1 Motivation
model, depends on investment and credit in period t-1, it is assumed not to be affected by
bank lending in the same period.2
Considering first the aggregate variables, we find that net domestic product (NDP) displays
a significant and positive reaction to a standard shock in the bank lending variable, using
both identification approaches. We find a direct effect on GDP and an additional indirect
channel via its effect on investment. This finding is consistent with most papers on economic
history (see for instance Burhop (2006) for Germany, Levine (1997), King and Levine (1993),
Rousseau and Wachtel (1998, 2000) and Schularick and Steger (2010) for other countries), as
well as a large body of literature on finance and growth in the post world war two period, in
particular in today’s emerging markets (see Beck et al. (2000) for an overview).
In the sectoral analysis, we find that all subsectors also react significantly to an unexpected
shock in aggregate lending. It is interesting, however, that the importance of these shocks
varies substantially across sectors. In a variance decomposition of the forecast errors, we
find that for the mining sector, the industrial sector and the trade sector, shocks from the
banking system only play a minor role. The agricultural sector, the transportation sector and
the service sector, on the other hand, are substantially more affected. Although our findings
confirm previous empirical studies on the aggregate impact of bank lending on growth, they
therefore challenge the conventional view on the role the banking system has actually played
in promoting growth. Our results indicate that rather than speeding up the structural change
towards the industrial sectors, the importance of the bank lending was to to allow other sectors
to keep up with its pace. In a period of rapid technological change, it seems to have allowed
for a more balanced growth path than it otherwise could have been. This result appears to
be at odds with the hypothesis that the industrial sector benefited most strongly from the
development of lending in the banking sector, but is consistent with Edwards and Ogilvie’s
view that German banking system was primarily engaged in small firm financing.
The importance of sectoral information, when analysing the effects of financial deepening
on growth, has also been emphasized in Tornell and Schneider (2004), who point out that
aggregate measures on output often mask deep sectoral asymmetries in credit constrained
economies.3 It is interesting that the sectoral patterns in 19th century Germany are indeed
2In our empirical exercise a ’period’ would be a year, as higher frequencies were unavailable for this time
period.
3See also Rajan and Zingales (1998).