assume that the launch of a website is uncorrelated with the respective print medium’s
unobserved characteristics and they also both use nested logit-type demand models for
circulation.
Filistrucchi (2005) studies Italian national newspapers’ launch of companion
websites and shows that print demand statistically and economically decreases once an
online outlet channel is introduced.
Kaiser (2006) also estimates overall negative effects
but shows that these vary substantially across different consumer age groups and across
time. He claims that time may have mattered since publishers may have become better
at positioning the online companion.
George (2008) also underscores the importance of readership characteristics in the
relationship between the Internet and the demand for US newspapers. She estimates
reduced-form equations for local Internet penetration and per-capita local newspaper
circulation. Like
George (2008) and Gentzkow (2007), Hong (2007) also uses consumer
survey data to estimate the effect of the Internet on media demand. His dependent var-
iable is household expenditures for different types of entertainment goods, among others
newspapers and magazines. He estimates reduced-form equations and tries to identify
causal effects by running difference-in-difference regressions, treating general growth
in Internet penetration as exogenous.
Another strand of the literature uses time-series variation to explore the mapping
between online and offline media. In earlier work,
Deleersnyder et al. (2002) test for struc-
tural breaks (the introduction of the companion website) in monthly circulation time series
of British and Dutch newspapers, observed between 1990 and 2001. They find that few
newspapers experience a drop in circulation and advertising demand due to the existence
of a companion website. The effects are, however, disperse across newspapers and eco-
nomically fairly small. More recently,
Kaiser and Kongsted (2012) run Granger causality
tests on German magazine data. They find that online companion page visits decrease total
sales. This result is driven by a decrease in kiosk sales, which is not compensated by an
increase in subscriptions. Like
Kaiser (2006), they show that the relationship between
the online companion and the printed magazine depends on reader characteristics.
Cho et al. (2014) use a cross-country data set to study how Internet adoption affects
print newspaper circulation and the survival of newspaper firms. Their data covers over
90 countries for the years 2000–2009, which encompasses the most rapid period of Inter-
net adoption, but unfortunately ends just before the dramatic slowdown in newspaper
circulation following the financial crisis. Cho et al. show that Internet adoption directly
contributes to newspaper shutdowns in a number of countries, although the Internet
appears to have little effect on the net circulation of those firms that survive.
While existing research has primarily concerned itself with the effect of the Internet
on print demand, less is known about the reverse effect, which seems surprising given the
rapid gain in the importance of online advertising.
Kaiser and Kongsted (2012) do not
find any evidence that print circulation affects page visits, which contrasts with an earlier
study of a panel of 12 Spanish newspapers by
Pauwels and Dans (2001), which finds evi-
dence that print circulation increases website visits.
435Newspapers and Magazines