Every bank faces risk. No technological advancement will ever make risk disappear completely, but the way to manage risk in 2021 is different than it was even a decade ago. In this age of digital transformation, every industry is benefitting from advances in technology designed to make lives and jobs easier, and banking should be no exception.
Because risk management is both complex and extremely crucial to the bank’s identity and ability to exist, it is understandable why so many rules and regulations are in place, some of which have always relied heavily on manual work in order to comply.
If a system isn’t broke, there’s no reason to fix it, and some would argue that the risk management system in place in most banks is not particularly broken. As long as the bank is still standing and properly dealing with its risk, the system must be working, right? Not entirely true.
The risk management systems in place in most banks are antiquated, require hours of manpower, and are still not foolproof. In this paper, we will review the three main pillars of risk management and how technological solutions can be the game changer risk managers have been waiting for.
1. Financial Risk
Banks are built on financial risk. Without taking any risk, there would be no reward and banks would be out of business. At the same time, too much risk or unmitigated risk can lead to disaster. There are three main areas in which banks deal with financial risk: loans, investments, and financial calculations.
A central role of a bank is to loan money to individuals and businesses. These loans can take many different forms, but terms of each one are calculated carefully and fastidiously based on a risk assessment. Assessing the probability of which the loan will be paid back and on time is key to determining the terms of the loan – a riskier client must be charged higher interest rates and given less favorable terms than one who is more likely to make timely payments.
One tiny miscalculation in the complex formulas used to make these determinations can have a cascading effect leading to cash flow issues and – in the most extreme case – the collapse of the entire system.
This is an especially timely issue now, as non-performing loans (NPL) are becoming more common in the wake of the Covid-19 pandemic. In some cases, governments are stepping in to guarantee loans, but this is likely to end in the near future, putting banks at great risk for a glut of unpaid loans. Bank managers need to be well aware of their NPL risk and factor it into their decisions regarding new mortgages and increasing rates.
Imagine a world in which risk managers did not have to sit for hours poring over spreadsheets and searching out the magic number that determines how risky or safe a loan is? Tools exist today that not only automatically calculate the risks and resulting terms of a loan, but that also allow for the inputting of endless variables and scenarios. Whatever scenario the risk manager can imagine, a clear level of associated risk can be presented within minutes. A record of all calculations is stored for easy retrieval for regulatory purposes.
Digital solutions exist for NPLs as well. One way for banks to deal with NPLs is to bundle them into a package and sell to other institutions. Building these packages means analyzing each loan to create the optimal mix combining some low-risk loans with the unsecured high risk loans. Rather than testing scenarios by building a new Excel spreadsheet each time, cutting and pasting in data for each loan until the right mix is found, an automated system can do the hard work in record time with even better results.
Banks must carefully monitor their own investments. Money deposited in short term checking or savings accounts can be used to lend out as long-term mortgages in return for interest or in other investments. The risk here is in ensuring that there will be cash readily available when the owner of a checking account wants to make a withdrawal.
Here too, one small miscalculation could lead to a glut of short term liabilities with a lack of funds available to cover them. If customers are unable to access their deposits, a significant drop in confidence will ensue, leading to customers taking their business elsewhere and can ultimately result in a bank run.
A digital solution can help balance the intertemporal risk inherent in banks’ investment decisions. Of course, it’s impossible to predict the future with 100% accuracy, but a machine can get much closer than a human with a much lower margin of error when it comes to predicting which way a market will move.
Loans and investments both rely on a variety of financial calculations. Before making a loan, there is a calculation to determine the probability of that loan being repaid. The decision whether to buy a particular asset is dependent on a calculation that shows the probability of that asset being liquid when the time comes to sell it. Other calculations are used on a daily basis to measure the probability of the bank needing an influx of cash or a host of other potential issues. In addition to using the calculations effectively, there is also risk derived from faults in the calculations themselves. A wrong variable here or a mistaken plus instead of minus there and the entire foundation of the bank’s risk portfolio is at stake.
Rather than putting all of the stress on a human being who is more likely to inadvertently make a mistake, digital tools can be used to build the calculations necessary for risk assessment and decision making. When the process is digital, it becomes much simpler and less time consuming to readjust the calculations to account for different variables, such as the size of a company requesting a loan or the location of a particular asset.
2. Operational Risk
As if there was not enough to worry about with financial risks, banks also must concern themselves with operational risk. This refers to exposure of the bank to any type of risk based on the way daily operations are conducted.
The classic example of operational risk is one that leads to a PR nightmare. When a bank calculates prices used for loans, the data is put into a Loan Origination System (LOS) which other vendors can access in order to provide the price. For example, when someone purchases a car, the car dealership may offer a loan and the price will be based off of the bank’s LOS. The data that is in the LOS has generally been typed in manually – which means, is open to human error.
One decimal point in the wrong place can cause a disaster for a bank. This one little decimal can result in a borrower being offered outrageous terms. Rather than questioning them, the borrower could go directly to a watchdog organization or the media. As everyone loves a good banking scandal, this is an instant PR nightmare for the bank – one that could have been saved by not misplacing that one tiny decimal point.
Taking operational risk even further, in some countries it is illegal to charge interest rates over a certain limit. Even if the number is calculated and published by mistake, the head of pricing can still be held criminally liable. This is clearly not a situation that any bank wants to be in, and many safeguards are put in place to mitigate operational risk.
These safeguards include tasking two separate teams with data input – one to do the work and the other to check for errors. A pricing committee approves the rates before they are deployed and they are re-checked at the time of delivery. This entire process generally takes 2-3 days and the manpower of 8 employees.
The hassle of protecting against operational risk can be eased with technology that automatically transfers pricing and rates directly from the bank’s server to the loan provider. Eliminating the human intermediary also eliminates the risk of human error. Systems like this have been used to reduce a 4-person team down to 1, freeing up those employees for other tasks.
Banks rely on models when it comes to making decisions regarding mortgages and other types of loans. All of the information about the prospective borrower – such as collateral, income and other factors – is entered into the model to produce a risk score. Prices of financial products and loans are also based on models, and they need to be exact in order to ensure that fair pricing is offered.
Most banks employ data scientists who build the models and then a system is put in place to oversee their work and validate the models. Regulations require a strict documentation process that keeps track of every last detail including how the model was built, which data was used to test the model, what software was used, the testing methodology, and how to replicate any testing in case of an audit.
Banks are exposed to model risk if at any point a model is not properly documented or if improper variables are used. For example, it is discriminatory and illegal to use gender in a risk model, so a risk manager would want to check and make sure that this variable was not used.
Similar to operational risk, the processes put in place in order to protect against model risk are unwieldy, time-consuming and themselves open to risk of human error.
Through the wonders of technology, the mountains of paperwork synonymous with model risk management are consolidated into one report accessible at the click of the button. Anyone – from a risk manager to an auditor – can easily see all of the details that went into designing and testing any model. If there is a problem, it is easy to unravel the data and determine exactly where the problem occurred and who is responsible.
When a model does need to be updated, the changes can be made universally across the system so that anyone using the model will automatically have the new version.
Bringing Technology into Risk Management
Now is the time for risk managers to embrace the digital transformation that is sweeping the financial industry. There are tools out there that will easily turn risk management from a time-consuming highly manual job to one that is efficient, reliable and digital.
Hundreds of excel files can quickly be replaced by a comprehensive system that can be accessed and managed by team members no matter where they are, and that stores all of the historical data for any future use.
These tools are typically easy to master and intuitive to operate. When a bank literally has millions of dollars on the line every day, why take the chance of human error when there is a better, more accurate, flexible and scalable way?