With interest rates near record lows, bond investors and lenders might be guilty of "chasing" spreads, hunting for returns by taking more risks. How should credit spreads be determined?
When interest rates rise, bond values decline. Bond investors are frightened off by a spike in interest rates, because fixed-income portfolios plunge in value. When interest rates are low, after an initial rise in portfolio values, returns languish and frustrate investors who want something better than a percent or two on corporate bonds. Some investors flee debt markets and consider accepting the risks in equity markets and "alternatives" (private investment, real estate, etc.).
Or they chase credit spreads--showing a willingness to take more risk in loan and debt markets in arush find a better return.
Interest rates and yields on bonds and debt incorporate a "credit spread," a percentage spread (quoted in basis points) above a basic risk-free (U.S. Treasury, e.g.) rate that is intended to reward debt investors and corporate lenders.
In an era of low interest rates for over a decade now, market analysts have looked for evidence that debt investors and lenders are "chasing" those credit spreads, aggressively taking risks to accept higher credit spreads (even if it is not the appropriate risk-reward spread). Investors seek to achieve any return better than a U.S. Government-bond return or a bank CD.
Some markets analysts the past decade occasionally conclude lenders and investors have been too willing to take risks to get a decent reward (yield or credit spread). And in doing so, they have pushed down the spreads for some risk, non-investment-grade issuers. (Demand for risky investments pushes up the prices of those bond values, reducing the credit spread on the bonds' yields to maturity.)
Credit spreads in bond markets fluctuate in the same way equity prices vaccillate. Markets try to interpret all factors that could weigh in on a borrower's ability to meet debt obligations (principal and interest payments).
For a single issuer or for debt rated similarly, spreads can rise and fall swiftly. They certainly did that in March-2020. With the onset of the pandemic, bond spreads (investment-grade and high-yield) rose sharply. And then during the year, spreads declined. Some high-yield spreads climbed above 800 basis points and months later fell to 300 basis points.
Investors comfortable that selected risky names will not defaul will "chase" spreads or buy issues of non-investment-grade borrowers, but arguably will have overpaid for value.
What is the right correct spread?
Debt and bond analysts resort to fundamentals and principles of credit analysis to determine. But they resort to past cases and data, too: historical default rates, recovery rates on defaulted debt, etc.
Much of the determination of an appropriate credit spread (reward for debt risk) starts with the basics of assessing creditworthiness of the borrower and doing so on an ongoing basis.
Debt investors and banks also provide funding for special-purpose entities known for securitizations, or better known as asset-backed securities that include the alphabet soup of debt offerings (CMOs, ABSs, CLOs, MBSs, CBOs, CDOs, CMBSs, etc.).
The analysis of these structures examines cash flows from financial instruments (mortgages, credit-card receivables, auto loans, etc.) to pay down several levels (or tranches) of debt structures. Similarly, credit spreads are applicable to the tiers of debt based on expected loss and based the logic of payout in a waterfall liquidation scenario.
In fact, some investors have found these structures more appealing because for the same rating (say, AA) the credit spreads might be higher than the spreads on a corporate bond. Spreads on some investment-grade issues rose far above 300 basis points in early 2020.
Q&A
1. When we derive the credit rating of the borrower, how can you do so from the financial statements and financial disclosures?
How do we also relate ongoing business risks and projections to an analysis of historical information?
One of the primary goals of debt analysis is to derive a final rating for the borrower/company. Ratings agency use an alphabet AAA-to-D scale. Banks use a numerical scale. For example, the borrower may be rated on a 1-10 scale as an overall assessment of its financial condition, its expected performance in the future, and its ability to remain solvent by meeting all cash obligations on debt, expenses and liabilitires. Some banks and financial institutions use different scales or apply a qualitative rating ("excellent, very good, etc.").
We can map our overall ratings and criteria to that which is used by the ratings agencies, partly to compare our own assessments with those provided by the ratings agencies and partly to be able to use past default frequencies (by ratings groups) to determine a probability of default for the company we have just rated.
The ratings criteria and categories are based on many financial metrics, as explained (profitability, earnings, cash flow, liquidity, leverage, debt capacity, projected cash flow, etc.). But the business and industry risks will have helped us determine a rating for each category of above. Business and industry risk analysis help us finetune a rating of the company's historical earnings or liquidity. (Business and industry risk analysis will certainly prepare us to project the earnings, cash flows and balance sheet of a company.)
One of the primary goals of debt analysis is to derive a final rating for the borrower/company. Ratings agency use an alphabet AAA-to-D scale. Banks use a numerical scale. For example, the borrower may be rated on a 1-10 scale as an overall assessment of its financial condition, its expected performance in the future, and its ability to remain solvent by meeting all cash obligations on debt, expenses and liabilitires. Some banks and financial institutions use different scales or apply a qualitative rating ("excellent, very good, etc.").
We can map our overall ratings and criteria to that which is used by the ratings agencies, partly to compare our own assessments with those provided by the ratings agencies and partly to be able to use past default frequencies (by ratings groups) to determine a probability of default for the company we have just rated.
The ratings criteria and categories are based on many financial metrics, as explained (profitability, earnings, cash flow, liquidity, leverage, debt capacity, projected cash flow, etc.). But the business and industry risks will have helped us determine a rating for each category of above. Business and industry risk analysis help us finetune a rating of the company's historical earnings or liquidity. (Business and industry risk analysis will certainly prepare us to project the earnings, cash flows and balance sheet of a company.)
The business and industry risks help explain why profitability has declined or company liquidity is decreasing or the debt burden is too high, but will also help us assess how the company will perform going forward in varying scenarios and market or industry conditions.
In some ways, the rating of each category and the overall rating are an assessment of historical performance, but should also reflect our perception of how we see the company performing in the future. The business and industry analysis influences our perceptions of the future.
Some rating agencies (and not all analysts) will rate the industry after having performed an industry and business risk analysis. They will then use that rating as a benchmark to rate a company for each financial category.
Some analysts will apply a "qualitative overlay" to their financial assessments; hence, they rate the company's past performance and current financial condition reviewing metrics and ratios, but allow those assessments to be influenced by an overall assessment of the business/industry and other factors. Those can be company's response to environment factors, social and political influences, etc.
2. A bank or debt investor will price the rewards or returns of the debt based on credit risks. Credit risks can be quantified by the concept of "expected loss." And expected loss has traditionally been tied to the probability the borrower will default on obligations.
In some ways, the rating of each category and the overall rating are an assessment of historical performance, but should also reflect our perception of how we see the company performing in the future. The business and industry analysis influences our perceptions of the future.
Some rating agencies (and not all analysts) will rate the industry after having performed an industry and business risk analysis. They will then use that rating as a benchmark to rate a company for each financial category.
Some analysts will apply a "qualitative overlay" to their financial assessments; hence, they rate the company's past performance and current financial condition reviewing metrics and ratios, but allow those assessments to be influenced by an overall assessment of the business/industry and other factors. Those can be company's response to environment factors, social and political influences, etc.
2. A bank or debt investor will price the rewards or returns of the debt based on credit risks. Credit risks can be quantified by the concept of "expected loss." And expected loss has traditionally been tied to the probability the borrower will default on obligations.
What other risks should be incorporated in determining credit spreads, credit risks and expected loss--e.g., currency risk, longer maturities. What about the administrative and operating expenses related to reviewing, analyzing, negotiating, structuring and booking debt exposures?
If a bank or debt investor uses risk-based pricing (and many do; not all do), then the loan should be priced with a credit spread, and the credit spread = (probability of default) x (loss-given-default). Loss-given-default = (1 - expected recovery rate on credit exposure). (More technically, an analyst may seek to determine the "net present value" of the expected recovery rate, since recoveries from default debt may come as a series of cash flows over time.)
The probability of default is a function of the underlying rating of the borrower (and is often based of historical default frequencies by ratings group.)
But it is also a function of the tenor of the credit exposure. The probability of default for an A-rated name will be lower than that of a B-rated name. But also, for the A-rated name, the probability of default for a one-year loan will be lower than that of a 10-year long. Therefore, the spread should incorporate the maturity of the exposure. Default probabilities are expected to increase because time factors introduce uncertainties and risk about performance in the future.
The base credit spread doesn't necessarily incorporate other types of risks: liquidity risks, currency risks, etc. Therefore, some financial institutions will adjust the computed credit spread to account for those additional risks. If for example, the bank determines it wouldn't be able to sell off a long-term loan or hedge it in the way it prefers, it might adjust the spread accordingly. (A liquidity premium is essentially added to the credit spread.) Others might suggest that liquidity risk is reflected in a higher probability of default for longer-term exposures.
Financial institutions and investors will also try to price the loan or debt (or adjust the spread) to account for operating expenses associated with transaction. (Often, however, banks struggle to determine what are the direct costs associated with booking and monitoring one corporate loan. But they try.)
Banks, also, have regulatory capital requirements for the loans they book. And they desire to meet a minimum return on that capital. They will, therefore, try to ensure the spread on the loan incorporates the credit risks, but at least permits them to reach a target return on capital.
Many financial institutions, therefore, incorporate a risk-based credit spread discussed above, but also account for (a) the operating expenses related to processing and monitoring that loan and (b) the earnings necessary to meet a bank-wide return on allocated capital (about 13%, for example, at many large banks).
If a bank or debt investor uses risk-based pricing (and many do; not all do), then the loan should be priced with a credit spread, and the credit spread = (probability of default) x (loss-given-default). Loss-given-default = (1 - expected recovery rate on credit exposure). (More technically, an analyst may seek to determine the "net present value" of the expected recovery rate, since recoveries from default debt may come as a series of cash flows over time.)
The probability of default is a function of the underlying rating of the borrower (and is often based of historical default frequencies by ratings group.)
But it is also a function of the tenor of the credit exposure. The probability of default for an A-rated name will be lower than that of a B-rated name. But also, for the A-rated name, the probability of default for a one-year loan will be lower than that of a 10-year long. Therefore, the spread should incorporate the maturity of the exposure. Default probabilities are expected to increase because time factors introduce uncertainties and risk about performance in the future.
The base credit spread doesn't necessarily incorporate other types of risks: liquidity risks, currency risks, etc. Therefore, some financial institutions will adjust the computed credit spread to account for those additional risks. If for example, the bank determines it wouldn't be able to sell off a long-term loan or hedge it in the way it prefers, it might adjust the spread accordingly. (A liquidity premium is essentially added to the credit spread.) Others might suggest that liquidity risk is reflected in a higher probability of default for longer-term exposures.
Financial institutions and investors will also try to price the loan or debt (or adjust the spread) to account for operating expenses associated with transaction. (Often, however, banks struggle to determine what are the direct costs associated with booking and monitoring one corporate loan. But they try.)
Banks, also, have regulatory capital requirements for the loans they book. And they desire to meet a minimum return on that capital. They will, therefore, try to ensure the spread on the loan incorporates the credit risks, but at least permits them to reach a target return on capital.
Many financial institutions, therefore, incorporate a risk-based credit spread discussed above, but also account for (a) the operating expenses related to processing and monitoring that loan and (b) the earnings necessary to meet a bank-wide return on allocated capital (about 13%, for example, at many large banks).
3. How can an analyst, lender or investor approach the derivation of probability of default for a particular borrower (obligor) more precisely?
The probability of default, of course, follows from the analysis of the company, which includes the analysis of performance, profitability, operating cash flows, balance sheet, liquidity, leverage, and capital structure.
The analysis, too, includes a proper assessment of business risks, industry risks, management, and other operating-environment factors (including, say, country risks). The analysis of the company should result in an overall risk-rating of the company. The ratings agencies, we know, derive a rating and report it as AAA, AA, A, BBB, etc. Banks, as mentioned above, also derive a rating and commonly report it numerically, say, from 1 to 10.
The ratings help us determine an appropriate default probability. We often look at past defaults (default frequencies) by rating as a benchmark to assign a default probability for the company. We can, too, adjust that probability of default based on other forward-looking factors.
Ratings agencies publish data that show for each ratings group the frequency of default, based on tenor. We would then expect the default probability to be higher as ratings worsen. We also expect default probability to be higher as we extend the tenor. (The probability of default is higher for one-year exposure vs. 10-year exposure.)
The probability of default, of course, follows from the analysis of the company, which includes the analysis of performance, profitability, operating cash flows, balance sheet, liquidity, leverage, and capital structure.
The analysis, too, includes a proper assessment of business risks, industry risks, management, and other operating-environment factors (including, say, country risks). The analysis of the company should result in an overall risk-rating of the company. The ratings agencies, we know, derive a rating and report it as AAA, AA, A, BBB, etc. Banks, as mentioned above, also derive a rating and commonly report it numerically, say, from 1 to 10.
The ratings help us determine an appropriate default probability. We often look at past defaults (default frequencies) by rating as a benchmark to assign a default probability for the company. We can, too, adjust that probability of default based on other forward-looking factors.
Ratings agencies publish data that show for each ratings group the frequency of default, based on tenor. We would then expect the default probability to be higher as ratings worsen. We also expect default probability to be higher as we extend the tenor. (The probability of default is higher for one-year exposure vs. 10-year exposure.)
For example, a rating agency such as Fitch or Moody's will compile data for a defined time period (say, from 1970-2020) that show for each rating a cumulative historical default rate or frequency by tenor. The data, for example, may show that for A-rated corporate debt rated by Moody's over that period for five-year debt defaulted at a frequency or rate of 1.5% and for 10-year debt a rate of 2.5%. This will likely apply on to corporate issuers rated by Moody's, and it may include only U.S.-based or global corporates.
We could, therefore, consider using 1.5% as the basis for default probability for five-year, A-rated debt (and 2.5% as the basis for default probability for 10-year A-rated debt.)
It is a mapping process, in so many ways. But the process starts with the analysis of the financial condition, proceeds to an assignment of a rating, and then to mapping to default frequencies in the past for that rating. Yet ultimately the analyst must determine the right default probability, given today's environment and expectations of the future (industry, macro-economy, sovereign risks, interest rates, etc.).
The loss-given-default (or sometimes called, "loss severity") is a separate assessment, although there are some analysts who contend the Loss-Given-Default could be a function of the probability of default. They argue that as default probability increases, then the amount of recovery from a liquidation of assets will decrease. The same assets in recovery, they contend, are the same assets that delivered insufficient operating cash flows and led to a default situation.
It is a mapping process, in so many ways. But the process starts with the analysis of the financial condition, proceeds to an assignment of a rating, and then to mapping to default frequencies in the past for that rating. Yet ultimately the analyst must determine the right default probability, given today's environment and expectations of the future (industry, macro-economy, sovereign risks, interest rates, etc.).
The loss-given-default (or sometimes called, "loss severity") is a separate assessment, although there are some analysts who contend the Loss-Given-Default could be a function of the probability of default. They argue that as default probability increases, then the amount of recovery from a liquidation of assets will decrease. The same assets in recovery, they contend, are the same assets that delivered insufficient operating cash flows and led to a default situation.
"LGD" ties into the recovery of loan amounts after events of default. LGD, therefore, is affected by (a) collateral, collateral type, collateral haircuts, (b) seniority of the loan: senior, junior, subordinated, hybrid, etc., (c) support: guarantees (explicit, implied).
LGD, too, could be affected by the jurisdiction, country or legal system where the insolvency, bankruptcy or liquidation of the company occurs. So while the concept of LGD and recovery are not difficult understand, determining the appropriate LGD may require making assupmptions. Liquidation and bankrupcty history also indicate recovery rates can vary significantly.
4. . When we calculate the probability of default of a company, we can start with a credit assessment and derivation of a credit risk rating.
4. . When we calculate the probability of default of a company, we can start with a credit assessment and derivation of a credit risk rating.
Do we consider a risk rating before a company takes on new debt or a new loan, or do we consider rating the company (and determining default probability) after new debt?
Analysts should assess the company's rating and financial condition based on how the company looks with the new debt. For example, a company may have an existing rating from ratings agencies based on the current level of corporate debt.
Analysts should assess the company's rating and financial condition based on how the company looks with the new debt. For example, a company may have an existing rating from ratings agencies based on the current level of corporate debt.
Let's assume that amount for a large company is $5 billion. The rating assigned is, say, A. The company is contemplating an acquisition and considering issuing $2 billion in debt to accomplish the merger. Ratings agencies will analyze pro forma what the rating will be after the new debt. Corporate credit analysts would do the same.
The new rating will require analysts to make assumptions about the company's use of proceeds of the new debt and how the new proceeds will improve projected operating cash flows.
Large companies do the same. As they contemplate taking on new loans or new debt, they approach ratings agencies to get input on how they may alter the rating after the new debt has been included on the balance sheet. It is indeed possible that a company may appear to be A-rated before the new debt comes onto the books, but becomes a BBB+-rated company after the debt is added. And there are frequent cases where the company decides not to issue debt (or will issue in lesser amounts) because of the impact on rating.
(Rating impact, especially a probable downgrade, is crucial, because the downgrade can influence future costs of funding--old and new debt, short-term funding. The downgrade may also trigger debt financial covenants that impose requirements on the company (more collateral, default event, etc.).
5. Credit spreads related to debt are applicable to loan and debt funding costs (interest rates, debt-issue pricing, yields to maturity, etc.). How should investors, traders, lenders and analysts interpret credit spreads?
How are ratings determined for asset-backed structures (securitizations) (mortgage-backed securities, collateralized loan obligations, e.g.)?
How do analysts and rating agencies determine recovery rates (after debt or a loan has defaulted)?
A bank will determine an appropriate credit spreads when pricing a new loan and new debt. Corporate loans are typically priced based on adding the spread to existing "Libor" rates (although Libor, as a benchmark, will eventually be replaced by other benchmarks in the years to come).
Bond issues are priced based on adding a spread to an existing risk-free (sovereign bond) rate. That is often referred to as the "coupon interest rate." However, banks adjust pricing (by issuing the bond at a premium or discount to face value) to derive a "yield to maturity" for investors that reflects the true cost of funding for the corporate issuer. Thus, a bond may have a coupon rate of 3%, but to appeal to investors corporate issuers may need to issue the bond a discount to face value such that the true yield to maturity for the investor is, say, 3.5%.
Credit spreads, in general, are based on default probabilities and creditworthiness of the borrower. But corporates and banks also consider other important factors, including competition (from other lenders and structures), maturity and liquidity. If the issue cannot be traded easily, investors expect a slightly higher yield to compensation for the risk of not being able to sell they prefer. (A bond with a 3% coupon rate may require pricing at discount to, say, up to 4% to account for both credit and liquidity risks.)
In determining a credit spread, normally, the steps taken are
In determining a credit spread, normally, the steps taken are
(a) first, analyze the creditworthiness and assess financial performance of the debt issuer (borrower), based on certain criteria while considering maco-economic and industry factors and evaluating the ability of the issuer (borrower) to be able to meet regular payments of principal and interest on the debt obligation,
(b) second, derive a credit rating based on the financial assessment and creditworthiness, and
(c) map that rating to historical events of default.
Rating agencies will report frequencies of default for a particular rating over a defined timeframe and using data going back decades. The historical rate of default by rating can be used as a baseline or benchmark to determine what might be a more refined calculation of probability of default going forward.
This all means that we derive a probability default after we have analyzed the company's ability to meet debt obligations.
In some special instances, instead of mapping the derived rating to a default historical rate. They may map a determined default rate to the rating. This is the approach for determining ratings for securitizations (MBS, ABS, CDO, CBO, CMO, CLO, etc.).
The rating agencies, in these cases, must derive ratings for tranches (tiers) or debt for a special-purpose-entity securitizations (mortgage-backed bonds, asset-backed securities, etc.). In those cases, they focus on default scenarios and probabilities within the structure and map those to a rating--i.e., outline a default case or scenario for a certain structure and then tie that to a specific credit rating.
For example, the rating agency will use historical data to conclude that a AAA bond (or a AAA tranche in a securitization), based on historical default rates, should default at a specific rate. The AA tranche should default at a higher specific rate. An so on. In these structures, assets and investments may have losses, but the higher tranches (AAA, e.g.) would not absorb the first losses.
In securitizations, they use the historical default rate to determine an "expected loss" by rating. "Expected loss" = default probability x loss-given-default (or loss severity).
For the securitized structure ("SPV"), the tiers of debt (from AAA to B) are tranches that specify which classes of debt investors will get paid first in the waterfall during liquidation. The rating agency will rate the tranches based on the "expected loss" for each tier.
Recovery rates are used to determine for both corporate and securitized structures the loss after default (or "loss given default," or loss severity).
Recovery rates are based on many factors: legal system, bankruptcy code, collateral type, collateral values and value volatility, collateral cushion, guaranteed support, implied support, seniority, ranking, subordination, etc. Analysts and rating agencies often determine this from empirical data--actual cases, events, bankruptcies, liquidations, legal codes, sovereign risks.
In previous defaults and bankruptcies, for certain types of debt (secured, unsecured, senior, junior), they determine a recovery rate and report the recovery rates over previous time period. Analysts can use a median or mean recovery rate from these past cases and then adjust the rate based on new information or trends (new laws, bankruptcy precedents, etc.).
For collateral, they focus on collateral values, the amount of collateral cushion (the amount above the loan value), the potential volatility in the value of collateral, the ability to claim, seize and possess and liquidate collateral, etc. It is quite possible for analysts and banks to update and change recovery rates frequently.
Tracy E. Williams
Tracy E. Williams
See also:
Posts
No comments:
Post a Comment