Market-based lobbying: Evidence from advertising spending in Italy

Stefano DellaVigna, Ruben Durante, Brian Knight, Eliana La Ferrara, 9 March 2014



The policy debate abounds with discussions on conflict of interests faced by politicians. The scope of the problem is global, in that conflicts of interest are found in industrialised as well as developing countries. And it is complex; there are multiple ways in which conflict of interest can manifest itself.
The first, perhaps the one we understand best, is that while politicians have incentives to represent the interests of their constituents, they also often face conflicting incentives with respect to special-interest groups. An extensive literature has studied this direct lobbying relationship; firms attempt to curry favour with politicians, via campaign contributions or other means, in exchange for policy favours (Grossman and Helpman, 2001; Ansolabehere et al, 2003; Bertrand et al, 2011; Querubin and Snyder, 2011; Blanes i Vidal et al., 2012).

An additional conflict of interest arises for politicians with business holdings. Politicians in this situation have an interest to distort policy to benefit the firms they have a stake in, and thus reap the private benefits in the form of profits. A significant literature has focused on this businessman-politician channel (Fisman, 2001; Cingano and Pinotti, forthcoming).

New evidence on indirect lobbying

In new research (DellaVigna et al., 2014) we consider a different channel for the conflict of interest for politicians with business holdings. Specifically, we study whether third parties attempt to curry favour with conflicted politicians by shifting their business towards firms controlled by a politician. The politician benefits financially from the increased business, and the third parties hope for favourable regulation in return. We label this channel, involving lobbying through business proxies and which has not received much attention in the literature, as 'indirect lobbying'.

Figure 1 illustrates how this channel differs from the literature. The direct lobbying channel involves firms lobbying politicians directly for regulation. The businessman-politician channel applies to the case in which politicians are stakeholders in a firm. These politicians reap the benefits of political decisions though the firm revenue. The indirect lobbying channel we propose operates through business proxies. Firms provide favours to politicians by directing business orders to the firm controlled by the politician. This channel, like the second channel, only applies when rules do not forbid the concentration of political and business interests.

Figure 1. Conflict-of-interest channels

To provide evidence on this third channel, we consider a particularly egregious case of concentration of business and political interests, i.e. Italy since the mid 1990s. In the Spring of 1994, Silvio Berlusconi, previously a successful entrepreneur and owner of Italy's main private television network, was elected prime minister. Unlike the US, Italy has no rules forbidding the concentration in one person of business interest and prominent political positions, and does not have the tradition of blind trusts for politicians with interests in companies. As such, Berlusconi retained control of his business holdings in the media, inducing a conflict of interest with his role as prime minister.

In this context, the indirect lobbying distortions take the form of advertising decisions. The Italian broadcast television is dominated by two groups, the public broadcasting corporation (RAI) and a private network, Mediaset, controlled by Berlusconi. The profitability of the three Mediaset channels, which are free-to-air, depends on advertising revenue. The indirect lobbying channel posits that firms attempt to curry favours with the government by shifting some of their advertising from public channels to Berlusconi's channels when Berlusconi is in power.

While our analysis is focused on Italy, the indirect lobbying channel identified here applies broadly, given that politicians have major business holdings in several other countries. In some cases, the business interests are in the media, as in Italy. For example, Thaksin Shinawatra, prime minister of Thailand between 2001 to 2006, owned the country's largest free-to-air television, and Sebastián Piñera, former president of Chile, owned Chile's most influential TV station. Additional examples include Andrej Babis, leader of Czech Republic's second largest party and owner of multiple newspapers and two national TV channels and, in the US, Michael R. Bloomberg, mayor of New York from 2002 to 2013 and main shareholder of the news conglomerate Bloomberg LP. In other cases, the holdings are outside the media sector, as in the case of Nitin Gadkari, leader of India's opposition party BJP between 2010 and 2013 and main shareholder of the Purti group, with interests in the energy, sugar, and alcohol sectors, among others.

A model of the advertising market

To illustrate the indirect lobbying channel in our context, we sketch a model of the advertising market. We consider two types of firms, regulated and unregulated, who must decide how to allocate their advertisements between the two networks. In addition to the economic benefits associated with advertising, regulated firms receive a political benefit from advertising on Berlusconi's network when he is in power. When Berlusconi comes to power, demand for advertising on his network thus increases. This shift in demand induces an increase in the price of advertising in Berlusconi's channels and also a change in the composition of advertising spending. Regulated companies shift spending towards Berlusconi's channels, while unregulated firms do otherwise (given the price change). This quid-pro-quo increases the profits of Berlusconi's companies and lowers the profits of the competing public network.

Evidence from sector-level and firm-level data

To test the predictions of the model, we use sector-level and firm-level data by Nielsen on quarterly advertising expenditure by firm and media outlet between 1993 and 2009. We then compare the advertising spending on the different TV channels when Berlusconi is in power versus when he is not. In this respect, we exploit the repeated switches in political balance. Berlusconi was prime minister in 1994, between 2001 and 2006, and from 2008 to the end of our sample. Further, to test the predictions on regulation, we conduct a survey of Italian economists eliciting measures of regulation by industry, and use the responses to construct a continuous measure of the degree of regulation in a sector.

In the first part of the analysis, we compare outcomes in the advertising market during periods in which Berlusconi is in power to periods when he is not in power. Figure 2 shows the main result. Consistent with the predictions, advertising spending on Mediaset, relative to the public network, is higher when Berlusconi is in power (indicated with the shaded areas). The result is clearly visible for both the second and third Berlusconi government (the estimates for the first government are noisier given its short duration). The result is driven by an increase in advertising prices on Mediaset and a corresponding reduction in prices on the public network. Consistent with a relatively inelastic supply of advertising slots, we find no changes in the quantity of advertisements on the two networks.

Figure 2. Share of advertising on Berlusconi’s televisions

Building upon this evidence, we conduct a difference-in-difference analysis, comparing more regulated industries to less regulated industries. Figure 3 shows graphically the main result. Consistent with the predictions of the model, we find that regulated sectors (continuous line), relative to unregulated sectors (dotted line), spend more on Mediaset, relative to the public network, when Berlusconi is in power. In contrast to the time-series evidence, which was driven by a price response, this shift is mainly driven by a quantity response, with regulated sectors, relative to unregulated sectors, purchasing more slots on Mediaset, relative to the public network, when Berlusconi is in power. This effect is stronger for the peak-hours programming, which is of higher value to the networks.

Figure 3. Share of advertising on Berlusconi’s televisions, by regulation index in the industry

Finally, we exploit a key advantage of our setting. We use the simple economics of TV advertising slots, given a fixed supply of seconds of advertising, to back out the estimated profits accruing to Berlusconi's company due to the quid pro quo. We estimate a profit increase of over €1 billion over the nine years of Berlusconi government, accounting for 20% of the market capitalisation of Mediaset in 1997. In turn, this provides a measure of the expected returns from favourable regulation for the regulated firms of €2 billion over nine years. The large magnitudes indicate the first-order role played by the indirect lobbying channel.

Our findings have important policy implications. We provide an additional rationale for rules on conflict of interest like the ones in place in the modern US congress, with a tradition of blind trusts for politicians with interests in companies. The traditional rationale for such separation is to avoid self-serving legislation (the businessman-politician channel). We point out that, in addition, the concentration of business and political interests allows for alternative forms of lobbying – through business purchases – which are harder to monitor and regulate.


This paper documents an important link between the two literatures. In the presence of businessmen-politicians, the lobbying process can take an indirect route. Firms hoping for regulatory favours may lobby the politician through business purchases towards the firm controlled by the politician, who benefits from the additional revenue. We provide evidence consistent with this channel in Italy, where we exploit the detailed advertising data, the frequent switches in power, and variation in propensity for regulation. We show that the magnitudes of this effect are very sizeable, in the order of billions of euros. Our results suggest a further rationale for rules dictating a separation between business and political interests.

While the paper has focused on a specific setting – advertising markets in Italy – we stress that the channel at hand applies to all cases in which there is a confluence between business and political decision-making. In a classical paper on Suharto (Fisman 2001), for example, the returns to firms close to the dictator surely reflect the traditional favouritism channel, but likely also capture the indirect lobbying channel highlighted in this paper. To start with, our findings are likely to be relevant in other advertising markets in countries where media outlets are owned by powerful families which, as Djankov et. al. (2003) document, is a common situation. We hope that future research will investigate more such settings.


Ansolabehere S, J M de Figueiredo, and J M Snyder (2003), "Why is There so Little Money in US Politics?", Journal of Economic Perspectives, 17(1): 105-130.

Bertrand, Marianne, Matilde Bombardini, and Francesco Trebbi. 2011. "Is it whom you know or what you know? An empirical assessment of the lobbying process." NBER working paper 16765. National Bureau of Economic Research.

Blanes i Vidal J, M Draca, and C Fons-Rosen (2012), "Revolving Door Lobbyists" The American Economic Review, 102(7): 3731-48.

Cingano F and P Pinotti (forthcoming), “Politicians at Work: the Private Returns and Social Costs of Political Connections”, Journal of the European Economic Association.

DellaVigna S, R Durante, B Knight and E La Ferrara (2014), “Market-based Lobbying: Evidence from Advertising Spending in Italy”, CEPR Discussion Paper 9813.

S Djankov, A Shleifer, T Nenova and C Mcliesh (2003), “Who Owns the Media”. Journal of Law and Economics, 46(2): 341-383.

Fisman, R (2001), "Estimating the Value of Political Connections", The American Economic Review, 91(4): 1095-1102.

Grossman, G M and E Helpman (2001), Special interest politics, The MIT Press.

Querubin, P and J Snyder (2011), "The Control of Politicians in Normal Times and Times of Crisis: Wealth Accumulation by US Congressmen, 1850-1880", Working Paper.

Topics: Politics and economics
Tags: Berlusconi, Italy, lobbying

Professor of Economics at UC Berkeley
Assistant Professor in the Department of Economics, Sciences Po
Professor of Economics, Brown University; and Research Associate, NBER
Professor of Economics, Bocconi University and CEPR Research Fellow

Most Read