A little over a year ago, I published predicting Indian stock market returns based on historical analysis of the NIFTY and the CNX500. I did this with tongue firmly in cheek, given the very concept of value investing accepts inherent unpredictability of markets, especially in the short-term (anything less than five years) as laid out in my calm investing principles.
“Anyone who says they know where the market will be a week / month / year from now is guessing (or has super powers)…“
Since I am yet to find evidence of my super powers, except for my ability to order the best thing on any restaurant menu, this “prediction” exercise is academic more than for any real investment decision-making. Now keep in mind, while the use of words like “predictive modeling” can stir up images like these…
the analysis I did is closer to…
I used 15 years of Nifty Price-Earnings and Price data to determine the relationship between the two. On paper, what the model does is to answer the question:
Given the NIFTY level and PE on any give date, what is the predicted Nifty level one year from that date?
With that background, since the NIFTY has gone from almost scaling 9,000 to barely holding on to 7,000 in the course of a year, I reopened my so-called model from Jan 2015 to see how it did.
- Predicted NIFTY levels much less volatile than reality. Mostly because of my overly simplistic assumption of current PE driving future levels
- Post-election surge in Apr 2014 up to Feb 2015 left predicted NIFTY performance far behind until the worm turned about this time last year
- In Mar 2015, Nifty PE hit 24x (for every ₹ 1 in annual earnings, you paid ₹ 24). Only 5% of the time in the last 15 years, has Nifty PE been at or above 24. Consequence: it didn’t stay long at that level
Even though the model is all but useless in predicting even approximate levels on any given day, it is mildly useful directionally (maybe?), given current PE levels to see what might be in store.
So what does the model say for NIFTY levels up to Feb 2017? (since this one looks at 1 year returns, it can only predict up to year in the future, similar models could be built to predict three and five year returns but then historic data available would be a limitation)
NIFTY 2016 prediction: Expect a fair amount of short-term volatility followed by a gradual correction, but no real fireworks. NIFTY to be about the 8,000 mark by Feb 2017.
Academic exercise completed, let’s look at why all of the above should be treated as nothing but a short entertaining read and nothing more:
- On an average, the error between predicted and actual values was anything between +20% to -15% of the predicted value, so really the NIFTY could be anywhere between 6,800 and 10,000 based on past performance. If anyone asked me to place a bet on my prediction, I wouldn’t
- The 2016 chart up to Feb 19th shows the predicted value to be 7,630 while the NIFTY is actually at 7,210, a good 400 points off. That seems to be a consistent over-estimation since Nov 2015 before which it was under-estimating the NIFTY
- This is an overly simplistic model which looks at just PE as the input. For the statistically inclined, the R-squared of the model is 53%, meaning the PE on any given day only explains half of the NIFTY level. That’s like saying how warm you feel is explained by how many layers of clothes you have on without considering whether you are sitting in an air-conditioned office or trekking in the Sahara desert
- Finally, from my post on the futility of forecasting: As of late 2014, all the big financial powerhouses were forecasting NIFTY at 10,000 by Dec 2015. By April 2015, those same banks had cut forecasts by 15 – 20%. And this from folks who have models if compared against mine, would be akin to comparing
So if all the “modeling” and “prediction” is useless, what does an investor do? Same as one always does.
Here’s what I’m doing: I am optimistic about the Indian economy as a whole and believe that at the end of the day, if the companies you invest in show consistent earnings growth not withstanding the odd blip, stock prices will move up, and more than most other asset classes. My approach is to keep investing regularly in a set of identified stocks that fulfill sound investment criteria, pass a set of hurdles knowing that what I can control is the process and let the talking heads on financial media worry about short-term movements.
Keep calm and carry on investing.