In credit risk management, it is common to distinguish between point-in-time (PIT) and through-the-cycle (TTC) estimates of default probability or expected loss. But amid a unique pandemic, TTC loss forecasts may have been too bullish, and there is now talk about whether this credit risk estimation tool needs to be fine-tuned to reflect financial institutions’ new reality.
Interpreting the meaning of PIT and TTC is usually straightforward. Unfortunately, however, nothing about the COVID-19 episode or its impact on risk management has earned this descriptor. The question we address here is whether TTC ratings and estimates, normally forever unchanging, should now be adjusted as a direct result of the pandemic and its insidious economic fallout.
We’ll get to that answer in a minute – but let’s first break down the differences between PIT and TTC.
PIT calculations attempt to capture current macroeconomic dynamics and thus produce a live forecast of future credit performance. TTC estimates, meanwhile, seek to abstract from the business cycle, and instead measure the innate creditworthiness of the target entity. TTC ratings will not change at different points in the cycle, while PIT estimates invariably will.
In general, PIT models are more difficult to construct, but will provide more timely and informative insights for managers and other stakeholders – if they’re done well. TTC estimates are far more stable in nature; various regulatory decrees, moreover, demand that TTC estimates be used as the basis for a range of metrics used widely across the financial sector.
One key example is the determination of regulatory capital. For many exposures, a small change in the computation of TTC estimates will have a material impact on the amount of capital needed to be held by the institution. This, in turn, could potentially affect the company’s motivation for engaging in particular businesses.
A Distinct Event
When considering what, if any, TTC adjustments need to be made, it all comes down to whether you view the COVID-19 crisis as a cyclical downturn. If you do, then TTC methodologies and estimates in use during 2019 will still apply now. If you do not, your TTC loss rate numbers need to be raised slightly to accommodate the new reality exposed by the 2020/21 pandemic.
Normally, the business cycle is driven by waxing and waning consumer and business sentiment. In the lead-up to the GFC, for example, households and financiers engaged in a credit-fueled lending boom, centered on the mortgage industry, that caused house prices to soar in a number of countries.
When the bubble burst, credit dried up and consumers sought to repair their finances. Large financial failures significantly damaged confidence, and the economy slumped into a deep downturn. Consumer and business exuberance then slowly returned over subsequent years.
You could apply a similar narrative, with varying details, timing and severity, to most other recessions in living memory. Indeed, the very practice of central banking is focused on controlling the economy’s boomtime sugar-high in order to forestall or to minimize the severity of the subsequent recession.
The COVID-19 event is distinct from this. While the economy was humming along nicely during 2019, you wouldn’t exactly say it was overheating. The downturn was then triggered by a mindless virus, rather than a bunch of people enjoying a party too much for their own collective good.
The other telltale sign of a cyclical recession is that the exact start date is generally hard to pin down, because economic malaise does not follow a clearly defined timetable. In contrast, in the case of the pandemic, we can state with some degree of certainty that recession became inevitable when the first local COVID-19 case was recorded – and was in full swing by the time the first lockdown was announced. The pandemic recession we’re now experiencing feels more like an idiosyncratic shock than anything that could be compared to a business cycle in any meaningful sense.
There are also differences in the nature of the expected recovery. A few months ago, there was a lot of speculation about whether a U-, V- or W-shaped recession would be forthcoming, but consensus eventually settled on a K-shaped, two-speed recovery with strong growth limited to particular sectors. We saw wild fluctuations last summer, as some activities resumed in earnest while other industries lagged.
Looking ahead, the first few post-vaccine months will be critical – but there is every chance that confidence will quickly recover, leading to rapid, near-term economic growth. This, again, will be quite distinct from recent cyclical recoveries that have skewed strongly in the direction of sluggishness.
Pandemics in Review: Lessons Learned, and the Future of TTC
Looking back through history, it’s fair to say that the threat of a pandemic has been omnipresent.
Of course, prior to the present crisis, few thought seriously about the economic upheaval caused by the 1918 flu outbreak – and those that did most likely conflated it with the nascent recovery from World War I. The SARS outbreak in the early 2000s, meanwhile, caused some disruption in East Asia, but other regions were spared from the worst economic fallout. Fast forwarding to the present, it’s probably down to dumb luck that a globally disruptive disease outbreak was avoided for slightly more than a century.
In terms of credit, though the risk was ever-present, there were no COVID-like pandemic periods covered by the 2019-era data. As such, TTC loss estimates in use back then were overly optimistic (assuming, for simplicity, that all entities experienced higher credit risk in 2020). Given this, the most appropriate thing to do today is to downgrade the TTC numbers slightly to reflect what we now know about the economic impact of pandemics.
The required shift, it must be said, will be difficult to calibrate. Bear in mind that expected credit loss numbers will vary greatly across different sectors, depending on how susceptible they are to infectious disease. But including the full impact of COVID-19 in future TTC estimations seems heavy handed, especially if the data series being used to calculate the metric is relatively short.
In other words, the 2020/21 data should be discounted, but the exact weight the pandemic should be afforded in an updated TTC calculation remains unclear.