Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


What is BBVA 2.0?

No description

Sergio Alvarez-Teleña

on 21 May 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of What is BBVA 2.0?

What is BBVA 2.0?
Synergies' exploitation
across desks: execution strategies, pattern detection, code within each platform...
across regions: straight heritage in LatAm and US from UE.
across assets: equitization.
In order to keep decentralized the systematic trading activity across desks (whether electronic or client-driven flow) a new, organic structure shall be set out. A new layer with a cross reach but no hard P&L.
Cross reach
Short-run resources:
It is the consequence of a natural evolution in the banking industry towards automation.
As such, it is one of the strategic focuses of our President

Who ordered the set up of

to help us keep up with the industry
Data Scientist
US will be short of 190,000 data scientists by 2018.
approach along GM:
delivery optimally harmonized across desks, regions and assets depending on the expected added value of each intervention. Resources maximally utilized.
Our ongoing solution: Strategies & Data Science
Oficina Comercial does have a data scientist in charge of wallet sizing, and academic literature suggests that further advances will evolve around cross-selling through recommendations-like approaches (following the lead of ).
Regulation requires to prove a prudent risk management. And prudence is achieved with back-and-stress testing policies.
of assets &
of products implies tighter spreads and the seek for more frequent flow (i.e. a systematic trading approach).
There is also a natural evolution towards automation in GM:
Need to change the current interaction across teams to set out a more collaborative environment.
that ought to be cleared-up:
A challenging short run towards a flexible long run.
Algorithmic trading
Market making
VWAP of VWAPS of a 'sensitive trade' (relevant client + US holiday in between) with 0 bps of deviation.

By proving that 'Little Data' was enough to estimate the intraday volume's profile we were able to propose a simple, flexible way to activate our execution algorithm.
Mid-run resources:
CIB's representative:
2 data scientists
Current resources: 1 person optimizing several trading teams
Co-founder -> 1st projects for free
Needs internships' programme
Long-run resour.:
(talent discovery)
Financial computing
Guest lecturer since 2010
Big Data Innovation Summit
Keynote speaker 2014
Zero-latency price
Speaker since 2012
Industrial adviser MRes students
MRes: Marta Díez
(Oficina Comercial)
Optimal ETFs' construction on emerging markets.
+ intern
Set out basic execution strategies for both assets (mimic Equities).
Exploit the fact that electronic market participants/contributors are identified (advantage upon Equities).
Eg: Normalize liquidity in the FX's order book as contributors can
reject a trade
within a time frame. These liquidity deviations may obey to certain patterns that we should try to track-down.
So, it seems that BBVA 2.0 = Retail 2.0
... but, what about CIB 2.0?
Start by coordinating the transference from UE to Mexico of the ORC's market making platform to minimize issues in its set up.
AT: improvement for VWAP
how to match
50% bid-offer spread drop
recently observed in the markets - Menkveld (2012)
proposed elasticity of the flow
synthetic delta for
large sizes
- a recent sales' claim.
PT: calibration advances on stop-loss, take-profit and maximum time horizon towards a more robust out-of-sample performance.
A proposed solution to outsource the production of prototypes onto academia. Two types of interaction with SciTheWorld are available to us:

At inception
: its founding team could research for us in one or two projects for a '1eur contract' - i.e. with the incentive of having officially BBVA as their first client. Motivation is key in aligning their interests with ours.
: we transfer privileged (but non-sensitive) data, an insightful idea (however, far enough from our strategy not to be affected by its disclosure) and a resource (an intern) to a spot-on academic (identified by SciTheWorld - their real activity) to research for us. The researcher will have a valuable idea, data and a resource to easily publish a paper in exchange of a fine code. It seems that all incentives are aligned.

We could this way set out a factory for diverse, advanced and bespoke source of innovative prototypes at low-cost (interns). Finally, we develop only those whose prototypes seemed to be most valuable.
New competitors in the market making space, more innovative than the traditional school:
Does their prediction
affect CIB?
This is an all-or-none challenge
. Between nothing at all and Renaissance Technologies there is a large range of states. We need brains to fine tune the optimal state within such range.
Strategies can be postponed
. If there is a rule there is a (not calibrated) strategy. Rules shall be calibrated first and innovated then. There is no reason why not to try to calibrate rules from now in order to minimize the learning-by-doing process (both in terms of P&L realized and opportunity costs).
Data science is miraculous
. No. It just optimizes our capacities. It may still be 'magical' though as there always are 'some more tricks to show'.
Systematic trading = HFT = ultra-HFT
. It depends on the equilibria: brains vs platforms. BBVA should not focus on strategies that leverage on stat-of-the-art technology.
Shall CIB miss retail's momentum in terms of innovation
or ride it instead?
Strategies & Data Science
thinking BBVA
Quant IT
Full transcript