The Big Short II: Why shorting high loan-to-value Canadian mortgages may be the (second) greatest trade ever

Hi everyone

I completed a pitchbook which presents a thesis on shorting high loan-to-value Canadian mortgages. My research claims this trade opportunity is comparative to John Paulson's shorts on subprime mortgages which he established from 2005 to 2007.

I plan on sending the pitchbook to hedge fund managers with the intention of persuading them to pursue the trade.

I am making the pitchbook available on slideshare so I can get the WSO community's feedback and comments. Please feel free to point out errors, omissions or flaws

Please head over to slideshare to view my slidedeck titled The Big Short II: Why shorting high loan-to-value Canadian mortgages may be the (second) greatest trade ever

You should be able to find my slide deck by searching for my username: ndsouza22. The slide deck is accessible on my profile

Edit:

I came across a report on Canadian CDS market by the Bank of Canada. I have made the report available on my sideshare account as well

Hi Anonymous Monkey, upload your resume and land a job

Members that upload a resume get 2.3x the number of interview invites through the Talent Oasis. Learn more.

Comments (59)

Nov 26, 2014

I appreciate the work but this trade has not exactly been off the radar for many people.

Also, I can't find your deck.

Nov 26, 2014

@mrb87. Know any hedge funds that are currently pursuing it? I know a few hedge funds are shorting Canadian banks and the Canadian dollar but I'm not aware of any hedge funds shorting high LTV mortgages via CDS

Nov 26, 2014
aggregate:

@mrb87. Know any hedge funds that are currently pursuing it? I know a few hedge funds are shorting Canadian banks and the Canadian dollar but I'm not aware of any hedge funds shorting high LTV mortgages via CDS

No. Biggest difficulty with the trade is that it's hard to get direct/pure exposure. I'm not a macro guy so it's not something I've spent much time trying to construct.

Can you post a direct link to the slide deck?

Nov 26, 2014
mrb87:

I appreciate the work but this trade has not exactly been off the radar for many people.

Also, I can't find your deck.

http://www.slideshare.net/ndsouza22/the-big-short-...

Nov 26, 2014

@"mrb87"

Unfortunately WSO does not allow me to post links as a new user

Head over to slide share. Search by my username ndsouza22 and you should be able to find it

Nov 26, 2014

I can't seem to be able to access slideshare... I don't think this is a particularly original "Big Short", but would be curious to see the logic and the expression. In the past, most of of the arguments I have seen have been rather superficial and generally underwhelming.

Nov 26, 2014

@martinghoul

See @MilitaryToFinance post with a direct link to my slide deck

Nov 26, 2014

Question, if I may...

Leaving aside the implementation questions, your thesis is based on the assumption that mortgage rates will rise from their recent lows. Why do you presume that this is going to happen? Moreover, even if it does happen, given that most of the expressions you're looking at are negative carry, it's absolutely crucial to estimate when you expect this to happen. Timing will have a pretty monumental effect on the attractiveness of the trade. For instance, what if nothing happens for 3 years?

    • 1
Nov 26, 2014

@Martinghoul

1. First, let me address the issue of timing:

Timing is crucial. Too early and you may miss the payout. Too late and you may be priced out of the trade. John Paulson had a 3 to 5 year time frame for his subprime trade and got in early so he was able to execute trades with attractively priced CDS.

So yes timing is everything, especially when dealing with negative carry trades.

2. Second, let's address timing along with expectation of rising mortgage rates or defaults

Theoretically, I can model the expected default rate if I have the following data on current high LTV mortgages:
- household income of each high LTV mortgage
- monthly expenses for each high LTV mortgage
- mortgage rates and terms/metrics on each high LTV mortgage
- non-mortgage debt expeneses and terms

Based on the above data, I could theoretically determine the default rate at each incremental increase in mortgage rates i.e. how many high LTV mortgages would default if rates rise 100 basis points. In reality, this data is either too granular and not available or accurate modeling of this is not feasible (at least not feasible in a timely/accurate manner)

it should be noted that default rates can theoretically increase without a corresponding increase in mortgage rates. Paulson's thesis was based on subprime mortgage holders defaulting on their mortgages when real estate prices stopped rising as they wouldn't be able to refinance their mortgages. This ultimately proved to be correct. In Canada, the subprime market is 7 percent versus 30 percent (approx) prior to the subprime crisis. Defaults in subprime mortgages in Canada will be sufficient to cause a rise in mortgage rates across the mortgage market.

So the next question is when will real estate price stop rising. Again, based on my research, theoretically this will only occur when all low LTV mortgages (below 80%) refinance or purchase new properties which would transition them into high LTV mortgages. Realisticly, it is very unlikely for low LTV mortgages (about 72% of Canada's mortgage market) to transition to high LTV mortgages as lenders will restrain thei risk exposure before saturation and all current owners are unlikely to sell/upgrade/purchase a home.

Also, mortgage rates don't have to rise substantially for defaults to increase. According to a survey by BMO, 20 percent of mortgage holders claimed they would have difficulty making mortgage payments if rates rose 2 percent.

Ultimately, the decision to execute this trade is based on the attractive of its pricing. If CDS on high LTV mortgages can be obtained at 250 bps or lower then this trade is a no brainer. If you're managing a $1 billion plus portfolio, buying insurance which theoretically can hedge your entire long exposure in a rising rate environment at 250 bps or less is extremely attractive.

Nov 26, 2014

Indeed, timing is of the essence, which is really what I am getting at here. I would suggest that you quantify your timing assumptions, i.e., for a given expected default rate, the risk/reward is X1, assuming your scenario is realised in Y1 years; X2, if it takes Y2 years etc. It would make it a lot easier to understand the sensitivity of your trade to timing. Needless to say, what I am trying to convey is that you have to take the "inspiring" Paulson "Big Short" story with a very large pinch of salt. Due to survivorship bias, all these dramatic "legends" may be a lot less impressive than they look.

I understand, although you still haven't quite addressed my question. Assuming mtge rates don't rise due to defaults and falling property prices (otherwise, you have yourself a circular argument), why are you so certain that mtge rates in Canada will rise within the trade's timeframe? 2% or 1% or any amount whatsoever?

    • 1
Nov 26, 2014

The quick and simple answer to your question is: I'm not certain.

There is no methodology or math that can prove mortgage rates will rise in 3 years or 5 years with absolute certainty. The goal of my analysis was to demonstrate the risk of default among high LTV mortgages when rates rise.

The timing of the trade is straight forward once you understand how the CDS contract pays out. You pay an annual rate (i.e. premium) and get paid out on defaults that occur during the period the CDS is in force. If no defaults occur then you lose 100 percent of the premium paid.

So the risk/reward is dependent on the pricing of the CDS contracts. So the return formula is (Default Rate x Nominal value of Bonds) / (CDS price x Nominal value of Bonds).

Time is only a factor in the payout in relation to the default rate. A long time horizon requires a higher default rate to breakeven on the trade. This formula is (Time Horizon x CDS premium) / Nominal Value of bonds covered. For example, if the CDS premium is $20 million annually to cover $500 million in high LTV mortgages and assuming your time horizon is 5 years: Default rate required = (5 x $20M) / ($500M) = 20%. In other words, you require more than 20% of covered mortgages to default within 5 years to generate a positive return. The required default rate will vary in relation to CDS pricing (bps of nominal covered) and time horizon.

Circling back to the question about rates. Mortgage rates can't stay at record lows indefinitely. Of course, there's no way to determine when they will rise with certainty. However, the Federal Reserve is targeting a rate increase in 2015 and other central banks are likely to follow. But again, there is no certainty here either.

A fund manager would simply have to evaluate the risk/reward based on the price of CDS covering high LTV mortgages and make a decision on the trade, Even John Paulson who was certain he was right, had to wait nearly 3 years before the trade panned out. Along the way he questioned his analysis and doubted if the trade would work. There will always be a degree of uncertainty that you can't eliminate mathematically; this makes finance simultaneously thrilling and terrifying. If it weren't for uncertainty mathematicians would be billionaires :)

    • 1
Nov 27, 2014

I know how CDS contracts work. I am suggesting that you perform this sensitivity analysis in terms of timing, rather than in terms of the default rate. Like I said, fix the default rate at, say, 20% and the cost of CDS is X. What is your risk/reward if the default rate realizes at 20% in year 1, 2, 3 etc? The reason I am suggesting this is that these trades only make sense as low-cost, "lottery ticket"-like options, which give you a lot of leverage at low cost if you're right. So I would always want to see the numbers like "if I'm right in 1y, this is a 5:1 bet", "if in 2y, this is a 3:1 bet" etc etc. This helps to assess the attractiveness of the bet, as well as its sensitivity to timing.

Why can't mtge rates stay at record lows indefinitely? Happened in a few places in the past and looks like it might be happening across the western world as we speak (it's a "beautiful deleveraging", to quote Ray Dalio). Even as we speak, both the severity and the timing of Fed hikes is being pushed back in the market and by forecasters. Apart from that, recall that, even if you assume that the Fed hikes, BoC is starting from a higher initial point, having already hiked in 2010. Furthermore, the BoC has told us that, in the distant future when rates do rise, they will not rise a lot. So the rise in mtge rates which is your basic premise, if it occurs at all, may not occur as quickly or as sharply as you require. Which, ironically, brings us back to my point about timing above.

Finally, let me offer you some bad news. There's no way to trade CDS on high LTV CAD mtges directly. At least, not to my knowledge and I have explored this issue previously.

    • 1
Nov 26, 2014

Regarding timing of rate increases, there are many people in the following camp:

- Fed doesnt give a shit about unemployment or other useless stats
- Instead, Fed only cares about devaluing currency to keep cost of borrow low
- Eurozone & Japan will be in ZIRP forever
- China continues to approach ZIRP to manufacture growth
- Therefore, rates will never rise (not anytime soon anyways)

Nov 26, 2014
Cries:

Regarding timing of rate increases, there are many people in the following camp:

- Fed doesnt give a shit about unemployment or other useless stats

- Instead, Fed only cares about devaluing currency to keep cost of borrow low

- Eurozone & Japan will be in ZIRP forever

- China continues to approach ZIRP to manufacture growth

- Therefore, rates will never rise (not anytime soon anyways)

Quick point here. Even with ZIRP policies the economy reaches a point of debt saturation.

For example, if the average person has an annual income after expenses of $12,000 and debt costs 1% annually (debt is never lent out at 0% even with ZIRP). Then the maximum debt a person can carry is $1.2 million. An individual's sensitivity to interest rates increases as they approach their debt saturation point. Theoretically, the maximum probability of default is at the debt saturation point. In reality, this point is never reached on aggregate as either lenders or borrowers will eventually restrain risk.

The longer interest rates remain loan, the closer we get on aggregate to the debt saturation point. The closer we get to the debt saturation point the more sensitivity borrowers are to interest rates and a higher rate of default occurs for every 100 bps increase in rates.

It's a slow motion train wreck that can take years and sometimes a decade to unfold

Nov 27, 2014
aggregate:
Cries:

Regarding timing of rate increases, there are many people in the following camp:

- Fed doesnt give a shit about unemployment or other useless stats

- Instead, Fed only cares about devaluing currency to keep cost of borrow low

- Eurozone & Japan will be in ZIRP forever

- China continues to approach ZIRP to manufacture growth

- Therefore, rates will never rise (not anytime soon anyways)

Quick point here. Even with ZIRP policies the economy reaches a point of debt saturation.

For example, if the average person has an annual income after expenses of $12,000 and debt costs 1% annually (debt is never lent out at 0% even with ZIRP). Then the maximum debt a person can carry is $1.2 million. An individual's sensitivity to interest rates increases as they approach their debt saturation point. Theoretically, the maximum probability of default is at the debt saturation point. In reality, this point is never reached on aggregate as either lenders or borrowers will eventually restrain risk.

The longer interest rates remain loan, the closer we get on aggregate to the debt saturation point. The closer we get to the debt saturation point the more sensitivity borrowers are to interest rates and a higher rate of default occurs for every 100 bps increase in rates.

It's a slow motion train wreck that can take years and sometimes a decade to unfold

Good explanation. Agreed

Nov 27, 2014

But this assumes that the average person's nominal income remains constant? This is a rather questionable assumption, don't you think? Actually, in Canada what has been occurring is precisely the sort of thing you'd expect to see at this point in the cycle. Specifically, the total DSR (Debt Service Ratio) is falling, as growth in median income is well above the nominal cost of debt.

A "beautiful deleveraging", indeed...

Nov 27, 2014

@martinghoul

To address your points

1. Okay. Showing the CDS payout at a fixed default rate is relatively straight forward.

For example, let's assume a 10% default rate on $1 billion bonds at a cost of $15 million per year. Your payout is $100 million and your returns per year are: year 1 = 666%, year 2 = 333%, year 3 = 222%.

I chose to highlight how the variation in default rate effects return rather than the variation in timing. But, I can include a graph that shows timing effects as well

2. In regards to low mortgage rates indefinitely. Indefinitely means in perpetuity. Surely you're not claiming that rates will stay low forever? Rates can stay low for a long period of time but they can't be low forever. This circles back to a point i made earlier about debt saturation.

Even at a near zero cost of debt the economy has a point of debt saturation. The longer the cost of debt is kept low the closer individuals are to their level of debt saturation. Canadians have the highest household debt to income ratio in the developed world at 163%. This contrasts with a 120% household debt to income ratio for Americans prior to the subprime crash. As Canadians total debt load increases their sensitivity to debt also increases. If Canadians approach maximum debt saturation then a small move in interest rates will be sufficient to cause defaults.

My analysis shows that the entire high LTV mortgage market will default at mortgage rates of 8.8 percent or higher. That's only 400 basis points above the current average mortgage rate. If Canadians continue to accumulate mortgage debt then the a lower relative rate increase is required for default.

In my opinion, the claim that rates will stay low forever is analogous to the claim that housing prices would never fall prior to subprime crisis

3. In regards to nominal income. My analysis determines the mortgage rate increase required for high LTV mortgages to default en masse assuming constant nominal income and expenses. I can model the rates required going out 5 to 10 years with a wage growth assumption. Obviously, this will result in a higher mortgage rate for en masse defaults. I don't expect there to be a material difference if we assume wage growth in line with the BOC's inflation target of 2%. But, I'll have to do the math to certain.

    • 1
Nov 28, 2014

Well, no, by "indefinitely", I mean that mortgage rates will stay low for such a long period of time that it would make the trade that you propose impractical (i.e. it's very expensive to buy a lottery ticket that's valid for the next 10 years of lotteries).

As to rates remaining low forever, do you know what the yield on a 30y CAD govt bond is these days? It's arnd 2.4%, which is all of 140bps higher than the current BoC base rate.

Welcome to the brave new world of "secular stagnation", "lowflation", etc etc... I just want you to be aware that this is the broad macro thesis that you're fading here. You could be right, but it's tough going, at least for the moment. Good luck, regardless!

Nov 28, 2014

@martinghoul

Thanks. Appreciate the feedback

Nov 27, 2014

getting bespoke cds done when most banks are scaling back or in DB's case getting completely out of single names will be a bit tricky and likely prevent them from having such lottery ticket payouts even if you can get them done.

Nov 29, 2014

I was about to send this to my friend who works for the #1 PM in Canada, but refrained to because your pitchbook does not look nice. If you can't invest the time to make it look nice, it makes me think that you didn' invest the time in what whatever research you did as well.

Nov 29, 2014
wikileaks:

I was about to send this to my friend who works for the #1 PM in Canada, but refrained to because your pitchbook does not look nice. If you can't invest the time to make it look nice, it makes me think that you didn' invest the time in what whatever research you did as well.

Can you be more specific. What aspects did not look nice. Additionally, it would be great if you could provide resources or an example of a nice presentation

Nov 29, 2014
aggregate:
wikileaks:

I was about to send this to my friend who works for the #1 PM in Canada, but refrained to because your pitchbook does not look nice. If you can't invest the time to make it look nice, it makes me think that you didn' invest the time in what whatever research you did as well.

Can you be more specific. What aspects did not look nice. Additionally, it would be great if you could provide resources or an example of a nice presentation

https://www.scribd.com/doc/187070283/UBS-M-A-Pitch...

I would suggest you use a similar template. Use the slide master view to design your own template. It only takes about 5 min.

Nov 29, 2014

@kyc133enydc

I have to signup for a subscription to view the slide. Can you email me the slidedeck? [email protected](dot)com

Nov 29, 2014

DB post

Nov 29, 2014

Seems like Canada has lots of commonalities with the Australian housing market.. overvalued property relative to disposable income, first-home owners locked out of the market, record low interest rates, etc etc. Have you thought about how well your thesis applies to Australia? Why is this trade better executed in Canada? Is it simply because Australia doesn't have as much securitization and CDS activity?

Nov 29, 2014

@"diverse_kanga"

Honestly, I haven't looked into the Australian market. It's very possible that Australia's real estate market may have worse fundamentals than Canada.

Looking at the numbers quickly;
- Australia's median house price is US$489,538
- Australia's average annual household income is US$99,264 (assuming 2 wage earners per household)
- This yields a price to household income of 4.9. Comparatively, Canada's ratio is 5.5 so on this metric housing is more overvalued in Canada. I would have to look at more data to determine if its overvalued on other metrics

The main issue with this trade. Outside of CDS on American subprime mortgages it is very difficult to obtain CDS as no one wants to take the other side of the trade. I was speaking with a PM in Canada and he said he couldnt find a CP for CDS on Canadian mortgages and he had $4.5 billion AUM. He has resorted to shorting CAD.

In order for this trade to work, you need to find the Canadian Greg Lippman. Lippman was a trader at DB that created the CDS market on subprime. I am a bit surprised that CDS on Canadian mortgages lack CPs. Canada's financial system has a reputation of being the safest in the world and the current default rate is less than 0.5%. I am not sure why Canadian banks or investment funds have an aversion to CDS. Maybe its risk aversion specific to Canadian culture.

    • 1
Nov 29, 2014

There is too much debt in Canada and there are housing bubbles in two of the three largest cities in the country. You could basically say the same thing for Australia, Norway, Sweden, Hong Kong, and many other countries. Such is the World we live in. That said, it is always good to see people actually producing independent research and trying to add value rather then just complaining about how nobody is hiring them or noticing them.

If you are writing this slide deck to get a job then I think this is a decent start, but there is no "smoking gun" that I haven't seen elsewhere. Others have done much more detailed work and so I dont think you will get much traction showing this to well informed hedge fund people in its present form. IMHO, you need to produce something that others don't have or find a story that is less well-flagged. People have been writing about this topic since before the 2007 blow-up in the US and Canadian macro economics, bonds, and bank equities are all products that are well-covered by analysts. This doesn't mean you cant get traction with this idea, but it means you will need to do more to stand out. Martinghoul's suggestion about pay-out scenarios is a good idea, but I also think you need something more unique on the story...the charts on household debt are great but I see them every week in XYZ Bank's Canadian Economics Weekly. Off the top of my head, maybe you could take the US scenario from 2007-09, see where financing costs started to cause defaults, plug in the Canadian forwards or your own guess as to the BOC hiking cycle (if it ever occurs), and try to spit out a date for the problem to come to a head. Try to throw out an "end of the world" date...this strategy will make your story flashier and besides you can always push back the date every good cult leader knows that. Also banks do not do that sort of thing so your presentation will by definition be unique. From there you could go get pay-out ratios for CDS and the like as martinghoul suggests. End product is "if my assumptions on rates are correct, we will have a crisis on XYZ date, and XYZ security will pay out 100-1. Buy the security and hire me to manage the trade." . If you do get hired, immediately begin the process of extricating yourself from this prediction and tying your employment to something other then a wild as$ guess about Canadian housing prices.

If you are writing this because you actually want to trade this collapse you foresee profitably...well that is an entirely different topic. Loading up on negative carry lotto tickets anticipating a rate rise wouldn't be my choice on how to play this at this point. Maybe do a bit of that stuff but don't have enough premium out to mentally crush you or kill your trading account if the story doesn't happen in the next few years. If 2005-2009 taught me anything it is that preserving capital (mental and physical) is the key to profiting from a financial crisis and getting caught up in the story this early is not the way to do that. FWIW the first guy who talked to me about subprime way back in 2005, who knew the story better then everyone including the guys in the Big Short, is now in the real estate business....being early is just as bad as being wrong.

    • 1
Nov 29, 2014
Bondarb:

FWIW the first guy who talked to me about subprime way back in 2005, who knew the story better then everyone including the guys in the Big Short, is now in the real estate business....being early is just as bad as being wrong.

Oh man that has gotta suck *so* hard.

Is it even possible to time these things properly, or is the difference between your friend and John Paulson luck?

Nov 29, 2014

@"Bondarb"

Great feedback. I agree absolutely. Often selling a narrative trumps selling an idea. And yes, my goal is to be hired by a hedge fund.

I attacked this issue from the income side. I determined the mortgage rate required to exhaust personal disposable income after non-discretionary expenditures. My analysis shows that an 8.8 percent or higher mortgage rate will spur widespread defaults in the high LTV mortgage segment (27% of the Canadian mortgage market). This assumes household income/expenses remains constant going forward. The average mortgage rate has tracked at roughly the BOC rate + 300bps. Thus, the BOC to would have to raise its rate from 1% to 5% unless mortgage rates increase independent on the BOC rate. Again, keep in mind that the 8.8% mortgage rate is the target required for the entire high LTV segment to default. Mortgage defaults will likely increase with each incremental hike in mortgage rates.

As you stated, how do we determine a date for when rates will hit this target? I agree that putting a date on the outcome will allow me to differentiate my report and develop a following. However, doing so will strip my analysis of its nuance and the inherent uncertainty of predicting a future financial event. While I agree with you, I am very hesitant to pursue this strategy as it simply feels intellectually dishonest.

Nov 29, 2014

it's probably better to wait for something more significant to start happening before putting on a trade. many people have lost money betting that interest rates were going to rise.

i think foreign investors in canadian real estate (chinese) are an important piece of the story

Nov 29, 2014

Interesting read. There was an article a year ago in Bloomberg about housing in Sweden, how inflated prices have become over the past few years, more and more IO loans are used to buy homes etc. Even simple one bedroom appartments in Stockholm (30-40sq.m) can reach prices as high as US$ 600k to 700k if I recall correctly. Same thing is happening in Luxembourg but that's different animal due to its tax-haven like characteristics and how tiny it is. Only places I can see an increase in prices happening is Greece, Cyprus, Portugal and Ireland. But these are special situations. Prices are depressed and for good reasons. If you look hard enough you can buy with 50k-60k USD a 60-80sq.m appartment in an upscale neighbourhood in Athens.

Colourful TV, colourless Life.

Nov 29, 2014

I think the main thing you're failing to address is, where are you going to find someone to take the other side of that trade??? I mean your analysis pretty much assumes that there's a whole pool of retards waiting to take on the credit risk of the Canadian High LTV mortgage market.

Do you even know the approximate price of a CDS on a nominal value of 100 million? No? Then your analysis is dog shit ;useless.

Anyone that's going to even consider that trade would have already looked at the same facts that you're presenting. The US housing crisis is still fresh in everyone's head.There's absolutely nothing novel about you're thesis, people have been looking at the numbers that your presenting for a while now.

I love how you repeatedly invoke the word "theoretically" to justify your thesis. "Theoretically" a person should be willing to pay every penny they have to play the St. Petersburg lottery, but that doesn't make it a good trade.

Please don't waste your time trying to disseminate this research. Anyone that would be impressed by it isn't worth their salt.

    • 1
Nov 29, 2014

Could we convince prime brokers and sales product structuring folks to write CAD high LTV CDS? That's how the bros before the crisis did it. Just gotta convince AIG to buy them again.

#Abacus 2.0

Nov 29, 2014
jcpenny:

Could we convince prime brokers and sales product structuring folks to write CAD high LTV CDS? That's how the bros before the crisis did it. Just gotta convince AIG to buy them again.

#Abacus 2.0

Without liquid underlying, the banks will not be willing to structure a trade where they offload risk or synthetically take the other side...

Nov 29, 2014

Then it's up to the folks on this forum to convince people to trade canada......

Nov 29, 2014

If there is no way to facilitate a CDS on High LTV mortgages then the risk reward profile is no where near that of Paulson's trade. Also, there should be analysis such as to the specific regions with the highest LTV's, lowest DSCR ratios, highest household debt, etc. This would allow you to specifically point out the areas/cities within Canada that would have the highest default rates.

You need to do a better job selling the trade. Think about it from an investor's perspective. The economic analysis is sound but what does that translate into? What are the specific returns for each type of trade under x% default in y1, y2,y3? I understand it is difficult to model.

Nov 29, 2014

Random thought. Think about how heavily low oil prices affect the local economy of a city like Calgary, which has been one of the hottest real estate markets in Canada.

Nov 30, 2014

There's no CDS market for the Canadian real estate market, let alone specifically high LTV Canadian mortgages. This investment opportunity doesn't exist.

Nov 30, 2014

a few points the Canadian mortgage market that makes it safer:
1. In Canada, homeowners are personally liable for the mortgages they default on
2. In Canada, interest on mortgage payments are not deductible for personal income
3. The Canadian secularization market is not liquid at all, so the risk is not spread throughout the financial system

Nov 30, 2014

I have researched the CDS market further. According to a report by the BOC, CDS is available on Canadian-based entities.

See below for an excerpt:

"Quotations for CDSs are available for as many as 160 Canadian-based reference entities. Trading activity among these 160 names can be broken down into three tiers. The top tier includes five to ten names that are extremely liquid and in which there is a regular two-sided market. Approximately 20 additional Canadian names trade on a semi-regular basis. The bid/ask spreads of the first and second tiers are typically around 5 basis points (however, this may be indicative only for small volumes). The liquidity of the remaining 130 Canadian-based entities, or the third tier, is essentially nil, with any trade in these names being difficult to find. Approximately 2,100 reference obligations trade globally (Fitch Ratings 2004b); therefore, CDSs written on Canadian-based entities represent only a very small fraction of the global market."

WSO still won't let me post links as a new user. If you want the entire report head over to my user account on slideshare (ndsouza22).

What's really strange here is that CDS positions are held by pension/insurance funds but not by any hedge funds. Banks did not have any CDS positions with hedge funds as of the report date (2004). Why would CDS be available for one segment of the buy side but not the other? Anyone have an answer for this question? Is it simply because no one in the hedge fund space is seeking to buy CDS on Canadian entities?

    • 1
Nov 30, 2014

Double post

Nov 30, 2014

because shit is not liquid enough for hedge funds

Nov 30, 2014
whatwhatwhat:

because shit is not liquid enough for hedge funds

Liquidity was an issue initially with the subprime CDS trade as well. It wasn't until CDS became standardized that liquidity was improved marginally. Even with standardization, brokers had issues with price discovery and CDS positions were not properly marked to market when the ABX (subprime) index crashed

Here's my point though. Liquidity is an issue if you want to roll forward your CDS positions or sell it on the open market. However, it's not an issue if you are holding the CDS until expiry with the intention of receiving a payout from the writer on default.

If a hedge fund is convinced that the Canadian MBS market is likely to face increasing defaults, why not purchase illiquid CDS swaps? Even its priced at 500 bps (twice of subprime in 2005) that's a 5 million negative carry to insurance 100 million in MBS/mortgages

Is the liquidity concern of hedge funds that much greater than pension/insurance funds to not get involved in a single CDS position??

Nov 30, 2014

dude you just came out and said "oh btw i actually read up on canadian cds after i wrote my tl;dr presentation" and now you're trying to lecture me on things? the market is not liquid enough and banks are not going to get into this shit right now. corporate CDS right now is borderline illiquid, a randomly newly created market is not going to exhibit higher liquidity for any reason.

your second paragraph about liquidity only being an issue under x y z comes from a completely naive perspective. liquidity is probably just as important as pricing nowadays and just hoping that shit defaults at the right rate to pay you out without having a backup plan is crazy. your second paragraph btw explains why pension/insurance funds are primary holders of canadian cds

Is the liquidity concern of hedge funds that much greater than pension/insurance funds to not get involved in a single CDS position??

liquidity is absolutely a concern. there is no liquidity in this market because it does not exist and banks will not get involved because it's not worth it for them from a regulatory, capital, or even plain old "hey this is not a bet i want to take" perspective.

    • 1
Nov 30, 2014

@whatwhatwhat

The entire trade depends on pricing of the CDS contracts. The illiquidity of the CDS market would have to result in a price where the asymmetry of the trade does not justify the risk of uncertainty

For example, assuming an illiquid market, let's assume specialized private OTC CDS can be obtained at 500 bps. Covering $1 billion in Canadian MBS would cost you $50 million per year. If you're managing a $1 billion plus long short portfolio, why wouldn't you lay out $50 million a year (5% or less of your AUM) and overcome the performance drag through your long positions.

Here's the real question. The market is not non-existent. According to the BOC report there is CDS on 160 Canadian-based reference entities with the top-tier being highly liquid (in relation to the entire CDS market). Canadian CDS exists but its held by insurance/pension funds not hedge funds. Why is this the case?

I am not sure what paragraph you are referring to which answers this question.

Nov 30, 2014

why wouldn't you pay out 5% on something you have 0 idea on a timeline for a catalyst and just hope longs make up for it? come on man.

if the market is not non-existent, please show me a contract on canadian mortgages. i'll wait. some random report saying canadian cds is highly liquid does not make canadian cds as an aggregate liquid. how much do these names turnover on a monthly basis?

i was saying you answered your own question re: pension and insurance funds when you were talking about rolling contracts. buy and hold investors will buy cds to hedge their bonds that they plan on holding to maturity. liquidity is not a primary concern for them. liquidity is a concern for hedge funds for a million and one reasons, i don't have the energy to go into it.

    • 1
Nov 30, 2014

@whatwhatwhat

Here's the math:

Assume you insure $1 billion of MBS at 5% when your analysis predicts a 25% default rate with a three year horizon. This results in a $50 million per year cost outlay ($150 million total) with an expected payout of $250 million. So your gain is $100 million when correct and your loss is $150 million when incorrect. Therefore you would require a probability of success greater than 60% to implement this trade. 60% x $100M = $60M and 40% x $250M = $60M. This math changes depending on your expected default rate and the cost of the CDS. You can also view the trade as a hedge. If you expect the default rate to hit your target if interest rates rise. You can implement the CDS as a hedge for your long positions.

The report is by the Bank of Canada. It's not a random report. The BOC states that out of the 160 CDS names, the top 10 are highly liquid, the top 20-30 are liquid and trade with a semi-regular frequency, and the bottom 130 are highly illiquid. References are listed on the report if you want to dig deeper.

Dec 1, 2014

So this is, at best, a 1:1 trade?

It's interesting how your logic seems to have evolved. Recall that initially your trade was supposed to be a cheap "lottery ticket", which would be eminently suitable for a HF manager who is looking for Paulson-style glory. Now you're talking about it in the context of offering protection for a real money mtge portfolio.

Problem is that your thesis doesn't really sound all that attractive for either of the two groups of investors. For the former (HF fast-money types), it's just not cheap enough, given the uncertainties arnd timing and the inability to structure the trade in a compelling way. For the latter, it's not going to tell them anything they don't know already about a small fraction of their portfolios.

In general, there's been a bit of a bubble in housing bubble talk recently. People appear so inspired by Paulson's experience that they're seeing housing bubbles everywhere and, most importantly, acting on it. For instance, the other day I read a story about a hedge fund type betting against Danish mtges (Owl Creek is the fund, IIRC). In my view, a lot of these analyses are superficial and, driven by their desire for glory, miss a lot of the relevant facts. FWIW and with all the usual caveats, the performance of the fund in question has been rather poor this year.

    • 1
Dec 1, 2014

@matinghoul

Not at all. I used those numbes merely to illustrate an example. The trade is entirely dependent on pricing pf and ability to obtain CDS on high LTV mortgages. I assumed a CDS price of 150 bps which is reasonable given that subprime was priced between 150-250 bps. Second, my analysis shows that high LTV mortgages will default en-mass with mortgage rates at or above 8.8 percent. Let's say your analysis determines a time horizon of five years to hit this rate target. Assuming $1 billion in covered high LTV mortages; you pay $15 million per year with the expectation of a 50% default rate once your target mortgage rate is hit. This results in a CDS payout of $500 million at a cost of $75 million over 5 years; a return of approximately 7:1 if defaults occur in the 5th year. The return ratio improves substantially if the default rate increases earlier then your expected time horizon.

As others pointed out, the primary hurdle of this trade is obtaining CDS. According to the BOC report, CDS Is available on Canadian-referenced entities with a top-tier of entities exhibiting high liquidity. If CDS is not available then you will have to implement this trade through other means i.e. shorting mortgage lenders or shorting CAD

Citing or calling for housing bubbles is fine if your analysis and evidence is sound. If the data supports the observation then its simply a matter of timing.

    • 1
Dec 1, 2014

The BoC report is probably talking about simple single-name CDS, i.e. your vanilla bread and butter derivative of the corporate credit world (even these things are getting difficult these days). You, on the other hand, are referring to a CDS on some equivalent of a bespoke subprime CDO. Given that this mkt in Canada is mostly nonexistent and definitely illiquid (for instance, there are no indices like ABX etc), you just won't be able to express this trade in an efficient way. I think that's what the others have been telling you.

Well, there's all sorts of data that seems to point every which way, innit? For instance, I don't recall seeing any data on the household assets, incl pensions, in your slides. What if CAD households are just purchasing houses/taking on debt, because they're generally sitting on tons of savings (that's one of the things that is commonly missing from the analysis of housing bubbles)?

Finally, pls take my word for it: matters of timing in this biz are NEVER EVER simple. Survivorship bias really obscures just how crucial timing is and the media does a terrible job at telling the whole story.

    • 1
Dec 1, 2014

@martinghoul

Look, I am not disagreeing with you. I agree that the CDS trade on high LTV mortgages is not viable without a market being created for it just as it was created for subprime. If the CDS trade is ruled unviable then you'll have to look at other trades such as shorting mortgage lenders or shorting CAD.

I am also not disagreeing with you in regards to timing. Timing is a HUGE hurdle to shorting a credit bubble. For every Paulson there are dozens of unknown investors who lost a lot of capital by being too early or due to inefficient trade implementation.

As for your question on assets. Although household debt to income at a record high, debt as a function of net worth and assets is relatively unchanged due to rising house prices. Real Estate has driven the average Canadian's increase in assets and net worth. As for savings, I can pull the data. Given the increase in debt servicing expenses and overall outstanding debt, we should see a decrease in savings as a percentage of disposable income.

Dec 1, 2014

Cool, as long as you're aware, it's all good...

And yes, you're correct... You need to look at household mortgage debt vs assets ex housing. It's important to include pension assets, though.

    • 1
Dec 1, 2014

@martinghoul

Non-financial assets (i.e. real estate) as a percentage of total assets is also at a record high along with real estate prices and household debt to income. Inversely, as a proportion of total assets, financial assets (i.e. stocks & pension assets) are low. Nominally, however, pension asset value is likely higher now than it was 10 years ago.

But, keep in mind, that real estate corrections tend to coincide with a decline in economic growth and a correction in equities. Research shows that recessions that coincide with real estate corrections are three times worse than the average recession. Basically, if housing prices correct, those pension assets wont be able to counteract the hit to the average Canadian's asset base.

Dec 7, 2014

-

Jan 29, 2016

Bump

Feb 16, 2016
Feb 16, 2016
Feb 26, 2017