The Early Growth of the Engineering Industry in Italy’s Regions

The reconstruction of the historical national accounts for post-Unification Italy is proceeding. The national time series most recently compiled are those for the all-important engineering industry; this paper presents their regional counterparts. The engineering industry is very unevenly documented in the historical sources. Data abound for the shipbuilding and railway-vehicles industries, where public procurement, subsidies, and regulation produced a regular stream of reports; the national and regional estimates for these two sectors have been presented elsewhere. This paper presents the complementary regional estimates for the other, poorly documented components of the engineering industry, and the resulting regional aggregates. The regional engineering-industry estimates, like their national counterparts, distinguish the fabricated-metal (hardware) industry, three general-equipment industries (shipbuilding, railwayvehicles, and the residual, covering other metal vehicles, agricultural and industrial machinery, and structural components), two precision-equipment industries (precision instruments; clocks and watches), and the precious-metal-products industry. For all but the last maintenance activity is distinguished from new production. At Unification Italy’s engineering industry was dominated by the fabricated-metal industry, in essence the manufacture and maintenance of tools and other hardware by traditional smiths. The general-equipment industry then involved little more than (wood) shipbuilding; but it grew rapidly, and overtook the fabricated-metal industry early in the twentieth century. The precision-engineering and precious-metal-products industries were ever relatively small. The estimated regional value added series are collected in Table 1. Again like their national counterparts, these series are all at 1911 prices, in essence physical quantity series weighted by unit value added in 1911. This is the best that can be done with the evidence so far recovered; as has been pointed out elsewhere, with respect to the proper measures at a constant price level they suffer from two distortions that must be kept in mind. The minor one is that the use of constant prices tends to overstate industry’s growth, though less so with late-year prices used here than with earlyyear prices (which is the only kernel of truth in the widely misunderstood "Gerschenkron effect"); the major one is that as one goes back in time the weight of technologically stagnant production (maintenance, save for ships and railway vehicles) is increasingly overstated next to that of technologically progressive production (ship and railway vehicle maintenance, all new production). The regional series allocate the corresponding national totals with sector- and activityspecific regional value added shares. In general, and with the exceptions noted below, these are estimated initially for 1911, using the labor-force data in that year’s demographic census, and the (partial) data on employment and power in use, by shop size, in the contemporaneous industrial census. The regional maintenance value added (and employment) shares are then extrapolated to 1861-1913, using the sector-specific indicators recalled below. Regional new production shares are then estimated initially for the further benchmark years 1871, 1881, and 1900, using the census labor force data net of estimated employment in maintenance (and assuming that the differences in value added per new-production worker calculated in 1911 were less marked in earlier years); in the other years, finally, the estimated shares of new production are obtained by simple (linear) extrapolation or interpolation.