kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Modelling of Capital Requirements using LSTM and A-SA in CRR 3
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematics (Div.).
2022 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Modellering av kapitalkrav med hjälp av LSTM och A-SA i regelverket CRR 3 (Swedish)
Abstract [en]

In response to the Great Financial Crisis of 2008, a handful of measures were taken to increase the resilience toward a similar disaster in the future. Global financial regulatory entities implemented several new directives with the intention to enhance global capital markets, leading to regulatory frameworks where financial participants (FPs) are regulated with own fund's requirements for market risks. This thesis intends to investigate two different methods presented in the framework Capital Requirements Regulation 3 (CRR 3), a framework stemming from the Basel Committee and implemented in EU legislation for determining the capital requirements for an FP. The first method, The Alternative Standardised Approach (A-SA), looks at categorical data, whereas the second method, The Alternative Internal Model Approach (A-IMA), uses the risk measure Expected Shortfall (ES) for determining the capital requirement and therefore requires the FP to estimate ES using a proprietary/internal model based on time series data. The proprietary model in this thesis uses a recurrent neural network (RNN) with several long short-term memory (LSTM) layers to predict the next day's ES using the previous 20 day's returns. The data consisted of categorical and time series data of a portfolio with the Nasdaq 100 companies as positions. This thesis concluds that A-IMA with an LSTM-network as the proprietary model, gives a lower capital requirement compared to A-SA but is less reliable in real-life applications due to its behaviour as a "black box" and is, thus, less compliant from a regulatory standpoint. The LSTM-model showed promising results for capturing the overall trend in the data, for example periods with high volatility, but underestimated the true ES.

Abstract [sv]

Efter finanskrisen 2008 vidtogs flera effektiva åtgärder av världens största finansiella myndigheter som ett svar på det tidigare icke transparenta klimatet inom finanssektorn med intentionen att förstärka de globala kapitalmarknaderna. Detta innebar att nya samt strängare regelverk etablerades med direktiv så som hårdare kapitalkrav. Detta examensarbete är en empirisk undersökning samt jämförelse mellan två metoder i regelverket "Captail Requirements Regulation 3" (CRR 3) som kan användas för att beräkna en finansiell institutions kapitalkrav. Den första metoden, så kallad "Den alternativa schablonmetoden" (A-SA), använder kategorisk data för att beräkna kapitalkravet medan den andra metoden, "Den alternativa internmodellen" (A-IMA), kräver en att först beräkna riskmåttet "Expected Shortfall" (ES), med hjälp av en internmodell baserad på tidsseriedata, för att sedan kapitalkravet ska kunna beräknas. CRR 3 innehåller tydliga riktlinjer om hur en sådan internmodell ska utformas och i detta projekt testas en modell baserad på "återkommande neurala nätverk" (RNN) med den specifika arkitekturen "Long Short-Term Memory" (LSTM) för att estimera ES. De slutsatserna som kan dras är att A-IMA med en LSTM-modell, ger ett mindre kapitalkrav än A-SA. Däremot är A-IMA mindre tillförlitliga inom riskappliceringar på grund av risken att neurala nätverk kan bete sig som svarta lådor, vilket gör modellen mindre kompatibel från ett regelverksperspektiv. LSTM-modellen visade sig kunna upptäcka den generella trenden i portföljdatan (exempelvis perioder med hög volaitet) men gav konservativa prediktioner i jämförelse med testdatan.

Place, publisher, year, edition, pages
2022. , p. 65
Series
TRITA-SCI-GRU ; 2022:368
Keywords [en]
Capital Requirements, machine learning, neural networks, financial mathematics, risk management, CRR 3, FRTB
Keywords [sv]
Kapitalkrav, maskininlärning, neurala nätverk, finansiell matematik, riskhantering, CRR 3, FRTB
National Category
Other Mathematics
Identifiers
URN: urn:nbn:se:kth:diva-337082OAI: oai:DiVA.org:kth-337082DiVA, id: diva2:1799827
External cooperation
Zeb Consulting AB
Subject / course
Financial Mathematics
Educational program
Master of Science - Applied and Computational Mathematics
Supervisors
Examiners
Available from: 2023-09-26 Created: 2023-09-25 Last updated: 2023-09-26Bibliographically approved

Open Access in DiVA

fulltext(2098 kB)323 downloads
File information
File name FULLTEXT01.pdfFile size 2098 kBChecksum SHA-512
41f3cb58d9f80f709eacabfa1fe44fe08a48e592ae0c73ae2f0c72588b6be10c3e49bc239d8065e487ee3f31eebf42080e401be7dd2e57a1f08ee40da5845791
Type fulltextMimetype application/pdf

By organisation
Mathematics (Div.)
Other Mathematics

Search outside of DiVA

GoogleGoogle Scholar
Total: 323 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 321 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf