In our previous article we considered why bonds should be held in an investment portfolio and attempted to answer the question of how many bonds should be held. If an investor is prepared to hold risk free government bonds and is prepared to hold the bonds to maturity, then it is feasible to hold just one, so long as the investment is large enough relative to the size of the investment portfolio, to maximise the diversification benefits of bonds.
Remember Markowitz’s ideal investment portfolio of just two assets: one that only pays out when the sun is shining; and another that only pays out when it rains. Thus, the investor will have a portfolio that always pays out no matter the weather.
We expanded this into a portfolio of ten risk-free government bonds with maturities staggered so that just one matures over each of the next ten years. This being the concept of a bond ladder that allows an investor to have capital returned each year, without selling a bond, and if the capital is not needed it can be reinvested in a new bond at current interest rates.
This has the added advantage of not locking in low or high interest rates for a long period of time. Remember government bonds pay fixed interest rates, not floating.
However, we noted that government bonds pay very low interest rates relative to other bonds. And this means, if the bonds are being held to generate an income, the capital invested will need to be much larger than that required if higher yielding (but not risk-free) bonds are held.
This introduces the concept of credit risk – the subject of this article. Minimising credit risk so that total return from a portfolio can be maximised requires a greater number of bonds be held: possibly a lot more than ten.
Again, as noted last time, credit risk sits on a risk continuum between risk-free and equity risk. Risk in this case consists of systemic risk and non-systemic risk.
Systemic risk is the risk of industry or economic collapse. Banks that are considered too big to fail are banks that could bring down the whole financial system if they did fail. This is systemic risk.
Non-systemic risk is risk that relates to individual companies – the risk that a company could fail in isolation. This is also known as idiosyncratic risk.
When dealing with credit risk and equity risk it is possible to diversify away non-systemic risk but not systemic risk. Ten well-chosen stocks can eliminate non-systemic risk in an equity portfolio.
The number of stocks can be so few because equities come with both downside and upside risk. The value of the stocks held can move up or down.
However, with bonds risk is very much to the downside.
Ignoring short term fluctuations in bond prices, the best thing that can happen is that coupons are paid on time and principal is returned at maturity. The downside risk is that the bond issuer defaults and all is lost.
The asymmetric risk profile of bonds means that many more must be held in a portfolio to diversify away non-systemic risk. And, if an investor wants to rely on statistical averages to minimise credit risk, then the more bonds held in a portfolio the better, especially if the bonds are not investment grade.
Default statistics compiled by Global Rating Agency - S&P Global Ratings (S&P) - over the period from 1981 to 2016 show that the five year cumulative default rated for investment grade companies is 1% and for sub-investment grade companies the rate is 15%. In other words, over a five year period 15 companies in a portfolio of 100 companies can be expected to default.
If a recovery rate of 50% is assumed, and this may be a big assumption, then over a five year period a loss of 7.5% of the value of the portfolio can be expected.
This may not be a big deal if that portfolio is returning 5% per annum but how many investors will hold a portfolio of bonds issued by 100 different companies? What if you hold a portfolio of 20 bonds but 15 of those are issued by the companies that default over a five year period?
Averages can only be relied upon if you can play with large numbers.
S&P’s numbers show that more than 11% of sub-investment grade companies defaulted in just one year – 1991. Almost 10% defaulted in 2002 and again in 2009.
Among the most lowly rated companies assessed by S&P, nearly 50% defaulted in 2009!
Another way to think about this is to consider a portfolio that is returning 10% per annum. You can hold ten bonds in that portfolio and afforded to lose one in one year.
The income for the other nine will just about cover the capital lost on the one that defaults (assuming no recovery).
If the portfolio only returns 5% per annum then 20 bonds will need to be held to just about cover the capital loss arising from one default in one year. 50 bonds will be needed if the portfolio only returns 2% per annum, and 100 bonds if the return is only 1%.
If playing in large number is not an option, then holding investment grade bonds only will greatly reduce an investors’ risk. Unfortunately, outside of government bonds, most bonds sold to retail investors are not investment grade.
The author Philip Bayley is the Principal of ADCM Services, publisher of The DCM Review an independent online commentary, analysis and data on Australia & New Zealand's debt capital markets. He is also a contributing editor to Banking Day and a director at Australia Ratings.