The Lykke Platform

In the global economy, financial markets are the equivalent of bridges in a transport system; today’s markets are straining under the pressure of a new century they were not designed for. We believe a new foundation is needed. This e-book, which is part of Lykke University, presents the Lykke platform in 27 short chapters. We explain the rationale and details of a new kind of exchange that is purpose-built for the needs of the 21st century. With only a few exceptions, each chapter is short and can be read in 1–2 minutes.

Table of Contents

  • 1 History
     2 Exchange Design
     3 Matching Engine
     4 Order Book
     5 Multisig Execution
     6 Settlement
     7 Tokens
     8 Liquidity
     9 Intraday Interest
    10 Price Discovery
    11 P2P Margin Market
    12 Financial Engineering
    13 High-Frequency Trading
    14 Identity and Trust
    15 Cryptocurrencies
    16 Crypto Assets
    17 Crypto Equities
    18 Ethereum
    19 Compliance
    20 Scalability
    23 Ecosystem
    24 Liquidity Funds
    25 Stock Exchanges
    26 Risk Management
    27 Central Counterparty
    28 Central Banking
    29 Coastline Trading
    30 Summary

All these innovations taken together constitute a new beginning for finance. At Lykke, we hope to build the foundation many more people will build on. This book explains how we came to these innovations and why they are important. You’ll find a paged version at Lykke University. Any time you want, you can come to Lykke.com and get involved — we hope you will.

About the Authors

Richard Olsen is the founder of Lykke, a pioneer of high-frequency trading, founder of OandA, and a well known figure in the world of quantitative finance. He believes that blockchain technology may bring us closer to a world with less conflict and more happiness.

David Siegel is the author of five books, a serial entrepreneur, lecturer, public speaker, and a blockchain enthusiast. He is a big believer in the P2P movement, self management, responsive organizations, the wisdom of crowds, and the digital empowerment of individuals worldwide.

“Traditional finance is comparable to the nervous system of dinosaurs. Blockchain can fix all this — we will build a seamless financial system.” 
— Richard Olsen, founder of Lykke

Academic Foundations
Dr. Richard Olsen holds a Licentiate in Law from the University of Zurich (1979), a Masters in Economics from Oxford University (1980) and a Ph.D. from the University of Zurich (1981). He worked in banking as a researcher and foreign-exchange dealer before founding Olsen and Associates in 1985. This consulting company brought served clients in the microstructure of market exchanges, trader behavior, and high-frequency trading. His company, under the name of Olsen Data, continues to sell high-frequency data sets and advise clients. He is a visiting professor at the Centre for Computational Finance and Economic Agents at the University of Essex and a start-up mentor to Swiss start-ups.

Founder of OANDA
In the early 2000s, Olsen and his partner founded OANDA, now one of the premier foreign-exchange platforms. This small start-up took FX trading from professional terminals to people’s browsers, giving them access to the high-frequency, low-spread world of FX trading previously unavailable to all but the largest firms. OANDA now has licenses in dozens of countries and employs almost 300 people. In these years, Olsen created many innovations for the FX industry.

Fixing the Financial System
Olsen left OANDA several years ago, wanting to fulfill his dream of creating a new worldwide financial system, where people trade the assets they own without intermediaries, with far better security and near instant settlement. With more resources, he would more quickly and more creatively destroy the current financial system in favor of a far better one.

Olsen explains that markets have been designed around the batch paper processes of the 20th century, big financial institutions, and the entire system is designed to maximize fees. Olsen envisions a world without the famous stock exchanges and traditional bond markets — all replaced by electronic messaging, high-speed matching engines, agent-based liquidity algorithms, and blockchain settlements.

Global Aspirations
Olsen wants a world where people are treated fairly, where guns diminish and economic opportunities increase, where old businesses die and new, global services spring up in their place. He envisions a world where people can send money effortlessly to anyone around the world, without the hassles and friction of trade today.

Olsen believes the world of central banking and government-issued currencies are not up to the demands of this new century. As Olsen explains, if we offer people a better way to trade, new financial products will appear, people may not need central banks as much, and demand for many of today’s fiat currencies could diminish as a new financial order takes shape.

At Lykke, our goal is to be the marketplace where people can buy and sell anything represented by a cryptographic token — in other words, anything that you can hold in a wallet. We are starting with currencies, but we will eventually include stocks, bonds, financial products, ownership of real assets, goods, and even services. Wherever there are buyers and sellers, we want them to use Lykke to find what they are looking for. This chapter explores the history and current state of crypto-exchanges, putting our architectural decisions in context.

The Early Days
The bitcoin network started in 2009. Initially, miners made most of the coins, and they traded them among each other. Soon, bitcoins had value, and people wanted to exchange their own currency to either buy bitcoins to use or for investment. Exchanges sprang up — they were essentially ecommerce sites for converting the main currencies into bitcoin and vice versa.

Centralized Exchanges
As with a broker, traders of these exchanges had to trust the exchanges to keep their assets safe. This did not work out so well. The more assets an exchange had, the more likely it was to be hacked and coins stolen.

The Lykke solution is to separate the matching engine from the exchange and let people retain control of their own assets at all times. This is called a semi-decentralized exchange. To see how that works, understand that any exchange has five parts:

Assets: currency or other tokens that represent tradable units of value.

Traders: who want to buy and sell.

Application: which handles all the onboarding, login, bank transfer, and session/communication mechanics — can be in a browser or a separate application on desktop, tablet, or phone.

Matching engine: which makes the match between the orders.

Settlement: which executes the match.

To understand the overall architectural decision to build an exchange with a centralized matching engine and decentralized settlement, let’s look at the options:

Centralized Matching Engine
A centralized matching engine is an app that runs on a dedicated, high-performance server. It’s very efficient, because it can hold all the data in fast databases and execute orders the same way the major exchanges do. A high-performance matching engine must work on the order of milliseconds, and, for some markets, sub-millisecond.

Decentralized Matching Engine
The concept of a decentralized matching engine is fairly new. The goal is to hand the execution of the order book to a network of independent computers that can make the match. As with Bitcoin, no single computer in the network can be trusted — you need to design a system that is robust and fault tolerant. In fact, you must design a system that is under continuous attack and yet still executes the orders at the scale mentioned earlier. This is a tall order, because the machines used for decentralized services are not optimized for speed, so you would have to design the network to reward speed and reliability. Some people are trying this approach, but we don’t believe the performance will be adequate for large markets.

Centralized Settlement
We know a lot about centralized settlement, because we’ve been doing it for so long. As a result of the 1973 paperwork crisis, the major exchanges on Wall Street created the Depository Trust and Clearing Corporation, which technically owns all the shares and issues IOUs to traders and beneficial owners. This was a patch on a system that was already breaking. Even though traders can trade a hundred times per second, each individual trade takes 1 to 3 days to settle, and the paperwork logjam is only getting worse.

Decentralized Settlement
In contrast, decentralized settlement takes place on a blockchain that has been shown to be highly resistant to attack, settles in minutes, and costs less than a penny per trade — regardless of size. The key to decentralized settlement is that each trader keeps her own assets in her own wallet, so the exchange doesn’t hold assets. Execution occurs through an atomic swap, which assures that both parties get what they bargained for. The Lykke exchange has nothing to hack, because we don’t hold your private keys. See the chapter on settlement to learn more.

Continuous Flow
Today’s financial system is designed around early twentieth-century manual paper processes. When the markets are closed, pressure can build, so the next opening can be surprising. As Richard Olsen says, closing a market for the weekend is like closing the hospital for the weekend. We know that many market anomalies are due to opening hours, lunch, holidays, and weekends. The only way to run a worldwide financial system is continuously, with traders using software to watch for problems so they don’t have to watch the market.

The Lykke Exchange: The Best of Both Worlds
The Lykke exchange pairs a centralized matching engine with a decentralized settlement mechanism. This gives us a high-performance, world-scale exchange with the safety of decentralized assets. When you trade anything, the actual ownership changes hands within minutes on the blockchain, eliminating many of the problems we outline in other chapters. The Lykke ecosystem will run continuously, facilitating the smooth operations of markets at all times.

The team at Olsen and Associates has looked with a microscope at trade time-series and discovered many surprising things about market mechanisms. Here we present several problems with today’s markets, then we present our solution: the Lykke matching engine.

A Short History of Matching
Up until just a few years ago, price matching was accomplished with humans shouting in pits, writing and tearing up bits of paper. Today an exchange has a list of electronic orders from buyers and another from sellers, and it uses a matching engine to pair these orders. Price has priority — the highest bid price and the lowest ask price will be executed first. If two orders come in at the same price, then the earlier order is filled first. This price-first approach helps bring buy and sell orders together with minimal spreads.

Over the past few decades, high-frequency traders have entered the market, placing hundreds of simultaneous orders, canceling many, moving their positions, and trying to capture small slivers of profit in short bursts. Far from being a bad thing, many of these traders help provide liquidity, so buyers and sellers get their market orders filled as soon as possible, which is what they want. On the other hand, this activity has created problems.

Problem: Front Running
When a large institutional order comes into the market, it has a price impact — that is, the order will fill in small chunks, at different prices. High-frequency traders pay extra to the exchanges, allowing their software access to the order book every hundredth of a second. When their software sees a big order coming, they put in an order with a better price, which they know will be executed first. Then comes the large order, which moves the price even more in that direction, and the high-frequency trading program gets out with a quick profit. This is loosely called front-running, and it’s legal. It’s now so advanced that many hedge funds use custom chips located just a few meters from the exchange’s computers to minimize time gaps.

Another form of front running is when a broker takes a large institutional order and puts his own order ahead of it. This is illegal. So we will use the term in its broadest sense, to mean people who try to profit from large orders quickly, rather than providing liquidity. (It’s hard to tell which is worse — both forms of front running are bad for the markets.)

Exchange-design matters. In 2001, the U.S. Securities and Exchange Commission’s changed to pricing stocks in pennies, rather than fractions of dollars. This extra fidelity facilitated front running, because front-runners could now pick up pennies on many trades, all done by computer.

While high-speed trading sounds obscure, and legitimate liquidity providers account for much of this activity, front running accounts for more than half of all trades on the New York Stock Exchange today (Source: Congressional Research Service Report). Not surprisingly, large institutions have countered with an array of countermeasures, including setting up their own “dark pools” — private exchanges — where they can just trade with each other, without middlemen.

Problem: Flash Crashes
One consequence of HFT is that if the algorithms get it wrong, the mistake can multiply, and we have a flash crash. HFT strategies were responsible for the flash crash of May 6, 2010, the October 15, 2014 volatility in treasury markets, and the August 24, 2015 crash. As we were writing this, a flash crash occurred on the British pound overnight, as Brexit concerns were magnified by program trading reacting to news stories.

We have analyzed financial shocks. If you look closely at the data, you’ll learn that a series of aftershocks- traders reacting to each other — can continue up to 2,000 times after the initial event. Perturbations last much longer than expected and are as disruptive to the market as the initial shock was. We believe that market design is largely responsible for these shocks and that a better market design can correct them.

Solution: Lykke — An Exchange Designed for the 21st Century
Today, the vast majority of traders are algorithms that will happily buy or sell, whichever is more profitable, in a very short amount of time. While most orders used to be limit orders to buy or sell, most of today’s orders are two-sided, with both a bid and an ask. The goal of the Lykke matching engine is to continue to encourage market-making and fast liquidity, while preventing flash crashes and hackers from stealing people’s money. So we designed a new exchange mechanism based on four principles:

  1. Price-Spread-Time order preference
  2. Multisig Execution
  3. An open order book anyone can see for free
  4. An agent-based liquidity program

The next three chapters describe orders and settlements.

Previously, we identified several problems with the current price-time matching engines used by all exchanges today. At the highest level, our goal at Lykke is to create smooth price transitions when traders react to news and events, reducing volatility. In the cases where truly disruptive news comes out about a stock or currency, our goal is to facilitate as much liquidity as smoothly as possible, to transition to the next natural equilibrium state without overshooting and causing oscillations. Our goal is to discourage front-runners and encourage liquidity providers.

Our innovation is price-spread-time matching. It works like this:

  1. Price is first — orders with the most desirable price get executed first.
  2. Spread is second — if two orders have the same price, market-making orders that include both a bid and an ask at that price go first; the order with the narrowest spread gets highest priority.
  3. Time is third — if two orders have the same price and spread, then the earlier order is executed first.
  4. Size doesn’t matter — large orders don’t go ahead of small orders.

Not only is there good research backing up this approach, it has proven itself in thousands of simulation runs using real data, and it’s in use today in the Lykke exchange. In practice, orders are separate, but the exchange can see who is making them. When someone enters a buy order and then a matching sell order, the exchange can see the spread and move the two orders forward or back in the queue. If one of those two orders is canceled, the spread no longer exists, so the remaining order goes back in the queue. This matches the way orders are made today. In the future, Lykke may be able to offer a two-sided order type that would make the system more efficient.

This priority system encourages competition and discourages front running. A trader who wants to get ahead of a large order will do so with a one-sided order to buy or sell ahead of a predicted price move. A two-sided order would be a “toxic” trade, as the trader would be exposed to selling what he wants to buy or buying what he would prefer to sell. A natural market maker, on the other hand, will have his orders executed ahead of the front runner, as he is willing to trade either way and provide liquidity. To get ahead and have his orders executed, the front runner will have to offer a better price, which will reduce or eliminate his profits.

Another feature is that our order book will be available publicly at all times. We don’t sell a “premium version” to traders who pay more. This is another Lykke innovation that helps level the playing field for all traders.

This design prevents flash crashes. By prioritizing low spreads, it’s hard for the market to move too fast too quickly. Even in a dramatic price move, low spreads and fast matching keep prices moving in small increments, reducing volatility. This approach keeps spreads ultra-thin and creates a virtuous cycle that benefits investors.

The price-spread-time order book is a major innovation for electronic marketplaces. It is also an experiment. The Lykke team will continue to make adjustments and innovations as we learn from the data what is working well and what needs improvement. This could lead to several new order types and even more efficient ways of designing markets and matching engines. Lykke’s approach promises to keep spreads ultra-thin and create a virtuous cycle that benefits investors. Please visit Lykke Labs to learn more about our ongoing experiments.

This set of rules should give regulators less to worry about, so they don’t try to implement unnatural regulations that interfere with natural market functions.

On most exchanges, the exchange mechanism swaps value between counterparties and hands the resulting trade over for settlement. The assets are on the exchange, so the swap is just a matter of updating the internal ledger. On the Lykke exchange, since our customers’ assets are always under their control and we settle on the blockchain, we use a different approach.

The general name for this process is an atomic swap: the blockchain itself ensures that the trade completes in both directions. This is effectively like a revolving door: a buyer will put her money on one side, the seller puts the desired asset on the other side, and when they both give it a push, each gets what he wants. In the Lehman crisis, a German bank shipped 500 Million EUR to Lehman (not knowing that Lehman was already bankrupt) without receiving the other side of the trade.

Lykke uses something called a two-of-two multisignature transfer. Here’s an example of how it works:

A Multisig Example
You hold 100 Lykke coins (LKK) in your wallet. You want to exchange for euros on the Lykke platform. You hold the private key to your coins in your wallet (it’s the 12 words you wrote down when you originally set up your wallet).

At the same time, the Lykke exchange holds a second private key that is tied to your Lykke coins.

The same is true for your counterparty, who wants to trade Euros for Lykke coins. All coins on the Lykke exchange have two keys: one for the owner, and one for the exchange.

The coins exist right now on the Bitcoin blockchain. The goal of this transaction is to transfer the coins from one bitcoin address to another.

Now, the trade requires four keys: one from each owner, and one key from the exchange for each token. When all four trades are present, the bitcoin blockchain accepts the trade.

It takes just a few minutes to settle, and generally around 30 to 60 minutes for a full confirmation.

The Refund Contract
What happens if something bad happens to the exchange? For example, hackers get the password and are able to discover a bunch of private keys. Two things:

1. The hackers can’t move any coins, as they require you to agree as part of the multisig transaction. So the private keys are useless to the hackers.

2. If the hackers somehow take the exchange down, destroy the keys, or hold them for ransom, a special refund contract unwinds the trade. Let’s see how that works. Assume the Lykke keys are gone:

First, you can’t trade. Your assets are stuck in your Lykke wallet. Since you registered with your email address, you will get a notice saying Lykke’s private keys have been compromised.

A special contract on the blockchain checks for the private keys every day. If it sees the keys, it resets. If it doesn’t see the keys, it keeps checking for 30 days. If, after 30 days, it still doesn’t see the keys, it automatically issues a refund. In your Lykke wallet, there’s a place in the settings page where you can give a bitcoin refund address. You can do that now, or you can wait and do it if you ever need to (we hope you never do). The refund sends all your bitcoins to that address, which you have the sole private key to — this address is off the exchange.

At the moment, this only works for bitcoins. We will soon announce the refund feature for all colored coins as well.

Not only does it take two days to settle trades today, but the system that does it is so arcane that few people understand it. Patrick Byrne’s talk explains why we need a new system. In this chapter, we’ll outline the problems and show how powerful a blockchain solution is.

Problem: Delayed Settlement
In the stock market, no one actually owns any stocks. In response to the paperwork crisis of 1973, a company called the DTCC was created to hold all stock certificates and be the ultimate custodian between dealers and market participants. Today, settlement takes place using ancient computers and batch overnight processes. The entire chain of custody takes two to three days to settle, and that includes all the high-frequency trades that happen in the blink of an eye.

Delayed settlement allows for ‘step by step’ transactions. This operational detail can have huge impact in extreme events, such as systemic crisis — precisely, when a financial system is not allowed to break.

Problem: forward contracts
In forex, when you buy a currency, you’re actually buying a forward contract to receive custody of that currency, which itself is another IOU. The contract requires the payment of interest on a daily basis. So the first 23 hours are free, but then you have to pay. Since the vast majority of people don’t want to own the currency anyway, traders have found various ways to take advantage of the system. One of those ways is interest-rate swaps.

Problem: Interest Rate Swaps
The same amount of money held in the same bank in the form of euros, dollars, or yen can carry a different interest rate. To get around this, many institutional traders use interest-rate swaps, which preserve their currency holdings but effectively reduce the interest rates they pay, all while making money for another middleman: the swap bank.

This happens in long-term loans; it also happens in forex. Because of the two-day settlement, many people purchase swaps at the end of the day to avoid paying interest on their loans. It’s so common that traders now buy more swap contracts than actual spot (trade) contracts. Most swap buyers are just trying to get around the rules and save money — they are not natural swap buyers.

Solution: Blockchain
A blockchain is an ideal settlement mechanism. Using blockchain technology, everyone holds her own currencies and assets in her own wallet. Once the match has been made, the exchange executes an atomic swap to ensure that the trade goes through. You can think of this as a revolving door that makes sure both parties get what they asked for.

Using the Bitcoin blockchain, a trade settles when it is recorded on a block. This typically takes around ten minutes. Then it’s customary to wait for at least three blocks to pass before confirming the trade, so settlement takes place in stages, usually completing within an hour. On the Ethereum blockchain, the trade is recorded within a minute and confirmed in under ten minutes.

Within a few years, we will be able to execute tens of thousands of transactions per second, and we’ll be able to settle in seconds, not minutes. Because Lykke has no trading fees and low spreads, you can just buy and sell what you want as often as you like. Each time you do, the asset actually changes hands and the trade is complete within a few minutes.

With the exception of bitcoin (and soon, ether), all the assets you see in your Lykke wallet are tokens. A cryptocurrency is essentially digital money. It can be passed from person to person, and it has its own intrinsic value. A token, on the other hand, is a representation of value, an IOU. A token could represent the ownership of a diamond, a gram of gold, a piece of land, a car, a dollar, a loan, stock in a company, etc. In fact, you could hold dollars from several issuers, all of them represented by their own tokens.

In your Lykke wallet, you’ll find tokens that represent many kinds of assets. A colored coin is a tiny fraction of a bitcoin (less than 1/1000th of a cent) to which the issuer adds code that corresponds to the ownership of the underlying asset. These are tokens: tiny pieces of a cryptocurrency that carry specifications for ownership of a specific thing. They have no inherent value. The value they carry is specified by the agreement between the issuer and the owner.

When you see Swiss francs, dollars, pounds, and other currencies on the Lykke platform, these are colored coins we issue. We purchase the same amount from a bank and issue our tokens to you, with our promise that these tokens can be redeemed for their stated currencies. At Lykke, we have chosen the Open-Asset Protocol for colored coins. With the possible exception of Ethereum-based coins (described in a future chapter), all tokens on our platform have exactly the same structure. This lets us build tools, dashboards, derivatives, and other features that will apply to all assets, regardless of what they represent. Learn more at www.coloredcoins.org.

For now, these colored coins are only tradable on the Lykke platform. Some day, that may change, but for now it’s best to think of them as units of value you keep in your Lykke wallet for trading only on the Lykke platform. You generally wouldn’t plan to send these colored coins to another wallet. In fact, our software prevents you from sending your colored coins to a normal bitcoin address, since that would erase the value!

In a previous chapter, we saw how a price-spread-time order book reduces both volatility and major market shocks, allowing market participants to maintain equilibrium as they adapt to a shifting world of news and events. This chapter explains the advanced liquidity program Lykke applies to create a smarter matching engine for the digital economy.

The Liquidity Problem
Markets are complex adaptive systems. They don’t adhere to hard rules. Sometimes a large price move is a legitimate response to important real-world events; other times, a move that looks almost identical can be the result of market manipulation. A large trade could be a signal of something about to happen, or it could just be a routine rebalancing of a larger position. In an ecosystem like this, there are natural buyers, speculators, liquidity providers, and opportunists. A single market participant can easily move from one role to the other, and margin calls can come out of nowhere, so the order book tells you surprisingly little about what’s really going on.

Liquidity is relative to what’s happening in the market. Ideally, at exactly the moment you want to sell your position, there is a buyer who wants all of it at the same price you want to receive. In reality, a large sell order is sliced into several small orders and nibbled away by whoever is buying, and a large buy order is pieced together from many small offers. As this plays out, prices change — sometimes dramatically. Even in the forex market — the most liquid market in the world — liquidity is a challenge. All together, over $3 trillion changes hands daily. Assuming everyone is trading on the same exchange, 3 trillion divided by 86,400 seconds per day translates into $34.7 million per second (on average) for all of forex. Since the EUR/USD accounts for half of all volume, then, on average, only $17.4 million is available for liquidity at any given time. Anyone trading $1 million or more is going to end up with an aggregate price that could be significantly different from the current market price.

The problem comes when prices start to move out of the normal range — the uncertainty causes spreads to widen. That’s when programmed market makers can make mistakes. At any given small slice of a second, there may not be so many buyers and sellers in the market. Rather than calming down, both humans and programs can overreact. These factors create jumpiness, price oscillations, and occasionally big surprises.

Because so many people use leverage and have so much money at work, even a small but unexpected price move can have dramatic domino-like effects. Profit taking takes place toward the end of the day, spread out over several hours, but traders with unrealized losses are often forced out of their positions or close them reluctantly at the end of the day. Positive feedback sets in. Margin calls increase. As more intra-day traders are forced out of their positions, even stalwart investors see the bad news and try to get out of the market ahead of whatever may be coming. This effect is magnified on Fridays and even more magnified on the last Friday of a quarter. Liquidity can evaporate in seconds.

In an extreme liquidity shortage, whole groups of market participants can be driven out of the market. Lehman Brothers was a very visible example, but liquidity traps have crushed many speculators who simply did not have access to funds when they needed them.

An Agent-Based Algorithm
In this environment, a rules-based approach doesn’t work. Neither does individual human judgment. Lykke has created a sophisticated system that uses software agents to provide liquidity, especially in times of high volatility. This approach is the result of decades of research by Richard Olsen, Anton Golub, and their team.

The basic idea is to create many different software agents that simulate a very diverse market, trading at different time scales, using their own mini algorithms, and let the “wisdom of the crowd” find the best prices. In this way, one agent may buy larger positions and hold them for longer, and another may buy very small positions and hold them very briefly. By not working in concert, the different independent decisions help minimize positive feedback. The software agents are designed specifically not to converge to herd behavior in extreme circumstances. This should encourage smoother transitions from one price to the next.

We have learned that forex markets do not react to news in a predictable way. Because of the complex market microstructure, leverage, and unseen margin call orders, news has almost no predictive value. Our program, which is open source for all to use, looks only at the price movements and nothing else. That way, it can provide liquidity for a large number of tokens on our system without having to know anything about them. In the future, we plan to add more intelligence, so we can provide liquidity for equities, which do respond to news events and online chatter. When many market makers compete on the same platform, we get an efficient marketplace, where spreads are low and transactions clear at or near the desired price.

Liquidity is critical to well functioning markets. As we’ll explain in another document, Lykke plans to be a major liquidity provider for all digital assets, not just currencies and stocks.

When Richard Olsen was CEO of OANDA, he launched an industry innovation: margin loans on a second-by-second basis. With Lykke, he plans to bring second-by-second interest rates to the world’s financial markets, not just forex. This chapter shows that market design has a huge influence on market activity, and that we can fix many of today’s pernicious problems by implementing an intra-day yield curve. We start by identifying the problems and then show how powerful the intra-day yield curve is.

Problem: Ninety Five Percent of Trades are Intra-Day
The daily transaction volume in today’s forex markets is close to 4 trillion USD or equivalent to 30% of the US GDP. Trading happens at a fast pace, with traders opening and closing positions using software and fast-moving data. It is estimated that 90 to 95 percent of the positions are held for less then 24 hours, typically for only minutes or a few hours. Only five to ten percent of the total volume of positions actually pay interest. This is a direct result of a skewed system that encourages intraday leverage (borrowing) for free.

Problem: Free Money Skews Markets
Technically, you can borrow on margin for 23 hours and pay no interest at all. Once the clock strikes midnight, however, you pay the interest on a 24-hour basis. Market participants repeatedly borrow and return money before the end of the day, to take advantage of the free money.

Why do exchanges lend money at zero interest? First, because it imitates the old, paper-based processes of the 20th century. Second, because the more people trade, the more the exchange makes on each trade. The only person they don’t want to lend to is someone who buys and holds, so they start charging interest after one day, which compensates the exchange for not trading. As 95 percent of trades take place intra-day, almost no one pays interest to leverage his position, traders have become hooked on leverage and trade excessively, and exchanges are hooked on the trading fees.

This turned out to be a great marketing tool for exchanges: give away free money, encourage them to trade often, and make a profit on the spread. No third parties are interested in lending free money to traders — only the exchanges do it. For them, it’s a marketing expense.

Problem: Spot trades are not Spots
Spot trades are supposed to be trades that settle immediately. In reality, they settle in two days, using a form of IOU. The IOU has to pay a certain amount of interest during those two days, and the interest is factored into the price. An instantaneous trade would be at a different price. It can be that the interest is more than the gain on the trade.

Problem: Local Daytime Price Decline
Another anomaly seen in markets is that foreign traders tend to trade foreign currencies more than local currencies. Traders are most active during normal working hours. They put on their most risky positions during the middle of the trading day. These combine to drive the price of local currencies down during local trading hours around the globe. We can see this as a wave of price declines that moves across various markets as daylight shines on each region. Whatever the cause, it’s likely that the incentives built into the market have much to do with it.

Problem: Tomorrow-Next Swaps
In currency markets, few people actually want to take delivery of a currency. As noted earlier, more than ninety five percent of trades are intra-day. There are times when a trader would like to hold overnight but doesn’t want to incur the normal interest charge. This essentially involves buying a swap contract that ensures getting the same price the next day without actually holding the currency overnight, which incurs less interest. It’s so common that traders now buy more swap contracts than actual spot (trade) contracts.

Problem: The Carry Trade Pushes Prices Up
The carry trade involves borrowing money in one currency, changing it into another, and then lending the new currency at a higher rate of interest, pocketing the difference. In theory, there shouldn’t be much carry trade: if New Zealand dollars pay 5 percent and Japanese yen pay 1 percent, then the exchange rate should even out that difference. But that’s not what happens. What really happens is that the trade works, so people sell more yen to buy more NZ dollars and the price of the NZ dollar goes up, increasing the gain on the trade.

This works well for everyone, until too many people are invested in the smaller currency. Then there’s a collapse in the price, and the last people to invest lose everything. Remarkably, even accounting for these collapses, the carry trade still makes a good profit in the long term. The value of high-yielding currencies tends to increase over time more than their mainstream counterparts. It’s believed that carry traders drive the price of smaller currencies about two percent per year higher than they would otherwise be.

Problem: The Intraday Short Pushes Prices Down
This system of free intraday money favors shorting high-yielding assets during the day and discourages intraday buying. How does that work? Let’s say you live in the US and hold dollars in your account.

Suppose you believe the South African rand, a high-yielding currency, will likely go up in the near future. Then you can buy rand, but since settlement takes place in two days, you have just purchased a forward contract, not the actual rand. You won’t receive any interest on your rand today, but you do have market exposure today, so you might not want to buy just now. You may want to wait and watch, and buy later if you see a really good opportunity. To compensate, the rand has to move about twice as far as it would if you were receiving interest for the first day. And there’s little incentive to buy and sell rand on the same day, since you are guaranteed to receive no interest at all. This is a natural, systemic bias against purchasing.

Now, suppose you believe the South African rand will likely go down against the dollar. You put in an order to short the rand and get dollars in your account. Since you don’t have to pay interest if you unwind the contract by the end of the day, you might as well go short now. If it goes down even a tiny bit, you can cash before having to pay interest, or you can roll forward with a swap. Better to jump in than to wait and watch.

That one little marketing trick — loaning traders money for free during the day — has led not just to higher profits for exchanges but also to a vast market of unnecessary trades, products, and activities that have nothing to do with the smooth functioning of a global market.

Solution: The Intraday Yield Curve
Richard Olsen is, as far as we know, the only person to create second-by-second margin lending. He did it at OANDA, the forex company he founded.

At Lykke, we will implement second-by-second interest charges to remove the scam and bring markets to normal functionality. Lykke will charge interest on margin starting at one second. This takes away power from the exchange and gives it to traders, who will soon see a more correct-functioning marketplace. It will help correct all the problems outlined above. It will give our traders an advantage over other traders.

This is not just for forex. The traditional stock and bond markets also have no intra-day lending, which leads to market illiquidity. On September 17, 2008, Lehman Brothers found itself with several billion dollars’ worth of bad loans on its books, and it had to pay its creditors or default. The bank wanted to borrow money. They had no problem paying interest. If the bank had been able to borrow quickly, it may have been possible to save the business.

We want a smoother functioning market, one that has as few discrete jumps and surprises as possible.

This slightly longer chapter presents the results of decades of research into the nature of the micro-structure of markets. We present several non-intuitive concepts that shed light on some of finance’s most difficult problems.

Definition of Price
The fair market price of an asset is the price that would turn most sellers into buyers or most buyers into sellers. More accurately: the fair market price should be the average price of all people who are willing to offer both a bid and an ask for the same thing at the same time.

We talk about a single price for an asset, but in reality there is both an ask and a bid price, and they are usually different. Technically, we could be talking about the ask, the bid, or an average between the two.

What the price is not: it’s not the price someone is willing to pay. That may depend on the person’s circumstances. The price of a taxi ride to the hospital is not what a person with a heart attack would pay, but what most people would pay at any given time. If a large fund has a liquidity problem, it may start selling assets, even if it doesn’t want to. Furthermore, a sale can just as easily be the opening of a short position as the closing out a long position. Just looking at orders is misleading for understanding price. As we say in economics, “never reason from a price change.”

Market Makers Set More Accurate Prices
Market prices are determined by market makers — people who are willing to take either side of a trade. At what price would you switch from being a buyer to being a seller? That’s the market price. Is the spread wide or narrow? That itself says a lot about what people think the price is or could soon be. On the Lykke platform, we encourage real-time price discovery by giving two-sided (buy and sell) orders higher priority in the queue. In general, spreads are a measure of confidence about future prices. The more traders offering a spread, the better defined the price.

Sellers Move Prices
You might think that price is a symmetrical function of demand between buyers and sellers, but that’s not accurate at all. More than 90 percent of all trading occurs intra-day. When a trade goes against a trader, she will want to close out that position. If she doesn’t close it, she may soon have to close it.

Anyone — and this is also true of algorithms — who opens a position thinks he is getting a good deal. The same is not true for selling. Look at the following dataset, which comes from the OANDA order book page:

The current price is at the green line. On the right are the current positions, where orange represents an unrealized profit and blue represents an unrealized loss. On the left is the order book, where we see sell orders on the left and buy orders on the right. That’s where the action is.

In the left figure, we see in orange a lot of buy orders below the current price, and a lot of sell orders above. These act as a natural “spring” to push prices back toward where they are now. Note that there is much more orange than blue, so anyone unwinding a position should find a willing counterparty. This looks like a fairly stable situation, with plenty of orders to bring price moves back toward the center.

But it isn’t.

The reason is that you can’t see margin calls here — they happen when a trader’s position gets too far away from his ability to cover the amount required to maintain his margin loan, and the lender autoliquidates the trade. See the buy orders around 1.09 on the chart? There are many of them, and few sell orders, so pushing the price down a bit is like rolling a ball uphill — it encounters resistance and goes back where it came from. The problem is that if there’s a price move that goes quickly past that number down to about the 1.085 area, now the ball has crossed over the threshold and — because of the many margin calls we can’t see in this chart — starts to run away down the hill on the other side faster and faster. As more and more traders are forced to sell at exactly the time they don’t want to, the knife accelerates. This is due to overconfidence and overcommitment on the part of traders — something you can count on for many years to come.

Short-Term Traders Move Prices
Prices don’t move in any predictable pattern. They move in fairly random small jumps around a given price until something happens. Then something does happen — news, commentary, rumors, macro-economic events, someone starts buying or selling. Often the “seed” of a price move is just a larger than usual random fluctuation. People immediately overreact, causing the price to jump further. As soon as it gets over that threshold we just saw, those in the money do nothing, watching their value go up, while those out of the money are busy selling. This causes more forced selling. Eventually, the price settles back down, and we arrive at a “new normal.” We call this the “Alice in Wonderland Effect” — you go through the looking glass of the price move, and now you’re in a new world. Everyone adjusts. People who wouldn’t have bought for $100 this morning are now buying at $110. People who wouldn’t have sold at $50 are now selling at $45.

Short-Term Sellers
If you take the two previous points together, you conclude that intraday sellers (or, more accurately, traders unwinding their positions), most often as a result of a (legitimate or random) “seed” event, contribute the vast majority of price change to an asset in a given year. Many intra-day micro-adjustments can add up to a yearly change of 30 percent or more (see the twelve scaling laws for details). While it may look like long-term fundamentals drive prices, that may be a case of mistaking correlation with causation. What most people think of as fundamentals may provide an envelope for prices, keeping them in check, but intra-day traders do most of the work.

Short-term sellers have a larger effect on prices than people think. Markets have memory, but short-term traders have short-term memories. They adjust easily to “the new normal” and reinforce a new pattern, waiting for the next event. They go home with no positions and wait to trade another day. Longer-term traders have a small percentage of the overall market. They have longer memory but little impact on price. This is why so many assets track the overall market — because the same event that can cause a jump in one price can cause a similar jump in many other prices.

Short-Term Fundamentals
Short-term fundamentals are based on average volatility, which is based on many factors. Think of a stock or currency pair chugging along, going up and down a bit for much of the day, staying in the fundamental range. Then comes some bad news, causing an acceleration downward, as traders are forced to unwind their positions as a result of margin calls. In this case, we say the price flies away from the fundamentals. Looking at the graph of most securities’ prices, you can see short vertical movements, often followed by restabilization at a new price level.

This is what we call the dynamic fundamental: the rate of average price change in either direction, at a relatively short time scale. If the volatility goes out of the fundamental (historical) range for an asset, you can expect positive feedback, then opportunistic profit taking, then back to volatility again. This pattern repeats at short, medium, and long-term time intervals. Neither the layman nor the long-term investor understands this activity.

In a normal (broker) margin loan, you put up some collateral and get some increased buying power (line of credit). You borrow money when you put a position on. On today’s FX trading platforms, there are three parts to what happens next:

  1. You can pay off the loan any time by unwinding your trade.
  2. You pay the average rate of interest for the time you have held. This is computed based on the rates of interest in the swap market for the same period. As mentioned elsewhere, you pay no interest when borrowing intra-day. Most margin traders unwind or swap out before midnight.
  3. You’re not fully in control. If the investment moves away from you, you have to put up more money, or the lender has the right to liquidate your position.

The difference on the Lykke platform is that interest-rate charges start at one second, so if you open a position and hold it for 73 minutes and 21 seconds, you’ll be charged a blend of the one-hour, the one-minute, and the one-second rate. We have explained elsewhere why this is important, and why “free” margin money isn’t free.

For now, Lykke is the only broker on the platform providing margin loans. Other registered brokers may apply to provide margin on the platform — we welcome them.

Eventually, there could even be a P2P lending market, where traders can lend each other money. That certainly won’t happen for several years, if ever. We want Lykke traders to have plenty of choices, so we encourage other brokers to come make margin loans alongside us.

Earning Interest
On most forex platforms, when you hold any currency, even your native currency, you are technically earning interest, because your broker holds your money in a money-market account with local banks and other brokers. It may not be much (and it could conceivably be negative), but you can expect to earn interest on the Lykke platform at some time in the future as well. We don’t offer this feature at the moment, but we expect to some time in 2017.

In a remarkable paper entitled Nuclear Financial Economics, written in 1993, William Sharpe describes how a fairly simple market for futures can be turned into a rich ecosystem of “bets” and insurance policies that help market participants get what they want through creating new financial “atomic particles” that could then be used to hedge, balance, or offset portfolio risk. In the past, these products were created by exchanges and organized groups.

Today, with digital tokens, we can more easily create new financial instruments. We are building the Lykke platform to enable a Cambrian explosion of new financial products that will help people manage risk, opportunity, and portfolios.

Prediction markets are starting to prove themselves as powerful tools in forecasting complex adaptive systems. If you want to know the future value of a number — unemployment, GDP, the number of cars Toyota will ship this quarter, the price of gold three months from now, or any other quantifiable outcome in the future — you could create a token for that and see what the natural market price is.

Example: you may want to create a token that represents the official inflation number for the end of the year. To do that, you could offer a token that pays $100 if the inflation figure comes in above 3 percent, and another token that pays $100 if the inflation figure is lower than 3 percent. Then, you auction these contracts and let them trade in the market. The relative difference between these two tokens gives you a prediction. You could use that prediction in other products, or to price, say, an insurance contract. This market doesn’t cost you much to set up, because you are taking in money for each token, so you only lose the difference between what people pay and the face value — this can be a good way to pay for quality information.

Example: we may see that many people are trading bitcoin for gold, or solarcoin for lykke coins. In this case, we can create a single trading pair, so buyers and sellers only have to cross a single spread.

When many different kinds of assets are on a single platform, new financial products may create markets that previously didn’t exist. You can easily by fractional ownership of a portfolio of collectible cars, paintings, or villas. We may find, for example, that there’s a market for using diamonds to secure rental contracts, or that people want to hedge real-estate using gold. Instruments that facilitate these markets will be easy to create on the Lykke platform.

Another exciting area of development is oracles. An oracle is a service that resolves an uncertainty using an API. This brings about new products, like parametric insurance. In parametric insurance, there are no claims. If you purchase an insurance contract that pays you if your flight is canceled, then an oracle will provide the data about the flight, and the smart insurance contract will pay you automatically, without a claim. People are working on parametric insurance for farming, travel, natural disasters, and other areas where a clear signal can resolve the contract without a claim.

One prediction we have is that markets will replace experts in pricing. Today, insurance actuaries look at data and price insurance products. On the Lykke platform, you’ll be able to buy and sell risk tokens that are not only insurance, they are dynamic pricing indicators that continually adjust the price to the expected risk. In general, dynamic risk hedging will become much more possible on our platform than in the past, because people will be able to create such products, issue them, and trade them in real-time. This is the new world of finance Lykke is building.

We have been involved in high-frequency trading since the 1990s. We are designing the Lykke platform to be a high-performance matching engine for all HFTs. We plan to create an environment where natural buyers and sellers can do business without the usual slippage associated with front running.

As we have learned, front-running accounts for as much as fifty percent of all trades on the larger stock exchanges. Lykke’s price-time-matching engine will make front running unprofitable. So the successful high-frequency traders on the exchange will be market makers, and this adds liquidity to the markets.

We will provide APIs, tools, and dashboards for high-frequency and program traders. We will have an online community of algorithmic traders and will be responsive to their needs.

We will add more here as it develops.

Online Identity
For many years, people building the online world have struggled to build a digital identity infrastructure. There have been many attempts, but nothing has taken hold. At the moment, the gold standard remains your basic set of centrally-issued documents: passport, driver’s license, social-security card, etc. For proof of physical location, it’s usually a recent utility bill.

At Lykke, we are very serious about licensing, compliance with regulations, and building trust with our customers. We are raising money to register as either a broker-dealer or an exchange in many countries. At the moment, we use a KYC (know your customer) process similar to what banks use: we ask for your government-issued ID, a photo, and a picture of a recent utility bill. This can be done using your phone’s camera in a matter of minutes. Then we run a routine, automated check to make sure you’re not on any lists, and soon you are approved for trading.

In the future, your digital identity will be different. It will probably be on a blockchain somewhere, so everything that’s critical to your identity will be recorded forever. You won’t be able to erase or delete it. There are three forms of digital identity:

Self-sovereign identity: you construct your own digital identity by yourself. This is a visionary approach that takes into account people moving around the globe more and more, and building a trust mechanism for validation.

Managed identity: a large institution issues you an identity and can revoke it according to its rules. This would be a digital version of what we have today. It’s already being done in Estonia, and many other governments are working on their own versions.

User-centric identity: a nonprofit governing body issues you an identity. This basic identity layer does not include trust information — its sole purpose is to validate that you correspond to your identity and there are no duplicates.

We will see how this develops. Lykke hopes to be part of these new digital identity movements and help create an identity infrastructure that functions as well as our matching engine does.

Trust and Reptutation
Once you have a digital identity, you will naturally use it. The public part of that will form the basis for your trust score. This will combine with your credit score and public criminal record to form a digital profile that others can use in doing business with you. You will actively manage your reputation online, as it will enable so much. This will become more and more important as we scale Lykke services beyond our trading platform.

Take out a bill from your wallet and notice something — your money isn’t backed by anything other than hope and faith in institutions. The value of your money isn’t in the paper, it’s in the unique number printed on that piece of paper, and the promise behind that number. Because currency is limited and issued by an institution people trust, it has value. This same principle works on cryptocurrencies, with the trusted institution replaced by a mesh network of “nodes” maintaining a single ledger that everyone agrees is the official record of exchange.

We won’t explain the mechanics of decentralized currencies. Here, we will explain the value of decentralized currencies.

Fiat Currency
Today, you can see dollars, euros, pounds, and yen on our platform. You can buy them and put them in your cryptographic wallet, which means they exist as colored coins on the Bitcoin blockchain. At Lykke, we create our own tokens that represent these units of value, then we accept your electronic bank transfer and issue you Lykke tokens. What makes each one worth a dollar, or a euro, is our promise to redeem your tokens.

This is a baby step, similar to what banks do when they use electronic records to keep track of accounts. Already, several countries’ central banks are looking into issuing their own crypto-fiat currencies, using their own blockchains. The UK has started studying options, and the US has had a trial demo. These early experiments could pave the way for true mainstream digital currencies in the not-too-distant future.

This is going to sound strange, but in a few years it won’t: eventually, many banks will issue their own fiat currencies as cryptocurrencies, the same way we have. That is, there will be a Lykke euro token, a Morgan Stanley euro token, a UBS euro token, a Credit Suisse euro token, a Citicoin, and even a European Central Bank euro token. These tokens will have slightly different characteristics and may bear different rates of interest from their respective institutions.

Alternative Currencies
The history of money includes many local currencies that have been quite successful, and many are still in use today. There are also ideological currencies and currencies to ease trading between partners that would otherwise have high exchange costs.

The first worldwide decentralized digital currency is bitcoin, which began in 2009. What started as a way for disenfranchised people to trade with each other and establish their own financial regime is slowly becoming a mainstream currency and investment. At the moment, bitcoin has a market cap of around $12 billion. It has shown remarkable resilience to attack. And it has paved the way for dozens of similar efforts. You can see the list of most popular cryptocurrencies at CoinMarketCap.com. So far, only a handful have market capitalizations higher than $100 million, so the entire universe of crypto currencies is still far smaller than the global market for potato chips.

Now there is an explosion of new currencies, based on attention, reputation, quality of content, participation, contracting, and more. A company called Colu is making local currencies for local trading using blockchain technology. Our goal at Lykke is to make many of these currencies tradable on our platform. Today, you’ll find Solarcoin, a token to help foster increased use of solar energy, on our platform. We plan to add many more currencies and tokens in the future.

Custom Currencies
Large international companies and trading groups are working on creating their own cryptographic tokens to trade, so they don’t need to exchange currencies so much. This has been going on informally using credits for many years. Now, using cryptocurrencies, we can make it easy and explicit to keep accounts using the single ledger of a block chain. The air-travel industry is working on their own coin, which will be on the Lykke platform. Others will follow this example.

It’s impossible to make the perfect coin for all needs. A coin that serves the air-travel industry will also leak into the food-service industry, the banking industry, electronic components, and so forth. When we think about custom currencies, we look for ecosystems, where coins can circulate.

Since around 2013, people have taken advantage of the comment field in the Bitcoin blockchain to add additional value. The comment field was meant just for comments, jokes, poetry, and was also used as a notary service. Then people started adding codes to it to designate ownership of all kinds of things: diamonds, gold, collectibles, art, even land. The term for this on the bitcoin blockchain is colored coins: you can take a small slice of a bitcoin that’s worth far less than a penny and add the codes that specify the value the coin carries. Since it’s written to the blockchain forever, and since there’s a paper agreement behind it, the coins can carry as much value as people are willing to exchange. Now, when you trade the coins, the ownership of those assets transfers. Now we have other blockchains, like ethereum, that have even more capabilities for creating tokens of value.

Obviously, this scheme is as weak as the legal framework and agreements behind it. But those frameworks and agreements are getting better; they are now used for many kinds of asset transfer. In some cases, the colored coins specify a legal right. In other cases, the coins have no legal claim. And in most cases, crypto assets exist in a gray area of the law.

Digital assets, like media, are particularly suited to this kind of ownership, as they can be programmed and will execute according to the wishes of their creator. Ujo Music is creating tokens that represent songs, so musicians can be paid for their content directly. Everledger is helping people record their value on the Ethereum blockchain using smart contracts. Lava is making smart tickets for music events. There are now smart coupons, smart property, smart bonds, real estate, homes, insurance, and more.

We expect to see a Cambrian explosion of tokens representing everything from reputations to plantations. Our goal is to put as much of these tokens onto our exchange as possible, so we can apply all our innovations to a growing universe of digital assets.

Ultimately — and we hope sooner rather than later — digital tokens representing real-world assets will not be known as crypto-assets or digital anything. They will simply be known as assets.

Along the path from startup to trading publicly, companies typically go through several rounds of funding from “sophisticated” investors. Each round can have significant legal and compliance costs. From venture capitalists to private-equity firms to investment banks, the system is slow, burdensome, and expensive.

A new wave of equity funding has started in countries where it is legal. New companies either write a white paper or start going, and sometime early they go public by selling their shares as tokens on the blockchain. This is called an ICO — initial coin offering. The Lykke ICO took place in September/October, 2016, raising about 1 million CHF. Each week now, there are one or two new issues.

In this new approach to raising equity capital, a company trades the private stepwise approach of raising money from rich people for a public, more gradual approach. Employees and customers can be investors. Rather than getting preferred shares, investors get liquidity. A handful of small companies is starting to sprout up to help and enable these new issuers. Lykke is one of those. We hope to help many companies create, register, and issue their securities on the Lykke platform, as we have done with our own stock.

This public approach could be approved by regulators in countries like the United States and Canada. It could replace much of the venture capital industry. And it could even make the private-equity world more efficient, offering more liquidity and more potential for companies to keep the capital they raise, rather than having it skimmed off by PE funds and investment banks. On Wall Street, it costs $10 to $20 million to go public. On the blockchain, it costs almost nothing.

Eventually, we hope to attract thousands of companies to list their equity tokens on the Lykke platform. Price discovery, trading, transparency, and settlement will all improve dramatically. This is the future of equity exchanges. Download your wallet and start trading today.

Ethereum is a blockchain designed to support general-purpose smart contracts. Like bitcoin, it has its own currency token, called ether. Lykke will soon support native ether on the exchange. When you purchase ether that transaction will settle on the Ethereum blockchain. There will also be Ethereum-based tokens. We will update this page when these features launch.

Our goal is to build a worldwide exchange that complies with regulations in all countries. At the same time, we will be working with and encouraging regulators to adapt to the digital world that generally evolves faster than regulators can keep up with.

The era of smart contracts ushers in an era of smart legislation and e-regulation. Just as with identity and trust, we will need a new kind of digital regulatory environment, one that ties into all the digital ecosystems Lykke spans. This is the challenge of the next generation of regulators, and they are starting to realize how important it will be.

There will be trend-setters, followers, and those who wait and see. Our goal will be to work with the trend-setters, use regulatory sandboxes to test new ideas, and build toward a more open, egalitarian financial system that is not designed to favor large institutions but instead empowers individuals around the world.

For an exchange like ours, there are four critical numbers that will need to scale up to our projected commercial volume:

  • Speed of execution
  • Number of transactions per second
  • Settlement time
  • Number of settlements per day

Here we discuss each of them briefly, to let you know we’re working on them.

Speed of execution
This is measured in the time for a round trip of placing an order and getting confirmation back. Currently (fall 2016), the shortest time for a trade is fairly fast for a normal-sized trade. The speed of trades will be limited more by liquidity than by technology.

Generally, all received coins can immediately be reused in a new trade immediately. Thus, trading can be as fast as the connection between a trader and the exchange permits (normally in the range of 10ms — 100ms).

Number of Transactions per Second
It’s possible that the number of trades can exceed the current limits of the bitcoin blockchain, leading to a delayed settlement. Currently, the bitcoin network has a theoretical limit of about 780’000 transactions per day. Today, the network processes about 200’000 transactions per day, mostly during daytime trading hours in the US and Europe. At night, it’s fairly quiet. Over time, we expect the blockchain core team to implement new protocols to allow it to scale up to hundreds of millions of transactions per day.

However, we are not waiting for them. We are already working on solutions called payment channels, where people who trade often can do so using a local ledger and then “net out” to the blockchain later, when things are quieter. We should be able to execute many more transactions on our platform than the blockchain can “handle,” and the settlements will catch up as the blockchain allows.

Settlement Time
Currently, settlement for bitcoin-based transactions takes place within an average of ten minutes and is confirmed, on average, within about an hour. At certain times, these can stretch out, but there is very little risk of a denied settlement. On the Ethereum blockchain, settlement is usually within two minutes, and confirmation within ten minutes or so.

Number of Settlements per Day
The bitcoin blockchain does have scalability issues; many people are working on various solutions. We expect the settlement capacity to increase dramatically over the next few years. We don’t foresee a problem with settlement volume. The bigger problem is getting prompt settlements.

Lykke is more than an exchange, more than a platform. It’s a central nervous system for what comes next. In fact, what comes next won’t have a central nervous system, and that’s the point of what comes next. The 21st century is all about autonomy, mesh networks, swarming, and continuous operation with no downtime. Our systems will do most of the work for us. The Lykke ecosystem will be an important part of many people’s everyday lives. It will be built into everything from phones to drones, from cash registers to hotel room doors.

And that’s where you come in. Have an idea for what you want to build on top of or next to Lykke? We invite you to join our community and build whatever you think will be needed next. We want to tie into your project, to make sure people get fair prices, fast execution, and keep control of their assets.

Lykke is an open-source exchange. Can you think of a use for an exchange in another project? Can we have several exchanges around the world that tie together? It’s our gift to the tech community — think how you can not just improve it but adopt it to your needs and create something new.

In 2017, we will launch a competition platform, to let people around the world bid, compete, and collaborate. We want your ideas to take us further. Please explore our world and join our conversations.

We plan to implement our agent-based liquidity algorithm in early 2017. This algorithm likes toxic order flow and liquidity traps. It operates best at world scale. We plan to roll it out on some of the larger trading platforms and show that it can make money continuously. Once we have done that, we plan to raise liquidity funds to give investors a consistent return. When the Lykke platform has enough volume, we will install it and let it work its magic there. Our capacity for liquidity funds exceeds $1 billion. We hope to show institutional investors that these kinds of funds provide strong returns.

Today’s stock exchanges were born as physical trading places and pits. They have made the jump to digital, but in truth they have mostly just digitized their old paper-ticket trading flows. They are restricted, charge significant fees, hide the most pernicious actors, encourage front-running, and take days to settle. We hope to fix all these problems with real-time trading and settlement, full transparency, an open-source exchange, a better matching engine, and much more. In this century, we believe a smart-phone app and open-source infrastructure can replace the entire apparatus of stock exchanges worldwide. And this is what we humbly plan to do, one step at a time.

Most people lose money in the markets. Those who don’t lose money themselves pay professionals to lose money for them. There are many misconceptions about markets. Combined with cognitive biases and overconfidence, they add up to losses for most people.

Many digital markets are generally a game of wolves and sheep. A single wolf can eat a lot of sheep, and where there are many sheep you will find a hungry pack of wolves, usually feasting on the hapless sheep. How many times have you watched the price of something, only to say to yourself, “Well, if it dips down just a little bit more, I’ll buy it,” or “I’m waiting for a correction.” Yet markets are also extremely efficient. Our mental models of what should happen in the future are often based on false assumptions. And the second we get in, a wolf is waiting for us.

Uncertainty vs Risk
It helps to understand that uncertainty and risk are not the same thing. If a currency or stock is volatile, we may be uncertain about its future price. In this case, uncertainty represents both volatility (the amount the price moves up and down) and skewness (chances of something very bad — or good — happening that moves the price to an entirely new level).

Risk is relative to the holder. Risk is the chance of a negative event happening to your portfolio. There may be uncertainty about the price of Bitcoin — that is, it could go up or down. If you own bitcoin, you have a risk of losing if the price goes down. However, if you’re short bitcoin, then you’re quite happy for the price to go down, and the more it goes down, the better. So your risk is relative to your position. You could even do nothing, and in that case you may risk missing an opportunity, regardless of the volatility.

Systemic Risks
We would like to fix structural problems in the markets. If you look at markets from an airplane view, you will see many more longs than shorts. This makes for the asymmetric situation characterized by George Soros’ quote: “Markets go up slowly and down fast.” This “black swan” effect comes because so many participants take long positions only.

Our matching engine rewards double-sided orders. This encourages market making. At what price would you switch to become a seller? If we can encourage more balance between long and short, then we can smooth out the vertical drops and surprises.

We want to create a market with fewer systemic risks. If you’re interested in this, please join our community and tell us how you can help.

Dynamic Hedging
Because everything on the Lykke platform is available continuously, 365/24/7, we’ll be able to offer more capable tools for risk management than we have had access to before. Hedging is a form of insurance — a transfer or limitation of risk. Professional traders use big-data analytics and second-by-second risk management that adapts to conditions. We hope to see companies providing such tools on our platform. The more people hedge, the fewer big surprises there should be.

In the same way that a central counterparty helps simplify other markets, a central-counterparty currency may emerge on the Lykke platform. Rather than people trading pairs back and forth, crossing the spread each time, people may gravitate to a “lingua franca” coin. For example, if you want to buy yen, you may buy them from several people who don’t want the currency you’re paying. This results in an inefficient market.

Instead, if everyone maintains reserves of a central coin, you could buy and sell what you want by paying the central currency. It doesn’t reduce risk, because counterparty risk on the Lykke platform is practically nonexistent, but it does help markets function better:

  • Since many people hold the central currency, most trades execute quicker.
  • There will be fewer spreads to cross, resulting in more efficiency.
  • There will be more liquidity, as the liquidity will flow to the central currency.
  • Market makers will converge on the currency, where the “action” is, and spreads will come down.
  • Thinly traded pairs will simply convert to pairing with the central coin.

We don’t know which coin on our platform may become the next central currency, but it would be interesting to see if this phenomenon emerges. Have a suggestion for helping make it happen? Join our community and get involved!

Central banks have a role to play in our economy, but too often they are in their own bubble of beliefs and blinded by reality. Most central banks have a dual mandate, to maintain a reasonable level of inflation and a reasonable level of employment. They also have to deal with exchange rates. This is confusing, because at times these objectives go in separate directions. The main (but not only) tool central bankers use to achieve their goals is an overnight interbank lending rate that in the US is called the Federal Funds Rate. Changes in the discount rate set expectations in the rest of the bond market’s yield curve and affect everything from commodity prices to stocks to various exotic strategies. As we will see, this is a very blunt instrument that often backfires.

Because the worldwide economy is a complex adaptive system, central bankers generally don’t understand how the levers and gears of the economy work. This causes them to push the wrong ones back and forth unnecessarily. First, we’ll look at the problem of interest rate setting, then we’ll look at the big picture.

Problem: 95 Percent of Trading is Intraday
When the FED changes the discount rate, that affects most rates along the yield curve, all the way out to thirty-year mortgages. It’s not so much the actual rate change but the signal from the FED that they are trying to loosen or tighten the money supply. That sets off a wave of reactions in the markets. The FED doesn’t change rates often, so when it does, there’s usually a big knock-on effect.

But the FED may not realize that 95 percent of trading, both in currency and stock markets, takes place intra-day. That’s right: 95 percent of trading isn’t affected by the discount rate, at least not directly. In fact, there is an effect, and it’s one central bankers may not have understood.

Problem: The Intraday Reaction
If a currency is weak, central banks increase interest rates to encourage foreigners to buy. We don’t see this as much with the major currencies, but it’s common when inflation is high (examples: Argentina, Mexico, Madagascar). Unfortunately, increasing the overnight interest rate has the same effect as we saw in the Intraday Interest chapter — it causes even more intraday shorting. Decreasing the rate causes less intraday shorting. This is because traders can borrow for free and short a high-yielding currency, while buying the currency incurs more risk. This wouldn’t be a problem, except for the fact that 95 percent of all trading is intra-day.

Central bankers aren’t aware of these effects. The tools they use are very blunt. They go for months not saying anything, then they suddenly raise rates by 1/4 of a point, and it causes turmoil and uncertainty in the markets. This is magnified by various financial crises that build up and then fail catastrophically. The central bank under assault will be forced as a measure of last resort to increase the one-day interest rate, which in turn is a harsh break on the economy. Typically, when a central bank hikes interest rates, the price of its currency plummets. Because more than ninety percent of trading is intra-day, and because traders pay no intra-day interest, the central bank action provides an added incentive for the intra-day traders to short the currency. This is like flying a plane with inverted steering — trying to go up causes the nose to dive, and vice versa. So initially, the action of increasing the interest rate will actually make things worse, and the currency will drop even faster. Central bankers who don’t understand what’s going on will then raise interest rates even more, causing an overshoot and a stalled economy. We have seen this happen many times — when a country far from Wall Street has a currency crisis, New York traders are able to make a good intra-day profit while paying no interest at all. Then they go home to their families and watch TV news, where they see people far away who aren’t able to pay their rent or buy groceries.

Solution: Shorter-Term Interest-Rate Adjustments
If the yield curve starts with the one-second interest rate, a central bank can increase rates as small as for one second. This gives bankers a wider range of options to work with. Rather than adjusting (or not) on a quarterly basis, they could go to weekly or daily adjustments, and make changes to much shorter-term rates than just overnight.

Because markets are complex dynamic systems, it makes sense for central bankers to have sharper tools. The one-second interest rate is a factor of 86,400 (number of seconds per day) away from the daily interest rates, thus there is a lot of time for the change to seek a new equilibrium level. An intra-day yield curve puts everyone on a level playing field, stripping out the reverse effect. A continuous process creates smoother market transitions.

Problem: Central Bankers Are Torn Between Adjusting for Inflation vs Unemployment
At the macro level, central bankers try to steer the economy using a small number of very heavy tools, sometimes to achieve conflicting results. This is the equivalent of driving down the street by bumping into the gutters, back and forth, overshooting the centerline each time. There is good evidence, for example, that the US banking crisis of 2008/9 was turned into a worldwide financial crisis by central bankers who didn’t understand they were in a tight-money situation and needed to loosen aggressively. This is the problem with human judgment — it’s often too difficult to get people in a room to agree on a complex topic. If they had algorithmically applied the right tools at the right time, we could have avoided years of worldwide slowdown and unemployment.

Solution: NGDP Level Targeting Using Prediction Markets and Algorithms
Just as we can give central bankers a more accurate instrument in the one-second interest rate, we can also give them a better target to steer by: NGDP. By changing from inflation vs unemployment to nominal GDP, central banks can manage both targets simultaneously, and in the appropriate proportion, which changes according to the situation. The magic of this is that central bankers only have to watch one number and can forget about conflicting directives.

It turns out, we can do even better than asking the open-market committee to watch NGDP. We can make the adjustments algorithmically, using a prediction market and level targeting. This has the effect of driving straight down the middle of the road, making small continuous adjustments to stay centered at all times. You can learn more about this from Scott Sumner, who has a paper and a talk about it. Once we have this in place, the prediction market will give a good signal of where NGDP is likely to be six months later, and we can set monetary policy simply by automating the response to this single signal. Central banks can use intra-day interest rates to keep fine tuning and getting feedback from the markets. In this future, the committee has much less to do, no one has to guess what’s coming, and incremental adjustments make it easy to stay on course.

Decentral Banking
We end this section with a short vision of where world currencies could go. We believe in the decentralized revolution — the power of many people acting in real time to solve problems using technology and mesh networks. Blockchain technology has the potential to revolutionize many industries, and central banking is one of them. It’s entirely possible that in just one generation we’ll see most fiat currencies go fully digital, and many could use something like today’s blockchains. In this future, money supply and interest rates are set not by committees and institutions but by markets and algorithms making small, continuous adjustments. We hope the principles we have presented here provide the basis for a new approach to the world financial system. If you find it inspiring, we hope you’ll join us to help us build this new infrastructure and continually improve it.

The average price movement depends on the granularity of the measurement. Imagine a financial instrument that stays in a narrow range of one percent up and down for a year. If you look closely enough, the price can actually vary by up to six percent per day! This is the fractal nature of trading: the closer you look, the more relative volatility (noise) you see. But just as there are organisms living near heat vents at the bottom of the ocean, there are plenty of traders working in this 6-percent range — it only takes 1/10th of a percent per day to be a world-class investor. That’s why high-frequency algorithm-based trading can be so profitable — because the relative moves are every bit as profitable in 1/10th of a second as they are in ten years of investing.

If you have read about our liquidity algorithm, you understand that agent-based trading can “fill in the gaps” as orders are filled, and this is done organically, with many small trader-bots taking small bites of an order. The same is true of trading in general. We see that play out on the major exchanges today.

However, coastline trading is different. Coastline trading uses intrinsic time to capture both trends and mean reversion opportunities. Coastline trading is trading based on the twelve scaling laws we have discovered.

Absolute vs Relative Frame of Reference
We believe too many investors look at the world from a normative point of view — that which should be. They often fixate on big round numbers and targets. They think of mean reversion as going “back to normal” to some fixed frame of reference.

We have learned that in complex adaptive systems, everything is relative. It’s better to look at where a price is now than where you think it should be later. As Richard says, “Every time the ball moves, the entire field moves as well, and the ball stays in the center.” Football (soccer) provides an instructive analogy: for some period of time, the ball goes back and forth between players in a fairly small area. Then, it is kicked 20–50 meters, and everyone goes to the new location. However, the goals haven’t changed. The ball is now farther or closer to one of the goals. In relative trading, the entire field moves and there are no fixed goals.

If you take this seriously, you’ll see that mean reversion in this context means a reversion not back to the same price as before but to the same amount of volatility as before. By understanding the scaling laws and using a relative frame of reference, you’ll start to see trading patterns you couldn’t have seen before. This is why we believe “technical” traders who use various visual patterns are unsuccessful in general, but that relative coastline trading has a far higher chance of being profitable in the long run.

Thinking in Intrinsic Time
To trade the coastline, you must think in intrinsic time. This gets some getting used to. We’ll use the overshoot law as an example.

In normal time series, one tick is a tick of the clock. Ticks are spaced evenly according to a pre-defined time scale. You may have noticed that time series look very similar whether you’re looking at the scale of one minute or one month. This is the fractal nature of trading.

Now let’s switch to defining a tick as the time when a price changes by a pre-defined amount. If a trendline reverses and goes back more than x percent, then we call that a tick. This picture illustrates:

What’s driving this diagram is the pre-set percentage-change threshold. Each threshold has its own set of “ticks,” just as a time scale would. Most people would consider the “bottom” to be significant, but to a coastline trader, it isn’t significant until there’s a reversal of a certain magnitude — otherwise, it’s noise. And here you can see the first scaling law in action: Whenever there’s a reversal of magnitude x percent or greater (the diamonds), the trendline usually continues another x percent or so — on average. So if x = 1.7 percent, then when there’s a 1.7 percent reversal, it’s usually followed by a further 1.7 percent move in the same direction. This law holds for most markets at most price-change scales, from small to large. You can only really see this if you’re looking at intrinsic ticks, not at physical time. This first scaling law is called the overshoot law.

Coastline Trading — How to do it
Now that we understand the overshoot law, we can trade on it and hope to do better than the traditional technical traders. We don’t expect the overshoot to happen every time, but we do expect it to happen enough to make a good return.

Lykke wants people to know about coastline trading and take advantage of it. This will smooth out the markets for everyone. We will develop tools and indicators to make it easier for traders to get the signals they need to trade all coastlines in the Lykke ecosystem.

[Editor’s note: This is a work-in-progress. We are working on the scaling-law description and will include a link here when it’s ready. To learn more about the tools we’re building for coastline traders, come to Lykke Labs.]

This, so far, is the Lykke platform. There’s more coming. We want to build a solid foundation for the worldwide financial system of the 21st century. We want to enable better payments and remittances. We are issuing tokens to help people save energy and buy music. We want people in the developing world to have access to world-class financial services. We are starting a nonprofit foundation for our philanthropic efforts. We are interested in supporting you in making Lykke the future financial platform for all people, worldwide.

To keep up with us, come to Lykke.com and join our community and join the conversation!

Richard Olsen and David Siegel