Lessons learned: Taurus and the ASX blockchain integration

image by Tamarcus Brown via StockSnap
image by Tamarcus Brown via StockSnap

London, 1993. A big decision was about to be made, that would send ripple effects across Europe and forward through time, acting as a warning against ambition and consensus.

For the past 10 years, the London Stock Exchange had been working on a significant upgrade of its securities settlement system. With paper-based systems groaning under the 1980s boom in share ownership, pressure was building not only from nimbler competitors but also from the regulators across the Channel. If London wanted to maintain its role as the continent’s money centre, it needed to upgrade.

The new system was called Taurus, and its goal was to remove as much physical documentation from the system as possible. It also planned a move to rolling settlement, reducing the payment period for equities from three weeks to three days.

Yet things were not going well. The first sign was the rhythm of missed deadlines.

From the outset, the project was complicated. It aimed to include as many sector stakeholders as possible, in spite of conflicting interests. Institutional investors wanted a fast, reliable service, while private investors wanted lower costs. Also, the existing registrars (dominated by large banks) were given a say in the development of a centralized registry, even though it would undermine their business model. Well into the development cycle, they torpedoed the idea.

What went wrong?

In the haste to get development off the ground, the project allegedly started without a clear roadmap. And delays gave more time for the various stakeholders to add requirements.

Even with clear and stable stewardship, that scale of development would have been tough. Yet the project management structure was not clearly defined, and the lack of centralized control meant that interlocking pieces were being developed out of sync, with sections of the process at different testing stages, while other functions had not yet been designed.

Also, given the long lead time (which ended up being more than double the initial estimate), the system – if launched – would already have been behind the competition from day one.

The final straw came when an investigation in 1993 revealed that completion would take another two to three years, at double the cost-to-date.

The decision was taken to scrap the whole project. The exchange’s investment of over £70 million (over £140 million in today’s money) was lost. The London Stock Exchange handed over responsibility for the development of a new stock trading system to the Bank of England, and its CEO resigned.

It wasn’t just the colossal waste of money and the damage to its reputation that made many fear for the exchange’s future. Hundreds of brokers had based their systems development on the assumption that Taurus would be the main platform, and thousands of employees had been trained. The total cost to London’s financial centre was estimated to be in the hundreds of millions of pounds.

Of course, it’s easy to see in hindsight where things went wrong. And it’s easy to believe that today, big systemic projects would be managed with different principles.

While that may be the case, the fate of Taurus serves to highlight the colossal complexity of introducing a new systemic platform. Throw in a technology that has yet to be tested “in the field”, and you have a potential powder keg of risk.

All change

I’m talking about the decision of Australia’s primary securities exchange, ASX, to upgrade its clearing and settlement platform to one based on distributed ledger technology.

Announced late last year, the news sent waves of excitement through the blockchain sector – it would be one of the first major public-facing applications of the technology, which many have touted as having the potential to decentralize finance.

Introduced with bitcoin, the blockchain offers a way of sharing data that removes the need for validation from a central authority. The elimination of redundancies and the speed with which information can be transmitted and acted on present significant cost reductions, especially intriguing in an era of diminishing margins and increasing competition in the financial sector.

It’s not yet clear whether the technology that ASX will use (developed with blockchain startup Digital Asset) will technically be a blockchain, in which information is stored in blocks that are irrevocably linked to previous blocks, ensuring data integrity. The official press release referred to “digital ledgers”, and while the two terms are often used interchangeably, some distributed ledgers don’t rely on linked blocks to share and verify inputs and outputs. However, since the boundaries of the new technology are being blurred as the concept evolves, the announcement was treated as a triumph by blockchain sector participants – official, public validation of the potential benefits.

Be careful

And yet, it is by no means the windfall that the headlines proclaimed.

First, it isn’t happening anytime soon. At the end of March, the ASX will reveal a potential live date for the new platform – it will most likely be years away. We won’t get a clear indication of the expected timing until the end of June.

And, as we saw with Taurus, in complex undertakings, deadlines are often extended. Hopefully the new system will be revealed within a much shorter timeframe than the failed British attempt’s estimated 13 years…

If it gets revealed at all. The ASX platform does need to be replaced – known as CHESS, it is 25 years old and is struggling to keep up with newer and nimbler competitors. But the decision to build on top of a relatively untested technology with uncertain scaling and bottlenecks is a brave one. And few development projects progress without setbacks.

It’s fair to assume that the planning will be meticulous and thorough. But will it manage to avoid the pitfalls of overwhelming systemic change?

Learning from the mistakes of Taurus will help. But the leap forward in technology with this development adds a new layer of complexity.

A large part of the problem will be managing expectations. While “blockchain” has been hailed as “the next industrial revolution”, we are not going to see a new decentralized stock exchange emerge before our eyes. As far as the public is concerned, things will continue pretty much the way they are.

For the financial and technology sectors, though, it is a big deal. If all goes well, back office costs will be reduced, new efficiencies will be explored and distributed ledger technologists will learn much from the real-world rollout.

The true change, however, will come years down the road, as other exchanges around the world take a look at their own clearing and settlement processes, as regulators encourage compatibility and connectivity, and as frictionless cross-border trading finally begins to look like a possibility.

But first, the ASX system needs to be successfully launched. And, as we’ve seen, it’s nowhere near as easy as it sounds. While the decision to migrate a country’s main securities settlement and clearing platform to a distributed ledger is good news for the blockchain sector, it is too soon to celebrate.

Bitcoin futures and the meaning of finance – how did we get here?

Photo by Brandon Morgan on Unsplash
Photo by Brandon Morgan on Unsplash

One outstanding note in the cacophony of the bitcoin futures debate is an intriguing claim that I confess I didn’t understand at first: that bitcoin has no “natural sellers”. What’s unnatural, I thought, about people wanting to sell to realize profits? It turns out that’s not what the phrase means.

It means that nobody needs bitcoin. So why hedge it?

To go deeper, let’s look at why capital markets exist. They were developed to enable firms to raise money outside of bank loans. Bonds and equities pair those that need funds with investors who want a return.

Derivative markets emerged to protect cash flows. This both gives producers more security, and helps to raise funds – investors are more likely to “lend” to a company with protected income than to one subject to the vagaries of nature.

In essence, capital markets exist to help businesses flourish. Old-school capitalism.

Here’s where the “natural seller” part becomes important.

Farmers need to sell wheat. It’s what they do. Oil producers need to sell oil. Steel manufacturers need to sell steel. Gold miners need to sell gold. So, they all should protect those sales in the derivative markets.

No-one needs to sell bitcoin.

So what income flows are the derivatives protecting? Mutual fund redemptions, maybe. Pension plan payouts. But do we really think that mutual funds and pension plans should have significant exposure to bitcoin?

This question is important for whatever side of the bitcoin debate you’re on. If you’re a sceptic and think that it’s all a ponzi scheme, surely you don’t want institutional funds heavily invested in an asset that will no doubt crash. If you’re a bitcoin believer, do you really want the “money of the future” stuck in funds? Where’s the decentralizing potential in that?

So, it could be that the constructive purpose of bitcoin derivatives is to protect flows for funds that are either taking irrational risks or hijacking the finance of tomorrow. This is a far cry from ensuring that farmers can make a living and oil producers don’t go bust.

We could argue that all this started to go awry back in the ‘80s with the creation of synthetic derivatives that had as their sole aim to make a profit at the expense of others (trading being a zero-sum game). We could also argue that back then we got ahead of ourselves by letting markets run far ahead of the infrastructure. We know what happened next. (Ok, I’m simplifying, but the point still holds.)

And we could ask ourselves what good bitcoin futures will do the economy as a whole. To what productive use will their markets contribute? Are they adding stability, as per the original intent of derivatives? Or could they be adding yet another layer of complexity that masks a deepening fragility?

Of course, playing the long game, this could be what true bitcoin believers have known would happen all along. That the world will see (again) how unstable the current financial system is. And to what will people turn when widening cracks send central banks scrambling?

True, the bitcoin price would also likely tumble. But the technology would still work. People would still be able to independently transfer funds. And the advantage to having an alternative to an interconnected and unstable system would become more apparent than ever.

The threat of bitcoin futures

photo by Jesse Bowser on Unsplash
photo by Jesse Bowser on Unsplash

The financial press has been in a flutter of excitement over the launch of bitcoin futures trading on not one but two reputable, regulated and liquid exchanges: CME and Cboe.

CME Group (Chicago Mercantile Exchange) is the largest derivatives exchange in the world, as well as one of the oldest, with roots going back to the 19th century. It will launch bitcoin futures trading on December 18th.

Cboe Global Markets owns the Chicago Board Options Exchange (the largest US options exchange) and BATS Global Markets (the platform on which the Gemini-backed bitcoin ETF would have been listed had it been approved). It plans to beat CME to the punch by launching bitcoin futures trading on December 10th.

In theory this opens the doors to institutional and retail investors who want exposure to bitcoin but for some reason (such as internal rules, or an aversion to risky and complicated bitcoin exchanges and wallets) can’t trade actual bitcoin.

And that expected flood of interest is, from what I hear, part of the reason that bitcoin’s price recently shot past $11,000 (which, considering it started the year at $1,000, is phenomenal).

I’m missing something. I don’t understand why the market thinks there will be a huge demand for bitcoin itself as a result of futures trading.

First, a brief primer on how futures work: let’s say that I think that the price of xyz, which is currently trading at $50, will go up to $100 in two months. Someone offers me the chance to commit to paying $80 for xyz in two months’ time. I accept, which means that I’ve just “bought” a futures contract. If I’m right, I’ll be paying $80 for something that’s worth $100. If I’m wrong, and the price is lower, then I’ll be paying more than it’s worth in the market, and I will not be happy.

Alternatively, if I think that xyz is going to go down in price, I can “sell” a futures contract: I commit to delivering an xyz in two months’ time for a set price, say $80. When the contract is up, I buy an xyz at the market price, and deliver it to the contract holder in return for the promised amount. If I’m right and the market price is lower than $80, I’ve made a profit.

Beyond this basic premise there are all sorts of hybrid strategies that involve holding the underlying asset and hedging: for instance, I hold xyz and sell a futures contract (I commit to selling) at a higher price. If the price goes up, I make money on the underlying asset but lose on the futures contract, and if it goes down the situation is reversed. Another common strategy involves simultaneously buying and selling futures contracts to “lock in” a price.

Futures contracts currently exist for a vast range of commodities and financial instruments, with different terms and conditions. It’s a complex field that moves a lot of money. The futures market for gold is almost 10x the size (measuring the underlying asset of the contracts) of the physical gold market.

How can this be? How can you have more futures contracts for gold than actual gold? Because you don’t have to deliver an actual bar of gold when the contract matures. Many futures contracts settle on a “cash” basis – instead of physical delivery for the sale, the buyer receives the difference between the futures price (= the agreed-upon price) and the spot (= market) price. If the aforementioned xyz contract were on a cash settlement basis and the market price was $100 at the end of two months (as I had predicted), instead of an xyz, I would receive $20 (the difference between the $100 market price and the $80 that I committed to pay).

Both the CME and the Cboe futures settle in cash, not in actual bitcoin. Just imagine the legal and logistical hassle if two reputable and regulated exchanges had to set up custodial wallets, with all the security that would entail.

So, it’s likely that the bitcoin futures market will end up being even larger than the actual bitcoin market. That’s important.

Why? Because institutional investors will like that. Size and liquidity make fund managers feel less stressed than usual.

The bitcoin market seems to be excited at all the institutional money that will come pouring into bitcoin as a result of futures trading. That’s the part I don’t understand.

It’s true that the possibility of getting exposure to this mysterious asset that is producing outstanding returns on a regulated and liquid exchange will no doubt entice serious money to take a bitcoin punt. Many funds that are by charter prohibited from dealing in “alternative assets” on unregulated exchanges will now be able to participate. And the opportunity to leverage positions (get even more exposure than the money you’re putting in would normally warrant) to magnify the already outrageous returns will almost certainly attract funds that need the extra edge.

But here’s the thing: the money will not be pouring into the bitcoin market. It will be buying synthetic derivatives, that don’t directly impact bitcoin at all. For every $100 million (or whatever) that supermegahedgefundX puts into bitcoin futures, no extra money goes into bitcoin itself. These futures do not require ownership of actual bitcoins, not even on contract maturity.

Sure, many will argue that more funds will be interested in holding actual bitcoins now that they can hedge those positions. If supermegahedgefundX can offset any potential losses with futures trading, then maybe it will be more willing to buy bitcoin – although why it would allow its potential gains to be reduced with the same futures trade is beyond me. And, why hold the bitcoin when you can get similar profits with less initial outlay just by trading the synthetic derivatives?

That’s the part that most worries me. Why buy bitcoin when you can go long a futures contract? Or a combination of futures contracts that either exaggerates your potential gains or limits your potential loss? In other words, I’m concerned that institutional investors that would have purchased bitcoin for its potential gains will now just head to the futures market. Cleaner, cheaper, safer and more regulated.

So, if the market is discounting an inflow of institutional funds into actual bitcoins, it’s likely to be disappointed.

What worries me even more is the possibility that the institutional funds that have already bought bitcoin (and pushed the price up to current levels) will decide that the official futures market is safer. And they will sell.

Now, it’s possible that the demand for bitcoin futures and the general optimism that seems prevalent in the sector will push up futures prices (in other words, there will be more demand for contracts that commit to buying bitcoin at $20,000 in a year’s time than those that commit to buying at $12,000 – I know, but the market is strange). This will most likely influence the actual market price (“hey, the futures market knows something we don’t, right?”).

And the launch of liquid futures exchanges increases the likelihood of a bitcoin ETF being approved by the SEC in the near future. That would bring a lot of money into an already crowded space.

Buuuut… it’s also possible that the institutional investors that are negative on bitcoin’s prospects (and there’s no shortage of those) may use the futures markets to put money behind their conviction. It’s much easier to sell a futures contract with a lower-than-market price than it is to actually short bitcoin. These investors may well send signals to the actual bitcoin market that sends prices tumbling.

And the leverage inherent in futures contracts, especially those that settle for cash, could increase the volatility in a downturn.

That’s pretty scary.

Let’s not even go into the paradigm shift that this development implies. The growth of a bitcoin futures market positions it even more as a commodity than a currency (in the US, the Commodity Futures Trading Commission regulates futures markets). And even more as an investment asset than a technology that has the potential to change the plumbing of finance.

So, while the market appears to be greeting the launch of not one but two bitcoin futures exchanges in the next two weeks (with two more potentially important ones on the near horizon) with ebullience, we really should be regarding this development as the end of the beginning.

And the beginning of a new path.

The envelope, please… Blockchain and shareholder voting

Who knew that shareholder voting could be so… suspenseful?

If you missed the news, Proctor & Gamble has been locked in a bitter battle with shareholder activist Nelson Peltz, who wants a seat on the board. This led to the biggest shareholder battle to date, with over 2.5 billion votes (for 2.5 billion issued shares) in play. Robocalls, social media ads and a flood of mailings… the tactics got fierce.

suspense

At stake is the structure of the consumer goods conglomerate. Peltz – CEO of asset management firm Trian Partners – wants a seat on the board, and to break P&G up into three distinct units, to streamline operations and add flexibility. P&G says that the recent restructuring is already showing positive results, and changing the composition of the board would bring unnecessary disruption.

At the annual general shareholders’ meeting, P&G announced that Peltz’ bid had been defeated. Shareholders had voted to not give him a seat on the board, by a margin of 6.15 million votes, which sounds like a lot but when taken in context of the overall number of outstanding shares, was only 0.2%. A statistician would argue that is well within the margin of error.

And she would be right. Yesterday the FT reported that a recount by an independent expert found that the margin was only 43,000, in favour of Mr. Peltz. Effectively, a dead heat. The final, definitive results are not yet in. But Mr. Peltz could well get his board seat.

Why the lack of clarity in the outcome?

As you have most likely seen with national elections, counting votes is cumbersome, and largely manual. Even today, there is no definitive way to ensure that votes are not double-counted or falsely filed. One of the main problems is collecting all the votes, which are still mainly submitted on paper, either at a company’s annual general meeting, or sent in via physical mail (although some firms allow online voting). Another is making sure that the count is not manipulated. This requires rigorous identity verification, and a decentralized process of tally.

A separate issue is identifying who has the right to vote – with shares held at central depositories and “ownership” represented by a type of cession of rights, this is often not as clear as it should be.

Could blockchain technology, with its security and automation, help? Several large proxy voting managers believe so.

Last year, Russia’s National Settlement Depository announced that it has tested a blockchain-based voting system. The Abu Dhabi Stock Exchange unveiled a blockchain-based voting service that allows shareholders to both participate in and observe the process. And Nasdaq ran an e-voting trial, which recorded stock ownership on a blockchain platform, and issued digital voting right assets and tokens.

Earlier this year, Broadridge – the world’s largest provider of proxy voting infrastructure – revealed that it is building a blockchain platform on ethereum to streamline the sharing of information between custodians. A pilot run was successfully executed (in parallel with voting using traditional software) with JP Morgan, Northern Trust and Banco Santander.

Around the same time, financial services company TMX group (operator of the Toronto and Montreal stock exchanges, among others) revealed that it had completed a proxy shareholder voting prototype built on Hyperledger

And just last week, a group of central securities depositories (CSDs) announced progress on a distributed ledger proxy voting platform. Swift is among the institutions participating, to assist in ensuring compliance with international financial messaging standards (which would open up the platform to uses other than voting).

Timelines on any of the above projects going into the production are at the moment vague. Will any of them even happen? With shareholder voting generally an in-network activity (with limited, if any, need for participation from outside organizations), why use blockchain at all? Why not just go for a robust, efficient database?

Because of the vulnerabilities of centralization, which in many cases doesn’t matter – but when it comes to voting, that’s a different story. First of all, even a distributed database can be hacked and manipulated. Second, shareholders need to be certain that the vote was fair, and that the company in question has not tried to influence the tally. If they are granted real-time transparency into the voting process are more likely to trust the system, and therefore more likely to vote.

Plus, as shareholder voting becomes even more important, propelled by improvements in the technology (reduction of friction) and increased activism, audits of processes by external parties are going to become even more of a regular feature. An access node would facilitate that, as well as reduce the costs.

And finally, proxy polls are not cheap. According to FactSet Research Systems Inc., a “typical” proxy battle costs about $1 million, mainly from printing, mailing and legal fees. For context, the P&G battle is expected to cost the company over $35 million (small change compared to its Q2 income of $2.2 billion, but still…).

And even with that expenditure, it might not win. Final results are yet to come, so the battle isn’t over yet.

But the drama and nail-biting suspense sheds light on the urgent need to reform shareholder voting technology. Current platforms are, in general, inefficient. And electronic voting systems run by any one organization, even audited ones, will always have a cloud of doubt over the controlling interests. The transparency and security of distributed ledger systems could offer a more robust, lasting and scalable solution. Widespread use is still a long way off, though, and they’re unlikely to be practical until the murky issue of stock registration is solved.

While perhaps not the decentralize-the-organizations disruption that blockchain technology originally promised, it would be a step towards a more democratic governance, enabling shareholders to participate in corporate decisions more frequently and with less upheaval. It could end up giving shareholder activists more firepower and motivation, even perhaps going as far as to change what we understand by “shareholder capitalism”. Or capitalism overall, for that matter.

And when it comes to letting the market decide, it must be galling for the P&G board to see the positive price reaction to the news that they might not get their way. If the close outcome of the voting doesn’t send the board a strong message (in other words, when almost half of your shareholders side with your most vocal critic, you’re doing something wrong), perhaps the voice of the market will.

FX trading fines and regulation catch-up

Photo by Ben Rosett on Unsplash
Photo by Ben Rosett on Unsplash

In my last post I mentioned some hefty fines incurred by foreign exchange dealers for trading infractions. The amount keeps climbing.

Yesterday the Financial Times reported that Credit Suisse has just been fined $135m by New York state’s financial regulator for “unsound” conduct between 2008 and 2015. Apparently the traders shared client information with other global banks to manipulate foreign exchange (FX) prices and maximize Credit Suisse profits.

What’s more, the bank was found guilty of front-running (putting your order in just ahead of clients’ orders to take advantage of resulting price movements) between 2010 and 2013.

And (this ties in to my previous post) between 2012 and 2015, traders took advantage of the “last look” feature of their electronic trading platform, which allows dealers to back out of a trade before execution, by applying it to all client trades. To make matters worse, it lied to clients about why trades were rejected.

It’s one thing to use the system to profit your own book over your clients’. It’s quite another to lie about it.

This is one of the reasons for the increasing volume of calls to reform the last look practice. Many traders think it should be banned. Others believe that it should be allowed, but that traders should be honest and upfront with the conditions in which a previously agreed trade would be rejected.

Going back to the Credit Suisse infringements, what blows my mind is that they openly talked about frontrunning and using proprietary information in electronic chat rooms. And get this: one of chat rooms was known as “The Cartel”. It’s not the stupidity that surprises me, it’s the arrogance. If everyone’s doing it, it’s fine, right? Why even try to hide it?

The sentences are coming thick and fast. In September, HSBC was fined $175m by the Federal Reserve for “unsafe and unsound” FX trading practices. In July, the Fed ordered BNP Paribas to pay $246m for charges relating to its FX conduct between 2007 and 2013. These fines follow others of more than $5.7bn levied on a handful of major banks in 2015 by the US Department of Justice, and over $3bn handed down by the US Commodity Futures Trading Commission and the UK’s Financial Conduct Authority in 2014.

While we may be blinded by the volume of fines, we need to put them in context of the overall size of the FX market. The largest market in the world, it trades over $5tn per day. Apart from the massive profits the rogue traders earned for their banks (no doubt largely reflected in handsome bonuses), the fines also reflect the gravity of the infringements. Damage to its reputation and a loss of trust would pose a risk to global commerce and trade.

The high profile of these cases could add momentum to the move towards trading platforms that offer greater transparency to clients and to the regulators.

The FX market, already of systemic importance, is likely to expand as world trade continues to grow. Until recently, large clients didn’t have much of a choice – to get big deals done, you went to the big dealers. Now, however, newcomers with shinier platforms are nibbling away at market share. Increasing compliance adds to costs, and the advantages of largely manual, opaque and relationship-based execution are becoming less apparent. Especially with increasing scrutiny.

The embarrassment for the regulators at the revelations that this was going on for as long as 10 years before anyone noticed will surely give them a good incentive to push for better reporting and greater access to trading records.

So, distributed ledger-based trading platforms in which the regulators have a node that allows them to see in real time what’s going on? Confidentiality issues aside (because they can be solved), it is likely to happen in some form.

Regulators monitoring electronic chat rooms? That’s a different story.

FX trading and last look doubts

photo by Veri Ivanova for Unsplash
photo by Veri Ivanova for Unsplash

The murky world of foreign exchange trading could be about to get a bit more transparent.

A few days ago, the sovereign fund of Norway (NBIM) – the largest in the world, with over $1tn under management – published a report calling for improvements in the foreign exchange (FX) market. It feels that the market’s friction and opacity tilts profits unfairly towards the dealers, and that the lack of transparency is weakening trust in the system.

It points to three particular aspects of FX trading, specifically: last look, algorithms, request for quote feeds and their relationship with interdealer prices.

While each is intriguing and worthy of further digging, I want to take a closer peek at “last look”, since it exemplifies how new technologies can both improve and complicate trading, and how evolving infrastructure requires a regular re-think of established processes.

Looking back

What is “last look”? It’s a “way out” for the dealer, who can renege on an agreed trade if certain conditions are not met. It could be that the client doesn’t pass the credit check. Or it could be that the price moves against the dealer.

This last aspect gives last look the whiff of unfair advantage. Critics claim that it can be used to “cherry pick” trades, only following through on the profitable ones, which would negatively impact market confidence and liquidity. It could also lead to “front running” of trades, whereby information from client orders is used for the dealer’s own profit.

Others argue that its use as a latency buffer – protecting against price moves between order agreement and order execution – is no longer necessary given technology improvements that make that time gap almost negligible. And the lack of information – often clients are not told why their trades fell through – weakens confidence, which could impact order size and even willingness to operate in the market.

Proponents claim that the practice allows dealers to quote better prices – with less risk in a trade, the spread can be narrower, which implies a better deal for the clients.

What’s more, the option of backing out of a trade enables dealers to post their price on several exchanges at once. Without that option, the dealers run a higher risk that the market will move against them. With posts on several exchanges, changing all of them takes time (seconds, but that’s a long time in FX). So, last look encourages a wider spread of trading venues, which in theory enhances liquidity.

Wait a second

Norway’s sovereign fund is not alone in its concerns about the practice, which has been coming under increasing scrutiny.

Vanguard (the world’s largest mutual fund manager), Citadel (one of the world’s largest alternative asset managers) and others have called for its elimination. Several exchanges have echoed that sentiment. XTX Markets Limited, one of the world’s biggest spot currency traders, officially stepped back from the practice in August. The Bank of England has been publicly questioning the practice since 2015.

Global regulators are also taking a closer look. In 2015, Barclays was fined $150m for what was deemed abuse of the practice – not only did the bank filter all trades in which the market moved against it (as opposed to using last look as a sporadic protective measure), but it denied doing so. (2015 was a ripe year for FX manipulation – Barclays was fined a further $2.3bn for other FX infractions, and penalties levied on Citigroup, JP Morgan, UBS, Bank of America and the Royal Bank of Scotland brought the total to almost $6bn.)

However, removing the practice will leave end users vulnerable to predatory manoeuvres, especially given the prevalence of high-speed trading. It could also constrict liquidity as dealers protect themselves against risk.

Rules, please

Why can’t the regulators step in and establish certain rules? Because the FX market is notoriously difficult to regulate, largely due to its cross-border nature. Which jurisdiction would apply?

In an enlightening example of self-regulation, the Global Foreign Exchange Committee (GFXC) was created in May 2017 as a forum for FX market participants (including central banks). Its first act was to issue an updated FX Global Code, a set of “best practices” for the foreign exchange community.

It does not rule out last look, but does ask practitioners to disclose the criteria, in order to allow end clients to make the appropriate adjustments. The GFXC simultaneously issued a request for feedback on the practice, demonstrating a willingness to contemplate adjustments.

So why are the Norwegian sovereign fund and others protesting now? Just two years ago, NBIM publicly came to the practice’s defense, citing its potential to improve available liquidity for investors.

It turns out that their positions are not inconsistent. Even now, it is not advocating the removal of last look. What it wants is more transparency.

Furthermore, it is no doubt aware of the deteriorating levels of trust in FX trading. The previously mentioned scandals and fines are probably the tip of the iceberg when it comes to abuse, especially since the rules have been vague and the FX market is opaque to begin with.

And, the protests could be influenced by fund managers’ need to increase revenue and lower costs through narrower spreads and more transparent pricing. Quoted in the Financial Times, the co-author of the report said:

“We want to be more explicit about the risk sharing between us and the dealer. The client is providing optionality for the dealer. We would like to be rewarded for this option.”

Large market participants no doubt understand that the system is changing, and so are expectations. Calls for market reform are both timely and self-serving, contributing to a cleaner image and hopefully a more robust system.

Looking forward

So, what would a solution look like?

While blockchain technology is by no means the solution to all things financial, it could offer a useful tool for a platform that allows transparency, immutability and decentralized (but permissioned) participation. A major drawback would be the latency – it’s not the fastest way to share data, and the FX market is used to split-second speed.

It is clear that enhanced disclosure is a stop-gap remedy. Once the goal posts start moving, it’s impossible to see where they will stop. What’s more, temporary solutions are not conducive to a lasting realignment of trust. And with self-interest up against community fairness, and a huge economic sector in play, a more durable solution is urgent.

Will blockchain technology end up playing a part? It’s possible, perhaps even probable, especially as new features emerge and work-arounds gain strength. It’s unlikely to be the only solution, though, as database technology and communications infrastructure also continue to evolve. And as long as speed remains a competitive advantage, decentralized resilience and transparency are unlikely to be the main priority.

What’s more, the FX sector is unlikely to see a sweeping change in the near future – it’s just too big and important for that. However, the processes that keep the system running need revision and updating, to continuously improve efficiency and trust.

And eventually, the patchwork of solutions to specific problems will point to a deeper evolution, one that favours interoperability over universality, reliability over speed and trust over profit.

Blockchain and capital markets: interest rate swaps

by Alex Jodoin on Unsplash
by Alex Jodoin on Unsplash

And now on to the next riveting stage of our exploration of the impact of blockchain technology in capital markets. Welcome to interest rate swaps (IRSs), one of the most powerful risk-management tools in the market.

They’re not as complicated as they sound. There are several different types of swaps, but the basic “vanilla” variety works like this: if I am paying a fixed interest rate on my debt and I’d rather pay a variable rate, and if you have the opposite situation, then we swap. Not the actual debt, because that would be either complicated or downright impossible (cross-border regulations, collateral requirements, etc.). What we swap is the payment – I’ll send you the equivalent of your variable interest payments, and you send me your fixed payments. I’ll then use what you send me to keep my lender happy (he wants fixed payments – I pay him with your money), but my actual payout is to you at a variable rate. With that, I have converted a fixed obligation into a variable one.

Why would I want a different interest structure than the one I contracted with my lender? Well, maybe I have a fixed rate, but I think rates are going to come down so I want to switch (and, of course, you think they’re going to go up and so would rather lock in a fixed rate). Or maybe I want a fixed rate but my bank will only offer me a floating structure, and you have the opposite problem.

While mainly used by banks and other financial institutions to hedge their interest rate exposure, IRSs can also be used as a tool for portfolio management, taking positions on interest rates at various points in time. (It’s also possible to do this by going long or short Treasury bills, but interest rate swaps require much less capital outlay.)

Big stuff

The market is huge, trading almost $1.9bn a day, which makes interest rate swaps one of the most actively traded instruments in the over-the-counter (OTC) market.

Interest rate swaps have traditionally traded OTC (directly between two parties) rather than via a regulated exchange – most contracts are drawn up to satisfy particular conditions, and are not standardised enough to list on exchanges. As with credit default swaps, the Dodd-Frank Act of 2010 radically transformed the market in the US, mandating that a wide range of IRS contracts (but not all) be traded on “swap execution facilities” (SEFs), rather than by phone. These newly-created trading venues aggregate order books (increasing market transparency), and allow participants to ask for quotes from several dealers simultaneously – they are similar to exchanges in function, but have a more limited scope and fewer listing requirements.

Also, all swaps traded on SEFs have to be cleared via a central counterparty (CCP). Once a trade is confirmed, the CCP acts as buyer and seller, taking on the settlement risk. This lowers the collateral required of the parties to the trade, but increases the trading costs.

And, all swaps trades have to be reported to swap data repositories (SDRs), providing volume and pricing information to the market. SDRs also enable regulators to gauge participants’ risk exposures.

In Europe, the changes are similar to those in the US. The European Market Infrastructure Regulation (EMIR) – passed in 2012 – mandates that certain classes of interest rate swaps clear through CCPs. Those that are not required to do so still have to comply with tighter risk compliance rules. EMIR also tightened the reporting requirements.

In the US, interest rate swaps fall under the jurisdiction of the Commodity Futures Trading Commission (CFTC), and in Europe under the European Securities and Markets Authority (ESMA).

Connected data

The Chairman of the CFTC, J. Christopher Giancarlo, has often said in public that blockchain technology could have a big impact on how swaps are handled, helping to smooth the complexities brought about by Dodd-Frank (not that the market was straightforward before). However, the blockchain activity going on in this sector is relatively modest, compared to other types of derivatives, which is strange given its size and the potential impact on execution efficiencies.

While the pooling of risk aspect of central clearing would be a complicated area to automate (given the necessary level of flexibility), the redundant processes and documentation requirements could be streamlined via a distributed ledger. And the chaos of data reporting, especially given its systemic importance, points to this area as one likely to attract the attention of blockchain executives and developers.

Also, the actual agreements recorded on a blockchain could be largely automated using smart contracts. For instance, they could pull the interest rates for floating swaps from an established oracle and automatically calculate the relevant payment.

And, the current lack of transparency in the market due to the legacy of OTC trading could be alleviated by putting all swaps on a blockchain platform, and giving the relevant parties (as well as the regulators) access via their own node.

Step forward

UK bank Barclays – together with blockchain consortium R3 and the International Swaps and Derivatives Association (ISDA), a trade standards body that has played a significant role in the standardisation of swaps – have trialled a similar solution. They developed a distributed ledger prototype based on Corda (which is not technically a blockchain, but that’s a different story) with the aim of recreating derivatives agreements using smart contracts.

The scheme envisions ISDA acting as a central repository for smart contract-enabled documents. Swaps dealers could use these to create new agreements, with all counterparties collaborating. The hash (compressed representation) would be uploaded to the distributed ledger, eliminating the need for all parties to store their own set of documents.

While the platform could ostensibly be used for a range of financial instruments, the first example tested was an interest rate swap.

Media giant Thomson Reuters also had interest rate swaps in mind when it designed a data stream (called BlockOne IQ) specifically to interact with smart contracts. The streams of both variable and fixed interest payments could be made much simpler with automatic calculations and adjustments linked to uploaded agreements.

While the firm’s more traditional APIs are available to market participants, a dedicated oracle would open up access to a growing range of decentralized applications. It is expected to reach a decision by the end of the year on whether or not to monetize the experiment.

It’s not just the incumbents that are taking a look at this segment. Synswap, set up by two ex-traders, hopes to challenge current post-trade processes by disintermediating central counterparties from the clearing process. Its initial focus is ostensibly on interest rate and credit default swaps, and a prototype currently in development will perform key post-trade functions such as matching, confirmation, collateral management and settlement.

In plain sight

I am surprised that there isn’t more IRS-focused blockchain activity going on, given its characteristics and needs:

  • They are easily automated
  • The market has a relatively limited number of participants
  • The data collection is complicated (and can be simplified)
  • The accounting is complicated (and can be simplified)
  • There is little overlap with other instruments (which means that dedicated solutions – which are easier to implement – could work)
  • The market is still relatively opaque, in spite of a push for greater transparency
  • Data collection is complicated (and can be simplified), and slow – the latest figures given by the Bank of International Settlements are from April 2016
  • The instrument is systemically important (which implies increased attention from the regulators)

Of course there are complexities that would be hard to integrate into a blockchain, such as the mutualisation of losses and the management of margin levels.

But the potential is significant, and worthy of investigation. It will be fascinating to see what other projects emerge in this space, especially given its importance to capital markets, and the lessons it could impart to the rest of the sector.

A slippery slope

by Matthew Henry via Stocksnap
by Matthew Henry via Stocksnap

I never expected to get riled up over accounting.

Alexandra Scaggs in the FT this morning expertly dissects a paper in the CFA Institute’s Financial Analysts Journal, in which the authors claim that profits are no longer important in company valuations.

As an ex-investment analyst, this is befuddling. I used to be pretty damn good at financial models, all of which were based on earnings projections. And as an entrepreneur, I cared a lot about cash flow.

However, times have changed. The stock prices of Amazon and Tesla do not reflect earnings forecasts (which no-one really has a clue about, anyway) – their price/earnings ratios are crazy. And yet, that does not mean that they are bad investments (this is not professional advice!).

This disconnect could well be why so few tech companies are going public – that, and the relative ease with which they can come by funding through other methods.

I have been getting increasingly concerned about the accounting standards for initial coin offerings (ICOs), and will wail about them more in another post.

But for now, let’s focus on asset value.

Research and development (R&D) costs are considered “expenses”, and hit the bottom line. Capital outlays and acquisition costs are considered “investment” and can be capitalized and amortized over time. So, it makes more sense (from an accounting point of view) to buy a research company than to do it yourself.

But then, assuming you consolidate the figures, and assuming your new R&D team keeps at it, you’re back to having high costs relative to the growth in assets.

Again, this could partially explain why tech companies are increasingly realising that IPOs are not for them.

ICOs, on the other hand… No-one seems to care about fundamentals such as future earnings there.

And in a free market, that is the investors’ choice. But it does represent a fundamental shift from the value-creating roots of corporate participation, which we should be aware of.

We also need to contemplate the long-term impact that this shift could have. When we focus on speculative gain (as with most of the ICO market today), we don’t really care what the company does, or how. We care about how “cool” it is, how “hot” the topic. We veer towards an alarmingly short-term bias which will, if it persists and spreads, affect the investment decisions of the companies themselves.

An economy focused on short-term deliverables and market appeal will become more volatile. This will give speculation a veneer of common sense. And, let’s go big here, possibly undermine the very essence of capitalism.

We need to think about that.

Blockchain and credit default swaps, Part 2 – the application

by Snufkin via StockSnap
by Snufkin via StockSnap

As we saw in the previous entry in this series, credit default swaps are ideal for blockchain testing because:

  • they’re complex yet with a “programmable” structure;
  • they’re increasingly standardised following recent changes in regulation; and
  • they operate in a self-contained market – although they reference other securities, they don’t actually link to them, and can operate solely on straightforward data inputs.

The largest project currently underway – not only in credit derivatives but also in the financial industry as a whole – is that of the Depositary Trust and Clearing Corporation (DTCC) in the US, which is working on rebuilding its credit default swaps processing platform with blockchain technology.

To appreciate how huge the launch could be, let’s take a closer look at the structure of the DTCC and what it does.

Too big to fail

Set up in 1999 to combine the Depository Trust Company (established in 1973 to hold security titles) and the National Securities Clearing Corporation (founded in 1976 to handle clearing and netted settlement), the DTCC is currently the largest securities processor in the world. It settles transactions of almost $1.7qn a year (that’s quadrillion, with 15 zeroes). There’s no point in trying to get your head around that large a number.

Since then it has acquired or created further subsidiaries to extend its services to include pan-European equities clearing, fixed income transaction processing, information management for trading institutions among other functions.

In 2006, the DTCC launched the Trade Information Warehouse (TIW) service, to centralize the storage of information regarding trades of over-the-counter (OTC) derivatives. One of its main functions is to maintain the “golden copy” − the unique, reliable and actionable record of transactions. It also manages post-trade processing such as payments and adjustments over the life of each contract (which, in the case of OTC derivatives, can be as long as 10 years). It currently handles the event processing services for 98% of the world’s outstanding CDSs.

Time for an upgrade

This is the platform that the DTCC wants to replace with blockchain technology. One of the main attractions is the possibility of making the “golden copy” accessible to all participants. Another is being able to automate the processing of lifecycle events via smart contracts (currently a largely manual process). Also, on the current infrastructure, settlement can take as long as a week to close, whereas on the new platform it could be almost instantaneous.

To this end, the DTCC started work on the redesign of TIW at the beginning of 2017, following a successful proof-of-concept executed in 2016. IBM is acting as project lead, blockchain startup Axoni will provide the technology, and R3 is acting as advisor. The platform is expected to go live in early 2018, at which time the underlying protocol will be submitted to opn-source blockchain consortium Hyperledger (of which the DTCC is a founding member) for others to also work on.

Given the systemic importance of efficient derivatives settlement, initially the new platform will launch in “shadow” mode and run alongside the current system. Participation will be optional, and participants will adapt their internal processes gradually, with large firms implementing their own nodes on the ledger while smaller ones hook in via the DTCC’s node.

To start with, the platform would only handle information and reconciliation. Payments would continue to move on traditional rails.

Thinking ahead

An interesting question is why the DTCC would do this. Are they not potentially writing themselves out of the picture?

What they are in effect doing is “disrupting” their own processes. As the largest CDS post-trade processor, they do have a choke-hold on the market. But the DTCC is a not-for-profit organization, owned by the industry. As such, its obligation is to the market participants, and includes future-proofing its service. What’s more, a reduction in reconciliation costs could boost transactions and liquidity, possibly helping to offset the post-crash decline in trading volumes.

Furthermore, its systemically important role gives it a clear view of how fast financial services can shift. By upgrading the principal post-trade platform and making it easier for derivates to be centrally cleared, the DTCC could be getting ahead of regulatory changes. With a node on the distributed ledger, regulators would have a complete and real-time view of the state of the market.

Big impact

When the platform goes live (expected to be early next year), it will be the largest project to date to enter production. Its effects will not be visible to the mainstream market, but the financial sector will be watching this closely, not only to see if the technology works, but also to gauge the impact of the cautious implementation strategy.

Blockchain technology is not the answer to all of the problems, structural and otherwise, that currently plague financial markets. But its potential is intriguing, especially the opportunity to affect how information is handled. That in itself could fundamentally change how the markets work.

With many more projects in the pipeline – from the DTCC as well as other significant players in the field – the launch of the CDS blockchain platform could well be the tipping point that triggers a host of implementations. With that, we will finally be able to say that the next era of financial infrastructure has begun.

Tiptoe through the tulips

by Aaron Burden via StockSnap
by Aaron Burden via StockSnap

While cringing through reviews of “Tulip Fever”, released in the US last weekend, I couldn’t help but wonder why the story still captures our imagination.

The bitcoin bubble has often been likened to the tulip bulb mania of the Amsterdam markets in the 17th century. That is wrong.

While the speculative fervour may have the same underlying root (cough, sorry), namely the me-too desire to get rich quick, there the similarities end. Historical analogies make for good copy, but unless you believe history does repeat itself (and I don’t), the usefulness ends there.

First of all, in the tulip frenzy, the actual price of tulips was volatile, but not nearly as much as the futures price. People were entering into contracts to buy tulip bulbs at a certain price a few months hence. In early 1637, the Dutch authorities decided that the futures contracts would become options – the purchase contract could be “cancelled” upon payment of a small percentage. In other words, speculators were buying the right (but not the obligation) to buy tulip bulbs at certain prices – these are the ones that went through the roof, not the actual prices at which the bulbs changed hands.

Second, bitcoins are obviously not tulips (not nearly as pretty), but even the nature of the investment (or speculation) is totally different. Bitcoin actually has a practical use. It can be used to transfer value without third party control or interference. It satisfies all of the characteristics of money, except for the (dubious) requirement that it be authorised by a central authority.

Tulip bulbs, on the other hand, have the potential to make people happy, but beyond that, they’re not useful, not even as a means of exchange. They are 1) not a store of value (they perish), 2) they are not fungible (no way is a lily-flowered the same as a viridiflora), and 3) they are not limited in supply. They shot up in price because people expected them to shoot up in price. The same can be said of bitcoin, true, but the difference is in the residual value – what is the underlying use worth? With bitcoin, it is potentially worth a lot. Tulip bulbs, not so much.

(Although as a caveat I would like to stress that I do believe that the enjoyment of beautiful things, even ephemeral ones, is worth paying for – although nature so often gives them to us for free. An interesting anecdote is that the most coveted bulb during the tulip mania was the Semper Augustus – it turns out that the flower’s beautiful striations were due to a virus that eventually killed the breed off. Even nature’s accidents can have aesthetic value.)

It remains to be seen what impact the growth of the bitcoin derivatives market will have. In the case of tulips, the emergence of a liquid options market lend the market some stability – the strike prices were volatile, the actual traded prices much less so. Could bitcoin’s development follow the same path? Or will derivatives undermine the financial incentives of proof-of-work consensus? As soon as I have the time, I’ll be looking into this some more.