# Safe Withdrawal Rates – ERN #2

Today’s post is our second visit to the mammoth series of posts on Safe Withdrawal Rates at the ERN blog.

###### The State Pension

Last time we worked out that the 4% rule should really be the 3% rule.

- Which means that rather than retire when you’ve saved 25 times your living expenses, you really need to wait until you have 33 times your expenses.

The fourth post in ERN’s series is about the State Pension, or Social Security (SS) as they call it in the US.

- Because the State Pension means that you can cut back on withdrawals once you have it, you can also increase withdrawals in the period leading up to it.

This is a similar argument to my approach to ISAs and SIPPs – even if you want to retire before 55 (when you can’t access a SIPP), you can run down your ISA pot(s) until you hit 55.

Here are chart showing what ERN means:

ERN is using a 3.25% baseline, but I reduced this to 3% pa to account for the very low investing costs (0.05%) that ERN uses.

ERN is older than Mrs. ERN, and qualifies for SS at age 62 (lucky him – I have to wait until age 67 over here in the UK).

- He plans to defer for a few years to let the payment build up (note that this is no longer a good idea in the UK).
- When ERN does take the SS, it will amount to 1% pa of his starting pot.
- Note that ERN will have been retired at this point for 25 years, with another 35 years of their planned 60-year retirement to go.

The basic maths of 1% * 35 / 60 = 0.58% pa, but the present value of the SS at the end of retirement is much lower.

- ERN also points out that front-loading the higher withdrawals (before SS kicks in) will increase the Sequence of Returns Risk (SoRR).

When ERN re-runs his simulations to include a 1% cut in drawdown from the age at which he will take SS, the median increase in the SWR is only 0.18%.

- The range is between 0.12% and 0.25%.

If my other half (OH) and myself were using a 3% SWR then our combined UK state pensions would be around 1% of our starting pot.

- We actually have a lot more money saved than that, so our SWR is lower.
- Note also that we capitalised things like the State Pension, to encourage a more aggressive asset allocation strategy.
- If we were targeting 3% SWR, I would need to reverse out this capitalisation to avoid double-counting.

The US state pension sounds as if it’s twice as generous as ours.

- That might not be so far from the truth, since the UK has the lowest state pension (as a percentage of average salaries) in the whole OECD.

Our average retirement period between the two of us is around 32 years.

- And we receive the State Pension after an average of 12 years.
- Leaving an average of 20 years to receive it.

That makes our back of an envelope baseline contribution from the State Pension 1% * 20 / 32 = 0.63% pa.

- By comparison with ERN’s, the real contribution is more like 0.19%.

In practice, I’ll stick with (less than) 3% pa withdrawals from the grossed-up pot (including capitalisations).

ERN has a table of how the SS impacts vary by size of SS income, when you get it, and today’s CAPE value.

- But they are all pretty small (well, the 1% level numbers are, and that’s all you’ll get in the UK).
- And they refer only to 100% equity portfolios used towards a 60-year retirement with a 50 final asset value target.

He also provides a table of failure rates vs withdrawal rates under two CAPE regimes.

If you’re unlucky and face adverse capital market returns early on in retirement and you keep withdrawing your initial rate then your portfolio will be so compromised by the time you reach your Social Security age that it won’t make much of a difference.

ERN says he would risk a 3.5% withdrawal rate, but I’ll stick with 3% (using higher investment costs than ERN).

There are more tables in the post, but nothing for below a 50-year retirement.

###### Cost of living adjustments

In post number five in the series, ERN looks at cost of living adjustments (COLA).

- His baseline assumes that withdrawals increase in line with CPI each year.

But what if he used a CPI-x% inflator instead?

- This would help the portfolio to cope with bad periods in the stock market.
- And it might better reflect the tendency for people to spend less as they get older.

As the chart shows, a lower COLA means a higher starting withdrawal rate.

ERN mentions that the original Trinity Study included simulations with no inflation increases at all.

The high inflation of the 1970s (and ERNs 60-retirement horizon) mean that he isn’t interested in these.

- $40K in 1970 would only have been worth $8K by 2000.

Deflation in the 1930s would have had the opposite effect, so ERN doesn’t see much point in calculating averages across such different scenarios.

The table shows how much COLA can shrink your purchasing power.

- Using my own retirement horizon of 35 years, even a 0.5% pa reduction erodes the value of £100 to $84.
- CPI minus 1.5% leads to a final purchasing power of just £59.

So I can only really look at reductions of up to 1% in COLA (terminal purchasing power of £70).

- With his 60-year retirement, ERN will only consider 0.5% pa ($74).
- Luckily, he also includes my 1% pa reduction in the next stage of the analysis.

He adds two more scenarios:

- 5Y no COLA – this is the often quoted solution to a bad sequence of stock market returns
- You shrink the real withdrawals by 2% pa for 5 years, then return to using CPI.
- From that point, real withdrawals will be 9.6% lower than in year 1.

- 5Y transition – this is the “gradual retirement” option
- You keep an alternative income source and withdraw only 75% of your target in year 1.
- You scale up by 5% pa to reach 100% in year 5.

Here’s how the five strategies look, going out to 60 years:

I’m closest to the 5Y transition strategy, as my OH has carried on working (with some interruptions) since I retired six years ago.

But other strategies are of interest:

- I suspect that the CPI – 0.5% strategy comes closest to retirees’ actual spending patterns.
- And I would consider 5Y no COLA if we had a bad run of stock market returns.

In practice, the picture is more complicated:

- I keep a lot cash on hand to cope with market downturns.
- Tax rules mean that I over-withdraw from my (and soon, our) pensions, then find a home for the money that we don’t spend each year.

What do ERN’s simulations show?

The table above shows the median SWR increase (100% equities, 60-year retirement) as with Social Security further above.

The unpleasant fact is that the COLA-x% works best when we need it the least. That makes perfect sense: we have such a front- loaded consumption pattern that we get hit by the dreaded Sequence of Return Risk (SoRR).

The takeaway for me is that COLA is no free lunch.

- It might make sense to use one of these alternative strategies as events unfold (and indeed, CPI – 0.5% might be a good fit for a natural spending pattern in retirement).

But COLA doesn’t mean that we can turn our new 3% rule back into the old 4% rule.

- We could call it the 3.25% or the 3.5% rule, though.

The next table shows the failure rates for the COLA strategies.

- These are all 100% equities, 50% final pot target, 60-year retirement.
- ERN uses all equities because that minimises failure rates.

Here’s a table of failure rates by equity weight for a 3.5% SWR.

- 60% equities is good, but 80% looks better.

With my starting SWR of 3% pa, there’s not much for me to worry about with COLA strategies.

- ERN interprets the tables as him being able to push his 3.25% SWR out to 3.5% for most strategies and to 3.75% for CPI -1%.
- Using the same logic I reach 3.25% / 3.5%.

ERN doesn’t plan to use COLA.

- He believes that people spend less as they age, but isn’t sure whether they want to or they have to.
- Most people have less than they would like in retirement.
- ERN is starting out frugal (backpacking) and is more likely to increase his spending (cruises).

I won’t use COLA either, for a different reason.

- Our spending is so modest (for our more advanced age and larger pot size) that we are quite unlikely to run out of money.

So a “complicated” approach to boost our SWR by 0.5% pa doesn’t have much appeal.

- As we have fewer years of retirement left to fund, I too am likely to increase spending rather than cut it.

But if we were closer to the grey area of being financially independent, COLA strategies would definitely be worth considering.

###### The Potemkin Retirement Village

In the sixth post in the series, ERN looks at industry support for the 4% rule (about which he – and I – are clearly sceptical).

- ERN’s focus is obviously a 60-year retirement, but Wade Pfau has been arguing that 4% might not work for 30 years if starting from today’s high valuations (we’ve discussed this previously when looking at the work of Abraham Okusanya).

ERN also mentions some holes in the “ratcheting” strategy described by Michael Kitces.

- To survive the conditions of the 1970s, short-term rather than long-term bonds need to be used.
- And in a post on post-2000 returns, the key chart uses nominal rather than real portfolio value.
- These might mislead the unsophisticated reader.

The meat of ERN’s post looks at real returns through the dot com bust and the 2008 crisis.

- ERN was sceptical that the 4% rule could have worked through this period, since the S&P 500 only returned 2.3% pa real (total return) from 2000 through 2016.
- A 50:50 stock bond portfolio returned 3.4%, but that’s still less than 4% pa.
- Note that ERN has deducted his 0.05% ETF fees – I would need to deduct another 0.25% to 0.35%.

So we know that 4% wouldn’t work from 2000 to 2016.

- ERN runs some simulations, but they just make things look worse:

The chart uses real Equity and Bond returns until Dec 2016.

- Looking forward the chart uses the long-term average real return for Equities, and a 0.5% real return for bonds for 10Y
- Long-term averages for both are used after that.

The next chart shows the 30 year SWRs from 2000, by equity allocation.

- With a 50% final value and a 50% equity allocation, the SWR is 3.9% pa
- With an 80% equity allocation that goes down to 3.5% pa
- With full depletion and 80% equities we’re back up to 4.2% pa
- And with full depletion and 50% equities we’re up to 4.8% pa

So full depletion does allow the industry types to say that the 4% rule works.

- But with zero depletion it looks more like the 3% rule once more.
- And of course, 4% wouldn’t work for ERN’s proposed 60-year retirement.

###### Conclusions

That’s it for today – we’ve covered another three posts in ERN’s series.

- But that still leaves another 18 to go.

I’m really enjoying the series so far.

- But at the same time, I’m hoping that not all the posts are as meaty as the six we’ve looked at.

Otherwise it will take me eight articles of my own to get through them.

Let’s recap what we’ve learned today:

- The State Pension doesn’t make much difference to your SWR.
- It isn’t large enough and you don’t receive it for long enough.
- It will probably add less than 0.2% pa to your baseline SWR.

- Cost of living allowance (COLA) strategies can add 0.25% to 0.5% pa to your SWR.
- If you are close to the minimum pot size to retire (30 to 33 times annual spending), then you should consider them.

- The 4% rule has worked since 2000 only if:
- You have a 30 year retirement horizon.
- You are happy to deplete the pot.
- You didn’t choose too high an equity allocation.

I’ll be back in a few weeks with some more from ERN’s terrific series of articles.

Until next time.

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