As Canada’s big banks wrapped up their past fiscal year and looked forward into a period of economic uncertainty, they forecast that things were going to be … pretty okay, actually.
I wasn’t eavesdropping in the boardrooms. Instead, I’ve surveyed the fine print of the banks’ securities filings, which detail some of the economic projections they use to forecast potential loan losses. All of the banks give some degree of insight into their estimates of Canada’s gross domestic product, unemployment rate and housing prices for the following 12 months.
And as of Oct. 31, only one – Royal Bank of Canada RY-T – thought the economy would contract, to the tune of 0.2 per cent. The rest were still putting GDP growth in their “base case” – or most likely – scenarios.
We could call those forecasts “the good news.” Or, perhaps not: Since the forecasts are used by the banks to set aside money for future loan losses, if they’re too optimistic, shareholders will pay the price if the cost of their loss set-asides has to shoot up to keep pace with a rapidly declining economy.
It’s happened before: Bank of Nova Scotia BNS-T went into the pandemic with some of 2019′s most optimistic economic assumptions – then crashed its profits by making huge provisions for potential loan losses.
Now, there are a lot of qualifications to this analysis, which I’ve developed after consulting with Veritas Investment Research analyst Nigel D’Souza, who tracks the projections. The expected-credit-loss models are highly complex and largely undisclosed. What bits of information we do have were extracted out of the banks by the Office of the Superintendent of Financial Institutions, the regulator who told the banks during the COVID-19 pandemic they needed to give the public detail about their projections quarterly, not annually.
The banks can and do use more indicators than the three I named above, but those are the only ones that all banks consistently disclose. They all attach these numbers to “base case,” “benign” and “adverse” scenarios, but they, importantly, don’t disclose the probabilities they assign to each.
And then, on top of all this management judgment on the inputs and probabilities, they add another layer of management assessment, where they can tweak their decisions on provisions even further.
So even if one bank seems more optimistic than another, the all-important matter is just how much they’ve already set aside for bad loans. A bank’s GDP forecast may look rosy, but if it has a big, fat cushion for the future, it might be able to afford a cheerier outlook.
With all that said, the banks clearly ratcheted down their economic numbers from the second quarter of 2022, ended in April, to the fourth quarter, ended in October. (One assumes they might have some downward tweaks again when they finalize the numbers for the first-quarter end of Jan. 31.)
Toronto-Dominion Bank TD-N, for example, took down its “base case” GDP scenario from 4.1 per cent in the second quarter to 1.3 per cent in the fourth quarter, and Bank of Nova Scotia went from 4.6 per cent to 1.2 per cent. All the banks that disclosed specific numbers cut GDP forecasts by at least half.
(RBC uses a line chart that does not label the points with specific numbers. Spokesperson Gillian McArdle says that approach is “easier to interpret and understand” because “it provides a full trajectory of the economic assumptions that feed our models.”)
Some of the most dramatic changes in the base-case scenarios, perhaps unsurprisingly, come from the home-price numbers. Bank of Montreal went from a gain of more than 25 per cent in the second quarter to a decline of 10 per cent by the fourth. TD went from a rise of 11 per cent to a drop of 14 per cent.
Those two, National Bank and Scotiabank all have double-digit home-price declines in their base-case scenarios. But, RBC has just a 1-per-cent decline, and Canadian Imperial Bank of Commerce has a 2.5-per-cent fall.
It is in the banks’ “adverse case” scenarios that the numbers get truly scary. RBC and TD project an eye-watering 30-per-cent house-price decline.
CIBC’s “adverse case” scenario of a 13.1-per-cent drop in housing prices, however, is better than TD’s base-case figure.
CIBC seems more positive across the three major economic statistics than any other bank, and it’s also downgraded its forecasts by less in the past two quarters than any other bank, by my math. That strikes me as a concerning but possibly unsurprising combination for a bank that has more exposure to the Canadian economy, proportionally, than its larger peers. (CIBC declined to comment.)
Again, we don’t know how much weight each bank is placing on each scenario. But Mr. D’Souza says some, not all, banks provide hypothetical extra-provisioning numbers in the case of an adverse scenario. By his calculations, then, Mr. D’Souza figures the banks’ base-case scenarios typically have by far the biggest weight.
That takes us back to the question of who has the provisions to allow for more optimism. When CIBC posted its fourth-quarter results, it stood out for having relatively higher provisions for credit losses, compared with other banks. So maybe it can handle being wrong about its forecasts.
It is Scotiabank, actually, that Mr. D’Souza says may need to set aside more money if the economy tumbles. A 100-per-cent weighting to Scotiabank’s “very pessimistic” scenario – a step below its adverse scenario – would result in a $1.1-billion increase in allowances for performing loans. (Scotiabank did not respond to a request for comment.)
Looking back, Scotiabank and its peers escaped the pandemic in much better shape than we thought. But it taught a lesson worth learning for bank-stock investors: It’s good to know whether a bank’s economic outlook is as sanguine as your own, and whether it can afford to be so optimistic.