The low pressure front that brought flooding to the city of Görlitz on Germany's eastern border with Poland developed over the Mediterranean, bypassed the eastern flank of the Alps, headed for Poland and finally came to a stop at the Erzgebirge (Iron Ore Mountains) along the German-Czech border. But Angelika Wirtz wasn't worried yet.
When the rains came and fell along the Neisse River, on ground that was already saturated and on steep hillside, Wirtz began to take notice. The flood reached Görlitz on the evening of August 7, when the river rose four meters (about 13 feet) in the space of three hours. Parts of the city were underwater and a state of emergency was declared throughout the entire district. When Wirtz found out about it, she typed a note into her computer. It wasn't a particularly long note. Wirtz had seen worse.
On the weekend when the flooding reached Görlitz, Pakistan reported more rainfall of apocalyptic dimensions, more than 800 fires were burning in Russia, a storm in Finland had cut off the power supply for 70,000 people, and at least 80 were killed and hundreds were missing following mudslides in China.
Wirtz diligently entered all of this information into her computer on the morning of Aug. 9. She is accustomed to the trials life has in store for human beings. She has been the head of the situation room at Munich Re, the world's largest reinsurance company, for 17 years, and in those 17 years she has repeatedly entered data that represent the catastrophes of the day into the company's databases.
An Encyclopedia of Life, Dangers, Justice and Coincidences
The databases include information about disasters that have already taken place as well as those that are just beginning or could occur in the future. They include data on every earthquake and every trembling of the earth's crust, on the height of ocean waves, air and water temperatures, and on the direction and speed of currents. Reports on glacier melting rates in the Himalayas and snowfall in the Arctic and Antarctic are also documented. New knowledge from the fields of nanotechnology, waste incineration, oil production, shipbuilding, reproduction and transplantation medicine is entered into Munich Re's computers. The databases also contain studies by the United States Centers for Disease Control and Prevention, the Chinese Health Ministry and tumor centers in Bavaria, as well as new information on piracy off the coast of Somalia, fluctuations in the European power grid and the worrisome eating habits of the Arab middle class in the Persian Gulf States.
An endless supply of data, probably unparalleled in its breadth and depth, flows from every continent to a cluster of buildings on the edge of the English Garden in Munich. An encyclopedia of life, its dangers, its injustices, its coincidences, is being assembled there. There is probably no other place on Earth where the risks of the modern world are being studied more intensively and comprehensively than at the headquarters of Munich Re, the world's risk center.
Munich Re insures insurance companies. It takes on risks that are too big for insurance giants like Germany's Allianz or Gothaer. Together with its subsidiaries, the company employs about 47,000 people on all continents, and more than a quarter of the world's population, or about 2 billion people, are indirectly insured through the company. The decisions these people make, the accidents they have, the circumstances of their birth and death, all of this information is transmitted to Munich, where data mining methods are used to examine the information, analyze it and constantly link it to other circumstances. The goal is to find patterns within chaos and probabilities in the improbable.
How great is the risk that a freighter accident in Germany's Midland Canal will cause a power outage in Italy? What might it cost to insure the entire supply chain of an international automobile manufacturer, a total of 4,000 companies scattered across all continents, against every conceivable delivery problem, from strikes to volcanic eruptions? These are the sorts of questions researchers at Munich Re address. Their task is to assess the risks as accurately as possible, because the level of risk determines how often a loss can occur, and the frequency of losses, or claims, determines the amount of the premium. For instance, if a given house is at risk of being flooded by a river once a year, the insurance premium will correspond to the value of the house.
These are the things that matter to Munich Re, whose researchers are ultimately engaged in a seemingly mad attempt to link everything with everything else, eventually, so that the company can then apply the supposedly impossible to the small print of an insurance policy and charge a premium for it. One way to describe this attempt begins with a die on a table. That's the way Heike Trilovszky does it.
Like A Game of Dice
Trilovszky is a mathematician and a senior executive at Munich Re. She has worked for the company for 23 years, and has looked after customers in North America and Eastern Europe while constantly climbing the corporate ladder to reach her current position. Her job title is "Senior Executive Manager, Corporate Underwriting," and she is responsible for the insurance policies that are concluded with other insurers.
Hers is a classic career within Munich Re, and Trilovszky is probably a classic representative of the company: intelligent, eloquent, thoughtful and as levelheaded as an accountant. Munich Re expects its employees to be detail-oriented and precise, which also applies to their interactions with one another and how they dress. After all, this is a company that sells certainty and reliability in an unreliable world.
The kinds of gamblers who typically work in stock markets are not to be found at Munich Re, which values reliability and competency and prefers staid over flashy. The hallways are filled with men wearing suits in every shade of gray and dark outfits with gold buttons. At Munich Re, the extroverts wear Joop designer cufflinks and oversized Tag Heuer wristwatches. Women like Trilovszky stick to classic business attire, especially when they meet with clients for negotiations or with journalists to explain how the business works.
Laws of Probability
"Determining the price of an insurance policy isn't all that different from evaluating a game of dice," says Trilovszky. Assume you have six euros, she explains. If you want to keep these six euros, no more and no less, how much do you have to bet on each roll of the dice? The answer is one euro, because the probability of rolling a specific number is one in six.
The actual ratio will likely be different once you start rolling the dice. But the more often one rolls the dice, the closer does the appearance of each number approximate a ratio of one to six. Munich Re's business is based on this law of probability. The more frequently a specific event occurs, the more accurately can the probability of its occurrence be calculated. Take, for example, an individual German's risk of dying within the next 365 days, which is 1 in 88. This means that one in 89 Germans will no longer be alive on Sept. 1, 2011.
Of course, rolling one die isn't enough to make a statement like this. A large number of dice has to be rolled an extremely large number of times to determine the probability of every possible cause of death. It gets even more complicated when the goal is not to make a statement about a large group like all Germans, but about individuals within this group, and about the probability that these individuals will die of a specific cause. In that case, one is no longer rolling a large number of dice on one table, but a large number of dice on as many tables as there are individuals -- around 80 million tables, in the case of the population of Germany.
And when the goal is not only to calculate the probability of one person dying of a specific cause, but also to determine the age at which this is likely to occur, things become indescribably complex. Nevertheless, it's still possible to make this and many other calculations. For example, the probability that a 40-year-old man will die of cancer in the next 10 years is 0.7 percent. It jumps to 2.5 percent if the man is 50, and to 3.7 percent if the person in question is a woman and she is 60.
What is the probability that Cologne's historic Old Town will be flooded a second time within the next year? Ten percent. Or that San Francisco will be hit by an earthquake this year? One percent. And what about the football World Cup in South Africa -- how likely was it that it would be cancelled because of a terrorist attack? That probability, according to the estimates by Munich Re's risk analysts, was also 1 percent.
Putting Theory into Practice
Those analysts are the people who put theory into practice at the global risk headquarters. Their job is to adapt the mathematicians' models to individual cases, to keep themselves informed about things like security strategies, the strength of fire walls, the performance of sprinkler systems and, if necessary, the stability of national governments. They view the world as a collection of risks and are constantly trying to quantify these risks.
If McDonald's wants to insure all of its restaurants worldwide, or Walmart all of its stores and warehouses, it is up to the analysts at Munich Re to examine the potential risks that confront every restaurant, every Walmart store and every warehouse.
Using geographical coordinates, they determine, for each building, the probability of earthquakes, major storms and floods. The entire world is stored in the databases at Munich Re in many different forms. Some data sets specify the probability of an earthquake occurring at any given point on Earth, while others contain detailed, site-specific data on the likelihood of floods and storms.
Other data can be used to draw conclusions on how storm-proof a given roof is, be it a flat roof, a pitched roof, a half-hip roof, a lean-to roof or a barrel roof. Every roof and every building is classified, and factors like mode of construction, materials, shape, height and width are entered into programs, evaluated and considered in relation to other factors, such as the type and quality of fire protection measures, the kind of product being stored, the manner of storage, the labeling of emergency escape routes and, finally, the human factor. Do the employees seem motivated? Is there trash lying around? Are tools put away at the end of the workday? How often is the building burglarized?
The answers are entered into databases. Deviations from the norm are calculated and taken into account. In the end, after everything has been weighted -- from fire protection to labor organization to the crime rate -- the premium appears. One premium. One premium for all of McDonald's restaurants or for Walmart's global empire.
An Authority on Fate
This is how the world's risks are dissected and evaluated at Munich Re. It is an ambitious attempt to provide a safety net for global capitalism, using the tools of the trade and without government intervention. It's an attempt that has already been underway at the company for the last 130 years.
Munich Re paid for the consequences of the San Francisco earthquake of 1906, for the sinking of the Titanic and the attack on the World Trade Center in New York. The sinking of the Deepwater Horizon oil rig in the Gulf of Mexico will find its place in the company's books, as will the earthquakes in Chile and Haiti and the flooding in Pakistan. There is hardly a disaster, natural or manmade, for which Munich Re is not called upon to pay claims.
Its expertise makes Munich Re an authority on strokes of fate and an arbitrator on one of the most important questions facing humanity today, climate change.
The Face of Munich Re
The man who delivers the verdicts on that issue is a professor with two doctorates and a penchant for funny ties. His name is Peter Höppe, and he is a physicist specializing in the evaluation of atmospheric processes. As head of Geo Risks Research at Munich Re, Höppe acts as the face of the company, even more so than its CEO. When a region or a country somewhere in the world is inundated by flood waters or ravaged by storms, Höppe invariably makes an appearance on television or in the press, and he is always answering the same question: Is this catastrophe evidence of climate change?
His answer is as follows: "It is difficult to attribute a single event to climate change. One has to look at the development over a longer period of time, and this development is clear. Our data indicate that climate change is already playing a role."
Höppe recites this information in a very levelheaded way, almost dispassionately. He lacks the urgency that characterizes many climatologists when they discuss this issue, and it is probably this approach that has placed him, a company spokesman, in the unusual position of arbitrating a dispute among scientists who sometimes forget to mention that they don't have all the facts yet. Höppe doesn't make this mistake. He is quick to point out where there are gaps in his knowledge and where the attempt to measure risk has failed, perhaps because interactions were not taken into account, underlying conditions changed more quickly than assumed or the impossible became possible, after all.
The Limits of Their Abilities
The 9/11 terrorist attacks in New York are a case in point. The mathematicians at Munich Re had calculated the cost of a fire on 10 floors of one of the Twin Towers caused by an aircraft crashing into the tower. This seemed possible, though not very likely. But it seemed impossible that two aircraft with full fuel tanks would be intentionally slammed into both towers, resulting in the total collapse of the two buildings.
They also calculated the cost of a direct hit on New Orleans by a hurricane. Munich Re was also aware of the consequences of major flooding in the city, which were reflected in the pricing of insurance policies for real estate in New Orleans. But what the company's mathematicians did not calculate was the cost of a hurricane hitting the city and the levies failing at the same time.
Insurers refer to events like this as accumulations. They hate accumulations, because they are the catastrophes that reveal the limits of their abilities and spoil the bottom line.
Possible and Impossible Risks
To discover accumulations, despite all of the challenges, and to protect its bottom line, Munich Re employs, in addition to mathematicians, risk analysts and scientists, another group consisting of people like Christian Bendel. Bendel, an engineer, is what one would call a foot soldier at Munich Re. His job is to come to grips with the impossible through experience and visual inspection.
Bendel has a background in construction. He has built bridges in Croatia and the Philippines, and these days he spends much of his time traveling in Africa and the Gulf States. He doesn't look like he gets much sleep.
Bendel's task is to gather on-site impressions of possible and impossible risks. He uses empirical methods and visual inspection. Bendel knows how to work with concrete on the top floor of an 800-meter (2,624-foot) building, and he also knows that construction workers have a tendency to connect sprinkler systems at the end of a project, which explains why so many buildings have been destroyed by fire shortly before completion.
One of Bendel's colleagues is extremely familiar with the world's computer chip factories, another specializes in fish farms in Latin America and a third is practically at home on oil rigs. There are specialists in tunnel building, deep drilling, the construction of nuclear power plants and new developments in heart transplantation. In the age of outsourcing, employing an expert on fish farms in Latin America seems like an anachronism. But Munich Re is willing to pay for this luxury, because management believes that this is the only way to detect remote but potentially costly risks.
For example, in 1973, 37 years ago, this approach enabled Munich Re to offer an early warning against the risks of possible climate change. Or in 2008, two years before the eruption of Iceland's Eyjafjallajökull volcano, an employee completed an analysis of the possible effects of a volcanic eruption on international aviation.
Keeping Track of the World's Risks
What all of these experts have in common is the bottom-up, inductive method with which they approach the unknown, which means drawing general conclusions on the basis of isolated incidents. But the opposite approach also works, at least most of the time.
As the "chief risk officer" at Munich Re, Joachim Oechslin's job is to maintain an overview of the world's risks for the company and to ensure that its employees' understanding and assessment of these risks is constantly improving. A financial mathematician, he applies general data to individual scenarios. At 40, Oechslin, a Swiss national and amateur mountain climber, is surprisingly young to be in such a senior position at a company like Munich Re.
Oechslin is particularly interested in a very specific type of risk that is as important to his employer's bottom line as accumulations. He is partly responsible for Munich Re AG's financial investments and the company's overall investment strategy. A major presence on Germany's DAX stock index of blue chip companies, Munich Re's investment portfolio is worth €180 billion ($229 billion).
Oechslin knows that he can rely on his data and his computers, and that the scenarios they spit out are among the world's most sophisticated. He knows that the combined computing power of Munich Re is used to support him in his work. The company analyzes the daily fluctuations on the world's stock markets as intensively as it does the readings on seismographs. But Oechslin also knows that people are the key factor. A person calibrates the programs, determines which risks are assessed in what way, what is and isn't acceptable, and decides which risk levels are acceptable when making investments. But if the people performing these tasks are too greedy, all of the programs and scenarios become worthless.
'We Learned Our Lessons from the 2001 Losses'
Munich Re made this painful experience itself at the end of the New Economy boom. At the time, about 20 percent of its financial investments were in stocks, making the crash an expensive -- and humiliating -- proposition. The global risk center had allowed itself to become caught up in an atmosphere of greed and, like most others, had believed that the boom would never end and that the bubble would continue to grow forever. The company's professional pessimists had suddenly mutated into optimists, even though they, of all people, should have known better.
But then, when the heyday of mathematicians in the financial world began, when the risks posed by obscure credit securities were downplayed, and when bankers around the globe relying too heavily on computers ended up taking dangerous risks, the pessimists at Munich Re decided to hold onto their billions and not to blow their investments on fly-by-night credit bets. When they analyzed many of the obscure new financial instruments that investment banks were touting as low-risk, high-return investments, the collateralized debt obligations (CDOs) and the credit default swaps (CDSs), they generally concluded that the risks were in fact too high.
When these securities fell apart like bad checks, and when the global economy slid into recession and many companies posted losses, Munich Re reported a profit of about €1.6 billion, which increased to €2.5 billion in 2009.
There were scenarios in Oechslin's department that predicted an abrupt end to the boom, but it wasn't the scenarios that were primarily responsible for the profits of the past year. It was something else. "We had simply learned our lessons from the 2001 losses," says Oechslin. After the end of the Internet hysteria, Munich Re drastically reduced the percentage of its capital invested in stocks, from 20 to 2.5 percent, while at the same time increasing the share of fixed-interest government bonds to about 40 percent. This reduced the risk, as well as the opportunity to achieve massive returns during the boom years.
It was not a victory of mathematics or of data gathering, but one of reason in an era of greed, an admission of one's own limits. At the time, the hot shots at the investment banks amused themselves to no end over their overly conservative counterparts at Munich Re. Today Munich Re wins accolades for its restraint, while its shareholders are eagerly awaiting the results of a new project. The goal of the project under development in Oechslin's department, more comprehensive than any other project before it, is to redefine the limits of knowledge by developing a global risk model. In addition to specifying the major risks in the world, it also seeks to identify their interactions and make them predictable.
Work on the project has been underway for more than two years, and results are expected next year. When that happens, Oechslin and his team hope that life on Earth will become a little more predictable.