## wp1791

## Source details

**Canonical URL:** [wp1791](https://www.imf.org/-/media/files/publications/wp/2017/wp1791.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp1791.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp1791.pdf.json)

---

### II. A brief primer on submarine fiber-optic cables
- Historical development:
  - 1842: Samuel Morse submerged a copper cable in New York Harbor.
  - 1850s: Gutta-percha–insulated cables connected Great Britain with the European continent.
  - 1866: Successful transatlantic copper cable after eight years of failed attempts.
  - 1890s: Science of transmitting higher frequencies began to be established.
  - Early twentieth century: Practical vacuum-tube–based repeater amplifier developed; commercialization delayed by two world wars and the Great Depression.
  - 1955–September 25, 1956: First modern submarine cable TAT-1 laid/ inaugurated; TAT-1 connected Oban, Scotland with Clarenville, Newfoundland; it had 36 channels enabling it to carry 35 simultaneous telephone calls along with 22 telegraph lines on the thirty-sixth channel.
  - 1960s: Development of coaxial cables using transistors.
  - 1980s: Coaxial cables superseded by fiber-optic cables.
  - December 1988: First submarine fiber-optic cable TAT-8 entered service; capacity of 40,000 circuits (tenfold increase relative to coaxial cables).
  - 1989: PTAT-1 fitted with shark shielding and was the first fiber-optic submarine cable financed entirely privately.
- Technical attributes:
  - Fiber made by stretching glass/silica to roughly the diameter of a human hair.
  - Transfer data at a speed of 180,000–200,000 kilometers per second (i.e., the speed of light in glass).
  - Latency per kilometer of 5 to 5.5 microseconds (a 10 to 11 millisecond delay for a roundtrip of 1,000 kilometers).
  - Increase bandwidth significantly relative to coaxial cables and reduce losses in signal transmission.
- Reliability and hazards:
  - Early fiber cables attracted sharks due to absence of electrical interference shielding; shark shielding introduced starting PTAT-1 (1989).
  - Landing points chosen for gently sloping, sandy or silty sea-floors without strong currents to minimize costs and risk of damage.
  - Geography and seabed topography heavily influence layout; cables can only connect terrestrial points with direct access to the ocean.
  - Network contours mirror earlier telegraph and coaxial networks due to information about seabed conditions and convenient landing sites.
  - Safety and strategic considerations and high installation and maintenance costs lead to routes being persistent and ownership by large telecommunications consortia.
- Role for electronic trading:
  - Availability of fiber-optic cables with shark shielding from 1989 is dated as the initial availability of ICT needed to support long-distance electronic foreign exchange trading.
  - 2010: Spread Networks unveiled an 827-mile terrestrial cable from Chicago to New Jersey reducing latency from 17 to 13 milliseconds; cited as first terrestrial cable laid for electronic trading purposes.
  - 2015: Hibernia Express is the first submarine cable laid for the express purpose of electronic trading.
  - Empirical strategy uses only submarine cables laid between 1989 and 2002, "almost a decade before investors sought to lay them with electronic trading in mind."
- Market share and data:
  - By 2006, 99 percent of international communications traffic was carried by submarine cables and the remainder by satellite.
  - TeleGeography’s interactive Submarine Cable Map; data on 368 submarine cables starting in 1989 including cable profile, name, year ready for service, length, owners, and geographical coordinates of landing points.

### III. Empirical framework and hypotheses
- Theoretical foundation:
  - Framework derives from microstructure literature on foreign exchange and the effects of macroeconomic news (Andersen et al. 2003); exchange rate treated as forward-looking asset price (Engel and West 2005).
  - Log exchange rate st = (1−θ) ∑ θ^i Et[ f_{t+i} | Ωt ] (i from 0 to ∞); returns rt = st − st−1 driven by changes in market expectations.
- Operationalization of news:
  - Standardized news: S_t^k = (A_t^k − F_t^k) / σ̂_k, ∀ k ∈ I, where A_t^k is announced value, F_t^k is market expected value (median forecast from Bloomberg), σ̂_k is estimated sample standard deviation of (A_t^k − F_t^k).
  - I includes U.S. real GDP, CPI inflation, industrial production, unemployment rate, trade balance, and the monetary policy interest rate.
  - Focus on U.S. news because exchange rates analyzed are bilateral rates vis-à-vis the dollar.
- Baseline regression:
  - Equation (1): r_{j,t} = α_j + ∑_{k∈I} β_k S_t^k + ε_{j,t}, where r_{j,t} is daily log exchange rate return against U.S. dollar for currency j, α_j currency fixed-effect.
  - Sample: 56 bilateral exchange rates vs. the U.S. dollar between January 1, 1997 and November 30, 2015, totaling 240,430 observations; daily fixings at New York closing time (17:00 p.m. E.S.T.). Data on macro announcements from Bloomberg.
- Modeling ICT effects:
  - Point-to-point connection defined as cable directly linking country j to matching servers of EBS and Thomson Reuters in New York, London, and Tokyo.
  - Equation (2): r_{j,t} = α_j + ∑ β_k S_t^k + γ Cables_{j,t} + ∑ [δ_k S_t^k × Cables_{j,t}] + ε_{j,t}, where Cables_{j,t} is dummy = 1 if point-to-point connection at time t.
- Hypotheses (two-sided testing implied):
  - "Thin-skinned" hypothesis null H0: |E[r_{j,t} | S_t^k, Cables_{j,t}=1]| > |E[r_{j,t} | S_t^k, Cables_{j,t}=0]|.
  - "Thick-skinned" hypothesis null H0: |E[r_{j,t} | S_t^k, Cables_{j,t}=1]| ≤ |E[r_{j,t} | S_t^k, Cables_{j,t}=0]|.
- Identification premise:
  - Exogeneity arguments: cable layout shaped by seabed topography, earlier telegraph/coaxial networks, safety and strategic considerations, and general telecom purposes (not laid for electronic trading between 1989 and early 2000s).
- Cable data:
  - TeleGeography data collected by Global Bandwidth Research; 368 submarine cables from 1989; includes year ready for service and landing point coordinates.

### IV. Basic results
- Estimation approach:
  - Equation (1) estimated by pooled OLS; standard errors robust to heteroskedasticity and clustered by currency.
  - Table 1: full sample columns (1)–(3); G10 in (4)–(6); non-G10 in (7)–(9). Column (1)/(4)/(7) include all observations; (2)/(5)/(8) restricted to news days; (3)/(6)/(9) include time and currency-fixed effects.
- Replication and magnitudes:
  - Positive U.S. news on policy interest rates, unemployment and trade balance leads to appreciation of the dollar against local currencies.
  - A one-standard deviation positive news shock to U.S. policy interest rates leads on average to a 0.07 percent appreciation of the dollar against local currencies (consistent with Andersen et al.).
  - Policy interest rate news effect (Table 1, full sample): 0.068*** (0.010).
- Technology interaction results (Table 2):
  - Currencies of countries connected to EBS and Reuters matching servers by fiber-optic cables react less to U.S. economic news than currencies not connected.
  - Example magnitudes:
    - One-standard deviation positive news shock to U.S. policy interest rates:
      - ≈ 0.07 percent appreciation of the dollar vs. currencies of non-G10 countries not connected by cable.
      - ≈ 0.04 percent appreciation vs. currencies of countries that are connected — "about 50 percent smaller."
    - Coefficients reported:
      - Cables × policy interest rate = -0.018 (All), 0.010 (G10), -0.035** (Non-G10).
      - Cables × unemployment rate = -0.034** (All), -0.053** (G10), -0.036** (Non-G10).
- Interpretation:
  - Attenuation of exchange rate response with cable connectivity consistent with "thick-skinned" view: ICT levels the informational playing field, easing access to information beyond public price signals and limiting trend-following behavior.
  - Results robust to controls for fixed and time effects.

### V. Extended specification and robustness checks
- Trading location heterogeneity:
  - Distinguish onshore vs. offshore trading using BIS data for 55 currencies in seven years (1995, 1998, 2001, 2004, 2007, 2010, 2013); turnover allocated by location of initiating sales desk; annual values linearly interpolated to daily observations.
  - Equation (3) augments Equation (2) with Offshore_{j,t} dummy and triple interactions.
- Main patterns (Table 3 and Figure 5):
  - Effect of cable connections significant mainly for currencies traded mainly onshore, not for those traded mainly offshore.
  - Key coefficients (full-sample column 1, Table 3):
    - Policy interest rate coefficient: 0.097*** (0.033).
    - Cables × policy interest rate = -0.078** (0.037).
    - Cables × offshore × policy interest rate = 0.080+ (0.059).
  - Predicted responses (onshore currencies):
    - One-standard deviation positive news shock to U.S. policy interest rates leads to a 0.10 percent appreciation of the dollar against currencies of countries that are not connected by cable and are mainly traded onshore.
    - The same shock leads to a 0.02 percent appreciation of the dollar against currencies mainly traded onshore that are connected to fiber-optic cables — "80 percent smaller."
  - Offshore currencies: differences with and without cable connections are insignificant.
  - Robust to excluding outliers, time and fixed effects, and excluding the U.K. and Japan.
- Other news types (Table 4):
  - Heterogeneous results across GDP, industrial production, CPI, unemployment rate, trade balance.
  - General pattern: where U.S. news has a positive effect on the exchange rate, that effect is dampened by cable connections; dampening attenuated for currencies traded mainly offshore.
  - Example coefficients (Table 4, full sample):
    - Cables × U.S. news for policy rate (G10): -0.081** (0.039).
    - Cables × U.S. news for policy rate (Non-G10): -0.027*** (0.001).
    - Cables × U.S. news for unemployment rate: -0.039+ (0.025).
- Asymmetries and market uncertainty:
  - No significant asymmetry by sign of interest rate news: estimates for negative vs. positive policy news are indistinguishable.
  - VIX conditioning (top quartile vs. rest): coefficient patterns similar; consistent with thick-skinned hypothesis.
- Controlling for local news (Table 5):
  - Include both local and U.S. monetary policy surprises.
  - Key magnitudes (full sample, Table 5 column 1):
    - Local policy interest rate coefficient = -0.122* (0.072).
    - U.S. policy interest rate coefficient = 0.097*** (0.033).
    - Cables × U.S. policy interest rate = -0.078** (0.037).
    - Cables × local policy interest rate = 0.106+ (0.081).
    - Cables × offshore × local policy interest rate = -0.164+ (0.123).
    - Cables × offshore × U.S. policy interest rate = 0.080+ (0.059).
  - Findings:
    - ICT dampens reaction to both local and global news for onshore-traded currencies.
    - One-standard deviation positive local policy news shock leads to a 0.12 percent depreciation of the dollar against currencies of countries not cable connected and traded mainly onshore, versus a 0.02 percent depreciation when connected — "roughly 85 percent smaller."
- Robustness:
  - Results robust across subsamples: news days, exclusion of outliers, time-fixed effects, exclusion of U.K./Japan.
  - Sample size and timeframe: 240,430 daily observations for 56 bilateral exchange rates between January 1, 1997 and November 30, 2015.

### VI. Conclusion and policy implications
- Core conclusion:
  - Cable connections dampen the response of exchange rates to macroeconomic news.
  - Currencies traded mainly onshore respond less sharply to news emitted in the issuing country or in foreign countries like the United States when connected by submarine fiber-optic cables.
  - Estimated reduction in reaction of exchange rates to monetary policy news due to cable connections is "50 to 80 percent."
- Mechanism:
  - ICT appears to level the informational playing field by easing access to information beyond public price signals, reducing trend-following behavior.
  - Large-bandwidth-high-speed internet connections enable onshore traders to access and act on analysis produced in major financial centers, making them better informed and less prone to amplify trends.
- Policy relevance:
  - Various proposals to limit volatility and the role of high-speed trading include speed bumps, taxes, and bans on algorithmic trading.
  - The results do not speak directly to merits/demerits of these proposals but "suggest that the wider and more comprehensive provision of information may have a dampening effect on volatility."

### Key sample and data facts (exactly as reported)
- Sample: 56 bilateral exchange rates vs. the U.S. dollar.
- Sample period: January 1, 1997 to November 30, 2015.
- Observations: 240,430.
- Cable data: 368 submarine cables starting in 1989.
- Latency in glass: 5 to 5.5 microseconds per kilometer (10 to 11 millisecond roundtrip delay for 1,000 kilometers).
- Spread Networks terrestrial cable (2010) reduced latency from 17 to 13 milliseconds.
- By 2006, 99 percent of international communications traffic carried by submarine cables.
- Representative coefficients and magnitudes:
  - Policy interest rate news effect (Table 1, full sample): 0.068*** (0.010).
  - One-standard deviation policy interest rate shock → 0.07 percent appreciation of the dollar (Andersen et al. replication).
  - Cables × policy interest rate (Table 2, Non-G10): -0.035** (0.017).
  - Cables × policy interest rate (Table 3, full sample): -0.078** (0.037).
  - Cables × offshore × policy interest rate (Table 3): 0.080+ (0.059).
  - Local policy interest rate effect (Table 5): -0.122* (0.072).

*Source: https://www.imf.org/-/media/files/publications/wp/2017/wp1791.pdf*

### Section II provides a brief primer on submarine fiber-optic cables. Section III reviews our

### wp1791 - Section II provides a brief primer on submarine fiber-optic cables. Section III reviews our

### II. A brief primer on submarine fiber-optic cables
- Historical development:
  - 1842: Samuel Morse submerged a copper cable in New York Harbor.
  - 1850s: Gutta-percha–insulated cables connected Great Britain with the European continent.
  - 1866: Successful transatlantic copper cable after eight years of failed attempts.
  - 1890s: Science of transmitting higher frequencies began to be established.
  - Early twentieth century: Practical vacuum-tube–based repeater amplifier developed; commercialization delayed by two world wars and the Great Depression.
  - 1955–September 25, 1956: First modern submarine cable TAT-1 laid/ inaugurated; TAT-1 connected Oban, Scotland with Clarenville, Newfoundland; it had 36 channels enabling it to carry 35 simultaneous telephone calls along with 22 telegraph lines on the thirty-sixth channel.
  - 1960s: Development of coaxial cables using transistors.
  - 1980s: Coaxial cables superseded by fiber-optic cables.
  - December 1988: First submarine fiber-optic cable TAT-8 entered service; capacity of 40,000 circuits (tenfold increase relative to coaxial cables).
  - 1989: PTAT-1 fitted with shark shielding and was the first fiber-optic submarine cable financed entirely privately.
- Technical attributes of fiber-optic cables:
  - Made by stretching glass/silica to roughly the diameter of a human hair.
  - Transfer data at a speed of 180,000–200,000 kilometers per second (i.e., the speed of light in glass).
  - Latency per kilometer of 5 to 5.5 microseconds (a 10 to 11 millisecond delay for a roundtrip of 1,000 kilometers).
  - Increase bandwidth significantly relative to coaxial cables and reduce losses in signal transmission.
- Reliability and hazards:
  - Early fiber cables attracted sharks due to absence of electrical interference shielding; shark shielding introduced starting PTAT-1 (1989).
  - Landing points chosen for gently sloping, sandy or silty sea-floors without strong currents to minimize costs and risk of damage.
  - Geography and seabed topography heavily influence layout; cables can only connect terrestrial points with direct access to the ocean.
  - Network contours mirror earlier telegraph and coaxial networks due to information about seabed conditions and convenient landing sites.
  - Safety and strategic considerations and high installation and maintenance costs lead to routes being persistent and ownership by large telecommunications consortia.
- Role for electronic trading:
  - The availability of fiber-optic cables with shark shielding from 1989 is dated as the initial availability of ICT needed to support long-distance electronic foreign exchange trading.
  - 2010: Spread Networks unveiled an 827-mile terrestrial cable from Chicago to New Jersey reducing latency from 17 to 13 milliseconds; cited as first terrestrial cable laid for electronic trading purposes.
  - 2015: Hibernia Express is the first submarine cable laid for the express purpose of electronic trading.
  - Empirical strategy uses only submarine cables laid between 1989 and 2002, "almost a decade before investors sought to lay them with electronic trading in mind."
- Market share and dominance:
  - By 2006, 99 percent of international communications traffic was carried by submarine cables and the remainder by satellite.
  - Fiber-optic cables remain principal conduit for internet and electronic trading because they have much lower latency, larger bandwidth and better reliability than satellite.
- Data source for cable network:
  - TeleGeography’s interactive Submarine Cable Map; data on 368 submarine cables starting in 1989 including cable profile, name, year ready for service, length, owners, and geographical coordinates of landing points.

### III. Empirical framework and hypotheses
- Theoretical foundation:
  - Framework derives from microstructure literature on foreign exchange and the effects of macroeconomic news (Andersen et al. 2003); exchange rate treated as forward-looking asset price (Engel and West 2005).
  - Log exchange rate st = (1−θ) ∑ θ^i Et[ f_{t+i} | Ωt ] (i from 0 to ∞); returns rt = st − st−1 driven by changes in market expectations.
- Operationalization of news:
  - Standardized news: S_t^k = (A_t^k − F_t^k) / σ̂_k, ∀ k ∈ I, where A_t^k is announced value, F_t^k is market expected value (median forecast from Bloomberg), σ̂_k is estimated sample standard deviation of (A_t^k − F_t^k).
  - I includes U.S. real GDP, CPI inflation, industrial production, unemployment rate, trade balance, and the monetary policy interest rate.
  - Focus on U.S. news because exchange rates analyzed are bilateral rates vis-à-vis the dollar.
- Baseline regression (replication and extension of Andersen et al. 2003):
  - Equation (1): r_{j,t} = α_j + ∑_{k∈I} β_k S_t^k + ε_{j,t}, where r_{j,t} is daily log exchange rate return against U.S. dollar for currency j, α_j currency fixed-effect.
  - Sample: 56 bilateral exchange rates vs. the U.S. dollar between January 1, 1997 and November 30, 2015, totaling 240,430 observations; daily fixings at New York closing time (17:00 p.m. E.S.T.). Data on macro announcements from Bloomberg.
- Modeling ICT effects via cable connections:
  - Point-to-point connection defined as cable directly linking country j to matching servers of EBS and Thomson Reuters in New York, London, and Tokyo.
  - Modified regression (Equation (2)): r_{j,t} = α_j + ∑ β_k S_t^k + γ Cables_{j,t} + ∑ [δ_k S_t^k × Cables_{j,t}] + ε_{j,t}, where Cables_{j,t} is dummy = 1 if point-to-point connection at time t.
- Hypotheses:
  - "Thin-skinned" hypothesis null H0: |E[r_{j,t} | S_t^k, Cables_{j,t}=1]| > |E[r_{j,t} | S_t^k, Cables_{j,t}=0]|. Rejecting H0 is evidence against thin-skinned.
  - "Thick-skinned" hypothesis null H0: |E[r_{j,t} | S_t^k, Cables_{j,t}=1]| ≤ |E[r_{j,t} | S_t^k, Cables_{j,t}=0]|. Rejecting H0 is evidence against thick-skinned.
- Cable data:
  - TeleGeography data collected by Global Bandwidth Research; 368 submarine cables from 1989; includes year ready for service and landing point coordinates.
- Identification premise:
  - Exogeneity arguments: cable layout shaped by seabed topography, earlier telegraph/coaxial networks, safety and strategic considerations, and general telecom purposes (not laid for electronic trading between 1989 and early 2000s).

### IV. Basic results
- Estimation approach:
  - Equation (1) estimated by pooled OLS; standard errors robust to heteroskedasticity and clustered by currency.
  - Table 1: full sample columns (1)–(3); G10 in (4)–(6); non-G10 in (7)–(9). Column (1)/(4)/(7) include all observations; (2)/(5)/(8) restricted to news days; (3)/(6)/(9) include time and currency-fixed effects.
- Key findings (replication and magnitudes):
  - Positive U.S. news on policy interest rates, unemployment and trade balance leads to appreciation of the dollar against local currencies.
  - A one-standard deviation positive news shock to U.S. policy interest rates leads on average to a 0.07 percent appreciation of the dollar against local currencies (consistent with Andersen et al.).
  - G10: news suggesting more vigorous U.S. economic activity (higher industrial production and lower unemployment) is good for the dollar.
  - Some news types (e.g., GDP, industrial production) less significant in full sample; CPI news negative in full-sample estimates.
- Technology interaction results (Table 2):
  - Currencies of countries connected to EBS and Reuters matching servers by fiber-optic cables react less to U.S. economic news than currencies not connected.
  - Example magnitude: a one-standard deviation positive news shock to U.S. policy interest rates leads to almost 0.07 percent appreciation of the dollar against currencies of non-G10 countries not connected by cable, versus about 0.04 percent against currencies of countries that are connected — "about 50 percent smaller."
  - Coefficients: Cables × policy interest rate = -0.018 (All), 0.010 (G10), -0.035** (Non-G10) as reported in Table 2; unemployment interaction Cables × unemployment rate = -0.034** (All), -0.053** (G10), -0.036** (Non-G10).
- Interpretation:
  - Attenuation of exchange rate response with cable connectivity consistent with "thick-skinned" view: ICT levels the informational playing field, easing access to information beyond public price signals and limiting trend-following behavior.
  - Robust to controls for fixed and time effects.

### V. Extended specification and robustness checks
- Trading location heterogeneity:
  - Recognize onshore vs. offshore trading differences: traders in large financial centers (New York, London, Tokyo) may already have information advantages.
  - Equation (3): augment Equation (2) with Offshore_{j,t} dummy (1 if share of trading occurring offshore for currency j above full sample median) and triple interactions: r_{j,t} = α_j + ∑ β_k S_t^k + γ_c Cables_{j,t} + γ_o Offshore_{j,t} + ∑ δ_k S_t^k × Cables_{j,t} + ∑ μ_k S_t^k × Offshore_{j,t} + ∑ ρ_k S_t^k × Cables_{j,t} × Offshore_{j,t} + ε_{j,t}.
  - BIS data on location of foreign exchange trading used; data for 55 currencies in seven years (1995, 1998, 2001, 2004, 2007, 2010, 2013); triennial central bank surveys; turnover allocated by location of initiating sales desk; annual values linearly interpolated to daily observations.
- Main patterns (Table 3 and Figure 5):
  - Conditioning on trading location, effect of cable connections significant mainly for currencies traded mainly onshore, not those traded mainly offshore.
  - Magnitudes (full-sample column 1, Table 3):
    - Policy interest rate coefficient: 0.097*** (0.033).
    - Cables × policy interest rate = -0.078** (0.037).
    - Cables × offshore × policy interest rate = 0.080+ (0.059).
  - Predicted response (onshore currencies):
    - One-standard deviation positive news shock to U.S. policy interest rates leads to a 0.10 percent appreciation of the dollar against currencies of countries that are not connected by cable and are mainly traded onshore.
    - The same shock leads to a 0.02 percent appreciation of the dollar against currencies mainly traded onshore that are connected to fiber-optic cables — "80 percent smaller."
  - Offshore currencies: differences with and without cable connections are insignificant.
  - Results robust to excluding outliers, time and fixed effects, and excluding the U.K. and Japan.
- Other news types (Table 4):
  - Heterogeneous results across GDP, industrial production, CPI, unemployment rate, trade balance.
  - General pattern: where U.S. news has a positive effect on the exchange rate, that effect is dampened by cable connections; dampening attenuated for currencies traded mainly offshore.
  - Example coefficients (Table 4, full sample):
    - Cables × U.S. news for policy rate (G10): -0.081** (0.039).
    - Cables × U.S. news for policy rate (Non-G10): -0.027*** (0.001).
    - Cables × U.S. news for unemployment rate: -0.039+ (0.025).
- Asymmetries and market uncertainty:
  - No significant asymmetry by sign of interest rate news: estimates for negative vs. positive policy news are indistinguishable.
  - VIX conditioning (top quartile vs. rest): coefficient patterns similar; consistent with thick-skinned hypothesis.
- Controlling for local news (Table 5):
  - Include both local and U.S. monetary policy surprises.
  - Key magnitudes (full sample, Table 5 column 1):
    - Local policy interest rate coefficient = -0.122* (0.072).
    - U.S. policy interest rate coefficient = 0.097*** (0.033).
    - Cables × U.S. policy interest rate = -0.078** (0.037).
    - Cables × local policy interest rate = 0.106+ (0.081).
    - Cables × offshore × local policy interest rate = -0.164+ (0.123).
    - Cables × offshore × U.S. policy interest rate = 0.080+ (0.059).
  - Findings: ICT dampens reaction to both local and global news for onshore-traded currencies; a one-standard deviation positive local policy news shock leads to a 0.12 percent depreciation of the dollar against currencies of countries not cable connected and traded mainly onshore, versus a 0.02 percent depreciation when connected — "roughly 85 percent smaller."
  - Interpretation: foreign traders better able to assimilate and process local news once cable connections established; improved access reduces trend-following by better-informed traders.
- Robustness:
  - Results robust across subsamples: news days, exclusion of outliers, time-fixed effects, exclusion of U.K./Japan.
  - Sample size and timeframe: 240,430 daily observations for 56 bilateral exchange rates between January 1, 1997 and November 30, 2015.

### VI. Conclusion and policy implications
- Core conclusion:
  - Cable connections dampen the response of exchange rates to macroeconomic news.
  - Currencies traded mainly onshore respond less sharply to news emitted in the issuing country or in foreign countries like the United States when connected by submarine fiber-optic cables.
  - Estimated reduction in reaction of exchange rates to monetary policy news due to cable connections is "50 to 80 percent."
- Mechanism and interpretation:
  - ICT appears to level the informational playing field by easing access to information beyond public price signals, reducing trend-following behavior.
  - Large-bandwidth-high-speed internet connections enable onshore traders to access and act on analysis produced in major financial centers, making them better informed and less prone to amplify trends.
- Policy relevance:
  - Various proposals to limit volatility and the role of high-speed trading include speed bumps, taxes, and bans on algorithmic trading.
  - The results do not speak directly to merits/demerits of these proposals but "suggest that the wider and more comprehensive provision of information may have a dampening effect on volatility."

- Key sample and data facts (exactly as reported):
  - Sample: 56 bilateral exchange rates vs. the U.S. dollar.
  - Sample period: January 1, 1997 to November 30, 2015.
  - Observations: 240,430.
  - Cable data: 368 submarine cables starting in 1989.
  - Latency in glass: 5 to 5.5 microseconds per kilometer (10 to 11 millisecond roundtrip delay for 1,000 kilometers).
  - Spread Networks terrestrial cable (2010) reduced latency from 17 to 13 milliseconds.
  - By 2006, 99 percent of international communications traffic carried by submarine cables.
  - Representative coefficients:
    - Policy interest rate news effect (Table 1, full sample): 0.068*** (0.010).
    - One-standard deviation policy interest rate shock → 0.07 percent appreciation of the dollar (Andersen et al. replication).
    - Cables × policy interest rate (Table 2, Non-G10): -0.035** (0.017).
    - Cables × policy interest rate (Table 3, full sample): -0.078** (0.037).
    - Cables × offshore × policy interest rate (Table 3): 0.080+ (0.059).
    - Local policy interest rate effect (Table 5): -0.122* (0.072).
- Overall quantitative summary:
  - Cable connections reduce the reaction of exchange rates to monetary policy news by 50 to 80 percent for onshore-traded currencies; median empirical estimates and interaction coefficients reported above illustrate the magnitude and statistical significance across specifications.

*Source: https://www.imf.org/-/media/files/publications/wp/2017/wp1791.pdf*

### REFERENCES

### REFERENCES

### Works cited
- Acemoglu, Daron, Camilo García-Jimeno and James A. Robinson. 2015. “State Capacity and Development: A Network Approach,” American Economic Review 105 (8): 2364–2409.
- Acemoglu, Daron, Simon Johnson and James Robinson. 2005. “The Rise of Europe: Atlantic Trade, Institutional Change, and Economic Growth,” American Economic Review 95 (3): 546–579.
- Andersen, Torben, Tim Bollerslev, Francis Diebold, and Clara Vega. 2003. “Micro Effects of Macro Announcements: Real-Time Price Discovery in Foreign Exchange,” American Economic Review 91 (1): 38–62.
- Brennan, Michael, and Henry Cao. 1997. “International Portfolio Investment Flows,” Journal of Finance 52 (5): 1851–1880.
- Brennan, Michael, Henry Cao, Norman Strong, and Xinzhong Xu. 2005. “The Dynamics of International Equity Market Expectations,” Journal of Financial Economics 77 (2): 257–288.
- Borio, Claudio, Robert McCauley, Patrick McGuire, and Vladyslav Sushko. 2016. “Covered Interest Parity Lost: Understanding the Cross-Currency Basis,” BIS Quarterly Review (September): 45–64.
- Brogaard, Jonathan, Terrence Hendershott, and Ryan Riordan. 2014. “High-Frequency Trading and Price Discovery,” Review of Financial Studies 27 (8): 2267–2306.
- Brogaard, Jonathan, Carrion, Allen, Moyaert, Thibaut, Riordan, Ryan, Shkilko, Andriy, and Sokolov, Konstantin. 2015. “High Frequency Trading and Extreme Price Movements,” mimeo.
- Committee on the Global Financial System. 2001. The Implications of Electronic Trading in Financial Markets, Report by a working group established by the Committee on the Global Financial System of the central banks of the Group of Ten countries, Bank for International Settlements, Basel.
- Cornell, Bradford. 1983. “The Money Supply Announcements Puzzle: Review and Interpretation,” American Economic Review 73 (4): 644–657.
- Dumas, Bernard, Karen Lewis, and Emilio Osambela. 2015. “Differences of Opinion and International Equity Markets”, mimeo.
- Dobrev, Dobrislav and Ernst Schaumburg. 2016. “High-Frequency Cross-Market Trading and Market Volatility,” Liberty Street Economics, Federal Reserve Bank of New York, 17 February 2016.
- Ederington, Louis, and Jae Ha Lee. 1993. “How Markets Process Information: News Releases and Volatility,” Journal of Finance 48 (4): 1161–1191.
- Eichengreen, Barry, Romain Lafarguette, and Arnaud Mehl. 2016. “Cables, Sharks and Servers: Technology and the Geography of the Foreign Exchange Market,” NBER Working Paper 21884, National Bureau of Economic Research, Cambridge, MA.
- Engel, Charles, and Jeffrey Frankel. 1984. “Why Interest Rates React to Money Announcements: An Explanation from the Foreign Exchange Market,” Journal of Monetary Economics 13 (1): 31–39.
- Engel, Charles, and Kenneth D. West. 2005. “Exchange Rates and Fundamentals,” Journal of Political Economy, 113 (3): 485–517.
- Evans, Martin and Richard Lyons. 2008. “How is Macro News Transmitted to Exchange Rates?” Journal of Financial Economics 88 (1): 2650.
- Hardouvelis, Gikas. 1984. “Market Perceptions of Federal Reserve Policy and the Weekly Monetary Announcements,” Journal of Monetary Economics 14 (2): 225–240.
- Fama, Eugene. 1984. “Forward and Spot Exchange Rates,” Journal of Monetary Economics 14 (3): 319–338.
- Fatum, Rasmus and Barry Scholnick. 2008. “Monetary Policy News and Exchange Rate Responses: Do Only Surprises Matter?” Journal of Banking and Finance 32 (6): 1076–1086.
- Faust, Jon, John Rogers, Shing-Yi Wang, and Jonathan Wright. 2007. “The High-Frequency Response of Exchange rates and Interest Rates to Macroeconomic Announcements,” Journal of Monetary Economics 54 (4): 1051–1068.
- Giroud, Xavier. 2013. “Proximity and Investment: Evidence from Plant-Level Data,” Quarterly Journal of Economics, 128 (2): 861–915.
- Hasbrouck, Joel. 2015. “High Frequency Quoting: Short-Term Volatility in Bids and Offers,” New York University, mimeo.
- Hau, Harald. 2001. “Location Matters: An Examination of Trading Profits,” Journal of Finance 56 (5): 1959–1983.
- Ito, Takatoshi, and Vance Roley. 1987. “News from the US and Japan: Which Moves the Yen/Dollar Exchange Rate?” Journal of Monetary Economics 19 (2): 255–277.
- Kirilenko, Andrei and Andrew Lo. 2013. “Moore's Law versus Murphy's Law: Algorithmic Trading and Its Discontents,” Journal of Economic Perspectives 27 (2): 51–72.
- Kirilenko, Andrei, Albert Kyle, Mehrdad Samadi, and Tugkan Tuzun. 2015. “The Flash Crash: The Impact of High Frequency Trading on an Electronic Market,” mimeo.
- Lee, Eun Jung, Kyong Shik Eom, and Kyung Suh Park. 2013. “Microstructure-Based Manipulation: Strategic Behavior and Performance of Spoofing Traders,” Journal of Financial Markets 16 (2): 227–252.
- Lo, Andrew. 2016. “Moore Law vs. Murphy’s Law in the Financial System: Who’s Winning?” BIS Working Paper 564, Bank for International Settlements, Basel.
- Love, Ryan, and Richard Payne. 2008. “Macroeconomic News, Order Flows, and Exchange Rates,” Journal of Financial and Quantitative Analysis 43 (2): 467–488.
- Mark, Nelson. 1995. “Exchange Rates and Fundamentals: Evidence on Long-Horizon Predictability,” American Economic Review, 85 (1): 201–218.
- McCauley, Robert and Michela Scatigna. 2011. “Foreign Exchange Trading in Emerging Currencies: More Financial, More Offshore,” BIS Quarterly Review (March): 67–75.
- Miranda-Agrippino, Silvia, and Hélène Rey. 2015. “World Asset Markets and the Global Financial Cycle,” mimeo, London Business School.
- Moore, Michael and Richard Payne. 2011. “On the Sources of Private Information in Private FX Markets,” Journal of Banking and Finance 35 (5): 1250–1262.
- Morris, Stephen and Hyon Song Shin. 2002. “Social Value of Public Information”, American Economic Review 92 (5): 1521–1534.
- Obstfeld, Maurice. 2015. “Trilemmas and Trade-Offs: Living with Financial Globalization.” BIS Working Paper 480, Bank for International Settlements, Basel.
- Osambela, Emilio. 2013. “Differences of Opinion and International Foreign Exchange Markets,” mimeo.
- Osambela, Emilio. 2015. “Differences of Opinion, Endogenous Liquidity and Financial Asset Prices,” Review of Financial Studies 28 (7): 1914–1959.
- Rime, Dagfinn and Andreas Schrimpf. 2013. “The Anatomy of the FX Market through the Lens of the 2013 Triennial Survey,” BIS Quarterly Review (December): 27–43.
- Rey, Hélène. 2013. “Dilemma Not Trilemma: The Global Financial Cycle and Monetary Policy Independence,” Paper presented at Global Dimensions of Unconventional Monetary Policy Symposium, Jackson Hole, WY, August 22–24.
- Rey, Hélène. 2016. “International Channels of Transmission of Monetary Policy and the Mundellian Trilemma,” IMF Economic Review 64 (1): 6–35.
- Passari, Evgenia and Hélène Rey. 2015. “Financial Flows and the International Monetary System”, Economic Journal, 0 (584): 675-698.
- Swanson, Eric and John Williams. 2014. “Measuring the Effect of the Zero Lower Bound on Yields and Exchange Rates in the U.K. and Germany,” Journal of International Economics, Supplement 1, S2-S21.
- Taylor, John. 1993. “Discretion versus Policy Rules in Practice,” Carnegie-Rochester Conference Series on Public Policy 39 (0): 195–214.
- Wenzlhuemer, Roland. 2013. Connecting the Nineteenth-Century World: The Telegraph and Globalization, Cambridge: Cambridge University Press.

*Source: wp1791 - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp1791.pdf_