Should we hold cash, or until when?

[roadtrippin69]

Since the past year, there have been increasing talks about impending recession, and the biggest question that I ask myself is should I hold cash, or continue my SIPs!

I decided to run a little experiment today, and see how this decision has fared in the past. I know, that past performance is not an indicator of the future, but I just wanted to check how this has varied for my investment horizon in the recent past.

I took the NIFTY 50 data from Jan 01, 1999 to Jan 01, 2020. The basic assumption was that you get a fixed amount of money to invest every month on the first date of the month (Rs. 1000), and the minimum units that you can buy is 0.01. I tried two basic strategies:

• You invest as soon as you have cash.
• You invest only when the index is at least x% from the all-time high.

These are the basic results that I could get:

Strategy
Final Holding
Max Cash Deployed
Max Cash Held Deploy Date

Buy ASAP
11,33,700.11
1,108.96
2018-09-03

Buy on or below (5%)
11,27,949.68
13,054.54
2018-02-06

Buy on or below (10%)
11,32,815.26
20,025.82
2015-05-07

Buy on or below (15%)
10,95,813.36
24,047.62
2015-09-07

Buy on or below (20%)
10,80,350.19
43,004.86
2016-02-11

Buy on or below (25%)
10,65,730.86
24,013.73
2006-06-08

Buy on or below (50%)
9,52,183.34
85002.78
2008-10-17

Here is the code (if you'd like to fiddle). I am new to python, so it is absolutely possible that I messed up in the code somewhere. If you find any issues, let me know, and I'd be willing to run the experiment again.

Edit:
I'm thinking of writing a blog entry with detailed graphs and other strategies. Would that be something that you guys might be interested in?

 

[trentak]

@roadtrippin69 Does idle cash earn FD or liquid fund returns in your backtest?

We ran an extreme version of the same with idle cash returns here - https://kuvera.in/blog/sip-beats-trying-to-predict-market-bottoms/

Our take, timing the market is incredible hard. So invest as soon as you can and hold for as Long as you can (asset allocation helps here).

 

[roadtrippin69]

@trentak The current version does not include returns on idle cash. Another data point to take into account while doing that is tax implications while moving cash out of bank account.

I'll try and incorporate those in the backtest!

 

[montse811]

@roadtrippin69

is tax implications while moving cash out of bank account.

What tax implications while moving cash of a bank acct?

 

[roadtrippin69]

@montse811 In the current model, you never cash out your holdings. So, there are no tax implications. But, if you consider bank interests, or FD returns on idle cash, you need to factor in taxes on those interests too.

For, example, you put Rs. 100 in a bank in a year, you'd get 106 at the end of the year. But, you can't take the entire 106 and put it into the index. You need to deduct taxes on the RS. 6 interest you earned.

Obviously you could directly use real, tax adjusted returns, but that depends on your tax bracket.

 

[rocha]

@roadtrippin69 FD interest is taxable in the year it is accrued.

 

[roadtrippin69]

@rocha Makes the model more complicated. I am thinking for tweaking it slightly, and using a fixed return rate for days cash is held between deployments. Might not be very accurate, but closer to reality.

 

[how2101]

@roadtrippin69 Definitely interested in a blog explaining this with visuals

 

[yuliza]

@roadtrippin69 Even god can't beat dollar average

Good to see this is applicable even in the Indian context.

 

[phouglas]

@roadtrippin69 I feel the best way to compare would have been to try multiple start dates with a fixed tenure. This analysis is very sensitive to the start date which is Jan 01, 1999.

 

[roadtrippin69]

@phouglas Good point! I'll try and do that!

 

[heliumskylark]

@phouglas The golden bull run from 2003 to 2007 does skew up a lot of data analysis. One has to keep in mind that a period like that may never return while investing for the future.

 

[heartcall]

@roadtrippin69 Why would you ever stop SIPs? Defeats it's whole purpose. The only thing you should do is change the equity:debt allocation.

 

[roadtrippin69]

@heartcall I have a source of income that I only put into N50, and NN50. That is money that I'm willing to just grow for the next 10 years or more, so I have no debt component for that. So, the point was to whether hold this component until the index recedes.

Edit: The reason was as I mentioned growing gloom across different subreddits, and the experiment was to see how this fares.

 

[heartcall]

@roadtrippin69 Point still stands. If your horizon is 10+ years then highs and lows should even out. SIPs are meant to be continued as long as you can afford it - ideally you tweak the equity:debt ratio as the market changes, but you never stop SIPs.

 

[roadtrippin69]

@heartcall Completely makes sense!

 

[resjudicata]

@heartcall This is wrong. Risk does not reduce over long term: https://freefincal.com/stock-market-investment-risk-long-term/

Negative returns are possible up to two decades of investing.

 

[rosebuds1111]

@roadtrippin69 Well this seems to be a good analysis. But I would seriously recommend this with an asset allocation portfolio which is balanced every year.

To get analysis how much asset allocation helps in booking profits as well as drawdowns.

 

[hosnloem]

@roadtrippin69 The one thing I did find was that your logic looks like this - "If the % fall is greater than my threshold %, then buy immediately. This buys at the current close price."

But the % fall was already calculated using the day's close price (this means the market has closed), so you won't be able to buy after the market closes. So, shouldn't you buy it at the next day's close price?

 

[roadtrippin69]

@hosnloem The classic look-ahead bias! I'll try and run with the buy tomorrow correction.