Country Risk Premium: Pricing the Jurisdiction

A country risk premium is the extra yearly return an equity investor demands because the cash flows are earned in a jurisdiction that carries more risk than a mature one. The premium is added to the mature market equity risk premium, and the sum is then multiplied by the beta. In this worked example the country premium is 1.50 per cent, taking the total premium to 5.00, and that 1.50 is an assumed figure rather than one read off a market.

Underneath that sits an admission worth making early. The standard route from a beta to a cost of equity prices one thing: how much a company’s equity return moves when the market’s does. The beta has nothing whatever to say about the risk that the market itself sits somewhere with slower courts, a thinner capital accountThe side of a country’s external accounts covering investment flowing in and out, as against the trade in goods and services., a currency that lurches, or a government that borrows on worse terms than a mature marketA country whose capital markets are deep, whose institutions have been settled for a long time, and whose government borrows on the finest terms going. one. Jurisdiction risks are real, they land on every business in the jurisdiction at the same moment, and the standard route prices none of them. The country risk premium is a patch, added outside the model rather than derived inside it, and every honest treatment of it starts by saying so.

Consider a bicycle repair shop that has run for twenty years on a lane outside a school. Its risk is one thing set against the tea stall three doors down, and quite another when the whole lane is scheduled for demolition. The first comparison is what a beta measures. The second is what a country risk premium is trying to price, and no amount of care with the first will ever produce the second.

What is the premium actually paying for?

The premium pays for the conditions under which the money is earned, not for the company that earns it. Five things sit inside it, and they overlap: whether a contract can be enforced and how long enforcing it takes; the chance of assets or rights being taken by the state, or expropriationA government taking assets or rights away from whoever holds them, with or without payment.; official limits on moving money across the border, or capital controlsOfficial limits on how much money may be moved into or out of a country, and by whom.; the government’s own creditworthinessHow likely a borrower is judged to be to keep meeting what it promised to pay, on the dates it promised to pay it. as a borrower; and the extra swing in returns that comes with all four together.

Every one of those five lands on every business in the jurisdiction at once. A company-specific adjustment cannot carry it. It has to sit in the rate every company in that jurisdiction is discounted at. If the courts slow down, they slow down for the valve maker and for the coatings shop and for the tooling associate simultaneously. There is no diversifying that away by holding more of them.

And what is it not paying for?

Three things get pushed into this premium that do not belong in it, and the first is currency. Currency risk is about the unit the cash flows are measured in. Currency risk is handled by matching the rate to that unit: rupee cash flows are discounted at a rupee rate, and how a flow in one currency becomes a flow in another is a subject of its own, covered separately. Country risk is about the conditions under which the cash flows are earned at all. The two travel together often enough that people treat them as one thing, and treating them as one thing produces either a gap or a duplicate every single time.

The second is a forecast. A country risk premium is not a prediction that anything in particular will happen. The premium is compensation for the spread of outcomes that becomes possible, most of which never arrive. Where the view is that one specific event is coming, that view belongs in the cash flows as a scenario with a probability on it, not in the rate as a premium.

The third is the company’s own operating risk. The beta is already carrying that risk, and adding it again duplicates something the model has priced. A valve maker with lumpy orders is riskier than one with a service contract book, and the beta is where that difference lives. The country premium does not know one company from another.

Where does the premium enter, and why is it added to the equity premium?

The premium enters at one place and one place only: it is added to the mature market equity risk premium, and the sum of the two is what gets multiplied by the beta. For Sankalp Industrial Systems Limited, an invented manufacturer, the mature market premium is 3.50 per cent and the country premium is 1.50 per cent, so the total equity risk premium is 5.00 per cent. Both figures are assumptions, chosen so the arithmetic stays round. A live estimate would sit somewhere either side of each of them.

Why inside the bracket rather than outside it? Two reasons, and both bite. The first is definitional. The base rate is meant to be what a lender earns on a promise that will certainly be kept, priced in the same currency the cash flows arrive in. Load a country adjustment onto that and the number stops being what its own name says it is, and nobody downstream can tell what has been done to it.

The second reason is arithmetic. Inside the bracket the 1.50 points gets multiplied by the beta of 1.25 and becomes 1.875 points of cost of equity. Outside it the 1.50 stays 1.50. The placement is not housekeeping, and the two are simply different estimates. Run it through: 7.75 plus 1.25 times 5.00 is 14.00 per cent, against 7.75 plus 1.50 plus 1.25 times 3.50, or 13.625 per cent. A gap of 0.375 points of cost of equity turns entirely on which side of a bracket a number was typed.

What is 1.50 points actually worth?

Here is where a small-sounding assumption stops sounding small. The levered beta of 1.25 acts on the premium before anything else does, turning 1.50 into 1.875 points of a cost of equity that comes to 14.00 per cent. Then the 75.0 per cent equity weight acts on that figure, and 0.75 times 1.875 gives 1.40625 points of a blended rate of 12.00. Those 1.40625 points are 11.72 per cent of the whole rate, just under one eighth. Nearly an eighth of this company’s cost of capital is a single assumption that took one cell to type.

The borrowing side is untouched by all of this. Debt costs 6.00 per cent once the interest deduction is taken, carries the 25.0 per cent weight and contributes 1.50 points. The country premium is an equity premium, so no part of it reaches the borrowing side. The country premium is therefore scaled up once by the beta and scaled down once by the weight. That is 1.25 times 0.75, or 0.9375, the exact multiplier from a point of country premium to a point of blended rate.

Now push it into a value. Take the locked free cash flow to the firm for Sankalp Industrial Systems Limited, invented: Rs 98,00,00,000 in Year 1, rising by exactly Rs 18,00,00,000 a year to Rs 1,70,00,00,000 in Year 5, then a Year 6 flow of Rs 2,04,75,00,000 growing at 5.00 per cent for ever. Discount all of it at 12.00 per cent, year-end, and it comes to Rs 21,28,13,79,094. In crore, and rounding, that is Rs 2,128.14. Strip the country premium out entirely and nothing about the business changes at all, yet the same stream is worth Rs 26,99,29,87,444, a difference of Rs 5,71,16,08,351 or 26.84 per cent.

Why does a rate change that small move a value that much? Because the rate does its work in the divisor of a perpetuity, and that perpetuity carries 77.99 per cent of the answer. The divisor of a perpetuity is the rate net of growth, so the width of that gap is what does all the work, and here it begins only 7.00 percentage points wide. Shaving the rate widens the numerator of the fraction and shrinks its denominator in one move. Stripping the premium out swells the terminal block until it carries 81.96 per cent of a much larger total.

Route one: what does the government’s own borrowing show?

So where does a figure like 1.50 come from? There are three defensible routes and they are genuinely different measurements, not three ways of doing the same sum.

The first route takes the extra yield the jurisdiction’s government pays over a mature market government borrowing for the same length of time, or the spread that its credit rating typically attracts. The spread is observable, it costs nothing to obtain, and it is where most estimates in practice begin. Somebody has already done the work of pricing that government’s promises, and the market’s answer is sitting there in public.

The objection to it is exact rather than vague. A sovereign spread prices the risk of a government failing to pay a bond. Equity holders are not bondholders. Equity holders rank behind bondholders, they get paid out of whatever is left, and their claim is on a business rather than on a promise. In the same jurisdiction, equity moves a great deal more than government borrowing does, so a number taken from the borrowing market is measuring something narrower than the thing an equity investor is being asked to carry. Where a live spread is needed, practitioners go to Aswath Damodaran, whose valuation material is published at pages.stern.nyu.edu.

Route two: how is a bond number turned into an equity one?

The second route answers the first route’s objection head on, and it is Aswath Damodaran’s. Take the same sovereign default spread, then multiply it by the ratio of the standard deviationOne number summarising how widely a series scatters around its own average. A wider scatter gives a bigger number. of the country’s equity market to the standard deviation of its government bond market. Equities swing more than bonds do. The ratio therefore comes out above one, and it scales the premium up by exactly the amount by which the equity market is the noisier of the two.

The scaling is doing one job: it converts a number measured on debt into a number appropriate to equity, using the two markets’ own relative movement as the conversion rate. The scaling is a real argument rather than a fudge factor, and it is why this route is the one most widely used by people who publish country premiums for a living.

Its objection is the ratio itself. A standard deviation has to be measured over some measurement windowThe stretch of past data an estimate is calculated over. Whoever makes the estimate chooses it, and the choice moves the answer., and whoever is doing the measuring chooses that window. Move it and the answer moves. Worse, the ratio jumps after any stretch of market stress and settles during quiet spells, so a market that has been calm for two years produces a smaller premium at precisely the moment risk may be accumulating. The route is defensible and its input is unstable, and both things are true at once.

Route three: what are today’s prices already implying?

The third route ignores governments and ratings altogether. Take a market index, take a forecast of the cash flows the companies in it will produce, and solve for the single rate that makes those cash flows worth the index’s current level. The solved rate less the base rate is the equity risk premium the market itself is implying. Do that for the jurisdiction and again for a mature market, and the difference between the two implied premiums is the country premium.

The third route looks forward rather than back, needs no rating agency and no bond market, and moves with prices on the day rather than with an opinion published last quarter. If sentiment about a jurisdiction turns, this route registers it immediately. The route is reading the prices in which that turn has already happened.

The objection is the forecast. Solving for an implied premium requires a projection of cash flows for an entire index, and whoever supplies that projection is supplying the answer along with it. Push the assumed growth up and the implied premium falls; push it down and the premium rises. The route is noisy, it needs data the other two do not, and its output inherits every weakness of the growth assumption fed into it.

Three defensible routes give three answers. Then what?

Nobody rules between them. An analyst can size the disagreement instead of arguing about it, and the arithmetic above makes that easy. Half a point of country premium is 0.5 times 0.9375, or 0.47 points of the blended rate. On this company, half a point below the base case is worth Rs 162.85 crore of value and half a point above it is worth Rs 142.23 crore. A disagreement between two respectable estimation routes is therefore worth roughly a hundred and fifty crore on a company of this size. The choice of route is a valuation input, not a technicality.

Notice that the two half-point steps are not equal. Going down half a point buys more value than going up half a point costs. The relationship is a curve rather than a line. The rate sits in a denominator, and every further point of premium bites a little less than the point before it. The table below is the same curve as a ladder, and it is worth reading across the rows before reading down them.

The shaded row is the worked example. Every rupee figure in the last column comes from discounting one unchanged stream of cash at the rate in the column beside it, so the whole ladder is a statement about a rate and never about a business.

Does every company in a jurisdiction carry the same premium?

No, and the question is also Damodaran’s. Exposure to a jurisdiction is about where the cash flows are earned, not about where the head office keeps its registered address. Two companies can share a country of registration, a stock exchange listing and an auditor, and still be exposed to that country to entirely different degrees. One of them sells everything at home. The other ships most of its output abroad and gets paid by customers in mature markets.

Think about two grocery shops on the same street. One sells only to the households on that street; the other has a delivery arrangement with offices across the city. If the street is dug up for six months, the first one is in serious trouble and the second one is inconvenienced. Same street, same municipal ward, very different exposure to what happens on it. Weighting the premium by the share of revenue actually earned in the jurisdiction is the standard way of handling this, and it changes the rate for the second shop while leaving the first shop’s alone.

Every one of Sankalp Industrial Systems Limited’s operations sits in one jurisdiction, so the whole 1.50 per cent is applied to the whole company. An unweighted premium is a simplification, and it is the right one for this company only because the exposure genuinely is total.

What happens when the premium is taken to nil?

The arithmetic above can be checked by hand. The control below moves one number and one number only, from nil to 3.00 per cent. The five cash flows, the beta, the borrowing cost, the tax rate, the weights and the terminal growth are all frozen at the values the worked example gave them.

Two things are worth noticing while the control is moving. The first is the readout for a further tenth of a point: it is Rs 46.93 crore when the premium is at nil and Rs 20.74 crore when it is at 3.00, so the same tenth of a point buys over two times as much at the left end of the control as at the right. The second is the direction of travel. The relationship only runs one way: more premium, less value, always.

What if the premium is typed onto the base rate instead?

The second failure is quieter and it happens in a single keystroke. The premium gets typed onto the base rate instead of onto the equity premium. Nobody notices. The model still runs, and the answer still looks like a cost of capital.

The slip costs two things. The definitional cost has already been named: the base rate stops being a default-free rate in the currency of the cash flows, and every premium stacked above it now starts from a floor nobody recorded. A premium sitting on the base rate never meets the beta of 1.25, and one sitting inside the bracket does. The arithmetic cost of the slip is 0.375 points of cost of equity and 0.28125 points of the blended rate.

How do practitioners handle this in a live model?

In practice nobody treats a country premium as a number to be discovered. Practitioners treat it as a stated assumption with a name attached, and the working habits follow from that.

An equity analyst publishing a target range will name the route used and the date the input was taken. A premium taken from an implied calculation last quarter and one taken from a scaled spread this morning are not the same object. A reader who cannot tell which was used cannot tell what the number means. The same analyst runs the model at two or three premiums rather than one, and reports the resulting spread of values as the output. A single value carrying a single premium hides the fact that the premium was the largest judgement in the whole model.

A corporate finance team setting internal hurdle rates for units in several jurisdictions faces the exposure question from the other end. The instinct is to give every unit the group rate. The habit is administratively easy and quietly wrong for any unit earning its revenue somewhere other than where the group is registered. The discipline is to ask, for each unit, where the cash actually arrives, and to weight the premium accordingly.

A buyer running an acquisition case does one more thing: it asks the seller which premium is in the seller’s own model. Two parties can agree on every line of a forecast and still be a quarter apart on value, exactly as the arithmetic above shows. When a negotiation stalls on price with no disagreement about the business, the country premium is one of the first three places to look.

And everyone in that list writes the assumption down where a reviewer will see it, next to the base rate and the beta, rather than burying it inside a formula. An assumption in a cell of its own can be argued with. An assumption inside a formula cannot even be found.

Sankalp Industrial Systems Limited and Marudhar Valve Industries Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.