Published on October 22, 2016
1. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 1 Hedging Jumbo 30-Year Fixed Rate Mortgages Evaluating the Volatile Period from 12/1/2015- 3/11/2016… And explaining the hedge that worked. Hedging any origination that is priced off a specific hedge is quite easy. Take for example, the process of pricing Agency or Ginnie loans off TBA’s. Best case, the loan is sold better. Worst case, deliver into the TBA. Pricing to the hedge and hedging the price—now that’s a tautological process that works. Jumbo 30-year loans are, no doubt, a bit trickier. Pricing seems, at times, ephemeral and ethereal and, as is said up and down the 405 – “whacked”. In the post-securitization era, pricing may be “follow-the-leader.” Nonetheless, lending is still a business of creating net interest margin. Without securitization and even without continuous take-outs, there exists some amount of spread that motivates lenders. “The spread” is the focus of this piece: isolating and locking “the spread” down. First, and not an easy request, there has to be Jumbo pricing data. In the absence of private label securitization and in the absence of frequently-traded and price-reported whole-loan transactions, there are few reasonable pricing sources. In the last six months, find-the-price is even more difficult (with expanded lending criteria utilizing bank/business statements and increasing, eh, “courage”) because LTVs are creeping up as are loan values. “What’s the rate on a jumbo 30-year?” is not as straightforward today as it was just 6-9 months ago. From Dec 2015 to Mar 2016, there was less noise in loan specifics. There also was a banging gong with spreads lurching out and rates bounding around. Generic Jumbo 30-year pricing is available through that stretch using loan rates compiled by BankRate.com™ on Bloomberg™:
2. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 2 To construct the hedge, the immediate requirement is to define the contributing components of the Jumbo’s risk. Like all mortgages, the major risk components are: Interest Rate Risk Credit Volatility And, perhaps, Home Price Appreciation/Depreciation (for long term loan-owners) All four have exchange-traded solutions—though the Case/Schiller HPA futures are not active (market-makers noticed we all want to go the same way at the same time on those). Interest Rate Risk There are two pieces to the rate puzzle: (a) short-term funding cost and, (b) full duration risk. For simplicity, call (a) 3-mo LIBOR and, (b) 7yr bullet Swap. (Because there is enough ambiguity in Jumbo pricing—let alone speed assumptions—we’ll stick to a bullet vs amortizing swap). Credit For a quantifiable and trade-able credit proxy, generic 5yr Investment Grade (IG) is the solution. (For anyone with deeper interest in bifurcated credit exposure, there is an amazing staff research piece that was updated by FRBNY about a year-and-a-half ago, I strongly suggest this: https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr674.pdf AND it offers absolute confirmation that generic corporate IG credit influences all-- even Agency-- mortgage spreads. A must-read). Volatility Did mortgage spreads widen because implied volatility increased OR did credit spreads widen, forcing mortgage spreads wider, because of volatility—or something else? Right- forensically, this gets difficult to isolate for data-light Jumbos. Implied volatility plays a role in everything from path-dependent models to general liquidity. Independently, these variables have “OK” to “Negative” correlation with Jumbo RATES (as sourced from BankRate.com via Bloomberg). In AGGREGATE, the story is quite a bit better: Intuitively, these numbers make sense and bear in mind the BankRate.com-reported Jumbo rates are completely generic—no LTV, no FICO, no Geographical info. A strong correlation to a generic equivalent-duration 7yr Swap, a lesser but meaningful relationship to short-term funding costs and generic investment-grade credit exists. And then there’s implied volatility—most of the time
3. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 3 it’s not a factor; except when markets experience unforeseen turbulence and everything that is not a U.S. Treasury gets geometrically punished. During the volatile period of Dec 2015 – Mar2016, the above amalgam of futures contracts proved to effectively correlate to Jumbo origination yields. To achieve the correlations (above graphic), percentage contribution to jumbo rate, breaks down into: Correlations are great indicators, but how about P&L?- real dollars? Defining a reasonable Jumbo 30-year fixed-rate proxy:
4. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 4 The assumption will have the first five years paying at an 8 CPR and a jump to 15 CPR thereafter. Given a hypothetical price of 103.5 and 4% Net / 4.5% Gross (on-market at the time), the calculated values produce a 5.79 year duration and a $600 per $mm rate sensitivity (dv’01). Note- the prepay speeds could be set at any market-clearing level and the hedge adjusted accordingly. Using a static dv’01 of $600/mm on $100mm hypothetical, the DAILY P&L swings are: Not shockingly, the performance of the JUMBO that corresponds to the BankRate.com yield levels is skittish—especially during Jan-Feb’16 when markets were volatile. “Hey, stocks are getting crushed! - raise those rates… We’re not doing any business, get those rates down!... Where is XYZ’s rate?—get on top of it.” How volatile were Jumbo rates?
5. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 5 The point of the exercise is to answer the question: “Can Jumbo rates be reasonably hedged using a basket of futures contracts? Will the Jumbo:Hedge relationship hold up in exceptionally volatile periods?” Consider the cumulative P&L over the 12/1/15-3/11/16 date range: Without making a SINGLE ADJUSTMENT to quantities of the hedge instruments, nor making any “roll”—literally changing nothing from 12/1/15-3/11/16—the above graph tells the story. The real “problem” is that this hedge is not a simple construct. “Just sell one of those and everything will be OK” is absolutely not going to work. Heck, the futures contracts it takes to create this hedge strategy are cleared on three different EXCHANGES. But, there is nothing really surprising in the constituent parts of the hedge: Rate and bulk duration exposure (Eris Swap Future) Credit spread sensitivity (ICE-Eris IG Credit Future) Cost of Funds sensitivity (CME 3mo LIBOR Eurodollar Future) Implied Volatility exposure in extremes (CBOE VIX Implied Volatility Future) NO DOUBT—THIS IS AN ILLUSTRATION FROM VERY GENERIC JUMBO PRICING. Nonetheless, if you originate, trade or carry Jumbo Fixed, the strategy should be customizable to your loan specifics. Give us a call—let’s take a look versus your price data. There are many newer instruments trading in futures on multiple exchanges. We’re only beginning to bind these products together to find new solution sets to old problems. Stay tuned. JC—for the Fixed Income Group at R.J. O’Brien www.fixedincomegroup.com firstname.lastname@example.org 800-367-3349
6. The Fixed Income Group at RJO: 800-367-3349 ©Copyright 2016 The Fixed Income Group at RJO www.fixedincomegroup.com 6