Bridges are never designed for average wind speed and average traffic. They are designed for the worst so that the design can withstand adverse conditions. Retirement planning is no different. A good retirement plan should be able to withstand and overcome adverse market conditions of the future.
In current retirement planning practice in India, financial planners assume steady state growth rate of equity markets and inflation. These assumptions allow financial planners to project cash flows into the future. Plan looks organised and everything appears to be in control.
But in real world, stock markets fluctuate and so does inflation. Stock markets do not only fluctuate but go through bull, bear and sideways cycles which could last 15 or more years in best or worst cases. Retirement plans made using average growth rate of portfolio and inflation make no allowance for market cycles, sequence of inflation and return investors face in real life.
The question that arises then is “Can these retirement plans survive bad market conditions of the future?”. We do not have crystal ball to tell us the future sequence of return and inflation to test robustness of retirement plans, but we can test these plans with actual historical data with real sequence of returns and inflation.
This is called “aftcasting” where historic data is used, as opposed to “forecasting” where we assume growth, inflation and interest rates 30 or 40 years into the future. Aftecasting can give us a range of outcomes based on market history ranging from unlucky to lucky and everything in between. A robust retirement plan should be able to withstand unlucky outcomes of the past if we hope it to withstand unlucky outcomes of the future. Remember the worst of the future could be worse than the worst of the past.
If retirement plans we design can survive bad market conditions of the past, current retirement planning practice could be considered as good. But if they can’t, then we might have to reconsider the way we do retirement planning.
I have used retirement calculator developed by Jim Otar, the author of “Unveiling the retirement myth” to run aftcast for historical data. Otar calculator allows us to run aftcast for US, Japan, Britain, Canada and Australia market history. I have used US market history for this article. Indian market history is too short for this exercise.
It could be argued that it is not proper to use US data, but since US emerged as the most successful economy in the world, a good retirement strategy should work fine at least with US data. If it doesn’t then it is likely to fail in India too.
Current retirement planning practice in India
Let us first understand how retirement planning is done in India.
Suppose a couple aged 35 years, want to plan for 30 years in retirement starting at age 60. Their estimated monthly expense in retirement in current value is ₹40,000.
If expenses are assumed to grow at inflation rate of 7%, the expense in the first month of retirement will be ₹2,17,097 i.e. ₹26,05,167 expense in the first year of retirement.
Once expense in the first year of retirement is known, all that is needed to calculate retirement corpus is the return retirement corpus will generate during retirement. Retirement corpus is simply the present value of all expenses in retirement discounted by the return retirement portfolio will generate.
Different financial planners take different return assumptions from retirement corpus. Some financial planners are conservative and assume that retirement corpus would not generate any real return post tax during withdrawal phase, while others are optimistic and assume that upto 3% real return can be generated from retirement portfolio. Most financial planners assume real return from retirement portfolio during withdrawal phase to be 1 to 2%.
Freefincal robo advisory template uses bucket strategy to calculate retirement corpus with conservative return assumptions from equity and debt. But retirement corpus it calculates is more or less same as that calculated with zero real return assumption from retirement corpus.
The retirement corpus calculated for different return assumptions from retirement portfolio in retirement for our 35 year old couple is as follows
Case 1 Real return of 3% from retirement portfolio in retirement : Retirement corpus required ₹5,23,82,500
Case 2 Real return of 2% from retirement portfolio in retirement : Retirement corpus required ₹5,94,09,930
Case 3 Real return of 1% from retirement portfolio in retirement : Retirement corpus required ₹6,78,80,260
Case 4 Real return of 0% from retirement portfolio in retirement : Retirement corpus required ₹7,81,54,920
Retiring in any year in US since 1900 with above retirement corpus
Let us now use actual US market history since 1900 with growth rate and inflation as they actually happened; with underlying correlation between stocks, interest rates and inflation, actual volatility and actual market panics to see how our couple’s retirement portfolio would have fared in real life. This will be the real test of the robustness of their retirement plan.
Let us assume that 40% of the retirement corpus is invested in S&P 500 index and 60% in bonds yielding 1% premium over 6 month term deposits. The portfolio is rebalanced annually if equity percentage deviates over 3% from the target. Monthly withdrawal at the start of retirement is ₹2,17,097. The expenses are assumed to increase with real sequence of inflation.
Inflation is as big contributor to the failure of retirement plans as bad sequence of return from equity is. I have used 40:60 equity debt allocation for aftcasting instead of more conservative allocation because we get better aftcasting results with higher equity allocation in retirement portfolios.
Many investors and advisers think that higher withdrawal can be done from retirement portfolios by increasing equity allocation. There are articles written by mutual fund distributors suggesting SWP from ICICI Pru. Balanced Advantage kind of funds as a good retirement planning strategy. This article can give an idea about how such aggresive strategies to fund higher withdrawal rates from retirement portfolios would have fared with actual historical data with real sequence of return and inflation.
The results shown in this article are for the best managed retirement portfolios in the most successful economy in the world. Behavioural mistakes and sudden large expenses like medical emergencies are discounted too. If a strategy fails inspite of that, it should be termed as a bad strategy.
Here is what would have happened had our couple retired in each year since 1900 in US at age 60.
Case 1 Real return of 3% from retirement portfolio in retirement
Retirement corpus: ₹5,23,82,500
Starting annual withdrawal: ₹26,05,164
Withdrawal rate at the start of retirement: 4.97%
The red line shows the asset value of bottom decile portfolio value since 1900. This is the unlucky outcome. It shows the portfolio value where 90% did better and 10% did worse. This is the line we should use for retirement planning to make sure that odds are on our side during the retirement.
The blue line shows the asset value of the median of all portfolios since 1900. The median is where 50% of all portfolios had a lower value and 50% had a higher value.
The green line on the chart shows the top decile (top 10%) portfolio value of all the portfolio values since 1900. This is what could have happen if our couple were really lucky and caught a secular bull market at the beginning of their retirement.
The black line shows the worst case scenario.
Portfolio depleted by age 80 in 10% of the cases and in 50% of the cases it depleted by age 87. The probability of depletion of portfolio one year before assumed age of death i.e age 89 is 58%.
This is terrible way of doing retirement planning. Remember we are considering a case where portfolio is managed better than any financial planner can ever hope to do and in the most successful economy in the world. The situation will be lot worse in most real life cases.
The results do not get any better by increasing equity allocation. In fact probability of depletion increases if we increase equity allocation because higher equity allocation brings with it higher volatility which increases the impact of reverse rupee cost averaging.
Retirement plans designed with 3% real rate of return assumption from retirement portfolio or close to 5% starting withdrawal rate offer no margin of safety. These plans cannot cover longevity, market or inflation risks. If you find yourself a situation at the start of a potential 30 year retirement where annual withdrawal rate is 5% or more, the sensible thing to do is to reduce withdrawal. Do not expect your adviser to perform miracle either because if he tries to do so, it is highly likely that it will backfire sooner or later.
Case 2: Real return of 2% from retirement portfolio in retirement
Retirement corpus: ₹5,94,09,930
Starting annual withdrawal: ₹26,05,164
Withdrawal rate at the start of retirement: 4.38%
The probability of depletion of portfolio one year before assumed age of death i.e age 89 is 31%.
While this is better situation than case 1, probability of depletion one year before assumed age of death is still 31%. This portfolio cannot withstand unlucky outcomes in retirement. This is like designing a bridge for average wind speed and average traffic which will collapse if wind speed or traffic moves away from average in the wrong direction.
Case 3: Real return of 1% from retirement portfolio in retirement
Retirement corpus: ₹6,78,80,260
Starting annual withdrawal: ₹26,05,164
Withdrawal rate at the start of retirement: 3.83%
The probability of depletion of portfolio a year before assumed age of death i.e age 89 is 6%.
This is safer way of designing retirement plans. It creates ample margin of safety for portfolio to survive unlucky outcomes in the retirement. But it may not be able to handle the risk of longevity.
Many financial planners who design retirement plans with 1% real rate of return assumption, take 85 years as expected age of death. The otar retirement calculator gives survival rate of 30% for male and 45% for female at current age of 35, and it gives survival rate of 34% for male and 49% for female for current age of 60 (Probability of survival increases with increase in current age). These are survival rates for developed countries. Survival rates for clients of most financial advisers in India may not be very different.
Age of death assumption of 85 years may lose all the margin of safety 1% real rate of return assumption creates in retirement plan. It is better to design retirement plans at least till age 90. Jim Otar take 95 as the default age to design retirement plans and uses 97 as age of death for most of his clients.
In this case the withdrawal rate at the start of retirement is close to 4% which is considered as the safe withdrawal rate. But things can go wrong even at 4% rate of withdrawal.
Case 4: Real return of 0% from retirement portfolio in retirement (freefincal robo advisory template)
Retirement corpus: ₹7,81,54,920
Starting annual withdrawal: ₹26,05,164
Withdrawal rate at the start of retirement: 3.33%
The probability of depletion of portfolio a year before assumed year of death i.e. age 89 is 1%.
Financial planners who are conservative with return assumptions from retirement portfolio are also likely to be conservative with their suggested equity-debt allocation in retirement. Therefore let us rerun the aftcast with 25:75 equity:debt allocation of retirement portfolio.
Even with 25-75 equity-debt allocation, result is no different. The probability of portfolio depletion one year before assumed age of death is still 1%. We get similar aftcast results for bucket strategy with 15 years expenses in fixed income with remaining assets invested with 50:50 equity debt allocation.
This is the safest way of doing retirement planning. Retirement plans so designed can cover longevity, market and inflation risk. The portfolio is likely to be accumulating during retirement. Portfolio can be expected to provide lifelong income with any reasonable asset allocation strategy.
Some readers may think that majority of unlucky years to start retirement must be before 1950s that saw two world wars and great depression. But unlucky outcomes need not necessarily be the result of terrible events like great depression or world wars. Had our couple started their retirement in year 1966 and lived for another 30 years until 1995, they would have run out of assets even when no significant international event happened during that period. In fact real return of S&P 500 for the whole 30 year 1966-95 was 5.3%; not too far below the historical US stock of return of 7% post inflation. The return from 1983-95 was spectacular.
The only problem was that for the first 17 years (1966 to 1982) the return of S&P 500 was 6.81% which was identical to the rate of inflation for the period, making the real stock return for the period 1966-82 zero.
This is how retirement portfolios designed with 3%, 2%, 1% and 0% real rate of return assumption would have fared had retirement started in year 1966.
Case 1: Retirement assets would have depleted by age 81 or by year 1987.
Case 2: Retirement assets would have depleted by age 86 or by year 1992.
Case 3: Retirement assets would have provided lifelong income.
Case 4: Retirement assets would have provided lifelong income.
Retirement portfolios do not grow in straight line at assumed rate of growth in real life. If it turns out that start of the retirement also marked the beginning of a long and brutal bear market, and there are terrible post inflation return in the first decade of retirement; a retirement corpus which is supposed to finance 30 years in retirement, can run out of money long before reversion to mean saves it from the initial bad luck.
The point is not to predict when bad luck may hit the retirement portfolio, but understanding that it does hit from time to time and to design the retirement portfolio strategies appropriately. Retirement plans designed with higher than 1% real return assumption from retirement portfolios do not provide any protection from the misfortune or bad luck that may befall in retirement. Zero post tax return assumption from portfolio in retirement is still the safest way to design retirement plans.
This article can be summed up with a quote from William Bernstein. “When all is said and done, I still know of no better risk analysis tool for retirees under the age of 70 than this simple narrative: At a 2 percent withdrawal rate, your nest egg will survive all but catastrophic institutional and military collapse; at 3 percent, you are probably safe; at 4 percent, you are taking real chances; and at 5 percent and beyond, you should consider annuitizing most, if not all, of your nest egg.”