#AI ChatGPT - OpenAI: Assigning so much validity to perceived credit risk has that not introduced a way too dangerous systemic risk in our financial systems?

I asked ChatGPT: 

"The world always lives uncertain times, when pandemics, war, inflation and so many other unexpected events can cause havoc, and when risks can so easily have been misperceived.

Nonetheless current bank regulators based most of their bank capital requirements on the certainty of perceived credit risk.

And that they do even though all large bank exposures that ended up causing major bank crises, were built-up with assets that, from a credit risk point of view, were perceived as safe.

Can you explain to us regulators’ rationale for doing so?

And excuse me for insisting: Assigning so much validity to perceived credit risk, augmented by having some few human fallible credit rating agencies much deciding what’s risky and what’s safe; has that not introduced a way too dangerous systemic risk in our financial systems?"

ChatGPT answered:

"The rationale behind current bank regulators basing their capital requirements on perceived credit risk is rooted in the belief that it is essential to allocate sufficient capital to cover potential losses arising from credit exposures. Credit risk refers to the risk of borrowers defaulting on their obligations, leading to losses for the banks.

Historically, regulators have focused on credit risk because it has been one of the most significant sources of risk for banks. By assessing the creditworthiness of borrowers and assigning risk ratings to various assets, regulators aim to ensure that banks have enough capital to absorb potential losses. This approach is based on the assumption that higher-rated or perceived safer assets are less likely to default, while lower-rated or riskier assets carry a higher probability of default.

The reliance on credit ratings agencies can be attributed to their expertise in evaluating credit risk. These agencies assess the creditworthiness of issuers and provide ratings that reflect the agencies' assessment of the likelihood of default. Regulators have traditionally used these ratings as a benchmark for determining the riskiness of assets.

However, it is important to acknowledge that the reliance on perceived credit risk and credit rating agencies has its limitations. The global financial crisis of 2008 revealed some of these shortcomings when highly rated mortgage-backed securities experienced significant defaults. This event highlighted the potential for misperception of risk and the failure of credit rating agencies to accurately assess the riskiness of certain assets.

The introduction of a systemic risk in financial systems due to the reliance on perceived credit risk is a valid concern. It is challenging to accurately predict and measure systemic risks, especially during times of uncertainty and rapid market changes. The interconnectedness of financial institutions and the reliance on similar risk assessment models can amplify the impact of misjudged risk perceptions.

In recent years, regulators have recognized the need to address these shortcomings and enhance the stability of the financial system. Efforts have been made to strengthen capital requirements, improve risk management practices, and reduce the overreliance on credit ratings. The Basel III framework, for example, incorporates a broader range of risk factors and encourages banks to use their internal models to assess risk.

While the financial system continues to evolve, regulators are working towards developing more comprehensive approaches to risk management. This includes considering a wider range of risks beyond credit risk, such as liquidity risk, operational risk, and systemic risk, to build a more resilient financial system that can better withstand unexpected events and crises."

I asked ChatGPT: 

"Since banks might have a vested interest in reducing its capital requirements, for the purpose of calculating the risk weighted bank capital, does it make any sense to allow banks to use their internal models to assess risk?"

ChatGPT answered:

"The use of internal models by banks to assess risk is a topic of ongoing debate and consideration among regulators and experts in the financial industry. There are arguments both in favor of and against allowing banks to use their internal models for risk assessment.

One of the key arguments in favor of internal models is that they can potentially capture the specific risk profile and characteristics of individual banks more accurately than a standardized approach. Banks have access to extensive data and sophisticated modeling techniques, which they argue allows them to better estimate the risks associated with their portfolios. Proponents argue that this can lead to a more accurate assessment of risk and better alignment of capital requirements with the actual risk exposure of each bank.

However, there are also valid concerns associated with the use of internal models. Critics argue that banks might have an incentive to manipulate their models or underestimate risk to reduce their capital requirements. This behavior was observed in the past, where some banks used overly optimistic assumptions and models that failed to capture the true risks, contributing to the financial crisis.

To address these concerns, regulators have implemented rigorous standards and requirements for the use of internal models. They impose strict validation processes, regular audits, and requirements for banks to hold sufficient data to support their models. Regulators also conduct their own assessments and exercises to ensure the models are robust and the risks are adequately captured.

Additionally, regulators often require a combination of approaches, where banks use both internal models and standardized approaches to assess risk. This helps strike a balance between the advantages of internal models and the need for a standardized and consistent framework across the industry. It allows regulators to compare and validate the results obtained from internal models and provides a safeguard against potential manipulation.

Overall, the use of internal models is a complex issue with pros and cons. While they can potentially provide more accurate risk assessments, it is crucial for regulators to establish stringent oversight and ensure that banks are using these models responsibly. Striking the right balance between risk sensitivity and robust regulatory oversight remains a continuous challenge for regulators."

I asked ChatGPT: 

"When it comes to determine how much more capital a bank would have to hold against an asset, on the margin, there were it most counts when taking a decision whether to hold the asset or not, is it not the risk weighted bank capital requirements that defines that?"

ChatGPT answered:

"Yes, the risk-weighted bank capital requirements play a crucial role in determining how much capital a bank needs to hold against its assets. The risk-weighted capital framework assigns different risk weights to various types of assets based on their perceived riskiness. These risk weights reflect the probability of default and potential loss associated with each asset class.

The risk-weighted capital requirements are typically expressed as a percentage of the risk-weighted assets (RWA). To calculate the RWA, banks assign risk weights to their assets based on the prescribed risk weightings set by regulatory authorities. The risk weights are generally higher for assets considered riskier and lower for assets considered safer.

By applying risk weights to their assets, banks determine the amount of capital they need to set aside as a buffer to absorb potential losses. The higher the risk weight assigned to an asset, the more capital the bank is required to hold against it.

This risk-based approach aims to ensure that banks hold sufficient capital to cover the potential losses associated with their assets. It provides an incentive for banks to assess and differentiate the risk levels of their assets accurately. Assets perceived as safer and having lower risk weights require less capital, while riskier assets necessitate higher capital allocations.

However, it's important to note that the risk-weighted capital requirements have been subject to criticism, particularly regarding their effectiveness and potential for manipulation. Critics argue that the risk weights assigned to assets may not always reflect their true riskiness, leading to potential mispricing of risk. The reliance on credit rating agencies for determining risk weights has also been seen as problematic, given the issues highlighted during the financial crisis.

To address these concerns, regulatory bodies have been working on improving the risk-weighted capital framework. Revisions such as the Basel III framework have introduced more sophisticated risk measurement techniques, including the use of internal models, stress testing, and a focus on capturing systemic risks. These efforts aim to enhance the accuracy of risk-weighted capital requirements and improve the resilience of banks in the face of potential financial shocks.

Overall, while risk-weighted capital requirements are an important tool for determining capital adequacy, ongoing refinements and scrutiny are necessary to ensure that they effectively reflect the underlying risks and contribute to the stability of the financial system."

Note: The answers are copied exactly from those given to me by OpenAI

Here, a comment I made at the World Bank in 2003:  “Nowadays, when information is just too voluminous and fast to handle, market or authorities have decided to delegate the evaluation of it into the hands of much fewer players such as the credit rating agencies. This will, almost by definition, introduce systemic risks in the market.” 

Here my many other questions to ChatGPT – Grok on bank regulations

Bank regulators should be forced to see “Hell on Wheels” and read John Kenneth Galbraith’s “Money: Whence It Came, Where It Went”

In the TV series Hell on Wheels, its main character, Cullen Bohannon, when asked to testify before the US Senate about all the obvious corruption of Thomas ‘Doc’ Durant, someone absolutely not Bohannon’s friend, someone absolutely not one having been sanctimonious or behaved according to any social norms, repeats, over and over again, to the great chagrin of his interrogators: “The Transcontinental railroad could not have been built without Thomas Durant”

And John Kenneth Galbraith wrote in his “Money: Whence it came where it went” 1975 the following: “For the new parts of the country [USA’s West]… there was the right to create banks at will and therewith the notes and deposits that resulted from their loans…[if] the bank failed…someone was left holding the worthless notes… but some borrowers from this bank were now in business...[jobs created]

It was an arrangement which reputable bankers and merchants in the East viewed with extreme distaste… Men of economic wisdom, then as later expressing the views of the reputable business community, spoke of the anarchy of unstable banking… The men of wisdom missed the point. The anarchy served the frontier far better than a more orderly system that kept a tight hand on credit would have done…. what is called sound economics is very often what mirrors the needs of the respectfully affluent.”

And Galbraith also opined in his book that: “The function of credit in a simple society is, in fact, remarkably egalitarian. It allows the man with energy and no money to participate in the economy more or less on a par with the man who has capital of his own. And the more casual the conditions under which credit is granted and hence the more impecunious those accommodated, the more egalitarian credit is… Bad banks, unlike good, loaned to the poor risk, which is another name for the poor man.”

Therefore I cannot but conclude in that bank regulators should be forced to see “Hell on Wheels” and read John Kenneth Galbraith’s “Money: Whence It Came, Where It Went”. That in order to, hopefully, be able realize that with their risk weighted capital requirements for banks, these will not finance the risky future, but only refinance the safer past and present and, as a result, the economy will stall and fall. 

To add insult to the injury, bank regulators are doing all this in the belief that bank crises result from excessive exposures to what is perceived as risky, which is utter nonsense. Bank crises have always, and will always, result from uncertainties; that which includes unexpected events, like devaluations earthquakes and regulators not knowing what they are doing, criminal behavior and excessive exposures to something ex ante perceived as safe but that ex post turned out to be very risky.

“If you see something, say something”. Someone should run to the Homeland Security of the Home of the Brave and denounce that, most probably, unwittingly; some serious terrorism is taking place by means of dangerously risk adverse faulty bank regulations.

Do bank regulators, now with “output floor” based on their standardized risk weights, keep on making fun of us?

A 75-percent output floor signifies that no matter which outcome the bank’s internal calculation yields, the risk weight that determines the capital required, can’t be more than 25 percent lower than the standardized risk weighting method designed by the regulators.

So let’s see what that really means. 

For the standard method’s 0% risk weighted sovereign, unless some bank’s internal calculation comes up with a negative risk, it will still mean 0%.

For the standard method’s 20% risk weighted private asset, it means the weight cannot be less than 15%.

For the standard method’s 35% risk weighted residential mortgage, it means that weight cannot be less than 26.25%.

For the standard method’s 100% risk weighted private asset without a credit rating, like loans to SMEs, it means that the risk weight cannot be less than 75%.

Really? Are bank’s internal models worse than your standardized weights? Do you really think the Medici’s would have assigned a risk weight of 0% to the Sovereign?

There’s absolutely nothing wrong with allowing banks to use their own risk models in order to try to minimize their cost of capital, but that regulators concoct a set of ex ante perceived standardized risk weights in order to determine how much capital banks should have in order to be able to confront, not perceived risks, but uncertainty, is just as crazy as it gets.

Here some questions bank regulators refuse to answer, something that should make us all nervous. They really might not have a clue about what they are doing.

Are some consulting companies, e.g. McKinsey & Co. Too-Big-To-Think?

A book, “No ordinary Disruption”, which I paid for, arrived with my mail today. The authors are Richard Dobbs, James Manyika, Jonathan Woetzel all belong to McKinsey Global Institute, the economics and business research arm of the management-consulting firm McKinsey & Co

From its introduction “An intuition reset” I quote:

“Dramatic changes come from nowhere, and then from everywhere… The fortunes of industries, companies, products, technologies, and eve countries and cities rise and fall overnight and in completely unpredictable ways.”

That is true but it makes me ask: Where was McKinsey & Co when bank regulators decided that their capital [equity) requirements for banks, those that are expected to cover for unexpected losses, were to be based on the predictable expected losses derived from the ex ante perceived credit risks?

Why on earth should banks need capital against perceived credit risks, when what is perceived cannot really be what is that dangerous? 

Or is it that McKinsey does not understand the difference between risk and uncertainty?

And the McKinsey authors identifying their “Four great disruptive forces” list: (1) the locus of economic dynamism shifting to emerging markets like China; (2) the impact of technology; (3) demographics; (4) “The final disruptive force is the degree to which the world is much more connected through trade and through movements in capital, people, and information. 

But they leave out that monstrous source of disruptive force that can emanate at any moment from sheer regulatory stupidities with global reach. Why? 

And I ask this because I am convinced that McKinsey & Co., somewhere deep in its bowels, must have known that: allowing banks to hold so little equity against some assets, only because these were perceived as safe, had to end in tears; and that allowing for different capital requirements for different assets, based on perceived credit risk already cleared for, had to dangerously distort the allocation of bank credit to the real economy.

The authors present us with the management imperative for the coming decade, namely: “To realize that much of what we thought we knew about the how the world works is wrong.”... 

Wrong! That’s no excuse, McKinsey & Co. involved in so many areas should have known that when regulating banks you must do two things: First define what’s the purpose of banks, something which was not done; and second analyze what caused bank crises in the past… and it sure was not what was perceived as risky but always what was ex ante perceived as safe but that ex-pots turned out risky.

So if there is a management imperative for the next decade that should be: To realize why so much we think about how the world should work could turn out to be so fundamentally wrong; and how to avoid to become a silly mutual admiration club prone to groupthink.

When a consulting group is no longer able to freely question what’s going on, to freely be able to call the bluff of what’s dumb, then it will have grown too big. It will be weighed down by too many conflicts of interests of all nature; which hinders it from speaking or even thinking the truth… and finally, very sadly, it will end up as a highly paid endorser of stupidities.

When a consulting group with global reach reaches a point of too much importance, then it also becomes a dangerous source of systemic risks.

So do we now need capital requirements for banks based on the size of the consultant group they use? J