Swap Data Repository and Networks: Buy Side Implications

Terry Leitch
16-September-2016

Swap Data Repository and Networks: Buy Side Implications

A series of connections

  • Networks and Risk
  • Swap Data Repository
  • “Flows”

Drawing

Drawing

Graph Credits

  • Demirer, M., Diebold, F.X., Liu, L. and Yilmaz, K. (2015), “Estimating Global Bank Network Connectedness,” Manuscript, MIT, University of Pennsylvania, and Koc University.

  • Sanjiv Das, Matrix Metrics (2016): Network-Based Systemic Risk Scoring. Journal of Alternative Investments, Special Issue on Systemic Risk, v18(4), 33-51.

Issue with this approach

  • A lot of assumptions
  • Inference implied from messy data
    • Stock returns
    • CDS
    • Credit ratings

Quantitative Rube Goldberg?

Drawing

Swap Data Repository

Examination of the SDR illustrated potential alternative

  • All bilateral trades required reporting
    • Market
    • Structure
    • Price
    • Counter-parties

Swap Data Repository

  • This data, can be translated into a bilateral directed network
    • Risk Exposure
    • Trade flows
      • Duration
      • DV01
      • Vega
      • Index delta
      • Convexity

Granularity

Unfortunately the bilateral details are not available to the buy side for analysis

  • Only regulators can examine node info
  • Broad network research will probably require government investment

But there is “anonymous” trade data reported throughout the day

  • Everything but the counter-parties
  • We know data is reported by dealers

Adding Some Knowledge

Data includes sell-side and buy-side

  • Sell-side has many more connections
  • Dealers report trades
  • If two dealers trade, fixed payer reports
  • Most buy side is receive fixed

Central core of dealers provide “flows” to buy side clients

Hub and Spoke Model

Drawing

Drawing

One Month of Reported Trade DV01 Flow

Drawing

One Day of Reported Trade DV01 Flow

Drawing

90 Minutes of Reported Trade DV01 Flow

Drawing

Other Flows

Could also analyze

  • Yield Curve Segments
  • Option/Swaption Vega
  • Convexity
  • Index Delta
  • CDS DV01

Questions Buy Side Should be Asking with respect to Public Data

  1. Can we improve execution with better knowledge of flow?
  2. Can the flow be acceptably modeled to improve execution?
  3. How have exogenous changes affected the flow?
    • Crisis
    • Policy
    • Entrance/Exits
  4. How have costs changed as new policies are implemented

Questions Buys Side Could ask of the Regulators

  1. Can you analyze and monitor the net flow from dealers to buy-side?
  2. How has the hub/spoke relationship changed through time?
  3. Can you measure policy change impact on
    • Prices
    • Volumes