A simple way is to invest in risk-free assets (Government bonds, PPF, EPF, bank fixed deposits). You will never see losses. However, by shunning risk completely, you might have to compromise a bit on the long-term returns.
And if you consider risky assets (gold or equities) for higher returns, you must prepare yourself for portfolio losses too. At the same time, avoiding losses (especially big losses) in the portfolio is important. Not only is watching deep losses in the portfolio painful, but it can also compromise investment discipline. Facing big losses, you might panic. You might stop making incremental investments or worse still, exit the investments at the worst possible time, making those losses permanent. The markets may subsequently recover, but your portfolio will not since you have already sold.
A way to reduce losses is to rely on Moving-average based or trend-based approach to time the entry into or exit from various risky assets. I have discussed a similar strategy in an earlier post. Not my cup of tea but you can try. The problem?
No strategy, no matter how good, works all the time. Therefore, the issue is such approaches (or any active strategy for that matter) is that there will be bouts of overperformance and underperformance. The periods of underperformance of difficult to digest. And it is not just about the underperformance compared to broader markets. It is also about underperformance compared to portfolios of colleagues, friends, or everyone we know. And this complicates matters. Here is the behaviour matrix.
Everybody else is losing money. We are losing money. (We are OK).
Everybody else is making money. We are making money. (We are OK)
Everybody else is losing money. We are not losing money. (We are HAPPY)
Everybody else is making money. We are not making money. (This is armageddon. Simply not acceptable or tolerable. We might shun strategy at the wrong time).
Apart from that, there will be a tax impact. Add to that, the constant requirement to monitor the market.
How do we reduce the losses then?
Personally, I prefer an approach that is simpler, easy to execute behaviourally, and requires lesser involvement.
I prefer to diversify the portfolio and work with an asset allocation approach i.e., bring different types of assets in the mix and hope that the losses will reduce.
Does this approach work?
Let’s find out.
Which assets shall we use?
We will consider domestic equities, international equities, gold, and fixed-income investments.
- Domestic Equity: Nifty 50 TRI
- International Equity: Motilal Oswal Nasdaq 100 ETF (There was no other international equity index that has a longer history). It would have been better if we had a more diversified indexing option in the international equity space.
- Gold: Nippon Gold BeES
- Debt: HDFC Liquid (We could have used any other debt mutual fund or bank FD rates)
Both Nifty 50 and Nasdaq 100 are equity indices. Hence, not really different assets
We consider the data from March 30, 2011 until December 31, 2020.
Let us first look at the calendar year returns.
Nasdaq 100 is the clear winner here, with no negative calendar year returns in the past 10 years. Note these are rupee returns. Hence, some of this super-performance can also be attributed to rupee depreciation in this decade.
Now, let’s look some the risk characteristics and return performance during the period.
Standard deviation is a measure of portfolio volatility. If portfolio volatility worries you, you must invest in a portfolio that has lower standard deviation. Lower the better.
Maximum drawdown is the maximum loss you would have suffered if you invested on any of the dates. For instance, if you invested Rs 1,000 in an index on a day and the lowest the value of the investment goes to ((in the future) is 900, then the maximum drawdown is 10%. Lower the better.
In previous posts, I have shown maximum drawdown for each of the dates in a chart. In this post, I have just picked up the biggest loss. For instance, the maximum drawdown for Nifty 50 is 38.27%. This happened in March 2020 (for the investment made on January 14, 2020).
It is never easy to digest losses in your portfolio. And that too big ones. It is painful. More importantly, it can compromise investment discipline. For instance, you might consider exiting your investment when it is falling sharply (since you feel it will fall more). It is easier to stick with strategies/investments where you lose less (lower drawdowns).
Rolling returns and CAGR indicate return performance. Higher the better.
CAGR indicates the annualized return you would have earned if you invested on March 30, 2011 (analysis start date) and held until December 31, 2002 (end date).
Rolling returns specifies holding period returns. You invested equal amounts on each day of the sample period and held on for exactly 3 years. Take a simple average of 3-year returns. You have the average 3-year rolling returns. In the previous posts, I have plotted the 3-year and 5-year rolling returns. In this post, I am just showing the average values. Higher the better. This is a better indicator of returns experience (than CAGR).
What happens when we mix two assets in the portfolio?
Higher returns, Lower volatility, neither or both?
When we mix two or more assets with low correlation (or negative correlation), you can expect a reduction in both standard deviation and maximum deviation (compared to individual assets). Reduction in standard deviation and maximum deviation is positive news.
The value of correlation coefficient can range between -1 and 1. Correlation coefficient of 1 means perfect positive correlation i.e., the two assets move in tandem. Both rise and fall together. You can see, from the table below, that the correlation of any asset with itself is 1.
Correlation of -1 means perfect negative correlation. When one rises, the other one falls.
If the intent is to diversify the portfolio, you must mix assets with negative or low positive correlation. We saw that in our post on mixing sectoral indices (Banking, Pharma, IT).
Let’s first look at the correlation between the 4 assets (sub-assets) considered.
The correlation coefficients are either negative or low positive. Therefore, you can expect that mixing these assets will add value (at least reduce risk).
Looking at the correlation coefficient values, you can be almost sure that there will be an improvement in risk parameters (standard deviation and maximum drawdown).
By the way, you cannot say the same about portfolio returns (on mixing two assets with lower correlation). The results from the portfolio may be higher or lower. We witnessed a rebalancing bonus (portfolio returns higher than returns from individual assets) when we invested 50:50 in equity and gold portfolio (March 2007-November 2020). But that’s no guarantee.
In the following tabulation, I have constructed portfolios using a different mix of various assets.
The first section is about single asset portfolios. We had seen the single asset performance earlier in the post. The same performance is reproduced here.
Subsequently, we combine various assets in different proportions. All multi-asset portfolios are rebalanced annually to target allocations on January 1.
We shall compare the performance of multi-asset portfolios with the performance of Nifty 50 on both risk and return characteristics.
Wherever the multi-asset portfolio has done better than Nifty 50 (lower standard deviation, lower maximum drawdown, higher rolling returns, higher CAGR), I have highlighted those characteristics in GREEN.
Where the multi-asset portfolio fares worse, such characteristics have been highlighted in RED.
You can see that, with any mix, there has been a sharp improvement in risk characteristics. In a few cases, the standard deviation and the maximum drawdown has fallen by half.
The return experience is better in a few cases, worse in others. Nifty 50 was the second-best single asset. Nasdaq 100 ETF was the best. Wherever we have mixed Nasdaq 100 to the mix, the returns have improved.
While my knowledge of statistics is limited, I would believe that risk characteristics of multi-asset portfolios are more reliable than return characteristics.
Have the losses gone down?
While we have shown the maximum drawdown for a set of portfolios, a single number does not present the complete picture. A maximum drawdown plot for all the dates gives a better idea.
At the same time, it is not possible to produce maximum drawdown charts for all the portfolio combinations.
I choose 2 portfolios for complete plot.
- 25% Nifty + 25% Nasdaq 100 + 25% Gold + 25% Liquid
- 33% Nifty + 33.33% Nasdaq 100 + 33.33% Gold
Easy to see sharp improvement.
The purpose is served.
- We are likely to have home bias in our portfolios. Hence, in all the portfolios considered, Nifty 50 has the highest allocation (or shares the top spot with other assets).
- We have used the data for a little over 9 years. Clearly, not enough. We could have done with data for longer duration. Motilal Nasdaq ETF was launched only in March 2011.
- The intent should be to look at the risk aspects in the multi-asset portfolios i.e., standard deviation and the maximum drawdown (maximum loss). Focusing on return aspects can be misleading. That will make you load up on Nasdaq 100. That may not be the best approach. The baton of the best-performing asset class keeps changing hands.
- Nasdaq 100 has been a runaway winner during the period under consideration. No other investment (considered) comes even close. Therefore, wherever we have added Nasdaq 100 in any proportion, it has provided a kicker to the returns. Remember, the past performance may not repeat.
- We have used Nifty 50 TRI for domestic equity. Total Returns index (TRI) includes dividends. In real life, if you were to invest even in a Nifty Index fund, there will be some expenses and tracking error. For other assets/sub-assets, I have considered NAVs for various funds schemes. Hence, Gold, Nasdaq and Liquid fund performance is after accounting for expenses. Hence, with this data, I have given advantage to domestic equity (Nifty 50).
- Standard deviations and correlations between two assets (or sub-assets) can keep changing. Note that diversification can sometimes fail you when you need it the most, especially when you mix sub-assets (domestic and international equity in our case). For instance, when the markets are falling, the correlation can increase. For instance, when the markets fell in March, both Nifty 50 and Nasdaq 100 fell very sharply (in sharp contrast to what their correlation would suggest).
- For the ETF used in this analysis (Motilal Nasdaq 100 and Nippon Gold BeES), we have considered the day-end NAV for price performance. Please note you can’t buy ETFs at the day-end NAV. ETFs must be bought in the secondary market. The ETF transactions may happen during the day at prices much different from the day-end NAV.
Never forget this about portfolio diversification
Diversification is NOT about having all your money in the best-performing asset class all the time.
Diversification is about NOT having all your money in the worst-performing asset class at any time.
Mixing different types of assets reduces volatility and drawdowns. However, this does not eliminate risk or losses. For instance, if you mix gold and Nifty in 50:50, the deepest drawdown is still ~20%. Better than Nifty 50 (38%), but no less by any stretch of imagination. Many would panic at that kind of loss. Hence, you still need massive investment discipline.
And you must have a portfolio approach. It is easy to question the utility of debt in the portfolio when equity markets are firing all cylinders. Or it may seem futile to keep gold in the portfolio when it does nothing for a few years. It is expected that all the components won’t do well at the same time. And not do badly at the same time. And that’s how you reduce sharp losses in the portfolio. Do not look at assets in isolation. Look at the portfolio together.
Image Credit: Unsplash