Sunday, January 2, 2011

Canadian Taxpayers and CMHC, How Much Are They on the Hook For?

Canadian Residential Mortgage Markets: Boring But Effective?
Mortgage insurance plays a big role in the Canadian mortgage market. Federally-regulated deposit-taking institutions, including all the chartered banks, can only hold “high ratio” loans (i.e., those with loan-to-value (LTV) ratios greater than 80 percent since April 20, 2007, up from 75 percent) if they are insured against default. Hence about 45 percent of all chartered bank-held mortgages are insured. Also, all of the mortgages that back NHA MBSs must be insured. The dominant mortgage insurer is government-guaranteed CMHC, accounting for about 70 percent of all outstanding insurance. Two private insurers, American International Group and Genworth Financial, account for almost all of the rest. Since 1988, the federal government has been providing a 90 percent guarantee to private insurers.

A VERY rough calculation

45% of mortgages are insured
75% are insured by cmhc
$1 trillion total mortgage debt

$338 billion cmhc insured mortgages

by definition all are less than %30 down

estimate 15% current equity (totally pulled out of the air) for this pool
assume 40% drop in prices
cmhc will have to cover gap of

$84 billion
(plus approximately 50k per house processed)

To an American, that sounds like a drop in the bucket.
Canada: 1.6 trillion gdp
so the CMHC bailout would be ~5% of gdp

7 comments:

jesse said...

"estimate 15% current equity (totally pulled out of the air) for this pool assume 40% drop in prices
cmhc will have to cover gap of $84 billion "


The assumption here is that all underwater mortgages will default. That is not realistic; even with negative equity we can assume most people will be able to perpetually carry over their loans.

In a similar analysis I assumed 20% default rate over 20 years and a higher equity %. (CMHC in its annual report indicated around 40% of their loans are of vintages over 10 years.)

Running with these numbers I end up with a "worst-case" default of between $10-20BB. Certainly a ton of money but this will be spread over many years. Think of it as a series of $1BB security budgets without the illegal mass arrests and other human rights violations.

jesse said...

I should add that while I think CMHC is an egregious use of taxpayer money, it has managed its affairs rather well. But that should come as no shock. Insurance is a racket: in this case it's a fantastic irony that debtors get the privilege of paying a crown corporation, part of a government supposedly representing their best interests, in order to facilitate their own demise.

Leith van Onselen said...

I wrote an article on the Canadian system and the CMHC's exposure in November (see: here)

As at 31 December 2009, the CMHC had only $9.3 billion CAD of shareholder capital but a whopping $473 billion CAD of outstanding insured loans. Further, 29% of CMHC-insured borrowers had less than 20% equity in their homes (see table in article for a break-down of LTV ratios).

jesse said...

Hi Leith,

Sure but with a leveraged asset having a small reserve isn't necessarily game over for a crown corporation. There is notional and net.

I still have not seen a convincing argument for Canadian taxpayers being on the hook for more than $10-20BB, even after a 40% nation-wide drop in prices.

That statement gets blown out of the water if someone can show me CMHC has taken on a large amount of second tier loans, in which case losses are going to be worse.

jesse said...

Here's my calculation for reference (the calc I did is a bit out of date but you can follow the logic and update the #s with whatever):

$8BB for $500BB+ total insured assets. Some additional amount of loan repayments will have been established in almost all of these properties. You could go through various stress tests and I'm sure under certain market conditions CMHC will be technically insolvent.

As a quick estimate, let's say the national market drops 30% from peak in 10 years. Assuming the provision remains untouched, let's assume the average mortgage is 15% below peak prices. That works out to about 15% average loss on $500BB, or $75BB notional shortfall assuming no principal payback (which there is). This would be the amount CMHC is on the hook for if all loans defaulted.

Now assume 20% foreclosure rate on insured loans over the 10 years, so that equals $15BB CMHC must pay out over 10 years.

They will make some up by increasing premiums, tightening approval criteria, and cutting staff, as well as likely turfing some of their actual housing affordability initiatives (which MI is not). If we also include the principal reductions I think they will probably stay solvent but will be under massive pressure to reconstitute their balance sheet if these losses materialize.

I just don't see how CMHC is a "massive" liability to the government. It may well cost some billions but it's not in the order of many tens of billions from what I can ascertain.

GG said...

>The assumption here is that all underwater mortgages will default. That is not realistic; even with negative equity we can assume most people will be able to perpetually carry over their loans.

Yes, very true. I was looking for total exposure to put some perspective on the issue. I suffer from immediacy effect from watching mortgage pools decline to 90% here in the states that makes it hard to pick a "realistic" actual default rate.

What I did leave out is the taxpayer's exposure in the pool of mortgages insured by AIG, et al. They are guaranteed up to 90%, but I have not had time to research the details of that. Whenever the name AIG comes up, that naturally screams "bailout" as well, although their traditional insurance arm was much better behaved.

And I agree, I don't think it's going to bankrupt Canada to cover this. What continues to irk me is this notion that adding money to the system will make housing affordable, as opposed to simply driving up prices and putting every household deeper in hock. If we can get one, simple lesson out of this mess . . .

GG said...

Thanks for the link, Leith.