long thread on where to look for excess returns in #nifty50 india index , data from 31 jan 2008 1/n
plain nifty returns , 701bps CAGR !! 
top 10 marcap weighted portfolio re-balanced at each year end , at 2008 31 st jan , the portfolio is
returns for for top 10 marcap weight portfolio at 798 bps , or about 1% excess returns per annum , call that as big becoming bigger !!
was always wondering , if those top 10 were dowjonesed , an example port at start of 2017 would look like ->
i will explain what is dowjonesing , the one that you do, in your watch list , whether it is KSERASERA or MRF you always buy 100 shares right ? thats how dow jones index is constructed
ok the top 10 cagr by dowjonesing would give you 664 bps , slightly underperforming nifty ->
the 10 marcap equiweighted at 889 bps , thats almost 2 percent extra than nifty returns !! ->
chasing the beta , that is high beta gets higher weight , an example 2017 end portfolio is ->
top 10 beta weight cagr since 2008 stands at 745 bps , slight 30 bps out performance against nifty ->
the nicer guys at nse takes the pain to publish r2 values as well , which is nothing but correlation multiplied by correlation , or by doing =SQRT(R2) one will get correlation value of a stock , for ex , on 29 dec 2017 the hdfcbank r2 was at 0.24 ,which means with nifty
it has a correlation of SQRT(0.24) = 0.49 , now basically correl can range from -1 to 1 , 1 indicates that the two variables are moving in unison. They rise and fall together and have perfect correlation. -1 means that the two variables are in perfect opposites
this is one thing i couldn't explain to either my mum or a 5 year old kid , may be one day i will master how to explain what is r2 to my mum in heaven and that 5 yr old kid !!
anyway for simplicity sake , if we pick up how r2 stocks we move more or less like nifty , and if we pick low r2 stocks our move will be bit different to that of nifty moves ..
an example of R2 weighted portfolio at end of 2017 looks like
yes a portfolio with R2 weight gives more or less similar returns in over 10 yr period , port returns are at 705 bps vs nifty returns at 701 bps !! ->
a volatility weighted portfolio at end of 2017 -> higher vol gets high weight , and vice versa ->
chasing volatility since 2008 , not a bad idea cagr at 882 bps vs 701 bps of nifty ->
now let me explain inversion , for a given value of x , if we invert using a mirror we get x , or literally write x and turn it upside down , we still get x , but we also a tight slap from our math teacher for such answers
, and our sat score will be zero , and its a different matter that we still get admission into harvard like hoggy boy !!
lets do plain mathematical inversion of instead of getting slaps from our math teacher, for a given value of x , 1/x is the inverted value , for ex : lets to do an inverted beta portfolio , that is take beta and do 1/beta and construct portfolio , for ex 2017 port looks like ->
wow !! inverted beta portfolio cagr is at 1035 bps vs 701 of nifty returns ->
now i dont have 100's phd's working for me , and am not after poor pension funds money , and neither am a charlatan like @CliffordAsness , but can write optimize equation on just those above 4 variables of marcap , beta , r2 , volatility !!
basically that optimized is " inverted marcap weight + inverted beta + inverted r2 ( that is low correl gets high weight ) + volatility weighted " ->
the problem you might have noticed , it gets highly concentrated , and thats the case as well in prior cases !!
anyway , lets call that as our magic formula , here the whopping cagr of 1330 bps vs 7.01 of nifty ->
you can get bit additional alpha by shorting equal amount of nifty futures , i will leave that exercise to the reader what are additional returns by doing so !! 2) or if you are not into these arb , i would say talk to your broker and give all your holdings to SLB mechanism
theoretically it should add an extra 300 bps to your portfolio !!
slight mistake in that out put , picked up wrong headings , here the corrected one ->
