In the attached file that you posted, the ASX300 is shown as P/E of 21 at 1st June 2014. What source are you using?

The Australian Financial Review 31/5/14 reports the S&P/ASX300 PE at 27/5/14 as 16.06. Now I think that this P/E does not include loss-making companies which makes it a bit unusual compared to international norms of calculating P/Es. with the ASX300 Index at 5451

They report the ASX200 P/E being 16.09 with ASX200 at 5511 - both on the 27/5/14.
The RBA offers a May P/E for the ASX200 for "May" of 17.1 when the ASX200 was 5493. I seem to recall from somewhere the RBA get their P/E data from the source Morgan & Stanley Indices, but I am not sure about that.

There seems to be a modest inconsistency between the P/E reported by RBA and Australian financial review, but the discrepancy between the ASX300 P/E and these other 2 is very large. 

Can you through any light on this? Such a large discrepancy throws some confusion into how expensive the Australian market is.

And 

Yes, I can confirm that RBA does source its P/Es from MSCI - it says so at the bottom of the RBA F7 document. 

My guess had been that the discrepancy between RBA data and S&P data was that S&P ASX 300 was excluding loss-makers.

?But the P/E that you have provided possibly throws my theory out the window.

Eoin Treacy's view

Thank you for this detailed question. The complexity of what should and should not be included in the calculation of financial ratios makes me glad that we focus on price as an absolute arbiter of the actions of investors and traders. All of our data originates in Bloomberg and they rely exclusively on operating earnings for their calculations. How appropriate operating earnings are is an additional topic of conversation which was particularly relevant in the aftermath of the credit crisis when temporary dislocations in earnings and liabilities caused P/Es to spike higher. 

In order get a definitive answer to your question I took the Danish KFX Index because it has 20 constituents which allows us to make some calculations without over complicating the issue. Bloomberg reports a P/E for the Index of 22.12. 

Nordea Bank had negative earnings in the last quarter which means Bloomberg does not report a value for the P/E. However the share represents 13.88% of the Index’s market cap. If we weight the P/Es by their respective market caps we get a sum value of 22.24 which is pretty close to the value reported by Bloomberg. If I include estimated P/Es for Nordea Bank and FLSmidth we get a value of 24.07. A number of indices are weighted using a market cap and free float method which further complicated the issue. 

However we can deduce that Index providers are taking a zero value for companies with negative earnings in their calculation, probably because this is the most expedient method. In the valuation tables we post, since the same calculation is used for every index it is probably better to look at their relative position on the table rather than focus on the specific number posted.