Fixed Income ETFs Add Flexibility and Liquidity

September 18, 2018

By Steve Sachs, Managing Director, Head of ETF Capital Markets, Goldman Sachs Asset Management

Fixed income ETFs are straightforward and transparent, offering highly liquid access to a core asset class and its sectors.

Exchange traded funds (ETFs) offer investors a number of advantages. The most important is liquidity which, crucially, lies at the underlying level of the structure not at the wrapper level. A deep and extensive market ecosystem exists for this reason.

Fixed income ETFs are at nascent stage in development compared to the vast, global equities ETF market that has grown up in the last 25 years. However, their appeal to institutional investors has increased over the past five years, and accelerated as they are forced to respond to client requirements during the recent up-turn in the interest rate cycle.

There is now almost $700 billion in fixed income ETF assets globally. It’s a small proportion of the total $4 trillion market, but it is growing as investors have become more comfortable with different ETF vehicles. In the early days, they mainly comprised market-capitalisation weighted stocks linked to broad benchmark indices, then evolved to country-specific and niche vehicles, before maturing in sophistication to smart-beta strategies about five years ago.

Fixed income ETF development has lagged equity ETF maturation for several reasons. Primarily, the equity portion of a portfolio is typically inherently larger than the fixed income portion and thus the focus on ETF development has traditionally focused on this larger slice of the pie and greater need. While bonds traditionally play a risk mitigation and income generating role in portfolio construction, they have, in the post financial crisis landscape, also driven absolute return in portfolios. With the unwinding of accommodative monetary policies that is taking place globally now, investors are rethinking fixed income allocations and portfolio construction and demanding a more robust set of tools to build allocations. This is driving more demand for the next generation of fixed income ETF’s.

Fixed income ETFs driven by client demand

As the days of low interest rates and depressed bond yields are numbered, investors can no longer expect predictable returns in their fixed income portfolios. They are facing heightened duration risk, possible negative returns and exposure to credit events.

Moreover, the problems for US asset managers are exacerbated by the country’s demographics: an aging population requires wealth and income preservation, yet the conventional asset class for this phase in the investment cycle is fixed income.

Fortunately, bottom-up demand for a better way to gain exposure to fixed income markets has met top-down drivers of the ETF industry which promote mechanical efficiency, regulatory guidance and greater professional participation. Basel III, the Dodd-Frank Act and the Volker Rule restrict the warehousing of bonds by banks, leading to new intermediaries providing alternative sources of liquidity, facilitating the construction of fixed income ETF vehicles, which in turn have become a source of price discovery.

Trading in fixed income markets is robust, but it remains opaque and liquidity is fragmented and difficult to access at times. In practice, it is difficult to buy a large, broad portfolio of bonds, especially if the intention is to replicate a benchmark index or achieve diversification to mitigate risk. While this has improved over the past ten years as more technology has been applied to bond trading, challenges remain.

An ETF is an attractive alternative for many investors.

The ETF structure and its components are transparent and indices don’t need to be entirely replicated, as optimization lends itself well to fixed income portfolio construction. Other advantages of the ETF structure include lower dealing costs, greater tax-efficiency (in the US, because there is no capital gains tax payable), and, perhaps most importantly, they are tradeable as discrete units. Purchases and sales do not necessarily impact the value of the underlying fixed income securities.

Sale or redemption

Around $80 billion in ETF notional value are traded a day (including about $25 billion of fixed income ETFs), but only $10 billion is created or redeemed each day. For a sale, for instance, a dealer can hold the ETF in inventory, either taking an uncovered position or hedging their exposure. They can either find a buyer of the ETF, or might eventually redeem the ETF position with the issuer, who provides a basket of the underlying bonds in exchange.

Sometimes, the intermediary might use those bonds to construct a new ETF, warehouse the bonds as inventory, or simply sell the bonds. The idea is that the dealer has choices in which to mitigate risk and find liquidity.

Investors can take advantage of this flexibility. For example, an insurance company might hold a high yield ETF to gain immediate exposure and earn incremental yield, then redeem the fund through a dealer and take delivery of the underlying bonds.

Fixed income ETFs, like equity ETFs, benefit from the dual layers of liquidity that exists: secondary market and primary market liquidity. These two layers work together to not only provide liquidity, but also the ability to arbitrage any price difference between the two, thus keeping secondary market prices in a normal fair value range.

Of course, investors can always build a fixed income bond portfolio directly, with individual bond selections. However, there is a trade-off between costs and access, implementation and efficiency.

Future trends

At this stage, most fixed income ETFs are linked to broad benchmark indices, which is a constraint for investors that require more specific or thoughtful exposures. Niche fixed income ETFs are in the early stages, with issuers and providers examining the potential for a next generation of products, such as smart beta fixed income ETFs. In the future, we may see this continue to evolve as the market responds to the ever-changing needs of the investors in the asset class.

Fixed income ETFs are straightforward and transparent as well as offering highly liquid access to a core asset class and its sectors. The structures and investor choices will evolve as the client demand determines and as an ecosystem of participants, their skills and experience, grows.

Global Volumes Shifting Toward the Close

GlobalTrading

September 18, 2018

By Niamh Golden, Director, Analytics, ITG

The move of trading volume to the end of the day in Asia and globally highlights a large natural liquidity event even for those not benchmarked to the close directly.

The relative share of market-on-close (MOC) auctions around the globe has increased significantly over the past two years. In some of the larger developed markets, closing auctions can now account for as much as 20% of volume traded.¹

The trend toward using the closing facilities—spurred by the continued growth of passively invested funds and the relatively low cost for liquidity at the close—raises a number of important questions about how best to interact with this volume.

This article serves as a primer for comparing developed Asian markets with other global markets, focusing particularly on Hong Kong and how volume patterns have changed since the Closing Auction Session (CAS) began in 2016. We start with a quick survey of relative share of closing auctions, as well as the last 15 minutes of continuous trading, in various markets.²

NORTH AMERICA

Closing volumes in S&P/TSX Composite names have grown by greater than 1.5 percentage points to 4.6% over the past two years. Meanwhile, trading in the last 15 minutes in the same period, grew and then reverted to historic levels. The 5% of share level is the lowest, by far, of any of the markets we examine in this paper. This is likely best explained by both the relatively high exchange fees of the Canadian closing facility and the high level of intraday turnover in composite names, resulting from arbitrage between fungible listings of like names in the US market.

The Canadian close runs at 16:00 Toronto time, with orders entered during the regular trading session. The closing auction has a single imbalance publication at 15:40. Updated imbalances are given at 16:00 only if a given name is going to move more than 3% from the last tick.

The market in the US, like those in Japan and Canada, has higher levels of trading in the final 15 minutes than the auction itself. That said, auction share has grown dramatically over the past year. The US is unique in that it has two large exchanges running slightly different closing auctions (i.e. the NYSE and Nasdaq). Both facilities run alongside the continuous trading books, but the NYSE facility has designated market makers participating in the price-setting process, while Nasdaq does not. The two markets have surprisingly similar levels of MOC participation, suggesting that structural differences have little impact on investor use of an auction.

We also sliced the US via a few other attributes and observed a few interesting trends:

S&P large-cap names have grown considerably more in MOC share than small- or mid-cap stocks. Large, mid and small market caps saw a similar percentage traded in the close back in Q2 2016; large-caps ended over a full percentage point higher than mid- and small-caps in Q1 2018.
ETF volumes traded a modest 3.4% in Q2 2016 and increased to a high of 4% in Q3 2017. While still much lower than what we see in the equity space, this is nearly an 18% increase.
Perhaps our most interesting finding in the US market was the uptick in trading during the final minutes of the continuous session.³

EUROPE

MOC relative share has grown by more than 40% in CAC 40 names over the past two years, but trading share in the final 15 minutes has remained virtually flat. Like most European exchanges, the MOC facility at Euronext Paris is separate from continuous trading. Continuous trading ends at 17:30 Paris time, when the closing auction opens. The closing cross occurs at a randomized time between 17:35 and 17:35:30.

MOC share grew roughly 20% over the two years in DAX names. This is more modest than most of the markets we studied, although still very meaningful. During the period, we saw a very minor decrease in relative share traded during the final 15 minutes of continuous trading.

The Deutsche Börse auction, like Euronext Paris, runs post the continuous trading session and has a randomized end time. The auction begins right at 17:30 and runs through 17:35 Frankfurt time. As is the case with most European closing auctions, the Deutsche Börse publishes updated indicative size and price data during the auction.

MOC share of FTSE 100 stock trading is up more than 25% in our study period, while trading in the final 15 minutes is down very modestly. The LSE closing facility is similar to many in the European markets as explained above. The auction is a separate book that runs from 16:30-16:35, after the close of continuous trading.

APAC

Stocks in the ASX 200 have seen a nearly 30% rise in closing auction market share over the two years. Meanwhile, relative share during the last 15 minutes of trading has been virtually flat. The Australian Stock Exchange closing mechanism runs separately from the continuous trading facility. Orders are entered during the Pre-Closing Single Price Auction from 16:00-16:10 Sydney time, with the actual match occurring between 16:10 and 16:12.

The MOC facility has lower relative share than most developed markets and has witnessed far more modest growth than many peers. This may be explained by the continued high levels of volume weighted average price (VWAP)-style trading in the Japanese markets.

The closing auction runs at 15:00 Tokyo time and is a separate book from that of continuous trading, although continuous trading orders will roll into the auction. Market orders placed in the auction are assigned a limit price equal to the calculated daily limit price for a given stock.

The Hong Kong Stock Exchange introduced a closing auction in two tranches, starting with the most liquid names on 25 July 2016. Previously, the closing price was calculated as the median of five prints, taken at 15-second intervals during the last minute of trading and a large percentage of volume completed in the last few minutes of the day. Since the introduction of the CAS, there is a clear shift in volumes toward the auction and, by Q2 2017, the auction accounted for as much volume as the last 15 minutes of continuous trading. The auction now accounts for just over 10% of volume in the Hang Seng Index.

Hong Kong has taken a European approach to the close, with the auction being run separately from the continuous trading session. The auction starts at 16:00 Hong Kong time, with a randomized close between 16:08 and 16:10.

The last thing we examined as part of the comparison was price dislocation. By comparing the last 30-minute VWAP to the closing price, we can look at relative performance across markets.

This chart suggests that Hong Kong has the most price elasticity of the markets we looked at, while the North American closing auctions are more efficient.

While dislocation is one important factor in market efficiency, there are others, such as price drift and ability to easily replicate a price. To this point, markets including Shanghai and Mexico use a VWAP price to determine the official close.

It’s clear that volume is shifting to the end of the day in Asia and globally, which highlights a large natural liquidity event even for those not benchmarked to the close directly. At ITG, we are aiming to create a better framework for thinking about both close liquidity and market impact in a dynamic fashion. We welcome client feedback and questions in order to better inform our own research.

¹ITG’s analysis focused on stocks included in market-level indices.

²A median of the individual stock’s bin trading as a percentage of daily stock’s trading, weighted by the VWAP price.

³Aggregate share volume in minute bins across stocks during the quarter as a percentage of aggregate daily share volume. Universe includes stocks traded on average above $1 and $50k daily volume during Q1 2018 in the U.S. (universe includes 3,400 stocks)

Castles In The Sand

GlobalTrading

September 13, 2018

By Huw Gronow, Head of Dealing, Newton Investment Management

A standardised, consolidated transaction tape utility would provide consistent and complete liquidity data which would improve forecast future outcomes.

Nearly 200 years ago, the scientist John Brown made the startling discovery that particles inside pollen grains in a drop of water moved around due to their collisions with the water molecules, themselves fast-moving, in a process later proved by Albert Einstein.

Brown’s discovery gave rise to the eponymous description of random motion that we know today. To the casual observer, a beaker of water remains a still object in equilibrium; without closer inspection, this assumption does not see that at the molecular (and even quantum) level there is, to put it in layman’s terms, a lot going on.

The extension of the discovery of this phenomenon gave rise to what we now describe as stochastic processes, with several extensions into the world of finance, and as a way of applying these to and thinking about the way securities markets work.

As it is, our current equity market structure in Europe, and to a greater extent in the US, reflects this. The advent of ultra-low-latency access to markets, with trades now measured in nanoseconds from order to entry to arrival at the exchange matching engines, has developed along with the new fragmented market structure. It is now necessary to atomise the intended trade into much smaller particles and direct to many different venues simultaneously to retain a high degree of probability of success, designed so that a significant number of “messages” are sent to the market in excess of the desired intent to trade.

This inevitably means that the data seen at one level on the computer screen accounts for just a fraction of the activity that comprises the ecosystem of fast, interconnected exchanges, multilateral trading facilities (MTFs), systematic internalisers (SIs) and the rest.

Contributions and benefits

In Europe, it is a matter of sometimes-heated debate about whom the development of market structure benefits, in terms of the advances in technology and regulation. The issue most discussed concerns the short-term, ultra-low-latency proprietary trading participants or high-frequency trading (HFT) firms, as they are broadly labelled. These HFT firms are seen as generally either good for the market, claiming to supply liquidity, or predatory, seeking to detect the signals of predictable trading strategies.

The reality is that all participants in the ecosystem make a contribution, of whatever value, and that it is the job of regulators, as well as the responsibility of market participants themselves, to police what is deemed illegal, or potentially so in the environment.

The aim should be to eliminate any informational advantage given by the composition of the ecosystem itself, where those advantages are identified and considered detrimental to the role of capital markets – which is to allocate capital efficiently and fairly.

The exponential increase in data that the market has produced by ever diminishing trade sizes and frequency of trading means that the equity market overall can arguably now be viewed in better definition than 20 or even 10 years ago. It is unsurprising that the interest from the academic world, of mathematicians, engineers and applied statisticians, schooled in Markov, Monte Carlo, Wiener and so on, has grown substantially.

Balancing alternative trading approaches

The task of the institutional trader is to navigate two difficult paths: one is between seeking significant size inventory, and therefore (at least for the early part of the risk transfer process of liquidity consumption) eschewing the exposure to pre-trade transparent, displayed, “lit” environments, and thus incurring possibly punitive transaction costs. The other path is followed by participating actively in all parts of the ecosystem and therefore subjecting the portfolio to potentially deleterious levels of market impact costs.

At all times, the balance of these two approaches is determined by the urgency and progress of the trade.

The undertaking for large inventory transfer is therefore, a complex one to pre-programme and is better managed in vivo by skilled and experienced human traders rather than predictably and deterministically. It follows that the higher aim, and the opportunity provided by the recent changes in regulation in Europe, is to apply this learning to the contribution of transaction costs to the investment process and, in particular, to the efficiency of portfolio construction.

What tools are necessary to accomplish the task effectively? Aside from the required efficacy and efficiency of routing capabilities, and exposure and access to all desired execution venues, the ability to analyse, forecast and implement overall trading strategy is vital to delivering the best possible result.

The latest revisions to Europe’s Markets in Financial Instruments Directive (MiFID) and Regulation (MiFIR) have the central tenet of transparency coursing through the text, whether it is via the restriction on non-displayed trade activity under large-in-scale size, or the share trading obligation, as well as the requirement to publish most trades as close to real-time after the event as possible, among other stipulations. Whatever the debate about transparency and the impact of real-time disclosure to large trades and the cost of capital incurred by broker-dealers who may provide liquidity on this scale, the central desire must be consistency in data labelling and disclosure.

Accessible liquidity

At present, much debate centres on what is “accessible liquidity”; what one may define as the ability for any market participant to take part in a trade at any given moment. This is easy to delineate for a public exchange’s order books or those of an MTF; arguably less so for a periodic auction where “broker preferencing” may be a feature, or an SI source of liquidity, just to take two of a myriad examples.

And therein lies the issue. Most if not all measures of the ease of investment in a security have liquidity as a major factor. While it may be marginal if that liquidity is not clearly and consistently defined, even if after the event, the inputs into the liquidity forecast become unsound to a greater or lesser extent.

When one then attempts to forecast the market impact associated with the decision, the portfolio construction assumptions may suffer. The highest aim for the next revision of the legislation, unless an enterprising entity takes the opportunity beforehand, is to grasp the opportunity to fill this gap. This is by way of a mandate of a standardised, consolidated transaction “tape” utility.

This is clearly within the bounds of possibility for equities, but may be some years away for other asset classes. It is a truism that any attempt to try to model or forecast future outcomes based on inconsistent data and classification amounts to an approximation, however narrow the distribution of these outcomes is.

Incorporating this inconsistency into the important task of integrated efficient portfolio construction for the benefit of end-investors and their returns runs the risk of building castles in the sand.

We’d love to hear your feedback on this article. Please click here

Short-Term Alpha Signals

GlobalTrading

September 12, 2018

By Andrew Royal, Head of APAC Autobahn Analytics and Algorithms, Deutsche Bank

Identifying short-term alpha signals in the market can improve trading strategy logic.

The application of agency algorithms can help buy-side firms improve trade execution performance by detecting trends at the micro-scale. Order book and trade flow are two key indicators that can be used to improve algorithm child order placement. In this paper, we examine the neural network model of these short-term alpha signals deployed to enhance the strategy logic on the Deutsche Bank Autobahn Equities platform.

Order book imbalance

The order book signal at time t(Q_t) can be defined by the difference in bid size (q^B_t) to ask size(q^A_t):

This outcome is on an interval (-1,1), where numbers close to +1 indicate that we are bid side heavy relative to the ask side. Order book imbalance theory (see Cont, Stoikov and Talreja, 2011) suggests that low ask sizes indicate high probabilities of the ask queue becoming zero before the bid queue, and thus the next price we see should be towards the ask side. The reverse holds if the imbalance is negative.

Trade flow imbalance

A second signal we can look at is where the trades are occurring. Trade flow can give us good information as to where hidden orders and large meta orders are priced in the market. For example, if there are multiple trades on the ask side, we might conclude a hidden order to be responsible. Hidden orders imply autocorrelation in the trades, and hence multiple trades in a row give us some expectation that the next trade will also be on the same side.

We define the trade imbalance at time t(T_t) by using the volume of trades at either the bid (v^B_t) or the ask(v^A_t):

This outcome is in the interval (-1,1), with a “+1” value indicating all trades are occurring on the ask side. When T_t is nearing “+1”, all else being equal, we might expect volume to continue at this price level – there is clearly demand at this price, but there might not be enough supply to satisfy it – that is, the mid-price should tick upwards.

The model

Our intention is to determine where the next mid-price at time t+1(P_t+1) is going to move. We let:

The task is to find a function f so that we can fit the model:

We find f using a feed-forward one-layer neural network. This is quite simple to fit and doesn’t assume any linearity or functional form in the solution.

A fitted solution

This model was fitted to stock of HSBC for the six months from 1 January 2018 to 30 June 2018, and data was split into training, testing and validation data. We found the following predictions, which are defined as the function g(x,y) = f(x,y,…) and plotted below:

What can we conclude about this surface? Firstly, the signals can be conflicting in nature – a positive trade imbalance (momentum) and a negative orderbook imbalance can reduce the expected price increase.

Secondly, the five-second signal gives the strongest prediction. The strength of the signal declines as the time interval increases. This suggests that some reversion is going on.

Thirdly, the fitted function is non-symmetrical for the five-second prediction – we expect that short selling rules here produce this, making it more difficult to observe an imbalanced book. More research is required to prove this.

Backtesting

Here we analyse the predictability of short-term alpha in the stock price for different time intervals, namely: 5, 10, 15, 30, 60, 120 seconds using out-of-sample data for HSBC for the six-month period.

Broadly speaking, the out-of-sample data showed a small decline in the predictability of the stock compared with in-sample data, but it was pleasing to see we were not overfitting the data.

Turning the prediction surface into actions necessarily means a tradeoff between false positives (we let in too many signals) and false negatives (we don’t let in enough). The following ROC graph demonstrates the tradeoff and allows us the appropriate way to quantize our prediction surface into actions. We tend to go for a small type 1 error (small number of false positives). The interesting feature is how quickly the prediction declines with time – five seconds can produce an informative prediction that will help with the strategy as the ROC curve is above the 45-degree line. However, 120 seconds does not create a useful action so we are better off just guessing.

Usage in the algo suite

We can use these results in our algo strategy by defining a set of simple rules on top of the best quantization of the prediction surface. Graphically, for a buy order we have the following three possible actions: amend, stay, cross.

For a buy order, the cross action means we should cross the spread immediately if we can, as the price is almost certainly moving away. The stay action means we do nothing as the signals are not strong enough. The amend action means that the price is coming towards us so we should either amend the price down if we are ahead of the volume, or amend the quantity up if we are behind the volume.

Going forward

Deutsche Bank Autobahn Equities continues to introduce more alpha signals into our algo strategy logic. Identifying order book and trade flow imbalance is proving to capture short-term alpha and therefore optimize trading performance. We are testing several statistical models that have important applications and expected performance improvements for global markets, leading in APAC.

Applying Technology To High-Touch Trading

GlobalTrading

September 12, 2018

By Lee Bray, Head of Asia Pacific Equities Trading, J.P. Morgan Asset Management

Traditional single stock trading processes have suffered from underinvestment, but new technologies are now being adopted that will create a more effective and structured environment.

Over the last few years, there has been a significant push from the financial industry to create a trading landscape around the implementation process of buy-side desks. There are a few distinct driving forces behind these changes. One is regulation, such as the Markets in Financial Instruments Directive (MiFID) II. Another is modern technology – transaction cost analysis (TCA) products are constantly changing, and algorithmic usage is now a significant portion of most trading desks’ flow.

As we have seen these changes take effect, certain areas of buy-side trading desks have been impacted more than others. Traders who can now be described as “low-touch” have benefited from significant investments in tools to help enhance their decision making, whether this takes the form of in-house models or “wheels” or accurate feedback from TCA products that can store down every parameter that the trader has applied to the stock when implementing their strategy. This has created an invaluable feedback loop in the move towards a more quantifiable implementation in their world.

Investment in single stock trading

Unfortunately, more traditional single stock trading processes have suffered from underinvestment. This critical trading role has historically been more manual in nature. As a consequence, it has been difficult to create a structured trading workflow able to provide meaningful feedback to the traders. Given that this area by definition involves trading in more costly stocks, building out this feedback loop will have positive benefits. Fortunately, we are beginning to see these changes.

The market is beginning to apply some of the techniques developed in the low touch world to the “high-touch” world. Also, high touch traders are becoming the recipients of more corporate technology spends and putting it to good work by creating innovative solutions to supplement their daily workflow.

We have seen this notably in the indications of interest (IOI) space. IOIs have been a big focus for the industry over recent years, enabling buy-side firms to assess their content much more accurately. This has been overlaid with technology and smarts from the quant world to increase the information that can be derived from what was historically a single-stock standard tool.

All IOI data is now stored down at JP Morgan, constituting millions of data points a year. Quantitative models are now being developed to highlight recommendations to our single stock traders, facilitating more efficient counterparty selection and providing more transparency to our traders around the trading landscape.

It doesn’t simply stop at IOIs. The empowerment of the single stock trader armed with information to aid decision making will continue. We are starting to see new products coming to the desk that incorporate aspects of natural language processing. Although it remains at an early stage, in time this may be significant in supplementing our traders when making key strategic decisions.

We believe the workflow of the future for the single stock trader will link changes in their strategy back to key events, while simultaneously providing information to enrich the traders’ knowledge on trade outcomes and TCA numbers achieved. In other words, the process will ultimately enable the trader to answer much more detailed questions on how their actions in trading a stock contributed to the quality of the executions.

Adding value

It is not just the internal framework for the single stock trader that is evolving; now more than ever, investment in this space is impacting the assessment of the brokerage community. Crossing rates – a proxy for measuring the value added by many single stock traders – can now be monitored and assessed much more systematically with the help of technology.

It is rarely the case in my experience that a single stock trader doesn’t want to cross stock, but the fear is that by talking to the sell-side there can be negative effects on the stock price which often proves an adverse influence. By providing up-to-date trading metrics using smarter technology, these fears can be confirmed or dismissed by actual data assessed over time. As a result, traders may feel more confident when taking the important decisions they are called upon to make daily.

The single stock trader is an important component of the modern, forward-looking trading desk. After a period of underinvestment by the industry as a whole, we are embarking on a period of change which will see significantly more empowerment of the community.

New technologies being adopted will create the kind of structured trading environment that has been used for some time in the low touch space. Far from disenfranchising the trader, these new technologies will optimise their existing workflows, helping to aid quantitatively based decisions which will ultimately lead to lower costs, in what is a notoriously tricky segment of the market to trade.

Limitations On The Use Of Artificial Intelligence

GlobalTrading

August 28, 2018

By Ambrose Tan, Head of Dealing – Asia Pacific, Aberdeen Standard Investments

The successful application of the latest technologies to the trade execution process requires clear and consistent rules and standards.

Diverse global regulatory regimes and contrasting trading procedures in different asset classes restrict the effectiveness of new technologies in the trading process. Custom, inertia and the difficulties in enforcing change in dynamic markets mean differences in asset class trading practices are likely to remain entrenched for some time. On the other hand, a common policy shared among regulators might facilitate a more confident deployment of the latest technologies, notably artificial intelligence (AI).

The application of AI can offer buy-side dealing desks undoubted benefits, notably by reducing the time it typically takes for a human trader to gather and analyse information required to facilitate best execution. AI can access data faster, and then recognise trading patterns in order to provide the basis for a better informed trade decision. In future, it is feasible that dealing desks may have AI specialists working alongside human traders.

However, currently, it has fundamental limitations: AI is only able to access and evaluate the data that is available, recognise patterns and factors within pre-established parameters, and therefore cannot make recommendations based on inherently incomplete material, subjective instructions and constraints. AI can suggest a “most likely” outcome, not predict a definitive result.

On the face of it, a statistically accurate assessment and decision forecast might seem attractive. But, both regulators and clients could feel justifiably uncomfortable. Regulators, especially in Europe following the introduction of Markets in Financial Instruments Directive (MiFID) II this year, might conclude that if a platform’s software is provided by a sell-side firm, and it promotes a decision-making process based on AI, then it could be an inducement to trade, which would potentially hinder best execution and not work in the interest of clients’. The MiFID II requirement for buy-side firms to provide LEI numbers designating discrete, individual accounts tightens the scrutiny even further, but promotes greater trade transparency and reporting. It is a fine line between enticement and inducement, and the advertisement of a buy or sell signal might fall on the side of the latter.

Yet, the unbundling of research provision and trade execution in itself is unambiguous, so it is important that a fund manager avoids stepping over that line, or is even suspected of doing so. The onus is on the fund management firm and its traders to be fully cognizant of brokerages that have been selected as authorised providers of (qualitative and quantitative) research material and ensure that there is no suspicion that it is being used as an inducement to trade.

Moreover, the regulator is also wary of a buy-side firm that relies on an in-house developed AI system for its trading and investment decisions, rather than a process featuring directly accountable individuals. Ultimately, humans, not machines, must be held accountable.

Meanwhile, clients too might legitimately complain that a trading decision, which is integral to a buy-side firm’s investment process, should not be the product of a machine-generated computation. After all, they are paying the firm for the skills and experience of its fund managers and traders.

Nevertheless, AI is a useful tool for helping trade some, but not all, asset classes. The main determinant is the comprehensiveness of accessible data. Information about exchange-traded equities is usually sufficiently extensive and reliable for AI systems to operate, but far less so for fixed income and foreign exchange markets where most trading occurs over-the-counter (OTC), directly with sell-side counterparties or anonymously via broker intermediaries. Executed transactions and post-trade information can be opaque. Price makers are conscious of market impact, especially during bouts of illiquidity, hence there is an emphasis on protecting market flows.

Fixed income markets are characterised by brokerages’ inventory supplies, risk capacity and niche expertise; although AI might identify trading patterns, ultimately a human trader’s networks and experience are a more reliable way of achieving best execution.

Foreign exchange markets are generally opaque, especially spot trades where price activity is determined by myriad influences. Speculative behaviour remains a powerful force, but less so now that hedge funds are under tighter regulatory scrutiny. Instead, “genuine” transactions that support or hedge investment and corporate treasury decisions dominate the spot market and AI will struggle to determine useful patterns among those trades.

MiFID II seeks to address some if not all the above issues as it aims to create a level playing field for all market participants and all markets with a push for greater disclosure and transparency. Now, transactions that involve foreign exchange products – options, non-deliverable forwards and currency and interest rate swaps – are reportable to the market within 15 minutes of trading. Trading data on various venues including Multi Trading Facilities (MTF), aka multi-dealer platforms, will also be captured. New legislation seeks to monitor not just post-trade but pre-trade analysis as well.

While these developments may spell a positive for AI trading, we go back to the fundamental issue of human accountability. Will AI create an environment in which volatility spikes up in the face of speedier data crunching?

Disparate jurisdictions and rules
A more fundamental problem is the confusingly disparate regulatory regimes throughout the world that force buy-side firms, in particular, to be circumspect about the adoption of some technologies – such as AI – that should in theory benefit the investment and trading processes.

Some financial regulators in Asia do manage to combine a firm oversight (across all asset classes), while also being receptive to global regulatory trends and open to the development of new trading technologies. Elsewhere in Asia, there are several emerging markets with diverse regulatory regimes which are a challenge to institutional investors with a global or regional mandate.

Yet, perhaps a standout example of regulatory anomaly is actually between the rules in two developed jurisdictions. In a direct contradiction to the unbundling requirements of MiFID II throughout Europe, in Japan, brokers are not allowed to charge fees directly for research, but instead bundled in transaction charges.

If the various worldwide regulatory bodies could agree on common policies and standardised rules, then they would help money managers and their clients achieve their objectives more easily. More importantly, consistency would allow a successful adoption and implementation of the new technologies that are increasingly available.

Aberdeen Standard Investments is a brand of the investment businesses of Aberdeen Asset Management and Standard Life Investments.

The views expressed in this article are the author’s and not the company’s.

We’d love to hear your feedback on this article. Please click here

Electronic Trading In The Brazilian Markets

GlobalTrading

August 28, 2018

By Dr. Christian J. Zimmer, Quantitative Research, Itaú Asset Management and Chair LatAm, FIX Community, and Leandro Pereira, IT Manager, Itaú Asset Management

Fixed income trading in Brazil is now much more efficient, and attention is moving to equity order matching and stock-lending.

Brazil witnessed significant advances in electronic fixed income trading in 2017. The country’s notoriously high inflation rates means fixed income is an especially important asset class in Brazil, so improvements in trading infrastructure and systems have a substantial and supportive impact on the country’s savings and investment industry.

However, electronification in the fixed income market historically always lagged a considerable way behind equities. There are several reasons for this, including fragmented liquidity, an entrenched human trading culture, buy-and-hold investment preferences and a requirement for pre-trade submission of cash and securities to meet T+0 settlement.

Progress was made following a FIX Community event in 2016, where representatives from the buy-side and sell-side in Brazil recognised that innovation was needed in the local fixed income markets. In particular, sell-side firms agreed to move towards electronic trading as long as the buy-side was supportive. A local FIX Protocol (FPL) committee decided in 2017 that it should focus on three major issues for enhancement. These were: fixed income electronic trading, intraday trade matching for equities, and stock lending.

At this time, fixed income securities were tradeable on the BM&FBovespa and Cetip (later merged into B3) and on a Bloomberg platform, fragmenting an already illiquid market in private and public debt. As Cetip was the most liquid market at this time, a pilot project between Itaú Asset Management (IAM) and Cetip was initiated to route fixed income orders.

Pilot project
So, an arrangement was made with Cetip, where IAM can route the orders to the market, such as send orders to the brokers in the market. Basically, there are three types of fixed income order:

• Market Order
The market order is the most traditional kind of trade, used by the entire market, in this kind of order the buy side sends the order to a broker indicating the limited price or limited yield (usually limited yield). This kind of order is defined by the value 0 (zero) at the tag #828.

• On behalf of
Orders on “behalf of” are similar to the above market orders, the only difference is the tag #448, where the buy side identifies the broker that will appear at the Cetip platform. This kind of order is used only when the buy side decided to create an order in a order book.

• Cross opportunity
This type of order is chosen when the seller and buyer are management desks of the same institution, trading the same security, but on opposite sides of the transaction.

In the Brazilian market the central bank doesn’t allow trades between two accounts without a broker intermediating the trade even for an internal crossing opportunity the buy-and the sell-order must be sent via a broker.

From the sell-side perspective, an order originated from an internal crossing opportunity is very different to a market order. The FIX message for this kind of order has special tags filled by the buy-side, in order to indicate to the sell-side the details of the trade as follow: #828: Trade Type: 3 (Cross opportunity). #548: Cross ID: Identifier for the cross order. (The ID is used to link the buy and sell orders, and only two orders can have the same ID per session).

One of the problematic workflows for the fixed income industry is in the post-trade process, because all trades must be registered with the central bank for T+1, and the cut off time is 7:00pm.

1. Allocation
Every trade scenario has a specific FIX message, but all orders are pre-allocated within FIX 4.4 using repeating groups with the tags #79 and #80. That was the practice for the market at the beginning of the project, but was abandoned without any major issues for the participants.

2. Settlement command
Due to a convention of the Brazilian market, the seller is responsible for the origination of a number (unique in the day) to be used in the settlement process at the central bank. This number is called “command”. In this project we used the custom tag #8000 in the repeating group, to accommodate the settlement command.

3. Settlement account
The settlement account is a number defined by the broker, and represents the agency or principal account to be used in the settlement process at the central bank. During this project we included this number in the custom tag #8001 and the broker provides that information for each order fill sent.

4.Validation
Since the post-trade validation is included during the trading process, the broker has to implement the following validation before acknowledging the message:

Verify if all accounts indicated at the repeating group of the tag #79 are allowed by the brokerage.
Verify if the security identifier of the tag #48 match with the other securities information in the message, that is maturity date and ticker.
Verify if the yield of the order is fair, according to the spreads currently executed in the market.

Once the order is executed, the broker has to send the execution price for the order, even when a target yield was provided as a reference for the execution. This allows the buy-side to verify if the price is correct, corresponding to the yield provided.

The participating firms may then send the order to the central bank for registration, without any order post-trade process.

And what are the plans for the next few years? Although there is always place for improvement, we understand that the process for fixed income trading is now relatively well designed and effective. Now, it is time for the rest of the market to follow the trend and board the electronic trading train.

We believe that there is also a pressing need to improve the real-time matching of Brazilian equity trades, which is still a remarkably manual process.

Matching of equity trades in Brazil
The process for matching equity trades in Brazil is the following:

When trades are completed – that is, after the market closes – the buy-side sends a flat file (a “tordist”) to the sell side. This file specifies how all trades made on the omnibus accounts should be distributed to the final accounts that trade under this umbrella.
The Brazilian exchange, as the central counterparty clearing (CCP) for all equity trades, demands an allocation of all trades that are executed on a trade-by-trade level. This means that if an omnibus account sends one trade to the exchange and receives 10 fills, then 10 distributions must be indicated for each fund participating in the trade. So, if five funds participate in the trade, the buy-side sends 5*10=50 designations. It is not permissible to work on an average price basis.
The sell-side receives the distribution and inserts it into a system of the exchange, called Sinacor. After processing all trades and distributions, the exchange “clears” all trades and thus allows the sell-side to close the day.
When all trades and allocations are approved, the sell-side sends a confirmation file to the buy-side (“pesq”).
If nothing is wrong, the buy-side matches the tordist-file with the pesq-file and then inserts the accepted allocation into the middle- and back-office systems.

This process is only for physical allocation: a separated file is sent from the sell-side to the buy-side with all the transaction data, and depending on the buy-side systems, these values are confirmed with the internally calculated result based on the physical allocation file.

Sometimes trade allocations do not match – perhaps because of an execution error – and maybe the error is only identified after the market has closed. This causes an end-of-day panic, putting pressure on trading desks and middle offices. But, creating a point-to-point FIX-allocation process is cumbersome and some players do not have the systems to post their information into a FIX engine.

Stock lending and borrowing
The liquidation part of the stock lending and borrowing process is easily automated, with APIs available from the CLBC (the exchange’s entity responsible for equity custody). Simple commands can be used to inform new borrowing and lending operations, renewals or cancelations.

For the buy-side, the problem is less in the operations aspect of executing a deal, but more on the deal-arrangement aspect and finding liquidity. A first step in the direction of creating a deal-arrangement platform is with the intraday price (rate) broadcasting of deals sent to the CBLC.

We hope that within the next year and a half, we will have devised a more efficient solution for equity trade matching, and also have found ways to improve and automate the stock lending process.

We’d love to hear your feedback on this article. Please click here

The Future Of Social Investing – And It Is Not About Blockchain

GlobalTrading

August 21, 2018

By Tan T-Kiang, Chief Technology Officer, Grasshopper

Artificial intelligence and machine learning should be commoditised for the financial industry, but implementation suffers from technology weaknesses, data inadequacies and human inertia.

Big data and machine learning have already changed the way we receive information. Advertisers no longer rely on billboards or television to reach us. The recent events surrounding Facebook, Twitter and Google have shed a very public light on the way those mammoths have been using the data of their users to serve the commercial and political purposes of paying third parties. By simply reading the user profile and applying machine learning tools, they can, more easily than ever, identify the user’s needs and wants, with an almost scary accuracy.

The pervasiveness with which artificial intelligence (AI) has come into our daily lives is indeed impressive. What if we could also apply this to trading and investing? Actually, this is already possible. Some hedge funds and quantitative funds have already developed AI driven trading bots. But how do we make it easily available for everyone?

In general, we would like to automate the following investing process:

Idea generation
Idea validation
Trade execution
Evaluate steps 1 – 3 to iteratively improve them. This is a very logical place for AI to help us improve our own trading biases.

Most of these processes could be automated with AI bots. We could implement bots that are able to identify changes in prices, volumes or even related news from reputable media outlets. Automation with bots will allow investors to be able to expand their trading universe without spending more time in front of a computer to find and analyse the information that an AI bot would be able to process more accurately and tirelessly. Sometimes we also need automation to ensure that we are not driven by our emotions especially when we have been monitoring a position closely for long periods of time.

I believe that AI bots would be able to play a large role here to help our investors to be more reliant on automation and scalability of these bots to become better at their trade. It is almost like creating a trading team without hiring a team.

Next Steps

For it to happen, there are still many challenges that AI will have to face before becoming commoditised and easily utilised by everyone.

The first challenge lies in ensuring the perfect understanding by the AI of the user’s request. To be used for trading, the AI bots would need to understand with 100% certainty the demands of the trader. This 100% capable AI bot is still far from being achieved. Even Siri is still struggling when you request the weather or the nearest gas station. Imagine letting Siri give trading orders to your broker…. I would still be afraid today.

The second main impediment for a good investing bot lies in the quality of the data and the data collection method. On the former, the current data available to the general public is limited and with minimal checks to ensure validity. If the source of data is not good or the data sets are corrupted, there is no chance for the AI bot to operate a successful investment strategy. In short: “garbage in, garbage out”.

On the latter, we replicate the data collection processes over and over again between the different data vendors, hedge funds, individual traders and big institutions. Wouldn’t it be to the benefit of all to have a single data collection and validation facility to be shared open source with all the market participants? I am appealing to all market participants to be “green” about how we are collect and store our financial information. Currently, we are all doing it in a highly non-sustainable, non-environmentally friendly manner.

One of the main reasons invoked by market participants to run the data collection individually is the capability to differentiate themselves from the competition, by obtaining a “better” set of data. Well understood, but I think focusing on the raw information is the wrong battle. The focal point should be growing the capability to harvest the knowledge from the data. This would be akin to the commoditisation of computer power with cloud computing, so why not the data too? After all, data is merely oil within the engine of race car. A good driver is still necessary to ensure that you can reach the finish line.

New interfaces

The third obstacle resides in the user interfaces that currently exist. We see chatbots being used successfully in customer relationship management or as virtual assistants, but not yet in the area of finance. Why? First and foremost, we need a way to transform our financial industry’s reliance on outdated user interfaces such as spreadsheets, watch-lists and chart reading skills to identify signals.

The next generation of traders grew up as smart phone users with apps and games that leverage touch screen interfaces and notifications. A simpler user interface that encourages idea generation, validation and sharing would be essential in the social media age. In general, most online brokers have perfected the user interfaces for execution but have neglected the investors’ discovery journey.

Once those challenges are solved, the next step would be to increase the complexity of the bots that one creates, by using machine learning tools to develop predictive capabilities and enhance simple regression analysis with neural networks

For now, what would make most sense is for everyone to get involved collectively and build community-based investing platforms so that all can benefit. Automation and machine learning is within the reach of many investors today because of the birth of cloud computing, open source code and pre-trained AI models. Let’s make it happen for all.

The Future Of Bitcoin

GlobalTrading

August 21, 2018

An interview with Elizabeth Stark, Cofounder and CEO, Lightning Labs

The keynote speaker at the 16th FIX Asia Pacific Trading Summit discusses bitcoin, blockchain and the tech industry.

What is the greatest transformation in technology you’ve seen in your career?

While I was fairly young at the time, watching the internet and World Wide Web spread was fascinating. I then went on to teach about the intersection of law and technology at Yale and Stanford Universities, looking at how the internet is changing society, culture, and the economy. It’s so interesting to see that there’s an entire generation that has never known a world without widespread internet. I believe we’re seeing the same for the ability to transact on the internet with bitcoin and cryptocurrency, and in many ways it has felt like the early days of the internet.

Sounds like you are a firm believer in blockchain and cryptocurrencies?

When I first heard about bitcoin, I thought it was fascinating, though I was admittedly sceptical if it would ever take off. A few months later, that scepticism faded and I was hooked!

However, I still consider myself a blockchain sceptic. That’s not to say that I don’t think this technology is important, but for so long there was a movement to minimize the importance of bitcoin and cryptocurrency and instead call the true innovation: “the blockchain.” For me what is so special about this technology is we have the formation of new public, decentralized networks that are similar to how the internet developed. In the early days of the internet, many companies also created corporate intranets, but they were not interoperable or connected to the broader network. I view much of the blockchain discussion as the equivalent of the corporate intranets, or just general hype around use cases that aren’t actually feasible.

The beauty of bitcoin is that it has a built-in incentive structure through proof-of-work to keep the network decentralized and secure, along with making public/private key cryptography more widespread, so it’s really not just about a “blockchain” alone.

The true potential of this technology is to create a new financial infrastructure, an internet of value that is open and decentralized, and that anyone can build on top of. This is what I believe bitcoin can achieve.

Why did you create Lightning Labs?

At Lightning Labs, we’re building a software layer on top of bitcoin that can bring it to the next wave of people. We’re making bitcoin faster and more scalable, by enabling instant, high volume transactions. We’re making it easy for developers to build on bitcoin, with easy-to-use interfaces for developers. And we’re enabling interoperability between blockchains, as Lightning can enable cross-blockchain swaps between currencies like bitcoin and litecoin where you don’t have to trust an exchange to trade. In short, we’re creating the building blocks for this new financial infrastructure.

We are building out the Layer 2 of blockchains, but it is also possible to build a third layer, be it a smart contracting protocol or algo trading protocol on top of Lightning itself. I can’t wait to see what other people build on top of our technology, as the potential for new use cases is massive. Check out this talk I gave at the Blockstack Summit for more on the importance of Layer 2 technologies.

The good, bad and ugly of blockchain?

Everything in our industry moves so quickly that it can be extremely hard to keep up, even when you work in the space full time. I see that as a positive sign though, as it means more and more people are getting interested in building this future. Right now there’s a massive opportunity for people to get up to speed and learn about this new technology and where it may be headed. As I like to say, there’s no one in the world with a decade of “blockchain experience” (except for Satoshi Nakamoto, of course). What that means is if someone is motivated to learn about the technology they can start reading any number of resources online and get up to speed fairly quickly.

That said, the downside is there are a lot of people who have gotten involved lately with a “get rich quick” mentality, not to mention many questionable projects and outright scams. Those who will succeed, though, will have a long term vision and understand that this is a marathon, not a sprint, and this technology will take years of hard work to build out.

Tell us about your leadership style and philosophy in life.

As I’ve been involved with open source communities for over a decade, my leadership style is very collaborative. I believe in giving people agency and autonomy, not just telling people what to do. That said, communication is of utmost importance, especially with a distributed team like our own. So I would say my style is a mix of collaboration, communication, and guiding people to work on their passion.

In terms of philosophy, I’m very much a doer. If I see a problem, my instinct is to fix it. I don’t like it when people tell me I can’t do something. When I first started my company, lots of people told me we were crazy or that it would never work. My approach is to ignore what people say and keep on building.

How could the tech industry be more inclusive for women?

First off, it’s crucially important to highlight the achievements of women and people of colour. The more other people can see role models that look like them, the more they will be encouraged to get involved. Second, creating more opportunities to learn, be it on the technology front or the business front. Along those lines, I have been involved in funding scholarships for female software engineers to learn more about the bitcoin protocol with bitcoin developer and educator Jimmy Song, which has been extremely successful in bringing more technical women into the fold. I’m also co-organizing a conference this autumn in California that will feature leaders in the industry giving technical and business talks about their work – who just happen to be women.

In terms of challenges, right now there’s a huge opportunity to reshape the global financial system as we know it. We need participation from all different kinds of people from all over the world, and part of that will be engaging more women in the process. It can be difficult if you’re the only woman in a room, at a conference, or event, as I have often been in the past. One good example was an event I attended recently in Morocco. For the first version of this event four years ago, there were very few women in attendance, and numerous people pointed it out. At this year’s event, it was 52% women, after the organizers made an effort to reach out to female leaders. So, for women who are feeling like they’re the only ones, know that there are many of us out there, and more to come.

I make it a point to refer to great women when it comes to speaking, press, or other opportunities because it’s important that their stories are told. The more a younger generation can see role models, the more inspired they will be to get involved.

1...517518519...675 Page 518 of 675

.cls-8{fill:#fff;stroke-width:0px;}

.cls-8{fill:#fff;stroke-width:0px;}