AIG for Dummies

In a nutshell, here’s what I understand to be happening with American International Group (AIG):

1. AIG is a major player in the market for “Credit Default Swaps (CDS’s).

a. A CDS contract involves the transfer of the credit risk of municipal bonds, emerging market bonds, mortgage-backed securities, or corporate debt between two parties.

b. A CDS provides its buyer with protection against default, a credit rating downgrade, or some other negative “credit event” (e.g., being put on a credit watch list by a rating agency).

c. The CDS seller (in this case, AIG) assumes the credit risk that the buyer does not wish to bear in exchange for a periodic protection fee similar to an insurance premium, and is obligated to pay only if a negative credit event occurs. Essentially, AIG’s CDS business is a form of financial insurance.

2. What’s AIG’s “problem”?

a. As the “credit crunch” has continued to worsen, AIG’s CDS’s are declining in value and becoming increasingly less liquid. Consequently, regulatory and rating agency capital standards are essentially forcing AIG to collateralize these assets by raising more capital. However, it is very difficult (particularly for an already credit-impaired firm like AIG competing in an adverse financial market environment) to raise the capital needed on a timely basis in order to be able to comply with these standards.

b. To make matter even worse, a substantial share of AIG’s business involves exposure to commercial insurance products, whose buyers (primarily private and public sector risk managers) expect to place their business with high credit quality insurers. As AIG’s credit rating continues to decline, this substantially puts at risk AIG’s valuable commercial insurance franchise, which in turn makes it all the more difficult for AIG to raise capital.

It can be very tempting to criticize someone else’s actions when you have the benefit of hindsight. Having said that, the credit crisis does expose a fundamental flaw with AIG business model; specifically, it probably doesn’t make much sense to expose a business which has tremendous franchise value (i.e., AIG’s commercial insurance business) to an inordinate amount of credit risk. It is interesting to note that by and large, most firms in the financial guarantee business (e.g., companies like MBIA and AMBAC) are monoline companies; i.e., they market and distribute financial insurance products only.

Prediction Markets Update (September 15, 2008)

By comparison with real-world financial markets, prediction markets have lately grown quite boring. As was the case yesterday, the changes in the 2008.PRES.McCAIN Intrade contract (down .9 at 51.6) and 2008.PRES.OBAMA (up .4 at 47.2) Intrade contract are very marginal. There is also very little movement in the state-by-state contracts, so based upon my (somewhat arbitrary) cutoff price point of 55 for allocating Electoral College votes, it still appears that Mr. Obama and Mr. McCain are virtually tied (with 264 and 265 Electoral College votes respectively), with only Colorado (and its 9 Electoral College votes) “undecided”.

Similarly, Nate Silver’s PECOTA model (see FiveThirtyEight) shows little change from yesterday, allocating 280 (instead of 281) Electoral College votes to Mr. McCain and 259 (instead of 258) electoral College votes to Mr. Obama.

Observations concerning some of today's events in the worlds of finance, insurance, and risk management

1. Lehman Brothers and AIG: As John Markman cleverly notes, “…the Federal Reserve stood up to the big Wall Street financial houses on Sunday and essentially told them, ‘thanks but no thanks’ on their request for a bridge loan to nowhere”. Consequently, Lehman Brothers (the 4th largest investment bank in the world) has filed for bankruptcy protection, and its shares are trading today for 23 cents per share. Bill Gross made an interesting comment concerning Lehman and American International Group (AIG) this morning on CNBC; he noted that while AIG is technically solvent, it is highly illiquid, whereas Lehman is technically insolvent but otherwise quite liquid. Just one year ago, AIG had a market capitalization of nearly $200 billion, making it one of America’s most valuable publicly traded corporations. Today, AIG’s market capitalization stands at $18 billion. The problem AIG faces is that if this company can’t resolve its liquidity problems really soon (like sometime within the next couple of days), then it risks being downgraded by the credit rating agencies. If AIG’s credit rating becomes impaired, it stands to lose a substantial share of its client base. AIG is particularly heavily exposed to commercial insurance lines of business, and their clients have no interest in doing business with credit-impaired insurers. If this scenario plays out, expect to see AIG follow up with its own bankruptcy filing.

 
2. Hurricane Ike: The three major catastrophe risk modeling firms (Applied Insurance Research (AIR), EQECAT, and Risk Management Solutions (RMS)) project insured losses of between $6 – $18 billion, primarily in the Texas counties of Brazoria, Harris, Galveston, Chambers and Jefferson. Since the Texas Windstorm Insurance Association (TWIA) is heavily exposed to insured properties in these counties, I suspect that claims may very likely exceed TWIA’s capital base of $1.87 billion, which consists $370 million in a Catastrophe Reserve Trust Fund and $1.5 billion in reinsurance. Depending upon the magnitude of TWIA’s Ike-related losses, private insurers in Texas could face assessments which would have the potential for setting in motion a financially vicious cycle in which future earnings and capitalization are impaired, which in turn adversely affect insurer credit ratings. Matters can only get worse if the Texas gulf coast gets hit with any more hurricanes any time soon.
 
3. Oil Speculation: Apparently so-called “oil speculators” were not behind the large run-up in oil prices during the first half of 2008, any more than they were behind the subsequent retreat in oil prices during the past couple of months. Last week, the Commodity Futures Trading Commission (CFTC) issued a report which provides compelling empirical evidence to the effect that financial trading hasn’t been driving price moves in oil markets after all. For a synopsis of this study, see the article from today’s Wall Street Journal entitled “See You Later, Speculator”. As I noted in my July 10, 2008 blog entry entitled “Rebuttal of ‘An open letter to all airline customers‘”, this finding comes as no surprise. Thankfully American Airlines has finally gotten around to removing the link from its home page urging consumers to “stop oil speculators by sending Congress an S.O.S.”…
]]>

Observations concerning some of today’s events in the worlds of finance, insurance, and risk management

1. Lehman Brothers and AIG: As John Markman cleverly notes, “…the Federal Reserve stood up to the big Wall Street financial houses on Sunday and essentially told them, ‘thanks but no thanks’ on their request for a bridge loan to nowhere”. Consequently, Lehman Brothers (the 4th largest investment bank in the world) has filed for bankruptcy protection, and its shares are trading today for 23 cents per share. Bill Gross made an interesting comment concerning Lehman and American International Group (AIG) this morning on CNBC; he noted that while AIG is technically solvent, it is highly illiquid, whereas Lehman is technically insolvent but otherwise quite liquid. Just one year ago, AIG had a market capitalization of nearly $200 billion, making it one of America’s most valuable publicly traded corporations. Today, AIG’s market capitalization stands at $18 billion. The problem AIG faces is that if this company can’t resolve its liquidity problems really soon (like sometime within the next couple of days), then it risks being downgraded by the credit rating agencies. If AIG’s credit rating becomes impaired, it stands to lose a substantial share of its client base. AIG is particularly heavily exposed to commercial insurance lines of business, and their clients have no interest in doing business with credit-impaired insurers. If this scenario plays out, expect to see AIG follow up with its own bankruptcy filing.
 
2. Hurricane Ike: The three major catastrophe risk modeling firms (Applied Insurance Research (AIR), EQECAT, and Risk Management Solutions (RMS)) project insured losses of between $6 – $18 billion, primarily in the Texas counties of Brazoria, Harris, Galveston, Chambers and Jefferson. Since the Texas Windstorm Insurance Association (TWIA) is heavily exposed to insured properties in these counties, I suspect that claims may very likely exceed TWIA’s capital base of $1.87 billion, which consists $370 million in a Catastrophe Reserve Trust Fund and $1.5 billion in reinsurance. Depending upon the magnitude of TWIA’s Ike-related losses, private insurers in Texas could face assessments which would have the potential for setting in motion a financially vicious cycle in which future earnings and capitalization are impaired, which in turn adversely affect insurer credit ratings. Matters can only get worse if the Texas gulf coast gets hit with any more hurricanes any time soon.
 
3. Oil Speculation: Apparently so-called “oil speculators” were not behind the large run-up in oil prices during the first half of 2008, any more than they were behind the subsequent retreat in oil prices during the past couple of months. Last week, the Commodity Futures Trading Commission (CFTC) issued a report which provides compelling empirical evidence to the effect that financial trading hasn’t been driving price moves in oil markets after all. For a synopsis of this study, see the article from today’s Wall Street Journal entitled “See You Later, Speculator”. As I noted in my July 10, 2008 blog entry entitled “Rebuttal of ‘An open letter to all airline customers‘”, this finding comes as no surprise. Thankfully American Airlines has finally gotten around to removing the link from its home page urging consumers to “stop oil speculators by sending Congress an S.O.S.”…

Prediction Markets Update (September 14, 2008)

Tonight, (at 8:10 p.m. Central time), the McCain Intrade contract (2008.PRES.McCAIN) is trading at 52.5, whereas the Obama Intrade contract (2008.PRES.OBAMA) is trading at 47.1; 2008.PRES.McCAIN is 1 point higher than it was the last time I posted a blog entry about the election (on Saturday, September 12), whereas 2008.PRES.OBAMA is .9 points lower.

Examining the state-by-state contracts, I find that by using a cutoff price point of 55 for allocating Electoral College votes, that this continues to “give” Barack Obama 264 Electoral College votes and John McCain 265 Electoral College votes. However, I couldn’t help but notice that many of the bid-ask-last price quotes and volume statistics for the various state-by-state contracts haven’t budged since Friday, so apparently there hasn’t been much investor interest in these contracts this weekend!

Turning our attention next to Nate Silver’s PECOTA model (see FiveThirtyEight), Mr. Silver’s model now has McCain winning the Electoral vote by a margin of 289 to 249.

Prediction Markets Update (September 12, 2008)

As of 4:40 p.m. Central time today, the McCain Intrade contract (2008.PRES.McCAIN) is trading at 51.5, which is 1 point up from yesterday’s close of 50.5, whereas the Obama Intrade contract (2008.PRES.OBAMA) is trading at 48, which is .4 points down from yesterday’s close of 48.4.

Examining the state-by-state contracts, I find that by using a cutoff price point of 55 for allocating Electoral College votes, that this currently gives Barack Obama 264 Electoral College votes and John McCain 265 Electoral College votes. Basically, the story today on a state-by-state is the same as it was yesterday; out of the 102 total contracts (Democratic and Republican) offered (for the 50 states and the District of Columbia), only the two Colorado contracts (COLORADO.DEM and COLORADO.REP) remain within my arbitrarily defined “swing state” price interval, since the last trades on these contracts were 52 and 48 respectively.

I like to compare the prediction market data with Nate Silver’s PECOTA model which is published on the FiveThirtyEight website. Today (for the first time), Mr. Silver’s model has McCain winning the Electoral vote by a margin of 277 to 261.

Addendum: September 12, 2008 Electoral College Vote allocation

John McCain (265): Alabama (9), Alaska (3), Arizona (10), Arkansas (6), Florida (27), Georgia (15), Idaho (4), Indiana (11), Kansas (6), Kentucky (8), Louisiana (9), Mississippi (6), Missouri (11), Montana (3), Nebraska (5), Nevada (5), North Carolina (15), North Dakota (3), Ohio (20), Oklahoma (7), South Carolina (8), South Dakota (3), Tennessee (11), Texas (34), Utah (5), Virginia (13), West Virginia (5), and Wyoming (3)

Barack Obama (264): California (55), Connecticut (7), Delaware (3), District of Columbia (3), Hawaii (4), Illinois (21), Iowa (7), Maine (4), Maryland (10), Massachusetts (12), Michigan (17), Minnesota (10), New Hampshire (4), New Jersey (15), New Mexico (5), New York (31), Oregon (7), Pennsylvania (21), Rhode Island (4), Vermont (3), Washington (11), and Wisconsin (10)

Prediction Markets Update (September 11, 2008)

As of 3:10 p.m. Central time today, the McCain Intrade contract (2008.PRES.McCAIN) is trading at 51.5, which is 1.7 points up from yesterday’s close of 49.8, whereas the Obama Intrade contract (2008.PRES.OBAMA) is trading at 47.7, which is 1.2 points up from yesterday’s close of 48.9.

Examining the state-by-state contracts, I find that by using a cutoff price point of 55 for allocating Electoral College votes, that this currently gives Barack Obama 264 Electoral College votes and John McCain 265 Electoral College votes. Since yesterday, the VIRGINIA.REP and NEVADA.REP have both rallied by roughly 5 points (to 60 and 59.9 respectively), while the NEWHAMPSHIRE.DEM has rallied 8 points to 65, so it would appear that Virginia and Nevada are moving more solidly to McCain’s advantage, whereas New Hampshire is moving even more in Obama’s direction. Out of the 102 total contracts (Democratic and Republican) offered (for the 50 states and the District of Columbia), only the two Colorado contracts (COLORADO.DEM and COLORADO.REP) remain within my arbitrarily defined “swing state” price interval, since the last trades on these contracts were 55 and 48 respectively. Other things equal, at this point it appears that Colorado is absolutely pivotal, since its 9 Electoral College votes could put either candidate over the 270 vote total required to win the presidential election.

Finally, for what it’s worth, Nate Silver over at FiveThirtyEight currently puts the Electoral Vote for Obama and McCain at 270 versus 268, which shows once again that a simple, cursory inspection of state-by-state prediction market prices produces a result which is qualitatively quite similar to the result obtained by Mr. Silver’s much more sophisticated PECOTA model.

Addendum: September 11, 2008 Electoral College Vote allocation

John McCain (265): Alabama (9), Alaska (3), Arizona (10), Arkansas (6), Florida (27), Georgia (15), Idaho (4), Indiana (11), Kansas (6), Kentucky (8), Louisiana (9), Mississippi (6), Missouri (11), Montana (3), Nebraska (5), Nevada (5), North Carolina (15), North Dakota (3), Ohio (20), Oklahoma (7), South Carolina (8), South Dakota (3), Tennessee (11), Texas (34), Utah (5), Virginia (13), West Virginia (5), and Wyoming (3)

Barack Obama (264): California (55), Connecticut (7), Delaware (3), District of Columbia (3), Hawaii (4), Illinois (21), Iowa (7), Maine (4), Maryland (10), Massachusetts (12), Michigan (17), Minnesota (10), New Hampshire (4), New Jersey (15), New Mexico (5), New York (31), Oregon (7), Pennsylvania (21), Rhode Island (4), Vermont (3), Washington (11), and Wisconsin (10)

Prediction Markets Update (September 10, 2008)

Today (for the first time in a long time), the McCain Intrade contract (2008.PRES.McCAIN) is essentially trading at parity with the Obama Intrade contract (2008.PRES.OBAMA); the last recorded prices for these contracts (at 5:12 p.m. Central time today) are 49.9 and 50 respectively. Yesterday, 2008.PRES.OBAMA closed at 52.4 and 2008.PRES.McCAIN closed at 47.4, and at the time the state-by-state contracts (see http://blog.garven.com/2008/09/09/forecast-of-the-2008-presidential-election-on-the-basis-of-state-by-state-prediction-market-contracts/) seemed to imply an electoral college advantage for Obama over McCain of 260-247, with Virginia (13 electoral college votes), Nevada (5 electoral college votes), Colorado (9 electoral college votes), and New Hampshire (4 electoral college votes) fitting into the “swing” state category by virtue of how close their prices were at the time.

Naturally, after seeing today’s price changes for 2008.PRES.McCAIN and 2008.PRES.OBAMA, I couldn’t help but wonder what might be going on with the .DEM and .REP versions of the Virginia, Nevada, Colorado, and New Hampshire state contracts. Today, VIRGINIA.REP and NEVADA.REP are more valuable than VIRGINIA.DEM and NEVADA.DEM respectively (55 and 55 versus 50 and 45 respectively), whereas COLORADO.DEM and NEWHAMPSHIRE.DEM are pricier than COLORADO.REP and NEWHAMPSHIRE.REP (52.2 and 57 versus 47.9 and 47 respectively (technical note: when added together, prices can exceed 100 because they are not necessarily synchronous)). If Virginia and Nevada went McCain’s way while Colorado and New Hampshire favored Obama, this would translate into a 273-265 win for Obama. Interestingly, this lines up very closely with today’s version of Nate Silver’s PECOTA model (see http://www.fivethirtyeight.com; as I understand it, the PECOTA model is not based upon prediction markets data and leans more toward some combination of state-by-state polling data, simulation, and econometrics), which currently predicts a 275-263 Obama victory.

There is a scenario in which the Electoral College ends up in a tie; this could happen if Virginia, Nevada, and New Hampshire went to McCain and Colorado went to Obama. If a tie occurs in the Electoral College, then the election is decided by the House of Representatives. Under this scenario, the most likely result would be an Obama presidency, since the Democratic Party currently controls the House of Representatives and the HOUSE.DEM.2008 contract (which pays 100 Intrade points in the event that the Democrats control the House of Representatives after 2008 Congressional Elections) currently indicates a 90 percent probability that this will continue to be the case after the general election on November 4.

Forecast of the 2008 presidential election on the basis of state-by-state prediction market contracts

A colleague pointed out to me today that Intrade hosts prediction market contracts on a state-by-state basis. These contracts have very simple names; e.g., TEXAS.DEM is a contract which will pay 100 Intrade points ($10) if the Obama/Biden ticket wins Texas and 0 otherwise, whereas TEXAS.REP is a contract which will pay 100 Intrade points ($10) if the McCain/Palin ticket wins Texas and 0 otherwise (FWIW, these contracts are currently quoted at 5 and 95 Intrade points respectively). This afternoon, I downloaded these data to a spreadsheet and sorted by price. Currently, the most valuable GOP contract is ALABAMA.REP, whereas the most valuable Democrat contract is ILLINOIS.DEM; in both cases, the last trades were at 98. The least valuable GOP contract is currently HAWAII.REP, which can be had for 2.30 Intrade points, whereas the least valuable Democrat contract is currently UTAH.DEM, which can be had for 2.5 Intrade points.

The next step in my analysis involved determining what an appropriate cutoff point might be for allocating Electoral College votes. I arbitrarily selected a price point of 55. Using this criterion, it would appear that the Obama/Biden is currently “winning” the following 21 states (I am including the electoral votes for each state in parentheses after its name): Illinois (21), Hawaii (4), Delaware (3), District Of Columbia (3), Vermont (3), New York (31), Maryland (10), Massachusetts (12), Maine (4), Rhode Island (4), Connecticut (7), California (55), Washington (11), New Jersey (15), Oregon (7), Minnesota (10), Wisconsin (10), Iowa (7), Pennsylvania (21), New Mexico (5), and Michigan (17). Applying the same criterion to the McCain/Palin ticket, I count the following 26 states: Alabama (9), Kentucky (8), Utah (5), Idaho (4), Alaska (3), Wyoming (3), Oklahoma (7), Nebraska (5), Arizona (10), Kansas (6), Louisiana (9), South Carolina (8), Texas (34), Arkansas (6), South Dakota (3), West Virginia (5), Mississippi (6), Tennessee (11), Georgia (15), North Carolina (15), North Dakota (3), Indiana (11), Missouri (11), Montana (3), Florida (27), and Ohio (20). If you add all the electoral college votes up, this analysis gives Obama/Biden a current electoral college total of 260, whereas McCain/Palin total 247 electoral college votes (270 votes are needed in order to win the election). The prediction markets also seem to indicate that the election is likely to be won or lost depending upon what happens in the following four battleground states (whose current prices are between 45 and 55 Intrade points): Virginia (13), Nevada (5), Colorado (9), and New Hampshire (4). If the 260-247 tally holds, Obama/Biden could win the election by also carrying Virginia or Nevada and Colorado, whereas McCain/Palin would need to pick up 3 of these 4 states.

Stay tuned; we live in interesting times!

Best Bet for Next President: Prediction Markets

Going into the Democratic convention this week, most polls of registered voters were showing the presidential race to be a statistical dead heat between Obama and McCain (although today’s Gallup Poll Daily tracking poll is beginning to reflect the typical convention “bounce” that Gallup has observed in most party conventions in recent decades). However, a more reliable indicator of what’s really going on is provided by intrade.com, which trades “all or nothing” futures contracts on a number of different contingent events, including who is likely to be elected the President of the United States. These contracts pay 100 points (where 1 point = $.10) if a specific contingent event occurs and 0 points otherwise otherwise; consequently the prices for these contracts represent discounted, risk neutral probabilities.

According to intrade.com, Obama currently has a nearly 3:2 edge over McCain (which is actually down considerably from the 2:1 advantage which Obama had a few months ago after he became the presumptive Democratic presidential nominee). An important problem with national polls such as Gallup is that these polls are designed to provide an overall snapshot at a given point in time of the nation’s political mood. However, the nation’s political mood doesn’t elect presidents; the electoral college does. While the race may appear to be close nationally, once you do the math state-by-state, you get a very different picture. This information is provided by the intrade political futures contracts, since they represent unbiased bets on the electoral college outcome. While information on overall popularity amongst registered voters is not without interest, the election obviously will come down to who actually turns up to vote, how the undecided segment is swayed, and how all this is distributed on a state-by-state basis. These important subtleties are implicitly captured by prediction markets prices but are completely lacking in the national polling data.

For more information concerning the topic of “prediction markets”, I recommend a Wall Street Journal essay entitled “Best Bet for Next President: Prediction Markets” by Wharton professor Justin Wolfers. Wolfers and Zitzewitz also published the following article in 2004 in the Journal of Economic Perspectives:

Wolfers, Justin and Eric Zitzewitz, 2004, “Prediction Markets”, Journal of Economic Perspectives, Vol. 18, No. 2 (Spring), pp. 107-126.

A blog exploring the intersection of finance, economics, risk, public policy, & life in general