Business fluctuations in Italy, 1861–1913: The new evidence

Abstract

Band-pass filters and structural time-series models are applied to the new estimates of Italy’s domestic product from 1861 to 1913. These indicate a strong four-year cycle, derived from the agricultural sector, which curiously (and perhaps spuriously) vanishes after 30 years. Over the longer term GDP and the services reflect the long swing in industrial production, tied to the investment cycle. Agriculture seems marked instead by a further cycle of some 12–15 years, and also by a long wave related to the sector’s terms of trade.

Introduction

Italy’s national historical accounts were reconstructed just over a half-century ago by Istat (the Istituto Centrale di Statistica) and Ornello Vitali (the statistician of the “Ancona Group” organized by Giorgio Fuà). The Istat–Vitali estimates for the decades between Unification (1861) and World War I have long been considered seriously distorted; preliminary alternative estimates that combine Giovanni Federico’s constant-price value added series for agriculture and the corresponding Fenoaltea series for industry and the services point to higher and above all much steadier long-term growth.¹ This paper examines business fluctuations in post-Unification Italy in the light of these new estimates. Section 2 considers the short-term business cycle, and Section 3 the longer-term fluctuations in GDP; a brief conclusion summarizes the main empirical results, and the research agenda to which they point.

The business cycle in the new GDP series is identified through alternative standard decompositions; these yield very similar results, which are correspondingly robust. Post-Unification Italy was largely agricultural, and the short-term cycle in GDP derived overwhelmingly from the movements of agricultural production. The estimated cyclical deviations were unusually moderate over the quarter-century before the war; but this comparative stability is not clearly confirmed by the available historical data, and it may be a statistical artifact.

Over the longer term trend growth appears to have been above average in the 1880s, and even more so after the turn of the century; alternative decompositions again yield very similar results, and these fluctuations too appear to be robustly identified. The agricultural series displays a further cycle of some 12–15 years, and an even longer fluctuation apparently related to the sector’s terms of trade; but these movements are relatively mild. The estimated fluctuations of the services sector are even milder, and the major longer-term movements in GDP paralleled those in industrial output. The statistically significant turn-of-the-century acceleration of GDP growth is tied to the rising industrial production of durables; it appears to reflect a sustained upswing in the investment cycle induced by the varying supply of foreign capital, rather than a break in the deep structure of the economy.