Dan,
I'm happy to engage in this debate because I feel that you are covering an issue that I have had considerable concern about for some time now.
We know that many retirees in Ireland hold an ARF rather than an annuity and we know that the reasons for this are some combination of a desire on the part of retirees to "protect" the value of the pension and the incentive effects of commission payments on brokers (for the avoidance of doubt annuities pay less commission than an ARF). Note that you referenced fees vs commission in your original question.
We also know that many of the ARFs in Ireland hold a considerable amount on deposit (like your friend) in the belief that this a "conservative" strategy. As I have highlighted in this thread half of all retirees can be expected to live longer than the average and they may find out with the benefit of hindsight that such a strategy fails to preserve their income throughout their retirement. The reference I made to to "reckless conservatism"
As a point of reference the oldest known annuity payment in the world is an American Civil War pension that was being paid to a soldier who married a lady 50 years his junior and she is still alive today and continuing to receive the widow's pension - least anyone claims that an annuity is bad value! As an Actuary put it;"its an income for life, however long you live and if you die after 5 years, well it still did the job that it is intended to do"
So, the first observation I would make is that, all things being equal, the primary purpose of an ARF is to provide an income NOT an Inheritance.
The reason for this preamble is because retirees clearly face a series of difficult choices with imperfect information and this is made worse because some advisers do not understand these issues fully (for full disclosure I teach advisers this subject) and they have the conflict of interest due to commissions. As Upton Sinclair said;"its difficult to get a man to understand something when his salary depends on his not understanding it"
My second observation is that, at the conservative end of the risk return continuum, a portfolio with less than 20% risk assets is currently unlikely to cover the additional costs of managing an investment portfolio. This is just simple addition, add up all the fees and they are more than the expected return. Implicit in this is that any additional investment return is being lost in investment costs yet the investor is still taking on some additional investment risk. This doesn't seem right to me.
Therefore, any adviser acting as the client's fiduciary should stress that there is probably very little additional expected benefit to the client of pursuing a very conservative investment strategy but there is additional risk to the client. Included in this would be insurance company "secure" or "cash/deposit" funds, With Profits funds, Structured Products (tracker bonds) and so on.
As I have already noted the primary objective of an ARF for most people should be to maintain the income and hence financial security of the pensioner and (if relevant) their spouse for the whole of the rest of their combined lives - however long they both might live. Under these conditions, longevity combined with inflation are in reality the biggest financial risks for ARF investors. Yet the focus of many ARF investors is on short-term volatility so there is a mismatch between the problem and the approach taken by many people.
Consider that Insurance Companies hedge their risks in both the Annuity and Life Assurance markets using Long-term bonds. Therefore both Annuities and Guaranteed Whole of Life Assurance can be thought of as an investment asset with predictable risk/return characteristics like an investment directly in a bond.
For illustration purposes only: consider a "balanced" retiree who invests half of their ARF in an Annuity with a guaranteed spouses pension and the other half in equities.
A 60 year old with a 60 year old spouse 10 year guarantee plus 100% spouses pension can currently obtain an annuity rate of 3.155%pa. Source Irish Life.
So, rather than put half of the ARF in cash and fixed interest with a very low expected return currently we could put half of the ARF into an annuity with a guaranteed return of 3.155%pa and have no longevity risk. We have swapped investment risk for default risk and we still have inflation risk.
Here is a calculation that I rarely see anyone make.
For an initial €100,000 annuity purchase what is my equivalent future interest rate at any point in time?
Annual pension is €3117.20pa
After 20 years I have received total payments of €62,344 so my remaining "fund" is €37,656 but my pension is still €3117.20pa so my "interest rate" is now 8.278%pa
After 30 years I have received total payments of €93,516 so my remaining "fund" is €6,484 but my pension is still €3117.2pa so my "interest rate" is now 48.07%pa
Compared to a deposit strategy this could look good value in the future if one or both spouses lives a long time, which in 50% of all cases is going to be true. You either live longer than average or you don't.
The inheritance aspect is protected to an extent by the 10 year guarantee (if both spouses die early the minimum value of the annuity to the estate is €31,172) plus the value of the other half of the ARF - €50,000 plus or minus market risk. By limiting the exposure of the ARF to 50% equities you still have a "balanced" portfolio, its just that the defensive part is now the annuity rather than cash and fixed interest.
For anyone who has less than say 40% to 50% exposure to risk assets, this sort of thinking might be a better option than a traditional cash, fixed interest and equity allocation.
Equally, if gives a sense of the "critical yield" that is to say the return the ARF must earn every year just in order to match the annuity income forgone.
What I hope I have shown is that for a 60 year old today, that number is very close to the 4.25% with charges differential target.
Now to address your questions:
1. Are my calculations from Friday accurate or in need of correction?
Here is a summary of the process for deriving the assumptions.
Investment Objectives: Low Risk & High Return
At its most basic level, the two key objectives in investing are to maximise returns and minimise risk. Unfortunately, when it comes to asset classes, these two objectives conflict with each other. Less risky asset classes, like cash or bonds, tend to have lower expected returns than riskier asset classes, like emerging markets equities. If you want higher returns then you’ll generally have to accept higher risk.
The key benefit of building portfolios is diversification. In essence, because asset classes aren’t perfectly correlated (i.e. they don’t always move up and down together), combining asset classes together allows us to build portfolios that have both lower risk and higher expected return than any single asset class by itself. There is, however, a great deal of art and science to how this is done.
Traditional Mean Variance Optimisation
Traditional Mean Variance Optimisation, as introduced by Harry Markowitz in 1952, is a method of attempting to maximise the benefit of diversification i.e. trying to get the portfolio with the maximum expected return for any defined level of risk.
Estimates of each asset class’s expected return, volatility (a proxy for risk) and correlation (how it is behaves in relation to other asset classes) are fed into a computer optimiser. The output is a suggestion of the optimal portfolio for every risk level - those portfolios for which there is no other possible portfolio that has the same risk and yet a higher expected return. Such portfolios are said to lie on the efficient frontier.
The main problem with this method is that forecasting future returns, volatilities and correlations is very difficult and that this traditional method doesn’t handle these uncertainties very well (particularly with regard to expected returns estimates). These weaknesses lead to portfolios that are:
· unstable – for small changes in initial estimates of expected returns, the optimum portfolios change drastically
· not as diversified as they should be (often most asset classes tend to fall out of the final recommendation)
The net result is that traditional mean variance analysis is almost never used in isolation by asset management practitioners.
Black-Litterman
Black-Litterman (B-L) handles the problem of traditional MV optimisation by adding two extra steps. Risk and correlation numbers are still estimated (usually based on historic returns) but, instead of starting from a blank slate to estimate future returns on each asset class, B-L uses the views implicit in the market as a starting point. To a great extent this allows the portfolio manager the opportunity to identify asset classes where the historic return is unlikely to be a good estimate of the future return –such as in the Long-bond market.
The model then allows the portfolio manager to apply their own research to adjust these consensus views. Significantly each “view” must be assigned a confidence interval. Even so, the model weighs more heavily the consensus view of the market in recognition that markets are extremely efficient at pricing risk.
Using market consensus views of expected returns as our default starting point has a number of advantages. Firstly, knowing what the market/average investor “believes” is useful in challenging our own views. Secondly, where our own views differ from consensus, the B-L model provides a framework for combining our own views with market consensus, taking account of how confident we feel in our view. Overall this approach leads to portfolios that are:
· more robust – they don’t change drastically if we slightly alter our views
· very diversified – more asset classes are represented in the final portfolios
After we have the new adjusted expected returns we can follow the traditional MV analysis to create a set of efficient portfolios.
2. The charts from Friday no longer appear. Do you know why?
I am adding some warnings
3. What assumptions are made regarding currency movements in your portfolio? Is hedging involved and at what cost?
Historic volatilty of currency on average around 10 so it swamps fixed interest and cash. Therefore we always hedge currency risk in global bond portfolios. This is typically achieved through a one month forward currency swap between the fund manager and investment banks. The cost is negligible.
Equity volatility is greater than currency volatility so there is little or no benefit in hedging an equity fund.
4. If certain equity classes have higher expected returns than other equity classes, how is this higher risk taken into account in your modelling?
see 1
5. Similar question regarding bond investments - how is the added risk of higher yielding bonds taken into account?
see 1
6. Again regarding the modelling methodology used. Let's say:
- equities enjoy a stellar patch in 2015, 2016 and 2017 with returns in each of the three years of 15% p.a.;
- bonds returns and other variables do not change; and
- you were to update your model (at the start of 2018) so that it then examines the period 1970 to 2017
.......how would this impact on the model? Would it suggest that equities are more attractive than the model currently calculates?
There are two "models" that we use
1) Historic to determine the realised return of a given strategy in the PAST
2) Forward looking expectations using Black-Litterman
Of course as we have been in a 30 year bond bull market expected returns in the future are lower than returns in the past but that still doesn't mean you should declare that you expect no return from the bond market. Ask Bill Gross at Pimco!
Similarly, starting Shiller P/Es are probably the least worst predictor of equity returns in the future but they still only have around a 40% explanatory factor. There is still lots of room for uncertainty and we don't ever want to imply that we are certain of a particular course of events. Equities have a positive expected return and just because they are expensive or even very expensive relative to the average doesn't mean they won't go up more.
Remember Keynes line that; "the market can remain irrational longer than you can remain solvent"
See answer to 1