AFG’s LVR revisions

As my readers would know I try to follow AFG’s loan reports every month. They claim to be a good leading indicator for the ABS data that doesn’t appear until 6 weeks later. I have however had a long term issue with their LVR data because it made absolutely no sense.  I posted about this point back in March.

While I was checking out their latest report I was amazed to notice that the LVR table has been significantly revised upwards. What is even more amazing however is that AFG seem to have retrospectively updated all of their previous reports with the newer data, yet as far as I can tell haven’t provided an announcement about these revisions.

Here is a screenshot of the LVR chart I took from the May report back in May.

and here is how that chart now appears in the same May report.

Funnily enough they have neglected to update the comments in the reports they have retrospectively updated. Comments specific to the old data still appear in the February report

…. in addition, the AFG Mortgage Index shows that Loan to Value Ratios (LVRs), the value of loans expressed as a percentage of the value of properties, fell to 53.2% in February. This is the most conservative LVR figure AFG has recorded in six years, showing that mortgage buyers borrowed only around half the value of the properties they were buying or refinancing. LVR figures in the long term have tended to be around mid 60%.

The LVR actually went up in February according to the revisions.

So how are we supposed to interpret this statement in the SMH from Tuesday?

Australians are borrowing less as house prices ease, further entrenching the culture of reducing debt set in train by the global financial crisis. People taking out mortgages in June signed up for loans representing a lower percentage of their property’s value than at any time during the previous six months, according to data from Australia’s biggest mortgage broker, AFG.

The average loan to valuation ratio (LVR) was 64.2 per cent, with average LVRs falling in NSW, Victoria, South Australia and the Northern Territory.

Should I believe them this time ?

Latest posts by __ADAM__ (see all)


  1. I’d like to see that figure broken down. Can’t imagine that FHB’s in NSW, ACT and NT would have LVR’s that ‘good.’

    • these are not FHB LVRs, but total – including upgraders and “investors”

      FHB LVRs are probably at 95-97%

  2. Doesn’t that indicate a higher property value sell-down? Sell a $5m property and buy a $3m one on the trade-down, and you need less finance, if any at all. So the great demographic sel- down may be upon Australia.

  3. Oops! Took them a fair chunk of time to pick up that mistake! I suppose it wasn’t until the LVR was below 50% that they realised that it couldn’t possibly be right. That’s what happens when you use Excel to do your stats 🙂

    That’s strike two for them in my mind. Strike one was not understanding the seasonal impact of Easter on their own numbers, and talking up the jump in May.

    They must be missing four phds – it looks like they already have a half.

  4. Bloody Excel formulas, don’t you hate it when you set it to manual recalculate and then forget to hit F9… also shows that you should save company’s predictions / data as you go along. Those OESR Qld vacancy rates from March 2011 came out in June without a date on them.

  5. I heard rumours of a software problem that has only recently been fixed, such that data that was missed during the problem has only just come through – hence the revision.

  6. Sandgroper Sceptic

    Nice catch. It is pathetic they did not own up to their errors.

    That makes the bear case for property a lot worse. When property prices fall those LVRs are going to be ratcheted up markedly and then the screws will be turned on the banks, do they insist on margin calls with resulting mortgage defaults and repossession, opt for extend and pretend (40, 50 yr mortgages), or whine piteously to the government for help. My money is on the latter!

  7. I grabbed everything under that page with wget to check the file date changes — all PDFs named “mortgage-index-” followed by jan, feb, mar, apr, may, or jun were modified on 24 June at 13:13 (except jan11-sa which still has the dodgy figures). Very sloppy and underhanded.

  8. The smoking gun…

    Why would the banking system LVR on new loans 65% suddenly collapse toward the National LVR of 30%?

    40% of all new loans refinance

    During the same peiod an additional 100,000 dwellings are for sale and the national dwelling price is declining.

    The answer is… the original data was correct and the ammendments are an attempt tp cover-up a financial crisis!

  9. I do not believe their story. They seem to be getting many MSM stories now and something is not right. The data could not have been that far out!

  10. Do those stats include all the interest-only loans out there? The small writing at the bottom of their PDFs list common loan types but don’t mention interest-only ones?

  11. Refinance is 40% of all new AFG loans, therefore bank lending (as AFG are a broker i.e. a conduit).

    To qualify for refinance (we turn a blind eye to your sub-prime household finances… exagerated income – understated expenditures i.e. your in-ability to service your existing debt… sub-prime) and asset prices are flat or falling, so we subtly suggest you ask mum & dad to put up their title as security to refinance your loans…

    Voila… Banking system 65% LVR plummets toward national LVR 30% as AFG refinaces sub prime borrowers.

    If Mum & Dad say no way, you are on your own… 100,000 additional listings in 12 months.

    Household mortgage interest service is 12% of disposable household income (according to the RBA) vs 9% in 1990 when mortgage interest rates were 16%

    Anyone not think a heap of sub prime lies are NOT being swept under the carpet?

  12. Leith, do you have the data on the number and amounts as well. If so could you check it for alterations for me. I want to ensure it was just the LVR’s that altered.

    I was advised by AFG that they had a software upgrade in December 2010 which affected LVR’s after that date, and the error wasn’t picked up for some time.

    Their mistake was in not advising of the error in their next media release. It is always worse if you don’t come clean.

    Better still you could email a copy to me if possible.


  13. Peter Fraser: I used pdftotext and diff to check the differences for the Jan to May files (national and qld, my old ones vs the ones updated on 24 June, I’m missing jan11-national). The only differences are in the LVR lines.


    Peter W, you make very interesting points and I am impressed by your posts, particularly the guest post you wrote for DE. But I am a mathematical rather than a financial type and your writing seems very “compressed” and “terse” for me. I’m sure Deep T and regular financial professional types who post on ftalphaville / zerohedge / nakedcap can follow you, but I can’t.

    You’d have to spell out everything for me as if I were a first grader (like for Steve Eisman in “The Big Short”) because I can’t guess what the steps of logic are between “interest service rates are higher despite interest rates being much lower” and “there must be hidden sub prime lies” (presumably based on the corrected AFG figures).

    In other words, it may be that the people who understand your writing already understand the issues you’re talking about so you don’t need to write for them 🙂

  14. richard b

    Interest service…

    ‘Household finances’

    So when mortgage interest rates were 16% in 1990 the household interest service was 9% of household income

    Today mortgage interest rates are 7.5% and the household interest service cost is 12% of household income.

    We all get to see how much subprime (if any… I could be wrong) during the next 12 months.

    I don’t own any Australian bank shares!

    • It is clear from Peter Ws sharp and analytical writings that he was not educated in Australia. Hence the brevity but in place we get deep insights.

      I have followed him at debtdeflation and his sparring with another bright chap with the moniker bb (who again seems not to be educated from Australia )

      Definitely could be Europeans 🙂