In Search of Safe Havens: The Trust Deficit and Risk-free Investments!

In every introductory finance class, you begin with the notion of a risk-free investment, and the rate on that investment becomes the base on which you build, to get to expected returns on risky assets and investments. In fact, the standard practice that most analysts and investors follow to estimate the risk free rate is to use the government bond rate, with the only variants being whether they use a short term or a long term rate. I took this estimation process for granted until 2008, when during that crisis, I woke up to the realization that no matter what the text books say about risk-free investments, there are times when finding an investment with a guaranteed return can become an impossible task. In the aftermath of that crisis, I wrote a series of what I called my nightmare papers, starting with one titled, "What if nothing is risk free?", where I looked at the possibility that we live in a world where nothing is truly risk free. I was reminded of that paper a few weeks ago, when Fitch downgraded the US, from AAA to AA+, a relatively minor shift, but one with significant psychological consequences for investors in the largest economy in the world, whose currency still dominates global transactions. After the rating downgrade, my mailbox was inundated with questions of what this action meant for investing, in general, and for corporate finance and valuation practice, in particular, and this post is my attempt to answer them all with one post.

Risk Free Investments: Definition, Role and Measures

    The place to start a discussion of risk-free rates is by answering the question of what you need for an investment to be risk-free, following up by seeing why that risk-free rate plays a central role in corporate finance and investing and then looking at the determinants of that risk-free rate.

What is a risk free investment?

    For an investment to be risk-free, you have feel certain about the return you will make on it. With this definition in place, you can already see that to estimate a risk free rate, you need to be specific about your time horizon, as an investor. 

• An investment that is risk free over a six month time period will not be risk free, if you have a ten year time horizon. That is because you have reinvestment risk, i.e., the proceeds from the six-month investment will have to be reinvested back at the prevailing interest rate six months from now, a year from now and so on, until year 10, and those rates are not known at the time you take the first investment.
• By the same token, an investment that delivers a guaranteed return over ten years will not be risk free to an investor with a six month time horizon. With this investment, you face price risk, since even though you know what you will receive as a coupon or cash flow in future periods, since the present value of these cash flows, will change as rates change. During 2022, the US treasury did not default, but an investor in a 10-year US treasury bond would have earned a return of -18% on his or her investment, as bond prices dropped.

For an investment to be risk free then, it has to meet two conditions. The first is that there is no risk that the issuer of the security will default on their contractual commitments. The second is that the investment generates a cash flow only at your specified duration, and with no intermediate cash flows prior to that duration, since those cash flows will then have to be reinvested at future, uncertain rates. For a five-year time horizon, then, you would need the rate on a five-year zero default-free zero coupons bond as your risk-free rate.

    You can also draw a contrast between a nominal risk-free rate, where you are guaranteed a return in nominal terms, but with inflation being uncertain, the returns you are left with after inflation are no longer guaranteed, and a real risk-free rate, where you are guaranteed a return in real terms, with the investment is designed to protect you against volatile inflation. While there is an appeal to using real risk-free rates and returns, we live in a world of nominal returns, making nominal risk-free rates the dominant choice, in most investment analysis.

Why does the risk-free rate matter?

    By itself, a risk-free investment may seem unexceptional, and perhaps even boring, but it is a central component of investing and corporate finance:

1. Asset Allocation: Investors vary on risk aversion, with some more willing to take risk than others. While there are numerous mechanisms that they use to reflect their differences on risk tolerance, the simplest and the most powerful is in their choice on how much to invest in risky assets (stocks, corporate bonds, collectibles etc.) and how much to hold in investments with guaranteed returns over their time horizon (cash, treasury bill and treasury bonds).
2. Expected returns for Risky Investments: The risk-free rate becomes the base on which you build to estimate expected returns on all other investments. For instance, if you read my last post on equity risk premiums, I described the equity risk premium as the additional return you would demand, over and above the risk free rate. As the risk-free rate rises, expected returns on equities will be pushed up, and holding all else constant, stock prices will go down., and the reverse will occur, when risk-free rates drop.
3. Hurdle rates for companies: Using the same reasoning, higher risk-free rates push up the costs of equity and debt for all companies, and by doing so, raise the hurdle rates for new investments. As you increase hurdle rates, new investments will have to earn higher returns to be acceptable, and existing investments can cross from being value-creating (earning more than the hurdle rate) to value-destroying (earning less). 
4. Arbitrage pricing: Arbitrage refers to the possibility that you can create risk-free positions by combining holdings in different securities, and the benchmark used to judge whether these positions are value-creating becomes the risk-free rate. If you do assume that markets will price away this excess profit, you then have the basis for the models that are used to value options and other derivative assets. That is why the risk-free rate becomes an input into option pricing and forward pricing models, and its absence leaves a vacuum.

Determinants

    So, why do risk-free rates vary across time and across currencies? If your answer is the Fed or central banks, you have lost the script, since the rates that central banks set tend to be short-term, and inaccessible, for most investors. In the US, the Fed sets the Fed Funds rate, an overnight intra-bank borrowing rate, but US treasury rates, from the 3-month to 30-year, are set at auctions, and by demand and supply. To understand the fundamentals that determine these rates, put yourself in the shoes of a buyer of these securities, and consider the following:

1. Inflation: If you expect inflation to be 3% in the next year, it makes little sense to buy a bond, even if it is default free, that offers only 2%. As expected inflation rises, you should expect risk-free rates to rise, with or without central bank actions. 
2. Real Interest Rate: When you buy a note or a bond, you are giving up current consumption for future consumption, and it is fitting that you earn a return for this sacrifice. This is a real risk-free rate, and in the aggregate, it will be determined by the supply of savings in an economy and the demand for those savings from businesses and individuals making real investments. Put simply, economies with a surplus of growth investments, i.e., with more real growth, should see higher real interest rates, in steady state, than stagnant or declining economies.

The recognition of these fundamentals is what gives rise to the Fisher equation for interest rates or the risk free rate:

    Nominal Risk-free Rate = (1 + Expected Inflation) (1+ Real Interest Rate) -1 (or)

                                            =  Expected Inflation + Expected Real Interest Rate (as an approximation)

If you are wondering where central banks enter this equation, they can do so in three ways. The first is that central banking actions can affect expected inflation, at least in the long term, with more money-printing leading to higher inflation. The second is central banking actions can, at least at the margin, push rates above their fundamentals (expected inflation and real interest rates), by tightening monetary policy, and below their fundamentals by easing monetary policy. Since this is often achieved by raising or lowering the very short term rates set by the central bank, the central banking effect is likely to be greater at the shorter duration risk-free rates. The third is that central banks, by tightening or easing monetary policy, may affect real growth in the near term, and by doing so, affect real rates. 

    Having been fed the mythology that the Fed (or another central bank) set interest rates by investors and the media, you may be unconvinced, but there is no better way to show the emptiness of "the Fed did it" argument than to plot out the US treasury bond rate each year against a crude version of the fundamental risk-free rate, computed by adding the actual inflation in a year to the real GDP growth rate that year:

As you can see, the primary reasons why we saw historically low rates in the 2008-2021 time period was a combination of very low inflation and anemic real growth, and the main reason that we have seen rates rise in 2022 and 2023 is rising inflation. It is true that nominal rates follow a smoother path than the intrinsic risk free rates, but that is to be expected since the ten-year rates represent expected values for inflation and real growth over the next decade, whereas my estimates of the intrinsic rates represent one-year numbers. Thus, while inflation jumped in 2021 and 2022 to 6.98%, and investors are expecting higher inflation in the future, they are not expecting inflation to stay at those levels for the next decade.    

Risk Free Rate: Measurement

    Now that we have established what a risk-free rate is, why it matters and its determinants, let us look at how best to measure that risk-free rate. We will begin by looking at the standard practice of using government bond rates as riskfree rates, and why it collides with reality, move on to examine why governments default and end with an assessment of how to adjust government bond rates for that default risk.

Government Bond Rates as Risk Free

    I took my first finance class a long, long time ago, and during the risk-free rate discussion, which lasted all of 90 seconds, I was told to use the US treasury rate as a risk-free rate. Not only was this an indication of how dollar-centric much of finance education used to be, but also of how much faith there was that the US treasury was default-free. Since then, as finance has globalized, that lesson has been carried, almost unchanged, into other currencies, where we are now being taught to use government bond rates in those currencies as risk-free rates. While that is convenient, it is worth emphasizing two implicit assumptions that underlie why government bond rates are viewed as risk-free:

1. Control of the printing presses: If you have heard the rationale for government bond rates as risk-free rates, here is how it usually goes. A government, when it borrows or issues bonds in its local currency, preserves the option to print more money, when that debt comes due, and thus should never default. This assumption breaks down, of course, when countries share a common currency, as is the case with the dozen or more European countries that all use the Euro as their domestic currency, and none of them has the power to print currency at will. 
2. Trust in government: Governments that default, especially on their domestic currency borrowings, are sending a signal that they cannot be trusted on their obligations, and the implicit assumption is that no government that has a choice would ever send that signal. (Governments send the same signal when they default on their foreign currency debt/bonds, but they can at least point to circumstances out of their control for doing so.)

The problem with these assumptions is that they are at war with the data. As we noted in our country risk discussion, governments do default on their local currency borrowings and bonds, albeit at a lower rate than they do on their foreign currency obligations. 

If you are wondering why a government that has a choice of not defaulting would choose to default, it is worth remembering that printing more money to pay off local currency debt has a cost of its own, since it debases the currency, pushing up inflation. Inflation, especially when it becomes stratospheric, causes investors and consumers to lose trust in the currency, and given a choice between default and debasement, many governments choose the latter.

    Once you open the door to the possibility of sovereign default in a local currency, it stands to reason that a government bond rate in the local currency may not always yield a risk-free rate for that currency. It is also worth noting that until 2008, investors had that door firmly shut for some currencies, believing that some governments were so trustworthy that they would not even consider default. Thus, the notion that the US or UK governments would default on their debt would have been unthinkable, but the 2008 crisis, in addition to the financial damage it created, also opened up a trust deficit, which has made the unthinkable a reality. In fact, you would be hard pressed to find any government that is trusted the way it was prior to this crisis, and that loss of trust also implies that the clock is ticking towards expiration, for the "government bonds are risk free" argument.

When and Why Governments Default

    Now that we have established that governments can default, let’s look at why they default. The most obvious reason is economic, where a crisis and collapse in government revenues, from taxes and other sources, causes a government to be unable meet its obligations. The likelihood of this happening should be affected by the following factors:

1. Concentrated versus Diversified Economy: A government's capacity to cover its debt obligations is a function of the revenues it generates, and those revenues are likely to be more volatile in a country that gets its revenues from a single industry or commodity than it is in a country with a more diverse economy. One measure of economic concentration is the percent of GDP that comes from commodity exports, and the picture below provides that statistic, by country:

   

   Source: UNCTAD

   
   
   As you can see, much of Africa, Latin America, the Middle East and Asia are commodity dependent, effectively making them more exposed to default, with a downturn in commodity prices.
2. Degree of Indebtedness: As with companies, countries that borrow too much are more exposed to default risk than countries that borrow less. That said, the question of what to scale borrowing to is an open question. One widely-used measure of country indebtedness is the total debt owed by the country, as a percent of its GDP. Based on that statistic, the most indebted countries are listed below:
   
   As you can see, this table contains a mix of countries, with some (Venezuela, Greece and El Salvador) at high risk of default and others (Japan, US, UK, Canada and France) viewed as being at low risk of default. 
3. Tax Efficiency: It is worth remembering that governments do not cover debt obligations with gross domestic product or country wealth, but with their revenues, which come primarily from collecting taxes. Holding all else constant, governments with more efficient tax systems, where most taxpayers comply and pay their share, are less likely to default than governments with more porous tax systems, where tax evasion is more the rule than the exception, and corruption puts revenues into the hands of private players rather than the government.

There is a second force at play, in sovereign defaults. Ultimately, a government that chooses to default is making a political choice, as much as it is an economic one. When politics is functional, and parties across the spectrum share in the belief that default should be a last resort, with significant economic costs, there will be shared incentive in avoiding default. However, when politics becomes dysfunctional, and default is perceived as partisan, with one side of the political divide perceived as losing more from default than the other, governments may default even though they have the resources to cover their obligations.

    As a lender to a government, you may not care about why a government defaults, but economic defaults generally represent more intractable problems than defaults caused by political dysfunction, which tend to be solved once the partisan pounds of flesh are extracted. In my view, the ratings downgrades of the US government fall into the latter category, since they are triggered by a uniquely US phenomenon, which is a debt limit that has to be reset each time the total debt of the US approaches that value. Since that reset has to be approved by the legislature, it becomes a mechanism for political standoffs, especially when there is a split in executive and legislative power. In fact, the first downgrade of the US occurred more than a decade ago, when S&P lowered its sovereign rating for the US from AAA to AA+ in 2011, after a debt-limit standoff at the time. The Fitch downgrade of the US, this year, was triggered by a stand-off between the administration and Congress a few months ago on the debt-limit, and one that may be revisited in a few weeks again. 

Measuring Government Default Risk

    With that lead-in on sovereign default risk, let us look at how sovereign default risk gets measured, again with the US as the focus. The first and most widely used measure of default risk is sovereign ratings, where ratings agencies rate countries, just as they do companies, with a rating scale that goes from AAA (Aaa) down to D(default). Fitch, Moody's and S&P all provide sovereign ratings for countries, with separate ratings for foreign currency and local currency debt. With sovereign ratings, the implicit assumption is that AAA (Aaa) rated countries have negligible or no default risk, and the ratings agencies back this up with the statistic that no AAA rated country has ever defaulted on its debt within 15 years of getting a AAA rating. That said, the number of AAA (Aaa) rated countries has dropped over time, and there are only nine countries left that have the top rating from all three ratings agencies: Germany, Denmark, Netherlands, Sweden, Norway, Switzerland, Luxembourg, Singapore and Australia. Canada is rated AAA by two of the ratings agencies, and after the Fitch downgrade, the US is rated Aaa only by Moody's, whereas the UK is AAA rated only by S&P.

   In a reflection of the times, there have been two developments. The first is that the number of countries with the highest rating has dropped over time, as can be seen in the graph below of countries with Aaa ratings from Moody's: 

Second, even the ratings agencies have become less decisive about what a AAA sovereign rating implies for default risk, especially after the 2008 crisis, when S&P announced that not all AAA countries were equal, in terms of default risk, thus admitting that each ratings class included variations in default risk. 

    If you recognize that default risk falls on a continuum, rather than in the discrete classes that ratings assign, the sovereign CDS market gives you not only more nuanced estimates of default risk, but ones that are reflect, on an updated basis, what investors think about a country's default risk. The graph below contains the sovereign CDS spreads for the US going back to 2008, and reflect the market's reactions to events (including the 2011 and 2023 debt-limit standoffs) over time:

As you can see, the debt-limit and tax law standoffs created spikes in 2011 and 2012, and, to a lesser extent, in early 2023, and that these spikes preceded the ratings changes, and were not caused by them, and that the market very quickly recovered from them. In fact, the Fitch ratings downgrade has barely registered on the US CDS spread, in the market, indicating that investors are neither surprised nor spooked by the ratings downgrades (so far). 

Dealing with Government Default Risk

     No matter what you think about the Fitch downgrade of US government debt, the big-picture perspective is that we are closer to the scenario where no entity is viewed as default-free than we were fifteen years ago, and it may be only a matter of time before we have to retire the notion that government bonds are default-free entirely. The questions for investors and analysts, if this occurs, becomes practical ones, including how best to estimate risk-free rates in currencies, when governments have default risk, and what the consequences are for equity risk premiums and default spreads.

1. Clean up government bond rate

    Consider the two requirements that have to be met for a local-currency government bond rate to be used as a risk-free rate in that currency. The first is that the government bond has to be widely traded, making the interest rate on the bond a rate set by demand and supply in the market, rather than government edict. The second is that the government be perceived as default-free. The Swiss 10-year government bond rate, in July 2023, of 1.02% meets both criteria, making it the risk-free rate in Swiss Francs. Using a similar rationale, the German 10-year bund rate (in Euros) of 2.47% becomes the risk-free rate in Euros. With the British pound, if you stay with the Moody's ratings, things get trickier. The government bond rate of 4.42% is no longer risk-free, because it has default risk embedded in it. To clean up that default risk, we estimated a default spread of 0.64%, based upon UK's rating of Aa3, and netted this spread out from the government bond rate:

Risk-free Rate in British Pounds     

= Government Bond Rate in Pounds - Default Spread for UK = 4.42% - 0.64% = 3.78%

Extending this approach to all currencies, where there is a government bond rate present, we get the riskfree rates in about 30 currencies:

Since the US still preserves a bond rating of Aaa (for the moment), with Moody's, the US treasury rate of 3.77% on July 1, 2023, was used as the riskfree rate in US dollars. 

    As you look at these rates, especially in some emerging market currencies, you should be cautious about the numbers you get, especially since the liquidity is light or non-existent in government bonds in these markets. Thus, it is possible that the Vietnamese Dong has the lowest risk-free rate in the world in mid-2023, among all currencies, or it may reflect distortions in the Vietnamese government bond.   One way to check these riskier rates for reasonableness is to extend on the insight that the key driver of the risk free rate is inflation, and that in a world where capital moves to equalize real returns, the differences in risk-free rates across currencies come from differential inflation In my post on country risk, In fact, as I argued in my post on country risk, you can convert a riskfree rate in any currency into a risk-free rate in another currency by adjusting for the differential inflation between the currencies: 

Thus, using the IMF's forecasted inflation rates for the US (3%) and Vietnam (5.08%), in conjunction with the US dollar risk-free rate of 3.77% on July 1, 2023, yields a Vietnamese Dong risk-free rate of 5.87% (or 5.85% with the approximation).

    If you believe that S&P and Fitch are right on their default risk assessments for the US, and that it should get a rating lower than Aaa (say Aa1), from Moody's, the path to getting a US risk-free rate has an added step. You have to net out the default spread for the US treasury bond rate to get to a risk-free rate:

Riskfree Rate in US dollars = US Treasury Bond Rate - Default spread on US T.Bond

Using the sovereign CDS market's estimate of 0.30% in August 2023, for instance, when the US treasury bond rate hit 4.10%, would have yielded a risk-free rate of 3.80% for the US dollar.

2. Risk Premia

    If you focus just on risk-free rates, you may find it counter intuitive that an increase in default risk for a country lowers the risk free rate in its currency, but looking at the big picture should explain why it is necessary. An increase in sovereign default risk is usually triggered by events that also increase risk premia in markets, pushing up government bond rates, equity risk premiums and default spreads. In fact, if you go back to my post on country risk, it becomes the key driver of the additional risk premiums that you demand in countries:

You will notice that in my July 2023 update, I used the implied equity risk premium for the US of 5.00% as my estimate of a premium for a mature market, and assumed that any country with a Aaa  rating (from Moody's) would have the same premium. 

    Since Moody's remains the lone holdout on downgrading the US, I would use the same approach today, but assuming that Moody's downgrades the US from Aaa to Aa1, the approach will have to be modified. The implied equity risk premium for the US will still be my starting point, but countries with Aaa ratings will then be assigned equity risk premiums lower than the US, and that lower equity risk premium will become the mature market premium, to be used to get equity risk premiums for the rest of the world. Using the sovereign CDS spread of 0.30% as the basis, just for illustration, the mature market premium would drop from 5.00%, in my July 2023 update, to 4.58% (5.00% -1.42*.30%).

When safe havens become scarce...

    During crises, investors seeks out safety, but that pre-supposes that there is a safe place to put your money, where you know what you will make with certainty. The Fitch downgrade of the US, by itself, is not a market-shaking event, but in conjunction with a minus 18% return on the ten-year US treasury bond in 2022, these events undercut the notion that there is a safe haven for investors. When there is no safe haven, market corrections when they happen will not follow predictable patterns. Historically, when stock prices have plunged, investors have sought out US treasuries, pushing down yields and prices. But what if government securities are viewed as risky? Is it any surprise that the loss of trust in governments that has undercut the perception that they are default-free has also given rise to a host of other investment options, each claiming to be the next safe haven. While my skepticism about crypto currencies and NFTs is well documented, a portion of their rise over the last 15 years has been driven by the erosion of trust in institutions. 

Conclusion

    I started this post by noting that we pay little attention to risk-free rates in theory and in practice, taking it as a given that it is easy to estimate. As you can see from this post, that casual acceptance of what comprises a risk-free investment can be a recipe for disaster. In closing, here are a few general propositions about risk-free rates that are worth keeping in mind:

1. Risk-free rates go with currencies, not countries or governments: You estimate a risk-free rate in Euros or dollars, not one for the Euro-zone or the United States. Thus, if you choose to analyze a Brazilian company in US dollars, the risk-free rate you should use is the US dollar risk free rate, not the rate on Brazilian US-dollar denominated bond. It follows, therefore, that the notion of a global risk-free rate, touted by some, is fantasy, and using the lowest government bond rate, ignoring currencies, as an estimate of this rate, is nonsensical.
2. Investment returns should be currency-explicit and time-specific: Would you be okay with a 12% return on a stock, in the long term? That question is unanswerable, until you specify the currency in which you are denominating returns, and the time you are making the assessment. An investment that earns 12%, in Zambian Kwacha, may be making less than the risk-free rate in Kwachas, but one that earns that same return in Swiss Francs should be a slam-dunk as an investment. In the same vein, an investment that earns 12% in US dollars in 2023 may well pass muster as a good investment, but an investment that earned 12% in US dollars in 1980 would not (since the US treasury bond rate would have yielded more than 10% at the time).
3. Currencies are measurement mechanisms, not value-enhancer or destroyers: A good financial analysis or valuation should be currency-invariant, with whatever conclusion you draw when you do your analysis in one currency carrying over into the same analysis, done in different currencies. Thus, switching from a currency with a high risk-free rate to one with a much lower risk-free rate will lower your discount rate, but the inflation differential that causes this to happen will also lower your cash flows by a proportional amount, leaving your value unchanged.
4. No one (including central banks) cannot fight fundamentals: Central banks and governments that think that they have the power to raise or lower interest rates by edict, and the investors who invest on that basis, are being delusional. While they can nudge rates at the margin, they cannot fight fundamentals (inflation and real growth), and when they do, the fundamentals will win.

YouTube Video

Governments and Value III: Bribery, Corruption and other "Dark" Costs

In this last post on the effects of government on valuations, I want to return to the value destructive effects that corruption, bribery and other "illegal" side-payments to government officials can have on value. In many countries, business people know that to keep doing business, they have to grease palms and provide “gratuities” to the gatekeepers of officialdom. A spate of news stories in the last few weeks should alert us all to the reality that the problem is not only still prevalent but that companies everywhere are exposed to its costs.

1. In a reminder to natural resource companies that the countries where these resources are most abundantly found are often also the ones with the most corrupt government officials, Cobalt International Energy, an energy company backed by Goldman Sachs, saw $900 million of its value wiped out, after revelations that three powerful Angolan officials held concealed interests in the company.
2. India is the second-largest telecom market in the world, with hundreds of millions of subscribers. The regulatory uncertainty that has always bedeviled companies competing in the sector was augmented to by a tainted telecom auction in 2008, which resulted in the resignation and arrests of a cabinet minister. The saga played out in the Indian Supreme Court's recent ruling taking away licenses awarded to eight companies in that auction; the fact that six of these eight companies were foreign suggested a nativist spin to corruption. Put in blunter terms, the ruling seemed to suggest that bribery of Indians by other Indians was par for the course, but bribery by foreign nationals was an abomination.
3. Finally, from the other great growth story in Asia, China, came the story of Bo Xilai, a prominent member of the party elite, and his family: his wife, who is accused of murdering a British businessman, and a son, Bo Guagua, who goes to the Harvard's Kennedy School of Government, drives a Ferrari and has the lifestyle of a top notch capitalist. While the story is filled with the kinds of details that tabloid newspapers love, the real story that the Chinese government wants to keep a lid on is that Bo is not alone among government officials, in accumulating wealth out of proportion to his "income" as a government official.

I am not an expert on corruption but here is what I see as the ingredients that allow it to flourish. First, for official gatekeepers to have power, you need gates: the more licenses, permissions or other official approvals you need to operate, the greater the potential for corruption. Second, it is a lot less risky being corrupt if you have political hegemony (whether it be of the dictatorial variety or one party rule), an ineffective legal system (making it impossible to challenge biased official acts) and an apathetic or controlled media (that either cannot or will not view corruption as a good news story). Third, the odds of corruption increase if the system is designed on the premise that corruption is the rule rather than the exception. Thus, setting the salaries of public employees at well below what the market would pay them, given their qualifications, on the assumption that they will augment these salaries with "side payments", will ensure that you will attract the "most corrupt" people into government and a continuation of the system.

Rather than debate the basis of corruption and whether culture and history play a role, I want to first focus on “objective” measures of corruption . Transparency International, an organization that tracks corruption globally, releases an annual listing of corruption across the world. Just to provide a summary, the following is a list of the ten least corrupt and the ten most corrupt countries in the world, based on their ranking.

Least corrupt countries in the world	Most corrupt countries in the world
1. New Zealand	1. Somalia
2. Denmark	2. North Korea
3. Finland	3. Myanmar
4. Sweden	4. Afghanistan
5. Singapore	5. Uzbekistan
6. Norway	6. Turkmenistan
7. Netherlands	7. Sudan
8. Australia	8. Iraq
9. Switzerland	9. Haiti
10. Canada	10. Venezuela

Just for information, the United States came in as the 24th least corrupt country out of 182 countries, China was 75th and India was 95th on the list. While I am sure that there are countries where you and I may disagree with the rankings, there are clearly regions of the world where operating a business without "paying off" government officials is close to impossible.

If you are valuing a company that operates in these dens of iniquity, how do you incorporate the costs of corruption into your value? Here are a three alternatives:

1. Treat bribes as operating expenses: From a valuation perspective, it would be easiest to deal with bribes if they were out in the open and  treated as a separate line item in the expenses. So, in your operating expense breakdown, you could have a line item titled "Bribes and payments to corrupt officials" with the expense associated with it. Perhaps, we can then assess firms on the efficiency of their bribery and treat it as a competitive advantage for companies that are exceptionally good at getting results for their money. Unfortunately, even in countries where corruption is endemic, it remains "under the surface" and unreported.
2. Treat corruption as an implicit (and unreported) tax: In the more likely scenario, where corruption exists but is not explicitly reported, it may make sense to consider the expenses associated with it as an implicit tax levied by the government. The fact that this tax revenue goes to the government officials and not to the taxpayers is deplorable, but that makes little difference to the company paying it. While this idea may seem farfetched, PWC did exactly this in an "opacity index" that they computed for dozens of countries and converted into tax rates. In 2001, for instance, they estimated the added cost of operating in China was the equivalent of facing an effective tax rate of 46%. Unfortunately, this listing is almost a decade old and while the opacity index itself has been updated by others, the effective tax rates no longer seem to be computed by country.
3. Increase the cost of capital to cover "government" partners: When corruption occurs at the highest levels, you can argue that as a private business owner, you have "corrupt government officials" as partners who provide no capital but get a share of the income. Consequently, you have to generate a higher return on your capital invested to cover the cash outflows to your implicit partners. You can find interesting attempts to quantify this effect here and here.

There are two complicating factors. The first is that the United States (among other countries) has laws on the books that forbid companies from paying bribes not only to US officials but to officials in other countries . As a consequence, the costs of bribery may be far greater than the actual expenditures incurred and include the penalties that these companies will be face, if the bribery is exposed. The second is that the "right connections in high places" in countries with extensive corruption is a significant competitive advantage in itself. Odious though we may find this proposition, the firms that understand how the system works (or who to pay and how much to pay to make it work) will generate excess returns and higher value than their more virtuous counterparts. 

Governments and Value II: Subsidies and Value

In my last post, I looked at the negative effects on equity value of the threat of government expropriation (nationalization). In this one, I want to focus on the more benign (and perhaps positive) impact that governments can have the values of some companies, through subsidies in one of many forms: providing or facilitating below-market rate financing, special tax benefits, revenues or price supports and even forcing competitors to provide direct benefits to a subsidized entity. Note that my intent in this post is not to examine the wisdom of these subsidies and whether governments should be tilting the playing field. While I do have strong views on the topic (and you can guess what they are from the subtext), I want to focus on the mechanics of how best to value businesses that benefit from these subsidies. This post was, in part, triggered by the recent news story on First Solar, where the company announced its intent to both scale back its operations and return a $30 million subsidy it had received from the German government.

Subsidy Variants
Governments, through the ages, have played favorites with businesses, either providing help to their preferred companies or, in some cases, handicapping their competition. Broadly speaking, there are at least four ways in which governments can try to benefit a subset of companies:

1. "Low or no cost" financing:  The cost of borrowing (debt) for a company should reflect its default risk. In some cases, governments can step in the fray and either provide or facilitate "cheap" or "below market rate" financing, ranging from grants (effectively free financing) to low-interest rate loans (Airbus) to acting as a loan guarantor with banks (Tesla). The net effect is the same: the company is able to borrow more money at lower interest rates than it otherwise would have been able to, which, in turns, lowers its overall cost of financing its operations. You can argue that bailouts are a variant on this subsidy, insofar as it offers a financial lifeline to distressed (usually too-big-to-fail) firms that otherwise would have faced default.
2. Tax holidays, credits and deductions:  The tax code has long been a favored device for the government to bestow benefits on chosen sectors or companies. In some cases, this can take the form of a lower tax rate on income (than the tax rate paid by other businesses) or a tax holiday, and in others it can take the form of more generous expensing and depreciation rules. Fossil fuel companies in the US, for instance,  have been allowed to expense a portion of exploration costs, granted tax credits amounting to 15% of investment costs related to enhanced oil recovery and gas pipelines can be depreciated over 15 years instead of 20 years. These benefits translate into higher after-tax cash flows (from paying less in taxes)  or timing benefits on tax savings (with expensing and depreciation breaks).
A side note: One oft-used proxy of which businesses get subsidized the most is the difference between the effective tax rate paid by these businesses and the marginal tax rate. I report the average effective tax rates on my website, by sector. However, I think that the dominant factor driving effective tax rates now is not tax subsidization but foreign sales. The more revenues a company (or sector) generates from overseas (where corporate tax rates are lower), the lower the effective tax rate will be.
3. Revenue or price support (Higher and more predictable revenues): In some cases, governments step in to both stabilize and increase revenues of businesses by providing price support to companies. For instance, the US government, among others, has provided price supports for some agricultural products, such as sugar. In other cases, governments benefit firms by handicapping foreign competition and imposing tariffs on imported goods. Sometimes, government can indirectly support revenues by providing the subsidies to the customers of preferred companies; an example would be credits offered to homeowners for using solar panels on their houses.
4. Indirect subsidies: Rather than provide benefits directly to a company, the government can also force competitors to sustain the company by either paying a cash subsidy to the company or by buying its products at an arranged price. The Zero Emissions Vehicle Program, a California state mandate requiring that auto manufacturers failing to produce a certain number of zero emission vehicles buy credits from those who did, resulted in Tesla receiving millions of dollars in payments from other auto companies.

Ways of dealing with subsidies
There are two ways of dealing with subsidies. One is to build them into your discounted cash flow valuation inputs and let them flow into your estimated value. The other is to ignore subsidies in a DCF valuation and to value subsidies separately and add them on.

1. Build into valuation 
Each of the subsidies, described above, can be incorporated into a DCF valuation input:
a. "Low cost" financing: Enter the subsidized cost of debt and/or the subsidized debt ratio into the cost of capital, which will yield a lower cost of capital and higher value. Thus, if a firm like Tesla that normally would not have been able to borrow money, since it is a risky, money losing company. and would have been all equity financed (say with a cost of equity of 11%) may be able to borrow a portion of its capital at a "low" interest rate (because of implicit or explicit government subsidization) and end up with a cost of capital of 10.8%.
b. Tax holidays, credits and deductions: Subsidies that take the form of a tax holiday or special tax rate will lower the effective tax rate and increase after-tax cash flows. To the extent that the tax subsidized operations can be kept separate from non subsidized business, the company may be able to still get the full tax benefits of borrowing. More generous expensing and depreciation rules don't increase the nominal tax benefits across time but the value of the tax benefits will increase because they occur earlier in time.
c. Revenue or price supports: These subsidies can show up in two places. First, the price support increases revenue to producers who can sell at the support price, which is higher than the market price. Second, to the extent that these subsidies make revenues more stable, they may reduce the operating risk in the business and increase value.
d. Indirect subsidies: The transfer payments from competitors will boost revenues and cash flows and increase the value of the subsidy-receiving company.
The advantage of this approach is that the subsidies then get baked into the valuation, with no need for post-valuation garnishing or augmentation. The disadvantage of this approach is that it is easy to forget that subsidies don't last forever and that the firm will eventually lose them, either because governments cannot afford them anymore or because the company loses its preferred status.
If you do decide to go this route, keep in mind at least two issues. If you build subsidies into your DCF valuation, think through how long these subsidies will last. For instance, the "low cost" financing subsidy may cease to be one, if your company becomes a larger, more profitable entity. In addition, check to see what the value of the company would be, with no subsidies. In other words, break the company's value down into its operating value and its subsidy value.

A valuation of Tesla
To illustrate the process, let me try to value Tesla Motors, the electric car company founded by Elon Musk, one of the co-founders of Paypal. Tesla Motors got a subsidy from the US government, in the form of a Department of Energy loan facility that it utilized to borrow about $250 million in 2011, at an interest rate of 3%. (You can download my excel spreadsheets, with the valuations, if you want):
Step 1: I valued Tesla Motors, with the subsidized financing. The company's borrowing gives it a debt ratio of about 10%, which with its subsidized interest rate, results in a cost of capital of about 10.8%. The valuation, where I do assume that Tesla's revenues will climb to about $ 5 billion in 10 years and that the pre-tax operating margin will converge on 12% (much higher than the average margin of 7% across automobile companies in 2011), yields a value per share of $10.40/share.
Step 2: I valued Tesla Motors without the subsidized financing, by assuming that the firm would have to raise the debt at a market interest rate of 9% (instead of the 3% subsidized rate). The resulting value per share is $9.60.
Step 3: The interest rate subsidy can be valued at $0.80/share, the difference between the valuation with the subsidy and the valuation without.
This is the narrowest measure of the subsidy. If we expand the subsidization to include tax credits for future investments (reducing reinvestment needs for the future) and perhaps less risk (if the government supports revenues or requires competitors to pay Tesla), the value per share would increase (and so would the subsidy value). In this final valuation, I expand the Tesla valuation to include broader subsidies and generate a value per share of $18.17/share.

2. Separate valuation
In this approach, the discounted cash flow valuation is done with inputs that the firm would have had in a non-subsidized world, and the value of the subsidy is assessed separately. Thus, in the case of Tesla, you would value the company using the 12% cost of equity (or capital) that the firm would have had in a non-subsidized world, and then value the effect of the low cost financing separately. Thus, if Tesla is able to borrow money at a lower rate, as a result of the government support or backing, the savings each year from the subsidy amount to the difference between the market and the subsidized interest rates. Taking the present value of these savings over time should generate a value for the subsidy, which can then be added on to the value obtained using the non-subsidized cost of capital.
While this approach requires more detailed information on the nature of the subsidy and what the firm would have looked like in its absence, it has two benefits:
a. The analyst can value the subsidy for only the period that he or she thinks it will be offered and discount it at an appropriate rate. Thus, if Tesla has $250 million in debt at a 3% subsidized rate, when it should have been paying 9%, it is saving $15 million a year because of the subsidy (9% of 250 - 3% of 250). Assuming that the subsidy is likely to continue for only 10 years and that the only risk of not getting it is if Tesla defaults, the present value of $15 million a year for 10 years, discounted back at the unsubsidized cost  of debt of 9%, yields a value today of $96.26 million.
b. If the subsidy from the government requires the company to offer something in return (build a manufacturing plant with higher cost labor), separating the effects of the subsidy from the valuation allows you to assess the costs and benefits of taking the subsidy. If the net benefit is negative, the company may be better off rejecting or returning the subsidy to the government.

Implications for investing/valuation
A company that gets significant subsidies from the government will have a higher value, in most cases, than one that does not. In some sectors, say green energy, the subsidies can account for a significant portion of the overall value of the firm (and its equity). As an investor, I have always been uncomfortable investing in these companies at prices that require the continuation of subsidies to justify the investments. Governments, especially in these times of budget constraints and sovereign defaults, are both fickle in their choice of favorites and unreliable subsidizers. Thus, if I can buy Tesla at a price that is less than its unsubsidized value, I will do so, and view the subsidies as icing on my investment cake. If, on the other hand, making money on Tesla requires me to count on the government's continuing indulgence, you can count me out. In this case, I am spared the choice, since Tesla at the prevailing stock price of $ 30 looks overvalued, even relative to the most generous subsidized value.

Governments and Value: Part 1 - Nationalization Risk

I have been writing about valuation for a long time and for much of that time period, I chose to ignore the effects, positive or negative, that governments can have on the value of businesses. Implicitly, I was assuming that governments could affect the value of a business only through the tax code and perhaps through regulatory rule changes (if you were a regulated firm), but that  a firm's value ultimately rested on its capacity to find a market for its products and generate profits from these products. The last five years have been a wake-up call to me that governments can and often do affect value in significant ways and that these effects are not restricted to emerging markets.

The news story that brought this thought back to the forefront was from Argentina, where Cristina Fernandez, the president, announced that the Argentine government planned to nationalize YPF. The ripple effects were felt across the ocean in Spain, where Repsol, the majority owner of YPF, now stands to lose several billion dollars as a consequence. Not surpringly, the stock price of YPF, already down about 50% this year, plunged another 21% in New York trading. If you own YPF stock, my sympathies to you, but it is too late to reverse that mistake. However, there are general lessons that we can take away from this sorry episode about how best to incorporate the possibility of government capriciousness into what you pay for shares in a company.

1. Intrinsic value and nationalization risk
There are three components to intrinsic value: cash flows (reflecting the profitability of your business), growth (incorporating both the benefits of growth and the costs of delivering that growth) and risk. If you have to value a company in a country where nationalization risk is a clear and present danger, the obvious input that you may think of changing is the risk measure. After all, as investors, you face more risk to your investments in countries with capricious heads of state or governments, than in countries with governments that respect ownership rights (and have legal systems that back it up).
There are three options that you can use to incorporate the effect of this risk on your value:


Option 1- Use a "higher required return or discount rate": If you are using a discounted cash flow valuation, you could try to use a higher discount rate for companies that operate in Argentina, Venezuela or Russia, for instance, to reflect the higher risk that your ownership stake may be taken away from you for less-than-fair compensation. The problem that you will face is that discount rates are blunt instruments and that the risk and return models  are more attuned to capturing the risk that your earnings or cash flow estimates will be volatile than to reflecting discrete risk, i.e., risks like survival risk or nationalization risk that "truncate or end" your investment.


Option 2: Reduce your "expected cash flows for risk of nationalization: You can reduce the expected cash flows that you will get from a company incorporated in a "nationalization-prone" market to reflect the risk that those cash flows will be expropriated. While this may be straight forward for the near term cash flows (say the first year or two), they will be much more difficult to do for the cash flows beyond that time period.

Option 3: Deal with the nationalization risk separately from your valuation: Since it is so difficult to adjust discount rates and cash flows for nationalization risk (or any other discrete risk), here is my preferred option.
Step 1: Value the company using conventional discounted cash flow models, with no increment in the discount rate or haircutting of the cash flows. The value that you get from the model will be your "going concern" value.
Step 2: Bring in the concerns you have about nationalization into two numbers: a probability that the firm will be nationalized and the proceeds that you will get if you are nationalized.
Value of operating assets = Value of assets from DCF (1 - Probability of nationalization) + Value of assets if nationalized (Probability of nationalization)

To illustrate, consider Dominguez & Cia, a Venezuelan packaging company, which generated 117 million Venezuelan Bolivar (VEB) in operating income on revenues of 491 million VEB in 2010. A discounted cash flow valuation of the company generates a value of 483 million VEB for the operating assets. Assuming a 20% probability of nationalization and also assuming that the owners will be paid half of fair value, if nationalization occurs, here is what we obtain as the nationalization adjusted value:
Nationalization adjusted value = 483 (.8) + (483*.5) (.2) = 435 million VEB
Subtracting out the debt (291 million) and adding cash (68 million) yields a value for the equity of 212 million VEB. At its traded equity value of 211 million VEB, the stock looks fairly priced. If you download the valuation, you can see that I have incorporated the high operating risk (separate from nationalization risk) in Venezuela with a higher equity risk premium (12%) and the higher inflation/interest rates in Venezuela with a higher risk free rate of 20%. In particular, play with the nationalization probabilities and the consequences of nationalization to see how it plays out in your value per share.

Note, though, that my 20% estimate of the probability of nationalization is a complete guess, in this case. If I were interested in investing in Venezuelan (Russian, Argentine) companies,  I would spend more of my time assessing Hugo Chavez's (Vlad Putin's, Cristina Fernandez's) proclivities and persuasions than on generating cash flow estimates for companies. Since my skill set does not lie in psychoanalysis, I am going to steer away from companies in these countries.

2. Relative value and nationalization risk
How would you bring in the concerns about nationalization, if you value companies based upon multiples? One is to use multiples extracted from the country in question, on the assumption that the market would have incorporated (correctly) the risk and cost of nationalization into these multiples. To an extent, this is reasonable and it is true that companies in countries with high nationalization risk trade at lower multiples.

Note that while Russian and Venezuelan companies trade at a discount to their emerging market peers (and my guess is that Argentine companies will join them soon), you have no way of knowing whether the discount is a fair one.

The problem, though,  becomes more acute when you are not able to find enough companies in the sector within that country to make your valuation judgment. With Dominguez & Cia, for instance, you have the only publicly traded packaging company operating in Venezuela. If you decide to go out of the market, say look at US packaging companies in 2011, the average EV/Operating income multiple is about 10.51 in January 2012. Applying this multiple to Dominguez's operating income would generate a value of  1230 million VEB, well above the market value of 211 million VEB. However, you have not incorporated the higher operating risk in Venezuela (separate from the nationalization risk) and the risk of nationalization.

The bottom line with multiples is simple. If you do not control for nationalization risk, companies in countries which are exposed to this risk will often look absurdly cheap on a PE ratio or an EV/EBITDA basis. But looking cheap does not necessarily equate to being cheap..

Implications
While it is too late to incorporate the risk of nationalization in the value of YPF, you can adjust the estimated values of other Argentine companies. While the government of Argentina may argue that YPF was unique and that they would not extend the nationalization model to other companies, I would operate under the presumption of "fool me once, shame on you... fool me twice, shame on me" and incorporate a higher probability of bankruptcy into the valuation of every Argentine company. The net effect would be a drop in equity values across the board: that is the consequence of government action. There are other repercussions as well. A government that is cavalier about private ownership is likely to be just as cavalier about its financial obligations: no surprise then at the news that the default spreads for Argentina have surged after the YPF news.

In closing
While this post is about the "negative" effects of government intervention, it is possible that the potential for government intervention can push up the value of equity in other companies. In particular, the possibility that governments may "bail out" companies that are "too large or important to fail" may increase the value of equities in those companies as will the potential for government subsidies to "worthy" companies. I will come back to these questions in subsequent posts.

Returning again to the Argentina story, Ms. Fernandez was quoted as saying, "I am a head of state, and not a hoodlum". Someone should remind her that the two are not mutual exclusive, and the problem may be that she is both.