# Protection for Sale (PFS) Model Analysis

Published: Last Edited:

Disclaimer: This essay has been submitted by a student. This is not an example of the work written by our professional essay writers. You can view samples of our professional work here.

Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.

The Grossman - Helpman's "Protection for Sale'' model is adopted in this paper for several reasons. First, the PFS model provides clear predictions for the cross-sectoral structure of the tariff protection. Second, the model states that the cross-sectional differences in protection can be explained by three variables: if the industry is organized or not, import penetration ratio, and import demand elasticity. Those variables introduce a background intuition for further empirical examination of the determinants of trade protection in various settings.

The model predicts as follows from Equation (\ref{Equation3}), all other things being equal: (i) protection exists for all organized industries (i.e. if $I=1$), (ii) if all citizens are members of an organized special interests group ($\alpha_{L}=1$), the political equilibrium is free trade, therefore the departure from free trade arise because SIGs can exploit non-members, (iii) protection level is high if inverse import penetration ratio ($\frac{y_{i}}{m_{i}}$) of industry $i$ is high and industry is organized (i.e. if $I=1$), (iv) protection level is high if the absolute value of the import demand elasticity $|e_{i}|$ is small.

First, I check whether predictions of the PFS model are consistent with the Russian import tariff data for 2001, 2005, 2009 as well as for 2010. Second, the model provides microeconomic foundations to the behavior of lobbies and politicians (the way it is designed). Thus, the PFS model could help to see if there are substantial systemic differences in the behavior of the government (the welfare mindedness of the government a'') in crisis and economic stability times in the determination of trade policy. Residual analysis identifies additional explanatory variables for the cross-sectional differences in protection outside the PFS setting.

Protection for sale'' model considers an economy, where prices are given endogenously. Individuals are assumed to have identical preferences. This economy produces a numeraire good with labor as input under constant returns to scale. The other $n$ non-numeraire goods use labor and specific input to a particular industrial sector. A quasi-linear utility function is assumed for an individual as shown in Equation (\ref{Equation1}). The model assumes that the politicians use only tariffs and subsidies as trade policy instruments. In the politically organized sectors, a specific factor owners can lobby the government for the trade protection of their own sectors and even lower protection for the other ones. Organized interest groups can offer political contributions which politicians value for their potential use in the coming elections. Therefore, a politician maximizes a weighted sum of total political contributions and aggregate social welfare, as she knows that her reelection depends both on money for reelection campaign and the utility level of an average voter achieved. It is a two staged non-cooperative game. At the first stage lobbies simultaneously chose their political contribution schedules, at the second - government sets trade policy. The government redistributes the revenue from this policy uniformly to all its citizens.

I borrow here from \cite{Goldberg-et-all-1999} the simpler version of the G-H model which yields the same predictions as the original model. The society consists of a continuum of individuals and those individuals have identical preferences, given by:

\begin{eqnarray}

U&=&c_0+\sum_{i=1}^nu_i(c_i) \label{Equation1}

\end{eqnarray}

where $c_{0}$ is consumption of numeraire good, $c_{i}$ is consumption of good $i$ and $u_{i}$ is an increasing concave utility function.

The government objective function is a combination of welfare of the society ($W$) and contributions by lobbyists ($C_{i}$):

\begin{eqnarray}

U^G&=&aW+(1-a)\sum_{i\in L}^nC_i \label{Equation2}

\end{eqnarray}

where $a\in[0;1]$ captures the weight of welfare, and $(1-a)$ represents the weight that government puts on contributions from lobby groups.

After calculating the equilibrium trade policy \cite{Goldberg-et-all-1999} present the simple equation of PFS model describing the determinants of trade policy:

\begin{eqnarray}

\frac{t_{i}}{1+t_{i}}&=&\frac{I_{i}-\alpha_{L}}{\frac{a}{1-a}+\alpha_{L}} * \frac{y_{i}}{m_{i}}\frac{1}{e_{i}}\label{Equation3}

\end{eqnarray}

where $t_{i}$ is an ad-valorem tariff on good $i$, $e_{i}$ is the import demand elasticity of good $i$, $\alpha_{L}$ is a fraction of population represented by a lobby, $a\in [0;1]$ captures the weight of welfare in the government policy, $I_{i}$ equals 1 if industry is organized (and 0 otherwise), $y_{i}$ represents domestic output for good $i$, and $m_{i}$ is imports from the world.

In this paper, I use industrial information on 57 industries at three digits ISIC Revision 3 code from UNIDO INDSTAT4 database. It is the most-detailed aggregation of industry data available. I have analyzed in total 266 business associations which are registered in the Chamber of Commerce and Industry of Russia (RF CCI) as well as in the Russian Union of Industrialist and Entrepreneurs (RSPP) and concluded that all industries are represented by one or more business associations and therefore all industries in my data set get the value $I=1$ at this level of aggregation. Furthermore, the oligarchic nature of the Russian economy allows to assume that the ownership of a specific factor is highly concentrated in all sectors which implies for simplicity that $\alpha_{L}=0$. Those assumptions were also used in \cite{Gawande-et-al-2009} as well as in \cite{Gawande-et-al-2012} where the authors assumed that all sectors in 54 analyzed countries were politically organized at this level of disaggregation, and the proportion of the population of a country that is represented by lobbies was negligible. Taking into account the assumptions described above the version of the PFS model for the Russian case yields the following form derived from Equation (\ref{Equation3}):

\begin{eqnarray}

\frac{t_{i}}{1+t_{i}}=\frac{(1-a)}{a}*\frac{y_{i}}{m_{i}}\frac{1}{e_{i}}\label{Equation4}

\end{eqnarray}

I use this simplified version of the PFS model for the estimation with the Russian data for 2001, 2005, 2009 and 2010.

\section{The Empirical Strategy and Data}

\label{Section:EmpiricalStrategyandData}

\subsection{The Empirical Strategy}

\label{Subsection:TheEmpiricalStrategy}

The paper by \cite{Baldwin-Evenett-2012} provides comparison of the commercial policy reactions of governments around the world in various crises, i.e. global depression of 1930s, Asian crisis of 1997 as well as the latest world economic downturn in 2008. Although the Asian crisis did not have the volume of the latest downturn and had different origins and roots, it still had the enormous effect on Russia and other economies. The falling demand on oil in Asian markets in 1997 led to falling oil price and as a result to the bankruptcy of the Russian government in August 1998. President Putin was elected in March 2000 and took the office in May 2000 right after the Russia's economic crisis (17 August 1998). The years 1999-2001 became economic recovery times blessed'' by increasing oil prices as a result of economic recovery of the Asian economies. After GDP collapse of 5.3 percent in 1998, Russia has demonstrated the economic growth of about 6 percent on average between 1999 and 2002. It is well known that this growth was led by rapid devaluation of the Russian domestic currency (ruble) in August 1998, high price of the energy exports as well as by low cost of energy inside Russia. Since 2000 the set of prudent laws has followed as a result of the crisis learning. High import prices due to domestic currency depreciation in combination with the new legislation have stimulated the development of the domestic import substituting industries and particularly the food-processing domestic industry. For the first estimation I use the import tariff schedule for the year 2001 to test the PFS model, the times right after the Russian economic crisis.

The evidence from the business surveys from 2002 onwards have shown the significant improvements in the Russian business environment in that time, which led to the relative economic stability period in 2004-2007, when a series of so called national projects'' were under consideration by the Russian government. In March 2004 Vladimir Putin was elected as President for his second term. He has served as President of the Russian Federation from May 2000 to 2008 and from May 2012 onwards. In September 2005 Vladimir Putin announced Human Capital Development Programme'' which includes a set of the national priority projects in the following areas: health system, education, housing, as well as agriculture. Most of those programmes were actively introduced from 2006 onwards, until the world economic crisis in 2008-2009. The year 2005 is the second period that I use for the empirical analysis in this paper. This is the year of the relative economic stability which became possible through the accumulation of the excessive oil income in the Russia's Stabilization Fund. In 2008 Putin took the office of Prime Minister while Dmitry Medvedev became President for 2008-2012 period. This political cooperation between Putin and Medvedev is known as Putin-Medvedev tandem'' and represents one political line of executive power over more than 12 years.

Since the global financial crisis 2008, governments around the world have implemented various measures to stimulate their economies. The measures that states have been taking were not limited to macroeconomic stimulation only. The Global Trade Alert (GTA) initiative has collected the evidence for the years 2008-2012 on government interventions around the world in the area of commercial policy and particular trade policy. As described in \cite{Aggarwal-Evenett-2012}, those interventions in crisis era have been industry specific and often discriminatory in nature. Moreover, using an extensive database of non-macroeconomic interventions (GTA database) during the crisis, the authors provided the qualitative evidence on various forms of selectivity such as promotion of certain sectors, certain firms within sectors, the selectivity against foreign commercial interest among major economic powers. Therefore, the revival of interest to industrial policy in recent years based on this trend of government selectivity is no surprise. Indeed, the government selectivity in various forms is the main feature of industrial policy, which seems to become the trend of a crisis trade policy making around the world. Russian Federation is not an exception, it also combined its anti-crisis management with the announced economic modernization programme (industrial policy) from 2008 onwards (see \cite{Gerasimenko-2012}).

The evidence presented above leads to two competing hypotheses on the welfare mindedness of the Russian Government in 2001-2010. Thus, the common political trend in the form of Putin-Medvedev tandem'' from 2000 onwards leads the hypothesis $H_{0}$.

\emph{Hypothesis $H_{0}$: there is no change in government welfare mindedness (parameter a'' ) of the Russian government in 2001-2010 during Putin-Medvedev tandem'' despite the difference in economic performance of the country during this time period.}

However, the government selectivity in anti-crisis policies shown in the GTA database might lead to an alternative hypothesis $H_{1}$.

\emph{Hypothesis $H_{1}$: there is a change in a'' in a crisis (including post crisis recovery) vs. economic stability times. The intuition here comes from the observation of anti-crisis policies around the world from 2008 onwards including the Russian case (see \cite{Gerasimenko-2012}).}

The trend of this change in parameter a'' could be as follows: the government weight on the welfare of the society parameter a'' in its objective function during crisis as well as in post-crisis recovery is relatively lower than the government weight on the welfare in economic stability times. The intuition here comes from the sectors and firms selectivity observation in Global Trade Alert database over the last years. I argue that the owners of the production factors, facing financial difficulty in crisis times and the reduced demand both at home and abroad, would appeal to get support in any form available at the government disposal such as subsidies (bailouts), import tariffs, government procurement, export subsidies and others. Therefore, one could observe an increase of the weight on contributions from lobbies in crisis relative to the economic stability times in the government objective function .

The empirical part is based on Equation (\ref{Equation4}). However, its estimation brings about two technical problems. The first one is the endogeneity of the import demand elasticity variable $e_{i}$. Therefore, this variable is moved in the estimation of Equation (\ref{Equation5}) to the left-hand side. The second issue is the potential endogeneity of the inverse import penetration ratio $\frac{y_{i}}{m_{i}}$. As \cite{Trefler-1993} showed, import tariffs have an effect on (inverse) import penetration ratio, implying that $\frac{y_{i}}{m_{i}}$ has to be treated as an endogenous variable. Therefore, I also estimate a two-stage OLS model using a set of instrumental variables for the inverse import penetration ratio $\frac{y_{i}}{m_{i}}$ to solve this endogeneity problem. The econometric model has the following form:

\begin{eqnarray}

\frac{t_{it}}{1+t_{it}}e_{i} &=& \frac{(1-a_{t})}{a_{t}}*\frac{y_{it}}{m_{it}}+\varepsilon_{it}\\

&=&\beta_{t}\frac{y_{it}}{m_{it}}+\varepsilon_{it}\label{Equation5}

\end{eqnarray}

\begin{eqnarray}

\frac{y_{it}}{m_{it}}&=&\phi_{t}Z_{it}+\epsilon_{i} \label{Equation6}

\end{eqnarray}

\begin{eqnarray}

\beta_{t}=(1-a_{t})/a_{t} \label{Equation7}

\end{eqnarray}

$i=1,...n$ and time period $t=2001, 2005, 2009, 2010$.

Vector $Z_{i}$ consists of variables that I use in the specification for the inverse import penetration ratio in Equation (\ref{Equation6}). Those variables are the number of employees, wages, and value added per industry for 2001, 2005, 2009, and 2010. Those are not the perfect instrumental variables but . It is still important to use even a non-perfect IV test to have at least two econometric settings to compare the results. The error terms $\varepsilon_{i}$ and $\epsilon_{i}$ are assumed to be distributed normally.

The Protection for Sale'' model in this form implies that $\beta_{t}>0$. I will then use those parameter estimates $\beta$ for each year to compute the implied weight of welfare in the government objective $(a)$ relative to the government weight on political contributions $(1-a)$ per year and at the second stage I will compare parameter a'' over the years of Putin-Medvedev tandem'' in crisis vs. economics stability times.

I use separate estimations per year (not polled OLS) in order to keep as much original information on import tariffs and tariff lines as possible. Each year of analysis uses a different classification of the HS code schedule (HS 1996, HS 2002, HS 2007). Thus, bringing them into one classification and estimating the model in a polled way would result in an unnecessary loss of valuable information.