I parted ways with my previous employer in 2015, which was a significant financial hit because it had a “gold plated” defined benefit (DB) pension. If I had stayed with that employer until I retired, I was on track to receive a pension of $70,000 per year, significantly higher than the average employment income for a male in Canada (see appendix). However, when I did leave, the value of my DB pension only equated to $7,200 per year, about 90% less than what I should have received. Needless to say, this was a huge point of stress for me as it significantly impacted my retirement plans. Notwithstanding the drop, with such a long time horizon, I asked myself: “Is it worth it to take the $600/month salary in 25+ years, or can I do better?” From that question was borne The Great Pension Experiment.
To recap from previous posts, I felt confident that I could earn a better pension than $600 a month, if I started managing my DB pension myself. To “manage” the DB pension meant that I would have to take a lump-sum payout, and invest the money into securities I felt would give me a better rate of return. My strategy was to leverage a variation of the very popular couch potato portfolio to grow my assets, and when I reached the age at which I could withdraw the funds, convert those into an annuity.
For the annuity, I assumed that I would receive a 4% cashflow on the annuity. Reviewing the latest (as of August 2020) annuity rates from Life Annuities, the lowest available annuity provides $391.19 in monthly income per $100,000 invested, which equates to $4,694 per annum or a yield of 4.7%. With that in mind, a 4% payout ratio still seems like a reasonable metric. (If anything: targeting a 4% payout and receiving a higher payout puts us even further ahead). The original DB pension was paying $600 per month, or $7,200 per year. At a 4% payout, that means we need a lump sum of $180,000 to invest into an annuity to break even with the DB pension.
|Company||Age 60||Age 65||Age 70||Age 75||Age 80|
The original payout of the DB pension was a little under $65,000, which means I need to triple my money over a 25-year timeframe. With that in mind, key measures of success are:
- Portfolio income. How much is it currently generating, and what is the probability that we will generate at least $600/month in the future?
- Portfolio performance. What is the net value of the portfolio relative to the present value of the pension I would have received, and how is the portfolio performing relative to the professionals?
With those metrics in mind, how are we faring for 2019?
The first metric for portfolio income is to review the actual income that the portfolio received. For the 12 months ending October 31, 2019, the portfolio generated $1860 in income, which equates to $155/month – that is 75% less than what my DB pension would provide. However, that $600 won’t be paid for another 21 years, so I have some time to catch up. To that end, when I review the yearly income numbers for my portfolio, the results look promising:
|Year Ending||Income Growth||CAGR Income Growth|
With 21 years left, $155/month income at the current CAGR will generate $608 in income. However, that number is based off of the assumption that CAGR will remain at or above 6.4% going forward.
The second metric is to check how much income I could receive from an annuity today if I were to cash out the entire portfolio. The portfolio did well in the year ending October 31, 2019, and if I were to cash it out into a 4% annuity it would generate $3,392 in annual income, or $283 in monthly income. Again, this is a far cry from the $600 I would receive from the DB pension, but it is as of this moment, and I expect the portfolio to grow in value over time; more on that in the performance section.
Those first two measures are summed up succinctly in this graph:
The red line represents the value of the DB pension (i.e. the $600/month) in today’s dollars. That $600 per month in the future is worth a lot less today, so for an accurate measure we should compare the discounted value to what the portfolio is currently generating. With that in mind, the point of success will be when one of the black lines crosses the red line: that means that the income the portfolio is generating (real income in the case of the solid line, or annuity income in the case of the dotted line), is exceeding that of the DB pension.
The third, and what I consider biggest, test to ascertain income in the future is in the use of a Monte Carlo simulation. With this method I run a number of simulations (250,000 simulations to be precise), and see how the value of the portfolio changes over time based on the randomness of expected returns. The reason I do this is that using a historical average is not necessarily the best way to estimate the future value of a portfolio: an average is just that, an average of high and low values. As an example, if the historical average return of a portfolio is 5%, that means that some years it could have lost 2%, but some years it could have gained 12% – the 5% is just that mid-point between the low and high return values. For that reason, we should take into account some variability in portfolio returns. While not the only way to consider this variability, the Monte Carlo simulation is straight forward to implement: we see how the portfolio performs over a given time period (in this case, from 2020 to 2040), make note of the results, do this a quarter of a million times, and then determine what the most likely outcome is based on those 250,000 simulations.
For this year, I used an annual average return of 8.15% and an annual standard deviation of 8.36%. The standard deviation is important because it gives me the effective range around which the annual average is focused. These values were calculated by taking the average returns of the benchmark portfolio from 2005 to 2019.
The results of the simulation are captured in this histogram:
The histogram illustrates the most likely monthly income based on converting the value of the portfolio to a 4% annuity. Reviewing this, the peak (i.e. most likely) scenario is monthly income between $1006 and $1256, based on cashing out the portfolio and purchasing an annuity with a 4% return (which implies a portfolio value between $301,000 and $377,000). If we poke at this a little further, this implies that the probability of generating at least $600 in income is 90%.
The other metric is how well the portfolio is performing, especially against professional fund managers such as my previous employer.
|Period Ending||TTM Return||Since Inception Return||OPTrust Returns||OPTrust since inception|
Since I started this experiment, my total return has been 31.00%, vs. my former employer’s 30.36% return – so based on numbers alone I am exceeding what my investments would have made with my former employer.
Due to time commitments I was unable to do an update in 2018, so the results here are relative to my 2017 update. That said, overall, the experiment to date has been a resounding success:
- Based on simulations of average returns, the portfolio has an 89% chance of exceeding the income I would have received from my employer
- The total return of the portfolio is slightly better than what I would have received if I left my money with them
- To date, the actual and projected income are inline with what I would expect, given the time horizon remaining
This emphasizes the fact that you don’t have to take what is given to you. If you have the patience and the stomach to weather the markets, you have a higher probability of coming out ahead than if you settle for what is given to you.
I am worried about the 2020 update; with COVID-19 the markets have just now started recovering from the massive drops at the beginning of the year, and the dividend income landscape is constantly shifting as companies revisit their dividend policies. However, that will be a discussion for later this year.
Onwards and upwards!
Average Income for Males as of 2018
Source: Statistics Canada.
- Age group: 16 and over
- Income source: Employment Income
- Sex: Males
Another year has passed, which means it is time to review The Great Pension Experiment to see how we are doing.
This will be my second year of results to review, and as such I’ve had some time to mull over exactly how we can measure performance, to see how the experiment is running. When I first discussed the experiment, I decided to undertake it because I thought I could do better than what my previous employer was offering me for a pension when I turned 65 years old. Based on my pension with that company, I would receive a monthly pension of $600, or an annualized pension of $7,200. I argued that I could take the lump sum that the company would give me if I cashed out my pension, and over time, end up with a portfolio that would pay me more than $600/month. To that end, there are a few key metrics to see how well we are doing.
First and foremost, I want to check my actual returns over the previous 12 months, and compare them to the annual returns of OPTrust. This is mainly a source of pride: I’d like to see if I, as a small time investor, can beat the performance of a pension fund with over $19 billion in assets, which pays money managers $45 million in investment administrative expenses (See Note 10b of their 2016 annual statements). As an aside, for the purposes of my own portfolio I use a October 31 year-end, since I started the portfolio in November 2015. OPTrust uses a calendar year-end — so I will be comparing my November 1 to October 31 results to their January 1 to December 31 results.
Second (and arguably the most important), we have to compare the monthly income my portfolio would give me, versus what my pension from the company would have given me. The best measure of this is to determine what the present value of the monthly pension from OPTrust would be, and compare that to the currently monthly income that my portfolio is giving me.
Finally, using Monte Carlo Analysis (MCA), we can run a model to see how well the current performance of my portfolio will project into the future. To do this, we will use actual returns up to the point of review (in this case, two years of actual returns), and then use MCA for the remaining years. For the experiment, the total duration (i.e. from when the portfolio started, to when I turn 65), is 26 years. Since we are two years into the portfolio, that means we have two years of actual returns, and will use MCA for the remaining 24 years. From there, we can see if the probability of beating $600/month in income at age 65 (i.e. what my company pension would have been) is still favourable.
So, how did we do for 2017?
The returns will be discussed in a subsequent section. For 2017, I pretty much left the portfolio on auto-pilot. Using synthetic drips, my brokerage continued to purchase shares for me whenever a dividend/distribution was issued. The composition of the portfolio is now:
All in, we picked up 12 shares of VCN.TO, 18 shares of VXC.TO, and 14 shares of VAB.TO. Because a synthetic drip cannot fully invest all proceeds, there has been a slow buildup of cash: in 2016 we had $106 in cash and we now have $370 in cash. When the cash account hits the $1,000 mark, I’ll redistribute it to one of the ETFs. There is little value in doing so right now with such a small amount: to do so would 2.7% going to transaction fees.
|Period Ending||Returns||OPTrust Returns||Beat (Miss) vs. OPTrust||Assertive Couch Potato Returns||Beat (Miss) vs. Couch Potato|
Total returns for the period ending October 31, 2017, were 12.05%. This yields a 18.53% return since inception, or an 8.87% return compounded annually over the past two years. Since The Great Pension Experiment is based on the Assertive Couch Potato portfolio, it is useful to compare my results to the Assertive Couch Potato, since that is effectively my benchmark. Ignoring transaction fees, the return on the Assertive Couch Potato was 10.6%. All in, I have beat the benchmark by 13.67%. Unfortunately OPTrust hasn’t yet published its 2017 results, so I do not yet know if I have beat them, but I will publish an update once their results are in.
Returns aside, the monthly income of the portfolio is the real “meat” of the investment: since this is what is supposed to support me in retirement. To that end, the following graph highlights the salient points:
|Year||Present Value of
OPTrust Annual Income
|Annuity Income||Real Income|
For 2017, discounting back the OPTrust pension at 2%, the pension is worth $4,471/year or $373/month. The Great Pension Experiment is currently producing $1,590/year in real income (i.e. from dividends and/or distributions), or a hypothetical $3,069/year if we were to cash out the entire portfolio and buy a 4% annuity today. The real test of The Great Pension Experiment will be when one of the black lines crosses the red, since that will signal that The Great Pension Experiment is now generating income greater than the pension that OPTrust had offered me. It is much too soon to tell how well we are doing, but as long as the Real Income and/or Annuity Income are increasing over time, we should be okay.
Monte Carlo Analysis
Rerunning the Monte Carlo Analysis (MCA) using real returns for 2016 and 2017 still produces a favourable graph:
With the newest MCA completed, our probability of exceeding $600/month in income by 2041 is as follows:
So, while we had a dip last year, we’re back on track at a 93% probability of making our income targets.
All things considered, I would consider 2017 to be a good year for The Experiment. While I am still a far ways away from beating the pension that OPTrust would have given me (I’m about two-thirds there when using the Annuity Income projection, and about one-third there when using the Real Income value), I do have a very long time horizon: I still have in excess of 20 years to make this work! So for now, I will sit back and let the portfolio do its thing.
Onwards and upwards!
This is part 3 of a 3 part series I have entitled "The Great Pension Experiment", which details my analysis on what to do with a defined benefit pension plan payout. The first two parts may be found here: Part I, Part II.
It has been one year since I started off on The Great Pension Experiment. I feel that this is a useful experiment because it leverages real world results in a closed environment. Because the funds are in a LIRA, I am unable to add or withdrawal funds, so any losses must be recouped “internally” by better investments.
In Part II I made the claim that, based on the Assertive Couch Potato Portfolio, over a 25 year time horizon I would be able to grow the portfolio to a point where I would be able to generate over $600/month in passive income, assuming a 4.00% yield. It has been 12 months since that claim, so lets see how we’ve done after one year.
When the portfolio was first opened, I set it up with a blend of 25% fixed income, 25% Canadian Equities, and 50% of non-Canadian Equities:
Originally I had intended on re-balancing the portfolio semi-annually (i.e. every six months), and in retrospect, this was a stupid idea. Cash was building up in the portfolio slowly, but sitting there idle until I had a chance to re-balance. Due to the already high number of shares, any dividends and/or distributions from the holdings would be in excess of the current price of those shares. Because of this, I should have been using synthetic drips right from the start! Accordingly, in September 2016 I set up my brokerage to re-invest any dividends received directly into additional shares. This ensures that money is not sitting idle, and because the investments are via synthetic drip, I receive additional shares commission free.
That said, as of October 2016 (one year), the portfolio sits at:
For the year ending October 31, 2016, I received $1,485.24 in dividends/distributions ($123.77/month). The portfolio as a whole has grown from an initial investment of $64,723.32 in cash to $68,468.29, which represents a 5.648% total return. This is actually pretty impressive, since the first half of the fiscal year (From November 2015-April 2016), the portfolio was in negative territory:
|Period Ending||Open NAV||Close NAV||Return %||TTM Return %|
I had bought into the portfolio at a peak in 2015, and the market had a clawback shortly after. As a result, the portfolio lost money for the first six months. I did take the opportunity during that period to top off the VXC shares whilst I had some excess cash from distributions, which in retrospect was a wise choice. Buying when the market is in a downturn helps to dollar cost average down, ultimately increasing future returns.
The portfolio is more or less at the target weighting if you round to the nearest whole number, and because of this I see no need to buy or sell additional shares at this time. Hypothetically, if the portfolio were to keep this pace for another 24 years, even at a 5.000% return we would generate approximately $736/month in passive income. So based on simple extrapolation (i.e. assuming constant returns for the next 24 years), we are right on track.
That said, I am going to ignore the portfolio for another six months, and will revisit in May of 2017 to see if any rebalancing is required. Until then, the next update will not be until December 2017. Here’s hoping that the portfolio continues to provide excellent returns at that time.
Onwards and upwards!
This is part 2 in a 3 part series I have entitled “The Great Pension Experiment”. This will detail my analysis on what to do with a defined benefit pension plan payout, after leaving my previous employer.
I am a big fan of Monte Carlo analysis (MCA). In a nutshell, when performing an MCA we run a test a fixed number of times, using a controlled set of inputs, and observe the results.
I had mentioned on my previous entry that I felt I could do better than the $600 per month defined benefit pension that OPTrust was offering after I left the company. But, how would I quantify this?
I wanted my LIRA to be as simple as possible, to run virtually effortless. Because of this, I elected to use one of the Couch Potato model portfolios, specifically the one using ETFs. Within these options, given my 25+ year time horizon, I elected to use the Assertive Portfolio, which has the following composition:
- 25% of the Vanguard Canadian Aggregate Bond Fund ETF, VAB
- 25% of the Vanguard FTSE Canada All Cap Index ETF, VCN
- 50% of the Vanguard FTSE All-World ex Canada ETF, VXC
The weighted average expense ratio of this portfolio is 0.18% as of November 2016, and the returns are pretty impressive:
- 1-year return: 7.20%
- 3-year return: 12.61%
- 5-year return: 8.91%
- 10-year return: 6.24%
- 20-year return: 7.10%
- Lowest 1-year return (2008-03 to 2009-02): -24.95% (2008 financial crisis)
To perform my analysis, I elected to take the lump sum from OPTrust, and run an MCA on it to see what the projected monthly income would be in 25 years, if I had used the Assertive Portfolio. To do this, I ran 100,000 iterations on 25 years of growth in the portfolio, using randomized returns. Here is an example of one set of returns:
|Year||Start Value||% Gain||End Value||Implied Annual Income||Implied Monthly Income|
In the above, the % gain is a random gain for the portfolio based on an average return of 8.29%, and a standard deviation of 7.96%. The implied annual income assumes I could take the ending value of the portfolio, and buy an annuity or other similar (set of) instrument(s) to generate 4% of annual income. I believe 4% annual income, if you are not concerned with growth, is incredibly doable in the market.
But wait, where did that 8.29% average return, and 7.96% standard deviation come from? And what do they mean?
If you take 15 years of returns for the couch potato model portfolio, and do some statistical analysis on that data, you will end up with an average daily return of 0.032%, which equates to an average annual return of 8.287%. Moreover, those same daily returns have a standard deviation of 0.503%, which is 7.959% annualized. (For annualizing, I assume there are 250 active trading days in the year: 52 weeks @ 5 days/week, less 10 days for various holidays). Now, because the model portfolio asks for VAB, VCN, and VXC, and those ETFs are relatively new, I used iShares Core S&P/TSX Capped Composite Index ETF (XIC), iShares Canadian Universe Bond Index ETF (XBB), iShares MSCI World Index ETF (XWD), and iShares Core S&P 500 Index ETF (CAD- Hedged) (XSP) as proxies:
- For the period of October 29, 2009 to December 31, 2015, I used a blend of 25% XIC, 25% XBB, and 50% XWD.
- For the period of April 15, 2002 to October 28, 2009, I used a blend of 25% XIC, 25% XBB, and 50% XSP.
The ETFs listed are iShares ETFs. The XWD is the equivalent to VXC, but prior to October 29, 2009, there were no ETFs I could find that were all-world excluding Canada ETFs. Moreover, the period I chose covers the tail end of the dot-com bubble, as well as the massive 2008 financial crisis. Doing this provides more real world examples. In fact, looking at the above table you can see that in this iteration, in years 18 and 20 the sample has massive declines of -14% and -8% respectively! I feel this is an accurate representation of what could happen.
The average and standard deviation play into each iteration of 25 years. In the table above, the “% Gain” column will be, on average, 8.29%, and vary with the standard deviation of 7.96%.
Now, if we pull this all together:
- Create the table above 100,000 times
- Take the final 25 year implied monthly income from each iteration
- Plot a histogram
We get the following:
The above histogram shows that we have a 16% probability of having total monthly income less than or equal to $786. Put another way: we have an 84% probability of having implied monthly income greater than $786. If we take this a little further, we have an approximately 7% probability of making at most $600, or a 93% probability of making at least $600.00. But wait, OPTrust’s defined benefit income would be $600/month. I have only a 7% chance of not beating OPTrust’s defined benefit pension! Screw you OPTrust, I’ll take my chances.
So, based on my Monte Carlo analysis, using 13 years of historical returns on a portfolio of 25% Canadian Equities, 50% non-Canadian Equities, and 25% Canadian Fixed Income, I took the plunge to invest all of my money from OPTrust into my LIRA, using the couch potato portfolio. That was almost a year ago. In my next post, I’ll speak to the first year of actual results.
Onward and upward!
This is part 1 in a 3 part series I have entitled "The Great Pension Experiment". This will detail my analysis on what to do with a defined benefit pension plan payout, after leaving my previous employer.
Up until 2015, I worked for the OPSEU Pension Trust, a major pension plan in Ontario, and had the benefit of a “golden pension”, or a Defined Benefit Pension. These pensions are considered the gold standard in the pension industry because the total payment that one receives upon retirement is guaranteed, regardless of what happens in the markets.
However, towards the end of my tenure at the company, a new CEO came in, and there were a number of changes with upper management. As a result of these changes, there was a large amount of restructuring as the company made attempts to streamline processes and change the long term strategic goals and organization of the firm. As with any restructuring, there are causalities, and my department was no different: my entire department was explained away, and I left the organization.
Due to pension law in Canada, when an employee leaves a company, they are given three options with the money that is with the company for their future pension:
- Leave the money with the company, and at age 65, start taking a pension.
- Transfer the money with the company to your new employer’s pension plan. In this way, any deductions which were taken off of your paycheque whilst with employed with the former company, would carry forward with you to your new company. The caveat to this is that it is not the actual deductions which would transfer, but the commuted value of those deductions.
- Transfer the commuted value of your pension to a Locked In Retirement Account.
Before continuing with the background information, it would help to get some terminology out of the way..
The commuted value of a pension is an actuarial calculation which states what the total value of your pension in the future is worth today. To give a very simple example, assume you received a pension of $1,200/year ($100/month) starting at age 65, and given your actuarial profile, you would live to the age of 85. Given that, the total payments you would receive would be:
But, that $24,000 is the value of the pension as you receive in the future; to be accurate, you have to discount that back to the present day. If we extend our example to assume that we have a discount rate of 5.25%, and we are currently 50 years old, this means that the first annual pension payment of $1,200 (i.e. when you turn 65) has to be discounted back 15 years at the discount rate of 5.25%, or . The next payment would have to be discounted 16 years, or . If we do this for all of our payments to age 85 (our assumed death age), then the present value of the pension is then:
And if we do the math on this, we find that our $1,200 annual pension at age 65 is worth $7,353.54 in today’s dollars.
The example above is very simple, and a more true calculation would discount each (monthly) payment by the discount rate. Also, the assumption that the pension would last for 20 years is based on a number of actuarial tables, etc. But at the end of the day, the commuted value is the present day value of all future pension payments.
The other item to discuss is the Locked In Retirement Account, or "LIRA" for short. A LIRA functions much like an RRSP, with the following key differences:
- Under normal circumstances, funds in the LIRA are locked in until either your retirement age, or some other specified age as specified by the province the LIRA is registered in. For my purposes, I am basing my calculations assuming the LIRA is locked in to age 65.
- Under normal circumstances, money added to a LIRA can come only from another LIRA, or registered pension. You cannot add money to a LIRA at a time of your choosing as you would with an RRSP.
- Depending on the circumstances, you may be able to transfer your LIRA to your pension plan with another employer in the future.
Put another way: when you open the LIRA, any and all money is inaccessible until the prescribed age of withdrawal, or if available to you, you transfer the funds to another registered pension plan. If the market goes south and your LIRA loses 90% in value, you cannot add money to bring it "back up". Similarly, if it grows in value over time, you cannot withdrawal that money until some time in the future.
For my own personal situation, after leaving my previous employer I elected to incorporate myself and become an independent consultant. This mean that Option 2 was not possible, as I would no longer be employed by a company with a pension plan to transfer the funds into. The other options would be Option 1, leave my money with OPTrust, and Option 3, transfer it to a LIRA.
The formula for OPTrust when calculating the defined benefit pension is:
Based on my years of service, and average salary, the pension I could have expected to receive at age 65 would be approximately $600/month, or $7,200/year. But, that is $7,200/year in 28 years 26 years!! The Bank of Canada has an inflation target of 2.00% annually. If we discount back the pension offered by my previous employer at the inflation target of 2.00%, that would equate to $426.67/month $352.12/month in today’s dollars. I’ll be honest: I think I can do much better than that.
That said, because the pension offered by OPTrust would be worth relatively little in the future, and because I did not have an employer to transfer my pension to, I went with Option 3. The second part of this series will outline the quantitative analysis I performed to come to this conclusion.
Edit – January 2, 2018
I realized that there was an error in my above calculations. Since the pension kicks in at age 65, when I wrote that post there were 26 years remaining, not 28. Moreover, there was an error in my original worksheet. The discounted monthly pension is actually $352.12, not $426.67, in today’s dollars (well, at that time!). Fortunately, this error is in my favour: the $352.12 threshold is even more attainable than the $426.67 threshold.