Chapter 03

Long-term
Return
Expectations By Peter Eerdmans,
Antoon de Klerk
& Grant Webster

Introduction

The reversal of capital from emerging market equities and debt in early 2014 reminded the world of the nascent nature of capital markets in many emerging markets, despite large and growing savings. How do we think emerging market debt will grow in the next five years? In this piece, Grant Webster, Peter Eerdmans and Antoon de Klerk, who are part of Investec Asset Management’s Emerging Market Debt team, focus on return potentials, current account developments, and the growth of new asset categories within emerging market debt.

Local currency denominated emerging market debt (EMD) rallied strongly in the run-up to 2013, returning 12.2% per annum, on average, since the decade beginning December 2002. EMD was buoyed by strong fundamentals, a resilience of emerging markets to the developed world’s credit crisis and the quantitative easing-driven hunt for yield. Arguably this rally had pushed valuations into expensive territory especially as, on the margin, fundamentals had deteriorated.

Additionally, strong domestic demand and a reversal in commodity prices had eroded current account balances and, in some key economies, pushed current accounts into deficit.

During the summer of 2013 sentiment drifted and, after many years of inflows, emerging markets started experiencing significant outflows. Market participants feared that easy global liquidity had led to asset bubbles and diminished the momentum of structural reforms. EMD sold off as yields rose from historically low levels, while currencies also weakened.

This market reaction highlighted that a long-term sustainable approach to investing is crucial in the development of local EMD markets. The key questions we want to address are: where we are now and what we can expect from investing in local EMD on a five-year view? Have markets overshot towards cheap levels?

What might be deemed to be “fair value”? What returns can we expect as markets revert to these levels over the medium term? In particular, we will analyse the long-term growth potential in emerging markets and what this means for long-term real yields. We will also analyse if we can expect continued improvements in inflation and what the equilibrium levels for currencies might be based on our adjusted Balassa-Samuelson/PENN State valuation work.

“EMD was buoyed by strong fundamentals, a resilience of emerging markets to the developed world’s credit crisis and the quantitative easing-driven hunt for yield.”

EMD

Structural real GDP and what it means for real yields

Central to any discussion on the “attractiveness of a financial asset” is its ability to generate economic returns. The attractiveness of an investment in any company is, of course, greatly influenced by its potential to return cash to investors. Similarly, the attractiveness of an investment in the debt and/or currency of a country is greatly influenced by its potential to generate economic growth.

Furthermore, strong theoretic and empirical evidence exists on the positive relationship between real economic growth and real interest rates – clearly important for bond returns (see Figure 1). The first chart shows the average 5-year interest rates and GDP growth rates for a range of developed markets (DM) between 1975 and 2013. Emerging markets have demonstrated a similar relationship over the past 10-15 years.

One important difference, however, is that yields in emerging markets have generally been lower than their GDP growth rates may have implied. More recently, this has been a function of the economic environment, namely higher global liquidity driving yields lower. However, over the longer term the reason can be found in savings rates, which have been much higher in emerging markets than in their developed market counterparts. The specific reasons for this difference remain contested, but generally it is thought that this is due to an underdeveloped credit sector and a lack of social safety nets – specifically, no state pension, underdeveloped insurance markets and poor basic healthcare, which leads to the need for self-insurance through saving. In fact, it could be argued that developed markets were in a similar situation in the post-World War Two period, with the same phenomenon experienced in their bond markets, as highlighted in the third chart in Figure 1.

Social
Graph 1
Graph 2
Graph 3

Over the coming years, we expect that real yields in emerging markets will gradually trend closer to real GDP, as savings rates fall from their historically high levels. This makes economic growth an important metric in assessing future yields in emerging markets. Figure 2 looks at 5-year annualised growth averages by geographic regions (with developed markets grouped together).

Although global growth is expected to be lower over the coming five years, than it was during the early 2000s, emerging markets are

generally expected to grow at 3.8%, faster than developed markets which may grow at approximately 2% a year. Corresponding to this level of growth, we expect that emerging market real yields will gradually trend upwards from their current level of 3% to approximately 3.5 - 4%.1

To further assess the validity of these estimates of higher growth for emerging markets (EM) relative to developed markets (DM), we consider a number of supply-side and demand-side factors.

Growth Averages

Supply-side dynamics

From the supply side we investigate the production inputs of capital, labour and technology (total factor productivity). From the graphs the following observations are offered:

Although it has greatly increased, the stock of capital (e.g. machinery) available per worker in EM is still well below DM levels. In fact, even Asia, which has by far the most capital in the EM universe, only now has the same stock of capital that DM had in the late 1980s. (Figure 3). Therefore, despite the well-documented high investment levels and growth in EM supply (Figure 4), growth can, from a pure stock perspective, be further boosted by continuing this trend.

Similarly, investment in human capital (Figure 5) has also been a notable contributor to increased supply, but has not yet (excepting Central and Eastern Europe) reached developed market levels.

As the investment in human capital in emerging markets increases, this will no doubt have an influence on the quality of growth in emerging markets.

The combined increase in the stock of physical capital available per worker, as well as the heavy investment in EM, can clearly be seen in the growth of labour productivity (Figure 6).

At the same time, given the distance to developed market averages, the data strongly suggests that these avenues have not nearly been exhausted as potential sources of supply growth.

Finally from the supply side it is almost disappointing to see the slow pace of technological convergence of EM to DM (Figure 7). On the other hand it also suggests that EM has not yet reached the production efficiency frontier and could still benefit from adopting existing technologies.

Graph 1
Graph 2
Graph 3

Demand-side dynamics

On the demand side, growth is derived from four areas: investment (private or public), government consumption expenditure, household consumption expenditure and net exports. While the contributions from each of these drivers will be varied, we believe that they still remain supportive of growth in EM:

Exports, although marginally less important than during the first decade of this century, will remain an important source of demand growth. Developed markets have clearly struggled for growth during the past five years although recent data does seem to suggest that the economies of the US, Germany and Japan, among others, are looking relatively stronger.

Domestic consumption, both public and private, is the component which we believe will increase in relative importance. Government spending, as a share of total demand (Figure 8), in EM remains significantly lower than in DM. As these countries continue to develop, governments are increasingly able to provide social safety nets to

the population, thereby resulting in a more stable middle-class consumer. Historically, rising per capita GDP has been associated with increased levels of government current expenditure on social goods and services.

Private consumption is also expected to grow. This will be driven by the triple dynamic of access to insurance (both public and private), greater labour mobility and higher disposable income. These all increase the marginal propensity to consume, which will boost sustainable growth in EM. This dynamic comes on top of the well-documented positive population growth in the great majority of emerging markets.

Investment, as the final potential source of demand, has already played a key role in EM growth, especially in Asia. Even so, as seen from the supply-side analysis, stock of capital per worker still has a long way to go in order to catch up to DM levels and for this reason we see investment as continuing to play an important role.

Graph 1
Graph 2
Partly Funded

There are two important and related consequences to the expected model of growth which we have described here. First, we have said that we expect yields to rise gradually towards growth levels due partly to lower savings rates, but we also expect that levels of investment will remain high. Since investment relies on savings, the implication is that EM investment will have to be partly funded by foreign savings. This will have important implications for EM current accounts, which we will describe later.

The second implication is that EM growth will be more reliant on credit growth, since savings are expected to be lower, while consumption continues to grow. Figure 10 highlights how credit-dependent growth in EM has been over the past decade (the 2010 data is slightly skewed due to the high debt levels during the financial crisis). We expect this trend to continue, except with diminishing marginal returns to credit, as has been experienced in developed markets. All else being equal, this is an important reason why we should expect growth overall to be structurally lower going forward.

Figure 10

Long-term inflation expectations

Since New Zealand first initiated inflation targeting in 1990, at least 50 countries and monetary unions have incorporated inflation targets into their monetary policy framework, a large number of which are emerging markets. This has had the effect of bringing down both inflation and the volatility of inflation across emerging markets as can be seen in the graphs below.

There is reason to believe that inflation may continue to fall structurally lower in EM. Generally, there is a relationship between the level of inflation and the level of national income which we can observe empirically by looking at countries that have moved from low GDP per-capita levels to those associated with developed markets. In Figure 12, we can clearly see the pattern of lower inflation as income per capita grows.

Of course, the direction of causality here is likely to run both ways: certainly higher incomes are associated with a higher quality

institutional framework, political stability and a more developed and diversified economy – all of which are elements of long-term sustainable growth.

But lower, more stable inflation also lowers nominal interest rates, reduces the volatility of interest rates, attracts capital and makes long-term investment more predictable and attractive. These are then part of a positive feedback loop towards lower and less volatile inflation.

Higher incomes also lead to inflation baskets, which have less exposure to food and transportation, both of which have a higher sensitivity to energy prices and currencies and thus tend to have more volatile prices. More developed countries have inflation baskets that are skewed towards services, medication, personal care, culture and education, all of which have more stable prices. We show the baskets of selected EM in Figure 13.

“For these reasons we believe it is likely that EM will continue to experience lower inflation going forward, and through the cycle.”

For these reasons, we believe it is likely that EM will continue to experience lower inflation going forward, and through the cycle. At the same time we expect that the differential between EM and DM is also likely to contract as central banks in developed markets continue to target higher inflation through excess liquidity provision. Overall, the differential between EM and DM inflation has been around 3% over the past decade, where we believe this could drop to around 2%. Assuming average inflation for DM of 1.5-2%2 is repeatable into the next decade, a fair assumption for EM inflation would be 3.5-4% which is very much in line with the IMF’s expectations for inflation over the coming five years.

Inflation
Inflation
Inflation
Inflation
Inflation



“If an exchange rate implies that goods in a given country are much more expensive than in the country...



...on the other side of the exchange rate, then the rate can be said to be overvalued in terms of the first country.”

Potatoes

Currency valuations and expected returns

Currency returns are of critical importance to emerging market bond investors. Since 2002, local currency emerging market debt has returned 10.1% annualised and 4.7% of that return has been due to the returns from currencies3 (including both currency appreciation and the yield earned on currencies). Over the coming five years, we expect that currencies will offer a similar return, although almost all of the return will be due to currency yields, with nominal spot currency valuations remaining largely unchanged.

We estimate long-term currency returns using the most basic driver of value: purchasing power parity (PPP). If an exchange rate implies that goods in a given country are much more expensive than in the country on the other side of the exchange rate, then the rate can be said to be overvalued in terms of the first country. Overvalued exchange rates tend to reverse – sometimes sharply (and painfully) so.

The one important adjustment we make to PPP is to allow for the changing differences in per-capita income levels. The reason for this adjustment is that higher income level countries have shown a very resilient empirical tendency to run relatively more expensive exchange rates than that which would be suggested by PPP alone. This phenomenon – known as the Balassa-Samuelson effect – can be ascribed to rising incomes, which lead to higher inflation in the non-tradable goods and services sector where price differentials cannot easily be corrected by currency valuations alone.4

In Figure 14, we run though these calculations, for each country, to get an overall average currency return. First we calculate the currency correction to the Balassa-Samuelson adjusted PPP fair value, typically judged to adjust by half the difference over a three- to five-year period.

“Higher income level countries have shown a very resilient empirical tendency to run relatively more expensive exchange rates than that which would be suggested by PPP alone.”

Then we estimate the return due to the relative adjustment in per-capital income of each country. This gives us an expected real (i.e. inflation-adjusted) currency return to which we can add back the effect of inflation to achieve an expected spot return in nominal space i.e. the observed currency return. The final component is the interest rate differential, or carry, which the currency is expected to earn.

While this approach to currency valuation is well-founded in economic theory, we must still acknowledge that offshore capital flows also matter, particularly in the shorter term and given our assumptions that domestic savings rates in emerging markets are likely to fall while levels of investment and domestic consumption remain high. Capital inflows lead to a build-up of external debt which, in the past, has led to currency instability (we refer particularly to the Asian crisis of the 1990s).

With this risk in mind, we analyse the external funding needs of EM regions below (Figure 15) and find that, in fact, Asia is likely to continue to be an exporter of capital (running current account surpluses) although not quite to the same extent as in recent years. The other EM regions are expected to require between 2% to 4% of GDP capital inflows per year. In order to assess the impact of these financing needs, and using the IMF forecasts of external financing needs, total public and private current external debt and economic growth, we map the expected evolution of external indebtedness in each of the EM regions. We also make the very punishing assumption that all external financing will be in the form of debt financing – whereas, historically, foreign direct investment and equity portfolio flows have always been a bigger part of the story.

“Balassa-Samuelson effect – can be ascribed to rising incomes, which lead to higher inflation in the non-tradable goods and services sector where price differentials cannot easily be corrected by currency valuations alone.”
balassa
balassa2

Bringing it all together

As highlighted before, local currency emerging market debt is driven by bond and currency returns. These returns can also be seen to comprise a yield contribution, as well as capital gains or losses, which we have considered separately based on our analysis above. Over the medium term we expect real yields to average between 3.5-4.0%. This is based on our long-term potential growth estimation, as well as savings rates and the composition of these economies. As emerging markets continue on their path of development we expect inflation to continue to range between 3.5-4.0%. Taking these expectations together we see a fair value yield range for nominal bonds of about 7.0-8.0%.

Our analysis of currencies reveals that they are currently about 3.5% cheaper on a real effective basis. Allowing for some resumption to fair value and additional returns from the expected higher productivity gains (based on the Balassa-Samuelson theory) leads us to expect a real effective appreciation of EM currencies of about 1% per annum over the next 5 years. Taking expected inflation differentials of about 2% into account, we would expect spot currencies to depreciate by approximately 1% per annum even though currencies themselves, including carry, should return 3% to 4% per year.

“The key conclusion is that although the double-digit returns of the past are unlikely to be repeated, the future expected returns from local EMD remain attractive and positive even under more extreme negative scenarios.”

Total Return expectations

We can now bring this all together and calculate expected returns for local EMD using these inputs. Our base case assumptions are for yields to rise to 7.5% over five years and currencies to return -1% per annum on a spot basis. Obviously, as yields rise, coupons and maturing bonds that come due before the five years are over, can be reinvested at better yields. So the correct way to calculate the expected returns is to analyse the full local currency bond universe (proxied by the JP Morgan GBI-EM Global Diversified Index) on a cashflow basis, assuming a steady path of yield rises. We show the results in the table below, where we highlight our base case, but also show returns for different scenarios.

The key conclusion is that although the double-digit returns of the past are unlikely to be repeated, the future expected returns from local EMD remain attractive and positive even under more extreme negative scenarios. We believe that emerging markets remain on an attractive sustainable development path and that their bond markets and currencies will continue to benefit, offering investors a relatively attractive medium to long-term investment.

Antoon de Klerk, Peter Eerdmans and Grant Webster are part of Investec Asset Management’s Emerging Market Debt team. Antoon and Grant are portfolio managers and Peter is co-head of the Emerging Market Debt team.

Annualised

1 IMF, IAM Calculations.

2 IMF data, CPI expectations, 2014.

3 J.P. Morgan, IAM Calculations.

4 The Big Mac index is good for calculating currency (mis) valuations on a like-for-like basis, but just how easy is it to buy a Big Mac in Rio de Janeiro and eat it in New York? Big Macs are a non-tradable good. Rising productivity in Brazil causes incomes to rise too, increasing demand for Big Macs, which Brazilians cannot physically buy more cheaply in the US. Comparing the average of tradable and non-tradable prices across countries makes currencies seem mis-valued on a purely PPP basis.

Download PDF
3.2
Private Equity in Africa: The Challenges & the Promise by Nick Tims