I want to revisit the topic of hedging home price exposure to cities not referenced by CME Case Shiller futures with HPHF over-the-counter ("OTC") agreements^1. HPHF Ratio Agreements might be one tool that solves many of the challenges of a traditional long/short trades.(i.e. one where a price is agreed and each user gains and losses 1:1 for every move in the index relative to a trade).^2 Please allow me ~five lines to explain why I prefer this approach.
*****
A key challenge to OTC agreements is counterparty risk.^3 I've tried to address this is outright agreements by putting boundaries on one-year agreements (typically +/-10%), and having each side post collateral to cover maximum possible price moves. However the last year has shown that 10% moves are possible even inside a year, and some users want to hedge for longer timeframes.^4
Another challenge is that the boundary conditions are set at the beginning of an agreement, and initial levels (on one agreement)might not be seen as neutral should prices change. For example, an index agreement struck at a price of 200, might have had 180-200 price collars, but then next user might want price collars of 205-245 if the next agreement is struck 225.^5 That is, absolute price agreements with different boundaries are not fungible.
Another issue is the lack of fungibility across regions. If I am long and short outright agreements on a combination of ten individual cities, with ten different counterparties, I have to post margin on all ten exposures -even if those ten exposures might be highly correlated.^6
Net individual absolute price agreements would consume a lot of capital to handle a large number of trades (if each has different boundaries and reference cities).
Ratio agreements strip out the majority of the absolute price risk and focus on the performance of one index versus a benchmark (in this case the Case Shiller 10-city index). Let me illustrate using Atlanta as an example.
*****
The graph below shows the ratio of the NSA Case Shiller Atlanta index divided by the Case Shiller 10-city index for the last seven years. The year-end values are highlighted a) to reduce the impact of differences in seasonality, and b) as agreements will settle on year-end values. ^7 From 2014-2020, the ratio rose as the Atlanta home price index had higher percentage gains than the 10-city index. Recently gains in Atlanta home prices have lagged the 10-city index and some have expressed concerns about home price metrics in Atlanta (see below).
Using today's numbers, the Atlanta index is 174.49 and the HCI index is 264.77 for a ratio (always region divided by 10-city index) of 0.659.
Since the ratio captures changes in a regional index versus the 10-city index, if the Atlanta index rose 1% (to 176.24), and the 10-city index remained unchanged, the new ratio would also be 1% higher, at 0.6656. Thus if someone had a neutral view on the absolute level of prices, but thought that one region would outperform the 10-city index (or others^8) ratio trades might be a useful tool.
Note that views on relative gains into the future should get priced into the levels where ratio trades are bid and offered. It happens that I've quoted Atlanta ratios at levels that straddle "unchanged" (versus the ratio) but that would likely not be the case for Chicago (trending lower) or Salt Lake (trending higher).
Note further that the ATX/HCI ratio has stayed in a very narrow range (between 0.62-0.68), despite large moves in both indices. That is, while the absolute price move in the ATX index has been ~46% since 2014, the ratio has only moved ~5%. This is due to the high correlation between ATX and HCI index values. Most of the price move in ATX can be attributed to movements in the HCI index. Separating the "intrinsic" risk of price moves in the HCI (10-city) index, allows a user take exposure to the much less volatile relative price movements. This "less riskiness" can be extracted into ratio agreements with smaller percentage boundaries, requiring less capital.
As an example, I would "buy" the ratio at 0.655 and "sell" it at 0.664, with boundaries of 0.62 and 0.72 (+/- ~0.4) using the same format as in other HPHF agreements. That is, the long would be a call with a strike of 0.62 and a cap of 0.70, and the "short" would be a put with strike of 0.70 and a floor of 0.62.
Note that the bid/offer on the quotes is .009 or 1.4%, compared to much wider bid/ask spreads on outright agreements. Note further, that the boundaries here are ~+/- 6% (lower than an outright trade) and that those boundaries have not been approached in the last three years.
A third-party buyer would pay .044 points (quoted offer minus call strike, or 0.664-.062) and a seller would pay 0.045 points ( put strike minus quoted bid, or 0.700-0.655).
Points would be negotiated based on the notional exposure a user wants. For example, if each .001 is worth $100, then the buyer would pay $4400 in the example above. The maximum payout is the difference between the floor and cap, of 0.08, so the total maximum payout would be $8,000. At a ratio of 0.66 the notional value is this example is $68,000.
Why do this, and how does it help one hedge Atlanta (or any other city)?
The key to this ratio approach is that one might better address the overall risk of a move in the ATX index, by separating the two components -the absolute price risk of 10-city index moves (that may dominate the overall risk), and the relative performance of a region vs. the 10-city index move- into two more liquid, less capital-intensive instruments - a trade in CME 10-city index futures ^9 combined with an HPHF Ratio Agreement^10. I would be happy to coordinate both legs, but there's no reason that the two trades have to be done simultaneously.
Note that in the example above, the notional value of the Ratio Agreement was $68,000. That's about the notional value of one 10-city index contract, so the notional values would line up. If Ratio Agreements are to be combined with 10-city index contracts, they will only line up with multiple values of 10-city index contracts. (If successful (with a proof of concept demonstrated here) I will petition the CME to introduce mini-10 city contracts to allow for a broader range of notional value hedges and maybe even ratio contracts on an expanded list of cities!?!)
Note that while both legs will cash settle -so there's no obligation to unwind- the ability to reverse the 10-city index risk (the historically much larger risk) can be done on the CME at any time, and on markets where bid/ask spreads tend to compress as the number of months to expiration slows.
Further, a side benefit of ratio trades is that if users are looking to hedge absolute price risk, more activity will be directed to the CME 10-city index contracts. That increased interest will likely increase the depth of market, further benefiting other CME contracts via calendar and intercity spread trades.
Both agreements and 10-city index futures can be done in granular amounts, so that the hedging decision does not have to be made in one 100% Buy vs 100% Sell vs 100% Hedge decision.
The above HPHF Ratio Agreements should work well for the ten cities that constitute the balance of the Case Shiller 20-city index (i.e. Atlanta, Charlotte, Cleveland, Dallas, Detroit, Minneapolis, Phoenix, Portland, Tampa and Seattle) as the index values for these indices and the 10-city index are released at the same time, and historical correlations have been high. (Besides there's no other publicly available home price index hedging product that I'm aware of.)
Intercity spread trades are available for the current ten regional contracts, but each contract has a different notional value (and is subject to margin changes) so I'd be open to ratio trades referencing those cities. For example, someone with a long-term view that Miami will underperform the rest of the country over the next five years (or in theory, longer) due to rising sea levels, might negotiate a short MIA/HCI ratio agreement.
An RMBS issuer that has "too much" exposure to San Francisco (which might penalize a deal's capital structure) might negotiate a short SFR/HCI agreement. in theory, one should be able to write puts and calls on ratios, which, given the low historical volatility - in SOME cities -might require relatively low premiums.
Beyond referencing Case Shiller indices, I'd be open to agreements on other large cities (e.g. Austin, Houston, Nashville, Philadelphia, and Salt Lake) that are referenced by Freddie Mac home price indices, with adjustments (to negotiate) as indices are released on different dates.
I'm showing levels for trades that expire on the Feb '22 and Feb '23 index releases, but would be open to longer-dated agreements (but with wider boundaries).
Finally, the reason I used Atlanta is that Jonathan Lansner published a piece in his "Bubble Watch" column at the Orange County Register . While his primary focus was his backyard of California, Jonathan noted that Atlanta showed up (by his tally) as the riskiest city in the US.
I don't have a view on his call, but his statement does suggest that someone might want to hedge their Atlanta home price risk exposure (and this other top 10 picks), hence this blog.
Please feel free to contract me if you have any questions about this post, if you have any region on which you'd like to explore an agreement, or if you'd just like to learn more about using home price index derivatives in hedging strategies.
Thanks, John
^1 -Over the counter agreements, i.e, where there is no exchange acting as a middleman and the parties have exposure to each other.
^2 - This assumes that a user has a futures account that can trade the 10-city index contracts. In addition to the benefits described above, opening a futures contract evidences someone being able to get through a FCM's KYC process.
^3 -There are others, to include price discovery and lack of being able to offset the trade with an exchange, and inability to assign the agreements to third parties.
^4- A benefit to HPHF Agreements with boundaries is that users are not exposed to increases in margins. Recall that during Spring 2020, a large FCM raised margins on CME Case Shiller futures 10x, which lead some parties to unwind positions.
^5- The issue of calculating unwind prices also gets involved when index levels move close to boundaries.
^6-Percent price changes across cities tend to be highly correlated in the top 25 cities. (More to follow).
^7 -That is, like CME contracts, ratio agreements will settle on the indices released on the last Tuesday in February.
^8- For now, one could accomplish long one region/short another by entering into two ratio trades with offsetting exposures to the 10-city index.
^9 -The symbol for the 10-city index contracts may be either CUS or HCI depending on your trading platform.
^10-Note that there may be scenarios (e.g. should ratios move beyond boundaries) where a combination of an RP agreement and a long/short in the 10-city index futures, is not equivalent to an outright position in the regional exposure.