Tracking Asia’s Black Swans
Remarks prepared for the IISS Global Strategic Review
“Managing Global Security and Risk”
Geneva, Switzerland: September 7~9, 2007
2nd Plenary Session: Asia and Managing New Global Risks
James Shinn
Asian and Pacific Security Affairs
U.S. Department of Defense[i]
“A little risk management can save a lot of fan cleaning.”[ii]
If something goes massively wrong in Asia, it is probably going to be a Black Swan.
I’m not talking about some disease-ridden bird carrying avian flu. You’ll be hearing about pandemic risks in a moment from Alan DuPont.
I am a big fan of Nassim Nicholas Taleb, the options trader-turned-philosopher, and author of Fooled by Randomness. Taleb uses the Black Swan metaphor to describe an event with three features: “rarity, extreme impact, and retrospective (though not prospective) predictability.” [iii]
This summer’s meltdown in sub-prime mortgages, with its knock-on effect on risk spreads in general and hedge funds in particular, is exactly the sort of Black Swan event that Taleb writes about. No question about rarity and impact. Now we’ve been reading about its “retrospective predictability” in various Financial Times columns and op-eds for more than a month. Apparently everybody could see it coming, except that almost no-one did.[iv]
At the DOD’s APSA unit, we are seriously concerned about Black Swan events in Asia – low probability, high impact security events.
Here are a few Black Swans possibly flapping our way
- Another Tsunami in Southeast Asia, somewhere along the Belt of Fire
- Human to human avian flu breaks out of Sumatra or Java
- Another EP-3-type military accident, in the air, afloat, or even under the sea – like maybe a Chinese submarine surfaces under a US aircraft carrier, rather than just astern.
- A shooting war starts across the Taiwan Straits
- A nuclear warhead is mounted on a Taipodong missile on a launch pad in North Korea
- A nuclear device appears to be on board a merchant ship in the Pacific headed who knows where.
- A coup in Thailand, the Philippines, or Pakistan.
We are firmly in Taleb-land here. Good or bad, these Swans are all fairly rare events, they have a big impact on security in Asia, and they are rarely, if ever, predicted.
I was much concerned with Swan events while I was working in the financial markets and in Silicon Valley for a couple of decades. I used to spend a lot of time figuring out how to make money from such events, or at least how to avoid being crushed by them. When I joined DOD, it seemed prudent to think through these possible Asian security events with my team – “risk management contingency planning” in corporate-speak, or what my naval aviator colleagues refer to (charmingly) as “pre mishap analysis.”
For each event, we asked ourselves five questions:
- What caused the most similar recent event?
- What scenarios could trigger a repeat?
- Who in the U.S. government actually did what last time?
- What were the lessons learned?
- And (most to the point) what is our game plan for next time?
For added realism, we conduct these analytic sessions in the emergency management room in the Pentagon, a room whose ambiance perfectly creates the sense of claustrophobia and panic so appropriate to national security decision-making.
The details of our contingency planning would be of technical interest to some of you, but there are some risk management observations that are common to most of the countries and almost all of the security problems we are discussing this weekend – with the possible exception of sustained low-intensity conflict.
First, there is usually little or no warning for really bad things. Black Swans usually arrive at 2 am on long weekends, with little or no “hard” warning – none for a tsunami, hours or days for missile shots or coups, a few weeks at most for events like a cross-Straits war, as much as a month for avian flu.
EP3-type incident none
Tsunami none
Asymmetric terror hit none
TD-2 nuclear shot hours or days
Military coup d’etat days or weeks
Attack on Taiwan 2~3 weeks
Failed State Melt-down 2~4 weeks
Avian Flu Breakout 3~6 weeks
Hard warning relies on harried diplomats, furtive spies, and very expensive satellites. The problem is that these Swan events often run on a faster clock than intelligence collection.
For man-made (as opposed to natural) Swan events, the initiating party usually has an incentive to use “denial and deception” to further narrow the warning window.
So much of the post-crisis confusion dealing with these security events is how long it takes to figure out you actually have a Black Swan on your hands – and of what magnitude.
Second, fast reaction is at a premium in managing these security problems. Diplomacy and de-escalation have to move fast: if they fail, military force must be applied quickly and in appropriate scale. Humanitarian or terrorist relief must come quickly, in hours or days, or it is too late. Swans won’t wait.
Third, it is extremely expensive to train and equip a military to deal with these events – unilaterally, or within an alliance or multilateral structure, or some combination of all three. It is also incredibly expense to actually react to these events in the timely manner noted above. For example, the U.S. military relief operation in response to the 2006 Asian tsunami was like the mobilization for a small war. Within days, the US moved 26 ships, 45 helicopters and 15,000 personnel to provide almost 4 million pounds of food and 10 million pounds of relief supplies[v]. Inside DOD, we had not set aside funds for anything this size a response, and were frantically shifting accounts around to cover the costs in the first few days.
In the limit, defense budgeting is a function of threats, and the threats are largely a function of – what else? – Black Swans. So the first question a rational military planner or budgeter asks is, give me a probability of each Swan type.
Alas, fourthly, it is a migraine-inducing exercise to assign probabilities to Swans.[vi] They are “fat tail “events in traders’ parlance, referring to a sort of low asymptote at the ends of the distribution (not necessarily or even usually a bell curve distribution).
Swans do show up from time to time, though. There are about 18 earthquakes greater than scale 7 every year, most of them in Asia, and a serious tsunami somewhere every 18 months on average. [vii] North Korea shot off missiles periodically, though TD-2’s only about once a decade. Thailand and Pakistan usually go more than a decade-and-a-half between coups. The last deadly global flu pandemic was almost a century ago.
Annoyingly, intelligence analysts absolutely refuse to give point or distribution probability estimates of Black Swan outcomes.[viii] To try to get around this problem, I even set up a small betting pool inside the intelligence community in advance of the North Korean missile shot last June. [ix]
I also tried using financial markets as an early warning system. It is empirically proven that orange juice futures markets are more reliable forecasters than the US Weather Service, and that election markets outperform all the opinion polls in choosing presidents.
But it turns out that it’s hard to pick out the security signal from the financial noise in Asia, at least based on my anecdotal analysis. For those of you who are used to staring at Bloomberg screens, three cases are particularly, and graphically, eloquent:
The Korean equity markets (the KOSPI index in this case) did not anticipate, and then barely yawned at the North Korean Taipodong-2 launch (actually, a salvo of missiles) on July 4 , 2006. The North’s nuclear explosion 3 months later managed to elicit less than a 2% drop, if in fact the explosion had any effect at all.
By the same token, the palace coup in Bangkok that removed PM Thaksin Shinawatra in September 06 was apparently not anticipated by the Stock Exchange of Thailand (SET) markets, and then barely budged the SET50 index after the fact (by 2 or 3 %). In contrast, the arbitrary imposition of reserve controls by the Bank of Thailand in December caused a negative swing 8 times as large!
As for avian flu, traders appear to be watching this closely, in particular the WHO announcements of lab analysis of various avian flu cases. A severe influenza pandemic (similar to the one that started in 1918 that killed 50 million people worldwide) is widely predicted to be a quasi-apocalyptic event that could cost the global economy US$800 billion a year[x]. When the WHO issued a press release on the lab analysis of an apparent human-to-human case in Indonesia one afternoon of May 2006, S&P futures took an instantaneous 25 point dive. The only problem is, it was a false alarm.
Fifth, and finally, a realistic and prudent strategy (as opposed to risk management tactics) would anticipate a mix of some bad and some good Swans, too.
- The CCP and the KMT could negotiate a sudden “one country 3 systems” deal across the Taiwan Straits – through one of many back-channels that are periodically activated by the PRC and Taiwan.
- The Burmese junta could collapse and Aung San Suu Chi could be swept into office by popular acclaim – the reverse of a military coup d’etat.
- Sudden peace could break out on the Korean Peninsula – at an inter-Korean summit, for example, or on the margins of the Six Party Talks
- The rest of the Abu Sayyaf thug-terrorist gang in the Philippines could be rolled up and put away for good.
A wise security strategy would factor in the chance of good Swans showing up, and preparing to quickly capitalizing on them, as well as hedging the downside of bad Swan events. Though their appearance individually may be close to random, Swans may be linked causally and sequentially in time. An accidental collision between ship and sub could be the proximate trigger for a Taiwan Straits crisis of major proportions. Good and bad swans may be linked, as complements or substitutes: a peace regime on the Korean Peninsula would probably make TD-2 launches fairly unlikely.
Some – including my hero Taleb -- would consider our Asia risk management contingency exercise a triumph of hope over experience. I prefer to paraphrase Karl Marx: “History always repeats itself: first time as tragedy, second time as farce, third time as a Pentagon simulation.”[xi]
Thank you.
[i] These remarks reflect the author’s personal views and do not necessarily represent the views of the Department of Defense or the U.S. Government. The author welcomes comments and corrections.
[ii] Anonymous
[iii] “Before the discovery of Australia, people in the Old World were convinced that all swans were white, an assailable belief as it seemed completely confirmed by empirical evidence…..One single observation can invalidate a general statement derived from millennia of confirmatory sightings of millions of white swans. All you need is one single (and, I am told, quite ugly) black bird.” Nassim Nicholas Taleb, The Black Swan, page 1.
[iv] Except maybe the Bank of England’s prescient April 2007 Financial Stability Report.
[v] Source: U.S. Pacific Command, Operation Unified Assistance
[vi] Taleb argues that most traders simply assume away the fat tail events in their quantitative models. This simplifying assumption brought Long-Term Capital Management (LTCM) to grief in the summer of 1998, and blew up any number of hedge funds in the summer of 2007. See Roger Lowenstein’s When Genius Failed for a classic account of LTCM.
[vii] http://earthquake.usgs.gov/learning/faq.php?categoryID=11&faqID=69
[viii] For example, there is a standard (and unclassified) boilerplate page in every National Intelligence Estimate (NIE) that warns “Assigning precise numerical ratings to such judgments would imply more rigor than we intend”, substituting a color-coded chart loosely associated with five phrases [remote, unlikely, even chance, probably, almost certainly] for probabilities.
[ix] I didn’t get many takers. I did place a large wager myself that North Korea would light off the TD-2, and won, among other things, a case of excellent wine from the head of a friendly Asian intelligence service.
[x] http://web.worldbank.org. Even a minor pandemic (similar to those that occurred in 1957 and 1968) would reduce U.S. GDP by 1%.
[xi] Karl Marx, “Hegel remarks somewhere that all great, world-historical facts and personages occur, as it were, twice. He has forgotten to add: the first time as tragedy, the second as farce.”The Eighteenth Brumaire of Louis Bonaparte.
