House price growth cools from record levels

In the AFR I write that while some of the record heat is coming out of Aussie housing, which should be reinforced by out-of-cycle rate hikes from lenders and regulatory constraints on credit creation, the value of bricks and mortar will continue to climb.

Across metro and non-metro regions, Aussie dwelling values appreciated by a strong 1.5 per cent in September, and have risen a staggering 17.6 per cent in 2021 according to CoreLogic. Benchmarked against their previous high-watermark of April 2020, Aussie dwelling values are up almost 18 per cent. It is nevertheless noteworthy that there has been a steady decline in the pace of monthly house price increases since March this year notwithstanding the stellar ongoing growth.

When the pandemic first hit, we projected up to 20 per cent house price growth this cycle after a short 0 to 5 per cent correction in prices between March and September. This was predicated on a 100 basis point drop in mortgage rates from their April 2019 levels. Accounting for steeper falls in fixed-rates, we upgraded our forecast to 20 to 30 per cent capital gains this cycle. Recreating the Reserve Bank of Australia’s internal house price forecasting model gave us similar results.

Two important variables will dampen this ebullience. We've long suggested it is inevitable that the Australian Prudential Regulation Authority and the Reserve Bank of Australia eventually intervene to throw sand in the wheels of credit growth. It now appears that constraints on debt-to-income multiples and tougher interest serviceability tests are live “macroprudential” options under consideration.

Our chief macro strategist, Kieran Davies, cautions, however, that this is likely to punish first time buyers, who appear to be partly responsible for the sharp recent increase in the share of new home loans with debt-to-income multiples above 6 times.

Davies further notes that “new macroprudential measures would make it harder at the margin for the RBA to return inflation to the 2 to 3 per cent target band given that that the housing market is a key part of the transmission mechanism of monetary policy”.

A second handbrake on housing will be slightly higher mortgage rates, which are our central case even though the RBA is committed to not hiking its overnight cash rate until 2024.

APRA’s recent decision to force larger banks to move back in line with global best practice and hold government bonds as their emergency liquidity asset—instead of their peers’ lower-rated and far less liquid senior bonds and residential mortgage-backed securities (RMBS)—will force the cost of capital on the latter back-up to much more normal levels. (According to the RBA’s recent research, senior bonds and RMBS have a tiny fraction of the liquidity of government bonds.)

Since 2013 the average credit spread, or risk premium, that the major banks paid to borrow money via a 5-year senior unsecured bond and 3-year RMBS has been 83 basis points and 92 basis points, respectively, above the quarterly bank bill swap rate. Because the RBA lent the banks $188 billion of 3-year money at a cost of just 10 to 25 basis points, they basically stopped borrowing from markets. This dearth of issuance crushed their cost of capital on 5-year senior bonds and 3-year RMBS down to circa 27 basis points and 48 basis points, levels not seen since 2007.

Yet with the RBA’s lending facility due to be repaid over the next 2 to 3 years, banks are issuing large volumes of bonds again, as we had assumed (a coming bond supply tsunami was an idea rejected by many). Although this money is still super-cheap by historical standards, it would be reasonable to presume that over time the cost of capital will revert back to historical averages.

The crux for housing will be that bank and non-bank lenders will want to defray this cost by passing it on to borrowers. This will be evident via skinnier interest rate discounts and higher variable- and fixed-rate loan costs.

The RBA’s generous lending facility was the main driver of unusually low fixed-rates, and the huge jump in the share of borrowers fixing their interest rates rather than using variable-rates. With the facility closed and longer-term rates climbing again, variable-rate loans, which price off the RBA’s short-term cash rate, will become more popular again.

Greater bank debt issuance is being fuelled by both the need for banks to repay the RBA its $188 billion and the shift to holding more government bonds. Prior to APRA’s decision, banks were allowed to hold up to $139 billion of emergency liquidity via their own internal loans and other banks’ bonds. Internal loans made-up about 80 per cent of the total. Since APRA has said it expects banks to purchase government bonds to replace internal loans and other banks' bonds, they need to find cash to do so. And much will come through new borrowing, as we are now seeing.

In the decade prior to the pandemic, the major banks typically borrowed $141 billion each year. Our analysis implies the borrowing task ahead will be similar at about $168 billion per year on average over the next three years, which is materially higher than consensus expectations. (Note we are only talking about the four majors here—the banking system will need to borrow more.)

Since the major banks’ AA- rated bonds are among the most sought-after globally, this will be easy to achieve. But they will have to pay more to borrow, and senior bonds and RMBS will eventually become much more attractive to investors at some point. (While we have been historically active in these markets, we exited most of our exposure late last year.)

This is why the AusBond Floating-Rate Note Index recorded a rare monthly 0.06 per cent loss in September (senior bond spreads moved wider). Concurrently, the AusBond Composite Bond Index also registered a large 1.5 per cent loss in September for different reasons (higher long-term yields). The triple-header was a 3 per cent decline in global equities in the month. Given the reflationary environment, a positive correlation between equities and fixed-rate bonds, or interest rate "duration", has re-emerged.

A second big question for markets is the quantum of government bonds the banks need to actually buy. As with debt issuance, consensus views have been benign, ranging from very little (CBA) to circa $90 billion (UBS). Our research points to different numbers. Analysts appear to have overlooked the fact that when the banks repay the $188 billion they owe the RBA, this mechanically disappears an equivalent quantity of excess cash on deposit at the central bank that presently counts towards the banking system’s emergency liquidity metrics (or Liquidity Coverage Ratios). Over time, the RBA’s holdings of government bonds will also mature, which will slowly disappear more excess cash that is otherwise included in the LCRs. Then there is healthy balance-sheet growth, which needs to be funded, and these borrowings attract a liquidity charge.

Applying conservative assumptions, we find that the entire banking system will have to slowly accumulate between $250 billion and $450 billion of government bonds between now and the end of 2024. Given the quantum of government debt issuance, this should be straightforward. The advent of this demand is one reason why we have liked government bonds for some time.

The speed with which the RBA tapers its bond purchases also impacts these findings: the faster the taper, the less excess cash banks hold at the RBA, and the more government bonds they have to buy. While some are starting to argue for a faster taper, the RBA would be wise to stick to "nowcasting" and not predicate its decisions on unreliable forecasts of the future. When the RBA debates this decision in February, Australia is likely to be grappling with a striking increase in COVID cases as we begin to live with the endemic. This could dampen short-term growth, which might rationalise a smoother taper profile.

The Bank of England is fascinating in this context. There is a case for central banks normalising both short and long-term interest rates in parallel by tapering (rather than completely exiting) bond purchases while they lift their cash rates. This is exactly what the BoE has chosen to do. While it is too much to ask of the RBA given how committed it is to keeping its cash rate at 0.1 per cent for years to come, there are theoretical grounds for it to retain optionality with its bond purchase program and minimise the distance between a complete exit and the first cash rate hike.