S WAYAM MOOC · IIT KANPUR · PROF.

AB HINAVA TRIPATHI

Advanced Financial Instruments

for Sustainable Business &
Decentralised Markets
QUICK REVISION NOTES · 12 WEEKS · 36 LECTURES

W1–2 Basics & R W3 Risk & Return W4 Portfolio Theory

W5 Carbon Markets W6 Market Connectedness W7 Time Series

W8 Volatility Models W9 Portfolio Optimisation W10 APT & CAPM

W11 Crypto & Blockchain W12 ESG & SRI

── WEEKS 1 – 2 ──

W1–2 Course Overview & R Programming Basics ▾

C O U R S E OV E R V I E W

› Three core instrument classes: Carbon Markets, Digital Currencies

(Crypto/CBDC), ESG Funds

› Motivation: Kyoto Protocol (1997), Paris Agreement (2015), UN sustainability

goals

› Modern trading executed in microseconds via ML algorithms

› Risk & return framework + portfolio construction + financial risk management

R P R O G R A M M I N G E SS E N T I A L S

› R = open-source statistical computing language; works on Windows, macOS,

Linux

› RStudio = IDE for R; four panes: Console, Script, Environment, Plots

› Key data types: vectors, data frames, lists, matrices

› Arithmetic ops: +, -, *, /, ^ | Logical: &, |, !, ==, !=, <, >

› Install packages: [Link]("name"); load: library(name)

› User-defined functions: my_fn <- function(x){ ... }

› Key packages for course: PerformanceAnalytics, PortfolioAnalytics, quantmod,

rugarch

── WEEK 3 ──

W3 Risk, Return & Utility Theory ▾

UTILITY & RISK TYPES HIGH YIELD

UTILITY
TYPE U''(W) KEY PROPERTY
CURVE

Risk Concave <0 Needs risk premium to gamble; CE <

Averse E(W)

Risk Linear =0 U(E(W)) = E(U(W)); indifferent to risk

Neutral

Risk Convex >0 Accepts gamble even at lower

Loving expected value

KEY CONCEPTS

› Certainty Equivalent (CE): sure amount making investor indifferent to a

gamble

› Risk Premium = E(W) − CE (positive for risk-averse; they pay to avoid risk)

› Non-satiation: more wealth always preferred (higher return → higher utility)

› Risk-averse: diminishing marginal utility of wealth

Risk Averse condition:

U[E(W)] > E[U(W)]

MARKET MICROSTRUCTURE BASICS

› Quote-driven (Broker-dealer) markets: dealers post bid/ask, earn spread

› Spread = Ask − Bid = dealer profit = proxy for liquidity

› Liquidity: ability to trade large volume quickly with minimal price impact and
cost
› Deep & liquid market → low spread; illiquid market → high spread

── WEEK 4 ──

W4 Markowitz Portfolio Theory & Efficient Frontier ▾

T W O - A SS E T P O RT F O L I O HIGH YIELD

Expected Portfolio Return:

̄ ̄ ₁ + w₂R
Rₚ = w₁R ̄₂

Portfolio Variance:

σ²ₚ = w₁²σ₁² + w₂²σ₂² + 2w₁w₂σ₁σ₂ρ₁₂

C O R R E L AT I O N & D I V E R S I F I C AT I O N

Ρ₁₂ DIVERSIFICATION PORTFOLIO RISK

+1 None σₚ = w₁σ₁ + w₂σ₂ (max)

0 Moderate Intermediate (concave curve)

−1 Maximum (theoretical) Can reach σₚ = 0

› All feasible portfolios lie between ρ=+1 (blue) and ρ=−1 (black) curves

› With 15–20 stocks: substantial specific risk eliminated

EFFICIENT FRONTIER

› Feasible Region: egg-cut shape; all possible risk-return combinations

› Efficient Frontier (SS'): upper-left boundary of feasible region; max return

for given risk OR min risk for given return

› Global Minimum Variance Portfolio (GMVP): leftmost point S

› Maximum Return Portfolio: topmost point S'

› With short sales: frontier extends beyond S' upward (no upper bound on
returns)

R I S K- F R E E A SS E T & C M L

› Adding risk-free asset (RF): investors choose on the line from RF tangent to
efficient frontier
› Tangency Portfolio G: point where line from RF touches frontier at highest
slope

› Sharpe Ratio = (R̄ − RF) / σ ; maximised at tangency point G

› Separation Theorem / Two-Fund Theorem: all optimal portfolios = RF +

Portfolio G (regardless of risk preference)

› Risk-averse investors: lend at RF (invest partly in RF + G)

› Risk-lovers: borrow at RF and invest more in G

Exam tip: Convex shapes of efficient frontier are NOT possible above GMVP. Only
concave shapes are valid above GMVP; convex below GMVP.

── WEEK 5 ──

W5 Climate Finance & Carbon Markets ▾

C L I M AT E C H A N G E B A S I C S CONCEPT

› Climate change: long-term shifts in temp/weather; mainly human-driven

(fossil fuels → GHG emissions)

› UNFCCC: basis for international climate negotiations; 197 states

› Kyoto Protocol (1997) & Paris Agreement (2015): key milestones

› Climate Finance: attracts investment toward green/renewable tech by

pricing carbon scarcity

CARBON PRICING HIGH YIELD

› Abatement Cost: total cost (investment + operating cost − savings) ÷ tons

of CO₂ avoided = $ per ton of carbon not emitted

› Direct Pricing: ETS (cap & trade) + Carbon Taxes

› Indirect Pricing: fuel excise taxes, fossil fuel subsidies (not directly linked to
emission volume)

E T S V S C A R B O N TA X HIGH YIELD

PARAMETER ETS (CAP & TRADE) CARBON TAX

Price Uncertain (market-driven) Fixed/certain

High (cap = upper limit) Low (hard to predict)

 Emission
certainty

Cost Higher (cross-entity trading) Lower (uniform tax)

effectiveness

Ease of admin Complex (needs market infra) Simple (existing tax

system)

Price Volatile Stable

predictability

Revenue 2022 69% of $95bn global carbon 31% of $95bn

revenue

HOW ETS WORKS

› Regulator sets aggregate cap on emissions → issues tradeable allowances

› 1 allowance = right to emit 1 ton of CO₂

› Low-emitters have surplus → sell to high-emitters → market price emerges

› Stringent cap → fewer allowances → higher price → stronger incentive to

reduce

› ETS = quantity-based policy; guarantees environmental outcome (the cap)

CARBON PRICE DETERMINANTS

› Energy prices: rising fossil fuel prices → less consumption → lower emissions
→ lower carbon price

› Economic activity: higher GDP → more emissions → higher carbon price

› Weather: unexpected cold/heat → more energy demand → higher carbon

price

› Policy uncertainty → price volatility (via 3 channels: technology innovation,

energy effect, investment/output effect)

── WEEK 6 ──

W6 Carbon Market Connectedness ▾

WHY MARKETS ARE CONNECTED

› Economic Fundamentals Hypothesis: interconnected economic activity

links markets
› Financialisation of Commodities: cross-border trade increases global
market interaction

› Market Contagion Hypothesis: extreme events (e.g. COVID, debt crisis)

spread via investor psychology + herd behaviour

CARBON ↔ OTHER MARKETS

› Stock Markets: Higher economic activity → more emissions → demand for

allowances → higher carbon price; bidirectional spillover

› Cryptocurrency: Bitcoin annual carbon footprint ≈ New Zealand's; energy-

intensive mining → higher emissions → demand for carbon allowances; BTC =
carbon-intensive market

› Green Bonds: Green bonds finance low-carbon projects → reduce emissions

→ lower allowance demand → lower carbon price. Green bonds = complement
to carbon markets (low prices = short-term hedge); can substitute at high
prices (long-term hedge)

› Energy Markets: Fossil fuel price ↑ → firms shift to cleaner fuels → lower
emissions → lower carbon price (and vice versa)

── WEEK 7 ──

W7 Time Series: Stationarity, ARIMA Models ▾

S TAT I O N A R I T Y HIGH YIELD

› Strictly Stationary: joint distribution unchanged over time

› Weakly/Covariance Stationary (practically used): constant mean, constant

variance σ², autocovariance depends only on lag not time

› Non-stationary problems: shocks don't die away; spurious regressions

(high R² and t-stats even for unrelated variables)

WHITE NOISE & ACF

› White Noise (IID): E(μₜ) = 0, Var = σ² (constant), zero autocorrelation at all

lags

› Residuals from a correctly specified model should be white noise

› ACF (Autocorrelation Function / Correlogram): plot of autocorrelation τₛ

at lags s=0,1,2…; τ₀ = 1 always
› PACF (Partial ACF): correlation between yₜ and yₜ₋ₛ after removing
intermediate lags

AR, MA, ARIMA HIGH YIELD

› AR(p): current value depends on its own p past values; ACF decays gradually;
PACF cuts off at lag p

› MA(q): current value depends on q past error terms; ACF cuts off at lag q;
PACF decays gradually

› ARIMA(p,d,q): p=AR order, d=differencing to achieve stationarity, q=MA

order

› Why univariate? High-frequency trading data lacks macro variables (GDP,

policy not available intraday)

› Despite being atheoretical, ARIMA often outperforms structural models in out-

of-sample prediction

AR(p):

yₜ = μ + φ₁yₜ₋₁ + φ₂yₜ₋₂ + ... + φₚyₜ₋ₚ + εₜ

── WEEK 8 ──

W8 Volatility Models: EWMA, ARCH, GARCH ▾

V O L AT I L I T Y T Y P E S

› Historical Volatility: equal weight to all past observations — not ideal

(sudden drop when old obs excluded)

› Implied Volatility: backed out from option pricing models (e.g. Black-
Scholes); = market's forecast of future volatility

› Conditional Volatility: variance conditional on past information — gives

more weight to recent shocks

EWMA HIGH YIELD

Exponentially Weighted Moving Average:

σ²ₜ = λ·σ²ₜ₋₁ + (1−λ)·μ²ₜ₋₁ where 0 < λ < 1 (e.g. 0.9)

› λ close to 0 → all weight on recent data (noisy estimate)

› λ close to 1 → mostly historical data (less timely)

› Impact of old observations decays exponentially (no sudden drop)

ARCH(Q) MODEL

ARCH(q):

σ²ₜ = α₀ + α₁μ²ₜ₋₁ + α₂μ²ₜ₋₂ + ... + αqμ²ₜ₋q

› Captures volatility clustering: calm periods cluster, volatile periods cluster

› All αᵢ ≥ 0 (constraint for positive variance)

› Limitation: non-parsimonious for large q; risk of negative coefficients

GARCH(1,1) HIGH YIELD

GARCH(1,1):

σ²ₜ = α₀ + α₁μ²ₜ₋₁ + β₁σ²ₜ₋₁

› α₀: drives long-run mean reversion

› α₁: weight on recent shock (ARCH effect)

› β₁: weight on previous variance (persistence)

› α₁ + β₁ < 1 for stationarity; α₁ + β₁ ≈ 1 → high persistence (common in fin.

markets)

› More parsimonious than ARCH(q)

── WEEK 9 ──

W9 Portfolio Optimisation in R (PortfolioAnalytics) ▾

P O RT F O L I O C O N S T R U C T I O N W O R K F LO W

› Assets in course portfolio: EUA Carbon Futures, Bitcoin, ESG Fund, S&P
500

› Steps: create portfolio spec → add constraints → add objectives → optimise

› Functions: [Link](), [Link](), [Link](),

[Link]()

CONSTRAINTS

› Weight Sum: min 0.99, max 1.01 (fully invested)

› Box Constraint: min/max weight per asset (e.g. 5% min, 90% max)
› Group Constraint: min/max weight for groups of assets

O B J E C T I V E S & S O LV E R S HIGH YIELD

› Maximize Return: type="return", name="mean" → heavily weights highest

return asset

› Minimize Risk (StdDev): type="risk", name="StdDev"

› Minimize Expected Shortfall (ES/CVaR): type="risk", name="ES" (default

95% confidence)

› Maximize Sharpe Ratio: add both return and risk objectives; use maxSR=TRUE

› Solvers: ROI (deterministic), DEoptim (differential evolution ~2000 portfolios),

random (random portfolio generation)

RISK MEASURES

› Standard Deviation: common symmetric risk measure

› Expected Shortfall (ES / CVaR): average loss beyond VaR threshold; better
captures tail risk

› Sharpe Ratio: (Return − RF) / σ; higher = better risk-adjusted performance

── WEEK 10 ──

W10 Asset Pricing: CAPM & APT ▾

CAPM HIGH YIELD

Capital Asset Pricing Model:

̄ ̄ ₘ − RF)
Rᵢ = RF + βᵢ(R

› β (Beta): sensitivity of security to market; systematic risk measure

› (R̄ₘ − RF) = Market Risk Premium

› βᵢ > 1: amplifies market moves; βᵢ < 1: less sensitive

A P T ( A R B I T R A G E P R I C I N G T H E O RY ) HIGH YIELD

APT Multi-factor Equilibrium:

̄
Rᵢ = RF + λ₁bᵢ₁ + λ₂bᵢ₂ + ... + λⱼbᵢⱼ

› λⱼ = price of risk for factor j (≈ risk premium of portfolio with sensitivity=1 to

factor j, 0 to others)
› APT uses arbitrage argument (unlike CAPM's mean-variance framework)

› Return generating process (multi-index): Rᵢ = αᵢ + Bᵢ₁I₁ + Bᵢ₂I₂ + … + eᵢ

› If CAPM holds for index portfolios (I₁, I₂…), APT reduces to CAPM form — APT
and CAPM are consistent

A P P L I C AT I O N S

› Passive Management: construct small portfolio tracking index (e.g. Nifty

50) using APT factor sensitivities

› Active Management: take bets on specific sectors (e.g. oil & gas) while
keeping market beta constant; realise gains and liquidate

› Multi-index more efficient than single-index for tracking; avoids performance

mismeasurement due to unmatched factors

── WEEK 11 ──

W11 Blockchain, Cryptocurrency & CBDC ▾

C RY P T O G R A P H Y B A S I C S CONCEPT

› Hash Function (SHA-256): any input → fixed 64-char hex string; one-way,
deterministic, collision-resistant

› Digital Signature: 3 algorithms — GenerateKey (makes public + private key

pair), Sign (uses private key), Verify (uses public key)

› Private Key: sign transactions; Public Key: verify signatures & derive
address; circulated on network

› Unforgeability: knowing public key + seeing signatures ≠ ability to forge

› Asymmetric cryptography enables trustless ownership verification in

blockchain

B LO C KC H A I N & C O N S E N S U S HIGH YIELD

› Byzantine Generals Problem: how do distributed, mutually distrusting

nodes agree without central authority? → basis for consensus mechanisms

› Genesis Block: first block; hash of previous = 0; defines initial state

› Longer chain wins (most work done) if two valid chains presented

› Permissionless blockchain: open; no central authority; needs resource-

intensive consensus
› Permissioned blockchain: trusted participants; lighter consensus possible;
legal remedies available

PROOF OF WORK (POW) HIGH YIELD

› Miners compete to find nonce (random number used once) such that
Hash(prev_hash + transactions + nonce) starts with required number of zeros

› Hard to compute (~10²¹ calculations / 10 min per block) but easy to verify
(single hash)

› Winner mines next block & earns block reward (new cryptocurrency) +
transaction fees

› Double-spend problem: solved by PoW — miners verify transaction

integrity before adding to blockchain

› PoW coins: Bitcoin, Dogecoin, Bitcoin Cash, Litecoin, Monero

PoW Limitations: High energy consumption; centralisation risk (mining pools);

51% attack possible if majority computing power concentrated

C B D C ( C E N T R A L B A N K D I G I TA L C U R R E N C Y ) HIGH YIELD

› Retail CBDC: digital cash for general public; direct claim on central bank (like
physical cash, but digital)

› Wholesale CBDC: for interbank settlements

› Key difference: Bank deposits = claim on commercial bank; CBDC = claim on

central bank (no insolvency risk)

› Risk: too much CBDC adoption → deposits shift from commercial banks →
disrupts bank lending → harms economy

› 4 Motivations for CBDC: (1) Declining cash use, (2) Rise of private crypto,
(3) Central bank as payment innovator, (4) Global payment system control

› Live CBDCs: Bahamas Sand Dollar (Oct 2020), Nigeria e-Naira (Oct 2021),
Jamaica JAMDEX

› 87 countries / 90% global GDP exploring CBDCs

── WEEK 12 ──

W12 ESG, SRI & Sustainable Investing ▾

KEY DEFINITIONS CONCEPT

› SRI (Socially Responsible Investing): integrates social, environmental,

ethical factors into investment decisions

› ESG: Environmental, Social, Governance — criteria used to evaluate

companies

› Green Bonds: proceeds exclusively for green projects; same as corporate

bonds but labelled green; double-edged (tackle climate change + portfolio
diversification)

› Thematic Investing: top-down; invest in trends (e.g. clean energy)

› Impact Investing: financial return + measurable positive social/

environmental impact

SCREENING HIGH YIELD

› Negative Screening: exclude companies/sectors (weapons, tobacco,

gambling, alcohol, defense, poor labour relations, animal testing) from
investment universe

› Positive Screening: select companies meeting superior ESG standards;

often combined with Best-in-Class approach (top ESG performers in each
sector)

› 3rd Generation: triple bottom line — People, Planet, Profit (combined +ve
and −ve screens)

› 4th Generation: adds shareholder activism (voting at AGMs, direct

management dialogue)

› Norms-Based Screening: checks against UN Global Compact, OECD

guidelines, ILO standards, human rights declarations

SCREENING INTENSITY VS PERFORMANCE HIGH YIELD

› Relationship is curvilinear (not monotonic)

› Few screens (1–2): large universe → good diversification → market-level

performance

› Moderate screens (~7): minimum performance (worst of both worlds — less

diversification, not yet selecting only good stocks)

› Strict screens (~12): better stock selection (superior ESG stocks) but still
~2.4%/year below very diversified portfolios

S R I M U T UA L F U N D S — C O S T S & B E N E F I T S
 BENEFITS COSTS/CHALLENGES

Align with personal values Greenwashing risk

Lower long-term risk/volatility Inconsistent ESG rating

methodology

Positive societal change Limited diversification

Competitive financial returns Higher management fees

(possible)

P E R F O R M A N C E E VA LUAT I O N — FA M A - F R E N C H 3 - FA C T O R M O D E L HIGH YIELD

3-Factor Model (Green Fund Performance):

Rᵢₜ − RFₜ = αᵢₜ + β₁(Rₘₜ−RFₜ) + β₂·SMBₜ + β₃·HMLₜ + εₜ

› α (alpha): fund manager's stock selectivity ability; positive & significant =

superior performance

› SMB (Small Minus Big): size factor; return of small-cap minus large-cap
portfolios

› HML (High Minus Low): book-to-market factor

› Timing ability: add squared factors (RMT−RF)², SMB², HML²; positive λ =

manager increases exposure before factor outperforms

WHY SRI FUNDS UNDERPERFORM/OUTPERFORM

› Underperformance reasons: limited universe → less diversification → more

specific risk; frequent rebalancing costs; poor managerial skills in restricted
universe

› Outperformance reasons: stakeholder theory — better CSR firms have

competitive advantage; fund managers gain specialisation in limited ESG
universe; less ESG risk during crises

› Over time: SRI fund performance expected to improve as experience grows

T R A N S I T I O N T O S U S TA I N A B L E E C O N O M Y ( F R A M E W O R K )

› Channel 1: More investor inflow to sustainable funds → lower cost of capital

for green firms → encourages transition-aligned investment

› Channel 2: Engagement — investors influence companies through

shareholder activism, proxy votes, direct dialogue

── QUICK REFERENCE ──
 REF Key Formulas & Concepts at a Glance ▾

Portfolio Return:

Rₚ = Σ wᵢRᵢ

Portfolio Variance (2 assets):

σ²ₚ = w₁²σ₁² + w₂²σ₂² + 2w₁w₂ρ₁₂σ₁σ₂

Sharpe Ratio:

̄ − RF) / σ
SR = (R

CAPM:

̄ ̄ ₘ − RF)
Rᵢ = RF + βᵢ(R

APT:

̄
Rᵢ = RF + λ₁bᵢ₁ + λ₂bᵢ₂ + ... + λⱼbᵢⱼ

GARCH(1,1):

σ²ₜ = α₀ + α₁μ²ₜ₋₁ + β₁σ²ₜ₋₁ (stationarity: α₁+β₁ < 1)

EWMA:

σ²ₜ = λσ²ₜ₋₁ + (1−λ)μ²ₜ₋₁ (typical λ = 0.94)

Abatement Cost:

$ per ton CO₂ avoided = Total additional cost ÷ Avoided emissions

Risk Premium:

RP = E(W) − CE (CE = certainty equivalent)

M U S T- K N O W D I S T I N C T I O N S

› ETS vs Carbon Tax: ETS = quantity certainty, price volatility; Tax = price
certainty, quantity uncertainty

› Systematic Risk = market risk (cannot be diversified away); Specific Risk

= firm-level (diversifiable)

› CBDC vs Deposit: CBDC = claim on central bank; Bank deposit = claim on

commercial bank

› PoW = hard to compute, easy to verify; nonce = random number used once
› Negative screen = exclude bad; Positive screen = select best ESG
performers

› α (alpha) positive in 3-factor model = good selectivity; λ positive = good

timing ability

› Green Bond = regular bond but proceeds go only to green projects

› Contagion = crisis in one market spreads to others via investor psychology