1
2020 EXAMINATIONS

MSc/MEcon/PhD (integrated) in Economics
ECONOMICS
ECON511 – Public Economics and Political Economy (2 hours)
Candidates should answer one question from Part A and one from Part B. Each question
carries equal weight.





















(Please turn over)
2
Part A
1. In 2014, the UK government published a study estimating the own-price elasticity of
different alcoholic drinks. The table below summarizes the results. The category “on”
refers to alcohol sold to be consumed on the premises (pubs, bars, restaurants), “off”
refers to alcohol sold be consumed off the premises (shops) and “RTD” refers to “ready-
to-drink” (for instance a pre-mixed gin and tonic).
On Off
Beer -0.34 -0.74
Spirits -1.25 -0.45
RTDs -0.24 -0.52
Cider -0.49 -0.74
Wine -0.24 -0.08
Source: Sousa, 2014, HRMC Working Paper 16
Using the models and intuitions from the literature on optimal commodity taxation, how would
you use the information in the above table to choose the optimal indirect tax level for the
different categories of drinks listed?

Good answers should contain at least the following information:
? Inverse elasticity rule should tell us to tax the most the more inelastic goods (wine, in
particular off-premises)
? Main assumption however: no cross elasticity. Clearly not the case here (substitution
between alcoholic drinks and between on and off premises). We would need to know
the elasticity of a broader category “alcoholic drinks”
? Many-person Ramsey rule also tell us to treat differently the goods depending on the
income of those who consume it. If alcohol is a higher share of the budget of the poor,
it should lead to lower tax. If Wine is consumed by people with different income than
Beer, they should not be taxed at the same level
? It would be useful to provide some formalization either of the many-person Ramsey
rule or of the inverse elasticity rule, clearly stating the assumptions
Then, some additional points and connections with the different topics of the lecture can
be made (there are certainly more than the one I list below):
? Substitution with labour? Perhaps a point can be made than drinking “On” premises is
a better way to do social networking than drinking at home…
? Externalities? On a pure health basis, alcohol is unhealthy and the NHS is free at the
point of use, so taxation should be used to internalize the public health cost. In terms of
other externalities (noise, anti-social behaviour), there might be a distinction between
on and off-premises drinking
3
? Paternalistic taxation? One could consider that alcohol consumption presents the same
problems when consumers use quasi-hyperbolic discounting than cigarette, so the result
of Gruber and Koszegi should apply
? When discussing redistribution, it could be useful to provide some conditions
(Atkinson-Stiglitz) such that redistribution should not be done through commodity
taxation, but through income taxes.

2. The government of your country wants to develop a very large infrastructure project to
increase accessibility in the North region. The prime minister argues that the central
government should entirely fund the project and take the big decision on how to allocate
the money. The (fictional) opposition group “better for the North” has just released the
following statement:
“Local priorities: the budget allocated to the project should be directly transferred to local
authorities who would then decide how to use it, instead of bureaucrats in the capital city
with no knowledge of local priorities.”
The (fictional) “alliance for lower taxes” has released another statement:
“No more money wasting: as a rule, we should only fund projects for which the private
sector is willing to take a share of the cost. This is the only way to identify projects that
create enough social value.”
Using the models and intuition from the literature on the provision of public goods and fiscal
federalism, critically assess the above proposals
Good answers should contain at least the following information:
Public good characteristics: it is important to define whether the project really is a public
good (non-rival, non-excludable). It can be argued that building infrastructure benefits
everyone in the region, even those who do not directly benefit from the infrastructure.
Local priorities: there is a tradeoff between local preferences and spillovers. If we are talking
about a project on accessibility and infrastructure, it is likely that a big part of the project is
largely about spillovers (should be left at the central level). However, it is true that for the
details of how the infrastructure is developed, local preferences could matter. A good way to
illustrate the tradeoff is the result in Besley-Coate (2003)
No more money wasting: private provision of public good is inefficient (this is the Samuelson
rule). However, it may be possible to design mechanisms leading to the provision of the public
good such as the Vickrey-Clarke-Grove mechanism. A convincing argument for the private
provision however would have to include such a mechanism, in which it is indeed incentive
compatible for the private sector to contribute.


4

Part B
1. Based on the material discussed in class - namely on the papers by Besley and Case (1995)
and by Finan and Ferraz (2010a) - discuss how term limits can contribute to make politicians
accountable. Briefly describe the methodology used in each paper and its main findings.
Political agency models predict that politicians should behave better when they can run for re-
election than when they cannot. However, testing this prediction empirically is challenging
since simply comparing the average behaviour of term-limited vs. non-term limited politicians
will deliver bias estimates, and since which politicians get re-elected is not random.
Besley and Case (1995) is the first paper to rigorously tackle this important empirical question.
The authors use comprehensive data on the policies implemented by governors in U.S. states
between 1950 and 2000 (namely taxes and spending), and exploit variation across states in the
presence of term limits. Using an econometric specification including both state and year fixed
effects, they estimate the coefficient of a dummy indicating whether the incumbent face a
binding term limit. They find that, on average, term-limited governors raise higher taxes –
particularly sales and income taxes – increase overall public spending, and reduce minimum
wage. The average effects mask large differences between incumbents of different party
affiliation, with the increase in taxes and spending being driven by Democrats and the decrease
in minimum wage by Republicans. These results suggest that, when not facing re-election,
politicians tend to implement policies that are ideologically more extreme and potentially
unpopular with the median voter.
Finan and Ferraz (2010) examine to what extent politicians facing re-election are less likely to
engage in corruption. Using data from random audits of municipal governments in Brazil, they
compare the relative frequency of corruption charges for mayors in their first term and mayors
in their second term. They use a cross-sectional specification and control for a wide range of
municipal characteristics (e.g., population, education, GDP) and mayor characteristics (e.g.,
education, age, gender, party). They find that audits are more likely to detect corruption for
term-limited mayors than for mayors in their first term. They show their results are robust to
comparing second-term mayors with: i) first-term mayors with political experience, and ii)
first-term mayors that will later be elected for a second term. They also show similar results
hold when comparing mayors that barely won and barely lost their second-term election in an
RDD setting. Finally, they show that the term-limit effect is weaker in areas with well-
functioning media and effective judicial prosecutor, both of which help keeping elected
officials accountable even in the absence of re-election incentives.




5
(Please turn over)
2. Based on the evidence presented in Stromberg (2007), Eisensee and Stromberg (2007) and
Durante and Zhuravskaya (2018), explain how mass media can influence government policy,
and how policy-makers may attempt to avoid media scrutiny. Discuss the methodology used
in each paper and explain why, in your opinion, their findings can inform our general
understanding of the relationship between media and accountability.
Mass media inform voters about government actions and, by doing so, can impact the policies
that are ultimately implemented. First, more informed voters are more able to punish or reward
politicians for their actions. Knowing that, politicians tend to cater policies to segments of the
populations that are more informed. Stromberg tests this hypothesis by looking at the effect of
the introduction of the radio in the U.S. in the 1930s on spending decisions by U.S. governors
in the context of a New Deal program for unemployment relief. Using data on spending and
access to radio at the county level, he shows that, controlling for a range of county
characteristics and state fixed effects, counties where a larger share of households owned a
radio, were more likely to receive funds.

Eisensee and Stromberg (2007) study, instead, whether media coverage of specific events
affects government policy. Specifically, they analyse whether natural disasters that are covered
on U.S. news media are more likely to receive relief funds from U.S. government agency, US
Aid. Estimating the causal effect of news coverage is difficult due to problems of reverse
causality and omitted variable bias. To address these issues, the authors exploit exogenous
variation in the news coverage of natural disasters due to the presence of other newsworthy
events - that are likely to crowd out news coverage of disasters – which they use as an
instrumental variable. Using this approach, they find that natural disasters that do not coincide
with other events are on average 7% more likely to make the news and to receive funds. The
effect is much larger for “marginal” disasters that would not be covered at all in the presence
of competing news.

News coverage of public policies increases public scrutiny of government actions and
contribute to keep politicians accountable. Yet, sophisticated politicians can strategically time
their unpopular policies to moment when the media and public opinion are distracted by other
events so as to minimize the political costs of their actions. Durante and Zhuravskaya (2018)
test this hypothesis for the domain of military politics by focusing on the context of the conflict
between Israelis and Palestinians. Specifically, they test whether Israeli forces are more likely
to attacks Palestinians on days when U.S. media are dominated by other events, as this would
decrease the consequences of the attacks on U.S. public opinion. Following Eisensee and
Stromberg’s approach, they use the occurrence of other newsworthy events that can crowd out
coverage of the attacks on U.S. media. Their results indicate that indeed Israeli attacks are more
likely to occur, and deadlier, on the days before important news events. The effect is driven by
attacks that carry a higher risk of civilian victims, and by news about events that are predictable.
Taken together this result indicate that i) mass media can have tangible effects on policy, ii)
that policies tend to favour voters that are more informed, iii) that informing voters about
government policies contribute to keep politicians accountable, but that iv) strategic behaviour
by politicians can undermine the role of mass media as watchdog.
(Please turn over)
6

References
Besley, T. and A. Case (1995). “Does Political Accountability Affect Economic Policy
Choices? Evidence from Gubernatorial Term Limits,” Quarterly Journal of Economics, 110(3),
769-98.
Durante R., and E. Zhuravskaya, “Attack when the World is not Watching: U.S. Media and the
Israeli-Palestinian Conflict," Journal of Political Economy, 2018.
Eisensee, T. and D. Str?mberg (2007), News Droughts, News Floods, and U. S. Disaster Relief,
Quarterly Journal of Economics, 122(2): 693-728.
Ferraz, C. and F. Finan (2010), “Electoral accountability and corruption: Evidence from the
audit if local governments,” American Economic Review, 101, 1274-1311.
Stromberg, 2004, “Radio's Impact on Public Spending”, Quarterly Journal of Economics, vol.
119, n. 1, pp. 189-221.






























7

(End of paper)