Ibovespa P/E

I'm sick and tired of hearing those dumbs and biased analysts (or should I say experts?) claiming that Ibovespa is cheap, because its P/E currently is 9, or because its P/E is most attractive in relation to all emergents, blah, blah, blah...
So I decided to calculate myself the Ibovespa P/E. Here is my methodology:

• I've selected 231 companies to calculate the median P/E for the last 12 and 20 quarters. I prefer the median estimator, because it often better expresses the common-run, since it is not, as is the mean, affected by an excessively high or low figure, like a extraordinary result in determined quarter.
• For that group, I've excluded 31 companies that had median losses either for the last 3 years or 5 years, remaining 200 companies.
• For 5 years calculation, only 140 companies had enough quarters.
• I also managed to compute the current yield.
• All that median figures was then ranked by market value and liquidity. I don't give a shit for liquidity, but as it is too loved by most market "investors", it's interesting to calculate it anyway.
• All the data are of April 5, when Ibovespa was closer to 70k points.

The results are shown below:

• Current Yield: the current yield is just of 3.0% and 2.5%, when averaged by market value and liquidity, respectively.
• 3 years P/E: the median P/E for the first 50 companies, ranked by market value and liquidity, respectively, was 17.7 and 23.9.
• 5 years P/E: the median P/E for the first 50 companies, ranked by market value and liquidity, respectively, was 16.9 and 20.6.

And here goes my conclusions:

• First of all, it's straightforward to note that when the metrics was ranked by liquidity, it turned out to become more expensive, clearly denoting that when something is in fashion, you have to run away.
• I think it's not an evident bubble, but sure it's not also a bargain. Compare this figures with the Graham-Dodd 10 year P/E ratio of about 24 and the much lower US interest rate. (I didn't adjust by inflation, but considering an average of 5% per year, it would increase earnings in about only 10% in the average).
• The current yield is very low. Some would say that's because of growth expectation. It would be arguable that 3% yield + 5%  inflation + 5%  growth, for example, would equal 13% prefixed bond available now. It may sounds reasonable. But than I ask: ok, but what would happens if the prices double? Yield would go to 1.5% and it still would be arguable that 1.5% yield + 6%  inflation + 5.5%  growth would equal 13%. Similarly, if yield were zero, we could be told to expect 6% inflation plus 7% growth. So it's all analysts bullshit.

The moral of story: Ibovespa isn't cheap and you have to check everything is told to you. End of story.