14-07-2009 om 00:00 geschreven door pieter

Samenvatting:

Nuttige stukken:

There are some features that are unique to this crisis, such as the transfer of assets from the balance sheets of banks to the markets, the creation of complex and opaque assets, the failure of ratings agencies to properly assess the risk of such assets, and the application of fair value accounting P3

Two serious problems arise with the practice of structured finance. The first regards the incentive of the originator to screen debtors when the loans are destined to be placed off balance sheet. Reputational considerations would suggest that the originator would not want to compromise its standards. However, the fact that regulators and accounting standards required little disclosure about unconsolidated off-balance sheet entities made these entities opaque to investors and lowered the cost of reputational loss to the sponsoring institution. P9

Implications of Fair Value accounting: The first is that unrealized gains and losses impact owners’ capital. The second is that, like for any other accounting rule, fair valuation provides incentives to management to game the rules to boost earnings (or reduce losses) and management bonuses linked to earnings’ performance. Thus, during periods of rising asset prices, the incentives will be to move assets into the trading account categories, and conversely during declining asset prices. The third is that when markets become less liquid, valuation models based on internal information become more relevant than valuation models based on observables and the latter more relevant than the use of unadjusted quoted prices. In the presence of patently illiquid markets, an argument can be made that fair values should be based on reliable estimates of future net cash flows. Finally, the pro-cyclical bias of fair value accounting tends to magnify current financial trends and, consequently, exacerbates a financial crisis (Wallison 2008). Lower accounting asset values that impact on regulatory capital may trigger management to sell in illiquid markets. P13

Ander bruikbaar materiaal:

Errors in judgment were as glaring as assigning the same letter grade to a CDO and a corporate bond with sharply different default rates.6

Calomiris (2007, p. 19) quotes from the Bloomberg Market of July, 2007 that CDOs rated Baa by Moody suffered five-year default rates of 24 percent, whereas corporate bonds with the same rating had default rates of 2.2 percent.

14-07-2009 om 00:00 geschreven door pieter

Samenvatting:

This paper analyses credit rating agencies as one example of a ‘private authority’ involved in the governance of financial markets.

Nuttige Stukken:

These risk assessments are widely used for making investment decisions in the market place.private credit ratings have been used to make the regulation of financial market risk sensitive – for example – by restricting the investment activities of banks to instruments of low credit risk P5

rating agencies almost never have to justify their decisions, let alone provide compensation to others for the adverse consequences of their mistakes. The breach between the magnitude of potential damages for borrowers and the possibilities of a remedy gives rise to an ‘accountability gap’. P5

Rating agencies have not been exploiting the informational asymmetry between themselves and their users. The  reason is that information intermediaries are seen to be particularly vulnerable to a loss of reputation as credible sources of information (Mann 1999). Yet, monitoring by reputation can be less effective in times of financial crisis, when rating agencies start to copy each other in order to avoid being the only one off the mark (Kuhner n.d.). P7

Second, it is doubtful whether the power of the rating agencies can be explained solely by the fact that there is no alternative to their credit risk assessments. In fact, alternative sources of information on credit do exist, such as credit registers and export credit ratings (Estrella et al. 2000: 55-125). Beyond that, a wide range of sources exist that produce credit-related knowledge; for example, investment funds, banks, professional associations and even academia. P8

the accountability gap is a problem because it prevents the correction of errors, not because of a hidden ideological agenda P12

As a rule, the rating of a borrower will not be higher than the rating of the country of origin. Second, ratings depend on some properties of the borrower himself. For any type  of firm, credit-worthiness depends on the ratio of firm value to outstanding debt (Gordon 1991). The higher the value of a firm’s assets, the higher the income flows and the lower all debt obligations, the higher the credit rating is going to be….assess the stability over time and the likelihood of a declining performance under stress… P13

Rating agencies and other institutions are not distinguished for publishing their analyses of credit-worthiness ... The claim is thus that ratings allow a comparison of the credit-worthiness P14

What is more, rating decisions always have a qualitative side that is impossible to formalize P15

Since the regulatory use of ratings seeks to curb excessive risk-taking, it targets the investors or borrowers….Thus, in order to avoid being excluded from the credit market, lenders have a strong incentive to observe the rating agencies standard of credit- worthiness. Thus, in order to avoid being excluded from the credit market,

lenders have a strong incentive to observe the rating agencies standard of credit-worthiness.P19

An accountability gap arises whenever there is a breach between the power of a rule-maker and the possibility to attribute blame. P22

Certainly rating agencies decline any responsibility for their decisions: they maintain that their information on credit risks cannot pre-empt financial decision-making but only guide it (McGuire 1991: 83).6 There are a number of good reasons for this stance. First, investment decisions always have to be related to the entire portfolio. A security with a low credit rating is not necessarily a bad investment, since it promises a higher yield. Second, investment decisions inevitably comprise other risks beyond mere credit risk measured by the credit rating. Additional risks such as exchange-rate risks with foreign investments or liquidity risks are not included in a rating. P22

Limited Liability

Limited Competition

Limited Supervision

In times of crisis, changes of credit ratings seem to have a pro-cyclical effect on the dynamics of financial markets and thus aggravate rather than stabilise a crisis (Adams et al. 1999: 186). P27

18-07-2009 om 23:20 geschreven door pieter

Samenvatting:

De laatste 20 jaar heeft men in de financiële sector teveel gebruik gemaakt van risicomodellen die de realiteit niet kunnen vatten. Daarnaast zijn er ook nog enorm complexe producten ontstaan die soms dubbel verpakt worden. Er is iets fundamenteels mis met de waarderingsmodellen.

Nuttige Stukken:

The widespread use computer-based risk models in the financial industry in the last two decades enabled the marketing of more complex financial products to consumers, the growth of securitization and derivatives, and the development of sophisticated risk management strategies by financial institutions p1

These risk models failed spectacularly in the global financial crisis that started in the subprime mortgage market,2 and this outsourcing of regulation exacerbated the crisis. P2

This Article refers to the above software and computer-based risk models as the “new financial code.” P4

risk models make various assumptions in attempting to quantify these probabilities and often use historic loss data to model future losses. But, attempts to quantify these probabilities are inherently problematic, because future losses may not follow historic patterns P11

Credit risk: For a financial institution, credit risk means the risk that a borrower will default on payment of obligations to that institution.

 Market risk: Market risk, on the other hand, covers risks that the value of a firm’s investments or other assets will decline (or, that its liabilities will increase) due to changes in market prices.41 A firm and its investment portfolio may be subject to price fluctuations in different types of markets. Therefore, market risk includes several different subcategories of risk, including the following:

Interest rate risk, or the risk exposure from changes in interest rates;42 and

Equity risk, or risk arising from fluctuations in stock values.43

Other important categories of risk, such as operational, liquidity and systemic risk, prove harder to quantify as they are less directly reflected by historical market data. These forms of risk are defined as follows:

Operational risk. Operational risk is a broad term that conveys risk posed by a firm’s operations. The Basel II Accord defines operational risk as: “the risk of loss resulting from inadequate or failed internal processes, people and systems or from external events.”

Liquidity risk: This risk takes two forms. Trading liquidity risk is the risk that a firm cannot find a counterparty in the market willing to buy or sell the asset at fair market value.45 Funding liquidity risk means “the risk that [a] firm will not be able to meet efficiently both expected and unexpected current and future cash flow and collateral needs without affecting either daily operations or the financial condition of the firm.”

Systemic risk: Systemic risk arises from a broader market failure; this form of risk denotes potential losses that affect the entire market.

Other forms of risk: Financial institutions also must manage other forms of specialized risk, including concentration risk (potential losses to a lender due to a high percentage of its total loans concentrated in a small number of debtors),51 and reputation risk (potential losses stemming from a decline in public opinion of the institution).52 Reputation risk also captures the potential threat of bank runsP12-14

Nieuwe risicomodellen:

Simply stated, value-at-risk describes the maximum possible loss over a specified time period with a given level of confidence For example, a value-at-risk determination of a maximum of $1,000,000 of losses over a two week period with a 95% confidence interval translates into a 5% probability that losses will exceed $1,000,000 over those two weeks. But, note that value-at-risk numbers say nothing about the magnitude of losses above that confidence interval.58 In the above example, there is a 5% chance of losses exceeding $1,000,000, but the value-at-risk measurement does not specify how large those losses may be.P15

To calculate value-at-risk, modelers need to assume the basic distribution of losses, i.e. they need to determine the shape of the curve in the preceding diagram. They have three options P15

-normaalverdeing, may have no basis in reality

-historical losses, may have sample bias, markets do not always follow historical patterns

-monte carlo simulations – sophisticated random sampling driven by heavy computing power -> requires assumptions about relationship among different variables in real world market

One particular problem faced by value-at-risk models is “fat tails,” or potential large magnitude, but low probability losses. P17

Stress testing involves changing model assumptions, such as confidence level or the period of time measured, to see if value-at-risk determinations change… Modelers can also judge model performance through “back testing.” This involves modelers making several hypothetical jumps back in time, inputting into a model the historical data that was available at those respective times, and then comparing the predictions of the model with how losses actually unfolded P17

First, the models need to assess the risk of non-payment (i.e. credit risk) on the underlying assets.91 To accomplish this, these models require critical information from the originators on the underlying assets. Given the sheer number and variety of underlying assets, originators typically only provide to the sponsors of securitizations certain categories of information. P26

In many securitizations, rating agencies are paid by the SIV to issue credit ratings of the asset-backed securities P28

rating agencies rely on the originators to provide information on the underlying assets. P29

Internal Ratings-Based Approach” that allows certain banks to use internal models.P33

By giving banks the flexibility to adjust their regulatory capital according to a mix of rating agency ratings and the respective banks’ internal models, Basel II outsources significant regulatory authority to the models of rating agencies and banks. P34

To be sure, the accord sets standards for when national regulators may allow banks to use internal models and requires regulators to audit those models.125 But, these lengthy standards give bank regulators significant discretion in both deciding which banks qualify for the privilege to use internal models auditing the models of those banks. P35

Bewijs dat basel II normen (nog bezig met implementatie) significant lagere kapitaalvereisten opleveren :

After conducting a “Quantitative Impact Study” of the prospective effects of the Internal Ratings-Based Approach, U.S. bank regulators expressed concern that this new rule would dramatically lower regulatory capital. See e.g., Pamela Martin, QIS 4: What Do the Numbers Really Mean?. RMA J. (Sept. 1, 2005). See also Paul H. Kupiec, Capital Adequacy and Basel II, FDIC Center for Financial Research Working Paper No. 2004-02 (Dec. 2004) available at http://www.fdic.gov/bank/analytical/cfr/2004/wp2004/CFRWP_2004-02_Kupiec.pdf

Securities issued in CDO are often resecuritized themselves, creating what is called a “CDO Squared.”147 The iterative layering of securitizations of securitizations of securitizations became wildly popular in financial markets.148 Similarly, risks assumed in derivative transactions can themselves be hedged with other derivatives.149 P42

Just as with a basic securitization or credit default swap, modelers looking to measure the credit risk of the securities issued in a CDO squared face high costs in identifying – let alone finding information about the credit risk posed by – the numerous underlying mortgages and other cash-producing assets that ultimately back the CDO P42 squared. These modelers again make simplifying assumptions, such as relying on the credit ratings assigned to the securities immediately prior in the securitization chain. Using rating agencies as an analytic shortcut places a great deal of faith in the accuracy and integrity of those ratings. P43

Henry T.C. Hu, Misunderstood Derivatives: the Causes of Informational Failure and the Promise of Regulatory Incrementalism 1993

These flaws fall into two broad categories: flaws in the technical design of the models themselves, and flaws because of the skewed incentives of the parties who select, implement and use those models. P50

Technical Flaws

Non-robust assumptions, inadequate stress-and back testing

Flaws in Human Modelling Behaviour

-> individuals as rational actors

Spillover Effect

P54-56

Similarly, systemic risk can be reduced effectively only if institutions throughout the financial system make adequate risk management decisions. Otherwise, the system is susceptible to domino effects as one institutional failure could cause cascading failures of other firms due to counterparty and reputation risk. P65

Open source risk models -> assumpties, algorithms & structures in the models

è    promoting transparency,  auditable

è    counterparties could evaluate the adequacy

è    competitors can check on financial institution as well as on the regulators

è    many minds work on 1 thing

dangers ->

tendency towars homogeneity, more systemic risk

but this is already the case, just not possible to measure the degree of homogeneity (not disclosed models)

CRA Open source:

->license of SEC, so no problem there

-> oligopoly positions

->ratings are used in many regulations, like basel II so there is a need for them

-> bad track records

-> assumptions and algorithms should be shown

18-07-2009 om 23:21 geschreven door pieter

Samenvatting:

Nuttige Stukken:

Although a number of other credit rating agencies with a narrower focus operate in the United States, DCR, Fitch, Moody's and S&P are the only agencies that rate US corporate obligations across a broad industry spectrum. For example, Thomson Bankwatch and IBCA (in the United States) exclusively rate financial institutions and A.M. Best provides ratings on insurance companies' claims-paying abilities. P3

The Securities and Exchange Commission (SEC) currently designates only Moody's, S&P, DCR, and Fitch as "nationally recognized statistical rating organizations" (NRSROs) for rating all US corporate bond issues; 5 the other regulators rely on the SEC's designations. P4

Many researchers have documented that third, or optional, agencies on average assign higher ratings than Moody's and S&P. Our results suggest that observed differences in average ratings reflect differences in rating scales. They also call into question financial industry regulations that assume equivalence of rating scales P22

We also explored why firms choose to obtain additional ratings. We found that firms are more likely to obtain a third rating if they are large and experienced issuers in the capital market. However, there is little evidence that third agencies are employed either to resolve ex ante uncertainty or to clear regulatory hurdles. While the rating scales of third agencies may be higher than others, there is little evidence that the decisions of issuers to use them are influenced by that factor P22

20-07-2009 om 01:13 geschreven door pieter

Samenvatting:

Nuttige Stukken:

This paper studies a repeated principal-agent problem in which a regulator approves credit rating agencies. While credit rating agencies can observe an issuer’s type, the regulator cannot. Credit rating agencies offer each issuer a rating and are paid by the issuers  who demand a rating. The regulator cannot observe whether a credit rating agency assigns correct ratings. The regulator can only observe the default rate within a rating category for each credit rating agency. The default rate within a rating category does not only depend on whether a credit rating agency assigns correct ratings. The default rate can also be influenced by a common shock. P1

The model shows that there exists an approval scheme which can induce credit rating agencies to offer correct ratings. The model suggests that a regulator should both deter a credit rating agency from unilaterally offering inflated ratings, and provide an incentive to deviate from a collusive agreement to offer inflated ratings. The model indicates that a regulator should both threaten to deny approval in future periods if a credit rating agency’s performance is worse than its competitors’, and reward a credit rating agency which deviates from a collusive agreement to offer inflated ratings by reducing the number of approved credit rating agencies in future periods. P5

20-07-2009 om 15:29 geschreven door pieter

Samenvatting:

...

Nuttige Stukken:

The paper asks if credit rating agencies have incentives to misrepresent their clients’

credit quality during an ongoing systemic crisis. Two important elements are essential for

a systemic crisis: (1) Investors are not able to distinguish fundamentally healthy debtors

from fundamentally unhealthy ones. (2) Investors tend to cumulatively withdraw their

funds. Therefore, neither fundamentally healthy nor unhealthy debtors can be expected to

survive a creditor’s exit. P1

Their goal is to overcome asymmetric information between

both market sides by evaluating financial claims according to standardized quality

categories P1

The agencies’ most important job is the evaluation of fixed income securities. P2

rating agencies emphasize that their evaluation is, after

all, their subjective opinion, which is not verifiable by courts4. Also, damages due

to obvious rating errors are not part of the rating contract and not currently

enforceable by litigation5. P2

ð      enkel reputatie staat in voor kwaliteit

P3 -> allerlei argumenten tegen rating agencies

Procyclical bias: Especially in the aftermath of the Southeast Asian crisis in

1997/1998, agencies were blamed for behaving procyclically, i.e. for simply following

the majority opinion of market participants13: According to many

observers, the agencies did not give any warning signals until after the turbulence

in the Asian markets had begun. However, when the crisis was actually

spreading, the agencies reacted by cumulatively downgrading issuers invested

in the Southeast-Asian markets without taking in account individual portfolio

quality14. P3

Voor de rest was deze paper te moeilijk!

20-07-2009 om 15:30 geschreven door pieter

Rating Shopping

20-07-2009 om 15:37 geschreven door pieter

Samenvatting:

Zowel de FASB als de IASB hebben de laatste jaren standaarden ingevoerd die er op wijzen dat men stelselmatig naar full fair value voor financial instruments wil streven. Intussen gaat dit al niet meer door, dixit de IASB!

Een van de onderzoeksresultaten is: Second, they need to consider more broadly how best to minimise measurement error in fair values so as to maximise their usefulness to investors and creditors as they make their investment decisions, and how best to ensure bank managers have incentives to select those.  Om minder measurement errorste hebben past de IASB sinds kort die level-hierarchy toe, maar die stond reeds in FASB exposure draft in 2005! Dus of het effectief zal zijn is maar de vraag natuurlijk... 

Market to market implementation issues 

Marking-to-market financial instruments are relatively easy if they are actively traded in liquid markets. The problem becomes more complicated if active markets do not exist, particularly if the financial instrument is a compound instrument comprising several embedded option-like features, values for which depend on inter-related default and price risk characteristics. Moreover, Barth and Landsman (1995) makes the observation that in the absence of active, liquid markets, fair value is not well defined in the sense that an instrument’s acquisition price, selling price, and value-in-use to the entity can differ from each other.13 Stated another way, even if an instrument’s acquisition or selling prices are observable, these prices can only, at best, provide upper or lower bounds on its “fair value”. The FASB’s stated preference for using an instrument’s selling price as its measure of fair value is appropriate when fair value is well defined, but is somewhat arbitrary when it is not.

Dus in 1995 is er reeds een paper verschenen die vragen stelt bij het gebruik van fair value in illiquide markten.

Voor de rest gaat de paper vooral over waarom men fair value al dan niet een goede maatstaf is om te meten. Dit is eigenlijk irrelevant voor de thesis, aangezien we fair value als een gegeven beschouwen. De dingen die hierboven staan kunnen wel interessant zijn. 

22-07-2009 om 12:06 geschreven door pieter

Gray, S. J. (1980), "The impact of International Accounting Differences from a Security-Analysis
Perspective: Some European Evidence", Journal of Accounting Research, Vol. 18, 1, pp.
64-76. 

We kunnen deze formule misschien toepassen op het verschil tussen IFRS en onderliggend resultaat bij banken, vanaf 2006 bekeken. Wel met de bedenking dat de formule voorzien is voor een hele lijst van ondernemingen binnen een land (niet alleen de banksector). Maar het kan interessant zijn aangezien een groot deel van de verliezen van KBC (en waarschijnlijk ook Dexia) te wijten zal zijn aan IFRS voor financial instruments.

Ook zeker nog te lezen:

Jermakowicz, E. (2004), "Effects of Adoption of International Financial Reporting Standards in
Belgium: The Evidence from BEL-20 Companies", Accounting in Europe, Vol. 1,
September, pp. 51-70.
Jones, T. and R. Luther (2005),

22-07-2009 om 17:56 geschreven door pieter