SORC 18 Part 2 – Vanguard Dynamic Spending

Today’s post is our second visit to the Science of Retirement Conference 2018, held in London recently. We’ll be looking at Vanguard’s strategy of Dynamic Spending.

Dynamic Spending

The first lecture of the day at SORC 18 was from Garrett Harbron, who told us about Vanguard’s strategy of Dynamic Spending.

• Garrett warned us that the relevant paper hadn’t been published yet.

His slides are mostly pictorial, and I faced the prospect of recreating his lecture from memory.

But while it’s true that Garrett’s work on UK retirees can’t be found on the web, there are plenty of similar white papers and blog posts from Vanguard staff on the other side of the Atlantic.

• So what follows is a blend of the US material and Garrett’s slides.

You should be able to tell where a chart comes from by looking at the currency unit it uses.

A lot of the US papers involve Colleen Jaconetti, so it only seems fair to give her a name check, too.

Retirement in the 21st Century

Our retirement¹ is not the retirement that our parents had.

• Many are retiring earlier.
• Most are living longer.
• And DB pensions have been replaced with DC pots.

All of which means that the income you take from your portfolio is more important than ever.

And falling yields and interest rates mean that earning that income safely is more difficult than ever.

---

Here in the UK we have the lowest state pension (as a proportion of the average wage) in the OECD.

• Which means that we have the highest dependency on our private pension portfolio.

But how much can we take out, and not have to worry about running out of cash?

What people want

Investors in retirement want three things:

1. Stable income
2. Liquidity and growth
3. An income for life

Unfortunately, these three are mutually exclusive.

• You can’t have all of them at once.

It’s like the famous project triangle of quick, cheap and good.

• Only two (and in fact, any two) are possible from a single solution.

The 4% rule

As with most discussions of drawdown, we need to consider of the 4% rule.

• Based on US research from the 90s, this rule suggests that taking 4% of the initial balance from a portfolio gives you a 95% chance of your pot lasting 30 years.

Note that even the 4% rule means that 5% of people run out of money.

• There are few certainties in life.

Based on current economic conditions, increasing longevity and historically lower UK rates of return, UK retirees might want to consider a lower constant withdrawal rate (say 3.3% pa).

You also need to consider whether you would feel comfortable taking a constant drawdown after a market fall.

• And if your portfolio had a good run, wouldn’t you want to spend a bit more?

Four levers

There are four levers that affect portfolio withdrawal rates:

1. Time horizon
   • How long do you need your portfolio to last (how long is your retirement expected to be)?
2. Asset allocation
   • Are you prepared to stick with risk but rewarding equities, or do you want the illusory safety (stability) of cash and bonds?
3. Degree of certainty desired
   • The 4% rule works on a 95% chance of your money not running out.
   • Could you live with 90% or even lower?
   • Or do you need a 99% or even a 100% certainty?
4. Spending flexibility
   • Do you need to spend the same amount every year?
   • Or could you spend more in the good times, and tighten your belt in hard times?

Garrett’s presentation focused on the last of these, and the various possible approaches to spending.

Alternative approaches

There are two popular approaches – other than the 4% rule – to spending in retirement.

• Note that Vanguard often refers to the 4% rule as “dollar plus inflation”.
• This is because you are allowed to increase the year 1 spend in line with inflation.

1 – Percent of portfolio

• It’s easy to confuse this with the 4% rule, but it means that you take a set percentage of whatever your portfolio balance is each year.
  • This means that market fluctuations are taken into account.
• The percentage is normally calculated from a modified (extended) version of life expectancy.
  • So if you expect to live for 20 years, you might want to set your portfolio to last for 30 years.
  • Which might mean that you took out 3.3% of the portfolio to spend each year.

2 – Dynamic Spending

• Here annual spending fluctuates in response to market performance, but only within limits.
  • There is a ceiling on how much spending can increase after a good year.
  • And a floor on how much it can fall after a bad year.
  • Vanguard normally sets these limits at 5% upwards and 2.5% downwards.

Percentage spending

The drawbacks of percentage spending are:

1. Your spending might fluctuate wildly from year to year.
   • This is a particular problem for those whose fixed expenses (housing, food) are a large proportion of total spending (ie. less well-off retirees).
2. You might end up leaving a large pot behind if you die early.
   • This one is common to most spending strategies.

Dynamic Spending

Once you’ve chosen your initial spending rate (withdrawal rate), and your floor and ceiling, applying the rules is pretty straightforward.

1. Multiply your initial portfolio size by your spending rate to get the Year 1 spend.
2. At the end of the year, multiply the Year 1 spend by:
   • 1 plus the ceiling to get the Yr 2 ceiling
   • 1 minus the floor to get the Yr 2 floor
3. Now multiply the portfolio value at the start of Yr 2 by your spending rate to get the Yr 2 spend.
   • If the Yr 2 spend is within the floor to ceiling range, that’s what you spend.
   • If the Yr 2 spend is higher than the ceiling, then the ceiling is your spend.
   • If the Yr 2 spend is lower than the floor, then the floor is your spend.
4. Whatever you actually spend in Yr 2 becomes your input into the Yr 3 calculations.

An example

Let’s start with a $1M portfolio, a withdrawal rate of 4% pa, a 5% ceiling and a 2.5% floor.

• We’ll ignore inflation for now.

1. In Year 1, we spend 4% of $1M, or $40K.
   • During the year, the portfolio increases by 10% ($100K).
   • So our year end balance is $1,060K.
2. Our Yr 2 planned spend is 4% of $1.06M = $42.4K.
   • But the ceiling is $40K * 1.05 = £42K.
   • So we can only spend the ceiling.
   • For the record, the floor was $40K * 0.975 = $39K.
3. During Yr 2, the portfolio goes up by 5% ($53K)
   • The balance at the start of Yr 3 is 1.060 – 0.042 + 0.053 = $1.071M
4. Our Yr 3 planned spend is 4% of $1.071M = $42.84K
   • The ceiling is $42K * 1.05 = $44.1K, which is higher than our planned spend.
   • The floor is $42K * 0.975 = $40.95K, which is lower than our planned spend.
   • So in Yr 3 we go with our planned spend of $42.84K.

The chart above works through the same example with 3% pa inflation.

The results

The best way to think of Dynamic Spending is as a point on a continuum between the 4% rule (dollar plus inflation) and a fixed percentage (of portfolio) withdrawal.

• There’s no right or wrong answer, it’s all about what suits your situation.

That said, let’s take a look at a few of the effects.

As you might expect, success rates (certainty levels) are in between those from dollar plus inflation, and from percentage of portfolio.

• Note that these success rates are generally quite low.
• Vanguard used a (too-high) withdrawal rate of 5% for this example, presumably to enhance the contrast between the bars.

Here’s a UK version of the same chart. The assumptions here are:

• 10,000 “VCMM” simulations (I assume this is Vanguard’s version of a Monte Carlo analysis).
• 50% stocks (40% UK, 60% International) and 50% bonds (40% UK, 60% international).
• Initial withdrawal rate of 5% (too high for me), ceiling of 5% and floor of -2.5%.
• Time horizon 30 years.

The more detailed chart above also shows:

1. IRRs (broadly similar)
2. End balances (similar medians, falling averages and volatility), and
3. Spending multipliers (similar medians and averages, rising ranges).

Floor and ceiling sensitivity

Vanguard found the results were more sensitive to the size of the floor than to that of the ceiling.

• The effects of a larger floor were felt until it reached around 4.5%.
• But the effects of a larger ceiling were much smaller.

Target withdrawal rates

As noted above, the Vanguard paper uses assumptions that produce an 85% success rate for Dynamic Spending.

• It would be much better to use a 95% success rate.

The table above of suitable withdrawal rates also targets an 85% success rate.

• This give the false impression that high withdrawal rates are suitable.

I certainly won’t be using a withdrawal rate of 5% pa or more to target 30-year portfolio survival.

• My starting point is 3.3% pa – a third less.
• Which should cover me just in case I last another 50 years.

Conclusions

Dynamic Spending is an interesting third option for withdrawal rates alongside the 4% rule (dollar plus inflation) and the fixed percentage rule.

• It has pluses and  minuses, and only you can decide whether it best suits your situation.

There are also other spending rules that we will look at in future posts.

• And other key variables in drawdown, which we will cover in the remaining two articles about SORC 18.

Until next time.

Share this with Twitter, Google+, Pinterest, LinkedIn, Tumblr, Reddit and StumbleUpon.

1. For those of us who have stopped work [↩]