Do we need to time the market?

Last Updated on February 12, 2022 at 6:20 pm

For a while now, I have been wanting to do a full back-test of multiple market timing or tactical asset allocation strategies. I used a speaking invitation as an excuse to create a spreadsheet where multiple models can be quickly tested with 1000s of time periods.  While at it, I asked myself, do we need to time the market? I also recognised that question is different from, can we time the market?

The answer to, “can we time the market?” is a vehement yes, but we should recognise that market timing strategies have one common goal – reduce risk. So yes, we can time the market to reduce risk in portfolios. This risk reduction may or may not be associated with return enhancement. It all depends on the method chosen, the associated costs and tax.  Read more:

Want to time the market with Nifty PE? Learn from Franklin Dynamic PE Fund

Is it possible to time the market?

Axis Dynamic Equity Fund: Investment Strategy Analysis

Deconstructing the Motilal Oswal Value Index (MOVI)

Dynamic Asset Allocation Mutual Funds: Yield Gap vs. P/E Ratio

In addition to these,   I have also analysed tactical “dip” buying.

Nifty 200 DMA: Buying High vs Buying Low

Equity: Buying “High” vs Buying “Low”

Buying “low” vs Buying “systematically”: Surprise, Surprise!

Buying “low” with “active” cash vs buying systematically: still a surprise!

I am now building up to a series of tactical asset allocation back-tests with equity allocation swinging from 100% to 0%. And the results are pretty interesting. I am still thinking about how to present these results as they meant for the eyes of discerning DIY investors and can confuse newbie investors.

So I would like to first answer in this post: Do we need to time the market? Will I miss anything if I choose to avoid timing? The answer is: No, we do not need to time the market. No, we will not miss anything, BUT we need a proper strategy to systematically reduce risk in our portfolio.

This systematic risk reduction need not depend on any market signals. All we need is a clear goal. A clear date when we need the money and a clear target corpus. I had discussed this strategy in this post. Please read this in full as it will help you understand the context: How to reduce risk in an investment portfolio

Resolve is a series of steps on investing and portfolio management.  In the first step, we considered how to quickly select equity mutual funds and build a diversified (equity) portfolio. As a second step, we discussed how to quickly decide if I should stay invested in a mutual fund or exit it. Now, in the third step, we consider goal based risk management.

Let us consider a goal that is 15 years away. The current cost of that goal is 10 lakh. Assuming 10% yearly inflation, the future cost is shown below (blue dots). If we start investing for this in a mix of equity and debt (fixed income), the corpus assuming same returns (10% from equity and 7% from debt – both post-tax), the total corpus will be shown as below. The investment is assumed to increase by 5% each year.

Naturally, the aim should be for corpus (red line) to hit the target (blue line) on or before the end of the 15Y period. We start from 0, so it ends at 14. When the returns are fixed, you get a nice smooth corpus growth. Returns in real life fluctuate wildly as you will see below. This is known as sequence of returns.

Now, there are two way to invest in equity and debt. Use the same asset allocation for each year. Say 60% equity and 40% debt. We will refer to this as constant equity allocation

Or we can reduce equity gradually as we approach the goal. I prefer a step-wise decrease (on paper at least) and this is the approach recommended in the Freefincal Robo Advisory Software Template. We will refer to this as the reducing equity allocation.

To these asset allocation strategies, we need real market returns. So I have used past Sensex annual returns.

To do this we take Sensex closing price. For the preliminary results shown here, I have used data from Jan 1991 to Jan 2018. This is the first return sequence considered

So we use: 121%, -26%, 96%, -16%, 1%, 17%, -24%, 45%, -9%, -24%, -9%, 30%, 26%, 53%, 51% as the sequence of annual returns for our portfolio.

Then we change the first date in the above table (marked in red) to the next business day to create the next sequence and so on. Here is another example:

So we keep repeating this to create return sequences and this is the last sequence considered.

This gives us a total of 2670 return sequences to test. In the above linked post, I had presented a few examples. For: -18%, -12%, 27%, -27%, 52%, -13%, -22%, -3%, 69%, 23%, 43%, 54%, 35%, -55%, 86%. Each of these is a return after one of year of investing. So 15 annual returns for 15 years of investing.

Constant Equity method

Decreasing Equity method

Now the question is, how do the results look if we backtest for all 2670 return sequences?

Constant Equity method (60% for all 15 years)

Total test runs: 2670

No of runs in which final portfolio value was equal to or above target value: 2670 (100%)

Decreasing Equity method (stepwise reduction)

Total test runs: 2670

No of runs in which final portfolio value was equal to or above target value: 2417 (91%)

No of runs in which final portfolio was equal to or above 90% of the target value: 2670 (100%). Full results will follow in a separate post.

Now, that is mighty impressive. Please note that the investment amount used for the decreasing equity method was the same as the constant equity method (the robo template accounts for this correctly). That is, it was lesser than necessary (when equity is fixed at 60% you need to invest less). So even then the result is remarkable.

So do you need to time the market? You can time the market if you wanted to, but there is no need. All you need to so is simple Goal Based Risk Managment to get you home.

UPDATE: Why we need to gradually pull out of equity investments well before we need the money!