Taming cognitive biases in lending will help significantly reduce decision errors and thus NPA incidence
The government and the RBI have been exhorting banks to step up lending to revive the economy and put India on the fast track once again. But with heavy pile-up of NPAs, which reached a peak of over Rs 10.3 lakh crore in 2018, banks have been wary of lending. If it was exaggerated optimism a decade ago, it is exaggerated caution over the last 4-5 years.
Credit growth has been sluggish in single digits and some of the PSU banks even registered negative growth. Credit to industry suffered significantly and has been negative over the last three years. Imposing Prompt Corrective Action (PCA) on over 12 banks also constrained credit growth. Recovery through various channels like DRT, Sarfaesi, Lok Adalat, ARC and IBC has been tardy at around 15% of the monies sought to be recovered. This has also contributed to risk aversion.
Over the last few years, banks have invested resources in improving the appraisal, follow-up skills and processes to reduce the incidence of NPAs. This is necessary but not adequate to minimise NPA creation. Besides malfeasance and alleged frauds and crimes, cognitive biases lead to decisions errors and NPAs. But cognitive weaknesses have not been given due attention.
There is a body of well-researched literature that traces errors in decision-making to cognitive biases. These cognitive biases are far beyond hard and soft skills. These are quite common in several areas like policymaking, healthcare diagnostics, and financial services and relevant in lending decisions as well.
In simple terms, a cognitive bias is deviation from rational judgment. It is about processing information from a particular point of view, is one-sided and aligned with one’s own beliefs, values, opinions and even prejudices. Many a time, it is misinformation, a systematic error, a tendency and disposition. These may be subconscious and decision-makers are even unaware of their own biases. At times, it is deliberate. According to psychologists, 95% of cognition happens below the threshold of conscious thought.
Nobel laureate Prof Daniel Kahneman highlights the role of cognitive biases in contributing to decision errors. “Thinking fast and slow”, one of his bestsellers, dealt at length with cognitive biases that lead to decision errors. Some of the cognitive biases that lead to decision errors and suboptimal decisions in lending, which lead to the creation of NPAs are:
Heuristics approaches to decisions: This approach follows rules of thumb that are developed on past experiences and current assumptions. These decisions are shortcuts. They are snappy and quick. They may be educated guesses or pure intuition. The following are some of the heuristic approaches used in quick decision making, which may not be rational:
• Availability Heuristic is the most common heuristic approach. Snap decisions are made on the basis of immediately available information, which may be true or not. There is no time or inclination to deep dive to gather more information by research.
• Representative Heuristic decisions are based on one’s own mental prototypes or stereotype.
• Affect Heuristic is about making decisions based on how one feels at that moment, essentially mood swings, and these decisions are influenced by past experiences, be it positive or negative.
The heuristics approach may be useful in responding to some emergencies but not appropriate and rational in financial markets, including lending.
Confirmation Bias: One unconsciously seeks and favours information that conforms to their prevailing views, beliefs, values and knowledge. The information may be of low quality. Diverse or opposite opinions are not sought or listened to and at times even suppressed. One tends to disregard evidence that doesn’t conform to his/her beliefs, which leads to poor quality of decision and outcomes. This is a natural tendency to seek or emphasise with people that conform to existing conclusion or hypothesis. This blindsides several dimensions of risks. This is a widely prevalent bias and can be countered by seeking contrarian information. Einstein famously said that he sought diverse answers for the same question.
Dunning-Kruger Bias: One believes that s/he knows more than what s/he actually does. They think they are smarter than actually they are. This prevents one to acknowledge own limitations and weaknesses.
Bandwagon Bias: People draw comfort in a decision because other people do the same. Under this influence, people lose the ability to independently analyse, think and decide. This herd instinct is too common and many bubbles are attributed to this bias. We have seen either groups of lenders are active and go on a lending binge or demonstrate risk aversion in a group as is witnessed now. This ‘group think’ and ‘collective mindset’ is seen in syndications/consortiums. Lenders who took the road less travelled look better off.
Overconfidence Trap: Confidence is not a good indicator of accuracy of their judgements. One does not know the limits of their knowledge or expertise. Experience does not necessarily translate to expertise as the world is irregular and not linear. They have too much confidence in their beliefs and experience, more than warranted. They suffer from the illusion of control when actually they cannot control the outcomes of their decisions. They either do not seek feedback or ignore it.
Endowment Effect/Bias: People place a higher value on what they know and possess rather than on a skill they do not know.
Hindsight Bias: This is a tendency to see past beneficial events, even random ones, as predictable and bad events as unpredictable. These people are the ‘I told you so’ type. Worse the consequence/outcomes, greater the hindsight bias. Scrutiny of decisions much later with hindsight leads to risk aversion. As such, hindsight bias is not kind to decision-makers. Many decision-makers are blamed for not seeing the writing on the wall. But the writing is in invisible ink.
Anchoring Bias: One relies too much on the first piece of information. Subsequent negative or bad news is ignored or rationalised. This is close to Halo and Horn effect.
False Consensus Bias: One overestimates how much the other agrees with one’s own opinions and skills. This operates when a group that makes decisions has members of several layers of hierarchy. A Committee approach to decision-making is prone to suffer from this bias.
Sunk Cost Fallacy: This is a common trap. One does not like to recognise losses in the investments/lendings already made. Further decisions are made not on the future costs but on sunk costs fallacy on assumptions to avoid waste/protection of resources already committed. One often encounters this fallacy in situations like restructuring/forbearance decisions. At times, this fallacy may end up in throwing good money after bad money. This dilemma is as much emotional as economic.
Prof Kahneman packages all these biases into System 1 and System 2 thinking. System 1 is variously characterised by rule of thumb, quick, hasty, emotional, irrational and ill-informed thinking and decisions. System 2 is more deliberative, effortful, rational and careful thinking decisions. Both systems are contextually correct or wrong. In System 2, there is more objectivity and fewer errors. In the real world, it is more an interplay of both the thinking systems.
Cognitive biases are like silent killers, which impact the quality of decision making. But these biases can be tamed through a well-structured de-biasing strategy and execution. One can overcome these biases if they are identified, acknowledged and neutralised in time. One way of combating bias is to engage in conversations with people of diverse or even opposite biases. Another good practice is to document all relevant facts and then base the decision on the facts gathered. It is important to identify “what is not known” and fill the knowledge gap. Use of artificial intelligence and automation tools may also mitigate some biases. But it needs to be ensured that AI models themselves do not suffer from biases.
Credit and Risk Officers must undergo regular training on de-biasing. The entire credit cycle must be designed to identify and tame cognitive biases. Besides financial experts, lenders need psychologists to help reduce decision errors, risk and keep NPAs to the minimum thereby not threatening financial stability. The Chief Risk Officer should formulate and execute de-bias strategies.
Boards and policymakers need to be proactive. Cognitive biases are universal and around us but not obvious. Investors, lenders, insurers, policymakers are overwhelmed by cognitive biases, mostly subconsciously. But the buck stops at the lenders’ board and CEO. Coaching, counselling and mentoring at the top sets the tone.
Taming cognitive biases, at least, in lending, especially for large borrowers, would help significantly reduce decision errors and thus NPA incidence. Even if this helps less NPAs by say 20-25%, it is no small change. Don’t miss the opportunity to get less wrong. As Prof Kahneman says, it is risky to ignore ignorance.
(The author is former Chairman of a PSU bank)
Now you can get handpicked stories from Telangana Today on Telegram everyday. Click the link to subscribe.
Click to follow Telangana Today Facebook page and Twitter .