Sequencing Risk

Sequencing risk refers to the risk of running out of funds due to the order of the returns from your investments.

The average yearly return of your investments may be the same in two scenarios, however the order of the returns may result in wildly different outcomes. Generally if you receive good returns when you have a large amount of funds (generally near the beginning of your retirement), and poor returns when you are low on funds (generally near the end of your retirement), you will do well (and vice-versa)

Is sequence risk something we should worry about, and if so, what can we do about it?

In this post I will look at some hypothetical scenarios based on past real returns, examine some proposals on how to mitigate the risk, and then look at our situation in particular.

Sequencing Risk – An example

The real returns from AustralianSuper for their balanced option from 1987 (its inception date) to now (2018) are shown below:

You can see that in most years the return has been quite good, but there also have been some years with quite large negative returns (the GFC!).

Now, let’s assume you are 65 and have $1M in Super with AustralianSuper balanced option. For simplicity, we will assume there is no Aged pension.

In the first scenario, we assume we get the returns commencing in 1987 and ending in 2009 i.e. starting from AustralianSuper’s balanced option inception date and ending at the GFC.

In the second scenario, we assume returns in the reverse direction. i.e. starting from the GFC and going backwards to the inception date.

The average yearly return in Scenario 1 is the same as Scenario 2.

We’ll assume that we spend the same amount every year and it is the amount ($86,453) which makes Super zero at age 88 with the average return during the period for all years. (6.7%).

Here is the value of capital by year:

 

You can see that sequencing risk is real. Capital runs out during age 78 for scenario 2, but there is still plenty of capital left (about $388K) at age 88 for Scenario 1. A huge difference!

What can we do about Sequencing Risk

There are no easy solutions for mitigating Sequencing risk. Here are some I have thought about.

Keeping some of your funds in cash

The idea here is you keep some of your funds in cash and spend it prior to using the funds in your shares. If you have a bad return at the beginning (the time you are most vulnerable), then it is not quite so bad as the negative return does not affect cash. As you run down your funds, the impact of negative returns becomes smaller so there should be less need to hold cash.

The diagram below shows what happens if we hold $450K in cash and $550K in Super and spend $86,453 each year while running down the cash before the shares.

 

As you can see, this strategy barely improves on the GFC at beginning scenario, and makes the GFC at end scenario considerably worse. It doesn’t work!

Moderating your Spending

This strategy involves reducing your spending when your funds go down. The idea is you might be able to recover your funds by using the subsequent higher returns and a lower spending rate. In the example below, we have the GFC at the start, and we work out how much to spend each year based on assuming returns will be the average for remaining years and make funds at 88 equal to zero.

The average spending in this approach is about $64K (rather than $86K), or about 74% of the original planned spend. Compared with the GFC at end scenario, we also have $388K less funds at 88. This strategy doesn’t really work either, but it does at least ensure you do have some funds throughout your retirement!

Here is the spend per year:

Building up a buffer/Working extra years

The idea here is you build up more funds than you would require assuming average returns each year, just in case you have a sequence of bad returns at the start. You might have an idea that bad returns are due if you have had a recent sequence of very good returns, or if you have been using predictive tools such as those discussed in this post. In the diagram below, we assume we work 2 more years, and save $60K each year of working, and of course we save the run down of funds during those two years. We also assume we spend the standard $86K each year once we do stop work. The diagram below shows what happens here assuming we have the GFC at the start.

In this scenario we don’t actually run out of funds and we have about $147K at the end. This strategy works (at least in this limited case).

Timing the Market

Another strategy would be to retain some funds (or perhaps a percentage of funds) as cash and to use these funds to invest in Super (or shares if this is not possible) when the market drops more than a certain amount during a year. The idea is you keep some funds to buy low.

As an example, I have assumed that at the end of each year, the cash/shares ratio is re-balanced to a certain percentage. If the return drops to less than -10% over a year (only once in the last 30 years for AustralianSuper), the cash is used to purchase shares at the end of the year and we stop rebalancing I.e. we maintain 100% in shares moving forwards.

Here is a diagram showing the age at which funds run out assuming various cash percentages and commencing at different years between 1987 and 2009..

 

This strategy doesn’t seem to work because most of the time the spare cash is dragging down the returns. In general, the higher our cash percentage, the lower the Age at which funds run out.

Sequencing Risk in our Situation

OK, let’s take a look at our situation. Our situation is a bit different than the above example as we have the Age Pension, which should moderate to some extent the impact of Sequence risk (when funds go down, income from the Pension increases). In addition, we work out how much to spend each year on the basis of our planned returns of 3.2% and making assets at 90 equal to zero (i.e. the same as the “Moderating Your Spending section” above.)

I assumed that we experience the AustraianSuper returns, starting in a year as per the below diagram:

These are historical returns, but I’ve added in 2019-21 figures by adding in another GFC (to offset the recent good returns). For any given starting point, once we get past 2021, we start at 1987 again. I assume we don’t sell our house. Here is the average spend per year with a start date on the x-axis:

 

Notice that even in the worse years to commence, the average spending is about $107K per annum, much larger than our planned $90K as discussed in the 2018 in review post. The reason for this is that, for the purpose of modelling how much we will have to spend, we have assumed real returns of 2.8%, in accumulation mode and 3.2% in pension mode, while AustralianSuper returns are on average 6.7%. in accumulation mode and about 7.2% in pension mode. Effectively we have built a buffer by assuming average Super returns that are a lot lower than commonly occur in real life.

The result of assuming lower returns when planning spending levels but actually experiencing higher returns (although not necessarily in the desired order!) is that typically  we end up spending less in earlier years and more in later years, something we don’t really want because we would prefer to spend more when we are younger and healthier. This is the price to pay for protecting ourselves against sequencing risk.

The below shows the spending pattern that would result from retiring in 1987:

And here is the worst average spend (resulting from retiring in 2008):

The lowest annual spend is three years in and is about $75,240 or about 83% of initial expense. The worst average corresponds to this sequence of returns:

Why did we choose the return of 2.8% when we did the original modelling? Well, it’s the return the calculators on the net use! Maybe they are setting this return rate low to protect us from sequencing risk (or maybe to cause us to  deposit more funds with the company publishing the calculator!).

Going back to our “Sequencing Risk – An Example” section, if we use a return rate of 3.3% to work out the spend level that causes funds to run out at 88, our calculated spend levels would have been about $63K. If we spent at this level and received AustralianSuper returns starting from the GFC, we would not have run out of funds (at 88  we would have about $145K). i..e the strategy of building up a buffer based on using lower than real-life returns would have worked.

Our Situation and Working Extra Years

Ok, what happens if we were not happy at the prospect of the worst case average spend ($107K). Well, we could work an extra year. Let’s assume we do this and save $80K, while spending about $50K while working.

In this case, our average spend from the retirement age is $116K . The updated diagram is shown below.

Working an additional year (a total of 2 years) results in the an average spend of $133K, quite an improvement, and the updated diagram is again shown below:

Conclusions

Sequencing risk is real.

And there are not many things you can do about it.

The best strategy I am aware of is to build up a buffer. One way of doing this is to assume lower returns than those experienced in real life to work out your desired spend levels. In this case if you have a sequence of poor returns early in your retirement, you are still likely to achieve your desired spend levels because subsequent returns will be higher than expected and will frequently be enough to bring you back to your desired spending levels.

My investigations show that if you are using the real return levels used by online calculators (about 3.2% in Pension mode) to work out your spend levels, you have implicitly included a buffer, as the actual real returns (at least over the last 30 years) in Pension mode for a good Super fund is about 7.3% (*).

(*)  – Although, read the next post!