At the start of July, I updated my estimates of equity risk premiums for countries, in an semiannual ritual that goes back almost three decades. As with some of my other data updates, I have mixed feelings about publishing these numbers. On the one hand, I have no qualms about sharing these estimates, which I use when I value companies, because there is no secret sauce or special insight embedded in them. On the other, I worry about people using these premiums in their valuations, without understanding the choices and assumptions that I had to make to get to them. Country risk, in particular, has many components to it, and while you have to ultimately capture them in numbers, I wanted to use this post to draw attention to the many layers of risk that separate countries. I hope, and especially if you are a user of my risk premiums, that you read this post, and if you do have the time and the stomach, a more detailed and much longer update that I write every year. Country Risk - Dimensions     When assessing business risk from operating in a country, you will be affected by uncertainty that arises from almost every source, with concerns about political structure (democracies have very different risk profiles than authoritarian regimes), exposure to violence (affecting both costs and revenues),  corruption (which operates an implicit tax) and legal systems (enforcing ownership rights) all playing out in business risk. I will start with political structure, where the facile answer is that it less risky to operate a business in a democracy than in an authoritarian regime, but where the often unpalatable truth is that each structure brings its own risks. With democracies, the risk is that newly elected governments can revisit, modify or discard policies that a previous government have adopted, requiring businesses to adapt and change to continuous changes in policy. In contrast, an authoritarian government can provide long term policy continuity, with the catch being that changes in the government, though infrequent, can create wrenching policy shifts that businesses have to learn to live with. Keeping the contrast between the continuous risk of operating in a democracy and the discontinuous risk in an authoritarian structure in mind, take a look at this picture of how the world looked in terms of democracy leading into 2025: Source: Economist Intelligence Unit (EIU) It is worth noting that there are judgment calls that the Economist made in measuring democracy that you and I might disagree with, but not only is a large proportion of the world under authoritarian rule, but the trend lines on this dimension  also have been towards more authoritarianism in the last decade.         On the second dimension, exposure to violence, the effects on business are manifold. In addition to the threat that violence can affect operations, its presence shows up as higher operating costs (providing security for employees and factories) and as insurance costs (if the risks can be insured). To measure exposure to violence, from both internal and external sources, I draw on measures developed and updated by the Institute of  Economics & Peace across countries in 2024: Institute of Economics & Peace The Russia-Ukraine war has caused risk to flare up in the surrounding states and the Middle East and central Africa continue to be risk cauldrons, but at least according to the Institute's measures, the parts of the world that are least exposed to violence are in Northern Europe, Australia and Canada. Again, there are judgments that are made in computing these scores that will lead you to disagree with specific country measures (according the Peace Institute, the United States and Brazil have higher exposures to violence than Argentina and Chile, and India has more exposure to violence than China), but the bottom line is that there are significant differences in exposure to violence across the world.          Corruption is a concern for everyone, but for businesses, it manifests in two ways. First, it puts more honest business operators at a disadvantage in a corrupt environment, since they are less willing to break the rules and go along with corrupt practices than their less scrupulous competitors. Second, even for those businesses that are willing to play the corruption game, it creates costs that I would liken to an implicit tax that reduces profits, cash flows and value. The measure of corruption that I use comes from Transparency International, and leading into July 2025, and the heat map below captures corruption scores (with higher scores indicating less corruption), as well as the ten most and least corrupt countries in the world:  Transparency International As you can see from the map, there are vast swaths of the world where businesses have to deal with corruption in almost every aspect of business, and while some may attribute this to cultural factors, I have long argued that corruption almost inevitably follows in bureaucratic settings, where you need licenses and approvals for even the most trivial of actions, and the bureaucrats (who make the licensing decisions) are paid a pittance relative to the businesses that they regulate.           As a final component, I look at legal systems, especially when it comes to enforcing contractual agreements and property rights, central to running successful businesses. Here, I used estimates from the IPRI, a non-profit institution that measures the quality of legal systems around the world. In their latest rankings from 2024, here is how countries measured up in 2024: Property Rights Alliance In making these assessments, you have to consider not just the laws in place but also the timeliness with which these laws get enforced, since a legal system where justice is delayed for years or even decades is almost as bad as one that is capricious and biased.  Country Risk - Measures     The simplest and most longstanding measure of country risk takes the form of sovereign ratings, with the same agencies that rate companies (S&P, Moody's and Fitch) also rating countries, with the ratings ranging from Aaa (safest) to D (in default). The number of countries with sovereign ratings available on them has surged in the last few decades; Moody’s rated 13 countries in 1985, but that number increased to 143 in 2025, with the figure below listing the number of rated countries over time: Note that that the number of Aaa rated countries stayed at eleven, even while more countries were rated, and has dropped from fifteen just a decade ago, with the UK and France losing their Aaa ratings during that period. In May 2025, Moody's downgraded the United States, bringing them in line with the other ratings agencies; S&P downgraded the US in 2011 and Fitch in 2023. The heat map below captures sovereign ratings across the world in July 2025: Moody's While sovereign ratings are useful risk measures, they do come with caveats. First, their focus on default risk can lead them to be misleading measures of overall country risk, especially in countries that have political risk issues but not much default risk; the Middle East, for instance, has high sovereign ratings. Second, the ratings agencies have blind spots, and some have critiqued these agencies for overrating European countries and underrating Asian, African and Latin American countries. Third, ratings agencies are often slow to react to events on the ground, and ratings changes, when they do occur, often lag changes in default risk.     If you are leery about trusting ratings agencies, I understand your distrust, and there is an alternative measure of sovereign default risk, at least for about half of all countries, and that is the sovereign credit default swap (CDS) market, which investors can buy protection against country default. These market-determined numbers will reflect events on the ground almost instantaneously, albeit with more volatility than ratings. At the end of June 2025, there were about 80 countries with sovereign CDS available on them, and the figure below captures the values: The sovereign CDS spreads are more timely, but as with all market-set numbers, they are subject to mood and momentum swings, and I find using them in conjunction with ratings gives me a better sense of sovereign default risk.     If default risk seems like to provide too narrow a focus on countr risk, you can consider using country risk scores, which at least in principle, incorporate other components of country risk. There are many services that estimate country risk scores, including the Economist and the World Bank, but I have long used Political Risk Services (PRS) for my scores.. The PRS country risk scores go from low to high, with the low scores indicative of more country risk, and the table below captures the world (at least according to PRS): Political Risk Services (PRS) There are some puzzling numbers here,  with the United States coming in as riskier than Vietnam and Libya, but that is one reason why country risk scores have never acquired traction. They vary across services, often reflecting judgments and choices made by each service, and there is no easy way to convert these scores into usable numbers in business and valuation or compare them across services.      Country Risk - Equity Risk Premiums     My interest in country risk stems almost entirely from my work in corporate finance and valuation, since this risk finds its way into the costs of equity and capital that are critical ingredients in both disciplines. To estimate the cost of equity for an investment in a risky country. I will not claim that the approaches I use to compute equity risk premiums for countries are either original or brilliant, but they do have the benefit of consistency, since I have used them every year (with an update at the start of the year and mid-year) since the 1990s.      The process starts with my estimate of the implied equity risk premium for the S&P 500, and I make this choice not for parochial reasons but because getting the raw data that you need for the implied equity risk premium is easiest to get for the S&P 500, the most widely tracked index in the world. In particular, the process requires data on dividends and stock buybacks on the stocks in the index, as well as expected growth in these cash flows over time, and involves finding the discount rate (internal rate of return) that makes the present value of cash flows equal to the level of the index. On June 30, 2025, this assessment generated an expected return of 8.45% for the index: Download ERP spreadsheet Until May 2025, I just subtracted the US 10-year treasury bond rate from this expected return, to get to an implied equity risk premium for the index, with the rationale that the US T.Bond rate is the riskfree rate in US dollars. The Moody’s downgrade of the US from Aaa to Aa1 has thrown a wrench into the process, since it implies that the T.Bond rate has some default risk associated with it, and thus incorporates a default spread. To remove that risk, I net out the default spread associated with Aa1 rating from the treasury rate to arrive at a riskfree rate in dollars and an equity risk premium based on that: Riskfree rate in US dollars       = T.Bond rate minus Default Spread for Aa1 rating                                                             = 4.24% - 0.27% = 3.97% Implied equity risk premium for US = Expected return on S&P 500 minus US $ riskfree rate                                                             = 8.45% - 3.97% = 4.48% Note that this approach to estimating equity risk premiums is model agnostic and reflects what investors are demanding in the market, rather than making a judgment on whether the premium is right or what it should be (which I leave to market timers).        To get the equity risk premiums for other countries, I need a base premium for a mature market, i.e., one that has no additional country risk, and here again, the US downgrade has thrown a twist into the process. Rather than use the US equity risk premium as my estimate of the mature market premium, my practice in every update through the start of 2025, I adjusted that premium (4.48%) down to take out the US default spread (0.27%), to arrive at the mature market premium of 4.21%. That then becomes the equity risk premium for the eleven countries that continue to have Aaa ratings, but for all other countries, I estimate default spreads based upon their sovereign ratings. As a final adjustment, I scale these default spreads upwards to incorporate the higher risk of equities, and these become the country risk premiums, which when added to the mature market premium, yields equity risk premiums by country. The process is described below: Download spreadsheet The results from following this process are captured in the picture below, where I create both a heat map based on the equity risk premiums, and report on the ratings, country risk premiums and equity risk premiums, by country: Download equity risk premium, by country If you compare the equity risk premium heat map with the heat maps on the other dimensions of country risk (political and legal structures, exposure to violence and corruption), you will notice the congruence. The parts of the world that are most exposed to corruption and violence, and have capricious legal systems, tend to have higher equity risk premiums. The effects of the US ratings downgrade also manifest in the table, with the US now having a higher equity risk premium than its Aaa counterparts in Northern Europe, Australia and Canada. A User's Guide      My estimates of equity risk premiums, by country, are available for download, and I am flattered that there are analysts that have found use for these number. One reason may be that they are free, but I do have concerns sometimes that they are misused, and the fault is mine for not clarifying how they should be used. In this section, I will lay out steps in using these equity risk premiums in corporate finance and valuation practice, and  if I have still left areas of  grey, please let me know. Step 1: Start with an understanding of what the equity risk premium measures     The starting point for most finance classes is with the recognition that investors are collectively risk averse, and will demand higher expected returns on investments with more risk. The equity risk premium is a measure of the “extra” return that investors need to make, over and above the riskfree rate, to compensate for the higher risk that they are exposed to, on equities collectively. In the context of country risk, it implies that investments in riskier countries will need to earn higher returns to beat benchmarks than in safer countries. Using the numbers from July 2025, this would imply that investors need to earn 7.46% more than the riskfree rate to invest in an average-risk investment in India, and 10.87% more than the riskfree rate to invest in an average risk investment in Turkey.     It is also worth recognizing how equity risk premiums play out investing and valuation. Increasing the equity risk premium will raise the rate of return you need to make on an investment, and by doing so, reduce its value. That is why equity risk premiums and stock prices move inversely, with the ERP rising as stock prices drop (all other thins being held constant) and falling as stock prices increase.  Step 2: Pick your currency of analysis (and estimate a riskfree rate)     I start my discussions of currency in valuation by positing that currency is a choice, and that not only can you assess any project or value any company in any currency, but also that your assessment of project worth or company value should not be affected by that choice. Defining the equity risk premium as the extra return that investors need to make, over and above the risk free rate, may leave you puzzled about what riskfree rate to use, and while the easy answer is that it should be the riskfree rate in the currency you chose to do the analysis in, it is worth emphasizing that this riskfree rate is not always the government bond rate, and especially so, if the government does not have Aaa rating and faces default risk. In that case, you will need to adjust the government bond rate (just as I did with the US dollar) for the default spread, to prevent double counting risk.   Staying with the example of an Indian investment, the expected return on an average-risk investment in Indian rupees would be computed as follows: Indian government bond rate on July 1, 2025 = 6.32% Default spread for India, based on rating on July 1, 2025 = 2.16% Indian rupee risk free rate on July 1, 2025 = 6.32% - 2.16% = 4.16% ERP for India on July 1, 2025 = 7.46% Expected return on average Indian equity in rupees on July 1, 2025 = 4.16% + 7..46% = 11.62% Note also that if using the Indian government bond rate as the riskfree rate in rupees, you would effectively be double counting Indian country risk, once in the government bond rate and once again in the equity risk premium.     I know that the ERP is in dollar terms, and adding it to a rupee riskfree rate may seem inconsistent, but it will work well for riskfree rates that are reasonably close to the US dollar risk free rate. For currencies, like the Brazilian real or Turkish lira, it is more prudent to do your calculations entirely in US dollars, and convert using the differential inflation rate: US dollar riskfree rate on July 1, 2025 = 3.97% ERP for Turkey on July 1, 2025 = 10.87% Expected return on average Turkish equity in US $ on July 1, 2025 = 3.97% + 10.87% = 14.84% Expected inflation rate in US dollars = 2.5%; Expected inflation rate in Turkish lira = 20% Expected return on average Turkish equity Turkish lira on July 1, 2025 = 1.1484 *(1.20/1.025) -1 = 34.45% Note that this process scales up the equity risk premium to a higher number for high-inflation currencies. Step 3: Estimate the equity risk premium or premiums that come into play based on operations    Many analysts use the equity risk premiums for a country when valuing companies that are incorporated in that country, but I think that is too narrow a perspective. In my view, the exposure to country risk comes from where a company operates, not where it is incorporated, opening the door for bringing in country risk from emerging markets into the cost of equity for multinationals that may be incorporated in mature markets. I use revenue weights, based on geography, for most companies, but I am open to using production weights, for natural resource companies, and even a mix of the two.  In corporate finance, where you need equity risk premiums to estimate costs of equity and capital in project assessment, the location of the project will determine which country’s equity risk premiums come into play. When Amazon decides to invest in a Brazilian online retail project, it is the equity risk premium for Brazil that should be incorporated, with the choice of currency for analysis determining the riskfree rate.  Step 4: Estimate project-specific or company-specific risk measures and costs     The riskfree rate and equity-risk premiums are market-wide numbers, driven by macro forces. To complete this process, you need two company-specific numbers: Not all companies or projects are average risk, for equity investors in them, and for companies that are riskier or safer than average, you need a measure of this relative risk. At the risk of provoking those who may be triggered by portfolio theory or the CAPM, the beta is one such measure, but as I have argued elsewhere, I am completely at home with alternative measures of relative equity risk. The cost of equity is calculated as follows:  Cost of equity = Riskfree rate + Beta × Equity Risk Premium The beta (relative risk measure) measures the risk of the business that the company/project is in, and for a diversified investor, captures only risk that cannot be diversified away. While we are often taught to use regressions against market indices to get these betas, using industry-average or bottom-up betas yields much better estimates for projects and companies. For the cost of debt, you need to estimate the default spread that the company will face. If the company has a bond rating, you can use this rating to estimate the default spread, and if it is not, you can use the company's financials to assess a synthetic rating. Cost of debt =Riskfree Rate + Default spread Harking back to the discussion of riskfree rates, a company in a country with sovereign default risk will often bear a double burden, carrying default spreads for both itself and the country. The currency choice made in step two will hold, with the riskfree rate in both the cost of equity and debt being the long-term default free rate in that currency (and not always the government bond rate). Step 5: Ensure that your cash flows are currency consistent      The currency choice made in step 2 determines not only the discount rates that you will be using but also the expected cash flows, with expected inflation driving both inputs. Thus, if you analyze a Turkish project in lira, where the expected inflation rate is 20%, you should expect to see costs of equity and capital that exceed 25%, but you should also see growth rates in the cash flows to be inflated the same expected inflation. If you assess the same project in Euros, where the expected inflation is 2%, you should expect to see much lower discount rates, high county risk notwithstanding, but the expected growth in cash flows will also be muted, because of the low inflation.     There is nothing in this process that is original or path-breaking, but it does yield a systematic and consistent process for estimating discount rates, the D in DCF. It works for me, because I am a pragmatist, with a valuation mission to complete, but you should feel free to adapt and modify it to meet your concerns.  YouTube Video Paper Country Risk Determinants: Determinants, Measures and Implications - The 2025 Edition Datasets Equity Risk Premiums, by country - July 2025 Country Risk Links EIU Democracy Index Global Peace Index (Exposure to Violence) Corruption Index International Property Rights Index Moody's Sovereign Ratings Political Risk Services (PRS) Country Risk Scores Spreadsheets Implied Equity Risk Premium for S&P 500 on July 1, 2025

Escape newsletter inbox chaos and algorithmic surveillance by building your own enshittification-proof newspaper from the writers you already read

How to resist thinking of large language models as friends. Or sentient things. Or intelligences you have to treat like God.

Daily Journal has been accused of incorrectly accounting for software development costs by an activist investor seeking a consulting contract.

The Trump administration’s regulatory whiplash has left prosecutors scrambling with misattributed chat messages and questionable victim testimony