Fintech Policy EU

24/06/2024

Did you know that was the most targeted sector for ? It recorded 27 % of all breaches in 2023, rising fom 19 % in 2022.
➡ https://www.kroll.com/en/insights/publications/cyber/data-breach-outlook-2024

According to IBM the global average cost of a was $4,45 million.
➡ https://www.ibm.com/reports/data-breach

But in finance, that number amounted to $5,9 million per incident!
➡ https://www.ptsecurity.com/ww-en/analytics/financial-industry-security-interim-2023-en/

With one attack affecting over 1,000 institutions at the same time!
➡ https://fortune.com/2024/02/09/data-breaches-financial-industry-ransomware-gang-kroll-report/

With the rising use of and generally , extended rules on that only covered institutions and prepared a new, more comprehensive , Digital Operational Resiliency Act, which covers the entre supply chain in the financial sector, including, for example, data providers.
s are preparing for implementation:
➡ European Securities and Markets Authority (ESMA)➡
https://www.esma.europa.eu/esmas-activities/digital-finance-and-innovation/digital-operational-resilience-act-dora

➡ European Banking Authority (EBA)➡
https://www.eba.europa.eu/activities/direct-supervision-and-oversight/digital-operational-resilience-act/preparation-dora-application

➡ European Insurance and Occupational Pensions Authority (EIOPA)➡
https://www.eiopa.europa.eu/digital-operational-resilience-act-dora_en

But so should you!
To better understand the issues, the and the for
✅ register for a free conference with EU experts in the field that will be held in Brussels this Thursday at 18h00, organised by Life in Codes and their Gabriele Mele.

Our Tilen Cuk will present some breaches in finance to show the importance of proper measures.

✅ Link for registration in the first comment below! ⏬

Mira Zlati
Raphael Gyory

11/03/2020

🚨 BIG NEWS ALERT🚨
FCA, the UK financial regulator is launching a study on the use of market and alternative data in wholesale financial markets !

☝️DID YOU KNOW?
Fun fact #1
Some firms buy satellite images and use image recognition algos to count the number of cars in front of supermarkets and predict their future revenues even before the supermarket chain itself!
Is that ? Surely not in most cases but the advantages a data driven firm can get today are pretty strange.

Fun fact #2
Some HFT firms spend hundreds of thousands of dollars to buy privileged access to market-moving information just a second or two before others.

Fun fact #3
Did you know some companies track private jet flights and reseravations in fancy Wall Street restaurants in order to predict a merger?

🤓The FCA study is excellent news and we have already worked on the topic! A year ago, our member Tilen Cuk wrote an article on the use of in exposing several problems with the use of and 🥳
And also provided a tentative solution by using a sort of effects-based approach.

👉The article: T. Cuk, Big data sur les marchés financiers : vers une approche par les effets, International Journal for Financial Services, Vol 2019, Issue 3-4, pp.10-24.

DO YOU WORK IN THE FIELD AND WOULD CARE TO SHARE YOUR VIEW ON THE USE OF ARTIFICIAL INTELLIGENCE IN FINANCIAL MARKETS? Just contact us, we'd be thrilled to discuss and include your views in our consultation response and research.



👉Link to the FCA call:

LinkedIn: https://www.linkedin.com/posts/financial-conduct-authority_accessing-and-using-wholesale-data-call-activity-6642685011336273920-z6FQ

Web: https://www.fca.org.uk/publications/calls-input/accessing-and-using-wholesale-data-call-input

15/11/2019

⚠️CROSS-MARKET MANIPULATION ALERT
🆘MORGAN STANLEY potentially facing a record €25M fine in France

☝️A possible record fine on the horizon for Morgan Stanley by the French regulator AMF for spoofing the market for French bonds.

Crypto world, beware.



☝️For the first time in the world, a sanction for cross-market spoofing.
❌Morgan Stanley increased the price on an exchange, then liquidated positions in dark pools.



⚠️After Tower Capital Research $67,4 million fine yesterday by the CFTC for spoofing

👉see Fintech Policy EU analysis here

https://www.facebook.com/345539049135566/posts/981674842188647

15/11/2019

⚠️MARKET MANIPULATION ALERT
🆘RECORD SPOOFING FINE:
Tower Research Capital, one of the top 5 high frequency trading firms, ordered by the US Commodities and Futures Trading Commission CFTC to pay $67,4M charges to settle spoofing charges



📜Fine comprises:

$32,593,849 in restitution
$10,500,000 in disgorgement
$24,400,000 civil monetary penalty

〽️💔instruments traded:

E-mini S&P 500
E-mini NASDAQ 100
E-mini Dow

❌On thousands of occasions the Traders placed orders to buy or sell futures contracts that they intended to cancel before ex*****on at the time the orders were placed. Traders intentionally sent false signals of increased supply or demand designed to trick market participants.
🤖Big data (AI/ML) techniques are used for multi venue market making but exact calibration of algorithms (for ex. In terms of order size) might cause disturbances.

⚠️HFT firm Tower Research Capital is also facing a class action in the US for spoofing on Korean Exchange.
😱 Orders on the KRX night market are matched by a Chicago Mercantile Exchange Globex electronic trading platform in Illinois, USA.


👉Choi et al v Tower Research Capital LLC et al, 2nd U.S. Circuit Court of Appeals, No. 17-648

⚡️Spoofing had been prosecuted before Dodd-Frank Act 2010 but rather unsuccessfully under Section 10b-5 Securities Exchange Act 1934 in particular because of the scienter and fraud on the market theory.

💫Dodd-Frank made spoofing a specific offence requiring only beyond recklessness standard.
Spoofing is defined in Commodity Exchange Act, 7 U.S.C. § 6c.

☄️Regulators did not hesitate to use it and won a landmark criminal case in Michael Coscia. Other important cases were Navinder Singh Sarao and Igor Oystacher. A significant number of cease and desist orders were issued by the SEC and CFTC and fines by the exchanges are accumulating

❄️The Flotron and Thakkar cases however showed prosecution mainly by statistics might not always go through. The facts and indictments were different though
👉 https://www.facebook.com/345539049135566/posts/830389207317212

📜Links
👉Press release:
https://www.cftc.gov/PressRoom/PressReleases/8074-19

👉ORDER:
https://www.cftc.gov/media/2986/enftowerresearchorder110619/download

Interestingly there was a dissenting opinion on one of the points:
👉 Dissenting Statement of Commissioner Dan M. Berkovitz on the waiver of SEC “Bad Actor” Disqualifications
👉 Conflict od competence between the CFTC and SEC
👉 Link:
https://www.cftc.gov/PressRoom/SpeechesTestimony/berkovitzstatement110719

25/10/2019

Man vs machine

Wanna see how sees you?

Detectron 2, the recent upgrade of Facebook’s object recognition with its different modules/layers. Check the video for a machine POV!

https://ai.facebook.com/blog/-detectron2-a-pytorch-based-modular-object-detection-library-/

Don’t treat machines badly. Machines just don’t think like humans.
See, to a you’re so much more than just a .
Thread on Twitter with images.
👇
https://twitter.com/FintechPolicy/status/1185868177609318400?s=20

02/08/2019

💫SEC Commissioner finds crypto regulatory competition healthy

🐼🐻 « Renegade Pandas: Opportunities for Cross Border Cooperation in Regulation of Digital Assets »

Commissioner Hester Peirce
Singapore, July 30, 2019

Remarks before the SUSS Convergence Forum: Inclusive Blockchain, Finance, and Emerging Technologies

☝️Global rules are not a solution as per H.Peirce of the

🤷‍♂️Nothing really new:
The cross-border concerns in crypto track standard concerns of Cross-Border markets from 1980’s on

Original statement from Singapore:
bit.ly/2KdDbHE

🤯1. about what the in any particular are
🤯2. pinning down a domicile for is difficult
🤯3. No on the precise nature of the object/ : , , ,

⚔️Regulatory competition is healthy and an opportunity for regulators to learn from each other

🎯SEC’s enforcement action targeted fraudulent not just unregistered

🦚🐲🦜 “Laboratories of regulation” bring wild animals under control
, , ,

💫As FPEU has been saying for over two years
🐚 are the best “A non-exclusive would permit issuers to offer under an alternative regime with robust requirements.”

H.P. “ framework may not be the appropriate one for all ”

FPEU also made links to algo description mandatory in and
“a plain English description of the ’ functionality and the functionality as written in the code”

🌞Healthy

«While a single would be unwise can create a healthy environment for this new market to grow by sharing that will smooth crossborder while stamping out &other harmful activity.»

Remarks before the SUSS Convergence Forum: Inclusive Blockchain, Finance, and Emerging Technologies by Commissioner Hester M. Peirce in Singapore on July 30, 2019

20/07/2019

30/05/2019

Great discussion in Malta on legal personality for DAOs!

Cheers to Leonard Bonello, Steve Tendon, Amandine Doat and Irina Heaver!

Our take: DAOs seem like they’re algorithmic monsters because the existing regulations are based on now outdated conceptions of capitalistic firm! They now allow great opportunities for legal engineering!



Sea, sun and food for thought!

https://twitter.com/fintechpolicy/status/1134101906442338304?s=21

29/04/2019

Nils Nilsson, 86, Dies

💪The Scientist Who Helped Robots Find Their Way
https://nyti.ms/2vknB4J

💫 who helped develop the first general-purpose

💫Co-invented for the to move efficiently & perform simple tasks

💡Member of Stanford Research Institute who pioneered like or speech-control systems like Siri.

😂Pentagon asked if his robot could carry a gun. “How many do you need? I think it should easily be able to handle two or three.”

😂👋The robot “shook like hell” when it moved, so it was named Shakey.

👑In 1970, Life Magazine, called Shakey “the first electronic person”

👨‍🦳Great grand-father of Sophia, 50 years before! (see picture 2)

🥇In 1965, published Learning : Foundations of Trainable -Classifying Systems. One of the 1st books on

🤓”A academically inclined,” Dr. Hart said. “He didn’t have any interest in or, as he would call it, industry.

RIP master!

15/04/2019

Liability of / for their
😇Shooting stars: from expert advisor to spoofing criminal
👉Spoofing algo programmer partially acquitted

🤦‍♂️CFTC and FBI fail with indictment of the programmer who developed a customized spoofing algorithm at the heart of the infamous 2010

Jitesh Thakkar, the founder and president of Edge Financial Technologies, one of the leading firms for software development for and high-frequency trading, was acquitted on April 9th, 2019 of the count of to commit .

👉There are still counts two and three of criminal (1:18-cr-00036) pending for spoofing and a civil case (1:18-cv-00619) brought against Thakkar by the CFTC for aiding and abetting in spoofing.
Interestingly, Thakkar served as an advisor to the SEC on high-frequency in 2012 and 2013.

⚠️Thakkar’s case is the first in the US where a programmer is prosecuted for his involvement in the development of a and shows ’ willingness to prosecute programmers in addition to the users of such algorithms.

☝️The alleged facts took place in 2013 but since then things have changed as NFA (oversight of futures markets) issued Interpretive Notice 9055 (in effect since September 19, 2016) and FINRA (oversight of securities markets) issued the Regulatory notice 16-21 (in effect since January 30, 2017) making programmers also responsible for developing or helping develop algorithms used for market abuse.
In the futures , CFTC’s new proposed Regulation Automated Trading will soon impose similar obligations. The new will make it easier for the regulators to chase programmers down the algorithmic supply chain.

⚠️We first heard of the programmer, who remained anonymous at the time, who developed a customized spoofing algorithm in points 10-21 of the indictment of Navinder Singh Sarao.

〽️In November 2017 Navinder Singh Sarao pleaded guilty for spoofing the E-mini markets in 2010 on Chicago’s famous , CME, from his small room in the London suburbs. The case was famous because Navinder’s spoofing was advanced by the CFTC as the principal explanation of the infamous 2010 when $1 trillion of vanished in about 10 minutes.

While more often than not many differing explanations of algorithmic mishaps exist, US regulators are determined to find the “bad guys”.

♻️Navinder’s case was the second successful spoofing in the US, after Michael Coscia and his firm Panther Trading.

♻️However, lately they seem to be running out of luck. UBS Flotron was acquitted of conspiracy to commit spoofing and now Thakkar is partially acquitted. In the conspiracy to commit fraud the government has to show the defendant possessed of the algorithm user’s illicit purpose beyond a reasonable doubt which is particularly difficult.

⚠️It remains to be seen whether the regulators can prosecute cases of spoofing and whether spoofing is a problem of , of “bad guys” or both.