliable, but it is at least as difficult as predicting the long-term GDP growth rate. Instead, this article focuses on two interpretations(hope) can make a more intuitive and reasonable explanation of the excess returns-that is, the benefits brought by unexpectedly high economic growth and the gains from taking currency risks Compensation.
1. Unexpected economic growth pushes up stock prices
Although as shown in Figure 3, the relationship between real GDP growth and stock market returns is weak, and may even be negatively correlated, but growth is beneficial to the stock index of each country, just as growth boosts individual companies. (such as the fictitious company Pear). But just like Pear, the effect of growth on stock prices depends on which expectations of future growth have been factored into stock prices.
Even if it is not completely impossible, it is difficult to measure the market’s expectations of future GDP growth rates that are included in current stock prices. This article uses the average of the consensus economics two-year future growth forecast as a representative. Then, the value of the unexpected growth “shock” is calculated as the difference between the actual growth rate and the predicted growth rate.
Table 3 estimates a regression model (Formula 2), which tested the economic growth of 77 markets from 2000 to 2013 The relationship between growth shocks and stock index returns. As a control variable, the model reported in Table 3 also includes lagging index returns. Unlike the random effects of panel regression reported in Table 1, the Brausch-Pagan test of serial correlation rejects the null hypothesis that serial correlation does not bias the results (That is, including lagged GDP and lagged returns will lead to serial correlation in the error term). Therefore, the regression model in Table 3 applies the generalized moment estimation of Arellano and Bond(1991) to the panel data.
Formula 2
Annualized index return it=β0 constant + β1 annualized index return i,t−1+β2 actual GDP growth it +β3 unexpected growth “shock” t−2 it +εit
Table 3: Generalized Moment Estimation (GMM) Annual Return vs. Real GDP Growth (2000-2013)
The results in Table 3 are consistent with the assumption that the expected growth rate has no statistically significant impact on the return on equity, but is unexpectedly high(low)< /span>The growth rate will produce a positive (negative) return on equity. Similar to Table 2, the first column of Table 3 shows a positive but not statistically significant relationship between real GDP growth and index returns. However, the second and third columns show that when the actual GDP growth rate exceeds the expected growth rate by 1% per year (ie, when there is a positive economic shock Hour), the stock yield increases annually from 1.97% to 7.95%. These results are statistically significant.
The main meaning of Table 3 is that the stock price already contains information about the expected growth rate. However, more than expected growth has a positive impact on prices. These results should not surprise supporters of efficient markets in (broadly). If the International Monetary Fund (IMF)’s five-year forecast of real GDP growth (IMF, 2013) reflects the consensus view of the market, then this means that only the growth rate in emerging and frontier markets (in general)The growth of these markets will make a positive contribution to the return on equity when it exceeds 5% per year. Unfortunately, correctly predicting the size and timing of the “shock” is a challenge for many forecasters.
2. Investors are compensated for taking currency risks
In addition to unexpectedly high growth rates, the second major source of returns in developing countries’ stock markets is exchange rate changes. Figure 4 depicts the standard emerging market index return rate (blue) and the equivalent but currency hedged (grey) index. MSCI emerging market foreign exchange hedging uses a one-month forward to eliminate currency risk in the MSCI emerging market index. Consistent with the financial model in the textbook, that is, investors receive a premium for taking certain risks. Since the second quarter of 2004, the unhedged rate of return has received a positive premium over the hedged rate of return. From the second quarter of 2004 to December 2013, the unhedged emerging market index produced a return of 208%, while the hedged emerging market index returned only 173%. During the same period, the developed market index returned 57%.
Perhaps more surprising is the proportion of the overall return of the MSCI Emerging Market Index that can be attributed to exchange rate risk factors. Table 4 reports the results of ordinary least squares regression (Formula 3), which attempts to clarify the role of currency risk in emerging market stock returns . The dependent variable is the daily return of the MSCI Emerging Market Index, and (similar to Figure 4) the independent variable includes the purchase of a one-month forward(Investment in MSCI foreign exchange hedge index) to hedge currency risk (currency hedge) Earnings and returns of the MSCI World Index (i.e. developed market returns).
Formula 3
Emerging market index return t = β0 constant + β1 emerging market currency hedge t + β2 developed market index return + εt
In the complete time series from March 2004 to December 2013, the R2 value in the regression (including currency hedging variables only) Explains the 64% change in the entire market index (column 1). Currency hedging is not only a powerful predictor of index returns, data shows that investment hedging itself will be more capital efficient. The estimated value of the coefficient is 3, which means that an investment of US$0.33 in currency futures can produce the same expected return as an investment of US$1 in the index. Adding the stock market returns of developed market indexes, R2 will increase to 73%(column 2).
These results are consistent across different time periods. Columns 3 and 4 report the results of the same regression before the economic crisis (before 2007), while columns 5 and 6 use crisis Use the data to estimate the regression model. R2 is higher in the later stages, indicating that equity returns in emerging markets are increasingly driven by currency exposure rather than equity exposure. Was this result driven by fundamental changes in the market, or was it a short-term (but not fleeting) caused by the financial crisis and central bank policies. rise, it remains to be seen.
This means to investors that excess returns from emerging market equity investments are largely and may be increasingly driven by currency risks. Since currency risk is a nominal factor, long-term investment in emerging markets is actually equivalent to a long-term bet on emerging market monetary policy. For many people, there is a more capital-efficient way to bet on monetary policy than simple stock investment (such as sovereign debt and currency) .
< /p>
Table 4: Emerging Market Returns vs. Emerging Market Currency Risk and Developed Market Returns
Figure 4: Emerging market returns vs. emerging market returns after currency hedging
4. Enhanced correlation between regions
Some investors see emerging market stocks as a way to diversify risk. Mathematically speaking, adding additional assets to the portfolio, according to Markowitz’s modern portfolio theory, will improve the effective frontier of the optimized portfolio (if additional If your assets are not completely related to existing assets). However, the possible diversification benefits of investing in emerging markets seem to be gradually diminishing.
1. Since 1992, the relevance of developed and developing markets has increased
Over time, the correlation between developing countries, emerging markets, and frontier markets is not stable, but the correlation seems to be on the rise. Figure 5 depicts the correlation between the returns of the MSCI Developed Markets Index and the MSCI Emerging Markets Index for five consecutive years (blue line). The return is calculated on a monthly basis. Not surprisingly, since 1997