Category: Blog

Basics – CUS Calculation

It might be useful to walk through the calculation of the CUS (The 10-city Nationwide) index once, as the index presents an opportunity for possible trading strategies.  To start, the table (below) lists the weights of the index and shows the weighted average price of the 10-cities for the May CS release.  (Weights are from S&P report on Case-Shiller index archived here under reports tab.)  Thankfully, the weighted average for this exercise equals the published index result.

 With these weights in mind, one can attempt to trade forward values of the index or the key underlying cities in an attempt to take advantage of discrepancies, trends or relative liquidity.  For example note that two cities (NY and LAX) make up more than 50% of the index (weighted by price index).  Changes in those two cities alone may dominate the value of the CUS -particularly if others (say Denver) remain relatively quiet. 

Recently NY and LA have been moving in opposite directions (NY one of two cities that fell in value last month) while LA, and the West Coast cities of SF and SD that in total make up 40% of the index by weighted average, have been the stars.   Those offsetting effects have somewhat dampened price moves in the CUS.  As such there may be trading opportunities if the bullishness of certain markets spills 1:1 into the CUS index while NY lags.  However should California and NY ever begin to move strongly together (up or down) look for the CUS to become more volatile.

Basics – Home Price Indices

There have been some questions about  – “Why use the Case Shiller index”, and “How does it differ from other indices”.  Deutche Bank summarized the differences between the Case Shiller (CS), National Association of Realtors (NAR), and OFEHO indices in the attached writeup from March 2009.   Please contact your DB coverage for copies of the full report.

Definitions of three widely used home price indices

The S&P/Case-Shiller Home Price Index (HPI) measures the prices of repeated paired sales of single-family homes in the U.S., excluding refinancings, home improvement and investment purchases.  The data underlying the index is retrieved from house deeds recorded at county courts.  The S&P/Case-Shiller national HPI is reported quarterly, with a tw0-month lag.  At the MSA level, S&P Case-Shiller provides data for 20 MSAs, which is available monthlywith a two-month lag.  (This nationwide data is available for free for both national and MSA data series at http://www2.standardandpoors.com).  For expanded MSA coverage, additional data can also be purchased from Economy.com (http://www.economy.com/home/products/housepriceforecasts.asp)

The OFHEO House Price Index (HPI) measures paired repeated “sale prices” of single-family homes in the U.S., but also includes refinancings. For a refinancing, the value of the home is based on an appraised value, not an actual buy-sell transaction. Because appraisals can lag the actual market, including appraised values in a home price index along with actual selling prices may create an upward bias when prices are falling. The data underlying the index also covers only conventional conforming mortgages provided by Fannie Mae and Freddie Mac.  By excluding jumbo mortgages, the index may also understate deterioration, as (traditionally) higher-priced homes are more volatile. The OFHEO HPI is reported quarterly, with a two-month lag. (Data is available for free at http://www.ofheo.gov/HPI.asp.)

The National Association of Realtors (NAR) House Price Index (HPI) measures the median sale prices for existing single-family homes sold in the U.S. The data underlying the index is retrieved from local associations/boards and multiple listing services nationwide. The NAR HPI is reported quarterly at the MSA level, with a two and a half month lag. The data is alsoavailable monthly at the national and regional level (i.e. Northeast, Midwest, South and West), with a one-month lag. (Data is available for free at http://www.realtor.org/research.nsf/pages/ehspage.) While median data has its limitations, this source is also the most timely.

 

Basics -Front Contract Expiration

A good way to learn how CME Home Price futures contracts work is to watch as the front contracts reach expiration.  Recall that the contracts have settlements in the current year at 3-month intervals (Feb,  May, Aug, Nov) and that the contracts “Cash-Settle” – that is the value used to settle an open position is the index number published on the last Tuesday of that month.

The May 2010 contract is approaching expiration as the S&P Case-Shiller index results for March 2010 will be released on May 25th.  Since any open interest will be cash settled (absent a closing trade) on the numbers released on May 25th, bids and offers tend to converge toward the expected index numbers.  That tendancy has the effect of narrowing the bid/ask spread.  As of today, May 13th, the bid/asked spreads in the May ’10 contracts range from 1.20 to 4.60 points (or 120 to 460 as quoted by the CME).    Recall that the CME contracts trade in 20 cent intervals and that 100 points (or 1.00 on the CS index) is worth $250.

The New York contract has the distinction of being the only contract where the current offer is lower than the prior month’s index.  (There are no higer bids than last month’s CS index). 

 

 

 

 

 

 

 

 

 

Any long/short position that needs to be rolled forward could do so either by legging the two sides of a trade (e.g. buy/lift/cover May ’10 short while trying to set a new short position in a more distant contract).  Given the illiquidity, a better approach (if one has to roll) is to enter orders in the calendar spread markets.  That is, one could place an order to execute a Buy May ’10/ Sell Aug ’10 at a certain price spread, but only if both orders were executed at the same time.  Again, while the calendar spread execution is better than legging each side of the trade, the bid/asked spreads have tended to be wide.  (I’m working on that next and will respond to inquiries.)

Finally, while contract expiration may be interesting, and there may be opportunities for trading (and thereby providing liquidity) at the end of the day most core real estate investors and hedgers are looking much into more distant contracts 18-36 months out to express their views. 

 

 

Basics -Geographic exposure

The question has frequently been asked – what specific geographic exposure am I getting when I buy the New York,  Miami,  San Fran, or any other contracts. 

Page 8 of the S&P/Case-Shiller Home Price Indices Methodology report (found in the Reports section here) lists the areas covered, by county, for each regional index.   In reviewing the lists, one can see that these are not strictly “City” indices in that home prices some distance from each city’s downtwon area are included.   (Anyone intending to trade any product referencing the S&P Case-Shiller index shoud read the complete report to understand not only the reference geographic areas, but all other aspects of the index calculations.)

In addition I’ve attached here a map from MacroMarkets (Robert Shiller’s firm) that highlights the span of the New York index, as an example.  A New Yorker may appreciate that home prices in an  area that stetches from New Haven, Ct. to Duchess County, north of NYC,  down to Ocean County in NJ on the shore, may not all move in unison with the rest of the regional index area.  There’s a tradeoff  in that the larger the reference area, the greater the number of possible repeat-sales, and the more current, robust and accurate the index.  The advantage of fewer, larger regions, is that you not only elimate nuances of a particular town, but you avoid fracturing trading into an excessive number of regional contracts, each of which would have much less liquidity. 

While somewhat dispersed, the NYM index performance has been genearlly reflective of broad changes in local real estate repeat sales, for many areas, as much of the area’s home price movements may be tied to the fortunes of the New York City economy.  (As an exercise map what you think your own home has been worth at various point in time over the last 10 years and see if it’s tracked the index.)

Does a large regional index introduce some basis risk between a house in one town versus the larger region?  Of course, but many indices and commodity contracts (such as the CME Home Price futures contracts) have that basis risk in either how reference obligations are defined, or in the seller’s ability to deliver different assets.  It’s just the trade-off that’s made to create an index (or a contract) that can be viable for hedging regional systemic risk.

The same could be argued for each of the other 9 regional indices.  Please refer to the MacroMarket website http://www.macromarkets.com/real-estate/ for maps of the those regions.

 

 

 

 

Basics -Calendar Spreads

If you look at the CME Website for prices (see link) you’ll see a pulldown menu for quotes for either calendar spreads or inter-city spreads across all 11 indices.

 http://datasuite.cmegroup.com/dataSuite.htmltemplate=hsng&leadMonth=NYMK0&strategyType=SP&category=Housing&exchange=XCME&selectedProduct=NYM&selected_tab=real_estate_tab 

The calendar spreads are useful for expressing views about the timing and magnitude of price changes. For example, if one thinks that prices will be higher in Nov. 2013 than Nov. 2011, one could sell the spread (sell the front contract and buy the back contract) at “flat” (the contracts at the same price) or for some positive spread (spreads are always quoted in terms of the front contract relative to the back contract so a positive spread implies that the front contract is trading at a higher price). If between now and then the market starts to price in a rise in the index between Nov. ‘11 and Nov ‘13 the spread could go negative -resulting in a profit.   

                          
One caution on calendar spreads – at expiration the front contract will cash settle leaving you holding just the forward postion.          

                                          
Another note, some calendar quotes are just computer generated combinations of a correseponding bid and offer. For example if Aug ‘10 is bid 17000 and Nov ‘10 is offered at 17240, a spread offer in Aug – Nov. ‘10 of -240 will likely be showing. If executed, a computer program will “lift” the offer at 17240 and “hit” the bid at 17000, backing into the -240 spread.          

                                                                                                                                    
Finally, calendar spreads can be entered as GTC while inter-city spreads must be entered (and renewed, if desired) as day orders.

Basics – Closing prices

How does the CME post closing prices on 11 home price contracts (the 10-city index and the 10 individual regional prices) over 10-11 expirations so quickly after the close?  They have a rules-based methodology that they follow, that sometimes generates some odd-looking results.

One of the challenges of a thinly traded futures contract is that closing prices, particularly on the more distant contracts,  may appear “out of whack”.  For example, the 25-point  higher closing price (relative to other expirations) for the San Diego Nov. 2012 contract (as of 4/30/2010) may not be due to any fundamental reason.  The CME is just following their methodology of calculating the “closing price” based on either the: 1) last trade, or 2) a subsequent higher bid, or lower offer.   Higher bids in the May 2012 contract or the Nov. 2013 contract, or lower offers on the Nov. 2012 contract would “smooth out” this closing curve, but as there is no open interest in those expirations, quite a bit of time may have elapsed since the closing price changed.

(Qualifier: I’m not saying that May 2012 is “too low” or Nov. 2012 is “too high”.  I’m just highlighting the differences between the two to explain the CME methodology to posting closing prices.)

The kinks in closing prices (such as those in the displayed San Diego contract) become more pronounced when the market has trended in one direction, but where trading (or quoting) has not kept pace with market moves.  More frequent trading and/or higher bids/lower offers will more likely keep closing prices closer to “reality” and eliminate the kink we see in this graph.

What are CME Home Price futures telling us? (And what they are not)

While the monthly release of the S&P Case-Shiller(CS) Home Price indices always results in a tremendous amount of number crunching of what has happened (with a two month delay), there is often little quantitative review of where “the market” says home prices
are going to be over the next three months or even three years. Sure there are talking heads who note that inventory levels are high, housing is more affordable, and the home prices might rise or fall if the GSEs lend (or not), or if modification programs are implemented (or not), but none if this information is pulled together into a collective market view on what “the market” thinks home prices will be. 

Such information –a set of market based forward home prices –might be tremendously useful to investors, model builders, regulators, policy makers, rating agencies, potential home buyers, RMBS traders, lenders, mortgage servicing companies, and entities that guarantee such exposures.

Such information, if traded in a public fashion, would be a much more direct inference on home
price expectations that builder and bank stocks, or losses implied by the level of ABX trading.

I would submit that such a forum m already exists -the CME Home Price futures mmarket. (See
http://www.cmegroup.com/trading/real-estate/index.html for more information.) *

March Madness and Credit Default Swaps (CDS)

My 13-year old son Jack asked me today if betting money on the NCAA basketball pool was dangerous. As a parent –and a geeky one – and knowing that every question is a potential teaching moment – I launched into an explanation of the various betting structures that I had encountered in my career on Wall Street. I covered the strategies of filling out brackets and the joys and angst of draws in a lottery format.

But knowing that I had a captive audience –we were in the car – I steered the conversation into the most dangerous element of NCAA pool betting –the trading of individual teams. I explained that while a $10 draw in a $640 lottery is exciting as the team names get pulled, that traders quickly got bored. They wanted “action” to demonstrate their analytical skills as the tournament progressed and they took to offering to buy teams. I told him that how in the mid-1980’s people would pounce on the secretary who had pulled the hot ticket and say “Oh you got Duke, I’ll buy it from you for $64!”, or to others “Oh you got Gonzaga, I wouldn’t give you $4”. That quickly lead to a discussion of odds –how Duke was valued at 10:1 and Gonzaga at less than 100:1.
But traders quickly learned that a Duke long would not surrender her ticket for any reasonable price, and while the Gonzaga fans were loyal, that they already had $10 invested in their one ticket. At that point, a guy named Charlie (a trading assistant with the mindset of a future CDS trader) wandered over and explained “Oh you don’t have to own the ticket to sell it. You can sell it synthetically. Watch. Who wants to buy Gonzaga at $3?”

I asked him if he was worried that if Gonzaga won, he’d have to pay out $640. He said, “No, he’d followed the tournament for years and it was a statistical near-certainty that a 12 wouldn’t make it past the quarter-finals” as he sold three more units. “Free money” he noted. “I’ll just keep selling any of the lower seeds. I’ve made money doing this the last three years.”

My son asked me – “Wasn’t there a limit to how many teams this guy could sell”. I told him “no” and in fact Charlie got so involved in every bet that he became an office star and started crossing sellers of teams from his friends outside the office and to guys on our desk. He’d had friends who’d sell him 10 Gonzagas at $2.75 and then he’d resell them to buyers in our office at $3.25 netting $5.00 on a 10-lot.

With money involved my son was all ears, and he asked “So, did the guy get rich”?

I explained that as the tournament progressed, Gonzaga kept winning. “Well couldn’t Charlie buy Gonzaga back”, Jack asked. I said that there were some offers first at $10 and then $25, but that Charlie would have had to take a loss, he wasn’t keen to do so, and no one could make him buy.
Gonzaga kept winning and by this time people were willing to pay $60 just to get involved in this Cinderella success story. One morning, after a frantic midnight overtime win by Gonzaga, Charlie wasn’t to be found anywhere in the office. A round of “poor Charlie’s” over breakfast coffee lead to a whispers of “how much will he owe you if Gonzaga wins” by morning break. With growing questions as to who his outside sellers were, a concern grew into a panic by lunch as the tally of Charlie’s potential IOU’s grew from $50,000 to $100,000 to numbers that would have taken Charlie a lifetime to re-pay.No one knew how much Charlie had bet, or whether he, or his friends were “good for it”.
Some who had bought Gonzaga at $3 and sold at $10 feared that they might not only not get their $7 winnings, but that they might have to make good on the pool (maybe even to their boss ouch!) even if Charlie disappeared. Some senior traders insisted that they be paid the $7 profits upfront, but wouldn’t advance on any losses. Many of the Gonzaga longs were counting on their gains to offset other losses. They scrambled to do trades at crazy levels just to unwind exposures.

I explained to Jack that senior management had to step in and close down all betting as the office was totally focused on the “pool” to the point where no one was working. It was rumored that one VERY senior trader was pressed into taking over Charlie’s positions. We all actually breathed a sigh of relief when Gonzaga finally lost.

My son noted -”That’s messed up. They wouldn’t let that happen today, right?”

I smiled silently as we drove the last mile thinking (remember I’m a geek), that if my 13-year old son could grasp these issues, that maybe some Congressman could see all of the parallels to CDS: the lack of margin, the problems of price discovery, the absence of netting, the flaw of short-sighted models, the issues of counterparty risk, the daisy chain of IOUs, the fragility of the system to tail risk, and the prayer that someone larger than the market will be there to rescue it in a crises. OK, maybe Jack didn’t get all of those messages, but he now knows the concept that someone can bet beyond their capital limit, and that unlikely events can happen with consequences to those even not directly involved.

So in an effort to advance the discussion of centralized clearing of CDS along, I cheer – GO BIG RED (Cornell)! If AIG didn’t work, maybe we need another (NCAA) credit scare to teach this generation the risks of OTC, un-margined, synthetic trades.

P.S. I read years later that Charlie became a very successful hedge funder manager…..for a while.

Home Price Futures – Are we there yet?

(This was part of my inaugural effort in Jan. 2009 to comment on the CME Home Price futures contracts.  I’ve reformatted the tables but my observations are the original ones.)

Jan 9, 2009 – With home prices at the center of our economic problems, I’m surprised that there haven’t been more quantitative discussions about home price forecasts.  Most investors, like the kid in the back seat on a long car ride have two great concerns that dominate their outlook:

1)      “Are we there yet?” (Have home prices hit bottom?)

2)      “How long ‘til we get there?” (When will they bottom?)

While the “we’ll get there when we get there” may suffice for a toddler, it’s a challenge to run a mortgage investment or servicing platform with that degree of uncertainty.

While many forecasters predict generally lower home prices of 10-25% there has been very few direct ways to react to or disagree with those estimates.  Many investments (equity investments in banks, purchases of mortgage-backed debt) hinge on an implicit view of home prices over the next few years (and in theory should be hedged against getting that home price forecast wrong).  That is why it surprises me that more people are not trying to express their divergent views (trade) in the CME’s CS Home Price Index futures.  In that context I offer the following material for discussion purposes*.

The table below highlights key features of a subset of the eleven home price index contracts traded on the CME for settlement in Nov. 2009, and the corresponding CS indices, as of Jan. 16, 2009. 

An interpretation would be that between now and September 2009 (as evidenced by Nov 2009 contract prices) that the CS index will decline by another 6-18% across various regions.

If June 2006 is viewed as the top in home prices (see graph for five indices) than the price declines from top to bottom will approach 50% (San Francisco is the worst).  Those indices that had the highest appreciation (Los Angeles and Miami) are priced (implied futures price) to fall the furthest.  The price declines projected by the Nov ’09 contract would indicate that the indices will retreat to levels seen in 2002-2003 and then remain flat for a few years.

The implied index price declines suggest that those areas that have fallen the hardest already (e.g. Miami) are the closest to finding a bottom (e.g. 90% “done” on the prior table.  On the other hand, New York, which has held up well so far, should expect to bear the brunt of the financial crises with almost half (only 55% “there”) of its total price declines still ahead. 

So, we’re not there yet, but we’re getting close. (And it looks like our ride will be a round trip to 2002).

_________________

*I may have positions or pending bids and offers in some of the contracts.