## Dataset overview — ch2 (GFSR Chapter 2 Oct. 2020)

## Source details

**Canonical URL:** [Dataset overview — ch2 (GFSR Chapter 2 Oct. 2020)](https://www.imf.org/-/media/files/publications/gfsr/2020/october/english/data/ch2.xlsx)

## Other formats

- [Markdown version](/-/media/files/publications/gfsr/2020/october/english/data/ch2.xlsx.md)
- [Structured JSON version](/-/media/files/publications/gfsr/2020/october/english/data/ch2.xlsx.json)

---

### Coverage and structure
- Workbook contains multiple worksheets tied to "Chapter 2. A GREATER SET OF POLICY OPTIONS TO RESTORE STABILITY".
- Key worksheets sampled: "GFSR Chapter 2 Oct. 2020", "Table of Contents", and multiple figure worksheets labeled Figure 2.1.–Figure 2.11. covering fiscal, monetary, central bank, reserves/FX, asset purchases, local stress indices, bond yields, frontier market debt, and related regression/specification results.

### Major themes and data series (by worksheet / figure)
- Figure 2.1. — Fiscal and policy response panels:
  - Panel headings include "1. General Government Deficit", "2. Monetary Policy Rates", "3. Central Bank Policy Actions".
  - Example series entries:
    - "2020 Deficit forecast in Oct. 2019" for selected economies:
      - Brazil: 6.943681259321692
      - Ghana: 6.093062769777843
      - South Africa: 6.74932363500371
      - India: 7.18996161779752
      - China: 6.316728720146668
      - Poland: 2.5263669893111045
      - Romania: 3.495397383692146
      - Colombia: 0.8692199327684694
      - Peru: 1.4095057611705624
      - Chile: 2.0657979398276822
      - Hungary: 1.56554479073743
      - Philippines: 1.7075921811753276
      - Turkey: 4.68488733845281
      - Egypt: 7.029485652198692
      - Indonesia: 1.8042493280989274
      - Thailand: 0.202630149122455
    - "Change in forecast of 2020 deficit" for selected economies:
      - Brazil: 9.83763363643866
      - Ghana: 10.323267683661227
      - South Africa: 7.293916593065105
      - India: 5.886999085840548
      - China: 5.5702335378621255
      - Poland: 7.937511538982416
      - Romania: 6.097735851647028
      - Colombia: 8.608260093600057
      - Peru: 7.998658213184742
      - Chile: 6.639101678709045
      - Hungary: 6.71274627552704
      - Philippines: 6.355571021592489
      - Turkey (row label): "TUR"
      - Central bank sample note: "Number of central banks, total sample = 50"

- Figure 2.2. — Reserves and FX Interventions; regression coefficients for EM FX surprise drivers:
  - Example median coefficient values (January 2020–May 2020):
    - FXI: 1.169, -2.795, -0.813 (presented across columns including "FXI" and "Median")
    - Policy rates: 0.242, -0.994, -0.376
    - VIX: -0.325, -0.4, -0.525
    - US FFR: -0.6391, -0.6536, -0.9659
  - Comparable coefficients for 2015 EM sell-off (April 2015–February 2016) shown in parallel columns (examples):
    - FXI: 0.7840000000000003, 2.824, 2.196
    - Policy rates: 0.16200000000000014, 3.252, 1.788
    - VIX: 0, 2.6992, 0.7928
    - US FFR: 0, 0.5, 0.2013

- Figure 2.3 — Central bank asset purchases and changes in local-currency government bond holdings (end-Feb–June 2020):
  - "Sovereign bonds" central bank purchases (Percent of GDP) examples:
    - UK: 11.76750951176992
    - US: 8.956514232633296
    - JPN: 7.751513583841474
    - EA: 5.153551716839853
    - POL: 4.537196898214792
    - CHL: 0
    - COL: 0.2813250715840684
    - THA: 0.5571388032303789
    - HUN: 0.5163897311871777
    - PHP: 5.739395891539129
    - IDN: 3.6999481761162993
    - IND: 0.8857688639790393
    - MYS: 0.6479170963187354
    - ROU: 0.49245912971056643
    - ZAF: 0.6331401758999717
    - TUR: 1.5635234898056793
  - For Turkey (row label "TUR"), holdings breakdown example:
    - Total Change in LC Debt Outstanding (example): 5.09370116978004
    - Domestic banks: 4.193895628638311
    - Central bank: 1.5033879709669997
    - Non-residents: -0.6303093270869049
    - Other domestic: 0.026726897261636563

- Figure 2.4 — Local Stress Indices (LSI), bilateral spillovers, and components:
  - Time series with dates and index values; sample date entries include "2005-07-05T00:00:00.000Z" and "2008-08-31T00:00:00.000Z".
  - Example FX LSI / Bond LSI components on "2008-08-31T00:00:00.000Z":
    - FX LSI: 0.13447007542698
    - FX LSI components: 0.123299755120213 (Realized volatility), 0.110098140687264 (Risk reversals), 0.0857111111111111 (Bid-Offer), 0.115354245216903 (Implied volatility), 0.0676866689448232 (FX basis)
    - Bond LSI: 0.160281252406425
    - Bond LSI components: -0.373647674790736 (Correlation), 0.0376666666666666 (Volumes (flows)), 0.133058493175782 (Asset Swap spreads), 0.137121549421193 (Realized volatility), 0.0804060461956521 (Curve term premium), 0.145676171737866 (Bid-offer)

- Figure 2.5 — Local Stress Indices: APP versus Non-APP economies; large time series sample (over 1,038,587 rows in sheet):
  - Example initial values on "2005-07-05T00:00:00.000Z":
    - APP sample: FX LCSI = 0.0958263270168256
    - APP sample: Bond LCSI = 0.118746301048089
    - Non-APP sample: FX LCSI = 0.0299189753290044
    - Non-APP sample: Bond LCSI = 0.487162250158416
  - Country sample labels appearing in the sheet include: "COL", "POL", "TUR", "IND", "ZAF", "HRV", "THA", "HUN", "CHL", "MYS", "IDN", "PHP", "RON"

- Figure 2.6. — EM 10-Year Government Bond Yields, 5-Year ACM Term Premiums, EM Currencies, and intraday asset price movements on APP announcement days:
  - Figure panels labeled:
    - "1. EM 10-Year Government Bond Yields"
    - "2. EM 5-Year ACM Term Premiums"
    - "3. EM Currencies"
    - "4. Intraday Asset Price Movement on the Days of Asset Purchase Program Announcements for Selected EMs"
  - Axis/annotation examples include "(Indexed at 0 on t = 0; basis points; days on x-axis)" and "(Indexed at 1 on t = 0; days on x-axis)".

- Figure 2.7 — Specification results: effects of Variable X and Variable Y on bond yields:
  - Multiple specifications reported; sample regression output row (coeff, std, up, low) for t = 0:
    - Specification 1 (X = Domestic APP Announcements): coeff = -0.1508945, std = 0.0553209, up = -0.09557359999999998, low = -0.2062154
    - Specification 2 (Y = Domestic APP Announcements or alternative Ys) sample: coeff = 0.1059749, std = 0.066124, up = 0.1720989, low = 0.039850899999999995
    - Additional specification sample: coeff = -0.0520504, std = 0.1137811, up = 0.0617307, low = -0.1658315
    - Another sample (coeff, std, up, low): coeff = -0.1778473, std = 0.0601313, up = -0.11771600000000002, low = -0.2379786
    - For alternative outcome: coeff = 0.0368309, std = 0.0160677, up = 0.052898600000000004, low = 0.0207632
    - Sample: coeff = -0.145163, std = 0.1364157, up = -0.008747299999999986, low = -0.28157869999999996

- Figure 2.8. — Asset Purchase Program Announcements and Domestic Currencies; Bond Spread Misalignments:
  - Currency effect regressions (coeff, std, up, low) sample rows:
    - t = 0: coeff = -0.2972496, std = 0.2465937, up = -0.050655900000000004, low = -0.5438433
    - t = 1: coeff = -0.2439005, std = 0.2952002, up = 0.05129970000000003, low = -0.5391007
    - t = 2: coeff = -0.3989921, std = 0.4946539, up = 0.09566179999999996, low = -0.8936459999999999

- Figure 2.10 — Frontier Market Debt: creditor composition, debt service, private vs official:
  - Time series and composition panels with date columns "2006-01-01T00:00:00.000Z" through "2018-01-01T00:00:00.000Z".
  - Example aggregated series ("Total" percent of GDP and sample values):
    - Total (selected years series sample): 17.610843195226263, 16.204322341959088, 13.694850692595256, 15.16656213535326, 14.231951416565005, 13.375376613092909, 13.201666804465795, 14.141868212350179, 14.530496801420888, 16.80951813213943, 19.174463567756288, 22.07090274897113, 22.903426139993464
  - "Bonds" series sample (percent of GDP):
    - 2.1726499839078315, 2.082786466089872, 1.6808444418058142, 1.3508090363545036, 1.3299295782474847, 1.2342073638574005, 1.455487001752645, 1.9915889364152641, 2.4858504298296022, 3.2196896258348264, 3.676210793802533, 4.9646136616044725, 5.695512883771124
  - Creditor-share labels present in sheet: "Bonds and Loans", "Bilateral (Official)", "Multilateral", "Paris Club Bilateral", "Non-Paris Club Bilateral"
  - Country labels appearing in context include "AGO", "BEN", "CMR"

### Data characteristics and notes
- Date-time stamps use ISO 8601 format, e.g., "2020-10-20T00:00:00.000Z", "2005-07-05T00:00:00.000Z".
- Numeric precision is high throughout (many values recorded with many decimal places, e.g., 11.76750951176992).
- Some sheets include rich text labels and formulas (examples: formula references to other sheets and formula results such as K11+L11 -> -0.09557359999999998).
- Several worksheets contain very large time series (Figure 2.5 contains 1,038,587 rows), indicating panel and high-frequency data coverage.

*Dataset: ch2 (GFSR Chapter 2 Oct. 2020) — extracted worksheet samples and key numeric series as present in the provided workbook*

---


_Source: https://www.imf.org/-/media/files/publications/gfsr/2020/october/english/data/ch2.xlsx_