A Band Spectral Analysis of Exports and Economic Growth in the United States

• Author: Jon Vilasuso,H. Sonmez Atesoglu
• Date: 01 February 1999
• DOI: http://doi.org/10.1111/1467-9396.00152
• Published date: 01 February 1999

A Band Spectral Analysis of Exports and Economic

Growth in the United States

H. Sonmez Atesoglu and Jon Vilasuso*

Abstract

This paper examines the role of exports in aggregate economic growth in the United States using band spec-

tral regression. The findings reveal a predictable relationship between long-run frequency components of

real export growth and real GDP growth over the post-Bretton Woods period of flexible exchange rates.

The study fails to uncover a significant relationship between long-run frequency components of the terms

of trade and real output. Overall, the findings support the export-led growth hypothesis, and dismiss long-

run movements in the terms of trade as an important determinant of real output growth.

1. Introduction

The purpose of this paper is to report our findings examining the empirical relation-

ship between exports and economic growth in the United States using band spectral

regression. In recent years in the United States, exports rather than domestic aggre-

gate demand is increasingly considered to be the vehicle of economic expansion and

growth. The commitment of the Federal Reserve to non-inflationary policies and the

reluctance of the federal authorities to expand government spending owing to persis-

tent federal budget deficits leaves exports as the primary source of economic expan-

sion. An additional advantage of export-led growth is the favorable effect of increased

exports on the net-foreign wealth of the United States. Whether exports can propel

the economy depends on the nature of the connection between exports and real output

growth.

The export-led growth literature stresses productivity as the key factor linking

exports and real output growth.1Increased production in the export sector,

for example, improves resource allocation and productivity which supports higher

aggregate output growth rates. A number of explanations have been offered for

the relationship between exports and productivity, including productivity spillovers,

greater economies of scale opportunities, and higher rates of technical change in

the export sector.2Although much of the empirical work examining the export-led

growth hypothesis focuses on developing countries, Marin (1992) argues that there is

no inherent reason why export-led growth should not apply to developed countries

as well.

In this paper, we employ band spectral regression techniques to determine the rela-

tionship between export and real output growth in the United States across different

frequency bands. We find that there is a significant, positive relationship between

export and real GDP growth in the United States since the passing of the Bretton

Woods system of fixed exchange rates. This relationship, however, is concentrated at

long-run movements in exports, while short-run variations in exports appear unrelated

to real output growth. We find little evidence to indicate that the terms of trade affect

Review of International Economics, 7(1), 140–152, 1999

© Blackwell Publishers Ltd 1999, 108 Cowley Road,Oxford OX4 1JF, UK and 350 Main Street, Malden,MA 02148, USA.

*Atesoglu: Clarkson University,School of Business, Potsdam,NY 13699-5785, USA. Tel:(315) 268-7981; Fax:

(315) 268-3810; E-mail:atesoglu@icarus.som.clarkson.edu.Vilasuso: Clarkson University,School of Business,

Potsdam, NY 13699-5785, USA.Tel: (315) 268-6609; E-mail: vilasusj@icarus.som.clarkson.edu.We wish to

thank an anonymous referee for helpful comments on an earlier draft. The usual caveat applies.