The PayPal Gang Summit: Big Data, Research, Credit Ratings, and Cybersecurity

In the United States, the PayPal gang is busy ramping up the system to a new and stratospheric level of super-powerful data processing of everything that moves. They are working for the government, hedge funds, large banks, and consumer companies. A leading-edge player here—and probably the best example of the future—is Palantir. Founded by Peter Thiel several years ago, Palantir has had explosive growth in the past three years. The initial funding for Palantir came from the CIA. The CIA's venture capital firm is called In-Q-Tel, and it was one of the angel investors. Palantir provides software that can search, cross-reference, and interpret large amounts of data from many sources. Palantir's software has been used to detect fraud and insider trading by law enforcement agencies and banks. Its clients include U.S. government agencies such as the CIA and FBI, Bank of America, J.P. Morgan Chase, and News Corp. In 2013, it had estimated revenues of US$450 million. The current value of Palantir is estimated to be US$9 billion. (Let's also keep in mind that Amazon has a US$600 million contract with the CIA, presumably for international sales, since it is illegal for the CIA to operate domestically.)

Having started as an entity that was funded by the government, Palantir now has 60 percent of its revenues from the private sector. It was surmised that the financial impropriety at HSBC with drug cartels in Mexico was discovered by way of the CIA. It would not be a stretch to think that Palantir might have had a role to play in this investigation.

Palantir's reach is extensive, to say the least. The company now has a right and a left hand. The right hand is Palantir Gotham. This entity analyzes structured and unstructured data (for example, data found in e-mails, news reports, books, and websites) into a single information model providing relationships that can be understood by nontechnical staff. It enables users to Palantir to search multiple large datasets simultaneously to identify relationships. Gotham's clients include U.S. government agencies the CIA and FBI, who use the software to seek patterns in large amounts of data to track terrorists, drug trafficking, and cybercrime.

The left hand is Palantir Metropolis. This entity is the financial services arm and provides powerful quantitative financial analysis software. This software distributes bank data in a centralized fashion to technical and nontechnical users. Customers include banks and hedge funds. As an example, Steve Cohen at Point72 Asset Management (formerly known as SAC capital) hired Palantir to assist in its compliance and surveillance. It has an “Unauthorized Trading” algorithm that assigns riskiness scores to traders by examining correlations between key risk indicators in the context of overall trading activity. Citi uses Palantir Capital Market software to merge proprietary and vendor data into one platform for equity analysis.

Palantir is a cutting-edge company that is taking financial software and big data to new heights. It helps banks manage internal risk. It helps law enforcement monitor the potential illegal activity of financial institutions with regard to violations of money laundering and antiterrorism. It gathers data for hedge funds to help to discover previously undiscovered correlations among social, economic, and commercial phenomena.

It is almost amusing that top executives at Palantir admit that they have trouble articulating what exactly the company does. They say that it “extracts insight from information.” They respond to human-driven queries. They can summarize large datasets, and they visualize datasets by articulating what is going on inside the data. Palantir lauds itself for being able to do the work quickly.

One example of a company that used large data to discover how consumers behave is the following. One company was asked to analyze all of the purchases at Walmart (tens of millions) and find the one common purchase of those that had the best credit rating. Interestingly, the company discovered that people who buy door-stops had the best credit rating. What was the most common purchase of those with the worst credit rating? The deadbeats most often bought mouth restraints for dogs. Is this a matter of mixing up correlation with causation? Maybe, but the experts in big data say without batting an eyelid that when you have millions of data points offering you some fairly sound and compelling correlations, it is folly to ignore these conclusions. In their book called Big Data, Kictor Mayer-Schonberger and Kenneth Cukier make the point that correlations can be found faster than causation and correlation is backed up by millions of data points. He makes the point that we still need controlled experiments with carefully handled data. For everyday needs, however, knowing what—not why—is good enough. In addition, big data correlations can be a harbinger of promising areas in which to explore causal relationships.^*

The above is an excellent example of the kind of work that Palantir might do. Also mentioned in Jaron Lanier's book Who Owns the Future? are examples of entities that search the globe thousands of times per second for the lowest price for a certain book. A company like Amazon will use this price (whether it is in Belgium, Biloxi, Beijing, or Bangalore) as a new benchmark and beat the previous price. This is known as a bot. This bot is an algorithmic program that searches for the lowest prices everywhere and at all times. This world of the bot is one where great fortunes are made by driving prices down rather than up. It drives out the middleman and forces out the marginal producer. The curse of humans is that they love getting good deals. (Amazon is an example of a bot gone haywire. It is constantly driving down prices so fast that it seems that this bot is devouring the shareholders of Amazon. The company can't manage to make money after all these years!)

These bots will always only ever get the best deal possible (priceline.com lives by this) but this means that the bot will drive the price of one industrial or social good after the other to something close to free. This is designed to push all prices to the marginal cost of production. So, we must stay close to this powerful server, as it is the arbiter of prices. This is a super-deflationary phenomenon. It learns all it can about our behavior and then it offers all of the services we create at a price that is the closest to free that can be possibly be achieved. If we do not stay near the server and constantly improve innovation and imagination, we are out of a job. Hence the vital importance of companies such as Palantir, Google, Alibaba, and Amazon. Imagine what will happen as these companies hone their skills better in the world of lending. They will eat inefficient and high-cost banks alive!

Credit card companies, new financial technology, and innovative software by companies like Moody's Analytics are also joining this game. They will crunch millions of data points to find good creditors who are down on their luck and need a short-term loan, for instance. They will eventually squeeze out inefficient and high-cost banking entities. One interesting example of a type of high-quality credit is a woman who is almost finished with her nursing degree and needs a loan. She fits a certain kind of criteria of an excellent credit and will be given a loan in less than a day. Companies have created algorithms to track down women like this and offer them loans. The level of detail about who is most apt to pay back a loan from myriad details of our private lives is astounding. Slight changes in our habits are picked up by some of the most sensitive bots out there. These bots can detect divorce, pregnancy, bankruptcy, new additions in the family, changes in health, changes in mood, and other intensely personal details with startling accuracy.

In addition, Palantir has access to all open information platforms of the U.S. government: health, safety, traffic, budgetary items. It is safe to say that Palantir's systems are firmly embedded into the defense, counterterror, intelligence, and law enforcement establishments. It is emphatic that it abides by all civil liberties protections mandated by the federal government. It is safe to say that Palantir's network is something of a foundation for how data collection, integration, and analysis can help financial institutions in the future.

Another competitor is Kaggle. Kaggle has had PayPal veteran Max Levchin as its chairman since 2011. It is a platform that creates competitions for predictive modeling and analytics. Data miners compete to produce the best models for data posted by companies and researchers. Kaggle charges clients a fixed fee and offers monetary rewards to data miners who seek to answer the human queries of Kaggle's customers. In the past few years, Kaggle has paid out more than US$5 million in “rewards” to data scientists who have helped its customers crack a thorny issue or discover important patterns in certain behaviors. Clients include GE, Microsoft, NASA, and Tencent.

Companies like Palantir and Kaggle are important for all financial institutions because they can breathe life into the imaginations of IT teams at banks about how to data mine consumer behavior (e.g., professional development, changing tastes, patterns of purchasing door-stops, etc.) and find better ways to safely and prudently lend money to those who are most likely to repay the money and also to avoid the deadbeats. It's a very simple procedure. And so much of this is automatic. Furthermore, software is getting cheaper by the day, and there is more information available on the Internet on patterns of individual and corporate financial behavior. Banks who do not aggressively embrace this trend are dead in the long run.

I am not saying we are entering into a new Panglossian world of an end to credit problems, thieves, and con artists. We live in a world of alligators and boa constrictors in all walks of life. I am saying that this world in which we have extremely powerful tools to examine and parse data points to more efficiently allocate credit is very likely a better world than the one dominated by banks that time and again have shown their ineptitude when it comes to allocating capital and managing risk.

I am saying that the hardest nut to crack in the world of credit has been the small and medium enterprises that have traditionally been locked out of the credit world. Is there a reason for this? Are small and medium enterprises composed of perpetual liars and thieves? Of course not. Most work very hard to run family businesses and are honest people. But banks have traditionally not had the economies of scale to make money on this sector when they can make better money on big-ticket, high-margin, low-touch loans to large corporations and governments.

In a world of immense data that can be easily managed among and between the banks and the company, there are many ways to open up credit to hundreds of thousands of SMEs. If an SME is a true partner to a bank and offers its data on working capital, inventories, tax, and payroll, the bank can offer not only more credit but also better pricing for this credit. This is a revolution right in front of our eyes. And it is real. Just ask people who bank with Commonwealth bank of Australia. This bank is making this a reality now. And companies tell CBA that the software offered by CBA helps them to know their company better.

One last comment. There is a predilection for some strange reason to pooh-pooh the idea of financial technology as a scam or a passing phase—some adolescent idea that will fade away. I think this is a misplaced notion. I sense that there is powerful momentum and a kind of irreversibility about this. I strongly believe this will only grow larger. And regulators, users, governments, and banks need to get their heads around this sooner rather than later. To ignore this trend as a passing phase is foolish and costly.

Alibaba's Cloud Business: The Future of Banking

We go to the other side of the world and see that, in China, Alibaba is doing precisely this. It is aggressively morphing from a company that is a combination of PayPal, eBay, and Amazon into a company more like Palantir and Google. Alibaba is achieving this through a number of smart moves that may create one of the most interesting hybrid companies the world has ever seen. Alibaba not only looks like Amazon but also has shades of Palantir and resembles the entertainment element of Disney. So, Alibaba is morphing into a four-headed creature in e-commerce, entertainment, banking, and information analysis.

The bread-and-butter of Alibaba is the equivalent of Amazon and eBay. Alipay is a separate entity but is similar to PayPal. It is likely to list inside China and have an H share listing in Hong Kong in 2015. Like PayPal and eBay, it is probably wise to have a separate listing. This will allow it to have a higher valuation, and they are in fact different businesses and belong apart.

Alibaba is gluing other interesting entities onto this framework that revolve around consumer behavior: driving, dating, learning, fun, languages, lifestyle, and buying goods that can be mailed by Alibaba to the home. Figure 12.2 shows how Alibaba is becoming a lifestyle company. This diagram shows how Alibaba is already like Google, Dropbox, eHarmony, Amazon, Twitter, Spotify, Orbitz, Uber, and ING Direct.

AliCloud and Hundsun: The Mother Lode of All Financial Data

There is something else going on with Alibaba, and the other part of this entity that is less understood is more like Google and Palantir and less like eBay and Amazon. Alibaba has built its own cloud server from scratch and called it Aliyun. This is very important, because Alibaba decided to build all of its infrastructure internally with no outside entities. This is referred to as a “No IOE” policy. In other words, the technology that Alibaba used to build the infrastructure is not from Intel. It is not from Oracle. And it is not from EMC. It is, indeed, indigenous technology built internally.

There is an element of paranoia here—justifiable paranoia, that is. The Snowden leaks of NSA material showed that everything is open to the U.S. government.² The NSA reads just about anything and has cracked any code there is. Governments from China to Germany are now aware of this. So, countries like China that are overcoming a century of humiliating occupation (at one point in the early 1900s, China was occupied by 12 different countries) are flexing their nationalist muscles and saying they do not want any cyberimperialism. This is understandable and expected. China has a very capable spying service, which is very likely on a par with the NSA. It wants to preserve its secrets from the prying eyes of the U.S. government. So, it has concluded that it will build its own technological architecture. Furthermore, we shall see that the phenomenal technological firepower achieved by the likes of Palantir has scared China into creating its own equivalent.

Since its establishment only a few years ago, AliCloud (Aliyun) has more than one million customers and it has generated almost $100 million in the three quarters to December 2013. Cloud computing and Internet infrastructure are a powerful combination. See Figure 12.3. With Taobao and Tmall, Alibaba has access to the spending habits and price-points of hundreds of millions of Chinese people. With Alipay, Alibaba has access to the credit histories of hundreds of millions of people in literally thousands of cities across the country. With the dating service called Momo, Alibaba has access to preferences and demographic information of 140 million adults in nearly every city in the country. And with Weibo and Tango, Alibaba has access to a social network that makes Facebook pale in comparison.

Inside China, the cloud market is growing at more than 40 percent per annum and now accounts for only 3 percent of the global cloud market. In this space, Alibaba has a commanding share. Tencent, Shanda, and Baidu are all competing in this market. But the real competition here comes from Huawei. Huawei is the gargantuan technology company founded by a PLA Army colonel. Think of IBM, GE, and Apple all in one company! Huawei is the real competition for Alibaba in this space. Time will tell just where this competition goes. So far, the software infrastructure is 70 percent of the market. Infrastructure is about 20 percent of the market. In the software sector, Alibaba is the dominant player. In effect, then, Alibaba commands the field in the first inning of this buildout. Estimates show that this business will be close to US$1 billion by 2018.

AliCloud's customers are remarkably diverse. Customers include companies in pharmaceuticals, utilities, government security, tourism, weather trends, near-field communication mobile support, gas distribution, telecom, and IT solutions. It is basically the entire economy located inside the AliCloud. And this is only three years old. Imagine what will come next! (See Figure 12.4.)

What's next is Sesame! Sesame is Alibaba's new credit rating service for consumers. In the United States, 85 percent of people have some form of credit rating through a Social Security card, credit cards, phone data, sensor data, or browsing data. In China it is only about 25 percent. This is only 350 million. Alibaba's ambition is to have reliable credit ratings on all Chinese over the age of 14 fairly soon, so even youngsters can have cellphone accounts or credit cards for school or other purposes. The numbers being thrown around so far are in the neighborhood of 900 million people who could conceivably have reliable credit ratings and be able to get credit. So, a company like Alibaba (whose market share in phone e-commerce is more than 80% and whose market share in peer-to-peer e-commerce is 90%) can realistically hope to achieve these ambitions. Imagine a world where a company like Alibaba or Tencent could have reliable credit data on 15 percent of the earth's population. This is precisely why Alibaba needs AliCloud. In this way, we should conclude that Alibaba will become the world's largest ratings agency within three years.

Hundsun: A Vast Array of Information on Financial Services

Now we arrive at the mother of all information lodes. In a brilliant move, Jack Ma bought a controlling interest in a company called Hundsun. Hundsun was developed by 12 software engineers in the 1990s. The 12 sold their 22 percent interest to Jack Ma's personal company Zhejiang Rongxin for US$522 million. Because this investment involves sensitive technological infrastructure for the banking sector, banking laws prohibit foreign ownership; this is why Hundsun lies in the private company rather than inside the listed entity. The remaining 78 percent of the company is listed on the Shanghai Stock Exchange and its ticker is 600570. The Ministry of Commerce sanctioned the deal in late 2014, so the relationship has been cemented.

Hundsun is a unique company. Imagine a company that is like IBM and Cisco and that has built the infrastructure for PIMCO, Fidelity, J.P. Morgan, Prudential, and Bank of America Merrill Lynch. Presumably a good portion of this information will end up being stored in Alibaba's AliCloud. If we combine the industries for which Hundsun has built the backbone, it runs the full gambit of financial services. These include securities, banking, asset management, insurance, treasury management, and exchange solutions. And it also includes the infrastructure for the new derivatives markets in China, including futures and options markets that were just opened in early 2015.

Hundsun also has a great deal to do with internal security for China. For instance, like Palantir, it also is involved with anti–money laundering, internal compliance platforms for banks, and internal controls for trading. But it also serves functions akin to exchanges like NASDAQ. Hundsun has built the backbone for the futures and options exchanges to be launched in 2015. It also operates systems that manage fund sales, margin accounts, and security exchange platforms. As a result of the integration of these tools in the Alibaba arsenal, Table 12.2 shows the raw data that can be analyzed inside the AliCloud to discover where people shop, how they drive, what they buy, how much they want to spend, where they travel, how they insure themselves, when they will buy a house, how they want to entertain themselves, and what they want to learn about and read.

There really is no one company like it in the world. This company can conduct human queries on just about anything. How many insurance policies are there? How many miles did Chinese people drive last year, and where? What kinds of movies do people want to see? Where do people want to go on vacation? What kind of analysis do investors in equities and fixed income like to use? How do Chinese construct portfolios for their future? Who needs to get rid of certain inventories? Who is a good credit and needs working capital for 90 days? In what way will Chinese use the futures and options markets? This is a unique and massive amount of data on the largest population in the world, and it is generally closed. But this is just getting started! Perhaps most important of all: What if companies like Alibaba, Palantir, and Tencent are able to more accurately predict GDP trends better than any government entity?

Think about this further. Look at Figure 12.5. It shows the power of the information that is part of the Hundsun infrastructure built over the past several years, which conceivably (and in all likelihood is part of Alibaba's cloud) will now be able to be mined by Alibaba entities. If one sees corporate transactions going on live, it is possible to get a good real-time sense of where GDP is going. It is possible to get an up-close-and-personal sense of consumption trends, entertainment likes and dislikes, and fund flows within the economy. It will be possible for government entities to track corrupt officials and their cash movements. It will be able to seek out money laundering. It will be able to have a reliable sense of funds flows among and between various asset classes. The opportunities for data analysis (for better and for worse) are infinite.