Reference

Glossary

Every technical word in the course and what it means in plain language, with flashcards for practice.

Every technical term in the course, with a plain-language meaning. On every page these words are marked with a dotted underline: hover or tap one to see its meaning. The “Used in” links point to the sections where each term appears.

Flashcards — term ⇄ meaning
Ch 1 · Money, Time and Returns
TermMeaningUsed in
absolute returnThe total gain or loss over a holding period as a fraction of the starting value, $R=V_1/V_0-1$, with no time scale. Funds commonly quote it for periods under one year.1.3
amortisation scheduleThe month-by-month table of a loan showing, for each EMI, the interest part $iB_{k-1}$, the principal part $E-iB_{k-1}$ and the balance still owed.1.2
annuityA stream of equal payments at equal intervals. Ordinary annuity: payments at the end of each period; annuity due: at the start. Its value is a geometric series: $FV=A\frac{(1+i)^n-1}{i}$.1.2 2.3 8.1
CAGRCompound annual growth rate: the constant yearly rate that turns $V_0$ into $V_1$ in $t$ years, $(V_1/V_0)^{1/t}-1$. For a sequence of yearly returns it is their geometric mean.1.1 1.3 3.1 4.2
compound interestInterest that is added to the balance at the end of each period, so that later interest is also earned on earlier interest. With annual rate $r$, $P$ grows to $P(1+r)^t$ in $t$ years.1.1 1.2 1.3 2.1
continuous compoundingThe limit of ever more frequent compounding: $P$ grows to $Pe^{rt}$, and the effective annual rate is $e^r-1$. It is the natural language of finance theory because growth rates add.1.1 1.2 1.3 2.2
effective annual rateThe once-a-year rate that gives the same growth as a quoted nominal rate $r$ compounded $m$ times a year: $\text{EAR}=(1+r/m)^m-1$. Use it to compare quotes with different compounding.1.1
EMIEquated monthly instalment: the fixed monthly payment that repays a reducing-balance loan $L$ in $n$ months at monthly rate $i$: $E=\frac{Li}{1-(1+i)^{-n}}$. Early EMIs are mostly interest.1.1 1.2 2.2 8.1
flat interest rateInterest charged on the original loan amount for the whole term, $I=Lf\,n/12$, even though the loan is being repaid. The equivalent reducing-balance rate is close to twice the flat rate.1.2
geometric mean return$[(1+R_1)\cdots(1+R_n)]^{1/n}-1$: the yearly rate that reproduces the actual growth. It never exceeds the arithmetic average of the returns, and is roughly the average minus $\sigma^2/2$.1.3
IRRInternal rate of return: a rate at which the NPV of a set of cash flows is zero. It is unique when all outflows come before all inflows; otherwise there may be several or none.1.3 2.3 8.1
log return$r=\log(P_1/P_0)=\log(1+R)$, the natural log of the price ratio. Log returns of consecutive periods add up to the log return of the whole period.1.1 1.3 3.1 3.3
NPVNet present value: the sum of all cash flows of a project, each discounted to today at the discount rate, $\sum C_k/(1+k)^{t_k}$. A project adds value when its NPV is positive.1.1 1.3 5.1
part-prepaymentAn extra payment that reduces a loan's balance at once. The lender then either keeps the EMI and shortens the tenure, or keeps the tenure and lowers the EMI.1.2
perpetuityAn annuity that continues for ever. A payment of $A$ each period at periodic rate $i$ is worth $A/i$ today.1.2 2.3 5.1
point-to-point returnA return measured between one particular start date and one end date (a trailing return when the end is today). It can change sharply if either date moves across a crash or rally.1.3
present valueWhat a future amount is worth today: an amount $A$ due in $t$ years, discounted at the rate $r$ you could earn meanwhile, has present value $A/(1+r)^t$.1.1 1.2 1.3 2.3
purchasing powerWhat a sum of money can buy. With inflation $\pi$ a year, ₹$X$ received in $t$ years has the purchasing power of $X/(1+\pi)^t$ rupees today.1.1 2.2 8.1 8.2
real returnThe growth of purchasing power rather than of rupees: $1+r_{\text{real}}=(1+r_{\text{nom}})/(1+\pi)$ (the Fisher relation). It is approximately the nominal return minus inflation.1.1 1.2 3.1 3.3
rolling returnThe CAGR over every window of a fixed length (say 3 years) in a fund's history, summarised by its worst, median and best values. It shows the range of past outcomes instead of one start date.1.3
rule of 72A quick estimate of doubling time: at $r\%$ a year money doubles in about $72/r$ years (exact: $\log 2/\log(1+r)$). The rule of 70 is better at low rates and for continuous growth.1.1 1.2 1.3 3.3
simple interestInterest calculated only on the original principal: $I=Prt$, so the amount after $t$ years is $P(1+rt)$. A payout FD whose interest is spent behaves this way.1.1 2.3 6.1
SIPSystematic Investment Plan: investing a fixed amount every month (usually in a mutual fund). With payments at the start of each month it is an annuity due: $FV=A\frac{(1+i)^n-1}{i}(1+i)$ under an assumed monthly return $i$.1.1 1.2 1.3 2.2
step-up SIPA SIP whose monthly amount is raised by a fixed percentage $g$ once a year, usually to keep pace with income.1.2 8.1
XIRRThe annual rate $x$ with $\sum C_k(1+x)^{-t_k}=0$ for cash flows on arbitrary dates ($t_k$ = days/365 from the first flow). The right measure of the return on a SIP with redemptions; found numerically, e.g. by Newton's method.1.1 1.3 2.2 2.3
Ch 2 · Indian Markets and Instruments
TermMeaningUsed in
accrued interestThe part of the next coupon earned since the last coupon date, $C\times$ (days elapsed)/(days in the period). Bonds are quoted at the clean price; the buyer pays the dirty price = clean price + accrued interest.2.3
bid–ask spreadThe gap between the lowest price at which someone will sell (best ask $a$) and the highest price at which someone will buy (best bid $b$): $a-b$. A small market order costs about half of it; a round trip costs all of it.2.1 2.2 6.1 6.2
bonus issueFree new shares given to shareholders, $a$ for every $b$ held: the holding is multiplied by $(a+b)/b$ and the theoretical price by $b/(a+b)$.2.1
convexityThe curvature of the price–yield curve, $\frac1P\frac{d^2P}{dy^2}$. It adds $+\frac12\mathcal C(\Delta y)^2$ to the duration estimate, so gains from falling yields exceed losses from equal rises.2.3 4.3 5.1 8.2
coupon rateA bond's annual interest as a fraction of its face value, paid in equal coupons $cF/m$ ($m=2$ for half-yearly). It equals the bond's return only if the bond is bought at par.2.3
credit spreadA bond's yield minus the yield of a government security of the same maturity. It pays for expected default losses, uncertainty and lower liquidity; a spread $s$ with recovery $R$ covers a default probability of about $s/(1-R)$ a year.1.1 2.3 6.2
ETFExchange-traded fund: an index-tracking fund whose units trade on the stock exchange at a market price, which may be at a small premium or discount to its indicative NAV (iNAV).2.2 4.1 6.1 6.2
exit loadA charge, as a percentage of the redemption value, on units redeemed within a stated period after purchase. Units are redeemed first in, first out.2.2 8.1 8.2
expense ratioThe annual percentage of a fund's net assets charged for its expenses, accrued daily and already deducted in the NAV. A difference $\Delta c$ between two plans shrinks the costlier one's value by the factor $e^{-\Delta c\,T}$ over $T$ years.2.1 2.2 2.3 3.2
free floatThe part of a company's shares available to the public for trading, excluding promoter and other strategic holdings. Its fraction is the investable weight factor $f$; free-float market cap is $PQf$.2.1 4.2
G-secA government security: a bond issued by the Government of India through RBI auctions. Short-term ones are Treasury bills; longer dated securities usually pay coupons half-yearly. Prices are quoted per ₹100 of face value.2.3 5.1 5.2
IDCWIncome distribution cum capital withdrawal: a fund option that pays out amounts per unit from time to time. The NAV falls by the amount paid, so it is a withdrawal of the investor's own money, not extra income.2.2
index divisorThe number $D$ by which an index's total free-float market cap is divided to give the index level, $I=M/D$. It is set at the base date and rescaled, $D'=DM'/M$, whenever the basket changes for a non-market reason.1.1 2.1 3.1 3.2
limit orderAn order to buy at no more than (or sell at no less than) a stated price. It controls the price but may fill only partly or not at all; the unfilled part waits in the order book.2.1 2.2
Macaulay durationThe present-value-weighted average time to a bond's cash flows, in years. It equals the maturity for a zero-coupon bond and is shorter for a coupon bond.2.3 5.1
market capitalisationThe market value of all of a company's shares: price × number of shares. In India SEBI's categories use its rank: roughly the top 100 companies are large cap, the next 150 mid cap, the rest small cap (check current rules).2.1 5.1 5.2
modified duration$D_{\text{mod}}=-\frac1P\frac{dP}{dy}=D_{\text{mac}}/(1+y/m)$: the approximate percentage fall in price for a 1-point rise in yield, $\Delta P/P\approx-D_{\text{mod}}\Delta y$.2.3 7.4
order bookThe exchange's list of all waiting buy orders (bids) and sell orders (asks) for a share, sorted by price. Incoming orders are matched against it by price–time priority.2.1 2.2 7.3
rights issueAn offer to existing shareholders to buy $a$ new shares for every $b$ held at a price $S$, usually below the market price $P$. The theoretical ex-rights price is $(bP+aS)/(a+b)$; the right is worth that minus $S$.2.1
rupee-cost averagingInvesting a fixed amount at regular intervals, so that more units are bought when the price is low. The average cost per unit is the harmonic mean of the prices, which is at most their arithmetic mean.1.3 2.2
stock splitDividing each share into $k$ shares: the holding is multiplied by $k$ and the price divided by $k$, so the value of the position is unchanged.2.1
STPSystematic transfer plan: moving a fixed amount every month from one fund (often a liquid fund holding a lump sum) to another (often an equity fund), turning a lump sum into a SIP.2.2
SWPSystematic withdrawal plan: redeeming a fixed amount $W$ from a fund every month. With corpus $C$ and monthly return $i$ it lasts for ever if $W\le iC$, otherwise $-\log(1-iC/W)/\log(1+i)$ months.2.2 8.1
T-billA zero-coupon G-sec of 91, 182 or 364 days, bought at a discount and redeemed at face value. Its yield on the price paid is $\frac{F-P}{P}\cdot\frac{365}{d}$.2.3 3.2 4.1 4.2
total return indexAn index that assumes all dividends are reinvested on their ex-dates: $\text{TRI}_t=\text{TRI}_{t-1}(I_t+d_t)/I_{t-1}$. It is the fair benchmark for a fund; the price index leaves dividends out.2.1 2.2
tracking differenceThe fund's return minus its benchmark total return index's return over a period. Usually negative, mainly because of the expense ratio, cash drag and trading costs.2.1 2.2 3.2 6.2
tracking errorThe annualised standard deviation of the daily differences between a fund's and its index's returns, $s_d\sqrt{252}$. It measures how erratically the fund follows the index, not how much it lags.2.1 2.2 3.2 7.3
yield to maturityThe annual rate that makes the present value of a bond's remaining cash flows equal to its price — the bond's internal rate of return if held to maturity with coupons reinvested at that rate. Found numerically.2.3
Ch 3 · The Statistics of Returns
TermMeaningUsed in
alphaThe return of a fund or share beyond what its beta explains: $\alpha=\bar R-[r_f+\beta(\bar R_m-r_f)]$ (Jensen's alpha), the intercept of the market-model regression.2.1 3.2 3.3 4.1
betaThe slope of the regression of an asset's returns on the market's: $\beta=\operatorname{Cov}(r_i,r_m)/\operatorname{Var}(r_m)=\rho\,\sigma_i/\sigma_m$. It measures market (systematic) risk only.2.1 2.2 2.3 3.2
confidence bandA range expected to contain a future value with a stated probability under a model. Under GBM the central band is $S_0e^{\nu t\pm z\sigma\sqrt t}$, which widens in rupees as the horizon grows.3.3 7.4
correlationA number between −1 and 1 measuring how closely two return series move together in a straight-line way: $\rho=s_{xy}/(s_xs_y)$. Compute it from returns, never from price levels; it says nothing about cause.3.2 3.3 4.1 4.2
covarianceThe average product of two series' deviations from their means, $s_{xy}=\frac1{n-1}\sum(x_t-\bar x)(y_t-\bar y)$: positive when they tend to be above or below average together.3.2 4.1 4.2 4.3
efficient-market hypothesisThe idea that prices already reflect available information, so that it cannot be used to earn consistent excess returns after costs. Weak form: past prices; semi-strong: all public information; strong: all information.3.3
fan chartA chart of the percentiles (for example the 10th, 25th, 50th, 75th and 90th) of a simulated quantity over time, showing a range of outcomes instead of a single forecast line.3.3
fat tailsA return distribution has fat tails if very large moves (beyond about 3σ) are much more frequent than under a normal distribution with the same volatility, i.e. its excess kurtosis is positive.3.1 3.2 3.3 6.2
geometric Brownian motionThe standard continuous-time price model $S(t)=S_0e^{(\mu-\sigma^2/2)t+\sigma W(t)}$ with drift $\mu$ and volatility $\sigma$. The mean grows at $\mu$, the median at $\mu-\sigma^2/2$.3.1 3.3 6.3 6.4
information ratioAverage active return (fund minus benchmark) divided by the tracking error, the volatility of the active returns: how large and how steady a manager's outperformance has been.3.2 7.3
kurtosisThe tail-weight measure $m_4/m_2^2$; it equals 3 for a normal distribution. Excess kurtosis is kurtosis minus 3; a positive value signals fat tails.3.1
lognormal distributionThe distribution of a positive variable $V$ whose logarithm is normal, $\log V\sim N(a,b^2)$. Median $e^a$, mean $e^{a+b^2/2}$; skewed to the right. Prices are lognormal when log returns are normal.3.3 4.2 6.3 6.4
maximum drawdownThe drawdown is the fall from the running peak, $V_t/\max_{s\le t}V_s-1$. The maximum drawdown is the largest such fall over a period; recovering from a fall $d$ needs a gain of $d/(1-d)$.3.1 3.2 3.3 4.2
Monte Carlo simulationEstimating a probability or average by simulating many random scenarios from a model and counting. Its standard error falls like $1/\sqrt N$ for $N$ scenarios; it measures simulation noise, not model error.3.3 4.3 7.2 7.4
R-squaredIn the regression of an asset on the market, $R^2=\rho^2$ is the fraction of the asset's return variance explained by the market; $1-R^2$ is the specific (diversifiable) part.3.2
random walkA sequence $S_n=S_0+X_1+\cdots+X_n$ built from independent, identically distributed steps. After $n$ steps the mean change is $n$ times the drift and the standard deviation is $\sqrt n$ times the step SD. The basic model of unpredictable price changes.2.3 3.3 6.3 7.1
Sharpe ratioExcess return over the risk-free rate per unit of volatility, $(\bar R-r_f)/\sigma$. Unchanged by mixing with the risk-free asset; from $T$ years of data its standard error is roughly $1/\sqrt T$.3.1 3.2 4.2 7.3
skewnessA measure of asymmetry, $m_3/m_2^{3/2}$ (central moments). Negative skewness: a long left tail — many small gains and occasional large losses. A normal distribution has skewness 0.3.1 3.2 7.4
Sortino ratioExcess return over a target $\tau$ per unit of downside deviation $\sqrt{\frac1n\sum\min(R_t-\tau,0)^2}$: only returns below the target count as risk.3.2
square-root-of-time ruleFor independent returns, volatility grows with the square root of the horizon (variances add), while the mean grows in proportion to it: $\sigma_h=\sigma\sqrt h$, $\mu_h=h\mu$.3.1 7.3 7.4
standard errorThe standard deviation of an estimate across possible samples. For the mean of $n$ independent returns it is $\sigma/\sqrt n$; for an annual mean from $T$ years of data, about $\sigma_{\text{ann}}/\sqrt T$ whatever the sampling frequency.3.1 3.2 3.3 4.3
Treynor ratioExcess return over the risk-free rate per unit of beta, $(\bar R-r_f)/\beta$: the right yardstick for an asset that is a small part of a diversified portfolio.3.2
volatilityThe standard deviation of an asset's returns, a measure of how much they swing. Usually quoted annualised: with independent daily returns, $\sigma_{\text{ann}}=\sigma_{\text{day}}\sqrt{252}$.1.3 2.2 2.3 3.1
volatility dragThe gap between the arithmetic mean return and the compound (geometric) growth rate caused by ups and downs: approximately $\sigma^2/2$, so the CAGR is about $A-\sigma^2/2$.1.3 2.2 3.1 3.3
Ch 4 · Portfolio Mathematics
TermMeaningUsed in
asset allocationThe split of a portfolio between asset classes such as equity, debt and gold. A strategic allocation is the long-term target mix; tactical allocation means temporary deviations from it.3.1 4.2 5.1 8.1
binding constraintAn inequality that holds with equality at the chosen point (it is 'active'); a constraint with room to spare is slack. Only binding constraints have non-zero shadow prices.4.3
capital allocation lineThe straight line of risk–return combinations obtained by mixing a risky portfolio $P$ with a risk-free asset: $\mu=r_f+\frac{\mu_P-r_f}{\sigma_P}\sigma$. Its slope is $P$'s Sharpe ratio.3.2 4.2
capital market lineThe capital allocation line of the market portfolio, $\mu=r_f+\frac{\mu_M-r_f}{\sigma_M}\sigma$. Under CAPM efficient portfolios lie on it; individual stocks lie below it.4.2
CAPMThe Capital Asset Pricing Model: under its assumptions every asset's expected return is $\mu_i=r_f+\beta_i(\mu_M-r_f)$, so only systematic risk (beta) is rewarded.4.2 5.1 5.2 8.1
covariance matrixThe $n\times n$ matrix $\Sigma$ with entries $\sigma_{ij}=\rho_{ij}\sigma_i\sigma_j$ (variances on the diagonal). A portfolio's variance is $w^{\mathsf T}\Sigma w$; $\Sigma$ is symmetric and positive semidefinite.4.1 4.3 7.4
diversificationSpreading money across assets that do not move in lockstep, so that part of each asset's risk cancels out. With correlation below 1, a portfolio's volatility is less than the weighted average of its assets' volatilities.3.1 3.2 4.1 4.2
efficient frontierThe set of portfolios with the highest expected return for each level of volatility (equivalently, the least volatility for each expected return). Without constraints it is the upper branch of a hyperbola in the risk–return plane.4.1 4.2
equal-weight portfolioThe '1/N' portfolio with the same amount in each of N assets. It needs no estimates at all, but equal money is usually very unequal risk.3.2 4.1 4.3 7.3
estimation errorThe difference between an estimated input (such as an expected return computed from past data) and its true value. Optimisers amplify it: for two similar assets one point of error in the return gap can move the optimal weight by tens of points.4.3 5.1 7.2
factor modelA regression model that explains returns by several factor returns — market, size (SMB), value (HML), momentum (WML) — with loadings as exposures; rule-based "smart-beta" funds tilt towards such factors.3.2 4.2
feasible regionThe set of all choices that satisfy every constraint of an optimisation problem. For linear constraints it is convex; if it is empty the problem is infeasible.4.3
glide pathA planned reduction of the equity share (or other risky assets) as a goal date approaches, as in age-based rules and target-date funds.4.2 8.1
linear programmeAn optimisation problem with a linear objective and linear equality/inequality constraints, $\max p^{\mathsf T}x$ subject to $Ax\le h$. Its optimum (if the feasible region is bounded and non-empty) is at a corner. Abbreviated LP.4.3
market risk premiumThe expected return of the market portfolio over the risk-free rate, $\mu_M-r_f$. It is an estimate, not an observed constant, and drives every CAPM required return.4.2 5.1
minimum-variance portfolioThe mix of assets with the smallest possible variance, $w=\Sigma^{-1}\mathbf 1/(\mathbf 1^{\mathsf T}\Sigma^{-1}\mathbf 1)$, found with a Lagrange multiplier. Every asset has the same covariance with it.4.1 4.3
planning returnA deliberately prudent annual return, below the expected return, used to work out how much money a goal needs today; it builds a safety margin into a plan.4.3 8.1
portfolio weightThe fraction of a portfolio's value held in one asset, $w_i=V_i/V$. The weights add up to 1; a negative weight is a short position.4.1
quadratic programmeAn optimisation problem with a quadratic objective and linear constraints, such as minimising portfolio variance $w^{\mathsf T}\Sigma w$ under budget, long-only and cap constraints. Abbreviated QP; convex when the quadratic form is positive semidefinite.4.3
rebalancingTrading back to the target weights after market moves have made them drift, on a calendar or when a weight leaves a tolerance band. It keeps risk at the chosen level; it is not a guaranteed source of extra return.2.1 2.2 4.2 5.1
risk contributionThe share of a portfolio's volatility due to one asset, $RC_i=w_i(\Sigma w)_i/\sigma_p$; the contributions add up to $\sigma_p$. Weights show where the money is, risk contributions where the risk is.4.1 4.3 7.4
risk parityA portfolio in which every asset contributes the same share of total risk, $w_i(\Sigma w)_i/\sigma_p$ equal for all $i$. It uses no expected returns; for two assets it is $w_1=\sigma_2/(\sigma_1+\sigma_2)$.4.1 4.3
security market lineThe CAPM line $\mu=r_f+\beta(\mu_M-r_f)$ in the beta–return plane. An asset whose forecast return lies above it has a positive alpha, below it a negative alpha.4.2
shadow priceThe improvement in the optimal objective value per unit relaxation of a constraint's right-hand side, valid while the same constraints stay binding. Zero for a slack constraint.4.3
single-index modelThe model $r_i=\alpha_i+\beta_i r_m+\varepsilon_i$ that explains each asset's return by one market return plus an independent specific part; it splits variance into systematic and unsystematic risk.3.2 4.1
systematic riskThe part of an asset's risk caused by market-wide factors and shared with other assets; in the single-index model it is $\beta_i^2\sigma_m^2$. Diversification cannot remove it.4.1 4.2
tangency portfolioThe risky portfolio with the highest Sharpe ratio, $w_T\propto\Sigma^{-1}(\mu-r_f\mathbf 1)$; its capital allocation line touches the efficient frontier. In theory every investor holds it, mixed with the risk-free asset.4.2
unsystematic riskThe part of an asset's risk specific to that company (its residual variance $s_i^2$ in the single-index model). It averages out in a well-diversified portfolio.3.2 4.1 4.2
Ch 5 · Valuing a Company
TermMeaningUsed in
cost of equityThe annual return shareholders require from a company, usually estimated with CAPM: $k_e=r_f+\beta\,(E[r_m]-r_f)$. It is the discount rate for dividends and free cash flow to equity.4.2 5.1 5.2
dividend yieldDividend per share ÷ share price; equal to the payout ratio divided by the P/E.2.1 5.1 5.2 6.1
DuPont decompositionThe identity ROE = net margin × asset turnover × equity multiplier (sales/assets and assets/equity), with a five-step version adding the tax and interest burdens. It shows whether ROE comes from margins, efficiency or leverage.5.2
earnings yieldEPS ÷ price, the reciprocal of the P/E: the current profit per ₹100 invested.5.2
enterprise valueThe value of a whole business to all its capital providers: the present value of its free cash flow to the firm, or in market terms market capitalisation plus debt minus cash. Equity value = enterprise value − net debt.5.1 5.2
EPSEarnings per share: (net profit − preference dividends) ÷ number of equity shares. Diluted EPS also counts shares that options or convertibles could create.5.1 5.2
EV/EBITDAEnterprise value (market capitalisation + debt − cash) ÷ earnings before interest, tax, depreciation and amortisation. It compares businesses independently of how they are financed, but ignores capital expenditure.5.1 5.2
free cash flowCash a business generates after taxes and the reinvestment it needs. To the firm: FCFF = $\text{EBIT}(1-t)$ + depreciation − capex − increase in working capital; to equity (FCFE): FCFF minus after-tax interest plus net borrowing.5.1 5.2
Gordon growth modelValues a share as the present value of dividends growing at a constant rate $g\lt k_e$: $P_0=D_1/(k_e-g)$. More generally, a dividend discount model values a share by discounting its expected dividends.5.1 5.2
intrinsic valueThe present value of the cash flows an owner expects from an asset, discounted at a rate that reflects their risk. It is an estimate that depends on the forecasts used, unlike the market price.3.3 5.1 6.2 6.3
margin of safetyHow far a price is below an estimated value, as a fraction of the value: $1-P/V$. A cushion against errors in the estimate, not a guarantee.5.1
P/B ratioShare price ÷ book value per share. $\text{P/B}=\text{ROE}\times\text{P/E}$; the justified P/B $(\text{ROE}-g)/(k-g)$ exceeds 1 exactly when ROE exceeds the cost of equity.5.2
P/E ratioShare price ÷ EPS (trailing or forward). In the constant-growth model the justified forward P/E is $(1-g/\text{ROE})/(k-g)$, which equals $1/k$ when ROE equals the cost of equity.5.2
PEG ratioP/E divided by expected annual EPS growth in percent (P/E 30 with 15% growth gives PEG 2). A rough growth adjustment that ignores ROE, risk and how long growth lasts.5.2
return on equityROE = net profit ÷ shareholders' equity: the return a company earns on the shareholders' capital. Growth adds value only when ROE exceeds the cost of equity.5.1 5.2
sustainable growth rateThe growth rate of earnings, dividends and book value when a company keeps a constant ROE and retention ratio $b$ and issues no new shares: $g=b\times\text{ROE}$.5.1 5.2
terminal valueThe value, at the end of an explicit forecast, of all later cash flows — usually the Gordon formula $CF_n(1+g)/(k-g)$ with $g\lt k$. It is often 60–85% of a DCF value.5.1
value trapA share that looks cheap on its multiples because its prospects are genuinely poor — low ROE, shrinking earnings or high risk — and so stays cheap or gets cheaper instead of re-rating.5.2
WACCWeighted average cost of capital: $\frac{E}{D+E}k_e+\frac{D}{D+E}k_d(1-t)$ with market-value weights. It is the discount rate for free cash flow to the firm.5.1
Ch 6 · Derivatives on NSE: Futures and Options
TermMeaningUsed in
American optionAn option that can be exercised at any time up to expiry. An American call on a share paying no dividends is worth the same as the European call; an American put can be worth more.6.2 6.3
backward inductionPricing an option on a tree by starting from the payoffs at expiry and moving back one step at a time, replacing each pair of values by their discounted risk-neutral average (and, for an American option, by the exercise value if that is larger).6.3
basis riskThe basis is the futures price minus the spot price, $F_t-S_t$; it shrinks to zero at expiry. Basis risk is the risk that it, or the gap between a hedged portfolio and the index, moves unexpectedly while a hedge is in place.6.1
binomial treeA model in which, over each short step, the price of the underlying moves up by a factor $u$ or down by a factor $d$. Options are priced on it by working backwards from expiry (backward induction); with many steps the price converges to Black–Scholes.6.3 6.4
Black–Scholes formulaThe price of a European call, $C=Se^{-qT}N(d_1)-Ke^{-rT}N(d_2)$, with $d_{1,2}=\big[\log(S/K)+(r-q\pm\tfrac12\sigma^2)T\big]/(\sigma\sqrt T)$, assuming geometric Brownian motion with constant volatility and frictionless markets. The put follows from put–call parity.1.1 3.3 6.2 6.3
call optionThe right, but not the obligation, to buy the underlying at the strike $K$ (at expiry, for a European option). Payoff $(S_T-K)^+$; the buyer pays a premium $c$ and can lose at most that.6.2 6.3 6.4
cash-and-carry arbitrageBuying the underlying with borrowed money and selling it forward with a short future, locking in $F_0-S_0e^{(r-q)T}$; the reverse trade sells the underlying and buys the future. These trades keep futures near their cost-of-carry price.6.1
cost of carryThe net cost of holding an asset until a future date: interest on the money tied up minus income (dividends) received. It sets the fair futures price $F_0=S_0e^{(r-q)T}$.6.1 6.3
covered callOwning the underlying and selling a call on it: premium income and a cap on the gain, but only a small cushion against a fall.6.2
deltaThe rate of change of an option's value with the price of the underlying, $\Delta=\partial V/\partial S$. For a Black–Scholes call it is $e^{-qT}N(d_1)$, between 0 and 1; it is also the number of units of the underlying in the replicating (hedge) portfolio.1.1 2.1 2.2 2.3
delta hedgingHolding $\Delta$ units of the underlying (or futures) against each option sold, so that small price moves leave the position's value unchanged, and rebalancing as delta changes. It removes direction risk but leaves gamma (volatility) risk.6.3 6.4
European optionAn option that can be exercised only at expiry. NSE's index options are European-style; the Black–Scholes formula prices European options.6.2 6.3 6.4
forward contractAn agreement made today to buy or sell an asset on a future date at a price fixed today, with nothing paid now. Payoff to the buyer at expiry: $S_T-F_0$.6.1
futures contractA standardised forward contract traded on an exchange, with a clearing corporation as counterparty and gains and losses settled daily through margin accounts. NSE index futures are cash-settled.6.1 6.2 6.4 7.4
gammaThe rate of change of delta with the underlying price, $\Gamma=\partial^2V/\partial S^2$: the curvature of the option value. It is largest for at-the-money options close to expiry, where hedges go out of date fastest.2.1 2.3 4.1 4.2
hedge ratioThe futures exposure taken per rupee of exposure being hedged. For a portfolio of value $V$ and beta $\beta$, the number of index futures to sell is $N^*=\beta V/(F_0Q)$.6.1
implied volatilityThe volatility $\sigma$ that makes the Black–Scholes price equal an option's market price, found numerically (Newton's method with vega, or bisection). It is unique when the price lies within the no-arbitrage bounds.6.3 6.4
India VIXNSE's volatility index, computed from Nifty 50 option prices to measure the market's expected (risk-neutral) annualised volatility over the next 30 calendar days. A reading $v$ implies a 1-SD 30-day move of about $v\%\times\sqrt{30/365}$.6.4
initial marginThe deposit each side of a futures (or short option) position must place with the broker and clearing corporation, a fraction of the contract value set by a risk model. With margin fraction $m$ the leverage factor is $1/m$.6.1
iron condorSelling an out-of-the-money put spread and call spread together: a net credit $C$ is the maximum profit if the price stays between the short strikes; the maximum loss is the wing width minus $C$.6.2
lot sizeThe number of units of the underlying in one derivatives contract (for an index contract, rupees per index point). Contract value $=F\times Q$. Set and revised by the exchange — always check the current value.6.1 6.2 6.3 6.4
margin callA demand from the broker to add money when losses take a margin account below the required level. If it is not met, the position may be closed, whatever happens to the price afterwards.3.3 6.1
mark-to-marketDaily settlement of a futures position: each day the long side receives $Q(F_t-F_{t-1})$ in cash (pays it if negative). The daily amounts add up to the contract's payoff.2.3 6.1 6.2 6.4
moneynessWhere the strike lies relative to the current price: a call is in the money if $S\gt K$, at the money if $S\approx K$, out of the money if $S\lt K$ (reversed for a put). Only in-the-money options have intrinsic value.6.2 6.3 6.4
option premiumThe price of an option, paid by the buyer to the seller (writer) when the trade is made; quoted per unit and multiplied by the lot size. It equals intrinsic value plus time value.6.2 6.3 6.4
protective putOwning the underlying and buying a put: a floor on the loss (strike minus premium) while keeping the upside — insurance for a fee.6.2
put optionThe right, but not the obligation, to sell the underlying at the strike $K$. Payoff $(K-S_T)^+$; the buyer pays a premium $p$. Buying a put protects a holding against a fall.4.2 6.2 6.3 6.4
put–call parityThe no-arbitrage identity for European options with the same strike and expiry: $c-p=S_0e^{-qT}-Ke^{-rT}=(F_0-K)e^{-rT}$. It needs no model of price movements.6.2 6.3 6.4
replicating portfolioA combination of the underlying asset and risk-free borrowing or lending that pays exactly what an option pays in every state. By the no-arbitrage law of one price, the option is worth what this portfolio costs.2.2 6.3 6.4
risk-neutral probabilityThe probability, such as $p^*=(e^{r\Delta t}-d)/(u-d)$ in a binomial tree, under which every asset earns the risk-free rate on average. Discounting expected payoffs under it gives no-arbitrage prices. It is a pricing device, not a forecast.6.3 6.4
straddleA call and a put at the same strike (a strangle uses an out-of-the-money put and call). Bought, it profits from a big move either way beyond the break-evens $K\pm(c+p)$; sold, it profits if the price stays near the strike and can lose without limit.6.2 6.4
strike priceThe fixed price $K$ at which an option holder may buy (call) or sell (put) the underlying.6.2
thetaThe rate of change of an option's value as time passes, $\Theta=\partial V/\partial t$, usually quoted per day. It is normally negative for option buyers (time decay) and accelerates like $1/\sqrt T$ for at-the-money options near expiry.2.1 3.3 4.3 6.4
vegaThe sensitivity of an option's value to volatility, $\partial V/\partial\sigma$, usually quoted per volatility point (divide the formula value by 100). It is the same for a call and a put with the same strike and expiry, and grows with $\sqrt T$.6.4
volatility smileThe pattern of implied volatilities across strikes for one expiry. Equity index options usually show a downward skew: low-strike puts have higher implied volatility than high-strike calls, reflecting crash risk and fat tails.6.4
Ch 7 · The Mathematics of Trading
TermMeaningUsed in
backtestA simulation of how a trading rule would have performed on past (or simulated) prices, using at each moment only the information available then.4.2 6.4 7.1 7.2
Bollinger bandsLines $k$ standard deviations (usually $k=2$) above and below the 20-day SMA, using the standard deviation of the last 20 closes. Their width measures recent volatility; closes outside the bands are common even for a random walk.7.1
bootstrapA resampling method: draw the observed returns at random with replacement many times and recompute a statistic, to estimate its uncertainty without assuming normality.3.3 7.1 7.3
crossoverThe day a fast moving average moves above (bullish) or below (bearish) a slow one, or the price crosses its average. On a trend the gap between the averages is proportional to the slope, so a crossover is a delayed estimate that the slope has changed sign. A 50/200-day bullish crossover is called a golden cross.7.1 7.3
data snoopingTesting many rules on the same data and reporting the best without allowing for the number of attempts; the winner's results are biased upwards.3.2 7.1 7.3
diversification benefitThe amount by which a portfolio's VaR (or risk) is below the sum of its holdings' standalone VaRs; it shrinks as correlations rise towards 1.4.1 7.4
equity curveThe value of a strategy's capital over time, $E_t=E_0\prod(1+R_k)$; its CAGR, volatility and drawdown summarise a backtest.7.2 7.3
expectancyThe expected profit per trade, $E=p\overline W-(1-p)\overline L$, or $pb-(1-p)$ in units of the risk per trade. A system is worth trading only if its expectancy after costs is positive.6.2 7.2
expected shortfallThe average loss on the occasions when the loss reaches or exceeds VaR; it describes the size of the tail that VaR ignores and, unlike VaR, is subadditive.4.3 7.4
exponential moving averageThe recursive average $\text{EMA}_t=\alpha P_t+(1-\alpha)\text{EMA}_{t-1}$, a weighted average with geometric weights $\alpha(1-\alpha)^k$. An "$n$-day" EMA uses $\alpha=2/(n+1)$, which gives it the same average lag $(n-1)/2$ as the $n$-day SMA.7.1
historical simulationEstimating VaR by applying the asset returns of the last $N$ days to today's holdings and reading the required quantile from the $N$ scenario losses.7.4
implementation shortfallThe cost of executing an order measured against the price when the decision was made, including the opportunity cost of any part left unfilled.7.3
Kelly criterionThe betting fraction that maximises long-run growth $g(f)=p\log(1+bf)+q\log(1-f)$: $f^*=p-q/b$. Betting about twice $f^*$ gives zero typical growth; fractional Kelly (e.g. half) keeps most of the growth with much smaller drawdowns and is robust to an overestimated edge.7.2
look-ahead biasUsing information in a backtest that would not have been available when the trade was made, such as the same day's closing price or later-revised data.7.1 7.3
losing streakA run of consecutive losing trades. With loss probability $q$ a given run of $k$ trades are all losses with probability $q^k$, and the longest streak in $N$ trades is typically about $\log(Np)/\log(1/q)$ — long streaks are normal, not a sign that a system has broken.4.2 7.1 7.2
MACDMoving-average convergence–divergence: the 12-day EMA minus the 26-day EMA, with a 9-day EMA of it as the signal line and their difference as the histogram. On a trend of slope $b$ it settles at $7b$ — a smoothed estimate of the price's rate of change.2.3 7.1
out-of-sample testAn evaluation on data that played no part in designing or choosing the rule, carried out once after the rule is fixed; a walk-forward test repeats this through time.7.1 7.3
overfittingTuning a rule to the accidental features of one data set, so that it performs well there and poorly on new data.7.3
parametric VaRVaR from an assumed normal distribution: $z_\alpha\sigma V$ (zero mean), with the portfolio volatility from the covariance matrix.3.1 7.4
position sizingChoosing how many shares to buy from the risk you accept: $Q=\lfloor fW/|E_0-S|\rfloor$ for equity $W$, risk fraction $f$, entry $E_0$ and stop $S$. Under fixed-fractional sizing each trade risks the same fraction of current equity.7.2
R-multipleA trade's profit or loss divided by its initial risk $1R$ (the planned loss to the stop). A trade stopped out exactly is $-1R$; one that makes three times the risk is $+3R$.7.2
reward–risk ratioPlanned gain to target divided by planned loss to stop; for a system, the payoff ratio $b$ = average win / average loss. The break-even win rate is $1/(1+b)$.7.2
risk of ruinThe probability that capital falls to a ruin level (zero, or a drawdown such as 50% after which one would stop). With fixed one-unit bets, win probability $p\gt\frac12$ and $a$ units of capital it is $(q/p)^a$.7.2
RSIRelative strength index: $100\cdot\text{AG}/(\text{AG}+\text{AL})$, where AG and AL are the average gain and average loss over (usually) 14 days. Above 70 is called overbought and below 30 oversold — descriptions of recent momentum, not forecasts.7.1
simple moving averageThe plain average of the last $n$ closes, $\text{SMA}_n(t)=\frac1n(P_t+\cdots+P_{t-n+1})$. On a steady trend it lags the price by $(n-1)/2$ days; it reduces independent noise variance by the factor $1/n$.7.1 7.3
slippageThe difference between the price at which a trade was intended and the average price actually obtained, caused by the spread, price moves and the order's own impact.7.1 7.2 7.3 8.2
stop-lossA price at which a losing position is to be closed, set in advance. It fixes the planned loss per share; gaps and slippage can make the actual loss larger. A stop placed within the share's normal noise is hit often: about $2\Phi(-d/(\sigma\sqrt T))$ of the time over $T$ days.2.1 3.3 6.1 7.1
stress testRevaluing a portfolio under a specified extreme joint move of the risk factors, without attaching a probability; a reverse stress test finds the scenario that would cause a given loss.3.2 3.3 4.1 5.1
subadditivityThe property $\rho(A+B)\le\rho(A)+\rho(B)$ of a risk measure: combining positions never looks riskier than the sum. Expected shortfall has it; VaR does not in general.7.4
support and resistancePrice levels where a share previously stopped falling (support) or rising (resistance), such as earlier lows and highs or round numbers. Mathematically they are past local minima and maxima; random walks produce them too.7.1
technical indicatorA number or line computed from past prices (and sometimes volumes) by a fixed formula — a moving average, RSI, MACD or Bollinger band. It summarises the recent past; whether it predicts anything is a statistical question to be tested.7.3
Value at RiskFor a horizon and confidence level $\alpha$, the loss that is exceeded with probability at most $1-\alpha$ — a quantile of the loss distribution, not the maximum possible loss.1.1 1.2 1.3 2.1
VaR exceptionA day on which the actual loss exceeds the VaR computed in advance; for a correct 99% model about 1% of days, scattered at random.7.4
VWAPVolume-weighted average price $\sum p_iq_i/\sum q_i$ over a period; a common benchmark for executing large orders.7.3
win rateThe fraction $p$ of trades that make money. On its own it says little: a system can win 90% of the time and still lose money if its losses are much larger than its wins.7.1 7.2 8.2
Ch 8 · Planning a Financial Life
TermMeaningUsed in
base-rate neglectJudging a case on striking evidence while ignoring how common the underlying type is. Bayes' theorem combines the base rate with the evidence: if skill is rare, most winning streaks belong to the lucky.8.2
behaviour gapThe difference between a fund's return (CAGR) and the return its investors actually earn (their XIRR), caused by the timing of their purchases and sales.8.2
bucket approachHolding the next one to three years of retirement withdrawals in low-volatility assets and the rest in growth assets, refilling the first bucket by rebalancing, so that growth assets need not be sold just after a fall.8.1
corpus factorThe ratio of a retirement corpus to the first year's spending, $F=(1+r)\,[1-(1+r)^{-N}]/r$ for withdrawals at the start of each year rising with inflation, where $r$ is the real return and $N$ the number of years.8.1
cost dragThe yearly loss of return to costs (expense ratio, trading costs, fees). Over $T$ years at gross return $R$ a cost $c$ removes about the fraction $1-e^{-cT/(1+R)}$ of final wealth.7.1 7.3 8.2
emergency fundMoney kept in safe, quickly accessible form to cover several months of essential spending during a loss of income or an unexpected expense, so that long-term investments need not be sold at a bad time.1.2 4.3 8.1
endowment planLife insurance that pays the sum assured on death during the term and a maturity benefit (sum assured plus any bonuses) on survival; its premium buys cover and savings together, and its savings part can be judged by an IRR.8.1
goal inflationThe expected yearly rise in the cost of a particular goal (education, a wedding, a house), which may differ from general inflation. A goal costing $C_0$ today costs $C_0(1+\pi_g)^T$ in $T$ years.8.1
human life valueThe present value of the income an earner would contribute to the family until retirement, after taxes and own consumption; a rough guide to the life cover needed.8.1
loss aversionFeeling a loss more strongly than an equal gain (often modelled as weighting losses by $\lambda\approx2$). Combined with frequent checking it makes a rising investment feel like a losing one.8.2
overconfidenceOverrating one's own skill or information, often after a lucky streak; it leads to over-trading and higher costs.8.2
Ponzi schemeA fraud that pays promised returns to existing investors out of new investors' money; it needs new money to grow at least as fast as the promised rate and must collapse when it does not.8.2
portfolio turnoverHow many times a year the whole capital is traded. With $m$ round trips at cost $k$ each, trading costs multiply the capital by $(1-k)^m$ a year, a drag of about $mk$.8.2
recency biasGiving too much weight to the latest results — buying last year's best fund or expecting a recent trend to continue — although short records are dominated by luck.8.2
retirement corpusThe amount needed at retirement to fund the planned withdrawals for the whole retirement under an assumed return; for inflation-indexed withdrawals it is the present value of a growing annuity.8.1
sequence-of-returns riskThe risk that the order of returns, not only their average, decides whether a plan with contributions or withdrawals succeeds. Bad years early hurt a withdrawal plan most; bad years late hurt a SIP most.2.2 8.1
tax deferralPaying tax when a gain is realised rather than every year. At the same rate, deferral always ends with more money, because the tax not yet paid keeps compounding.8.2
tax-loss harvestingSelling an investment at a loss to set the loss off against realised gains, then reinvesting. Because the repurchase resets the cost lower, it mostly defers tax rather than removing it; rules on set-off and repurchase vary and change.8.2
term insuranceLife insurance that pays the sum assured only if the insured dies during the policy term and nothing on survival, so its premium buys cover alone.8.1
withdrawal rateThe first year's withdrawal from a retirement corpus as a percentage of the corpus. The "4% rule" is a rule of thumb for this rate, not a law; the sustainable rate depends on the real return, volatility and horizon.8.1