1 Motivation
reminiscent of the sectoral growth patterns observed in today’s emerging markets. Tornell,
Westermann and Martinez (2003) have documented in a broad cross section of middle income
countries from 1980-2000 that there exists a pronounced shift toward small firms and non-
tradable goods producing firms in periods of rapid credit expansion. Tornell and Schneider
(2004) motivate theoretically that small firms in non-tradables goods producing sectors are
likely to benefit most from bank lending, while the tradables sectors typically consist of large
firms that have other forms of financial instruments available. In their model, the later
sectors can directly borrow from the (international) capital market and are largely unaffected
by the domestic banking system. Taking into account these characteristics of credit markets,
Ranciere and Tornell (2010) developed a two sector growth model, in which the non-tradable
sector is a bottleneck to economic growth as it is used as an input in the tradable sectors
production. Relaxing the credit constraints in the non-tradable sector therefore leads to
overall higher growth.
The empirical results in our paper seem to confirm this view. The industry, mining and
trade sector are classical tradable goods producing sectors. In particular, the industrial sector
displayed the highest export share during the late 19th and early 20th century in Germany.
Also the latter two sectors consist of mostly large firms. Transportation and services, on the
other hand, are clearly non-tradable. Although agriculture ranks among the more tradables
sectors today, it is plausible that due to the lack of modern refrigerating technologies as well
as high tariffs, its output was relatively non-tradable more than a century ago. The rapid
increase in productivity of small agricultural firms is documented in van Zanden (1991)4. Its
importance for the industrial revolution has been discussed for instance in Perkins (1981)
and Webb (1982).5 In the context of the Ranciere and Tornell model, it can be seen as an
input into the production process and the financial sector development helps to remove this
’bottleneck’ that prevents an overall higher growth path. Finally, the assumptions on credit
market imperfection in the Tornell and Schneider (2004) model are likely to be valid for our
sample period. Guinnane (2001) has argued that rural credit was a significant problem in
19th century Germany and pointed out that ”credit conditions in Germany sound similar to
4Van Zanden shows that the use of mechanical threshers, reapers or sowing machines was particularly high
in post-1870 Germany. The development of agricultural finance in the 19th century Germany has also been
documented in Blomer (1990).
5 This has also been documented for other countries. There is a consensus among economic historians that an
agricultural revolution has preceded the industrial revolution in several countries (see for instance Crafts
(1985) who documents growth in the agricultural sector in England, prior to 1820).