So, at this point, Bilk'em is not only a world-renowned philanthropist, he has also achieved political diversification of his assets. Naturally the Children's Fund person who authorized the transfer to Geneva is not going to talk, because that would cut off donations to the Children's Fund and she would also lose her well-paying, cushy job. . .

How to Launder Money in the Futures Market

A futures contract is a bet on the direction of price movement of some underlying commodity, financial asset, or index. The contract might involve crude oil, French government bonds, or the level of the S&P 500 stock index. The party that is "long" the futures contract is betting that the price will go up. The party that is "short" a futures contract is betting that the price will go down. If the price moves, there will be, with respect to the futures contract itself, a winner and a loser. If the price of whatever underlies the futures contract goes up, the long side will win, and the short side will lose. If the price goes down, the short side will win and the long side will lose. These losses are incurred every business day the contract is in existence.

For example, at the New York Mercantile Exchange (NYMEX), the crude oil futures contract is based on the price of West Texas Intermediate Crude. Each futures contract represents 1000 barrels of oil. So if the price of crude goes from $22.50 per barrel to $23.00, the long side of the futures contract will gain $(23.00 - 22.50) x 1000 = $500. The short side will lose $500. As another example, each live cattle futures contract at the Chicago Mercantile Exchange (CME) represents 40,000 lb. of cattle. So if the price falls from 64 cents to 63 cents, the short side of a single futures contract receives $(.64 - .63) x 40,000 = $400, while the long side loses the same amount.

The key way to use futures contracts to launder money is to arrange things so that the losing side of the contract involves "dirty" money. Michael Sindona, who ought to know, defined money laundering by the statement that "laundering money is to switch the black money or dirty money . . . to clean money." That definition is somewhat circular, but you get the idea. "Dirty" money is money that comes from illegal, or questionable, or unknown activities. Or it may just be ordinary money that one wants to make disappear. "Clean" money, on the other hand, comes from known and legal sources. For example, money earned from a futures contract is "clean" money: its source is well- defined, legal, and known. The terms "dirty" and "clean" are somewhat archaic. Money currently exists mostly in the form of electronic tokens: a digital series of ones and zeroes in bank computers. In this context, dirty money can be thought of as "naughty bits."

For example, in the late 1970s, Hillary Clinton made just shy of $100,000 in the cattle futures market. The source of the money was known: it was clean money collected in her account at the CME. She read the Wall Street Journal and rode the upward wave of cattle prices, she said.  (Of course, there is a problem with this explanation. An analysis of her trades shows that most of her profits were made by going short cattle futures. That is, most of her $100,000 was made in periods when cattle prices were temporarily falling during this period of overall rising prices. Tyson Foods' attorney Jim Blair apparently forgot to inform Hillary what trades actually took place in her futures account)

Now the obvious question arises: how do you arrange to get the price direction right? If prices fall, the short side will gain money and the long side will lose, so you want the long (losing) side to possess the dirty money. If prices are rising, you want the dirty money to be on the short side. Suppose you expect the price of the notional bond contract at the Paris MATIF to fall, so you arrange for the dirty money futures account to be long futures contracts. You also open another "clean" money futures account in another name to be short futures contracts. You then wait for prices to fall, expecting the dirty money to disappear into the black hole of futures losses, while clean money comes pouring into the "clean" account. But, instead, to your surprise, prices rise! The dirty money account thus collects more cash, while clean money is sucked out of the other account. It's a laundry in reverse: clean clothes in, dirty clothes out.

To prevent a problem like this, it helps to have the cooperation of a helpful futures broker. Let's call this broker-facilitator-prestidigitator by the name of Merlin Lynch. Merlin Lynch will create a long futures position to match the position of the party that is short futures, and a short futures position to match the party that is long futures. Merlin Lynch will, of course, extract a profit for itself -- let's say in the form of a nice, fat round-turn commission of $13 per contract.

Let's say the dirty money is held in an account under the name of Dirty Dick. It is intended that clean money be accumulated into an account under the name of Clean Jane. Now there will be price fluctuations during almost all trading days, so Merlin Lynch makes matched trades at two materially different prices during the day. Let's say Merlin has customer money that needs laundering, and decides to do it using the copper futures contract at the COMEX. At a random time in the morning, Merlin Lynch enters the market and buys (goes long) 100 contracts, say at a price of 94.20 cents per lb., while simultaneously selling (going short) 100 contracts, say at a price of 94.10 cents per lb. The price of 94.10 represents the market's bid price, the price at which the market will purchase the futures contracts which Merlin is selling. The price of 94.20 represents the market's asked price, the price at which the market will sell to Merlin the futures contracts Merlin is buying.

Later, at a random time shortly before the close of trading, Merlin Lynch sells (goes short) the 100 contracts, say at a price of 92.30 cents per pound, and simultaneously purchases (goes long) 100 contracts at 92.40 cents per pound. Now Merlin is ready to launder some money. Since the price has fallen, Merlin assigns the purchase of 100 contracts at 94.20 to the Dirty Dick account, along with the sale of 100 contracts at 92.30. Since each COMEX futures contract represents 25,000 lbs. of copper, the net loss to the Dirty Dick account is $(.9420 - .9230) x 25,000 x 100 = $47,500.

Meanwhile, Merlin assigns the sale of 100 contracts at 94.10 to the Clean Jane account, along with the purchase of 100 contracts at 92.40. The net gain to the Clean Jane account is $(.9410 - .9240) x 25,000 x 100 = $42,500. The $5,000 difference is the amount paid to the market for the laundering service. In addition, Merlin Lynch gets a brokerage fee of $13 x 100 x 2 = $2600, for a net cost of $7,600 to launder $47,500. This amounts to a laundering cost percentage of 16 percent. At these prices, the process is not terribly efficient, but workable in some cases. The natural evolution of the market will be to create more efficient, less costly laundering structures. This can be done in the current example by squeezing commission costs and bid-asked spreads.

To avoid raising eyebrows, the Dirty Dick account should be held by a party apparently unaffiliated with the holder of the Clean Jane account. For example, the Dirty Dick account could be held by Giant Copper Traders, Ltd., while the Clean Jane account is held by Uninsulated Wire Company. The clean cash accumulates to the benefit of Uninsulated Wire Company, who returns the favor by paying premium (above average) prices to Giant Copper Traders for supplies of copper. Uninsulated Wire Company, naturally, will want to gain something for its role in assisting the laundry process, so the premium price they pay for copper should not eat up the entire amount of its futures profit. This will increase Giant Copper Traders' cost of laundering funds.

Of course, if your intent is simply to bribe some politician, then the Dirty Dick account should be assigned to the company or other party paying the bribe, while the Clean Jane account should be assigned to the politician receiving the bribe (or to one of his trusted friends or relatives).

Now. Suppose you believe that all price fluctuations are random, and all people (other than, perhaps, the party with funds to launder) are honest. How do you arrange a structure so that stochastic price fluctuations will accomplish the laundry? No witting brokers, and no trades at artificial prices, are allowed. (This problem is left as an exercise for the student.)

Money Laundering Through CHIPS and SWIFT

Despite the recent hype given to electronic banking and electronic cash, most of the monetary system in the U.S. has been electronic since 1970. That's when the New York Clearing House began operation of the CHIPS system.

CHIPS stands for Clearing House Interbank Payments System, and is the primary site for settlement of U.S. international foreign exchange and eurodollar transactions, as well as international and domestic trade transactions taking place through New York banks, New York Edge Act subsidiaries of regional banks, New York branches of foreign banks, and Article XII investment companies.

CHIPS is an example of a Automated Clearing House. Clearing houses were created to simplify the paperwork and the process of banks making payments to each other.

In the old days when banks settled their claims on each other by writing checks, messengers were dispatched to deliver the checks and associated documentation from one bank to another. Now, if each of 10 banks sent messengers to each of the other banks, this would require 90 messengers. And even if they more efficiently only had one messenger that carried checks between any pair of banks, it would still require 45 messengers. To avoid this, the simple solution was for each bank to send a messenger to a central location at periodic times of the day (three times, in the case of the New York Clearing House). At that location each messenger would deliver checks directed to any of the other banks to the respective messengers of those banks. Only 10 messengers were required.

This central meeting place was the basis of the clearing house. The next step was to eliminate offsetting transactions. If the Citibank messenger had a $1 million check payable to Chase, and the Chase messenger had a $1 million check payable to Citibank, the two payments would cancel out. The simple solution is to tear up both checks. Or, more realistically, to record both checks without any payment actually taking place between the two banks. When this is done electronically, the result is called "netting".

The final step was to eliminate the messengers and replace them with telephone wires or fiber optic cables, and install computers to keep track of the payments and to automatically net transactions. This brings us to the present CHIPS system.

If, in the CHIPS system, Citibank pays $100 million to Chase, and Chase pays $250 million to Citibank, the net effect is a payment of $150 million from Chase to Citibank. So, in the CHIPS settlement procedure, the individual amounts that make up this $150 million net payment are recorded, but all that actually takes place is a single transfer of $150 million in reserves from Chase to Citibank in the account that CHIPS holds at the New York branch of the Federal Reserve. This account is funded (via Fedwire) by the banks making net positive payments, and depleted by the banks receiving reserve payments. At the end of each day, the CHIPS account at the Fed is brought to a zero balance.

Instead of there being three meeting times a day, payments may be entered into the CHIPS computer throughout the day. The CHIPS computer keeps a running net position for each bank in the system throughout the day, and makes a single net settlement at the end of the day.

Suppose a CHIPS member bank makes a payment to another bank. A terminal operator at the paying bank enters the relevant message information for a CHIPS payment transfer into one of its own computers--the one that interfaces with CHIPS. The bank interface computer is connected to each of the two CHIPS operating centers by separately routed fiber optic cables. At the operating center, the CHIPS central computer authenticates and stores the message, and sends a "store" acknowledgment back to the bank. The bank approves or rejects the stored message. If approved, the bank releases the payment transaction to CHIPS.

CHIPS then compares the payment transfer against variables it uses for risk control. These include bilateral credit limits, net debit caps (the maximum net total that a bank can pay to all other banks), and other information. If the payment transfer passes these tests, then CHIPS notifies the receiving bank (sends a "receive" message to that bank), and credits the account of the receiving bank (increases its running total of net receipts) and debits the account of the paying bank (decreases its running total of net receipts). This payment is irrevocable, and will be settled at the end of the day.

There are 105 banks that participate in CHIPS, but only 20 of them are "settling" banks, which are allowed to make the Fedwire transfers that settle all net debits or credits in the CHIPS system. These 20 banks include the 10 members of the New York Clearing House (which own the CHIPS system), as well as 10 others (the member banks of the New York Clearing House are Bank of New York, Chase Manhattan, Citibank, Morgan Guaranty, Bankers Trust, Marine Midland, United States Trust, Fleet Bank, European American Bank, and Republic National Bank of New York).

The 85 banks who are not settling banks must settle their net transactions with one of the 20 settling banks. The CHIPS network closes each day at 4:30 p.m. EST. The CHIPS computer then produces a report that shows the net debit or credit position of each of the 105 participants. There is a second report which shows the net settlement required of each of the 20 settling participants; in this report the net settlement position for each of the 85 non-settling participants is folded into the corresponding report of its settling participant representative.

At 5:30 p.m. EST the settling banks who owe a net payment (have a net debit position) pay their net obligation into the CHIPS escrow account at the Federal Reserve Bank of New York via Fedwire. CHIPS then uses these funds to pay the settling banks that have net credit positions, leaving behind a zero balance in the escrow account at the Fed. This process is completed by about 5:45 p.m. EST.

In 1996 there were an average of 213,105 CHIPS transactions per day, for an average daily dollar volume of $1.32 trillion. Since the U.S. M1 money supply averaged about $1.1 trillion in 1996, this means each dollar turned over 1.2 times a day through the CHIPS system alone. (A recent daily record was set on January 21, 1997, when CHIPS cleared 418,743 transactions totaling $2.178 trillion.)

The various debt and credit limits checked by the CHIPS computer become important if there is a failure of a bank which owes a net payment. This can be seen in the case of European American Bank--one of the 10 banks that currently make up the New York Clearing House. Back in the early 1970s, European American Bank was named Franklin National, and was effectively controlled by Laurence Tisch, the man who would later become CEO of CBS. The Long Island-based Franklin National was the 18th largest commercial bank in the U.S., and eventually went belly up due to foreign exchange trading losses in May 1974.

By 1971 Tisch had, through Loews corporation, acquired more than a 20 percent stake in Franklin New York Corp., the parent company of Franklin National. While not quite enough to get Loews declared a bank holding company by the Federal Reserve, it was enough to make them ask questions about Tisch's effective control of the bank. Meanwhile, bank examiners found more than 10 percent of the bank's loans--representing 90 percent of the bank's capital--in danger of default, so Tisch decided to dump the problem on someone else. He negotiated to sell his shares to the Vatican banker and renowned international money-launderer Michele Sindona for about $40 million. Sindona, the bank's chairman Harold Gleason, and the bank's president Paul Luftig (whom Tisch brought in from Banker's Trust right before selling out) would all go to jail over the subsequent failure, while Tisch would only pay $1.2 million to settle related lawsuits.

Under the current CHIPS system, each of the other settling banks have to take a pro rata share of any losses from failure of a participant to settle. For this purpose, they have to pledge collateral in the form of book-entry Treasury securities. Book-entry Treasury securities, like bank reserves, are transferred over Fedwire.

When a foreign bank settles its dollar transactions from foreign exchange or eurodollar trading through CHIPS, it does so through a New York subsidiary or a New York correspondent bank. It sends a payment instruction to the New York bank, which then transfers the instruction to the CHIPS system. The primary system for sending international payment messages is operated by SWIFT--the Society for Worldwide Interbank Financial Telecommunication. SWIFT is headquartered in Brussels, Belgium, and has global routing computers located in Brussels, Amsterdam, and Culpeper, Virginia.

If Bank of Montreal in Canada sells deutschmarks to Hong Kong Shanghai in London for dollars, Bank of Montreal will turn over to Hong Kong Shanghai a deutschemark deposit at some bank in Germany, while Hong Kong Shanghai will turn over to Bank of Montreal a dollar deposit at a bank in New York. The confirmation of the trade details, and payment instructions to the banks in Germany and New York, will be sent out over the SWIFT messaging system.

Meanwhile, Bank of Montreal will have a bank account in Germany in which it holds deutschmarks, while Hong Kong Shanghai will have an account in New York at which it holds dollars. These foreign bank accounts may be held at foreign branches of the same bank, or in an account with a "correspondent" bank. The accounts that banks hold with each other for settling international payments are variously called correspondent accounts, nostro accounts, or "due from accounts." These are equivalent to current accounts (checking accounts).

For example, Bank Danamon in Jakarta, Indonesia, holds--as nostro accounts--account number 36061913 at Citibank, New York, account number 93- 097-0046-2 at Hong Kong Chinese Bank in Hong Kong, and account number 95-042-05-B at Union Bank of Switzerland in Zurich. Through these accounts payments may be made or received in, respectively, U.S. dollars, Hong Kong dollars, or Swiss francs. International payment instructions will be sent to the banks involved via SWIFT. Payments made or received at the Citibank account will be cleared through the CHIPS system.

Another example of correspondent accounts are "payable through accounts". Some foreign banks offer their customers checking accounts in the U.S. dollars. They may have hundreds or thousands of local customers holding such accounts. The bank, meanwhile, will pool all these obligations in the form of a single account held at a U.S. bank. For example, Banco de Credito Centroamericano in Nicaragua offers payable through accounts. It holds a correspondent dollar account at Barclays Bank, Miami Agency. If Banco de Credito account holders write dollar checks to each other in Nicaragua, this only shows up as internal accounting entries on the books of Banco de Credito. It is not observable at Barclays. Only if a dollar check payment results in a transfer from Barclays Bank, Miami to another U.S. bank does it show up in U.S. records. Even then, it will typically show up as a simple bank-to-bank transfer-- say from Barclays Bank, Miami Agency to NationsBank.

Money laundering in this environment is a simple matter of setting up a virtual bank, or virtual account, or virtual dead-drop inside a nostro account, or a group of nostro accounts, or in the clearing system itself. Harried operations managers at banks have a hard enough time verifying that expected nostro net balance changes equal actual changes, without worrying about the individual bits and transactions that brought about reconciliation.

All net changes have to add up properly; or, if they temporarily don't add up properly, this has to go undetected in the clearing house settlement procedure, or during nostro reconciliation.

One can imagine, for example, a payment going from Unrecorded Nostro Account at Citibank to a Hidden Cyber-Money-Wants-to-Be-Free account in the CHIPS computer, on to a Mysterious Elsewhere Error Account also in the CHIPS computer, and finally to Unrecorded Nostro Account at Union Bank of Switzerland.