Decision support system for financial liquidity planning

  • View
    166

  • Download
    0

  • Category

    Science

Preview:

Citation preview

Master’s Thesis Tallinn University of Technology Department of Computer Engineering Computer Systems Design

DECISION SUPPORT SYSTEM FOR FINANCIAL LIQUIDITY PLANNING

Author: Erik Kaju Supervisor: Tarmo Robal (PhD)

15.06.2015

THE OBJECTIVE

The objective of this thesis is to use information technology facilities and build a minimum viable product solution that would potentially enhance a liquidity planning process in the world’s fastest growing money transfer service.

2

TRANSFERWISE

TRANSFERWISE

TransferWise (TW) is an international money transfer platform. It makes it up to 10 times cheaper to send money abroad compared to using similar services offered by banks. TW’s technology is based on peer-to-peer system and has helped customers to move more than £4,5bn – an approach that has saved customers£180m.

4

THE PROBLEM

IMBALANCE OF OF CURRENCY FLOWS

5

DIFFICULTY OF FORECASTING

6

DIVERSITY OF FACTORS

THAT INFLUENCE

THE DEMAND FOR MONEY

RAPIDLY CHANGING

ENVIRONMENT RANDOMNESS

FACTORS ANALYSIS

INTERNAL •  Liquidity buffer account balance •  Suspended transfers •  Liquidity in transit

7

EXTERNAL •  Recent volume descriptor •  Growth trends •  Weekly patterns •  Monthly patterns •  Exchange rate movements •  Frequency of extra large payments •  National holidays

WEEKLY PATTERNS

8

MONTHLY PATTERNS

9

TIME SERIES ANALYSIS

A time series is a collection of observations of well-defined data items obtained through repeated measurements over time.

COMPUTATIONAL MODEL

WEEKLY PATTERNS

12

MONTHLY PATTERNS

13

VISUALISATIONS

14

FORECASTING ALGORITHM

15

LIQUIDITY RECONCILIATION DECISION SUPPORT SYSTEM

THE IMPLEMENTATION

17

GRAILS GROOVY JAVA

THE PROTOTYPE

18

THE PROTOTYPE

19

THE PROTOTYPE

20

TESTING

21

TESTING, RESULTS

22

THE OUTCOME

23

THE PROPOSED NEXT STEPS 1.  Include more factors into

calculational model 2.  Enhance the proposal

algorithm 3.  Carry out tests and compare

the efficiency vs. current human factor

4.  Based on results decide whether more determinants need implementing

5.  If needed, repeat steps 1,2,3 6.  Go live 7.  Continuously improve the

solution

PLAN: WHAT TO DO NEXT

THANK YOU

24

QUESTION FROM REVIEWER

Q: Kuidas suhtute mõttesse, eemaldada andmetest põhitrendid ja siis katsetada jääki juhuslikkusele ja alles seejärel tuua sisse lisafaktoreid? How do you find the idea of removing main trends from data and test the irregular component and only then introduce extra factors. (perform seasonal adjustment)? A: Aegridade teooria järgi oleks selline lähenemine õige, aga kuna seadsin antud tööle järgneva tuleviku väljavaateks just olulisemate ettevõtteväliste ja sisemiste sesoonsusest sõltumatute lisafaktorite mõju uurimise ja vähendamise, siis põhitrendide analüüsi jätsin teadlikult kõrvale. According to the theory of time series, such approach is correct. I have set a goal for future development after this thesis to mitigate the impact of main internal and external determinants that are independent from seasonalities. For that reason I decided not to perform classical seasonal adjustment.

25

Reviewer: Enn Õunapuu

Recommended