Capital market assumptions: Asian equities and fixed income methodology

Capital market assumptions: Asian equities and fixed income methodology
Invesco Investment Solutions explains how we formulate capital market assumptions for various Asian equity and fixed income markets.
Key takeaways

We believe different Asian equity and fixed income markets need to be considered independently due to structural distinctions and differences in economic growth stages.


Analytical framework on how we view equity and fixed income across various Asian markets is key to developing outcome-orientated allocations with regards to the asset class.


Based on our research and understanding, we explain how we formulate our capital market assumptions for various Asian markets. This should be useful for investors residing in Asia, in making asset allocation decision between their home market and international markets.

Asia is the world’s fastest-growing economic region, as well as the largest continental economy by both nominal and PPP-based GDP today1. As Asian economies continue to open up and develop their financial markets, more and more investors are keen to explore opportunities in this region.

However, complex macroeconomic-level dynamics can drive stark differences in the pace of development among Asian markets, resulting in markets at varying stages of economic and structural development. As such, in-depth research focusing on individual markets in the region plays a fundamental role in asset allocation. These findings, in addition to our existing global capital market assumptions (CMA) framework, will provide local insights to international investors seeking to diversify their portfolios in emerging and developed Asian markets. In particular, Asia-based investors should find the framework helpful in their portfolio-construction process, in determining the appropriate levels of home-country bias and also in deciding their exposure to other regional and global markets.

In this piece, we discuss how we formulate our long-term CMAs for various equity and fixed-income asset classes in the region including China, India, South Korea, and ASEAN markets. Methodologies for these CMAs aim to reasonably account for the large cross-market differences in economic drivers as follows that make the region so unique among its global peers.

Growth: Economic growth has varied significantly across different Asian economies in the past 10 years (figure 1). China was experiencing close to double-digit growth for the past few decades until 2010, but since 2015 this has begun to slow to around 6% in terms of real GDP per capita growth. While we expect this deceleration to continue, we do view China’s transition to a domestic consumption- driven economy with higher-quality growth to drive opportunities going forward.

In Singapore, growth has been in steady decline since the mid-2000s, falling to 2.7% in 2018. Similarly, South Korea’s economic growth has been moderate, with a growth rate of 2.3% in 2018. Thailand has also experienced a slowdown in growth in the recent years, stemming from political unrests and natural disasters as well as from rising labour costs, resulting in a growth rate of 3.8% in 2018. During the same time, other emerging economies within the ASEAN region (such as Indonesia and Malaysia) have maintained stable growth rates.


Figure 1: Economic growth in Asia varied greatly from 2008 - 2018
Figure 1
Foreign exchange: For export-oriented economies like South Korea and Taiwan, currency exchange rates play a major role in determining overall economic performance. Exports-to-GDP ratios for South Korea and Taiwan have consistently been greater than other domestic driven economies such as China, India and the United States (figure 2). As such, we take this into consideration when building our valuation-change components for our South Korea and Taiwan equities in our CMA research.
Figure 2

Foreign reserves: Foreign-exchange reserves are essential for import-reliant economies to purchase key industrial inputs and other goods, as well as jurisdictions running account deficits required to repay debt. For example, in India, as steady economic growth continues to fuel demand for crude oil and natural gas, and as domestic production continues to fall due to ageing oil fields and a lack of major discoveries, the country is becoming more reliant on imports of these kinds of fuel – they accounted for 30% of India’s imports in 2017. As such, foreign reserves play an important role in the valuation-change component of India equity CMAs.

Methodology overview

We now turn to the framework that we adopt for the long-term capital market assumptions for Asian markets. This framework is consistent with the approach we apply for our research globally, that is – we use a “building block” based methodology to estimate returns for equities and fixed income. Our building blocks act as a “bottom-up” approach in which the underlying drivers of asset-class returns are used to form estimates, with estimated returns being divided into income and capital appreciation components.

Methodology for Asian equities

Our comprehensive research on equities for individual Asian markets has prompted us to:

  • Estimate the income component with the level of yield;
  • Establish market-specific approaches in formulating earnings growth estimates, and;
  • Incorporate macroeconomic data from multiple inputs while factoring in market- specific characteristics to calculate valuation change. 

Unlike the US, share buybacks are uncommon in Asian markets. Regulations surrounding share repurchases and the prominence of family and state ownership structures have resulted in Asian companies’ preference for dividend payments. As such, yields in these equity markets are mainly driven by dividend yields. We measure divided yield – our income component for equity returns – using the trailing 12 months’ dividend.

Earnings growth

Earnings growth is one of the two drivers of capital appreciation. Earnings growth is estimated using real GDP per capita growth adjusted for future inflation expectations. We use real GDP per capita growth – instead of historical data of earnings per share growth – to predict real earnings growth because it is a more stable time series while being highly correlated with real earnings growth in the long run.

The next question would be how we estimate real GDP per capita growth. For developed economies, we use long-term average real GDP per capita growth figures, as expected growth levels of developed economies are unlikely to deviate significantly from their long-term average growth levels.

However, the same methodology cannot be applied to emerging economies and certain Asian economies as we need to account for differences in phases of economic growth and adjust for idiosyncratic shocks to real GDP per capita growth. Asian economies can be grouped into three different groups based on the stages of their economic development:


  • Developed economies: for developed economies such as South Korea and Singapore, we expect real GDP per capita to maintain a stable growth rate. As such, earnings growth in South Korea and Singapore are estimated using historical real GDP per capita growth rate. This is the same approach we use for other developed economies such as the US, Japan or Europe.
  • Transitioning from developing to developed economies: a good example is China. To account for expected normalization of China’s economic growth rate, we look at other Asian economies such as Japan and South Korea as precedents to estimate China’s growth over the next 10 years.
  • Developing economies: countries like India, Indonesia, Malaysia and the Philippines all fall under this category. They generally are experiencing favourable demographic trends and better infrastructure. We expect these markets to maintain their fast pace of growth in the long term. We use long-term historical GDP (PPP) per capita growth as an estimate for earnings growth to reflect this continuing economic growth. 
Valuation change

Valuation change is our other driver of capital appreciation, reflected through changes in the price-to-earnings (P/E) or price-to-book (P/B) ratios over time.

Our research, in line with academic literature (Campbell and Shiller, 1998)2, suggests that valuation has mean-reverting characteristics in the long-term, adjusted for prevailing macroeconomic conditions. If the current valuation is deemed more expensive than the adjusted long-term mean, then the valuation change would be negative, annualized over the 10-year time horizon, in order to reflect its mean- reverting tendencies.
A vital consideration for valuation change is determining which are the macroeconomic factors that have a significant impact on valuations within each equity market. Our research shows that inflation and interest rates (risk-free rate) are decisive in determining valuations in developed economies such as the US, whereas for export-oriented economies like South Korea and Taiwan, foreign exchange plays a greater role in valuations (figures 3 and 4).

Figure 3
Figure 4

We now look at an import-reliant economy – India. As India’s economy is highly reliant on crude oil imports, we find that the country’s foreign-exchange reserve levels and inflation rates have a huge impact on valuation change in the NIFTY 50 (the flagship index on the National Stock Exchange of India) (figure 5). Similarly, India’s reliance on commodity imports means that we find a significant correlation between the strength of Indian rupee and valuations, as a stronger local currency means cheaper imports. However, to avoid the issue of multicollinearity, we exclude currency fluctuations from our model since the metric correlates highly with foreign-exchange reserves.

Figure 5

For China A-shares, we find that valuation change in the mid-small cap CSI 500 index (in terms of P/B ratio) is well explained by two macro factors: inflation and total social financing (TSF), which refers to the aggregate volume of funds provided by China’s domestic financial system to the private sector of the real economy within a given time frame. We also find that while money supply (M2) growth has significant explanatory power in explaining valuation change for CSI 500 index, it is highly correlated with TSF growth (figure 6). As such, we exclude M2 growth from our model to avoid multicollinearity issues. These findings are consistent with our findings for the large-cap CSI 300 index.

Figure 6

For markets like Hong Kong, Singapore and emerging ones such as Indonesia, Malaysia, and the Philippines, our research shows that valuation change is more correlated with inflation and interest rates. For the three emerging markets, even though foreign exchange correlates with valuation change in export-driven markets such as South Korea and Taiwan, the same phenomenon cannot be observed in Indonesia, Malaysia, and the Philippines. This is because a significant and increasing proportion of their exports are intra-ASEAN trades. In addition, in accounting for further integration of ASEAN markets in the future, our research suggests that valuation change is more correlated with inflation and interest rates than foreign exchange.
The following table summarizes the macro variables that we look at when analyzing valuation levels for various Asian equity markets:

Table 1

Methodology for Asian fixed income

Our drivers of returns for fixed income are also divided into two components: income (which is measured by yield) and capital appreciation (which is measured by roll return, valuation change, and credit loss). Therefore, for Asian-fixed income CMAs we:

  • Establish income component with the level of yield;
  • Estimate capital appreciation via the effect of roll return from a bond moving closer to maturity as time passes;
  • Form valuation-change estimates based on how expensive or cheap the fixed- income index currently is, and;
  • Account for the impact of potential bond migration or default loss.

Yield reflects the average income expected to be received when holding a fixed- income security until maturity. For the purpose of our CMAs, yield is calculated using the average of starting (current) and ending (estimated) yield levels.

To get ending (estimated) yield levels, we study how current (starting) yield curves might change over time as a result of two factors:


  • Changes to Treasury interest rates in each Asian market – as suggested in academic literature (Litterman and Scheinkman, 1991)3 – and;
  • Credit spreads over Treasury interest rates. 

In terms of credit spread changes, if the asset class is treasury or quasi-treasury local currency bond, we assume credit spread will not change and stay at zero. For example, for Singapore Treasury and Taiwan Treasury, credit spreads over Treasury interest rates are zero. Credit spreads of South Korea and Thailand aggregate are minimal since their proxy index constituents consist almost entirely of government and state-owned enterprise (proxy index for South Korea holds 0.60% of corporate debt). As such, yield is determined solely by changes to Treasury interest rates. If credit spread is substantial such as for the Markit ALBI HK$ Bond index, the change in credit spread is measured by the difference between the current spread and the 10-year average spread level.

A change in Treasury interest rates has a potential effect on the level and slope of the current (starting) yield curve. To estimate this effect, we first construct the ending (estimated) yield curve. Using the yield for three-month US Treasury Bills and the yield for 10-year US Treasury notes as two specific data points, polynomial interpolation – sourced from the consensus forecast of the Federal Reserve Bank of Philadelphia – is then applied to generate the ending (estimated) yield curve, after accounting for interest rate differential between the US and local markets. The impact of changes in Treasury interest rates is then calculated by taking the difference in yields between current and estimated yield curves, at a specified duration. A sample of yield levels in different Asian markets can be seen in figure 7.

Figure 7
Roll return

As one component in capital appreciation for fixed income, roll return reflects the impact of bond maturation on bond price, with all else being equal. If the yield curve slopes upward, movement along the curve (towards maturity) will have a positive impact on returns. For many Asian markets, flattening yield curves in recent years have resulted in a general decrease in roll returns (figure 8).

Figure 8
Valuation change

Valuation change is our second component of capital appreciation which captures the impact of yield curve movements on bond price. Valuation change is impacted by changes to Treasury interest rates, and changes to credit spreads over Treasury interest rates. However, since changes to credit spreads are often negligible in treasury or quasi-treasury assets, valuation change in these assets are mainly driven by changes in Treasury interest rates.

To estimate valuation change, we take a scaled figure of the difference between current (starting) yield curve and estimated (ending) yield curve. Note that for Markit ALBI HK$ Bond index, since the spread levels are not zero, we assume current spread levels will revert to the long term mean.

Credit Loss

Credit loss captures the potential impact on returns from a downgrade in credit rating or debt default. For treasury and quasi-treasury indices, since spread levels are at or close to zero, we assume no downward credit migration or credit default. For Markit ALBI HK$ Bond index, we use fixed percentage of the option adjusted spread (OAS) as the potential credit loss from rating migration.


Having a sound and robust analytical framework would help in developing outcome- orientated allocations with regards to different asset classes. However, the differences between Asian markets and in economic growth stages mean that we should look at each equity and fixed income market independently when estimating their long-term performance. Our research and understanding of Asian markets have guided us in coming up with capital market assumptions for these different markets, which are aimed at assisting in making asset allocation decisions.


Proxy Indices

Tables 2 and 3 list the proxy indices that we use in estimating our CMAs.

Table 2 & 3
Our current estimations

Here we share our latest long-term CMA (LTCMA) as of Sept. 30, 2019.

These estimates are forward looking, are not guarantees, and they involve risks, uncertainties and assumptions. These estimates reflect the views of the Invesco Investment Solutions, the view of other investment teams at Invesco may differ from those presented here.
These estimates are forward looking, are not guarantees, and they involve risks, uncertainties and assumptions. These estimates reflect the views of the Invesco Investment Solutions, the view of other investment teams at Invesco may differ from those presented here.