Folks,
Attached is the Nifty trailing 12 month PE (weighted) as at August 41, 2018. Nifty PE Q1 FY19 for valuepickr.xlsx (111.0 KB)

My computation shows a decrease of 4.6% in TTM Nifty EPS from 460 on May 31, 2018 to 439 on August 14, 2018.

Bloomberg on the other hand shows an increase of Nifty EPS to 480 (I presume it is TTM), an increase of 4.3%. However Bloomberg does not give its source of computation. So far as I’ve seen our computation has been in line with Bloomberg, except this time.

I have checked / rechecked / split whatever little hair I have, but could not see an error. I later analysed why this is so even though weighted PAT has gone up. I try to explain this below, but would appreciate if some members can chip in ( say @Yogesh_s and many others who are excel savvy) to validate what I have done or if some can try to source an alternate computation.

Consequently, as per my computations, Nifty PE has gone up from 23.35 end of May to 26.07 as of yesterday. This is again in keeping with the trend with increase in valuation ahead of increase in earnings.
My explanation of why Nifty EPS has gone down vs May 2018:
First, how I compute

We compute Nifty EPS as Nifty value / Nifty PE, Nifty value is the closing value taken from NSE. So Nifty PE is what we really compute.

Nifty PE is computed as (sum of weighted MCap of each constituent stock) divided by (sum of weighted PAT (consolidated) of each constituent stock)

Sum of weighted MCap of each constituent stock = Full MCap of each stock * weight of the stock. Total weights add up to 100. Weights are sourced from here. Market Cap is taken as the average of BSE / NSE from CMIE’s Prowessiq I am subscribed to.

Sum of weighted PAT (consolidated) of each constituent stock = Consolidated PAT of each stock * weights of the stock as sourced

Consolidated PAT (or standalone in case consolidated is not available) is sourced from company filings or Prowessiq, validated with filings. There are assumptions to be made sometime, which are generally not material and these assumptions can be found in my excel sheet. The biggest assumption I had to make was on HDFC which did not declare consolidated results this quarter because of Ind AS transition.

So computation of PE seems like a straight forward arithmetic affair. The excel shows each of the computation as well.
Next, explanation of decrease
Nifty EPS has declined because Nifty weights have shifted (qoq) from companies that added substantially to profits in quarter ended June 2018 vs June 2017 to companies that substantially reduced profits in quarter ended June 2018 vs June 2017. Weights can be thought of as quantities of each stock in a portfolio and shifts in weights among them are akin to selling some and buying ₹ equivalent of some. If the shares of those I sold earn more than the shares of those I buy, the my portfolio EPS will go down.
That’s what has happened with Nifty EPS. Of the 50 stocks, 16 increased their weights (by 4.35 out of 100) and 33 reduced their weights (obviously by the same amount), one remain unchanged. Now the 16 whose weights were increased had profits reduced by ₹ 42,054 million and the 33 whose weights were reduced had profits increased by ₹ 73,121 million. For instance SBI’s weights went up by 0.29 but its profits reduced by ₹ 72,623 million. Reliance’s weights increased by 1.43 but it produced increased profits of just ₹ 3,510 million or 4%. But Dr Reddy’s profits went up 6 times even as its weight dropped. (You may unhide Columns K to P in the sheet attached and see it stock by stock)
This can be better explained (oh really ) if you have say just 2 stocks (A and B) in a portfolio (of 1 unit), of 50 shares each. Say both have equal EPS of 1 and MCap of 500 and 1000. Next year, you sell 25 of stock A, whose MCap remains constant and use that to buy stock B, whose MCap is now risen by 20% to 1,200. Thus you end up buying 20.83 shares of B by selling 25 of A. If during this period A’s profits rose 25%, and B’s falls 10%, then the portfolio earnings per unit has moved down from 100 to 95 (95 = 25 shares1.25 EPS + 70.83 shares.9 EPS). But overall profits have grown up to 152.5 from 150.
P.S: My standalone PE computation matches with that shown in the NSE website (~ 28.02).