China Financial Services: Assessing impact from lower rates on existing mortgage

We expect 1) the pace of mortgage replacements to accelerate on a one-off rate cut on existing mortgages. Previously, we estimated ~4 more years to complete the mortgage replacement, assuming ~10% replacement completes in a half year and that the replacement cycle started in 2H22, and that ~20% of the mortgage book had been remortgaged by end-1H23.(Exhibit 1)

2) bank divergence is set to widen on mortgage growth. We expect larger banks to see more and earlier mortgage replacement, while smaller banks, with lower mortgage exposure, are likely to take more market share, though lower overall rates will hit topline growth. This was already evident in 1H23 bank results, where we note large banks like CCB, BOCOM and ABC reported decreasing mortgage balance, while PAB and BONB saw a net increase in their mortgage books.

3) lower rates should protect mortgage asset quality. We estimate ~76% mortgage balance was originated since 2018 and ~22% is accounted for by first-time buyers in tier 3 or lower tier cities. Drops in property prices in low tier cities would reduce home equity values, which could be offset by lower rate on the mortgage book.

4) banks could find it hard to maintain NIM without cutting deposit costs.

To quantify the bank NIM impact, we run two scenarios: with and without a deposit cost cut.

1) With a rate cut on the existing mortgages of 100bps and no deposit cost cut, bank NIM would decrease by 11bps across our coverage, or 1.3ppts ROE, on mortgages at the current level of 11% of interest earning assets (IEA).

In this scenario, we highlight banks like CCB/BOC would see large NIM decreases given their larger % of mortgage exposure.

2) With rate cut on the existing mortgage book of 100bps, and a deposit cost cut by 16bps, bank NIM would be neutral, on mortgages at 11% of IEA and total deposit at 75% of interest earning liabilities (IEL).

We expect divergence of NIM on bank funding mix. Banks with more deposits as % oftotal IEL and more time deposits would be better placed with lower requirement for deposit cost cuts to offset rate cuts on existing mortgages.

In this scenario, we highlight BONB/BONJ would require 5/6 bps deposit cost cut vs. 18bps on average of the banks we cover.

But we expect that with the base case of deposit cost cut of 16bps as per scenario 2 above, bank NIM pressure would persist, as other asset yields would go down further. This is because bank deposit rate functions as the system anchor rate, given the bank deposit size, which is 70% of banking assets, or 75% of banking liability. So to better offset bank NIM pressure, we would expect more than the base 16bps of deposit cost cut (see Exhibit 4).

The PBOC has guided banks to lower the deposit rate, and we note that the new deposit rate on 3yr/5yr has been decreased by 15bps, while bank deposit costs remain high for a sample of banks (those that had report results by Aug 29) reported deposit costs increase by 9/12bps hoh/yoy overall in 1H23. This is mainly driven by 1) deposit mix change, more time deposit taken by banks, despite a decrease in the cost of new deposits, and 2) extended deposit maturity, as seen for ABC (Exhibit 10), with increasing portion of 3yr/5yr deposits in the mix.

We estimate ~10 months as the average maturity of deposits, indicating a deposit replacement cycle would take ~10 months. But given more time deposits with longer maturity, the deposit replacement cycle could be longer than expected, thus leading to higher for longer deposit costs.

In summary, we expect lower rates on both existing mortgages and deposits as the base case, but remain less positive on bank NIM, owing to the structural growth of time deposits which slowdown the scope for deposit cost decreases to offset lower mortgage rates. Within our coverage, we reiterate Buy on BONB on relative protection for its NIM given its low mortgage exposure.