Wednesday, March 9, 2011

Household Debt: A Different Perspective

From today’s The Star (excerpt):

Malaysia's household debt on the rise...But mortgage NPLs at an all-time low

PETALING JAYA: Malaysia's household debt rose at a rapid rate of 11.1% per annum from 2004 to 2009, and from RM516.6bil at end-2009, it climbed by 8.4% to RM560.1bil as at end-August 2010, said CIMB Research.

The household debt to gross domestic product (GDP) ratio increased from 66.7% in 2004 to 76% in 2009 but is estimated to ease to 74.6% at end-2010.

The rapid growth of household borrowings is causing some worries that the excessive leveraging by households may make the economy and financial sector more vulnerable to instability and crisis...

...As at December 2010, mortgage NPLs stood at 3.3%. Its highest level was at 8.6% or about RM14bil in 2006. Since March 2007, this figure has been trending down...

...Certainly, rising property prices have fanned an increase in borrowings. The share of household loans to total bank loans rose from 35.2% in 2000 to 55.5% as at end-August 2010. Mortgage debt accounts for 48.5% of total household loans currently.

Another banking analyst said that while the household debt level was high, Malaysians also had a very high savings rate, at about 35% of GDP...

...CIMB head of economics Lee Heng Guie said the various financial indicators indicated that Malaysia's household balance sheet remained healthy with financial assets' coverage of total debt at 2.5 times, the liquid financial assets to debt ratio remained strong at 148.6%, and the non-performing loans ratio was low at 3.1% vs 8.5% in 2004.

“The strong balance sheet enabled the household sector to continue servicing its debt despite the rise of the debt service ratio to 49.2% in 2009 (38.4% in 2004),” said Lee.

As at end-August 2010, mortgage loans accounted for 48.5% of total household loans and 26.9% of total bank loans, followed by motor vehicle financing (25.9% of total household loans).

Credit card lending accelerated 16.4% per annum in 2000 to August 2010, pushing its share of total household loans from 5% in 2000 to 6.1% currently...

...He added that on the demand side, the rapid build-up of household debt stemmed from changing demographics, strong economic growth, a steady rise in incomes and low interest rates.

To be fair, I think this article is more an initiative of the paper than any comprehensive analysis by local economists.

But here’s the problem I see: if you take the numbers at face value, you’ll think there was a problem with household debt. And there’s no doubt that there has been a significant rise since the turn of the century.

But if you’re looking back just five years, a pretty different picture emerges.

Take for instance that first line, the numbers from CIMB Research. 8.4% growth in just eight months sounds pretty hefty, until you check the numbers for non-household debt – the growth rate is almost identical.

Here’s the ratio of household borrowing from the banking sector to total loans:

01_hh_r

Do you see the difference? Neither do I. As a share of total loans, lending to the household sector has been almost flat in the last five years. If you do a breakdown by loan purpose, the shares there are flat too – mortgages, auto, and credit cards. The only household loan category that’s climbing is “personal use”, which took in an extra 1% share since 2005 to about 4.8% – but that’s flattened in the last year as well.

Here are the ratios of lending to nominal GDP (left axis: household borrowing; right axis: total borrowing):

02_ratios

Not much of a difference here either – I could probably be do this more formally with a statistical test, but I’m not going to bother.

Now, if household borrowing were actually increasing more rapidly than total borrowing, then I agree we could be seeing a problem with increasing household debt. But that’s not really the case here.

What we really ought to be talking about is the increase in debt loads generally, but that distinction seems to be missing from this debate. If there is a concern with household debt, it should be with its distribution i.e. how much is being taken on by lower income households. The aggregate balance sheet numbers are comforting, but those could hide a lot of heterogeneity within the household sector itself. But these are numbers we simply do not have. For instance there’s no doubt that national savings are high – but that does not imply household savings are high. Most of the savings accrued over the past decade has been corporate savings, not household, which implies households that have managed to increase their savings have been those whose incomes come mainly from ownership of capital i.e. higher income households.

But getting back to the real issue, it’s higher credit expansion for both household and non-household sectors that should be the concern, not just household borrowing. And while we’re at it, total financing generally and not just bank loans – an increasingly bigger portion of bank finance is actually being funnelled through the capital markets:

03_pds

Just to make my point perfectly clear, here’s the ratio of bank lending to households against total financing (bank loans + bank holdings of PDS + bank holdings of MGS), with two different sample sizes:

04_hh_fin1

05_hh_fin2

If all you saw was the growth number from the beginning of the first sample to now, you’d think there was a problem. And if you all you saw was the growth number from the second sample, you’d think everything was hunky dory. The answer is probably somewhere in between.

12 comments:

  1. nice one...I knew ur gonna come up with a post to respond to that...

    Just to check does the Household data includes "margin financing" for Equity trades?

    D long term view from 01 paint a pretty diff pic, why the shift in portfolio composition?

    tks bro

    ReplyDelete
  2. http://is.gd/ZhLA7X
    http://is.gd/e7woBd

    ReplyDelete
  3. Bro satD,

    To the question of margin financing, yes. it's included. The biggest thing missing from my data is non-bank borrowing, which seems to be pretty hefty. I've no idea of the composition of that.

    The portfolio change came about from a couple of factors:

    1. Malaysian corporates suddenly discovered the joys of bond financing. It's cheaper for large scale borrowing, you can match maturities to projected income, and it's easier to put off a horde of bond holders rather than a single bank if you get into trouble :).

    2. Banks made a conscious decision to go after the household market post-1998 after being burned by corporate and SME lending.

    ReplyDelete
  4. Thanks for the links walla, it's clear The Star is cherry-picking from research papers to prove their point.

    ReplyDelete
  5. I just think your conclusion ratio of household debt to total loan is flat over years means no unhealthy growth in household debt is questionable, what if total loan increases a lot and so do household debt, with low interest rate and aggressive personal loans marketing by banks, household debt had snowballed. Statistics is good guidance, but it is also the most deceptive tool, a lagging indicator at best. I am seeing lots of dangerous spending behavior among friends, beware.

    ReplyDelete
  6. I didn't come to this conclusion in a vacuum. Loan growth has only started to pick up in the last year or so.

    Actually for the first half of the last decade, loan growth was pretty horrible. I remember BNM begging banks to lend, and they wouldn't comply.

    The funny thing is, people are now getting concerned about household borrowing, when corporate borrowing has been much higher the last few years.

    ReplyDelete
  7. If you zoom into the NPLs for household debts, it has risen in all segments from 2009 to 2010. Auto up from 1.2% to 1.5%; housing 3.3%-4.2%; personal 2.4%-3.6%; & credit cards 1.9%-2.2%. I would view the increase as significant. That was probably why certain measures were introduced by bnm. Is it too late? My worry is housing, the bulk of household debts. Completed stocks of property to be released into the market over the next few years will likely test the market as the loans would require to be serviced. I am not optimistic.

    ReplyDelete
  8. Hi Anon, sorry for the late reply, I wanted look at the latest official data first.

    I just don't see what you're seeing. Both quantum and ratio of NPLs in the household sector have fallen, not risen.

    What you're probably seeing is the change in reporting requirements and classification, not a real increase in bad loans per se. From Jan 2010, banking institutions are required to conform to FRS 139 which required new guidelines that superseded BNM/GP3.

    These new guidelines appear to have resulted in higher reporting of NPLs, or impaired loans as they are now called. There's a jump from 1.8% in Dec 2009 for net impaired loans under the old classification to 2.0% in Jan 2010 under the new. The new measure peaked at 2.5% in Jul 2010, and has since fallen to 2.1% currently.

    ReplyDelete
  9. Hi Hisham,

    Apologies, I read the report wrong way round. You are right. There were declines in NPLs. Maybe my "gut" is pessimistic and saw/read everything negative.

    Anyway, I noticed that many property launches recently have zero interests during development. They require no servicing/payment during construction. I can only assume that the interest is already priced-in. Given so, aren’t our bankers over-extended against the asset/collateral value? They are effectively putting up financing on the cost of developing the property plus developers’ margins plus interests for 2-3 years. In the past, only the first two are finance-able. This weakens the collateral cover. Insufficient cover can snowball quickly if the prices fall even marginally, like in the US. The ingenious schemes by the developers have fuelled significant speculative demand. The eventual risks will likely land on both the financiers and the buyers.

    The test is when these properties are handed over and the servicing of interests and principal begins. The next 3-5 years or so would be crucial and I am quite certain that NPLs will surge as a result. I read that as-it-is, 50% of the household disposable income goes to servicing of debts. This is apparently high and I wonder if there is anymore room to maneuver. Unless income grows quickly to service these debts, I don’t think things are going to work out well.

    Maybe someone should do a report on the high-end segment of housing to see if the present prices are sustainable.

    ReplyDelete
  10. Yes, we're going down the same route as the US and others. It's not a systemic problem yet, but the potential is there. Initiatives like the PR1MA are remind me of the Clinton/Bush initiatives to raise house ownership, with the results we see now.

    I'm less concerned about what the developers are doing than the apparent disconnect between BLR and actual lending rates. You're right in that the fundamental value of housing is being distorted, but cheap credit is more worrisome. The limited pass through between interbank overnight rates into longer lending rates is seriously limiting BNM's ability to manage asset price inflation.

    ReplyDelete
  11. Is there any place I can get malaysia yearly household disposable income and all the policies related or to affect household?

    ReplyDelete
  12. As fas as I know, that data is simply not available. I don't believe DOS compiles the national accounts using the income approach.

    ReplyDelete