Security analysis and portfolio management

In short

Security analysis is the appraisal of individual securities - equity, debt, hybrid - to estimate their intrinsic worth and expected return and risk; portfolio management is the construction, evaluation and revision of a combination of such securities so that risk is minimised for a target return. The two together form the core of the "SAPM" paper in almost every Indian MBA/PGDM finance specialisation and in CFA Level I-II, NISM and CA Final syllabi. It matters because risk falls when assets are combined but return does not - the single most powerful idea in finance - and because pricing models (CAPM, APT) give the discount rate every valuation exercise needs. It is examined through numericals on return and risk, two-security portfolios, beta, CAPM/SML and performance ratios, and in interviews at AMCs, PMS and wealth firms (HDFC AMC, SBI MF, ICICI Prudential, Motilal Oswal, Kotak, Nippon India), brokerages and research desks.

Explained

Return, risk and the investment setting

• Holding period return (HPR) = (ending price − beginning price + income)/beginning price. Expressed as a percentage per period; annualised for comparison.
• Expected return on a security = probability-weighted average of possible returns; historically, the arithmetic mean of past returns. Geometric mean (CAGR) is used for multi-period compounded growth and is always ≤ arithmetic mean.
• Risk = dispersion of actual return around the expected return, measured by variance (σ²) and standard deviation (σ); coefficient of variation (CV = σ/E(R)) compares risk per unit of return across securities of different scale.
• Types of risk: *systematic* (market-wide: interest rate, inflation/purchasing power, exchange rate, political, market risk) - cannot be diversified away; *unsystematic* (business risk, financial risk from leverage, management risk, default risk) - removed by diversification.
• Risk-return trade-off: higher expected return is the compensation for bearing higher *systematic* risk only; diversifiable risk earns no premium.
• Investment versus speculation versus gambling: investment is a longer horizon on fundamentals with reasonable risk; speculation is short-horizon price anticipation with high risk; gambling has negative expected value and created risk.
• Investment objectives: return, safety of principal, liquidity, hedge against inflation, tax efficiency (capital gains on listed equity under India's capital gains provisions - rates were revised in 2024 to 12.5% long-term and 20% short-term on listed equity with the LTCG exemption threshold raised; the governing statute is now the Income-tax Act, 2025 in force from 1 April 2026, which renumbered sections, so quote the provision by name rather than the old section number).

Fundamental analysis - the E-I-C framework

Top-down analysis in three stages, used to estimate intrinsic value, which is then compared with market price (buy if intrinsic > market).

1. Economy analysis: GDP growth, IIP, inflation (CPI/WPI), interest rates and the RBI repo rate, fiscal and monetary policy, Union Budget, monsoon, exchange rate, FII/FPI and DII flows, global commodity and crude prices. Leading, lagging and coincident indicators; diffusion indices; econometric and anticipatory surveys as forecasting tools.
2. Industry analysis: industry life cycle (pioneering → expansion/growth → stagnation/maturity → decline), Porter's five forces, nature of competition, cost structure, regulation, input availability, demand drivers, and classification into *growth, cyclical, defensive and cyclical-growth* industries.
3. Company analysis: management quality and promoter holding, business model and competitive moat, operating and financial leverage, earnings quality, and valuation through EPS, P/E, book value, P/BV, dividend yield, PEG, EV/EBITDA, RoE and RoCE; DuPont decomposition of RoE.
4. Valuation models used in security analysis: present value of expected cash flows - the dividend discount model (DDM) in its zero-growth, constant-growth (Gordon) and multi-stage forms; the earnings multiplier (P/E) approach; free-cash-flow models.
5. Bond/debt security analysis: intrinsic value is the PV of coupons and redemption value; yields (current yield, YTM, YTC); credit ratings (CRISIL, ICRA, CARE, India Ratings - SEBI-registered CRAs); interest-rate risk measured by duration. (Detailed bond mathematics belongs with fixed income.)
6. Limitations: depends on forecast quality, assumes the market will eventually correct mispricing, data lags, and ignores sentiment and timing.

Technical analysis

• Premise: prices move in trends set by the balance of demand and supply; history repeats because investor psychology repeats; price and volume data alone are sufficient.
• Dow Theory: three movements - *primary* (bull/bear, lasting years), *secondary* (reactions of one-third to two-thirds of the primary move), *minor* (daily); the averages must confirm each other; volume must confirm the trend; a trend persists until a clear reversal.
• Charts: line, bar, Japanese candlestick, point-and-figure. Patterns: head-and-shoulders (reversal), double top/bottom, triangles, flags, cup-and-handle, gaps; support and resistance levels; trendlines and channels.
• Indicators and oscillators: moving averages (simple and exponential) and the golden/death cross, MACD, RSI (above 70 overbought, below 30 oversold), stochastics, Bollinger Bands, rate of change, on-balance volume, advance-decline line, Elliott wave (5 impulse, 3 corrective waves), Fibonacci retracements (38.2%, 50%, 61.8%).
• Market sentiment measures: put-call ratio, short interest, India VIX, mutual fund cash levels, odd-lot theory, breadth indicators.
• Criticism: no theoretical foundation, subjective pattern reading, self-fulfilling in the short run, and directly contradicted by the weak form of the EMH; still universally used for entry/exit timing by traders and even by fundamental managers.

Efficient market hypothesis and behavioural challenge

• EMH: prices fully and instantly reflect all available information, so no analysis can earn consistent abnormal risk-adjusted returns.
• Weak form: past prices hold no information - defeats technical analysis; tested by serial correlation and runs tests, and by filter-rule tests.
• Semi-strong form: all public information is impounded - defeats fundamental analysis on public data; tested by event studies around results, bonus, split and merger announcements.
• Strong form: even private/insider information is reflected - no one, not even insiders, can beat the market; largely rejected, and insider trading is in any case prohibited under the SEBI (Prohibition of Insider Trading) Regulations, 2015.
• Anomalies: small-firm (size) effect, value effect, January/turn-of-the-year and Monday effects, post-earnings-announcement drift, momentum, overreaction - these motivated multifactor models (Fama-French three- and five-factor).
• Implication for practice: if markets are largely efficient, passive index funds and ETFs and low cost dominate; India's active-versus-passive debate turns exactly on this, and SEBI-mandated TRI benchmarking since 2018 made the comparison honest.

Portfolio theory - Markowitz, Sharpe and CAPM

• Markowitz (1952) mean-variance model: investors are risk-averse and judge portfolios only by expected return and variance over one period; the key insight is that portfolio risk depends on covariance/correlation, not just the individual σ's.
• Diversification: combining securities with correlation < +1 reduces σp below the weighted average of individual σ's; at ρ = −1 risk can be eliminated entirely; at ρ = +1 there is no benefit. Beyond roughly 15-25 well-spread stocks only systematic risk remains.
• Efficient frontier: the set of portfolios giving maximum return for each level of risk (or minimum risk for each return); the minimum variance portfolio is its leftmost point. Portfolios below the frontier are dominated.
• Utility and optimal portfolio: the investor's optimum is where their indifference curve is tangent to the efficient frontier - more risk-averse investors choose a point lower on the frontier.
• Capital Market Line (CML): introduce a risk-free asset and lending/borrowing; the straight line from Rf tangent to the frontier dominates it. The tangency point is the market portfolio (M); all investors hold M plus Rf in different proportions - Tobin's separation theorem.
• Sharpe's single-index model: Ri = αi + βi Rm + ei; it reduces Markowitz's n(n−1)/2 covariance estimates to n betas, making the model computationally workable. Total risk = systematic (βi²σm²) + unsystematic (σ²ei).
• CAPM: in equilibrium only beta is priced; the Security Market Line (SML) gives E(Ri) = Rf + βi[E(Rm) − Rf]. Securities plotting above the SML are underpriced (buy), below are overpriced (sell).
• Beta: sensitivity of a security's return to market return; β = Cov(i,m)/σ²m, estimated by regressing stock returns on index (Nifty 50/Sensex) returns - the characteristic line. β > 1 aggressive, β = 1 market, β < 1 defensive, β < 0 rare (gold-type assets). Portfolio beta is the weighted average of component betas.
• CAPM assumptions and criticism: homogeneous expectations, one period, perfect divisibility, no taxes or transaction costs, unlimited risk-free lending and borrowing, price-taking investors. Roll's critique (the true market portfolio is unobservable), beta instability, and empirical flatness of the SML are the standard criticisms.
• Arbitrage Pricing Theory (APT, Ross 1976): return is driven by several systematic factors - E(Ri) = Rf + Σ βik λk - with no need for a market portfolio or mean-variance utility; it relies only on no-arbitrage. Harder to apply because the factors are not specified.

Portfolio construction, revision and performance evaluation

• Portfolio management process: (1) specify objectives and constraints (return, risk tolerance, horizon, liquidity, tax, legal, unique needs) in an Investment Policy Statement; (2) set asset allocation; (3) select securities; (4) execute; (5) monitor and revise; (6) evaluate performance.
• Strategies: *active* (stock picking, sector rotation, market timing) versus *passive* (buy-and-hold, indexing); strategic versus tactical asset allocation; core-satellite.
• Portfolio revision: *formula plans* remove emotion - constant rupee value plan, constant ratio plan, variable ratio plan, and rupee cost averaging (equal rupees invested periodically, as in an SIP, buying more units when prices are low); rebalancing and portfolio-upgrading.
• Performance measures: Sharpe ratio (excess return per unit of total risk - for a whole portfolio), Treynor ratio (per unit of systematic risk - for a component of a diversified portfolio), Jensen's alpha (excess over CAPM-required return - measures selection skill), Information ratio (alpha ÷ tracking error), M² and Fama's decomposition (splitting return into risk-free, compensation for risk and net selectivity). Sharpe uses σ and Treynor uses β - that is the single most asked distinction.
• Mutual funds in India: governed by the SEBI (Mutual Funds) Regulations, 1996, with the sponsor-trustee-AMC structure; NAV = (assets − liabilities)/units outstanding; direct and regular plans; TER caps; categorisation and rationalisation circular (2017) fixing scheme categories; Riskometer disclosure; benchmarking against Total Return Indices; schemes must disclose portfolios periodically. Portfolio Management Services (PMS) have a ₹50 lakh minimum investment and Alternative Investment Funds are regulated under the SEBI (AIF) Regulations, 2012. Investment advisers are regulated under the SEBI (Investment Advisers) Regulations, 2013. (SEBI amends limits frequently - verify current thresholds and TER slabs before quoting.)

Key points

• Security analysis - estimating intrinsic value and risk of one security; portfolio management - combining securities to optimise the risk-return pair.
• Systematic risk - market-wide, undiversifiable, measured by beta, the only risk rewarded; unsystematic risk - firm-specific, removed free of cost by diversification.
• Covariance/correlation drives diversification - risk reduction needs ρ < +1; gains are largest when ρ is negative.
• Efficient frontier - maximum return per unit of risk; the minimum variance portfolio is its left tip.
• CML - risk-return line for efficient portfolios when a risk-free asset exists; its slope is the market price of risk.
• SML/CAPM - E(Ri) = Rf + β(Rm − Rf); applies to *every* security, efficient or not; above SML = underpriced.
• Beta = Cov(i,m)/σ²m = r(i,m)σi/σm; portfolio beta is the weighted average of betas.
• Single-index model - cuts estimation from n(n−1)/2 covariances to n betas; total variance = β²σ²m + σ²e.
• APT - multifactor, no-arbitrage alternative to CAPM; no market portfolio needed, but factors are unspecified.
• EMH forms - weak (past prices), semi-strong (public information), strong (all information including private).
• Fundamental versus technical - intrinsic value and long horizon versus price-volume patterns and short-term timing.
• Dow Theory - primary, secondary and minor movements; averages and volume must confirm.
• Sharpe versus Treynor - total risk (σ) versus systematic risk (β); use Sharpe for a stand-alone portfolio, Treynor for a component.
• Jensen's alpha - actual return minus CAPM-required return; positive alpha = superior selection.
• Formula plans - constant rupee, constant ratio, variable ratio; rupee cost averaging is the SIP principle.
• Markowitz assumptions - risk aversion, single period, decisions on mean and variance alone.
• Indian framework - SEBI (Mutual Funds) Regulations 1996, TRI benchmarking, Riskometer, PIT Regulations 2015, PMS minimum ₹50 lakh.

Formulas and examples

Example 1 - Expected return, risk and CV. A stock's returns and probabilities: 0.2 → 25%, 0.5 → 10%, 0.3 → −5%.

• E(R) = 0.2(25) + 0.5(10) + 0.3(−5) = 5 + 5 − 1.5 = 8.5%
• Deviations: 16.5, 1.5, −13.5. σ² = 0.2(272.25) + 0.5(2.25) + 0.3(182.25) = 54.45 + 1.125 + 54.675 = 110.25
• σ = 10.5%; CV = 10.5/8.5 = 1.24 (risk per unit of return).

Example 2 - Two-security portfolio and the correlation effect. Security A: E(R) = 18%, σ = 20%. Security B: E(R) = 12%, σ = 10%. Weights 50:50.

• E(Rp) = 0.5(18) + 0.5(12) = 15% - unaffected by correlation.
• σp at ρ = +1: √(0.25×400 + 0.25×100 + 2×0.5×0.5×1×20×10) = √(100 + 25 + 100) = √225 = 15.0%
• σp at ρ = 0: √(100 + 25 + 0) = √125 = 11.18%
• σp at ρ = −1: √(100 + 25 − 100) = √25 = 5.0%
• Minimum-variance weight at ρ = 0: wA = (100 − 0)/(400 + 100 − 0) = 0.20, so 20% A and 80% B → σp = √(0.04×400 + 0.64×100) = √(16 + 64) = √80 = 8.94%, with E(Rp) = 0.2(18)+0.8(12) = 13.2%.

Example 3 - Beta, CAPM and mispricing. Rf = 6.5%, E(Rm) = 13%. Stock X: σx = 28%, σm = 16%, ρxm = 0.80.

• βx = 0.80 × 28/16 = 1.40
• Required return = 6.5 + 1.40(13 − 6.5) = 6.5 + 9.1 = 15.6%
• Systematic variance = β²σ²m = 1.96 × 256 = 501.76; total variance = 784 → unsystematic variance = 282.24, so 64% of total risk is systematic.
• If analysts forecast 18% from X, alpha = 18 − 15.6 = +2.4%: the stock plots above the SML and is underpriced → buy.
• A portfolio of 60% X (β 1.40) and 40% of a stock with β 0.70: βp = 0.6(1.40) + 0.4(0.70) = 1.12; required return = 6.5 + 1.12(6.5) = 13.78%.

Example 4 - Performance evaluation of three funds. Rf = 6%, Rm = 12%, σm = 15%.

• Ranking by Sharpe: Q > P > Market > R. By Treynor: Q > P > Market > R. Fund R underperformed on every measure.
• P and Q have identical alphas, but Q achieves it with far less total risk - which is why Sharpe separates them (0.583 vs 0.50) while Jensen's alpha does not.

Example 5 - Intrinsic value by Gordon's model and a P/E cross-check. A company pays D₀ = ₹8, expects g = 6%, and investors require 14%.

• D₁ = 8 × 1.06 = ₹8.48 → P₀ = 8.48/(0.14 − 0.06) = ₹106
• Market price ₹92 → the share is underpriced by ₹14; expected return at ₹92 = 8.48/92 + 0.06 = 0.0922 + 0.06 = 15.2% against 14% required.
• Cross-check: if EPS₁ = ₹14 and the payout is 8.48/14 ≈ 60.6%, justified P/E = payout/(k − g) = 0.606/0.08 = 7.6×; market P/E = 92/14 = 6.6× - again cheap on this model's assumptions.

Where it is used

• Asset management companies (HDFC AMC, SBI MF, ICICI Prudential, Nippon India, Axis MF): fund managers set asset allocation and construct portfolios; dealers execute; risk teams monitor beta, tracking error and the SEBI Riskometer.
• Equity research and broking (Motilal Oswal, Kotak Securities, ICICI Securities, Jefferies, Nuvama): E-I-C analysis, target prices, buy/sell/hold ratings; technical desks publish levels and derivatives strategies.
• Wealth management, PMS and family offices: investment policy statements, model portfolios, SIP and rupee-cost-averaging advice, rebalancing and goal-based allocation for HNI clients.
• Insurance companies and pension/retirement funds (LIC, EPFO, NPS pension fund managers): asset-liability-driven allocation within IRDAI/PFRDA investment limits; NPS auto-choice lifecycle funds are age-based glide paths.
• Banks, treasury and corporate treasuries: surplus deployment across liquid funds, G-secs and fixed deposits using duration and risk-return criteria.
• Fintech and discount broking (Zerodha, Groww, Upstox, smallcase): screeners, factor and thematic baskets, index funds and ETFs built directly on portfolio theory; roles in quantitative and product analytics.
• Exams and certifications: MBA SAPM/investment management papers, CFA, CA Final SFM/FM, NISM Series V-A, X-A/X-B and XV, and selection tests at AMCs and research firms.

Interview questions

• Q: What is the difference between systematic and unsystematic risk? A: Systematic risk arises from market-wide factors, is measured by beta and cannot be diversified away, so it is the only risk that earns a premium; unsystematic risk is firm-specific and disappears in a well-diversified portfolio.
• Q: Why does diversification reduce risk? A: Because portfolio variance depends on covariances; whenever the correlation between assets is less than +1, offsetting movements lower portfolio σ below the weighted average of individual σ's.
• Q: What is the difference between the CML and the SML? A: The CML plots expected return against total risk (σ) and applies only to efficient portfolios combining Rf with the market portfolio; the SML plots expected return against beta and applies to every security or portfolio, efficient or not.
• Q: A stock plots below the SML - what do you do? A: It is overpriced: its expected return is less than the CAPM-required return for its beta, so you sell or avoid it and expect its price to fall until the return rises to the SML.
• Q: When would you use Treynor rather than Sharpe? A: Treynor, which divides excess return by beta, suits a fund being added to an already diversified portfolio where only systematic risk matters; Sharpe, which uses total risk, suits judging a portfolio held on its own.
• Q: Can a portfolio have a beta of 1 and still underperform the index? A: Yes - matching beta only matches systematic exposure; poor stock selection, high expenses and residual unsystematic risk can still produce a negative alpha.
• Q: What are the main criticisms of CAPM? A: Unrealistic assumptions (homogeneous expectations, costless borrowing at Rf), beta instability over time, Roll's critique that the true market portfolio is unobservable, and empirical evidence that size and value factors explain returns beyond beta.
• Q: How is beta calculated in practice for an Indian stock? A: Regress the stock's periodic returns on Nifty 50 (or Sensex) returns over, typically, 2-5 years of monthly or weekly data; the slope is beta, computed as Cov(stock, index)/Var(index).
• Q: If markets are semi-strong efficient, is equity research useless? A: Research on public information cannot earn consistent abnormal returns, but it still sets fair prices, and anomalies, information-processing speed and private insight leave room for skill - which is why passive funds and active managers coexist.
• Q: What is rupee cost averaging and why do SIPs use it? A: Investing a fixed rupee amount periodically buys more units when prices are low and fewer when high, so the average cost per unit falls below the average price and the timing decision is removed.
• Q: Difference between fundamental and technical analysis? A: Fundamental analysis estimates intrinsic value from economy, industry and company data for a longer horizon; technical analysis reads price and volume patterns to time short-term entries and exits.
• Q: What does APT add over CAPM? A: It allows several systematic factors instead of one market factor and rests only on the absence of arbitrage rather than mean-variance preferences - but it does not tell you what the factors are.
• Q: How do you pick between two funds with the same return? A: Compare risk-adjusted measures - Sharpe, Treynor, alpha, information ratio - plus expense ratio, consistency, benchmark (TRI), tracking error and the manager's mandate.

Common mistakes

• Using the CML when the question is about an individual security - an individual stock's required return comes from the SML, not the CML.
• Averaging standard deviations: σp is never the weighted average of σ's unless ρ = +1; the covariance term must be computed.
• Forgetting to square the weights, or omitting the factor 2 in the 2w₁w₂ρσ₁σ₂ term.
• Writing CAPM as Rf + β × Rm instead of Rf + β(Rm − Rf) - the beta multiplies the *premium*, not the market return.
• Mixing percentages and decimals in Gordon's model (D/(k−g) with k and g in % gives an answer a hundred times too small or large).
• Confusing Sharpe and Treynor denominators, or comparing a Sharpe ratio of one fund with the Treynor ratio of another.
• Treating a positive alpha from the past as guaranteed future outperformance, and ignoring expense ratios and survivorship bias in fund comparisons.
• Claiming a stock with high σ must have high beta - a stock can have huge total risk that is almost entirely unsystematic.

Related topics

Equity research and valuation: DCF and multiples · Cost of capital and capital structure · Financial markets and instruments · Fixed income: bond pricing, yields and duration · Behavioural finance · Financial risk management: market, credit and operational risk