Didier Sornette: How we can predict the next financial crisis

Translator: Joseph Geni Reviewer: Morton Bast

譯者: Yen Hung 洪晨晏 審譯者: Olivia Lo

Once upon a time

曾幾何時，

we lived in an economy of financial growth and prosperity.

我們生活在經濟與金融 的增長和繁榮中。

This was called the Great Moderation,

這就是人們所說的“大穩健”時期，

the misguided belief by most economists,

這種誤導性的看法被大多數經濟學家，

policymakers and central banks

政策制定者以及中央銀行所相信著。

that we have transformed into a new world

他們深信我們已經演進到一個新時代，

of never-ending growth and prosperity.

一個永不停止增長和永葆繁榮的世界，

This was seen by robust and steady GDP growth,

這個結論是由強健和穩定的GDP增長、

by low and controlled inflation,

受到良好控制的低通貨膨脹率、

by low unemployment,

低失業率，

and controlled and low financial volatility.

和受控制的低金融波動率而得出的。

But the Great Recession in 2007 and 2008,

但是2007年和2008年的大衰退，

the great crash, broke this illusion.

和大崩盤，打破了這個繁榮假象。

A few hundred billion dollars of losses in the financial sector

金融行業損失了數千億美元，

cascaded into five trillion dollars

整個世界的GDP

of losses in world GDP

流失了五兆美元，

and almost $30 trillion losses

還有全球股市大概虧損了

in the global stock market.

30兆美元。

So the understanding of this Great Recession

因此，人們對大衰退的理解

was that this was completely surprising,

是一場完全出乎常理的意外、

this came out of the blue,

是突然湧現的危機，

this was like the wrath of the gods.

就像是眾神的憤怒一般。

There was no responsibility.

沒有人有辦法為此負責。

So, as a reflection of this,

作為對這場危機的反饋，

we started the Financial Crisis Observatory.

我們成立了金融危機觀測局。

We had the goal to diagnose in real time

我們當時要達到的目標

financial bubbles

是能實時判斷出金融泡沫，

and identify in advance their critical time.

和預先判斷出它們的臨界點。

What is the underpinning, scientifically, of this financial observatory?

具體來說，是什麼架構了 金融觀測台的地基與科學方法呢？

We developed a theory called "dragon-kings."

是我們開發的一個叫“龍王”的理論。

Dragon-kings represent extreme events

"龍王理論"代表著各個種類中的

which are of a class of their own.

極端事件。

They are special. They are outliers.

他們是特殊的，也是異常值，

They are generated by specific mechanisms

他們衍生於某些特殊機制，

that may make them predictable,

因此是可以預測的

perhaps controllable.

還可能是可控制的。

Consider the financial price time series,

“龍王理論"考慮到了 金融價格的時間序列，

a given stock, your perfect stock,

和一支給定的股票，你的最佳股票

or a global index.

或者一個全球指數

You have these up-and-downs.

你能看到這些上下起伏的狀況。

A very good measure of the risk of this financial market

峰谷法則是一個金融市場 非常好的測量風險工具，

is the peaks-to-valleys that represent

從高峰至低谷呈現著

a worst case scenario

一個最糟糕的劇情，

when you bought at the top and sold at the bottom.

你在最高點買入，又在最低價賣出。

You can look at the statistics, the frequency of the occurrence

你可以看到統計資料--不同規模的峰谷法則

of peak-to-valleys of different sizes,

的出現頻率，

which is represented in this graph.

這些都會在表格裡呈現。

Now, interestingly, 99 percent

現在有趣的是，99%的

of the peak-to-valleys of different amplitudes

不同振幅的峰谷法則

can be represented by a universal power law

都可以由普通的冪次定律反映出來

represented by this red line here.

由這條紅線反映。

More interestingly, there are outliers, there are exceptions

更有趣的是，這還是離群值，這些屬於特例。

which are above this red line,

在這條紅線之上

occur 100 times more frequently, at least,

和那些基於其他99%峰谷法則校準的

than the extrapolation would predict them to occur

所預測會發生的推論相比

based on the calibration of the 99 percent remaining

發生的頻率

peak-to-valleys.

至少超過100倍以上。

They are due to trenchant dependancies

這可是個依聯性極強的準則，

such that a loss is followed by a loss

損失的後頭跟著一個損失，

which is followed by a loss which is followed by a loss.

又接續著另一個損失。

These kinds of dependencies

大多標準風險管理工具都遺漏了

are largely missed by standard risk management tools,

如此露骨的關聯性，

which ignore them and see lizards

當他們應該留意"龍王"的時候

when they should see dragon-kings.

常常是忽略它而只看到一些蜥蜴。

The root mechanism of a dragon-king

"龍王理論"的根本機制

is a slow maturation towards instability,

是一項緩慢而走向不穩定的成熟過程

which is the bubble,

它就是泡沫，

and the climax of the bubble is often the crash.

而泡沫的頂點往往被隨破滅的開始。

This is similar to the slow heating of water

這跟用文火燒開水很相似

in this test tube reaching the boiling point,

在實驗中當試管中的水達到沸點時，

where the instability of the water occurs

裡頭的水將產生不穩定的狀態，

and you have the phase transition to vapor.

轉變型態為蒸氣相。

And this process, which is absolutely non-linear --

而這個過程絕非是線性演變，

cannot be predicted by standard techniques --

故無法透過標準化技術去預測，

is the reflection of a collective emergent behavior

這是集體緊急行為的反射，

which is fundamentally endogenous.

是根深蒂固的內在根性。

So the cause of the crash, the cause of the crisis

所以，泡沫破滅的根源，危機的根源

has to be found in an inner instability of the system,

須從系統內在的不穩定中發覺，

and any tiny perturbation will make this instability occur.

且任何細微的擾動 都會讓這個不穩定現形。

Now, some of you may have come to the mind

現在你們當中某些人可能已經察覺，

that is this not related to the black swan concept

這和黑天鵝效應並不相關，

you have heard about frequently?

你經常聽到這個名詞嗎?

Remember, black swan is this rare bird

要記得，黑天鵝是稀有的鳥類，

that you see once and suddenly shattered your belief

那種你才看過一次，便會讓人突然粉碎

that all swans should be white,

過往"天鵝都該是白色的"那種信念

so it has captured the idea of unpredictability,

所以它也囊括著不可預測性、不可獲知性，

unknowability, that the extreme events

表示那些極端的事件

are fundamentally unknowable.

根本就無法得知。

Nothing can be further

幾乎沒有什麼能脫離

from the dragon-king concept I propose,

我提出的"龍王理論"範疇，

which is exactly the opposite, that most extreme events

它恰好完全相異於大多數的極端事件，

are actually knowable and predictable.

是可獲知且可預測的。

So we can be empowered and take responsibility

因此我們可以擁有權力並擔起責任

and make predictions about them.

去對這些事件進行預測。

So let's have my dragon-king burn this black swan concept.

現在就讓我的龍王來燒掉黑天鵝吧!

(Laughter)

(笑聲)

There are many early warning signals

有很多先期的警報訊號

that are predicted by this theory.

可以透過這個理論預測。

Let me just focus on one of them:

讓我為您舉例說明，

the super-exponential growth with positive feedback.

這個超級指數成長伴隨著正向的反饋，

What does it mean?

這是什麼意思呢?

Imagine you have an investment

假想你現在進行一項投資，

that returns the first year five percent,

第一年你獲得了百分之五的報酬，

the second year 10 percent, the third year 20 percent,

第二年百分之十、第三年百分之二十，

the next year 40 percent. Is that not marvelous?

再下一年有百分之四十， 這不是很令人驚奇嗎?

This is a super-exponential growth.

這就是超級指數成長。

A standard exponential growth corresponds

一種標準指數成長，相對於其他增長

to a constant growth rate, let's say, of 10 percent

比如說線性增長只維持在百分之十。

The point is that, many times during bubbles,

重點在於，在泡沫化期間重複演變了很多次

there are positive feedbacks which can be of many times,

諸多正回饋在多次演變中產生，

such that previous growths enhance,

過程中到達先前的成長增強幅度

push forward, increase the next growth

再更往前推進，增長到下一個成長級距，

through this kind of super-exponential growth,

直至超級指數成長的程度

which is very trenchant, not sustainable.

它是非常強而銳利的，並不是維持穩定發展的，

And the key idea is that the mathematical solution

主要概念是來自於

of this class of models exhibit finite-time singularities,

表現在這類有限時間奇異點模型的數學解法。

which means that there is a critical time

意思是在某個臨界時間上，

where the system will break, will change regime.

它的系統將崩解，改變其體系，

It may be a crash. It may be just a plateau, something else.

這可能是崩盤，也可能是平穩的，或其他可能

And the key idea is that the critical time,

關鍵放在臨界時間上。

the information about the critical time is contained

在早期發展超級指數成長時，

in the early development of this super-exponential growth.

有關臨界時間的資訊是包含在內的。

We have applied this theory early on, that was our first success,

我們前期將這個理論應用在實驗中，

to the diagnostic of the rupture of key elements

也是我們第一次成功地

on the iron rocket.

診斷出鐵火箭破裂的關鍵要素。

Using acoustic emission, you know, this little noise

我們採用聲發射器這種你可以聽見

that you hear a structure emit, sing to you

它發出小小噪音在對著你唱歌的東西。

when they are stressed, and reveal the damage going on,

當它承受壓力，開始顯現損壞時

there's a collective phenomenon of positive feedback,

便產生集體的正回饋現象，

the more damage gives the more damage,

也就是越來越多的損壞會伴隨著更多損壞。

so you can actually predict,

因此你可以在可能的範圍間

within, of course, a probability band,

去確切的預測

when the rupture will occur.

它的破裂何時會產生。

So this is now so successful that it is used

所以現在可以成功地用在

in the initial phase of [unclear] the flight.

確認飛機［能否］飛行的初期階段。

Perhaps more surprisingly, the same type of theory

也許更加出奇的是，相同的理論形式

applies to biology and medicine,

也能運用在生物學或是醫療上，

parturition, the act of giving birth, epileptic seizures.

像是在分娩、生產、癲癬發作時。

From seven months of pregnancy, a mother

從懷孕七個月，母體開始感到

starts to feel episodic precursory contractions of the uterus

陣陣的子宮收縮前兆，

that are the sign of these maturations

這類顯著朝向不穩定的、

toward the instability, giving birth to the baby,

將要生下寶寶的成熟過程，

the dragon-king.

龍王理論的徵兆。

So if you measure the precursor signal,

所以假如去測量這些前兆

you can actually identify pre- and post-maturity problems

就可以預先判斷

in advance.

事件發生前後的問題。

Epileptic seizures also come in a large variety of size,

癲癬發作也是有各種不同的形式，

and when the brain goes to a super-critical state,

當大腦進入了超臨界狀態時，

you have dragon-kings which have a degree of predictability

採用龍王理論能有一定程度的預見，

and this can help the patient to deal with this illness.

進而協助病患去處理自身症狀。

We have applied this theory to many systems,

我們已經在許多系統中應用這個理論，

landslides, glacier collapse,

走山、冰川崩塌，

even to the dynamics of prediction of success:

甚至是成功地預測熱門動態，

blockbusters, YouTube videos, movies, and so on.

像是賣座鉅片、YouTube短片、電影等等。

But perhaps the most important application

但或許最重要的應用

is for finance, and this theory

是應用在金融面。

illuminates, I believe, the deep reason

我相信這個理論闡明了 我們曾經歷過的金融危機，

for the financial crisis that we have gone through.

其深層的原因

This is rooted in 30 years of history of bubbles,

根自於三十年的泡沫歷史。

starting in 1980, with the global bubble

自1980年起

crashing in 1987,

全世界的泡沫在1987年破裂，

followed by many other bubbles.

接連在後的有其他諸多泡沫，

The biggest one was the "new economy" Internet bubble

當中最大的叫做2000年的"新經濟"網路泡沫

in 2000, crashing in 2000,

在2000年破裂。

the real estate bubbles in many countries,

之後許多國家的房地產泡沫，

financial derivative bubbles everywhere,

各處充斥著金融衍生產品泡沫，

stock market bubbles also everywhere,

股市泡沫也是無所不在，

commodity and all bubbles, debt and credit bubbles --

商品與各種泡沫，債務與信貸泡沫---

bubbles, bubbles, bubbles.

泡沫，泡沫，泡沫

We had a global bubble.

我們共有這全球性泡沫，

This is a measure of global overvaluation

這是全球市場被高估的程度，

of all markets, expressing what I call

顯露出我所指的，

an illusion of a perpetual money machine

一台永久印鈔機的幻覺

that suddenly broke in 2007.

突然在2007年崩盤了。

The problem is that we see the same process,

問題在於我們總看到相同的策略，

in particular through quantitative easing,

特別是透過量化寬鬆，

of a thinking of a perpetual money machine nowadays

各國政府老想著要永遠打開印鈔機 去解決金融危機

to tackle the crisis since 2008 in the U.S., in Europe,

打從2008年起，

in Japan.

在美國、歐洲、日本都是。

This has very important implications

這有著非常重要的意義

to understand the failure of quantitative easing

來幫助我們理解量化寬鬆

as well as austerity measures

以及緊縮政策為何會失敗。

as long as we don't attack the core,

除非我們直擊核心，

the structural cause of this perpetual money machine thinking.

解決這種永久印刷機的結構性原因。

Now, these are big claims.

這些都是龐大的主張，

Why would you believe me?

為何你需要相信我呢?

Well, perhaps because, in the last 15 years

嗯，也許是因為在過去的15年裡

we have come out of our ivory tower,

我們從象牙塔裡走了出來

and started to publish ex ante --

並開始發表了"事前"

and I stress the term ex ante, it means "in advance" —

我強調這個詞"事前"，意思指的是"預先"

before the crash confirmed

在確認泡沫存在以及金融過剩，

the existence of the bubble or the financial excesses.

在崩盤以前。

These are a few of the major bubbles

這些是近期歷史中，

that we have lived through in recent history.

我們已經歷的一些大型泡沫。

Again, many interesting stories for each of them.

這些泡沫都有其有趣故事。

Let me tell you just one or two stories

再一次地，讓我告訴各位其中一兩件

that deal with massive bubbles.

怎麼處理大規模泡沫的故事。

We all know the Chinese miracle.

我們都知道中國奇蹟。

This is the expression of the stock market

這是股市的大規模泡沫展示，

of a massive bubble, a factor of three,

僅在短短數年間，指數跳漲了三倍，

300 percent in just a few years.

300%

In September 2007,

在2007年的九月，

I was invited as a keynote speaker of a macro hedge fund

我受邀到一個總體對沖基金管理人會議

management conference,

去做專題演講。

and I showed to the conference a prediction

我在會議上做了預測，

that by the end of 2007, this bubble

在2007年底前

would change regime.

這顆泡沫就會發生質變。

There might be a crash. Certainly not sustainable.

可能是場崩盤，鐵定動盪。

Now, how do you believe the very smart,

你們相信這群絕頂聰明、積極、

very motivated, very informed macro hedge fund managers

又見廣多聞的總體對沖基金經理人們

reacted to this prediction?

對這個預測有何反應嗎?

You know, they had made billions

你們知道嗎，他們直到現在

just surfing this bubble until now.

已經丟下數十億在這棵泡沫上衝浪。

They told me, "Didier,

他們對我說: "Didier"

yeah, the market might be overvalued,

"是啦，這個市場確實是可能被高估"

but you forget something.

"但你可能忘了"

There is the Beijing Olympic Games coming

"2008年八月的北京奧運就要來了"

in August 2008, and it's very clear that

"顯然地"

the Chinese government is controlling the economy

"中國政府正在控制經濟"

and doing what it takes

"採取一切手段"

to also avoid any wave and control the stock market."

"也避免任何波動並控制股票市場"

Three weeks after my presentation,

在我發表完的三週之後

the markets lost 20 percent

市值就蒸發了百分之二十，

and went through a phase of volatility,

在經歷劇烈的波動後，

upheaval, and a total market loss of

到了當年年底，

70 percent until the end of the year.

總市值已經損失了百分之七十。

So how can we be so collectively wrong

我們怎麼會犯下這種集體錯誤?

by misreading or ignoring the science

怎麼會在發展成 即將轉變、系統已成熟的時候

of the fact that when an instability has developed,

誤解或是忽略了這樣的科學證據。

and the system is ripe, any perturbation

任何一丁點擾動

makes it essentially impossible to control?

都會讓其無法控制

The Chinese market collapsed, but it rebounded.

中國市場崩垮了，而後再次復甦

In 2009, we also identified that this new bubble,

2009年的時候，我們也辨識出新泡沫。

a smaller one, was unsustainable,

一顆規模略小、不穩定的泡沫。

so we published again a prediction, in advance,

我們就再次"預先地"公開預測

stating that by August 2009, the market will correct,

到2009年八月前，市場會開始修正

will not continue on this track.

不會繼續維持現在的軌跡。

Our critics, reading the prediction,

批評家讀到這篇預測說

said, "No, it's not possible.

"不，絕不可能"

The Chinese government is there.

"中國政府還在這兒呢!"

They have learned their lesson. They will control.

"他們已經學過教訓、中央會控制一切"

They want to benefit from the growth."

"他們想從這波成長裡撈些好處"

Perhaps these critics have not learned their lesson previously.

也許這群批評家 根本沒從之前事件得到教訓。

So the crisis did occur. The market corrected.

危機確實發生，市場轉向了

The same critics then said, "Ah, yes,

這群批評家又改口說

but you published your prediction.

"喔，對啦! 但你先公佈這項預測"

You influenced the market.

"害得市場受干擾了"

It was not a prediction."

"所以這不算真正的預測"

Maybe I am very powerful then.

我猜我大概有超凡的影響力吧!

Now, this is interesting.

而有趣的是，

It shows that it's essentially impossible until now

這顯示了到目前為止

to develop a science of economics

要發展一門經濟的科學幾乎不大可能，

because we are sentient beings who anticipate

因為我們都是有情感的物種，

and there is a problem of self-fulfilling prophesies.

我們會做預測、 還對會自己實現的預知感到障礙。

So we invented a new way of doing science.

所以我們發明了一個新科學方法。

We created the Financial Bubble Experiment.

創造了金融泡沫實驗

The idea is the following. We monitor the markets.

方法如下，我們監測市場

We identify excesses, bubbles.

辨識出過剩、泡沫，

We do our work. We write a report

盡己之責去寫下報告，

in which we put our prediction of the critical time.

放入我們預估的發生危機時機。

We don't release the report. It's kept secret.

我們不主動釋出報告，維持機密。

But with modern encrypting techniques,

但我們將報告以現代加密技術，

we have a hash, we publish a public key,

用hash算法，公開一個公用鑰匙。

and six months later, we release the report,

在六個月以後，我們會釋出報告

and there is authentication.

並且留有證據。

And all this is done on an international archive

這些一切都是在一個國際資料庫中進行，

so that we cannot be accused of just releasing the successes.

這樣就沒有人會說 我們釋出的報告影響市場。

Let me tease you with a very recent analysis.

讓我告訴各位一個最新的分析。

17th of May, 2013, just two weeks ago,

在2013年的五月十七日，也就是兩週前

we identified that the U.S. stock market

我們辨識出美國股市

was on an unsustainable path

正在走向即將轉變的道路。

and we released this on our website on the 21st of May

我們五月二十一日時 發布此訊息在我們的網站上

that there will be a change of regime.

說股市將發生權力更迭。

The next day, the market started to change regime, course.

隔天，市場就開始變化了。

This is not a crash.

這並非崩盤。

This is just the third or fourth act

只是大規模泡沫形成過程中的

of a massive bubble in the making.

第三或是第四步。

Scaling up the discussion at the size of the planet,

擴大討論到整個地球的規模，

we see the same thing.

我們也看到了相同的現象，

Wherever we look, it's observable:

無論我們在何處觀察都是可觀測的；

in the biosphere, in the atmosphere, in the ocean,

在生物圈、在大氣、在海洋

showing these super-exponential trajectories

都展現出這種超級指數變化軌跡，

characterizing an unsustainable path

展現出即將變化的路徑，

and announcing a phase transition.

預告了相同的變化。

This diagram on the right

這張右方的圖片

shows a very beautiful compilation of studies

顯示著非常精細美麗的研究，

suggesting indeed that there is a nonlinear -- possibility

明確指示'著未來數十年間

for a nonlinear transition just in the next few decades.

將有一個非線性的重大轉變發生，

So there are bubbles everywhere.

所以現在到處都充斥著泡沫。

From one side, this is exciting

一方面來說，對我而言實在很興奮，

for me, as a professor who chases bubbles

對一個追隨泡沫的教授、

and slays dragons, as the media has sometimes called me.

一個屠龍夫，有時候媒體這麼稱呼我。

But can we really slay the dragons?

但我們真的能夠"屠龍"嗎?

Very recently, with collaborators,

就在最近，和共同研究者

we studied a dynamical system

我們一起研究了一個動態系統，

where you see the dragon-king as these big loops

龍王在這裡是圈大循環，

and we were able to apply tiny perturbations at the right times

我們可以在恰當的時機點施加細微的擾動

that removed, when control is on, these dragons.

在控制手段運行下，拿下這些龍王

"Gouverner, c'est prévoir."

"Gouverner, c'est prévoir." (法語:預見執政)

Governing is the art of planning and predicting.

統治是計畫與預測的藝術

But is it not the case that this is probably

但這不或許就是

one of the biggest gaps of mankind,

人類至今最大的空缺之一，

which has the responsibility to steer

這些不就是讓我們有責任

our societies and our planet toward sustainability

使我們的社會與星球能夠

in the face of growing challenges and crises?

航向穩定發展的路上嗎?

But the dragon-king theory gives hope.

然而龍王理論帶來了希望。

We learn that most systems have pockets of predictability.

我們學習到多數的系統具有可預測性，

It is possible to develop advance diagnostics of crises

故有機會發展出危機預知診斷技術。

so that we can be prepared, we can take measures,

如此我們可以預作準備、採取對策，

we can take responsibility,

我們有辦法肩負起責任，

and so that never again will

進而阻止如同大衰退時期

extremes and crises like the Great Recession

或是歐債危機這樣的極端事件與危機

or the European crisis take us by surprise.

來驚嚇我們。

Thank you.

謝謝各位。

(Applause)

(掌聲)