Many months ago, I volunteered to write a blog comparing financial risk management practices in banking and insurance. Why did I make such a promise, why did I feel that I am in a position to comment on such a topic. The simple answer is that I work for an insurance company that also has a medium sized bank as a subsidiary and I happen to sit on the Asset Liability Committees for both the insurance and bank subsidiaries. As I am sure many of you recognise, after making my promise, my initial burst of enthusiasm waned and the task drifted towards the bottom of my priority list. What has happened to rocket this task up the list to a point where I feel a sudden urge to write on this topic, the answer, of course, is Silicon Valley Bank and the ensuing drama across the US banking sector.
The initial idea of this blog was to compare risk management practices, but I have decided to use the SVB crisis and take a “what if scenario” approach to illustrate how an Actuary may have helped SVB. The what if scenario is “what if SVB had an Actuary on its risk management team”. Bear in mind that SVB did not have a CRO for a large portion of 2022, never mind an Actuary.
As with most crises a number of risks coalesced to create the extreme event. In no particular order liquidity risk, concentration risk, policy holder behaviour risk and interest rate, with maybe regulatory risk thrown in for good measure, came together to bring SVB to the point of collapse. All of these risks are familiar to Actuaries working in the insurance sector and form an important role in the risk management.
I will take liquidity risk first. I am sure SVB had projections of outflows and compared these to the availability of highly liquid assets to check if liquidity tests could be passed. This particular risk is fundamental to banking and I am not sure if an Actuary would have added more value, indeed in my experience insurance companies can often learn from the more precise modelling of cash flows that banks perform. In the case of SVB a large proportion of their assets were held in highly liquid US Treasuries so I suspect our theoretical Actuary in the risk management function would have reviewed the results and ticked the OK box. It is interesting to note here that SVB was not considered large enough by Regulators to be subject to the full rigour of liquidity testing, I suspect this is an oversight that will be rectified once the dust has settled.
Next I will consider concentration risk and policyholder behaviour risk together as in this extreme event they are very intertwined. It is widely reported that SVB’s customer base was very concentrated to companies in the tech sector. The deposit base grew rapidly over the Covid Lockdown period as funding for tech startups flooded in, then in 2022 this funding dried up as rates increased dramatically and investors found alternative places to invest their money. With less new money coming in many of these tech companies started to withdraw large amounts of cash to fund their operations, which eventually led to a run on the bank. I do believe our theoretical Actuary would have been concerned about the concentration risk, but I am also sure that a regular risk manager in the bank would have raised this alarm as well. I am also convinced that the link between higher rates and a change to policyholder behaviour, what Actuaries might call dynamic policy holder behaviour, would have been recognised by the bank’s risk manager. But where I think our Actuary would have added value is in the modelling of policy holder behaviour and the assumptions used in the projection of outflows used to test liquidity. Statistical analysis is at the foundation of Actuarial Mathematics and analysis of data is at the heart of what Actuaries do. I would also add that in my experience banks concentrate more on shorter time periods and apply a narrower window of uncertainty when analysing risk compared to insurance companies. Adding longer time horizons and applying more extreme stresses may have helped the risk management function to spot the potential problem at an earlier stage.
The final risk on my list is when our theoretical Actuary comes into her own, the interest rate risk or the mismatch risk. It is well recorded that what pushed SVB to fail was when the outflow of deposits reached the point that SVB had to start selling its longer dated bonds and crystallise the unrealised losses caused by the sudden jump in rates. Matching assets with liabilities has been a cornerstone of Actuarial techniques for as long as anyone can remember, the capital regime for insurance companies in Europe will penalise mismatches and incentivise appropriate matching strategies. Based on very approximate calculations it would seem that the duration of assets on SVB’s balance sheet was about 6 years longer than its liabilities. I am sure that our Actuary would have raised the red flag. I do not want to give the impression that banks are not aware of this risk, far from it in my experience. But my experience is with a bank regulated in Europe where the interest rate risk associated with mismatches is a very high priority and managed very carefully.
I am sure there will be many investigations and reviews performed to understand what went wrong with SVB and what changes should be made to reduce the chance of such an event happening again. But at this early stage it does appear that modelling of cashflows and understanding interest rate risk were key factors, and these are skills that Actuaries have. It is easy to be wise after the event but maybe such an event provides the Actuarial profession with an opportunity to sell its skills to banking.
John McCrossan is a member of the SAI and of the ERM committee.
The views of this article do not necessarily reflect the views of the Society of Actuaries in Ireland, the Enterprise Risk Management Committee, or the author’s employer.