There is a United States based study called the trinity study that looked into how much someone can successfully withdraw from their investment portfolio over a 30-year period. The result of the study concluded that your portfolio had a 95% chance of surviving between 15 and 30 years at a withdrawal rate of 4% per year, adjusted each year for the cost of inflation. Also known as the 4% rule. The assumption of the study was 50% of your portfolio in stocks and 50% in bonds. The study was conducted based on stock/bond performance between 1926 and 1995.
If we experience 3% inflation for example, someone retired with $1 million ($500,000 in stocks, $500,000 in bonds) could withdraw $40,000 in year one, $41,200 in year 2 ($40,000 + 3% inflation), $42,436 in year 3 ($41,200 + 3% inflation) and so on.
There are a few limitations to the study
It was performed over a period of 30 years. If you are looking to retire early you could be retired for between 30 and 60 years.
Because the last year of research was 1995, to consider 30-year retirements they could only consider retirement dates from 1926 to 1966. Now that we are in 2018, we are missing out on an extra 22 years of more recent data (1966 to 1988).
The study was performed solely on United States data. How can we be sure that the data is relevant for New Zealand? Our markets perform differently, and it would be dangerous to base our retirement estimates solely on an international study.
This is why I have decided to run the numbers for myself, based on historical New Zealand market data. It is relevant to our market, it includes periods of over 30 years, and it is more recent data.
A national stock exchange was not formed in New Zealand until 1974. My data runs from the first full year of results in 1975 to 2016 and assumes a 2.5% inflation rate. This includes data for:
22 periods of 20 years
17 periods of 25 years
12 periods of 30 years
7 periods of 35 years
1 period each of 36 to 41 years
With 100% in stocks, all 3 withdrawal rates (3, 3.5 and 4%) were successful 21 out of the 22 (95%) 20-year periods. (1975-1996). The highest successful withdrawal rate was 14.5% if you started withdrawals in 1979, and the lowest successful withdrawal rate was 2.1% if you retired in 1986 just before the share market crash. The average successful withdrawal rate was 7.6% thanks to extremely high returns in the late 70’s/early 80’s. Has been a much lower 5% average since 1985.
With 75% in stocks and 25% in bonds, withdrawal rates of 3.5 and 4% were successful 95% of the time, with a withdrawal rate of 3% successful on all 22 x 20-year periods. The highest successful withdrawal rate was 11.4% in 1979 and the lowest 3.1% in 1986. Average withdrawal rate of 6.8%. Just 5.1% since 1985.
With 50% in stocks and 50% in bonds, withdrawal rates of 4% were successful 95% of the time, whilst 3 and 3.5% were successful in all 22 periods. The highest successful withdrawal rate was 9.8% in 1979 and the lowest 3.7% in 1986. Average withdrawal rate of 6.4%. Just 5.1% since 1985.
With 100% in stocks, a 4% withdrawal rate was successful just 12 out of the 17 (71%) 25-year periods. (1975-1991). All 5 failed periods were either side of the share market crash (1985 – 1989). A 3.5% withdrawal rate was successful 82% of the time (14 out of 17). Failures occurred between 1985 and 1987. A 3% withdrawal rate was successful 94% of the time (16 out of 17). Once again, failing in 1986. The highest successful withdrawal rate was 11.9% in 1979, and the lowest successful withdrawal rate was 1.6% if you retired in 1986 just before the share market crash. The average successful withdrawal rate was 6.9% thanks to extremely high returns in the late 70’s/early 80’s. Has been a much lower 3.9% average since 1985.
With 75% in stocks and 25% in bonds, withdrawal rates of 4% were successful 71% of the time, failing consecutively between 1985 and 1989. A withdrawal rate of 3% and 3.5% was successful 94% of the time. The highest successful withdrawal rate was 9.4% in 1979 and the lowest 2.6% in 1986. Average withdrawal rate of 6%. Just 4% since 1985.
With 50% in stocks and 50% in bonds, withdrawal rates of 4% were successful 76% of the time, 3.5% was successful 94% of the time, and 3% withdrawal rate was successful in all 17 25-year periods. The highest successful withdrawal rate was 8.1% in 1979 and the lowest 3% in 1986. Average withdrawal rate of 5.6%. Just 4.1% since 1985.
With 100% in stocks, a 4% withdrawal rate was successful 9 out of the 12 (75%) 30-year periods. (1975-1986). All 3 failed periods were just before the share market crash (1984 – 1986). If it follows the 25-year data then 1987, 1988 and 1989 will likely be failures in 30-year data that will be revealed in the next 3 years. A 3% and a 3.5% withdrawal rate was successful 83% of the time (10 out of 12). Failures occurred in 1985 and 1986. The highest successful withdrawal rate was 10.6% in 1979, and the lowest successful withdrawal rate was 1.5% if you retired in 1986 just before the share market crash.
With 75% in stocks and 25% in bonds, withdrawal rates of 4% were successful just 67% of the time, failing consecutively between 1983 and 1986. A withdrawal rate of 3.5% was successful 10 out of 12 times. 3% successful 11 out of 12. The highest successful withdrawal rate was 8.3% in 1979 and the lowest 2.2% in 1986.
With 50% in stocks and 50% in bonds, withdrawal rates of 4% were successful just 67% of the time, failing consecutively between 1983 and 1986. A withdrawal rate of 3.5% was successful 10 out of 12 times. 3% successful 11 out of 12. The highest successful withdrawal rate was 7.1% in 1979 and the lowest 2.5% in 1986.
With each of the three different portfolio allocations, your portfolio would have survived each of the 3 different withdrawal rates tested over the 7 x 35-year periods from 1975-1981. The highest successful withdrawal rate was 9.6% for a 100% stock portfolio and the lowest successful withdrawal rate was 4.5% in a 50/50 stock/bond portfolio.
Other trends of note
Using the 4% rule, the 50% and 75% stock/bond portfolio had the best results for 25 years or less, whereas 100% stocks had the best results for 30 + years. Bear in mind though that the 30 + year periods only went up to 1986 in the study. They didn’t experience the 1987 crash.
For 30-year periods, 100% stocks can have a difference of 9% between the most and least successful withdrawal rates. 50% stocks and bonds, the difference is just 4.5%. In other words, you can get much better results with 100% stocks, but you can also get much worse results.
For 30-year periods, the withdrawal success rate was only 67% for 50% and 75% stock/bond portfolios, and just 75% for a 100% stock portfolio.
If we weren’t unlucky enough to retire less than 4 years before the 1987 share market crash of 48%, our portfolio would have stood the test of time in all cases studied.
1983-1989 retirees would have experienced a 4% withdrawal success rate of 0%.
The share market experienced a drop of 40% in 1990, which had a large impact on those that retired before 1987. Two crashes adding to a total of 88% within 3 years of one another.
In 30-year periods, 100% in stocks performed best in 9 out of 12 periods. It also performed the worst in the other 3 periods (before the crash). 50% stocks and 50% bonds performed best in the 3 worst periods. 75% stocks and 25% bonds performed in the middle of all 12 periods.
It may seem strange that the 7 x 35-year periods enjoyed a greater success rate than the 22 x 20-year periods. The difference is that the 35-year periods started in 1975 and finished in 1981. There is no 35-year data yet for post 1981. EVERY SINGLE FAILURE occurred between 1983 and 1989. So, if you started your portfolio withdrawal in any of these years you would have needed much more than 25 times (4%) your expenses to outlive your portfolio. In a couple of unlucky instances, you would have needed 50 times (2%) your annual expenses saved up for your retirement.
The closer to the 1987 share market crash you retired the worse off you would have been. Even though portfolios that started withdrawals in 1975 still would have been active during this 1987 crash period, they would have had 10 years of gains first, before crashing. That is why they survived. It is the portfolios that experienced the market crash with 5 years of retiring that are hit hardest.
This is the very real risk called sequence of returns. When we FIRST retire we tend to have more money than we will ever have. Any sharp drops in the returns of our investments are most significant at this point. Not only are we experiencing significant percentage decreases in our portfolio, we still have the longest amount of time remaining and we are not able to purchase the cheap new shares because we are no longer in the accumulation phase. For these three reasons, our portfolio is more exposed and less likely to last.
No one can predict when there is going to be a crash. Our best strategy to minimise the risk of our portfolio not lasting is to take on a strategy where we have enough bonds or low risk investments. It is a fine balance as we don’t want too many low risk investments. Many of us will require growth in our portfolios for the money to last, so we need to take on some risk with stocks. Especially for those of us wanting to retire early. The dilemma is we need stocks to sustain our portfolio, but stocks are also responsible for the biggest falls in our portfolio.
To counter the sequence of returns risk, I will be using a strategy called increasing equity glidepaths, that I will be discussing in the next article. It is basically starting retirement with a low level of investments in stocks, and then increasing them as we age. This goes against the common wisdom of a declining equity glidepath, but it will help reduce the sequence of returns risk, whilst allowing for portfolio growth.
So, is the 4% rule relevant to New Zealand? I would like at least 50 years more data to be able to answer that more confidently. From the data we have though, a 30-year 4% success rate of 67% is not great. Using the 4% rule only two thirds of us would have been OK retiring between 1975 and 1986. You can read my article here on tips for reducing the sequence of returns risk. For without that 1987 crash, closely followed by 1990 crash, our numbers would be very similar to those in the United States study.
It also serves us a nice reminder to invest in more than one geography. For example, if the NZ market is down, there could be another market in the world doing well. A globally diversified portfolio will help reduce this risk.
It does go to show though how just two years of poor returns can impact our livelihood. This is why the 4% rule should not be taken as gospel. History is by no means a prediction of future market trends either. It is a good starting point for calculating your needs, but your decision on how much you need for retirement should run a bit deeper than accepting this as a rule.
In times of poor returns we need to remain flexible. In reality, we wouldn't spend at 4% if the market was experiencing a downturn. In hard times we tighten the belt. People generally adapt to conditions, so to withdraw exactly 4% every year is not really practical. Rules of thumb are just that. Rules of thumb.
Whatever you decide is best for you, it is always best to have back up plans, and back up plans for your back up plans. Flexibility is the key.
The information contained on this site is the opinion of the individual author(s) based on their personal opinions, observation, research, and years of experience. The information offered by this website is general education only and is not meant to be taken as individualised financial advice, legal advice, tax advice, or any other kind of advice. You can read more of my disclaimer here
Share your thoughts below. How are you calculating your retirement number? Did any of these results surprise you?