Cash Buffers, Sustainable Withdrawal and Bear Markets
The practice involves holding between one to five years of income as cash within a retirement portfolio to mitigate sequence of return risk. The rationale is very simple — since a typical bear market lasts around two years, holding an ample amount of cash means that you can avoid selling down equities at the worst possible times. To be clear, retirement cash buffer is a different concept to holding emergency fund - typically 3-6 months of expenses in event of ehrr… ehrr… unforeseen circumstances that life often throws at us.
There are claims of behavioural benefits associated with keeping a retirement cash buffer. It’s essentially a form of the bucketing strategy which is consistent with mental accounting — i.e. our natural tendency to keep separate pots/accounts for different purposes or to spend at different times. I want to focus on the practice of keeping a cash buffer in this article. (I’m working on a wider piece on retirement buckets in general).
The specific behavioural benefits of cash buffers often touted by advisers is that clients are far less likely to panic during bear markets. While I generally tend to agree with this view, I’ve not found any peer-reviewed study to corroborate the claim that clients with a cash buffer sleep better at night, or panic less than those without.
There is however ample evidence that a cash buffer does nothing to make retirement portfolios last longer in bear markets. Indeed, using cash buffers in a way that meaningfully reduces overall equity allocation in retirement portfolios can actually reduce the sustainability of the portfolio.
- Woerheide & Nanigian (2011): Sustainable Withdrawal Rates: The Historical Evidence on Buffer Zone Strategies
In this comprehensive study, Dr Walter Woerheide and Dr David Nanigian tested the success rate of one, two, three and four years of cash reserve versus fully invested portfolios, across several asset allocations and withdrawal rates. In the cash buffer portfolios, they assumed that withdrawals were taken from the portfolios when return was positive and from the cash reserve when the portfolio was down.
The results show that the fully invested portfolios produced better outcomes than corresponding cash reserve portfolios in about 80% of scenarios. In the other 20% of scenarios, the cash buffer approach fared no better than the fully invested portfolios.
- Michael Kitces (2012): Are Cash Reserve Bucket Strategies For Retirement Really Necessary?
- Abraham Okusanya (2017): Cash reserve buffers, withdrawal rates and old wives’ fables for retirement portfolios
This study by yours truly examined 11 methods of draining a retirement portfolio, including cash buffer and rebalancing strategies using success rate, portfolio longevity and legacy to evaluate their effectiveness. The conclusion is that a cash buffer does not actually mitigate sequence risk in retirement portfolios i.e. they offer no improvement whatsoever compared to a fully invested portfolio. Indeed, cash buffers typically lead to a reduction in the overall equity allocation, which ends up reducing the portfolio longevity.
…across all the strategies, replacing equity with cash offers no improvement in the success rate or portfolio longevity! Crucially, the success rate is meaningfully lower for 40-year retirement periods, and it resulted in a significant reduction in the size of the legacy at the end of 30 years in the non-failure scenarios!
- Javier Estrada (2018): The Bucket Approach for Retirement: A Suboptimal Behavioral Trick?
Prof Estrada at IESE Business School, a practising financial adviser, examined 3 different approaches to cash buffer and bucketing methods using comprehensive evidence from 115 years of capital market data spanning 21 countries.
Ultimately, the evidence shows that a bucket approach underperforms static strategies, and it does so based on four different ways of assessing performance. For this reason, however plausible, comforting, consistent with mental accounting, and easy to implement the bucket approach may be, simple static strategies, which call for periodic rebalancing and are just as easy to implement, would make retirees better off.
Estrada made an important point about why cash buffer and bucketing strategies don’t work. That is, while they tend to avoid selling equities in bad years, they often miss out on rebalancing which buys equities when prices are down. He notes further…
‘Most bucket strategies, and certainly those that park in bills a fixed number of annual withdrawals, avoid selling underperforming assets but refrain from buying more of those assets when their price goes down. Static strategies, which by definition imply periodic rebalancing, actively sell assets that have become relatively more expensive, and critically, also buy assets that have become relatively cheaper, thus enhancing performance.
Good Buffer, Bad Buffer, No Buffer
Regardless of the compelling evidence to the contrary, many advisers insist on holding cash buffers in retirement portfolios. Even if this increases the risk of running out of money faster in retirement.
Below, I use Timelineapp to illustrate how advisers might hold a cash buffer without having a negative impact on the sustainability of the client portfolio.
Let’s consider three strategies for £1m invested in a 60⁄40 global equity/bond portfolio, with withdrawal of £40k a year, adjusted annually for inflation, net of 1% in fees and charges.
- No Buffer: this is a 60⁄40 which is rebalanced whenever equity drifts by more than 10% up or down. Withdrawals are made evenly across the entire portfolio.
- Bad Buffer: this portfolio works on the basis that around two years of withdrawals are kept in cash — i.e. £100k, with the remaining £900k invested in a 60⁄40 portfolio. The upshot is that we end up with a 54/36/10 in global equities, bonds and cash. Withdrawal is only made from cash and never from the equities and bonds. The cash allocation is replenished only when it dips below 5% or rises above 11% of the entire portfolio, at which point we rebalance the portfolio and top our cash allocation back to 10%.
- Good Buffer: this portfolio consists of a 60/30/10 in global equity, bond and cash. Withdrawals are only ever taken from cash, which is replenished if the cash allocation drops below 2%. The portfolio is rebalanced if the equity allocation is more than 10% up or down from the original allocation. This approach draws on earlier research (Okusanya, 2018 referenced above) which shows that implementing a cash buffer by reducing ONLY the bond allocation is a more effective strategy.
Using empirical data we run all 30-year rolling periods starting between Jan 1926 and Dec 2019. The idea is that we can get a sense of how each strategy fares in retirement periods that includes bear markets. The period under examination included 10 bear markets, the most severe being the bear market of 73⁄74 when UK Equity declined a whopping 67% and took two and half years to recover.
To measure the effectiveness of each strategy, we look at four important metrics:
- Success Rate: this gives measures the chances of success under a wide range of historical market conditions
- Portfolio Longevity (10th percentile): this gives us a sense of how long the portfolio lasts in a bad scenario with poor return sequence. The logic of using the 10th percentile scenario is to look out how the strategy fared in the very poor scenarios with bad sequence of return
- Lifetime Withdrawal (10th percentile): this is the sum total of all withdrawals made from the portfolio, adjusted for inflation, in the 10th percentile scenario
- Legacy (50th percentile): here, we’re looking at the outstanding portfolio value, adjusted for inflation, at the end of the 30-year retirement. But rather than looking at the 10th percentile scenario, we consider the median scenario. Again, the idea here is to get a sense of how the strategy might fare in benign market conditions where the portfolio doesn’t necessarily run out.
Below we present the results for the three strategies, based on the afore mentioned metrics.
The results show that in scenarios with bad sequence of return (i.e. 10th percentile), a cash buffer is no better than the fully invested portfolio. Indeed, it results in poor outcomes, with the money running out about two years sooner.
Furthermore, a cash buffer hurts significantly in good scenarios, resulting in around £170k (inflation-adjusted) or 38% less in legacy in the median scenario, compared to the fully invested portfolio.
In essence, a cash buffer strategy is a heads you lose, tails you lose. It doesn’t mitigate sequence risk in retirement portfolios. And in scenarios with good sequence of returns, it leaves significant shortfall in legacy.
We can frame a cash buffer as a way to help clients sleep better at night, but should we neglect to tell them the price? The price tag is running out of money two years sooner, if market conditions are really bad. If market conditions are good, the price tag is 17% of their initial capital in real terms ( i.e. £170k for an initial capital of £1m). How you frame it is up to you. But the point is, everything has a price, including sleeping better in a two year bear market.
An alternative to the traditional cash buffer is the Good Buffer approach. This essentially keeps the overall equity allocation at the same level as the fully invested portfolio, by reducing the allocation to bonds. In essence, rather than a 60⁄40 portfolio, you end up with a 60/30/10. Withdrawals are always made from the cash, but with the key feature of preserving the allocation to equities.
Asset Allocation Glidepath
To understand why the Good Buffer strategy tends to work better, let’ss examine what’s going on behind the scenes. The image below illustrates the asset allocation glidepath for the 10th percentile scenario for the three strategies. Remember, we’ve modelled nearly a thousand scenarios; the purpose of looking at the 10th percentile is to get a sense of what’s going on behind the scenes in a scenario with poor sequence of return.
This glidepath for the No Buffer strategies shows that, by making withdrawal across the entire portfolio, you keep your equity allocation at around 60% mark, and have to rebalance less frequently — in other words, withdrawals serve as a means of rebalancing by selling more bonds and fewer equities in bad times.
The Bad Buffer approach has one big shortcoming, equity allocation mostly stayed below the 60% intended. This is a direct by-product of the need to maintain a minimum of a one-year cash buffer at all times, but ideally this needs to be topped up to over two years of a buffer. But as a retiree spends down the cash at the end of each year, you have to top it back up. In effect, this achieves no improvement over and above the No Buffer approach. This is because you constantly have to top up the cash buffer, since you don’t know whether a bear market is around the corner. But again, the cost of keeping all this cash is lower allocation to equities.
The Good Buffer on the other hand does something remarkable — keeping equity allocation mostly around the 60% mark. Reducing the bond allocation to allow for the cash buffer means that the equities can be left more or less intact. Of course, we don’t let this get out of control — hence why we set a 10% drift limit.
The Cost of Behaving Badly
Many advisers justify cash buffers on the grounds of its behavioural benefit, even if it damages retirement outcomes meaningfully in both good and bad scenarios.
Folks like me who crunch data are often accused of ignoring a client’s behavioural tendencies. ‘Advisers manage humans not robots’ I am often told. Apparently, data crunchers aren’t really human, so they wouldn’t understand the agony that clients go through in bear markets. Their love for data and rigour means that they’re just too emotionally inept to understand the natural human tendencies for clients to blow up their financial plans when things get rough.
Nothing could be further from the truth. I have always thought of a key role of a financial planner as being a behavioural coach to clients. There is no denying the fact that advisers play a crucial role in stopping clients from being their own worst enemy. The question is, given the real harm that a cash buffer does to a client’s retirement, is it really is an effective device for managing clients’ behavioural foibles or a manifestation of the adviser/client behavioural bias?
Cash buffers are a form of bucketing strategy and are consistent with known behavioural biases such as mental accounting. So I wonder, is a cash buffer merely a way for an adviser to confirm (conform to) their client’s behavioural tendency, or is it really an effective device for dealing with a client’s behavioural foibles?
I am no behavioural finance expert, so I put this to behavioural expert Dr Greg B Davies, who is founder of Centapse, a firm that specialises in applied decision science and behavioural finance, turning academic insight into practical applications. He told me that:
‘It’s both. But in most cases the behavioural costs outweigh the benefits. Not true for a few special case buckets, but these are the exception. All buckets, silos and accounts, whether mental, financial, or organisational are essentially coping mechanisms for complexity … as such they have benefits, but always add costly frictions and inefficiencies as a side effect’
Dr Davies added that a cash buffer is essentially ‘a coping mechanism for complexity. The problem comes when we start to treat a good way of explaining as the structural solution itself.’
All of this leads me to ask ‘is the job of a financial adviser as a behavioural coach to reinforce a client’s behavioural foibles, especially when it hurts them in the long term? Or is it to negate the behavioural tendencies which risk hurting a client’s long term interest, by finding better ways to frame (or implement) a strategy that would normally make clients uncomfortable?
And where we can’t find a structural solution that adequately negates a damaging client behavioural bias, do we pander to these behavioural tendencies, even if they hurt their financial futures?
My point is, if the purpose of the cash buffer is to avoid selling equities in bad years — well, you might win that battle. But, it may be at the cost of losing the war — making money last longer in retirement, particularly in bad scenarios.
Get the latest blogs sent straight to your inbox by subscribing to our mailing list.