Chapter 14

---

Expected Return and Risk

Forward-looking expected return, risk and correlation

‘Prediction is very difficult, especially if it’s about the future.’

Niels Bohr

Asset allocation is the process of appropriating capital to each asset class. It broadly sets the portfolio’s risk and return profile and, as such, it is the heart of investment strategy. It should match your return objective, risk tolerance and investment constraints, in particular time horizon and liquidity needs.

Asset allocation should mix assets based on how they are likely to perform in the future, not based on how they have performed in the past. To do so, we need a crystal-ball framework to predict the future.

Government’s assumptions

Each year, pension administrators need to give DC savers an illustration of the pension they may get when they retire. The illustration is known as Statutory Money Purchase Illustration. The law requires pension managers to provide it, based on assumptions laid down by the government. The assumptions should be detailed in your annual Pension Benefit Statement.

Whilst you can rely on these assumptions, we aim to generate independent forecasts that are better connected to current market conditions.

Capital market assumptions

Designing a long-term asset allocation aligned with your investment objectives starts with forming projections about the long-term prospective return, risk and correlations of assets that you consider including in your portfolio. These expectations are called capital market assumptions (CMAs) since they estimate how different capital markets (equities, bonds, cash) are likely to behave in the future.

It is important to know how assets tend to behave to understand their characteristics. However, past returns cannot just be simply projected into the future to estimate future returns as they differ from period to period.

Future likely returns depend on where we are now in the economic cycle. When assets are expensive, the expected return is low, and vice versa. This is what the CMAs are trying to capture.

CMAs are formulated for asset classes, not individual securities. When formulating an investment strategy, you choose in which assets to invest, not individual securities yet.

First make the beta decision on how to allocate to different asset classes. Then make the alpha decision on which securities to buy in each asset or outsource it to fund managers. Security selection follows asset selection.

In the next sections we explore how to estimate the future likely returns, risks and correlations of asset classes so you can blend them into an asset allocation fitting your needs. Prophesising about the future is challenging but, nevertheless, we will go through a relatively straightforward way to do so. It might be technical, but it is worthwhile. Use the expected returns to manage your expectations about what your portfolio is likely to achieve over the long term and plan accordingly.

Expected returns

Formulating CMAs is challenging. Unfortunately, looking at historic returns and extrapolating them into the future can be misleading.

Figure 14.1 shows the cumulative performance of the S&P 500 Index since January 1928. US markets have a reliable long historic track record, easy for studying the behaviour of assets over the long term with a statistically significant number of observations.

As Figure 14.1 shows, equities can move up, sideways or down at different times, for different lengths of time and at different paces. Can you tell how they are likely to behave over the next 10 years, based on how they have behaved in the past? It is a guess.

Figure 14.2 shows the annualised return of the S&P 500 Index during each decade. Whilst the average annual performance over the entire period since January 1928 is 9.4%, performance swings from decade to decade. The average performance in the last four decades deviated substantially from the long-term average. Past performance is hardly a reliable indicator of future results.

Since we cannot use past returns to formulate CMAs, we need a simple forward-looking methodology, based on readily available data.

Arguably, the simplest, most intuitive approach to formulate asset classes’ expected sustainable returns is breaking them down into their building blocks – their separate drivers. We estimate the expected return of each building block and aggregate them to formulate each asset’s expected return.

Our aim is using observable market data, telling us what the market collectively thinks about assets’ future performance over the long term (5–10 years). Not relying on our subjective views, we let the data speak for itself, implying expected returns for us.

Equities

Equities’ sustainable, long-term expected return is made of three building blocks:

1. Dividend yield (carry).
2. Earnings growth (momentum).
3. Re-pricing (valuation).

The first building block is straightforward. Dividend yield is cash payments to shareholders. When looking at a stock market, the dividend yield of the index representing the market is the relevant figure.

For example, the dividend yield of the FTSE 100 Index is currently about 4.0%.¹ When aggregating all the dividends of the index’s constituents and dividing the sum by the index’s price, the ratio is approximately 4.0%.

The second building block is earnings growth. Share price reflects the present value of future dividends. Dividends are paid out of company’s earnings. The pace of earnings growth is, therefore, part of equity’s return.

Whilst it is difficult to predict the rate of earnings growth for a single company or over the short term, it is easier doing so for the entire market over the long term. We assume that, over the long term, trend earnings growth is linked to general trend economic growth (nominal GDP growth rate).

However, earnings grow more slowly than GDP because growth of existing corporations explains only part of GDP growth. Historically, earnings growth in the USA lagged GDP growth by about 2% annually. This figure differs across different markets and across different regimes.² To simplify, we apply a 2% dilution factor between trend nominal GDP growth and earnings growth.

In the UK, for example, the average annual real GDP growth rate has been 2.6% since 1948.³ With BOE’s 2% target inflation rate, trend nominal GDP growth rate is about 4.6%. This means an estimated 2.6% earnings growth after deducting the dilution factor.⁴

Instead of trend nominal GDP growth rate, you can use economic forecasts. Future trend economic growth might be slower than historically due to slowing developing economies and aging population in the developed world. Free sources of economic forecasts are available online.⁵ However, forecasting GDP growth is challenging; many of the forecasts prove incorrect.

The third and final building block of equity return is re-pricing. It captures the price the market currently assigns to the stock market.

One metric indicating the valuation of the equity market is price-to-earnings ratio. P/E expresses the relationship between share price and earnings per share (EPS). For example, if the share price is £10 and EPS is £0.50, P/E is 20. Price reflects 20 years of earnings.

P/E tends to revert back to its long-term mean, unless there is a structural change, justifying the stock market to be more expensive or cheaper than it has been in the past. For example, new technologies, like the advent of the internet, can explain a higher market P/E since companies can permanently increase productivity.

If long-term average P/E is, say, 15 and current P/E is 20, the market is rich. If current P/E is 10, the market is cheap. P/E is likely to revert back to its 15 level. We do not know when it is going to happen. We do not know if it is going happen. But we assume it is likely to happen eventually.

Since we are estimating returns over the long term, including P/E mean reversion as part of expected returns is sensible. Empirical evidence shows that when P/E is very high, equity market performance over the next 10 years is much lower than when P/E is very low. It actually works over the long term.⁶

Figure 14.3 shows the rolling 10-year annualised performance of US equities and their cumulative performance superimposed. The level of rolling return at each point is average annualised performance during the previous 10 years. For example, the average return in the previous 10 years up to December 2015 was 8.3%.

The chart reveals how equity returns swing like a pendulum. Optimism pulls stocks to unsustainable highs and pessimism pushes them to unjustified lows. When markets are rich (high P/E), future returns are low and when markets are cheap (low P/E), future returns are high.

How do we translate P/E mean reversion into expected annual return? Here we need a simple formula. If long-term average P/E is 15 and current P/E is 10, we estimate the equity market will appreciate by 50% due to P/E mean reversion.⁷ To express this 50% as annual average return over the next 10 years we use the formula:

r = (P/E_{long-term average} ÷ P/E_current^(1/n) − 1

The annual compounded average return is 4.14% over the next 10 years.⁸ It matches a 50% overall return.⁹

The current P/E of the FTSE 100 Index is about 18.0.¹⁰ The index’s long-term average P/E is around 15.8. It suggests the index is expensive. P/E reverting back to its long-term average implies a −1.3% annual return over the next decade.¹¹

P/E average greatly depends over which time period it is calculated and which P/E is used (based on reported or estimated EPS). Use the longest history available since a last structural change and the same P/E methodology for both current and historic average. Assumptions about P/E can have a big impact on expected returns.

Putting everything together, assuming 4.0% dividend yield, 2.6% trend earnings growth, and −1.3% re-pricing, the total expected return for the FTSE 100 Index is 5.3%.¹² The expected return is above the index’s historic 4.9% return since January 1998.

This is an estimated average return over the next 10 years. During most single years, it is going to be wrong. However, it is a good indication for the level of return and the current valuation of the equity market. Expected return and valuation have an inverse relationship. Low expected return implies rich valuation. High expected return implies cheap valuation.

Since expected return is related to current observable variables (dividend yield and P/E) we can update it every month.

Expected returns of equity markets are not accurate; they rely on a number of assumptions. For example, using a long-term average P/E of 12 instead of 15.8 would swing the expected return from 5.3% to 2.6%.¹³ This illustrates how sensitive expected returns are to the assumptions we use.

The important thing is keeping the methodology consistent across different markets so they are all valued in a comparable way.

Bonds

Bonds’ sustainable, long-term expected return is made of three building blocks:

1. Yield to maturity (YTM).
2. Price movement.
3. Cost of default.

YTM is the return you will realise by holding a bond to maturity, assuming it does not default. It incorporates coupons and the positive or negative return you realise when buying a bond at a discount or premium, respectively.

When not holding a bond to maturity, price movement is total return’s second building block. Price movement is a function of changes in yield. The question is how to estimate whether the yield will move and if yes to which level.

For 10-year government bonds, an estimate for a yield’s fair-value level is the economy’s year-on-year (YoY) nominal GDP growth rate.¹⁵ The 10-year yield reflects expected inflation and expectations of future short-term interest rates, which are linked to economic growth.

For example, if current 10-year gilt yield is 2% and UK nominal GDP growth rate is 4.6%, yield may rise by a whopping 2.6%.¹⁶ If duration of 10-year bonds is 9.5 years, a 2.6% increase in yield means roughly a 24.7% capital loss.¹⁷ This is massive, in particular when the carry from the yield is only 2%. Govies’ downside risk can match that of equities.

We do not know when and if yields will normalise, reaching fair level. First, we need to estimate whether trend real GDP growth rate is likely to persist in the future. Second, we do not know whether the inflation rate will hit BOE’s 2% target. So the ‘normal’ level for yields is questionable. And third, we do not know the impact of central banks’ unconventional monetary policies on yields. QE distorts markets.

Assuming conservatively that within our 10-year horizon yields are likely to reach a level of 2.5%, price movement is −4.8%.¹⁸

The total expected return of 10-year gilts is therefore −2.8%.¹⁹ At this yield level govies are expensive. They are expected to generate negative total returns.

Buying gilts may still be sensible if they are held to maturity or if they add benefits of diversification and protection to your portfolio. In case of a flight to quality, yields might drop, not increase. Assuming they are held to maturity, the expected return of 10-year gilts is 2.0% (YTM).

The third building block of bonds’ expected return is cost of default. Whilst you should not expect the UK, US, German and Japanese Governments to default on their sovereign debt – if it happens, it is a financial Armageddon – corporations and emerging nations might do so. The expected cost of default should be deducted from expected return.

Cost of default (credit loss rate) on a diversified bond basket is a function of average credit rating, probability of default (default rate) and average recovery rate. When bonds default, lenders usually recover some of their loans – not all is lost.

The rating agency Moody’s publishes reports online with historical statistics on default rates and credit loss rates. Apply the average historic rate to the appropriate credit rating of your bonds. Table 14.1 includes the average credit losses of different credit ratings.

Table 14.1 Credit loss rates by letter rating, 1982–2010

Credit rating	Credit loss rate (%)	Rating description
Aaa	0.00	Highest quality, minimal credit risk.
Aa	0.02	High quality, subject to very low credit risk.
A	0.03	Upper-medium grade, subject to low credit risk.
Baa	0.13	Moderate credit risk, medium grade, may possess certain speculative characteristics.
Ba	0.73	Have speculative elements, subject to substantial credit risk.
B	3.01	Speculative, subject to high credit risk.
Caa to C	13.70	Caa poor standing, subject to very high credit risk. C typically in default, with little prospect for recovery of principal or interest.
Investment grade	0.06	Ratings Aaa to Baa.
Speculative grade	2.78	Ratings Ba to C.
All rated	1.01	Ratings Aaa to C.

Source: Moody’s Investors Service, Corporate Default and Recovery Rates, 1920–2010

The UK’s credit rating is Aa. A gilt’s cost of default is negligible (2 basis points). The cost of default becomes relevant for bonds with a credit rating of Ba and below.

Spread products

The first building block of IG credit, high yield and emerging market debt (spread products) is YTM. Spread products offer a spread over rates (yields) of comparable bonds of creditworthy governments as compensation for higher credit risk.²⁰

The second building block, price movement, is a bit tricky. Assuming spread remains constant, changes in the government bonds’ rates affect the spread product’s price, adjusted for duration (normally shorter than that of govies).

For example, the 10-year gilt rate is 2% and it is expected to rise to 2.5%. We wish to buy a portfolio of UK corporate bonds with an average duration of 6.7 years (compared to the 9.5-year duration of 10-year gilts).²¹ Assuming a constant spread and a parallel shift of the yield curve, the expected drop in price due to increasing rates is −3.4% when considering the corporate bonds’ duration.²²

However, increasing govies’ rates usually correspond with improving economic conditions, so spreads should narrow, absorbing some of the rise in rates. Conversely, spreads should widen when underlying rates fall, as the economy may be slowing.

It gets convoluted since, over the short term, increasing rates tend to correlate with narrowing spreads. However, over the long term, higher rates tend to cause spreads to widen. So, to make it simple, we assume the price of spread products moves by half the duration-adjusted yield movement (if for govies we expect rates to rise 0.5% from 2.0% to 2.5%; for credit we expect yields to rise by half of that, so 0.25%).²³

The third building block is estimated cost of default.

This all sounds knotty. An example can simplify. Say you want to formulate an expected return for UK corporate bonds. In a website of an ETF tracking the index, you find its YTM is 3.6%, duration (often called effective or modified duration) is 8.7 years and average credit rating is A.^24,25

The expected return is YTM 3.6% plus price movement of half yield change times duration of -2.2% minus cost of default for A rating of 0.03% (negligible).²⁶ The expected total return is 1.4%.²⁷

This low expected return echoes the expensiveness of gilts and credit. IG credit still offers better returns than gilts due to higher yield (carry) and shorter duration. Going through this methodology facilitates understanding why different assets offer different projected returns.

Cash

For cash, expected return is simply short-term current rate. If, for example, the base rate is 0.5%, expected return is 0.5%. Add a spread depending on the rate of a money market fund in which you invest or the saving rate on your cash ISA or bank’s savings account. A time deposit or a fixed rate ISA can offer higher rates, but you commit your money for the term.²⁸

When the BOE increases or decreases its base rate, expected return changes accordingly (immediately for variable rate accounts, with a lag for fixed rate accounts).

Property

Property’s sustainable expected return is made of three building blocks:

1. Net rental yield.
2. Trend nominal GDP growth rate.
3. Costs.

Similar to dividend yield for equities and YTM for bonds, net rental yield is the income component of property’s total return. Since property is a heterogeneous asset class, rental yield depends on each specific property or property fund.

Trend nominal GDP growth rate represents property’s expected long-term price appreciation. It should keep up with inflation and economic expansion. Apply this estimate to both residential and commercial property.

Since buying property can be expensive and ongoing costs are high, subtract costs from total return.

Some ongoing costs are already deducted to reach net rental yield. One-off charges can be annualised over the time horizon. For example, 5% stamp duty when buying a property is equivalent to 0.49% per year over 10 years.²⁹

Putting it all together, assuming a 3.5% net rental yield (after subtracting ongoing costs), 4.6% trend nominal GDP growth rate and 0.49% annual costs due to stamp duty, property’s expected return is 7.6%.³⁰

Other asset classes

Other asset classes in your portfolio are likely to be listed on the stock exchange. These include REITs, listed infrastructure and listed private equity. The same methodology as that of equities applies to them.

By choosing different indices and data for a relevant country, you can calculate expected return for assets classes across the world. For example, for US equities, use dividend yield and P/Es of the S&P 500 Index and US trend GDP growth rate.

Compare and contrast your expected returns with other sources.³¹ It is always advised to have a reality check, ensuring you are not out of whack.

Expected risk

Whilst past returns are not a good forecast of future returns, historic risk levels are useful to indicate assets’ sustainable risk level. Simply use annualised volatility of monthly returns as the risk yardstick.

Figure 14.4 shows annualised volatility of US equities during each decade since the 1930s. Whilst volatility changes, since the 1950s it has been range-bound between 12% and 16%. Annualised volatility since the beginning of the 1950s is 14.4%.

To annualise monthly volatility, calculate the standard deviation of monthly returns and multiply it by square root of 12.³² For example, multiplying a standard deviation of monthly returns of 4.3% by a square root of 12 gives an annualised standard deviation of 15%.

Since we aim to formulate CMAs over at least the next decade, use historic monthly returns over the last 10–20 years or longer, if available, since a structural change.

The 1930s and 1940s are examples of structural changes. Following the 1930s Great Depression and the 1940s Second World War, financial markets have been noticeably less volatile. Returns since the end of the War are appropriate to estimate future risk.³³

Expected correlation

Correlation is the most difficult to estimate since it lacks economic fundamentals predicting it. It can wildly swing from one period to the next. Luckily, the most important CMA is return, then risk and, finally, correlation.

Figure 14.5 shows rolling 36-month correlations between UK equities and global developed equities and between UK equities and emerging market equities (EME). Notice how correlation vividly changes over time.

For example, in December 2015 the correlation between UK equities and EME was 0.63 but, during 2009, when most risk assets fell together, it was close to 0.90. This demonstrates the fluctuating benefits of diversifications.

Use historic correlations between monthly returns. The time horizon should be 10–20 years or the longest available since a structural change.

This method of estimating CMAs is by no means the most sophisticated or exact. It has limitations. However, this is a quick and dirty way to estimate the long-term returns, risks and correlations of asset classes.

A professional financial adviser may do a better job at modelling assets. However, even advisers may not be able to more accurately predict the future. Use this methodology as a second opinion to challenge your adviser.

Data

Tables 14.2 and 14.3 include data for formulating CMAs of major asset classes. The data need updating to keep the CMAs up to date with current market conditions.

Table 14.2 Historic returns, risks and data

Source: Bloomberg. FTSE 100 Index, MSCI World Index, MSCI Emerging Markets Index, iBoxx £ Gilts Index, iBoxx £ Non-Gilts Overall Index, Barclays Capital UK Govt Inflation Linked Index, BofA Merrill Lynch Global High Yield Index (measured in $), JPM Emerging Markets Bond Index Plus EMBI+ Composite, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £ except when indicated otherwise

Table 14.3 Correlation matrix

Source: Bloomberg. FTSE 100 Index, MSCI World Index, MSCI Emerging Markets Index, iBoxx £ Gilts Index, iBoxx £ Non-Gilts Overall Index, Barclays Capital UK Govt Inflation Linked Index, BofA Merrill Lynch Global High Yield Index (measured in $), JPM Emerging Markets Bond Index Plus EMBI+ Composite, UK Cash LIBOR TR 1 Month Index, UK IPD TR All Property Index. January 1998 to December 2015. Based on monthly total returns, measured in £ except when indicated otherwise

A free data source on indices is websites of ETFs (Exchange Traded Funds). ETFs normally closely track underlying indices and provide their key facts.

Summary

• Capital market assumptions (CMAs) estimate sustainable expected return, risk and correlations of asset classes over the long term (5–10 years).
• CMAs are based on observable market indicators, inferring valuations and expected returns of assets as implied by the market’s collective wisdom.
• Formulating expected returns involves breaking down assets’ returns into building blocks and estimating each one’s expected return.
• Equity’s expected return consists of dividend yield, trend earnings growth and P/E mean reversion. Bond’s expected return consists of YTM, price movement and cost of default. Property’s expected return consists of net rental yield, trend nominal GDP growth rate and costs.
• Expected risk is historic volatility since a structural change.
• Expected correlations are historic correlations since a structural change.

Notes

¹ Source: factsheet of FTSE 100 Index at www.ftserussell.com. Dividend yields of US indices are available on the market data centre of the Wall Street Journal at markets.wsj.com.

² The paper ‘Earnings Growth: The Two Percent Dilution’ by William Bernstein and Robert Arnott is available for free online. It includes dilution factors for different equity markets. Its data is also available at www.efficientfrontier.com, the website of William Bernstein. Visit the website for a wealth of useful information on asset allocation and investing.

³ Source: Office for National Statistics at www.ons.gov.uk. Trading Economics at www.tradingeconomics.com is a free source of economic and financial data for different countries, including forecasts of economic growth. The Bank of England at www.bankofengland.co.uk is a source of historic economic data and forecasts for the UK.

⁴ 2.6% = 2.6% real GDP growth rate, plus 2% inflation, minus 2% dilution. Instead of using Bank of England’s target inflation, you can use the breakeven rate of 10-year linkers.

⁵ Sources for economic forecasts include the Office for Budget Responsibility at www.budgetresponsibility.org.uk, the Organisation for Economic Co-operation and Development (OECD) at www.oecd.org, the International Monetary Fund (IMF) at www.imf.org and Trading Economics at www.tradingeconomics.com.

⁶ Earnings yield (E/P) is the ratio of EPS over the last 12 months divided by current price. It is the inverse of P/E. Comparing E/P with bond yields indicates the relative valuation of equities and bonds.

⁷ 50% = 15 ÷ 10 − 1.

⁸ 4.14% = (15 ÷ 10)^(1/10) − 1.

⁹ 50% = (1 + 4.14%)¹⁰ − 1.

¹⁰ As of December 2015; based on monthly end-of-month P/Es since May 1993. Ask your financial adviser for long-term average and current P/Es. Current P/Es of equity indices are available on the CNN Money website at money.cnn.com. The factsheets of MSCI indices include P/Es. Check the website of MSCI at www.msci.com. The average P/E of the S&P 500 Index is 12.0 based on monthly end-of-month P/Es since January 1973.

¹¹ −1.3% = (18.0 ÷ 16.2)^(1/10) − 1.

¹² 5.3% = 4.0% + 2.6% − 1.3%.

¹³ P/E effect = |minus|4% = (12 ÷ 18)^(1/10) − 1; 2.6% = 4.0% + 2.6% − 4.0%.

¹⁴ Visit the website of Robert Shiller at www.econ.yale.edu/~shiller for long-term data on the US equity market with calculations of CAPE. The GuruFocus website at www.gurufocus.com offers a number of valuations for different equity markets. The two valuation methodologies on the website are market-cap/GDP valuation and Shiller P/E. The website is free and includes detailed explanations of the methodologies.

¹⁵ 10-year bond yields can deviate from nominal GDP growth rate due to supply and demand for bonds and capital, inflationary pressures and expected returns of substitute investments. These and other factors impact the ‘normal’ level of yields.

¹⁶ Current bond yields are available on Bloomberg’s website at www.bloomberg.com.

¹⁷ −24.7% = − 9.5 × (4.6% − 2%).

¹⁸ −4.8% = − 9.5 × (2.5% − 2%).

¹⁹ −2.8% = 2% − 4.8%.

²⁰ Credit spread is determined by credit risk, differences in tax treatment between govies and corporate bonds and liquidity (govies are more liquid than credit).

²¹ The duration of a portfolio is the weighted average duration of its constituencies.

²² −3.4% = − 6.7 × (2.5% − 2%).

²³ This is a rough, generalist simplification. The relationship between changes in rates and changes in credit spread depends on many factors, such as economic regime, slope of the yield curve, inflation expectations, duration, credit rating, callability, coupon rate, time horizon, liquidity and forces of supply and demand.

²⁴ For example, check the website of iShares at www.ishares.com/uk for iShares Core £ Corporate Bond UCITS ETF, tracking the Markit iBoxx GBP Liquid Corporate Large Cap Index.

²⁵ Effective duration considers the embedded options in the bonds. Embedded options typically represent callable bonds.

²⁶ −2.2% = − 8.7 × (2.5% − 2%) ÷ 2.

²⁷ 1.4% = 3.6% − 2.2% − 0.03%.

²⁸ For example, now that the base rate is 0.5%, a 5-year fixed rate ISA can offer a rate of 2.5%. A savings bank account for children can offer a 3% rate up to a certain savings amount (£3,000).

²⁹ 0.49% = (1 + 5%)^1/10 − 1.

³⁰ 7.6% = 3.5% + 4.6% − 0.49%.

³¹ GMO provides free 7-year asset class real return forecasts on its website at www.gmo.com. Register on the website for free access. Research Affiliates offers long-term capital market forecasts at www.researchaffiliates.com. BNY Mellon offers 10-year capital market assumptions at www.bnymellon.com. JP Morgan Asset Management provides long-term capital market assumptions at am.jpmorgan.com.

³² Historic returns are available on Google Finance at www.google.co.uk/finance and Yahoo! Finance at uk.finance.yahoo.com. You can use historic returns of ETFs tracking equity, bond and liquid alternative indices.

³³ Two improvements to formulating expected risk are incorporating the asymmetry (fat tails) of returns (as measured by skewness and kurtosis) and unsmoothing appraisal-based returns.

³⁴ YTM is BOE base rate: www.bankofengland.co.uk.

³⁵ Property rental yield is available on factsheets of property fund and trusts.

³⁶ Dividend yields are available on factsheets of indices. FTSE 100 Index: www.ftserussell.com. MSCI World Index and MSCI Emerging Markets Index: www.msci.com. Check www.bloomberg.com for 10-year gilt yield. Factsheets of ETFs include data for indices they track, including YTM, duration and credit rating. Check www.ishares.com/uk for ETFs and their key facts.