Monday, February 10, 2014

Cushioned Retirement Investing

Most of us believe that we should reduce the riskiness of our portfolios when we retire. However, there is little agreement on exactly how to do this. One common rule of thumb is to use your age as your percentage in bonds. However, such fixed rules just don’t take into account people’s unique circumstances. I prefer a technique I call Cushioned Retirement Investing to reduce risk. This technique is based on the simple principle that you shouldn’t invest money you’ll need in the next 5 years in risky investments.

There is nothing magical about the 5-year threshold. More daring types may choose 3 years, and more conservative types may prefer 7 years. I’ll stick to the 5-year figure for this discussion.

The main idea is that you keep any money you’ll need in the next 5 years out of the main part of your portfolio. Because the main part of your portfolio only holds funds that will be there for 5+ years, it can stick to your preferred asset allocation for your entire life. If you have a 75/25 split between stocks and bonds through your working life, you can keep the same allocation in retirement, as long as you maintain 5 years of living expenses safely off to the side.

By “safely off to the side,” I mean something like a high-interest savings account (HISA), short-term government bonds, or guaranteed investments at a bank. However, any bonds or guaranteed investments have to come due before you need the money.

Making Yearly Adjustments

If you’re using cushioned retirement investing, you’d need to make adjustments to start each year. First decide how much money you’ll need over the next 5 years (taking into account inflation). Because of your spending over the past year, you’ll likely have only about 4 years of spending set aside. So, you’ll have to gather some cash from your portfolio, which will likely require some selling. Once you’ve got a full 5 years of spending set aside again, you’re set for another year.

The process of setting aside money actually starts well before retirement. When you’re 4 years away from retiring and you look at your needs over the next 5 years, you’ll need to set aside one year of spending somewhere safe. The following year, you’ll need to set aside another year of spending, and so on.

How Much Can I Spend?

Most of us have no grand plans for leaving a big inheritance and want to know how much we can spend in retirement. In a previous post I offered a spreadsheet that allows you to input information related to your situation and calculate the percentage of your portfolio you can spend each year. Keep in mind that the answers are only as good as the inputs you provide. There is also a page on the spreadsheet for doing the calculation in the other direction: figuring out how much you need to retire.

An Objection

Some might object that it makes no sense to maintain the same asset allocation into retirement. Keep in mind that it is only what I call the main portfolio (excluding the cushion) that maintains the same asset allocation into retirement. If we look at total savings including both the main portfolio and the cushion, things are different.

In the 5 years leading up to retirement, the cushion builds. This lowers the volatility of the total savings. Further, as you draw down your main portfolio in retirement (as most people will), the cushion becomes a larger percentage of total savings, which lowers volatility further over time.

Other Applications of Cushioning

Cushioned retirement investing isn’t just for people looking to maximize spending. It can apply equally well to those who intend to leave an inheritance. If you’ve got a $5 million portfolio, but only spend $80,000 per year, you could just keep aside $400,000 and invest the rest with an asset allocation suitable for the long term.

In fact, the principles behind cushioned retirement investing can apply even before retirement. For example, they can be used for RESP investing. You could maintain your preferred asset allocation throughout the life of the RESP, except that starting 5 years before your children start post-secondary education, you begin setting aside some money in safe investments. The idea is that this safe money is still within the RESP until it gets spent, but you think of it as outside the main RESP holdings that are invested with your preferred asset allocation.

We can even think of an emergency fund as an extension of the idea of a 5-year cushion. The emergency fund is money that you might need over the next 5 years in case you have unexpected expenses or lose some income.

Cushioning can also apply to saving up for a house, cottage, or car. If you plan to buy in less than 5 years, the money should be kept safe.


Overall, I prefer cushioned retirement investing to hand-wavy advice on how to adjust your asset allocation as you enter retirement. A side benefit is that the concept of cushioning is general enough to apply to many other aspects of financial life as well.


  1. Are there any rules to the yearly adjustments with this methode? Surely you would withdraw less from your investment in a market downturn and catch up once the market is back up to avoir depleting your investment? If not, the cushion isn't very useful if you keep selling low your other investments.

    1. @Chuppe: This method involves recalculating your spending level each year. So if your portfolio drops so severely that it makes sense to reduce yearly spending from $50,000 to $45,000, your cushion (which had been at $250,000 to start last year and has been drawn down to $200,000) will only need $25,000 added to it to get back to 5 years of spending. So, the cushion needs less replenishing when your portfolio is down.

      If you follow the link in the post, it will take you to further discussion of the method (as well as links to all the other posts I've written on this subject).

    2. If you're drawing down the portfolio, does that also require reducing the spending or can you differentiate between reducing spending because of market performance and keeping it the same because the decline was expected?

    3. @Richard: The spending level is calculated so that if the market performs exactly as expected, spending will stay flat (adjusted for inflation) drawing down principal so that it's gone by the target age. So, any reduction in spending comes entirely from the market underperforming, and has nothing to do with the portfolio drawing down.

  2. I've been enjoying your blog Michael and this post touches on a very timely topic. The body of withdrawal research is still pretty young so it continues to evolve. Last fall, I wrote an article about two very different papers advocating two very different withdrawal strategies. What you've described is a "goals-based" or "bucket" approach. You'll see these terms in the article.

    1. @Dan: If I understood your article correctly, what I describe as cushioning is broadly similar to bucketing, but with an important difference. Bucketing seems to involve setting aside a base-level of spending for many years in safe investments (with the idea that it not need to be replenished). My cushion is just 5 years' worth of spending and gets replenished as portfolio returns permit. No doubt each of these strategies has its merits, but they seem to be different.

    2. Bucketing is generally the same strategy that you call cushioning - i.e. start with a cash cushion from which to draw and periodically replenish the cushion. This is also basically what we do for clients.

      Both papers I wrote about had effectively a very similar starting conservative positioning. The first one (that uses 3x leverage for equity exposure) does advocate a rules-based replenishment of the cash portion. The second paper simply states that you should start with lots of bonds and cash and spend it down without replenishing the cash. But both have a lot in common with bucketing.

      To read more on this look up articles and papers by Jim Otar, a Canadian researcher and former advisor. He has done some interesting work on this topic.

    3. @Dan: I read Jim Otar's book Unveiling the Retirement Myth. He might have set a record for the number of worked examples.

      I'm finding that the cushioning or bucketing has been the easy part compared to tax considerations. For example, as I look at my own situation, if I retire before I start drawing CPP and OAS, it seems that I should be drawing heavily from my RRSPs early on to live on plus a little extra to throw into TFSAs. I still have a lot of figuring to do, but it's clear that taxes make the job of maximizing safe spending levels quite tricky.

  3. Might not whether you need a 3 or 7 year cushion depend on stock market valuations?

  4. @Anonymous: I don't believe I have any useful insight into whether stocks are over- or under-valued. Anyone interested in CAPE should take a look at Jeremy Siegel's criticisms.

  5. Good post Michael.

    Your thinking also aligns with Daryl Diamond's in his book: Your Retirement Income Blueprint, one of my favourite personal finance books by the way. The concept of bucketing comes up a number of times...although I appreciate your angle is a bit different.

    I wrote a recent article on my site, looking at options for my parents to wind down their RRSPs. I'm thinking I'm going to withdraw from my RRSP before collecting CPP and OAS. They'll likely change the rules again and I'll be 70+ before OAS kicks in anyhow.


    1. @Mark: There isn't much new under the sun, but I wanted to give my strategy with its details a name to make it easier to write about in the future.

      Timing RRSP contributions and withdrawals is a dance involving trying to hit the highest possible marginal tax rate on contributions, the lowest possible marginal tax rate on withdrawals, and letting the assets grow tax free for as long as possible. I definitely think it can make some sense to take advantage of any low-income years to make RRSP withdrawals.

    2. Mike, I like this approach, especially the automatic nature of it. With our retired clients, we use a less elegant form of the strategy, although the spending reserve, as we call it, is much smaller.

      Where our approach differs, however, is on the yearly adjustments. We too prefer for it to be automatic, but we give ourselves some wiggle room to let the reserve run down if the markets have been particularly weak (i.e. we don’t sell stocks on weakness). That, of course, requires a judgment call.

      For either method, it’s unlikely that larger portfolios would ever have to sell stocks on weakness because of interest, dividends, fund distributions and bond maturities, and if they did, it would likely involve selling fixed income to rebalance the Main Portfolio back to its correct mix.

      But a question for you: If cushioning calls for an adjustment every year, why run with such a big cushion. If money is going to come out of the Main Portfolio regardless, the benefit of the cushion is largely offset. It seems like overkill.

      Tom Bradley

    3. @Tom: I'm trying to create a fully automatic approach that uses no judgement of whether markets are due to bounce back or fall. Obviously, as active investors, Steadyhand has an opinion on whether markets are likely to outperform or underperform over the next 5 or 10 years.

      I tend to think in terms of a large cushion mainly because my personal portfolio contains no allocation to bonds or cash. I agree that more balanced portfolios may have less need for a large cushion because they already contain some cushion. Probably the best way to decide this would be to run some simulations with different sizes of cushions and different asset allocations on historical stock data. A project for another day.

    4. Thx Michael. I understand the psychology of weak markets, but I sometimes wonder if we're trying too hard to never sell a share when markets are down. There is a significant cost to holding a lot of cash (our clients and your model). Would it kill the clients returns if once in their life they were forced to sell a few shares while markets were weak? I'm sure the math (that's your dept.) would say we can afford to have it happen, but then of course, there's the behavioral side and the sleep-at-night factor.

    5. @Tom: I think the only purpose of a large cushion is to protect against a sustained large drop in stock prices (1929-1932) or a very long period of poor returns (1966-1981). If we could guarantee that large drops would be quickly followed by a rebound (2008-2009), then cushions could be much smaller, and we could safely sell a few stocks when they're down.

      In my simulations of cushioning for a 1929 retiree, we still sell some stocks during the down years -- we just sell more after prices rebound.

  6. Love it, you even worked in a figure. Now we need to animate it so the cushion is like one of those bridging tanks, constantly taking planks from the main portfolio and inching it along... actually, it is at this point that I think I should go to bed.

    Anyway, a 5-year cash cushion and a lifetime portfolio is also how I set up my retirement spreadsheet. Makes sense to me (though I hadn't yet got to the level of refinement of adding rules-based adjustments to the spending budget).

    1. @Potato: Thanks. Your animation idea is way ahead of my abilities to create graphics. From Tom Bradley's comment, I think now that a 5-year cushion may be too conservative for investors who already have a sizable allocation to bonds. I'm tempted to say that the reason for this is that they were too conservative through their working years, but I'm still mulling it over.

  7. It seems like the advantage from the cushion only comes if you reduce spending when needed. Variable spending levels are one way to do this (and they are valuable to just about everyone anyways).

    Maybe there's some other automatic rule that could work based on the peak level of the portfolio value. Siegel's note about the maximum time to get back to the starting level hints at this.

    It could vary a lot based on the asset allocation. A portfolio that was 100% in the Canadian index in 2000 might take longer to recover.

    1. @Richard: You're right that the cushion only helps if you're prepared to reduce spending when the market disappoints. I'm sure there are many possible approaches into account expectations for a market rebound. For now, I prefer not to add any discretionary judgement or expectations of reversion to the mean. I might try adding some expectation of reversion to the mean at some point as long as it's completely automated and doesn't require any human judgement.

    2. Although I don't expect to rely on an automatic rule, there might still be some that work. For example if stock sales were reduced by an amount equivalent to the decline from the peak each year that would reduce selling in a down market. Since the maximum historical recovery time is a bit longer than the buffer time you're using, this could be smoothed a bit even though the cashflow would probably be enough without that.