Information Transmission between Financial Markets in Chicago and New York

• Author: Gregory Laughlin,Joseph Grundfest,Anthony Aguirre
• DOI: http://doi.org/10.1111/fire.12036
• Published date: 01 May 2014
• Date: 01 May 2014

The Financial Review 49 (2014) 283–312

Information Transmission between

Financial Markets in Chicago and

New York

Gregory Laughlin∗

University of California at Santa Cruz

Anthony Aguirre

University of California at Santa Cruz

Joseph Grundfest

Stanford University

Abstract

High-frequency trading has led to widespread efforts to reduce information propaga-

tion delays between physically distant exchanges. Using relativistically correct millisecond-

resolution tick data, we document a three millisecond decrease in one-way communication

time between the Chicago and New Yorkareas that occurred from April 27, 2010 to August 17,

2012. We attribute the first segmentof this decline to the introduction of a latency-optimized

fiber optic connection in late 2010. A second phase of latency decrease can be attributed

to line-of-sight microwave networks, operating primarily in the 6–11 GHz region of the

∗Corresponding author: University of California at Santa Cruz, Astronomy and Astrophysics, UCO/Lick

Observatory UCSC, Santa Cruz, CA 95062; Phone: (831) 459-3208; Fax: (831) 426-3115; E-mail:

laughlin@ucolick.org or glaughli@usc.edu.

Weare grateful to Darrell Duffie, Graham Giller, Neil Miller, Philip Nutzman, Matt Simsic, Max Tegmark,

and Jaipal Tuttle for useful discussions and constructive comments on the manuscript. We thank David

Levine for researching microwavetower availability rates. We thank Zachary Maggin of NASDAQOMX

for assistance with obtaining the historical ITCH4.1 tick data described herein on an academically fee-

waived basis. We thank Matt Frego of the CME for processing our Data Mine Academic order for

historical data, as well as for our ongoing standard Data Mine orders. We thank the Rock Center for

Corporate Governance at Stanford University for funding and research support.

C2014 The Eastern Finance Association 283

284 G. Laughlin et al./The Financial Review 49 (2014) 283–312

spectrum, licensed during 2011 and 2012. Using publicly available information, we estimate

these networks’ latencies, costs, and bandwidths.

Keywords: market behavior, market pricing, market structure

JEL Classification:D4

1. Introduction

On September 1, 1949, AT&T augmented its “Long Lines” telephone service

between New York City and Chicago with a 34-hop line-of-sight 4 GHz microwave

radio relay. The 838-mile route traversed six states, and carried voice and television

traffic between endpoint terminals located on AT&T’s New York City central office

building, and the Congress Street Telephone Building in Chicago. The New York

to Chicago segment was part of a larger 106-hop transcontinental microwave relay

that linked Chicago to Denver and San Francisco, and which cost $40 million to

construct (LaFrance, 2011).

Historically, new technologies, such as the principles of radio communication

underlying AT&T’s network, have rapidly been adopted for the advantage of partic-

ipants in financial markets. From 1847 through 1851, Paul Reuter employed carrier

pigeons between Brussels and Aachen to bridge a gap in telegraph stations on the

route connecting Berlin to Paris, thereby providing a low-latencydata feed for market-

moving events (Paul Reuter, 2012). In 1865, the financier James Fisk completed a

purpose-built telegraph line to Halifax in order to signal a fast steamship to cross

the Atlantic with instructions to short Confederate bonds at the close of the U.S.

Civil War(Swanberg, 1959). One of Thomas A. Edison’s best known inventions was

the stock ticker, which transmitted market pricing information using the then newly

practical transmission of information by telegraph.

In recent years, Moore’s law has driven computerization of the financial ex-

changes, with an attendant decrease in trading latency, and a strong growth in auto-

mated, algorithmic trading. The speed of trading has entered the special relativistic

regime, in which speed-of-light-related delays are a significant factor. Algorithmic

strategies are now routinely controlled by computer servers that are physically co-

located at the exchanges. Substantial effort and expense are invested in speeding up

the trading pipeline; end-to-end latencies between order generation from strategy

servers to order fulfillment and confirmation from the exchange matching engines

are now measured in microseconds or less. Exchange data (such as the NASDAQ

ITCH4.1 tick data analyzed here) are time-stamped to the nanosecond. Indeed, if

latencies continue to decrease at the current exponential rate, it is a matter of several

decades before novel general relativistic effects associated with time dilation in the

gravitational potential will begin to assume importance. This is because clocks run

at different rates depending on their location within a gravitational field (Einstein,

1908). Near the surface of the Earth, this is manifest through a clock that is elevated