Housing and Credit Data: Housing Starts, Mortgage Applications, Bank Lending Surveys and Money Supply

What you will learn

• How to read housing starts and permits as seasonally adjusted annual rates.
• What the weekly mortgage application index measures and how to compare it.
• How net percentages in bank lending surveys are built and what tightening implies.
• What M1, M2 and M3 are and how to read money-supply growth.

Explanation

Housing and credit are the parts of the economy most sensitive to interest rates: their data react first when policy changes and reach growth and employment with a lag of several quarters. They rarely cause one-minute market moves, but they shape the medium-term picture.

Housing starts and building permits (US)

The US Census Bureau publishes New Residential Construction monthly, usually in the third week: building permits, housing starts (foundations begun) and completions. All are reported as a seasonally adjusted annual rate (SAAR): the monthly count is divided by a seasonal factor and multiplied by 12, giving a figure such as "1.40 million units" (annualisation in general was covered in lesson 1). The monthly data are noisy (weather, large apartment projects) and heavily revised, so use the single-family series and three-month averages. Permits lead starts by one to two months, and starts lead completions by roughly six to twelve months. The NAHB housing market index, a builder survey on a 0–100 scale with 50 neutral, arrives a day or two earlier and often predicts the direction. Germany publishes Baugenehmigungen (building permits) monthly through Destatis; their sharp fall after 2022 shows how quickly higher mortgage rates reach construction.

Mortgage applications (US, weekly)

The Mortgage Bankers Association publishes a market composite index of applications every Wednesday, split into purchase and refinance indices, with the average 30-year fixed rate from its survey. The index was set to 100 in March 1990, so only changes matter. Compare year-on-year, because holidays distort single weeks; read the purchase index as a demand signal and the refinance index as a rate-sensitivity signal. When mortgage rates fall, refinancing jumps within weeks, long before starts respond.

Bank lending surveys

The Federal Reserve's Senior Loan Officer Opinion Survey (SLOOS) is quarterly. The ECB's Bank Lending Survey (BLS), running since 2003, asks about 160 euro-area banks four times a year and is published in January, April, July and October. Both ask whether banks tightened or eased credit standards over the last three months and whether loan demand rose or fell, for firms, mortgages and consumer credit. Results are net percentages: share reporting tightening minus share reporting easing (the BLS also publishes a diffusion index giving half weight to "somewhat" answers). Positive means net tightening. Several quarters of strong net tightening have historically preceded weaker investment and higher default rates; this is a lagged relationship, not a timing tool.

Money supply

M1 is currency plus overnight deposits; M2 adds savings and short-term time deposits; M3 (used by the ECB) adds repurchase agreements, money-market fund shares and short-term debt securities. The Fed publishes M2 monthly (release H.6); the ECB publishes monetary developments monthly, including loans to households and non-financial corporations, which many analysts find more useful than M3 itself. Read growth year-on-year against nominal GDP growth: money growing far faster than nominal output for years was part of the 2021–2022 inflation story, but monthly changes mostly reflect shifts between deposits and funds.

Worked example

Hypothetical example: 112,000 housing units were started in a month (not seasonally adjusted) with a seasonal factor of 0.95 (a seasonally weak month). Assumptions: round numbers; revisions ignored.

1. Seasonally adjusted monthly starts = 112,000 / 0.95 = 117,895.
2. SAAR = 117,895 × 12 = 1,414,736, reported as 1.41 million.
3. Permits the same month were reported at a 1.45 million rate and starts a month earlier at 1.30 million. Permits above starts by about 0.15 million suggest starts can rise in the next one or two months.

Credit: in the latest BLS, 25 % of banks tightened standards for corporate loans and 10 % eased. Net percentage = 25 − 10 = +15, net tightening. If the previous quarter was +5, tightening is accelerating; expect loan growth to slow in coming quarters.

Mortgages: the purchase application index fell from 235 to 215 over a year: (215 / 235 − 1) × 100 = −8.5 %, consistent with the tightening.

Money: M2 rose from 21,000 to 21,600 (billions, hypothetical units) in a year: (21,600 / 21,000 − 1) × 100 = +2.9 %. With nominal GDP growing about 4 %, money grows more slowly than output, which is not an inflationary signal on its own.

Common mistake

Practice activity

On paper: (a) 95,000 units were started in a month with a seasonal factor of 1.05. Calculate the SAAR. (b) In a lending survey 18 % of banks tightened mortgage standards and 22 % eased. Calculate the net percentage and name the direction. (c) Refinancing applications rose 40 % in four weeks while purchase applications were flat. Which tells you more about housing demand, and what does the other tell you?

Key takeaways

• Housing data are seasonally adjusted annual rates; permits lead starts, and single-family series are cleaner.
• Compare mortgage applications year-on-year; refinancing shows rate sensitivity, purchases show demand.
• Sustained net tightening in lending surveys precedes weaker credit growth with a lag.
• Money supply is a slow background indicator; compare it with nominal GDP.

Next step

In the next lesson you look outward: the trade balance, the current account and capital flows, and how external balances feed through to currencies and bond markets.