Monday, May 4, 2020

Calculating My Retirement Glidepath

While some people are busier than ever during the pandemic, like health care workers, many of us have extra time on our hands. I’ve used this time to clean up my plan for retirement investing and spending. Here I describe this plan.

Top Level

I’m an index investor with a portfolio invested in stock ETFs and bonds. By “bonds” I mean any type of safe fixed-income investment, including cash savings, GICs, and short-term government bonds; I have no interest in corporate bonds or long-term government bonds. At a broad level, I maintain chosen percentages of stocks and bonds. Currently, my portfolio is about 80% stocks and 20% bonds. However, I plan to increase the bond percentage over time.

When we adjust asset allocation percentages as we age, it’s called a retirement glidepath. The idea of a glidepath is far from new, but most recommended glidepath percentages seem to be just made up numbers, such as bond percentage equal to your age. I prefer to run my portfolio with a small number of fixed rules that make sense to me. Here is one of those rules:

Rule 1: Only invest in stocks with money I won’t need for 5 or more years.

This isn’t new on its own. However, I’ve used it to guide my asset allocation glidepath before and after retirement. It’s not obvious how this rule determines my asset allocation glidepath, but it does.

Before retirement, there was enough demand for my skills that I was confident in my ability to cover my family’s needs with my income. So, beyond some emergency funds to cover a short interruption in my income, I invested all savings in stocks.

Now that I’m retired, I maintain 5 years of my family’s spending in bonds, and the rest in stocks. By “family’s spending,” I mean the safe amount we can spend based on my portfolio’s size rather than some dollar amount we want to spend.

This safe spending level is something I can calculate based on a number of factors:
  • Portfolio size
  • Holding 5 years of spending in bonds
  • Making the money last until age 100 (I likely won’t live this long, but what matters is how long I might live.)
  • Conservative estimates of stock and bond returns
  • Expected future pensions such as CPP and OAS
  • Expected future large cash flows

As I get older and closer to 100 years old, my safe spending level rises as a fraction of my portfolio’s size. The fraction of my portfolio that is invested in bonds rises as I age as well, which determines my glidepath.

My goal is to plan for constant annual spending in inflation-adjusted dollars. However, if portfolio returns don’t match expectations, I adjust spending. So, my plan gives my spending levels as a percentage of my current portfolio size. Also, I’ll spend more from my portfolio leading up to the start of CPP and OAS, and then I’ll spend less from my portfolio.

I have a full paper, Withdrawal Rates for a Retirement Glidepath, that covers the math of calculating spending levels and asset allocation percentages based on Rule 1 above and the list of other factors.

This paper differs a little from a previous description of my spending plans. The older description was based on the idea that I’d recalculate my spending level and bond allocation once per year based on a set of percentages pre-calculated for each year of my retirement, much like the mandatory RRIF withdrawal percentages. However, I now have a spreadsheet that calculates my spending level and bond allocation in real-time. I now treat my bond allocation as an amount that needs rebalancing whenever it gets far enough away from its target percentage. This can happen in response to fluctuating investments or the slow draw-down of cash as I spend my money in retirement.

The main difference this change makes is that I now maintain roughly 5 years of spending in bonds at all times. My retirement spending used to draw down the 5 years in bonds to 4 years in bonds over the course of the year before I replenished the bonds at the end of the year. With the old plan, I averaged about 4.5 years of spending in bonds, but now I maintain 5 years’ worth. Luckily for me, I made this part of the change before the pandemic, so I benefited from having less money in stocks before the crash.

Unlike many investors, I don’t plan to switch to spending exclusively from bonds when stocks crash; I just rebalance mechanically. However, as I’ve explained before, the rebalancing process naturally shifts spending to bonds when stocks are down.

Although rebalancing a portfolio can produce profits, its purpose is to control risk. The reason we don’t rebalance very frequently is that trading costs can add up. This brings us to another fixed rule:

Rule 2: Limit rebalancing trading costs to 5% of rebalancing gains.

From this rule we can calculate rebalancing thresholds that limit costs but allow us to capture rebalancing gains as asset prices fluctuate. I recently updated my description of my threshold rebalancing strategy. The biggest change is an improvement to calculating rebalancing thresholds when there are two asset classes with significantly different allocation percentages. This applies to my stock/bond allocation.

I have a script that accesses my portfolio spreadsheet to send me an email when I need to rebalance my portfolio. Until recently, such alerts were rare. But during the pandemic, I’ve received several alerts to rebalance between stocks and bonds. It’s not easy to buy back into stocks after they’ve crashed, but doing so has produced profits for me. As stocks rose again, it was easier rebalancing back to bonds, but I still found myself not wanting to sell stocks when they seemed to be rising. However, I’m confident that my mechanical strategy is better than following my gut.

Stock Sub-Portfolio

I view my stocks as a sub-portfolio that consists of Canadian and U.S. stocks. I’ve chosen to hold 30% Canadian stocks (in Canadian-dollar ETFs) and 70% U.S. and international stocks (in U.S.-dollar ETFs). I keep them in balance using the same rebalancing strategy described above for my stocks and bonds.

One of the things I’ve done with the extra time on my hands during the pandemic is to improve the part of my spreadsheet that shows me the exact trades to make in each account when rebalancing. Without the spreadsheet, this can get tricky when I adjust both the stock/bond balance and the Canadian/non-Canadian stock balance at the same time.

U.S. and International Stock Sub-Portfolio

Within my stock sub-portfolio, I think of my U.S. and international stock ETFs as a sub-sub-portfolio. The 70% of my non-Canadian stocks are split 25% in VTI, 20% in VBR, and 25% in VXUS. Before I retired and bought some bonds, I wrote about my reasoning for this allocation. However, any reasonable allocation can work if you stick to it rather than deviate out of fear or greed. Again, the recent work I’ve done on my spreadsheet to calculate rebalancing trade amounts helps here.

Conclusion

I remain satisfied with the investment plan I’ve chosen. Recent changes I’ve made to automate my portfolio even more than it was before should help stop me from making costly mistakes. I tend not to even look at my portfolio value much. My spreadsheet calculates our safe after-tax monthly spending amount in real time. This amount tells me how my investments are performing, and helps us plan how much we can spend on our retirement activities and travel.

2 comments:

  1. Posts like this are the reason why yours is the best personal finance site I've found. I appreciate the details in the PDFs as well.

    ReplyDelete
    Replies
    1. Anonymous: Glad you like it. It's doubtful that many people could make use of the math I derived, but I'm hoping a few can, possibly including some advisors.

      Delete