KARACHI: Read any serious valuation of a Pakistani company — a broker's initiation report, an investment bank's fairness opinion, a privatisation adviser valuing a state enterprise — and somewhere in the machinery you will find a percentage labelled the "required return" or "cost of equity". It is the return investors are assumed to demand for holding that share, and it quietly decides what the whole company is said to be worth. More often than not, that number comes from one of the most famous and most argued-about equations in finance: the capital asset pricing model, CAPM. This guide states the model plainly, works an example with fictional Pakistani inputs, shows where practitioners actually use it, and then gives the critics their full hearing.
The equation and its three ingredients
CAPM says that the expected return on a share equals the risk-free rate, plus the share's beta multiplied by the market's risk premium: Expected return = Rf + β × (Market return − Rf). The first ingredient, Rf, is the return on a riskless asset — for Pakistani work, usually a government treasury bill or a longer Pakistan Investment Bond yield, depending on the horizon. The second, beta, measures how strongly the share has moved with the broad market, as explained in beta explained: the market's beta is 1 by construction, so beta scales a company's exposure to the risk everyone shares. The third, the equity risk premium — the bracketed term — is the extra return investors are assumed to require from the stock market as a whole, above the safe rate, for enduring its swings.
The logic underneath is elegant. CAPM argues that investors can diversify away company-specific risk — the factory fire, the lost contract — simply by holding many shares, so the market will not pay extra return for risk you could have removed for free. The only risk that earns a premium is the undiversifiable kind: sensitivity to the whole market's fortunes, measured by beta. A share's required return is therefore the safe rate plus a payment for its dose of that one systematic risk. Elegant, teachable, and — as the critics will shortly insist — a drastically simplified cartoon of how markets actually behave. Both things are true, and practitioners use the model in full knowledge of both.
A worked example with fictional Pakistani inputs
Suppose, purely for illustration, that the yield on long-term government bonds stands at 12 per cent, and an analyst estimates the Pakistani equity risk premium at 6 per cent — that is, investors are assumed to require 18 per cent in total from the average, beta-1.0 stock. Now consider two fictional companies. Example Foods, a steady staples business, has a beta of 0.7: CAPM gives a required return of 12 + 0.7 × 6 = 16.2 per cent. Example Cement, a cyclical builder's supplier, has a beta of 1.4: 12 + 1.4 × 6 = 20.4 per cent. The model is telling the analyst that a rupee of Example Cement's future earnings must be discounted more harshly than a rupee of Example Foods', because Cement's shareholders are being compensated for a wilder ride. Feed both companies the same hypothetical earnings stream and the steadier one receives the higher valuation — not because its business is better, but because its returns arrive with less market risk attached.
Watch how sensitive the answer is to the ingredients, because that sensitivity is where professional arguments actually happen. Raise the assumed risk premium from 6 to 7 per cent and Example Cement's required return climbs to 21.8 per cent; use a short-term T-bill yield instead of a long bond yield for Rf and both numbers shift again. Two honest analysts valuing the same PSX company can land several percentage points apart through input choices alone — which is why good reports disclose the Rf, the beta source and the premium assumption rather than presenting the final percentage as a fact of nature. When you read a target price, you are reading someone's CAPM whether they label it or not.
Where CAPM is actually used
The model's first workplace is valuation. In a discounted cash flow (DCF) analysis, forecast cash flows are discounted at a rate built from CAPM's cost of equity blended with the after-tax cost of debt — the weighted average cost of capital. Raise that discount rate and today's value falls; lower it and value rises. The dividend discount model, examined in the dividend discount model explained, is CAPM's close cousin: it asks what dividend stream justifies today's price at the required return. The second workplace is inside companies: a Pakistani conglomerate deciding whether a new plant, brand or acquisition clears its hurdle rate starts from a CAPM-derived cost of capital, so the model's assumptions ripple into real investment decisions, not just analyst spreadsheets.
The third workplace is performance judging. CAPM defines the return a fund should have earned given its beta; performance beyond that line — positive "alpha" — is the portion attributable to skill (or luck) rather than to simply riding a rising market with a leveraged character. When a local equity fund trumpets beating the KSE-100, the CAPM question follows automatically: what was its beta while doing so? A fund with a beta of 1.3 in a rising market should beat the index handsomely, and charging active fees for delivering arithmetic is the quiet scandal the model exposes. The risk-adjustment theme continues in the Sharpe ratio explained, which judges the same performance from the investor's seat.
Regulated industries supply a fourth, less visible workplace with direct consequences for household bills. When Pakistan's energy regulators determine the return a utility or gas company is allowed to earn on its asset base, the building blocks on the table are recognisably CAPM's: a risk-free anchor, a beta for the regulated business, and a premium for equity risk. The allowed return then flows into tariffs, which flow into the monthly bills of millions of consumers and the cost structures of every factory on the network. Few consumers reading a tariff notification realise a 1960s asset-pricing model helped set the number on it — but the model's fingerprints are there, one more reason its assumptions deserve public scrutiny rather than technocratic obscurity.
The Pakistani wrinkles
Applying CAPM in Pakistan requires judgement calls that textbook treatments glide past. Start with the risk-free rate: Pakistani government paper is the least risky rupee asset available, but with sovereign credit ratings deep in speculative territory and a history of inflation eroding real returns, it is "risk-free" only in the narrow sense of nominal repayment. Some practitioners therefore build the required return from a mature-market base and add a country risk premium for Pakistan — a method imported from international valuation practice, visible whenever foreign investors price PSX assets. Others keep everything in rupees with local bond yields. Both approaches are defensible; mixing them unknowingly is not.
Beta presents its own local problems, rehearsed in the beta guide: thin trading in many PSX names makes measured betas unreliable, and short listing histories leave some companies almost no usable data — analysts then substitute the beta of a comparable company or a sector average, another judgement layer. The equity risk premium is the softest number of all: estimated from Pakistan's own market history it swings with the chosen window, because the PSX has lived through euphoric and catastrophic decades; estimated from global data it may not reflect a frontier market at all. The honest conclusion is that a Pakistani CAPM output is a range wearing a point estimate's clothing, and decisions — privatisation reserve prices, tariff determinations, board hurdle rates — should be stress-tested across that range, as regulators examining utility returns have reason to know.
Assembling a Pakistani required return, step by step
It helps to watch the whole assembly once, slowly, with clearly fictional inputs. Step one: choose the risk-free rate. For a valuation in rupees over a long horizon, an analyst might take the yield on a ten-year Pakistan Investment Bond — say 13 per cent in our illustration — reasoning that a long-lived asset deserves a long-maturity anchor rather than this quarter's T-bill rate. Step two: obtain a beta. The target company trades thinly, so the analyst averages the betas of three comparable listed firms and arrives at 1.1. Step three: set the equity risk premium. Drawing on long-run local market history adjusted for survivorship flattery, the analyst settles on 6 per cent. The required return emerges: 13 + 1.1 × 6 = 19.6 per cent. Step four — the one amateurs skip — is sensitivity: at a premium of 5 per cent the answer is 18.5; at a beta of 1.3 it is 20.8. The deliverable is the range 18.5–20.8 per cent, and any valuation that flips from "buy" to "sell" inside that range is not a valuation conclusion at all, but a coin toss wearing a suit.
Two further assembly choices separate careful Pakistani work from careless. Currency consistency: a model built in rupees must use rupee cash flows and a rupee discount rate, while a dollar-based build — common when foreign funds appraise PSX assets — must use dollar forecasts and add Pakistan's country risk explicitly rather than smuggling it into the beta. And tax consistency: cash flows stated after corporate tax must meet a discount rate built from after-tax components, while the cost of debt entering a blended WACC uses the after-tax rate because interest is deductible. These sound like bookkeeping pedantries until you watch them move a privatisation valuation by a fifth — at which point they become the whole argument, and the reason fairness opinions devote pages to assumptions that press summaries omit entirely.
The critics: what half a century of testing found
CAPM's empirical record is genuinely mixed. Classic studies found the actual relationship between beta and return flatter than the model predicts: low-beta shares earned more, and high-beta shares less, than they "should". Later research identified other characteristics — company size, and the value-versus-growth distinction — that seemed to explain returns better, spawning multi-factor models used across the industry today; the value factor's Pakistani flavour appears in growth versus value on the PSX. Roll's famous critique adds a philosophical sting: the true "market portfolio" of all assets is unobservable, so the model can never be cleanly tested at all. Defenders reply that CAPM was never a law of physics — it is a disciplined language for separating the return you demand for time (the risk-free rate) from the return you demand for market risk, and even its rivals keep its core intuition that diversified investors are paid for systematic exposure, not for risks they could have diversified away.
For the individual investor, the practical residue is modest but real. Use CAPM thinking, not CAPM arithmetic: expect steadier companies to be priced for lower required returns and wilder ones for higher; be suspicious when a hot stock's story implies high expected returns with low risk, since the model names that combination as the thing markets arbitrage away; and when an analyst's target depends on a required return, ask what Rf, beta and premium produced it — the answer tells you how much of the target is analysis and how much is assumption. Keep one more perspective: the model also explains why diversification felt free in the theory — because, within its assumptions, it genuinely is, converting company-specific gambles into market exposure without surrendering expected return. Every later refinement of asset pricing is an argument about the assumptions; the habit of separating paid risk from unpaid risk survives them all. This article is educational content, not financial advice: models organise thinking, but they do not remove the need to read the actual company behind the equation.
Key takeaway: CAPM sets a share's required return as the risk-free rate plus beta times the market risk premium — payment for time, plus payment for undiversifiable market risk. In Pakistan every ingredient involves judgement, so treat any single output as the middle of a range.
