Should I switch some of my pension fund to cash as I approach retirement?

[NoRegretsCoyote]

The purpose of my paper was to make those high long-term returns available to all through an auto-enrolment scheme invested entirely in equities from cradle to grave, with returns smoothed to protect members from the risks of short-term (or not-so-short-term) equity underperformance.

Most pension funds are in the low hundreds of thousands at retirement. It's probably personally optimal to shift toward bonds at this stage, but not socially optimal. I agree that government should have a greater role in co-ordination here to keep pension holders more invested in equities. I am not sure that any government is big enough to absorb equity underperformance though over a sustained period though.

Much depends on your overall position of course. Suppose at retirement you have a mortgage-free house, kids through college, a full state pension, a partial DB pension, and opportunities to work part time. Then you can afford to take risks with your financial wealth.

Keeping costs as low as possible has also been an important element in the mix.

This is helpful of course, but self management isn't an option for most people, and will never be.

 

[Colm Fagan]

I am not sure that any government is big enough to absorb equity underperformance though over a sustained period though.

The paper mentioned above studied 120,000 simulated years of future experience (plus many years of actual past experience, including the horrific Japanese experience since 1990) and proved resilient in all circumstances. The government would never be on the hook. It would play the same role as the UK government does for its NEST auto-enrolment scheme, which was was set up by government but is run completely independently of government, by a Board of trustees. Some of those 120,000 years of future experience included some pretty frightening bad patches, including at least one where the return on cash flows over an entire 60-year period was negative, and the smoothing formula worked throughout, without any risk of the scheme becoming insolvent. Of course, results for contributors in that simulation weren't pretty. As I wrote, the smoothing formula cannot prevent a train crash; however, it can ease the pain.
Therefore, in summary, under the proposed smoothed equity approach to auto-enrolment, any government is big enough to absorb equity underperformance over a sustained period. Neither the Society of Actuaries in Ireland nor the Institute and Faculty of Actuaries, with whom the paper has also been discussed, disagree with that conclusion.

 

[Sarenco]

Everything I've read is completely artificial. For example, it assumes that everyone cashes their investments (to obtain an income) once a year rather than once a month as I do (or even once a quarter). Secondly, it completely ignores dividends and assumes investments are cashed every time. Thirdly, it assumes that retirees withdraw a specified amount each year rather than doing what is natural to all of us, of pulling in our horns a bit (giving a bit less to the grandchildren, etc.) if times are tough and splurging a bit if times are good.

There is nothing remotely artificial about sequence risk.

If you are drawing from a portfolio that suffers negative returns in the early years of your retirement, then the extent to which you can benefit from any subsequent market recovery will be limited by the fact that you have already spent a portion of the portfolio. That should be obvious.

Dividends are an important source of return but they are not magical. Failing to reinvest dividends is precisely the same thing economically as selling shares that produced those dividends.

You can't avoid sequence risk by simply increasing the frequency of drawdowns from a portfolio. Drawing less from a portfolio during down years (or months) is certainly one way of mitigating sequence risk. Another is to hold uncorrelated assets (bonds) to reduce the variability of returns on the overall portfolio.

You were extremely lucky to retire near the start of (what turned out to be) the longest bull market in history. Others may not be so fortunate.

 

[Brendan Burgess]

To keep the maths simple I’m going to assume a 100% equity allocation and a €1m pension fund.

back in 2008 global equities dropped 40% and in March last year almost the same

so that’s a €400k downside from a higher risk strategy

what’s the upside

lets assume a 10% pa return

so you are adding at a rate of €100kpa but losing at a rate of 4 times that.

Marc

That really is a terrible example and would mislead a lot of people.

It sounds as if there is a choice between + 100 or -400.

And is just wrong. The -400 occurred twice over a short period. The +100 in your example, is an annual figure.

Brendan

 

[Brendan Burgess]

I think most of us would agree that a person investing for the long term should have most of their assets in equities or property?

So if someone reaches retirement but will not be drawing down from their pension fund, then they should be almost fully invested in equities.

 

[Brendan Burgess]

So the question is whether someone who is drawing down their pension and spending it should have some of their funds in cash?

If they are just moving 4% a year from their ARF into their own direct investments, then it should make no difference to their investment strategy. It should remain in equities.

 

[Brendan Burgess]

But, most people will be actually running down their pension fund and spending it.

In other words, their annual expenditure will exceed their income.

They face two uncertainties - remaining years to live and investment returns.

I don't think that there is a correct answer to the question: Should they switch to cash?

I think that a mathematical model which purports to answer the question is likely to give a false sense of confidence in the answer.

Switching to cash is certainly not de-risking. Cash is at much greater risk from inflation than equities.

Brendan

 

[Brendan Burgess]

It seems to me that the primary reason to de-risk a portfolio is to reduce the variability of returns on that portfolio. That becomes particularly important when you start spending down your portfolio in order to help preserve the longevity of the portfolio.

If you get a particularly bad run of returns in the early years of retirement, you could well find yourself running out of money later in life. The extent to which you can benefit from any subsequent market recovery will be limited by the fact that you have already spent a significant portion of your portfolio.

This describes the dilemma very well.

But I am not sure that bonds will extend the real value of my portfolio. It may extend the nominal value, but that should not be a concern for me.

The only way to really extend the longevity of my portfolio is to cut my expenditure.

Brendan

 

[Brendan Burgess]

If you need it in 5 years time (which you do), I'd say move a portion each year into low risk.

Decide what fraction you want in level 6 equities in retirement say 30%

Each year between now and age 60 move 10% to low risk

But unless he is buying a holiday home , he won't need it in 5 years.

Taking out the tax-free lump sum in 5 years does not constitute "needing" it in 5 years.

The tax-free lump sum is not relevant. If the market is down 40% when he takes out the tax-free lump sum, well he will be investing the lump-sum in directly own equities at a lower price. He is just changing the label on the investments from "pension fund" to "owned by Spud".

Brendan

 

[Brendan Burgess]

The question of whether the market was elevated or depressed at my retirement date didn't really bother me, as it meant transferring equities from a tax-exempt account to a taxable one. It didn't matter much what their market value was at the time, as it was essentially an in specie transfer from one account to another.

That is what I am trying to say.

 

[Brendan Burgess]

one might be willing to assume a high degree of investment risk (risk tolerance) but objectively what matters most and certainly from a regulatory perspective is an objective assessment of one’s ability to bear losses (risk capacity) many people will find with the benefit of hindsight that buying an annuity at outset may well have been a better decision.

Yes, with the benefit of hindsight.

But with annuity rates so low at the moment it just seems very wrong to buy an annuity.

You may well have an income 30 years, but it might be worthless due to inflation.

Marc - this is the second time you have recommended annuities recently. I am surprised. Would you do a separate post on this issue? I would like to see your reasoning.

Brendan

 

[Brendan Burgess]

Drawing less from a portfolio during down years (or months) is certainly one way of mitigating sequence risk. Another is to hold uncorrelated assets (bonds) to reduce the variability of returns on the overall portfolio.

Agreed that drawing less from a portfolio extends the life of the fund.

Reducing the variability of the returns should not be an objective. Because people can't really measure risk, they use variability as a measure of risk and it is not a measure of long-term risk.

So holding bonds, might reduce variability. It reduces the risk that my fund will be worth less in a year's time. But it increases the long-term risk as the bonds are at greater long-term risk than equities due to inflation.

Brendan

 

[Sarenco]

So holding bonds, might reduce variability. It reduces the risk that my fund will be worth less in a year's time. But it increases the long-term risk as the bonds are at greater long-term risk than equities due to inflation.

Hi Brendan

Maybe a stylised example would help to demonstrate the impact of sequencing and how adding a slug of bonds to a portfolio can help.

Take a retirement portfolio of €1m that was withdrawn and spent over a 30 year retirement, at a fixed rate of €40k per annum.

Imagine over that 30-year period the portfolio had 29 years where the portfolio had a positive return of 5% and one terrible year where the portfolio dropped in value by 30%.

If the bad year happened at the start of the withdrawal period (so, 1 year down, then 29 years up), the remaining portfolio at the end of year 30 would be €223,741.

However, if the bad year happened at the end of the withdrawal period (29 years up, then 1 year down), the remaining portfolio at the end of year 30 would be €1,026,529.

Big difference!

Note that if there were no withdrawals, both sequences would produce identical returns over the 30-year period - sequence risk only arises where withdrawals are being made from a portfolio with variable returns.

Now, we obviously cannot control the sequence of returns on a portfolio – that is simply a question of luck.

However, we can add uncorrelated assets (bonds) to a portfolio of equities to reduce the severity of any draw-down in the early years of the withdrawal period.

Adding bonds to an equity portfolio will always reduce the expected (but in no way guaranteed) return on the portfolio. However, that might be a cost that an investor is willing to bear in order to mitigate (but not entirely remove) the sequence risk described above.

 

[Spud50]

thanks for the informative discussion ,one of the concerns i have is with drawing down 4% from the ARF , paying tax/stamp , AMC , brokers trailing fee etc., if i dont invest my pension Longterm in Global equities (s&p500) to counter the above secondary drawdowns then the sum will prob run out anyway , so to me there are risks on both sides ,

 

[Brendan Burgess]

If the bad year happened at the start of the withdrawal period (so, 1 year down, then 29 years up), the remaining portfolio at the end of year 30 would be €223,741.

OK, so I put 30% of the portfolio in bonds each year yielding 2%.

There was a 30% fall in the first year in the stockmarket as before.

The portfolio at the end of year 30 is €271k ( You might check my numbers.)

So sure it cushions the blow where you are unlucky enough to get the fall immediately you stop adding to your pension fund.

But if you had not invested in funds and the fall was at the end, you would have €1m

I think that this stylised example doesn't support the case for bonds.

 

[Brendan Burgess]

The case for bonds would be something like the following :

A sustained fall in the stockmarket over the first ten years of retirement.

Brendan

 

[Marc]

COMPOUND VS AVERAGE RETURNS

Imagine an investor that earns 10% on her portfolio in one year, and then loses 10% the next. The common mistake is to think that the investor would now be back to where she started. After all, the average of the two annual returns of +10% and-10% is simply 0%.

In actual fact, our investor would have lost 1% over the two years!

That’s an annual loss of about 0.5%

And the effect would have been even worse for more extreme movements. If the investor had instead gained 15% and then lost 15%, the net loss over the two years would have been 2.25%. 20% up and 20% down and the loss would have ballooned to 4% over two years.


VOLATILITY DRAG

This unfortunate effect is due to the fact that compound annual returns are always below average annual returns.

Mathematicians, using a branch of mathematics called stochastic calculus, have come up with an approximation for this effect:
g=µ-σ2/2


where g is the compound (or geometric) return of the investment, µ is its (arithmetic) average return, σ its volatility and thus σ2 its variance.

The difference between compound return and average return,approximated by half the variance (σ2/2), is known as volatility drag.


There is another generally more accurate but slightly more complicated approximation which we use in practice: g = (1+µ)(1+σ2(1+µ)-2)-1/2–1.

This particular approximation generally results in a higher (and more accurate) estimate of g than the arithmetic average minus half the variance and is exact if returns are distributed lognormally.

As required, it will also ensure the estimate of g is below µ. See: On the Relationship between Arithmetic
and Geometric Returns; Dimitry Mindlin, MAAA, PhD; CDI Advisors LLC; 14.08.2011

"We agree that Mandelbrot is right. As we can see when looking at the daily market returns, the distribution is fat-tailed relative to the normal distribution. In other words, extreme returns occur much more often than would be expected if returns were normal.

However most of what we do in terms of portfolio theory and models of risk and expected return works for Mandelbrot's stable distribution class, as well as for the normal distribution.

Our conclusion is that for long-term passive investors, the short term distribution doesn’t matter beyond being aware that outlier returns are more common than would be expected if return distributions were normal.”

Eugene F Fama

 

[Marc]

Rebalancing

The benefits of portfolio re-balancing derive from controlling risk exposure which ultimately benefits investors composure (less likely to sell during a crash) and not from an increase in the expected returns.

The proposition that rebalancing can increase the expected return of a portfolio is dubious. One thing is certain: rebalancing entails costs and costs reduce expected rates of return.

For rebalancing to increase expected returns over time, asset prices would have to be consistently mean reverting and clients would need to be able to accurately time their rebalancing decisions.

A portfolio’s asset allocation determines the portfolio’s risk and return characteristics.

Over time, as different asset classes produce different returns, the portfolio’s asset allocation changes. To recapture the portfolio’s original risk and return characteristics, the portfolio must be rebalanced to its original asset allocation.

 

[Marc]

Modern Portfolio Theory has four basic premises:

1) Investors are inherently loss averse. Investors are generally more concerned with loss than they are with reward.

2) Securities markets are competitive and drawn to a long run state of equilibrium.

This means that investors should assume that the price of any publicly traded security reflects the views and opinions of all market participants and is therefore probably a “fair price”.

3) The focus of attention should be shifted away from individual securities analysis to consideration of a portfolio as a whole, predicated on the explicit risk/reward parameters and on the total portfolio objectives. The efficient allocation of capital in a portfolio to specific asset classes is far more important than selecting the individual investments.

4) For every risk level, there is an optimal combination of asset classes that will maximise returns. Portfolio diversification is not so much a function of how many individual stocks or bonds are involved, as it is of the relationship of each asset to each other asset.

 

[Marc]

Why own Bonds?


Fixed-Income Investments

As the long-term returns figures show, an all-equity portfolio has attractive growth potential, but significant uncertainty about the exact outcome. For this reason, we describe an all-equity portfolio as being aggressive. It is most suitable for investors who are willing and need to take substantial risk in the pursuit of reward.

Investors with shorter investment horizons, a high level of risk aversion or less need to take risk should maintain portfolios that are significantly less aggressive than the all-equity strategy.

For these investors, some portion of the portfolio should remain in fixed-income instruments. Bonds provide income and help reduce the overall risk in a portfolio.

However, because of the fixed nature of the income stream from a bond, there is
comparatively little upside potential in a bond portfolio. Investors are sometimes
surprised to learn that bond prices can rise and fall with changes in interest rates, but the main source of investment returns from bonds are the interest payments they make.

A portion of an ARF portfolio’s assets should be invested in fixed-income investments.

Fixed-income investments will help reduce the overall level of risk in your portfolio, because fixed-income investments tend to be less risky than equities, and because the fixed-
income investments represent an additional diversification of your assets.

Fixed-income instruments should be used to reduce the overall level of risk to your comfort level. It is important to note that over the long term, fixed-income investments will typically have returns approximate to inflation.

in fact short-term fixed interest has been described as an inflation-hedge strategy and can work better than equites as an inflation hedge under certain economic conditions.

equally investors can purchase inflation-protected bonds which provide another source of inflation protection albeit at a price.

What Drives Bond Yields?

In this overview of the various factors that influence government bond yields, we show that both in theory and in the data, non-monetary policy factors drive significant variation in yields, particularly at longer maturities. Despite the exceptionally low yield environments we have witnessed...

www.aqr.com

Some takeaways
- bond risk remains two-sided
- bonds still provide diversification
- a fundamental approach to investing in bonds still makes sense


The bottom line: Building a prudent portfolio requires careful consideration of the unique characteristics of both equities and fixed income and what each can add to the portfolio