You are on page 1of 33

OctoberQuest 2019 notes

● (Please ignore typos)

Prof. Sanjay Bakshi 

Topic- Non ergodicity and


its implications for
businesses and investors 

● Idea about the term


called Ergodicity. This is
a term from physics.
● Ole Peters- has done
remarkable work on
ergodicity (his is a prof)
● Example of the term- a
favourable game of coin
toss where you win 50%
of the wager when you
get heads and you loose
40% of the wager when
it is tails. You play this
game for 1 minute - and
get TTHTH . If you play
this for 60 minutes you
can get multiple
possible trajectories
(many possibilities). If 10
people play the same
game for 1 minute they
too will have multiple
possible trajectories.
You will have a series of
winning and losing
streaks is such bets
(remember the coin has
no memory)
● In money management if
you are up 1000%
followed by down 100%
then life will really suck
● Similar in the coin toss
where there will be
some loosers and some
winners.
● “On average” a 20
sequence coin toss path
will be smoother vs.
each sequence. If we
look at it “on average”
for 1000 sequence we
see that the overall
payout improves
although led by noise
(sharp ups and down
seen in payoff through
the entire sequence) 
● Ensemble average - avg
experience of a bunch
of people (say 1000 or 1
mn) doing the same
thing over and over
again. This may be the
experience of a million
people but not my
experience and I am
more keen on my
performance (canʼt look
at parallel world for me
as I get only one
experience). Hence i am
not interested in
ensemble averages. 
● There is an imp lesson-
even if the aforesaid is a
good game there is
chance of a bad luck
which can take us out of
the game
● If i play the same game
for 24 hrs or 1 week or 1
year vs. earlier 1 hr we
see the noise which is
there is 24 hrs reduce
significantly in 1 week
and reduces completely
in 1 year
● Ensemble perspective-
over emphasises
exceptional success and
relevant to the collective
(“on average”) vs. time
prespective - reflects
what happens to me and
relevant for as an
individual 
● The ensemble
perspective is for
average result (one
could be an outlier and
very lucky and others
may be average or are
loosers) which is not in
line with what happens
in reality 
● “ individuals optimise
what happens to them
over time, not what
happens to them on
average in a collection
of parallel worlds.“Mark
Buchanan 
● How is this useful for
investors - investors too
think about multiple
scenarios and come to
value of the business
(ensemble prespective)
● “You are in the right
place and this is the
right time but I am afraid
you are in the wrong
alternate universe”. This
happens when our
hypothesis goes wrong. 
● “Six people playing
Russian roulette once
each is not the same as
one person playing it six
times”- Naval Ravikanth 
● You have to think about
a chance of a wipeout
when you are playing a
game; you do not take
the risk of ruin no matter
how good the upside 
● Concentrated
portfolios- in a strategy
that entails ruin benefits
never offesets risks of
ruin. It is iritational to
separate risk taking
from the risk
management of ruin-
Taleb 
● The investing game is all
about position sizing
and taking
disproportionate risk
can take you out of the
game 
● Non ergodicity in
business- the entire
NBFC space is seeing
this now 
● If you incur a tiny
probability of ruin as a
“one-off” risk, survive it,
then do it again (another
“one-off” deal), you will
eventually go bust with
a probability of of 100%-
Taleb
● Risk exposure in BFSI
should be though of in
terms of NW instead of
AUM
● Read Buffett essay-
Principles of Insurance
Underwriting 
● In your portfolio
ceaselessly search for
possible correlation
among seemingly
unrelated risks 
● Non ergodicity in
investing- 
● Two most imp word in
investing is that shit
happens. “Condoms
arenʼt completely safe.
My friend was wearing
one but he was hit by a
bus”- Robert Rubin 
● We have moved from 10
stocks to around 20
stocks portfolio 
● “The riskiest moment is
when you are right. That
when you are in the
most trouble, because
you tend to overstay the
good decisions. Once
you have been right for
long enough, you donʼt
even consider reducing
your winning positions
they feel so good” -
Peter Bernstein 
● My largest positions are
not the ones I think I am
not going to make
money. My largest are
one where I think I am
going to loose less
money- Joel Greenblat
● “I view diversification
not only as a survival
strategy but as an
aggressive strategy
because the next
windfall might come
from a surprising place. I
want to make sure I am
exposed to it”- Peter
Bernstein 

Parajit Garg

Topic- Algo Trading and


the future of investing 

● Myths- algo reading (1)


 has to be fully
automated, high
computer and intensive
programming, (2)
carries very high risks or
low risks, (3) is pure
arbitrage, (4) involves
front running, (5) cannot
do as well as value
investing
● What is it- a rules based
objective process to
generate buy and sell
decisions  
● Why it works- (1) less
prone to behavioural
mistakes, (2) back test
ability, (3) ideas are
often scalable and
replicable across
markets and asset
classes, (4)
improvements
accumulate but they
also get arbed away, (5)
speed advantage 
● Strategies- (1)
directional, (2) long/
short, (3) market
making, (4) arbitrage 
● Citadel does massive
volumes and with the
sophistication they do
very well. 
● Ways to segment algos
(1) daily market data, (2)
fundamental data, (3)
macro data , (4) news,
(5) high frequency data
(nse generates 20-25
GB of data everyday),
(6) alternative data
(satellite imagery; there
is a company who does
this. Social media,
weather data, etc....all
data machines can use)
● Putting it together- (1)
choose asset class and
universe, (2) choose
type of strategy and
inputs, (3) hypothesis
an alpha, (4) arrange for
data, (5) backtest, (6)
decide on risk limit, (7)
paper trade, (8) go live 
● A real strategy -
Momentum long/short
strategy which is a very
famous strategy- (1)
universe is large cap
Indian stocks, (2) long/
short strategy, (3) buy
the winners and sell the
loosers, (4) monthly
returns 
● High frequency trading-
stocks held for 10 s to 5
mins. The computer can
operate in a million of a
microsend and can send
order to the exchange.
You are pushing both
the hardware, software
and statistical limit to its
extremes. 
● Risks- (1) operational
(system crashes and
you cant close the
position), (2) regulatory
(Jet Airways ban F&O or
no more co-location is
allowed), (3) financial/
model (LTCM didnʼt
model what happens if
prices remains divergent
for too long), (4) human/
behavioural (humans
worry that models are
taking wrong risks and
close the position and
that could be
behavioural risks).
● Renaissance medallion
fund amasses annual
return of 80% before
fees (5/44) for last 20
odd years. 
● How can we leverage all
the above in investing-
what are the inputs and
what is the holding
period. 
● To screener, codify the
idea and use it in
investing. Best way of
investing vs. human
approach. Also can
make a portfolio via this
approach which
removes `human
subjectivity
significantly. 
● Whatʼs being automated
by most funds globally
to gain additional
advantage- (1) ESG-
assessment of metrics
across all of E, S and G,
(2) accounting
shenanigans, (3) public
perception-using social
media monitoring, etc.
(4) scuttlebutt- analysis
of export-import data,
Nielsen reports, pharma
data reports, litigation,
using online website to
check prices, etc, (5)
management quality-
automated analysis of
earnings calls, annual
reports for management
focus and reliability, (6)
and many more 
● The duality- Dataset
availability and ability to
analyse 
● Tails from the field- firm
named foursquare
predicted Chipotleʼs
earnings two weeks in
advance. The did this
using the check in data
of people. They
accurately predicted
30% decline in earnings.
This data today is
available on subscription
by Quandl. 
● Firm named Quandl is
selling daily estimates of
car sales by model and
manufacturer 
● Firm named Orbital
insights uses daily
photographs of the
entire planet to drive
insights 
● How to get started- (1)
validate your screens,
(2) analyse drivers of
portfolio returns, (3)
codify your decision
rules 

Sameer Kothari- CEO of


Hindustan Foods 

FMCG contract
manufacturing 
● Is a CA and pass out

from an Ivy League


business school 
● Exports - global FMCG

market is 750 bn usd


and US market is 120 bn
usd.
● US imports close to 50

bn usd mainly from


china which is a
opportunity for India
today as we are very
insignificant part of
global exports now and
are a net importer. 
● Who will manufacture all
of this? 
● Make in-house or
outsource- (1) cost of
the product, (2) secrecy
(IP), (3) competency
and realibiity of
suppliers, (4) small
volumes (donʼt want to
create capacity for peak
seasonal sales), (5)
capacity creation 
● Value chain of a FMCG
product 
● Why FMCG conctract
manufacturing is an idea
whose time has come
● Evolution of contract
FMCG manufacturing -
(1) Phase -I - SSI
resevations and tax
exemptions, (2) Phase-2
was area based
reservation (Horlicks
was sent from Baddi
Himachal to South
which is 40% of the
market). Focus had to
be on small, modular
size of the factory which
canʼt gain scale benefit,
(3) Phase-3 - GST,
change in distribution
network and evolution of
e-commerce and
modern trade (small
niche brands gaining
scale without the
distribution network),
Phase-4 - ongoing and
wish list- global
sourcing hub like
pharma generics,
explosion of small
brands who do not want
to invest in
manufacturing factilties,
most product in India
are duopolies or
oligopolies unlike in US/
Europe which has so
many more brands in
each category. Paper
boat created a niche in
the drinks industry and
there will be lot of
smaller brands being
launched. 
● Which large contract
manufacturers have a
gap? (1) project
execution, speed and
scale, brand trust,
vendor development,
decentralisation, QA and
R&D, flexibility,
dependability, product
development 
● Manage working capital
is key- need to work
with negligible WC to
grow. 

Rajeev Agarwal-Founder
of Innoviti 

Chat with the founder of


Innoviti on how payment
systems is evolving in India
● We are good a solving

problems vs.
discovering problem as
Indians. This has to be a
part of the ecosystem
which was missing in
India 
● They partner with

merchants and decide


to share the current
account balance of the
merchant to various
banks. 
● Have a fairly detailed
view based on data that
we generate across
merchant which no one
bank will have (based on
data from 100 cities and
multiple categories etc
that we cater to). We
donʼt provide
transactional data to
anyone but share
curated data based on
what data we have and
what is allowed based
on regulations
● SME lending- created a
product for e-commerce
B2B guys - we have
created a product which
brings together lenders
and wholesalers  (from
whom the SME is
buying). The credit line
of the SME is locked to
the wholesaler. For the
lender the end use and
risks is known and is
better for the entire
econsystem. We have
processed us 1 bn loan
so far and the yields is
19%+ and NIM is 10%+
and NPA is 0.7%. This is
a great product for the
entire value chain
● For a lender, experience
and vintage canʼt
replace any amount of
data. This has to be
looked at in the context
of Indian legal system 
● Payment is a lot about
human behaviour. In a
country like India you
will have multiple type of
payment systems and
technologies that will
evolve. For ex- e-
commerce, modern
trade, kiosk, transit, etc.,
are very different and its
all about human
behaviour. Hence a lot
of payment systems will
co-exist. 
● As consumers will have
3 cards, 3 payment
systems, cash which will
remain and be used.
● Rate of withdrawal of
cash from atm has
flattened out in the last
6 months as per RBI
data.
● Monthly transacting
digital users in India is
only 80 mn (60 mn
some months back) vs.
600 mn total. It will take
a decade for it to
penetrate to tier 3 is a
meaningful way. 
● Whatsapp has 400 mn+
active subscribers in the
country and could use
the data for multiple use
● Payments business may
not be pure payment  to
generate revenue but
may use to get data to
hook it for multiple use,
etc. Things will evolve
over next 5-7 years and
capital for the players
should remain
● His philosophy of
“happy work
culture”- key things
that he asks is that
will company create
pride among
employees and create
a happy work culture?
This has come out of
his past experience.
Pay the employees
the best and improve
their standard of
living and they will do
the best work and be
a pride for the
company. They will be
able to solve the most
complex problems
without worrying
about things which
they should not think
about while working
in our company. 
● Based on our data we
can heurestically
understand the
behaviour of the
consumer, age profile,
his choices, etc. this can
be of help
● RBI data, NPCI data,
data from ministries,
etc., are very useful 
Varun Gupta, promoter of
Ashiana Housing 
● RE going through

consolidation and will


get into fewer hands in
coming years. We will
gain market share 
● Business is driven by

cycle of supply which


has completely stopped
now + money is not
available and hence
consolidation will be
faster 
● Most projects are

getting sold the


maximum is still at
launche stage and
hence a myth that all
buyers only for ready to
move in project 
● We do Jaipur and

Bhiwandi where we have


created niche and
credibility and hence
larger guys wont find it
easy to make headway
here. Also the corporate
overheads of larger
guys is very high which
makes it difficult to
compete with us in
these cities. 
● Pre-tax ROCE of 20% is
what a Real Estate
company should target
in any particular
location.
● Better the location the
real estate company will
make lower return and
vice versa vs. the land
owner guy capturing the
value share. 
● Oberoi made big money
in Goregaon as not
many suitors were there
for that land and they
did high quality project
in the land and attracted
customers. 
● Tier-1 city we have to
play in niche areas like
senior living or outskirts
or kids centric home and
create value 
● Culture of the company-
purpose led
organisation (we have
lot of sweat equity than
financial equity). We
donʼt get a kick on how
are mcap is doing but
how people in our
projects are doing and
living. We have core
values and every
Monday morning we get
executives to share their
exp with what they did
to enchace core values 
● Senior living- the overall
scenario is getting
better. We got it wrong
earlier by launching it in
a younger city. We
launched it is Chennai
recently and it is doing
very well.
● The margins in senior
housing will be better
vs. normal housing. As
%age of revenue this will
be around 20% but
profitabiity share could
be 30-35%
● How did you handle the
downturn? Initially we
didnʼt handle this well at
a personal level and
cursed the Govt, etc.
later we realised no
pointing contemplating
on externalities and we
need to do something
internally which we
started 24 months back
like targeting Gurgaon
customers to Bhiwadi,
making the projects
better suited for
customers, etc.
● Admire HDFC Ltd a lot-
donʼt see any pressure
when you visit their
office but like clockwork
they have done well. In
their simplicity of thing
they have done well.
Donʼt need to be hyper
active to do well
● Didnʼt have any
investing in equity till 2
weeks back. Have only a
house and Ashiana
stock as holding 
● In a 300 crs revenue
project we should have
around 10-15 cr capital
exp at the max
excluding land costs.
You need to take
customer advance and
fund construction. If this
assumption goes wrong
then you need to fund
yourself and then you
do project in phases.
Phase 1 for us is 24-30
months and will have
say around 200-500
units and want to sell
25% in launch and rest
every month equally to
fund the entire project
on its own. Sometimes
you get good launch
and some take times to
sell.
● Commerical- we donʼt
know this well and
hence have not got into
it. We will do things that
we know well and pass
the rest that is our
philosophy  
● Situation on time and
cost of approval has not
changed much despite
all that is heard. EC
approval in India is a
completely broken
system- approval has to
be easy but compliance
should be stricter which
isnʼt happening. 

You might also like