National Bank Ahead of the start of second-quarter earnings season for Canadian banks next week, National Bank Financial’s Gabriel Dechaine trimmed his earnings expectations based a “flattening” of his net interest margins forecast, rising expense growth and a “more conservative view” on provisions for credit losses.
The result was a 3-per-cent decline in his earnings per share projections for both the second quarter and full-year 2023. His estimate for 2024 declined by 4 per cent.
“Big-6 bank stocks are underperforming the S&P/TSX by close to 400 basis points so far this year,” said Mr. Dechaine. “They are trading at a NTM P/E [next 12-month price-to-earnings] of 9.2 times on consensus earnings, or 10 percent below their five-year average. However, consensus earnings include an industry PCL ratio of roughly 30 bps over the next year, which is below the historical average of 35 bps and well short of the 70-80 bps seen in prior recessions. If we simply assume a doubling of PCLs (i.e., to 60 bps), the pro forma P/E of 11.2 times looks a lot less compelling (10 per cent above the five-year average).”
In a research note released Thursday titled Let the numbers do the talking (down), Mr. Dechaine said “it is is difficult for us to get excited about anything” in terms of stock calls, declaring “broken record alert: still too early.”
However, he does think the “short-term risk/reward” on Toronto-Dominion Bank is “compelling.”
“The stock is trading at a mid-single digit discount (compared to a typical mid-single digit premium),” he said. “We believe it will be a relative NIM outperformer this quarter. And we could see some modestly positive news on the capital deployment front (e.g., elimination of DRIP discount). Longer-term questions on capital deployment cloud TD’s outlook, though we do not believe this issue matters much next week.”
After his forecast reductions, Mr. Dechaine cut target price for TD shares to $94 from $98, maintaining a “sector perform” recommendation. The average target on the Street is $95.77, according to Refinitiv data.