The Road to Re-Election: Incumbency, Wear and Tear of Power and New Tactics

Gaël Lagadec

University of New Caledonia (LARJE)/CRESICA, Nouméa, New Caledonia.

DOI: 10.4236/tel.[phone]                

Abstract

Incumbency is frequently said to advantage re-election seeker incumbents, but incumbency disadvantage can also be highlighted in some countries in several kinds of electoral contexts. This paper takes a theoretical point of view and aims at shedding light on the capacity to remain in power and on the mechanisms at stake. In an electoral competition framework with election seeker politicians, contributing lobbies and voters divided into informed and non-informed voters, this paper analyzes the dynamics of political support in a two-period model. It is shown that the dynamics of electoral promises with two groups competing for political favor generates a mechanical downward trend of political support. This trend creates an incentive not to campaign on re-election about a balance sheet of the ending mandate but rather to campaign on new projects or new themes not connected to those of the previous election campaign or even to bring new tactics onto the political scene (including the promotion of voters’ misinformation).

Keywords

Elections, Lobbies, Voters, Electoral Promises, Political Support

1. Introduction

Advantages of incumbency have been studied for a long time. Many authors have identified the asset of incumbency at a local stage. See for example Stratmann (1995) who noticed that for congressional voting incumbents receive in average higher contributions than their opponents. See also de Benedictis-Kessner (2018) who have shown, using data on nearly 10,000 mayoral elections in cities over the past 60 years, that incumbency carries a substantial advantage for individual candidates.

But, except for the United States, for national elections, it’s more difficult to discern such a link.

In developing democracies, an incumbency disadvantage can be demonstrated in two ways: 1) in a first stage after the transition to democracy it appears a general anti-incumbency effect, unrelated to economic conditions (Bochsler & Hänni, 2019); 2) especially when the pool of candidates increases from one election to the next (Eggers & Spirling, 2014).

For Western democracies, one can observe the frequent difficulty of re-election for country leaders¹. Interestingly in France, where presidential election is a rare case of universal single-member ballot, re-election seems to have been an impossible mission for twenty years. Two French presidents managed to be re-elected, François Mitterrand in 1988 and Jacques Chirac in 2002, but both had lost the legislative election two years before, which had led twice to a situation of power sharing (a situation named “cohabitation” in France). Emmanuel Macron has also been recently re-elected, and became the first re-elected French president outside of a cohabitation period. This election took place in a specific context (war in Ukraine and second round of the election against a far-right candidate). Two months later M. Macron has suffered a big political loss in the legislative election and has become the first French president with no majority in the parliament after being elected.

This paper takes a theoretical point of view and aims at shedding light on the capacity to remain in power and on the mechanisms at stake.

We describe a political game, departing from the Grossman-Helpman micro-foundations model (see below in this introduction a summary of the literature connected to our concern, including Grossman and Helpman’s approach). Our model is an intertemporal model of endogenous economic policies, focusing on trade endogenous policies, taking into consideration both the side of the supply and of the demand. The model is built in an electoral competition framework considering political support² and embedding the re-election constraint in the policy-maker’s reaction function, so as to take into account the fact that lobbies’ contributions and electorate’s opinion may evolve in response to unsatisfied expectations from the incumbent and that the credibility of politicians may erode.

We do not consider a homogeneous electorate, and we integrate lobbies’ actions aiming at getting specific policies favoring their own interest. Secondly, we focus our analysis on the conditions in which politicians can disown their commitments, which explains the basic reasons for political lies as well as their consequences (re-election, time account).

One can notice that the prevalence of lies can be a convincing explanation when considering an electoral competition context in which each candidate anticipates, the others are going to lie. That is what makes Woon and Kanthak (2019) consider that “the nature of electoral processes can make dishonesty endemic to the democratic selection of leaders”.

However, Woon (2017) shows that the voters are quite capable of detecting the political lies. It even remains true in a political context with strong partisan forces exerting a powerful force on political attitudes and beliefs.

The capacity to detect political lies directly raises the issue of the power wear which is the core of the paper.

The wear of power effect is frequently observed and considered as a part of the political system see for example Baldersheim and Bàtora (2012); even for the United-States this effect can be highlighted in the sense that when the presidency has been held by a party for some time, that party tends to lose seats in Congress (Chatterjee & Eyigungor, 2017)³.

But as shown above, this effect does not automatically translate into an advantage for opponents in a re-election campaign. Obviously incumbents integrate this effect in their political and economic calculations and act to counter it. It tends to make the final balance unclear, all the more so as recent works on the sources of incumbency advantage emphasize the prominent role of the politicians’ personal characteristics, rather than a partisan effect (see for example Da Fonseca, 2017, or De Benedetto, 2019).

This paper’s value is to offer an explanation of mechanism of wear of power in a two-period electoral competition framework with contributing lobbies and voters divided into informed and non-informed voters.

Considering electoral promises play a central role, we show how these electoral campaign commitments help win an election but then tend to erode the (possibly new) incumbent’s electoral support.

Giving an explanation and detailing this power wear effect leads to discussing the means to counterbalance it and the implications for the re-election campaigns (new themes, new tactics and, possibly, positions at first sight paradoxical that have been recently observed notably in the US).

The article is structured as follows. In a second section we briefly present the main background literature. In a third section we describe the set of hypotheses, the agents’ behaviors and the political game. In a fourth section we present the micro foundations of the economic model and its interactions with the political support. In a fifth section we solve the model and emphasize the evolution of electoral support in a second step of a political cycle (re-election seeking). Section six concludes.

2. Background Literature

Following Stigler (1971) and Peltzman (1976) pioneering approaches, Grossman and Helpman (1994) have built an approach of endogenous protection, which has given rise to many articles falling within this framework.

However, major aspect of this approach is that they ignore the election process despite the lobbies’ role in the process.

Other works, following Brock and Magee (1978) make electoral competition the central aspect of the political game.

Looking at Ansolabehere et al. (2003) sheds light on the importance of taking into account the election.

Focusing on Political Action Committee (PAC) contributions they show relatively few effects on voting behavior. According to them, this could suggest that policies are not for sale. An alternative explanation is that lobbies want to affect election outcomes rather than change the position of politicians, as Ansolabehere et al. (2003) state, which shows the accuracy of an electoral competition approach.

But these two approaches appear to be complementary. Grossman and Helpman (1994: p. 834) themselves have considered that the electoral competition model was better suited to the analysis of the trade policy major shaping whereas their model was more suitable for a detailed analysis (by sectors) of this policy.

More importantly, relatively few works situated in the field of endogenous protection integrate the notion of repeated elections. Examples of approaches integrating at least partially the time aspect of endogenous policies can be found in Cairns (1989), Banks (1990), Wirl (1994), Grossman and Helpman (1998) in an intergenerational framework with competing groups, or in Aragonès et al. (2007) who demonstrate that the most accurate perspective to analyse political games is the dynamic one.

The corollary of the static analysis is that the phases of the political game are actually simultaneous. Conversely, in an intertemporal framework, nothing prevents a separation between these phases. Thus the payment by the lobby of a contribution, the announcement and then the implementation of the policy by the political party do not occur at the same time. In the same manner, as Ferejohn (1986) explored, it makes sense in a political model to separate the policy announced to the population (in the electoral campaign) and the policy implemented after the election.

So it seems necessary to introduce the fact that the game phases are linked by the players’ commitments, relying particularly on trust or credibility and also on disappointment as well as possible retaliation.

When the literature focuses on political retaliation, it is most often by modelling extreme cases as Persson and Tabellini (2000) notice. For example, Aragonès et al. (2007) make the hypothesis that when a candidate lies the electors punish her/him by never believing her future promises. Aragonès et al. (2007) consider that politicians’ interest is to keep their promises, so as to create a reputation towards voters playing trigger strategies. But this analysis takes place in an infinite-horizon model, whereas the time-horizon of politicians can end when they lose an election creating a strong incitation to lie if it facilitates winning the next election.

Banks’ (1990) analysis focuses on time aspect, which justifies the cost of the lie as the author considers that the voters can sanction the candidates for their past lies (Banks, 1990: p. 310).

Also, Callander and Wilkie (2007) develop Banks’ model (1990) by putting political lies at the core of the analysis. At that point, they notably modelize gratuitous lies as specific cases of usually costly lies.

Asako (2015) builds a political model in which a candidate who implements a policy that differs from his platform must pay a cost of betrayal, which increases with the size of the discrepancy.

Note that if the political lies appear to be a common concept in the literature, it can also be seen as a simplification. As Grossman and Helpman (1998) noticed when two groups compete for political favor the government is not capable to favor both “because the struggle over distribution allows at most one net beneficiary”. Then the concern is about whose group would be the more (less) favored.

3. Description of the Agents’ Behaviors

3.1. Set of Assumptions

We consider a three-agent political game, with one politician, one representative lobby in a representative economic sector and the population split between informed and non-informed voters. We consider two elections, occurring at the beginning of each period. Each period covers the campaign before election, the election proper and the ensuing mandate.

In this model context, we make some assumptions that can be categorized in three groups. The first group corresponds to standard assumptions in the political support literature (A1, A2 and A3). The second group’s assumptions precise the nature and extent of political support, especially the link between political support and election (A4 and A5). The third group corresponds to the agents’ behavior assumptions (from A6 to A9). We also add a technical (and possibly released) assumption (A10).

The assumptions are as follows.

First group (standard):

A1. Lobbies’ actions aim at getting specific policies favoring their own interest.

A2. We consider that the politicians want to be elected and re-elected (A2-a) but that they naturally prioritize the upcoming election (t period) and do not worry about the next electoral cycle until after the election to come (t + 1 period) (A2-b).

A3. The government exerts trade policies (creating a gap between domestic and international prices) for political support motives. Endogenous policies, that is to say specific interest influence-driven policy formation, are of general scope (see for example Soo, 2007). Endogenous trade policy, also frequently summarized as “endogenous protection”, is on the one hand a part of this field but also on the other hand its far more visible part because it is a typical area for lobbies actions which directly impact (tariffs) the consumers. A3 indicates the paper is concerned with endogenous policies through trade policies.

Second group (political support):

A4. We consider the election probability (p) to depend on the political support (PS) and that this probability tends to one when the political support (for example evaluated through polls) tends to infinity.

A5. The political support is supposed, in the Peltzman’s manner, to depend on two variables: the rent (Π) and the population’s well-being (W), to which we add an exogenous parameter $\varphi \in \left[0,1\right]$ to take in reputational effects specifically. More precisely, as we are in an electoral competition context (A4), what is at stake are the anticipated rent by the lobbies (determining their contributions) and the well-being anticipated by the population. These two groups compete for political favor.

Third group (agents’ behavior):

A6. We consider a “partisan proximity” parameter, $\beta$ , showing the candidate’s relative proximity with lobbies or consumers.

A7. A candidate can make different announcements concerning a topic that is identical for the lobby and the population (A7-a) but Political lies are costly (A7-b).

A8. Contributions evolution for the next campaign occurs in contrast with the disappointment felt by lobbies (difference between announcements and achievement).

A9. The evolution of the well-being anticipated by the population (apart from the lobby) is impacted by the variation of credibility of the candidate.

Technical assumption:

A10. General economic conditions do not change from one period to another. A10 is dedicated to facilitating the analysis of the proper dynamics of political support, by keeping at a distance possible exogenous or random economic disruption. The precise extent of this assumption is that for the representative economic sector the economic fundamentals do not change but over a long period: the level of output and the import demand elasticity are rigid in a short or medium period. A10 can be released by considering a random shock (for example two states of nature).

Status of the non-technical hypotheses.

Some of the above assumptions are very standard ones: A1, A2-a, A3, A5 and, perhaps more recently A7-b. A4 can be said of common sense. The remaining assumptions are specific to this paper: A2-b, A6, A7-a, A8, A9. It’s on the latter hypotheses which concern the agents’ preferences and behaviors that the originality of this paper is based.

3.2. The Agents and the Economic-Political Game

We consider a political game, including one politician, one representative lobby and the population, split between informed and non-informed electors (respectively represented by the proportions $\alpha$ and $\left(1-\alpha \right)$ ). Following Baron’s (1994) work informed voters can’t be influenced by the campaign expenses contrary to the non-informed voters.

Elections play a major role in the political game and determine the periods of the game. Period t + 1 starts at the end of the mandate of period t, at the beginning of the election campaign in which the incumbent will seek re-election. It means that the two periods partially overlap as the end of the first mandate also corresponds to the campaign for re-election.

From A2 and A4 we have: $p_t=\frac{P{S}_t}{1+P{S}_t}$

Reputational effects (A5) generate credibility for politicians. For a candidate deemed totally honest $\varphi =1$ and for a for a candidate deemed by no means trust worth $\varphi =0$ .

Consequently $\varphi$ and W are linked together in that well-being, in the form of well-being announced (or anticipated) in the context of an election campaign, will be weighted by $\varphi$ .

Profit and well-being depend on prices (P), as in Peltzman’s pioneering analysis: $\Pi =\Pi \left(P\right)$ and $W=W\left(P\right)$ , positively for profit, negatively for well-being. Whereas the “regulator” influences prices through tariffs, we have $P=P\left(r\right)$ , where r is the tariff, which allows Peltzman (1976) to write a political support function depending only on the price.

The price is then set through the tariff in order to maximize political support, so that the marginal gain of a tariff increase in terms of political support on the lobby side is precisely canceled out by its marginal cost on the population side.

But in our framework, in which the election occurs, the probability of election depends on promises and commitments made during the electoral campaign. Even considering a representative lobby, there is no reason to consider that it is the same policy (r)⁴ which is announced to the lobby and to the population. This is due to an information asymmetry: what is announced to the population is aimed at the greatest number and is therefore known to everyone, including the lobby, whereas the commitments in favor of the lobby, harmful in terms of global economic efficiency (see the Bhagwati’s DUPs) are more likely to be secret).

Therefore, we distinguish (A7-a) the vector of rates announced to the lobbies, $r^l$ , and the vector of prices announced to the population, $r^p$ , respectively $r_i^l$ and $r_i^p$ in the representative sector hypothesis framework we use.

4. The Model Foundations

4.1. Economic Micro-Foundations

We extend Grossman and Helpman’s standard model so as to consider that the government exerts (A3) trade policies (creating a gap between domestic and international prices) for political support motives.

We consider n economic sectors.

$P=\left(P_1,P_2,\cdots ,P_n\right)$ is the vector of domestic prices of non-standard goods. $\pi =\left({\pi }_1,{\pi }_2,\cdots ,{\pi }_n\right)$ is the vector of international prices of non-standard goods. $r=\left(r_1,r_2,\cdots ,r_n\right)$ is the vector of trade taxes (a tariff on imports or a subsidy for exports if superior to 1). So for each sector i we have: $P_i={\pi }_i{r}_i$ .

We have n interest groups (potential lobbies, corresponding to the factors owners); only one part of these groups is capable of overcoming the problem of collective action (Olson, 1966) and becoming lobbies.

The consumer’s surplus is the variable chosen to appreciate the voters’ welfare:

$S\left(P_i\right)={\displaystyle \underset{i}{\sum }U}\left(d_i\left(P_i\right)\right)-{\displaystyle \underset{i}{\sum }{P}_i{d}_i(\; P\; i\; )}$

and

$S\left(r_i{\pi }_i\right)={\displaystyle \underset{i}{\sum }U}\left(d_i\left(r_i{\pi }_i\right)\right)-{\displaystyle \underset{i}{\sum }{r}_i{\pi }_i{d}_i\left(r_i{\pi }_i\right)}$ ,

with d_i the demand in good i.

The tax policy also determines the tax revenue (R). We write that $q_i$ represents the produced domestic quantity of output for the sector i. We have the tax revenue:

$R\left(r,\pi \right)={\displaystyle \underset{i}{\sum }\left(\left(r_i-1\right){\pi }_i\left(d_i\left(r_i{\pi }_i\right)-q_i\left(r_i{\pi }_i\right)\right)\right)},$

where $d_i\left(r_i{\pi }_i\right)-q_i\left(r_i{\pi }_i\right)$ represents the demand minus the production, i.e. the imports ( $M_i$ ). A customs tariff exists when at least one factor of production specific in the industry is present. The tariff (subsidy) also determines the rent earned from this factor of production. Given ${\Pi }_i\left(P_i\right)$ the aggregated rent entailed by tariff (subsidy).

The general form of welfare (W) is the summation of these three elements to which we add the aggregate labor supply (L) as in the Grossman-Helpman standard model:

$W\left(r,\pi \right)=L+{\displaystyle \underset{i}{\sum }{\Pi }_i\left(r_i{\pi }_i\right)+R\left(r,\pi \right)+S\left(r,\pi \right)}$ (1)

From this equation population’s and lobby’s members’ well-being can be deduced. We distinguish several groups within the population (normalized to 1). In sector i, ${\chi }_i$ is the part of capital owners (interest group), ${\hat{\chi }}_i$ is the part of capital owners who have overcome the difficulties stemming from collective action and has become a “lobby”). We have ${\chi }_i>{\hat{\chi }}_i$ . The part of non-organized owners is $\left({\chi }_i-{\hat{\chi }}_i\right)$ and the population except the lobby is $\left(1-{\hat{\chi }}_i\right)$ .

$W_i^l$ is the lobby’s “welfare” and $W^p$ the welfare of the population except the lobby (see also Appendix 1).

This distinction makes it possible to calculate the tariff announced to the lobby, the tariff announced to the population (see A7-a) and the tariff implemented after the election.

4.2. Political Support

From A5, we have (for period t): $P{S}_{i,t}=f\left(\Pi \left(r_i^l\right),W\left(r_i^p\right),\varphi \right)$

The expected rent, $\Pi \left(r_i^l\right)$ , is reflected in contributions ( $C_i$ ), dedicated to influence the non-informed voters (if voters were unable to be influenced, we would be in the Black’s (1958) theoretical case with only informed voters and no lobbies).

The contribution schedule is then $C_i=C\left(\Pi \left(r_i^l\right)\right)$ , rewritten $C_i=C\left(r_i^l\right)$ , with $\frac{\text{d}{C}_i}{\text{d}{r}_i^l}>0$ , the contributions are a part of the rent diverted from an economic role (DUPs).

The contributions are anticipated from the tariff announced by the candidate to the lobby ( $r_i^l$ ).

The social well-being ( $W^p$ ) is anticipated from the tariff announced by the candidate to the population ( $r_i^p$ ).

The political support function will weigh all the more the announcements in terms of social welfare, and less the contributions as the share of informed voters ( $\alpha$ ) is high.

Then, the political support function is rewritten, having the shape of the Grossman and Helpman’s politician’s objective function (period t):

$P{S}_t=\left(1-\alpha \right)C_{i,t}+\alpha \cdot \left(\varphi \cdot W_{i,t}^p\right)$ (2)

As in Peltzman (1976) such an objective function exists for every economic sector, but our framework of a representative sector gives the function a broad generality.

The political support function includes terms with different variables, since, contrary to the approaches of Peltzman and Grossman and Helpman, we are in the context of an election campaign.

So, for each sector $MaxP{S}_t\iff Max{C}_t+Max{W}_t^p$

Remark: $\varphi$ is static in the short term, time or at least two periods being necessary to catch the reputational effects (see A5).

The tariff actually put in place after the election will lie between the one announced to the population (representative of a policy, see Trump) and the one announced to the lobby, according to a slider depending on the partisan aspects: one candidate closer to sectors threatened by imports should rather be in favor of protectionism, and conversely.

We consider (A6) a parameter representing the partisan proximity, $\beta$ , which shows the relative proximity to the lobby or to the consumers (population). (See also Trump who shows his closeness to both). It is interesting to notice (and this reinforces the accuracy of this notion) that this proximity goes both ways as the preferences of specific partisan groups of voters align with their preferred party’s position (see Grewenig et al., 2019).

We consider ${\beta }^l\in \left[0,1\right]$ the parameter regarding the proximity to the lobby. When ${\beta }^l$ tends to 1, the lobby’s interest and the politician’s positions merge. We consider ${\beta }^p\in \left[0,1\right]$ the parameter regarding the proximity to the consumers. When ${\beta }^p$ tends to 1, the consumers’ interest and the politician’s positions merge. We normalize: ${\beta }^l+{\beta }^p=1$ .

For the sake of simplicity, we won’t use the i index for the $\beta$ parameters.

5. Resolution and Interpretation

5.1. Resolution (Tariff Implemented (r* − 1))

We have: $MaxP{S}_t\iff Max\{C_{i,t};W_{i,t}^p\}$ , With:

$r_i^{\ast }-1={\beta }^l\left(r_i^l-1\right)+{\beta }^p\left(r_i^p-1\right)$ (3)

The announced tariff for the population is (see Appendix 1):

$\left(r_i^P-1\right)=\frac{q_i}{{M^{\prime }}_i\left(r_i\right)}\cdot \frac{{\hat{\chi }}_i}{\chi }\cdot \frac{1-{\chi }_i}{1-{\hat{\chi }}_i}$ (4)

$\left(r_i^P-1\right)$ is negative, which corresponds to an import subsidy or an export tax.

The announced tariff for the lobby is (see Appendix 1):

$\left(r_i^l-1\right)=\frac{q_i\left(1-\frac{1}{{\chi }_i}\right)}{{M^{\prime }}_i\left(r_i\right)}$ (5)

This tariff depends only on the basic characteristics of the lobby and its sector: sector production (positive link: See Yasar, Rejesus, Chen and Chakravorty (2011)), rigidity of the import demand (the more inflexible the import demand, (M' being low), the higher the tariff) and part of the interest group in relation to the whole population (negative impact on the tariff level, coming under the Olsonian small size condition of creation and efficiency of the lobbies).

From (3), (4) and (5), the implemented tariff is:

$\left(r_i^{*}-1\right)=q_i\frac{{\chi }_i-1}{{\hat{\chi }}_i-1}\left(\frac{{\beta }^l\left({\hat{\chi }}_i-1\right)+{\beta }^p{\hat{\chi }}_i}{{M^{\prime }}_i\left(r_i\right){\chi }_i}\right)$ (6)

To interpret this result we can start from the case of a politician equidistant between the interests of the lobby and the population.

We then have ${\beta }^l={\beta }^p=\beta$ , which leads to:

$\left(r_i^{*}-1\right)=q_i\beta \frac{{\chi }_i-1}{{\hat{\chi }}_i-1}\left(\frac{2{\hat{\chi }}_i-1}{{M^{\prime }}_i\left(r_i\right){\chi }_i}\right)$ (6')

$M^{\prime }\left(r_i\right)$ is a negative derivative (decreasing import demand function), the implemented tariff is then positive as soon as ${\hat{\chi }}_i<\frac{1}{2}$ , which is the standard case since $\chi \in \left[0,1\right]$ is already small and ${\hat{\chi }}_i<{\chi }_i$ .

This first interpretation means that for an equidistant politician (incumbent) there is no unilateral free trade, which is consistent with the observation.

Free trade is generally considered as the optimal policy for a small open economy. If free trade is associated with lump-sum transfers between individuals, then it is in the interest of every individual in an economy (Hillman, 1989: p. 1, for example), and it is a global pareto optimum.

If the collective well-being is maximized for zero protection, a positive protection is suboptimal.

International trade is actually hindered by many barriers and despite liberalization rounds made by the GATT, extended by the WTO, no country has chosen free trade (Magee, 1997: p. 542).

5.2. Re-Election and Evolution of the Political Support

Each period covers the campaign before the election, the election proper and the ensuing mandate (so the two periods partially overlap). At the next period (t + 1), the outgoing candidate’s re-election will depend on the evolution of the contributions he/she will be capable to collect as support to her candidacy as well as on the evolution of social welfare.

This indicates the evolution of political support:

$P{S}_{t+1}=\left(1-\alpha \right)\left(C_{i,t}+\text{Δ}{C}_{i,t+1}\right)+\alpha \left(\varphi +\text{Δ}\varphi \right)\left(W_{i,t}^p+\text{Δ}{W}_{i,t+1}^p\right)$ (7)

Which can be rewritten in terms of variations.

Economic fundamentals of the sector (output level and import demand elasticity) do not change (A10), meaning the anticipated social well-being doesn’t change. $W_{i,t}^p=W_{i,t+1}^p=\overline{W_i^p}$ : the social well-being anticipated by the population does not change (since from A10 $r_i^p$ is also unchanged, as is $r_i^l$ ) contrary to the credibility that the population gives to the incumbent politician).

In t + 1 the contributions have two components: an economic part, which doesn’t change ( $C_t$ , see A10), and a variable part ( $\text{Δ}{C}_{t+1}$ , or $\stackrel{\dot{}}{C}$ ) that takes in the effect of deception - psychological order mechanism.

It means that the economic determinants of the contributions and of the anticipated well-being being unchanged, the same context brings the same announcements as before for the lobby and the population.

We then rewrite: $\cdot PS=P{S}_{t+1}-P{S}_t$ , becoming:

$\cdot PS=\left(1-\alpha \right){\stackrel{\dot{}}{C}}_i+\alpha \left(\stackrel{\dot{}}{\varphi }\right)\overline{W_i^p}$ (8)

From A7-b and A8, we consider that the contributions evolution for the next campaign goes in the direction opposed to the disappointment felt by the lobby (difference between announcements and realizations).

We consider (A9) the evolution of anticipated well-being by the population to be impacted by the variation in credibility of the candidate, with so $\varphi$ varying in the opposite direction of the disappointment felt by the population (difference between announcements and realizations).

We consider two parameters $\tilde{\mu }$ and $\tilde{\nu }$ that capture the sensitivity to disappointment for the lobby and for the population. This sensitivity has an absolute dimension, $\mu$ and $\nu$ (both positive), and a partisan dimension ( $\kappa >0$ ), meaning that the disappointment is all the stronger because it comes from a supposedly politically close candidate. The absolute dimension is closed to Asako’s approach (2015) with a cost of betrayal increasing with the size of the deception and can also be seen as indicating how politically divided the society is.

We then have:

${\stackrel{\dot{}}{C}}_i=\tilde{\mu }\left(r_{i,t}^{*}-r_{i,t}^l\right)$ ;

$\stackrel{\dot{}}{\varphi }=\tilde{\nu }\left(r_{i,t}^p-r_{i,t}^{*}\right)$ ;

$\tilde{\mu }=\left(1+\kappa {\beta }^l\right)\mu$ ;

$\tilde{\nu }=\left(1+\kappa {\beta }^p\right)\nu$ ;

Which entails:

$\stackrel{\dot{}}{C}=\left(1+\kappa {\beta }^l\right)\left(r_{i,t}^{*}-r_{i,t}^l\right)\mu$ ;

$\stackrel{\dot{}}{\varphi }=\left(1+\kappa {\beta }^p\right)\left(r_{i,t}^p-r_{i,t}^{*}\right)\nu$ ;

To make the interpretation easier we consider the hypothesis of an equidistant politician ( ${\beta }^l={\beta }^p$ ).

The evolution of political support therefore becomes:

$\cdot PS=-0.25q_i\left(2+\kappa \right)\left({\chi }_i-1\right)\left(\frac{-\mu +\alpha \mu -\alpha \upsilon \bar{W}}{{M^{\prime }}_i{\chi }_i\left({\hat{\chi }}_i-1\right)}\right)$ (8')

We have $\cdot PS<0$ , meaning that a simple political game leads to a political support natural decrease in a re-election context.

Differently to the Grossman and Helpman’s analysis (1998), which focuses on redistribution and implies that at most one net gain is possible for two groups competing for political favor, promises are not binding (like a divide-the-cake approach) which, unless there is a positive exogenous shock, leads to unsatisfied expectations.

Proposition 1: Whatever the policy implemented after an election it leads to a mechanical decrease in political support, creating a hindrance for re-election.

This result in non-ambiguous, and doesn’t depend on the relative sensibility to deception for voters and for the lobby. We just logically find that the decrease of the political support is all the stronger as $\kappa$ is high, meaning that the more sensitive the agents are to betrayal the more politically costly the deceptions are.

Some comparative statistics provide further information.

From the capital owners’ side we can see that when the part of organized capital owners increases the loss of political support is stronger ( $\frac{\cdot \partial PS}{\partial \hat{\chi }}<0$ ) and, at the opposite, it is lesser ( $\frac{\cdot \partial PS}{\partial \hat{\chi }}>0$ ) when the part of non-organized owners increase, which makes sense given the lobby efficiency.

From the population side we can see that the higher the part of informed voter, the greater the loss of social support ( $\frac{\cdot \partial PS}{\partial \alpha }<0$ ), which means that though these voters are able to anticipate political lies they also are committed to punish these lies.

Proposition 2: Between two elections the incumbent’s political support decreases all the more as the industry is politically effectively organized and the part of informed voters is high.

These last elements may coalesce into a Proposition 3: a fruitful campaign for re-election should avoid presenting a balance sheet but rather focus on new proposals not or hardly linked directly to economic actions.

On can also notice that globally (both sides as a whole), if the society is more politically divided, the loss of political support is stronger ( $\frac{\cdot \partial PS}{\partial \kappa }<0$ ).

Proposition 4: Between two elections the incumbent’s political support decreases all the more as the society is more politically divided.

Finally, we can notice that if we were to consider a random positive (or negative) economic shock (A10 released), it would unambiguously positively (or negatively) impact the political support, as $\frac{\partial r_i^p}{\partial q_i}<0$ and $\frac{\partial r_i^l}{\partial q_i}>0$ . This is in accordance with most studies since Stigler’s (1973) pioneering analysis of the economic determinants of re-election.

6. Conclusion

Advantages of incumbency have been studied and accepted for many decades, although they differ depending on countries and on the kind of elections. For the latest point, the presidential elections show little evidence for advantages of being in charge—France being the typical case.

As soon as a candidate makes some electoral economic commitments to (groups of) actors with antagonistic interests (i.e. the whole electorate and sectoral lobbies), following unavoidable feelings of deception mechanically tend to diminish political support even if a group of actors is clearly given more advantages than the other through the policy enforced after the election.

This result, deriving from realistic behavioral hypotheses can illustrate at least two things. First, the political support decrease can be counterbalanced by improved economic conditions or by external reputational effect: typically political advertising to enhance the perception of a politician ( $\varphi$ ) applying for re-election (see Spenkuch & Toniatti, 2018, for the effect of political advertising on votes and especially concerning the alteration of the partisan composition of the electorate).

Second and more important, without denying the fact that elections usually repeatedly cover similar ground, i.e. the Left tax and spend more versus the Right tax and spend less, this natural downward trend of the political support explains why it could frequently be difficult to make a re-election political campaign based on a balance sheet rather than a new project. The balance sheet perspective inevitably emphasizes the disappointments that a politician cannot avoid creating for actors having antagonistic interests.

It can also partially explain why for presidential election incumbency could be seen as less important in the last years.

Some authors consider that an overall erosion of political support prevails in Western democracies (Russell, 2009). Incumbents are then the first ones to have to deal with this phenomenon (the opponents are less visible and sometimes unidentifiable for the electorate). This global effect would add to more specific effects in different situations.

The situation in France allows for valuable analyzes. Tombs (2017) considers that French presidents, holders of extensively powerful offices far greater than those of an American president, are subject to immense expectations and, “all ended despised and impotent”, unable to command enough public confidence for re-election. The dynamic of promises and disappointments gives a complementary explanation to this situation. The structure of power in the French constitution prevents the dilution of the political responsibility between several political levels or institutions (including the parliament and the government), leaving the president concentrating the power but also being responsible alone for the action report and commitments not kept and this creates a huge obstacle to re-election.

This mechanical downward trend can be also interpreted as an incitation not to focus campaigning on economic topics, and even less to present an economic balance sheet, but at least to widen the campaign scope with less measurable themes such as security, migrations, identity, and so on.

Finally, one can also notice that as a corollary to the difficulties to campaign on a balance-sheet, another fruitful way to win can be to develop misinformation (e.g. Freelon & Wells, 2020), resulting in misperception of the political world and so blurring previous broken promises.

Pushing this reasoning to the extreme, that is to say abandoning any idea of being judged on a balance-sheet, can also have as corollary not even trying to convince beyond the partisan electoral base. The electoral tactic then passing through a maximization of the participation of the supporters, notably with campaign themes dedicated to galvanizing them. Such a tactic should be all the more fruitful as the global participation would be weak (Grossman & Helpman, 2022) also show, in an electoral competition context, that fake news broadcasting by parties, in particular toward their partisans, entails suboptimal policies).

As political participation shows a long-term decrease in Western democracies (see Parvin, 2018), questioning the current factors determining re-election in this particular context (described by the author as “democracy without participation”) could augur well for fruitful research works to come.

Acknowledgements

The author would like to thank Etienne Farvaque and Claude Léostic for reviewing an earlier version of this article and for their suggestions.

Appendix 1⁵

The first two are virtual, more precisely they correspond to the electoral promises toward the economic actors.

The last one is a mix of the other two. It is a combination of the antagonistic interest according to the parts of informed and non-informed voters.

The first two are then determined before the election and the last one is set after the election.

We distinguish several groups within the population (normalized to 1). In sector i, ${\chi }_i$ is the part of capital owners (interest group), ${\hat{\chi }}_i$ is the part of capital owners who have overcome the difficulties stemming from collective action (pressure group or “lobby”)⁶. We have ${\chi }_i>{\hat{\chi }}_i$ . The part of non-organized owners is $\left({\chi }_i-{\hat{\chi }}_i\right)$ and the population except the lobby is $\left(1-{\hat{\chi }}_i\right)$ .

$W_i^l$ is the lobby’s welfare and $W^p$ the welfare of the population except the lobby.

The lobby’s welfare is expressed as:

$W_i^l\left(r,\pi \right)={\hat{\chi }}_i\cdot L_i+\frac{{\hat{\chi }}_i}{{\chi }_i}\cdot {\Pi }_i\left(r_i{\pi }_i\right)+{\hat{\chi }}_i\cdot \left[R\left(r,\pi \right)+S\left(r,\pi \right)\right]$

L represents the supply of labor, $\Pi$ represents the productivity of the specific factor, R represents the tax revenue, and S represents the consumer surplus.

The first term represents the part of labor offer coming from the lobby’s members, the second term represents the part of the rent of the specific factor’s falling to active members of the interest groups (that is to say to lobby’s members), the third terms is the part of consumers surplus and customs revenue falling to the lobby’s members.

The population’s welfare (except the lobby). is expressed as:

$W^P\left(r,\pi \right)={\displaystyle \underset{i}{\sum }\left(\left(1-{\hat{\chi }}_i\right)\cdot L_i+\frac{{\chi }_i-{\hat{\chi }}_i}{{\chi }_i}\cdot {\Pi }_i\left(r_i{\pi }_i\right)\right)}+\left(1-{\hat{\chi }}_i\right)\cdot \left[R\left(r,\pi \right)+S\left(r,\pi \right)\right]$

The first whole additive term represents the part of labor offer coming from the population except lobby’s members and the part of the rent of the specific factor’s falling to the members of the interest groups (and not to lobby’s members), the second terms is the part of consumers surplus and customs revenue falling to the population.

Resolution

Optimal tariff for the population except the lobby

In the typical situation of a sector politically organized since it is in competition with imports, the function is rewritten:

$\begin{array}{c}{W}^P\left(r,\pi \right)=\left(1-{\hat{\chi }}_i\right)\cdot L_i+\frac{{\chi }_i-{\hat{\chi }}_i}{{\chi }_i}\cdot {\Pi }_i\left(r_i{\pi }_i\right)\\ +\left(1-{\hat{\chi }}_i\right)\cdot \left[\left(r_i-1\right)\cdot M_i\left(r_i\right)+S\left(r_i,{\pi }_i\right)\right]\end{array}$

with M_i the quantity of imported good i.

Remark: for the sake of simplicity from here on we won’t indicate indice i for $\chi$ and $\hat{\chi }$ anymore, which doesn’t change anything to the calculus and the results.

$\frac{\partial W^P}{\partial r_i}=\frac{\chi -\hat{\chi }}{\chi }\cdot {{\Pi }^{\prime }}_i\left(r_i{\pi }_i\right)+\left(1-\hat{\chi }\right)\cdot \left[\left(r_i-1\right){M^{\prime }}_i\left(r_i\right)+M_i\left(r_i\right)+{S^{\prime }}_i\left(r_i,{\pi }_i\right)\right]=0$

and (by Hotelling’s lemma)⁷:

$\frac{\chi -\hat{\chi }}{\chi }\cdot q_i+\left(1-\hat{\chi }\right)\left(r_i-1\right){M^{\prime }}_i\left(r_i\right)=-\left(1-\hat{\chi }\right)\left(M_i\left(r_i\right)+{S^{\prime }}_i\left(r_i,{\pi }_i\right)\right)$

entailing:

$\left(r_i-1\right)=-\frac{\left(1-\hat{\chi }\right)\left(M_i\left(r_i\right)+{S^{\prime }}_i\left(r_i,{\pi }_i\right)\right)+\frac{\chi -\hat{\chi }}{\chi }\cdot q_i}{\left(1-\hat{\chi }\right)\cdot {M^{\prime }}_i(\; r\; i\; )}$

This expression can be simplified. For one sector, the consumer surplus is:

$S_i\left(P_i\right)=U\left[d_i\left(P_i\right)\right]-P_i{d}_i(\; P\; i\; )$

The surplus is derived in relation to the price:

$\frac{\text{d}{S}_i}{\text{d}{P}_i}={d^{\prime }}_i\left(P_i\right)\left[U^{\prime }\left[d_i\left(P_i\right)\right]-P_i\right]-d_i(\; P\; i\; )$

A unit of standard good (price equal to 1). is produced with a labor unit, we have:

$U^{\prime }\left(d_i\left(P_i\right)-P_i\right)=0$ , entailing: $\frac{\text{d}{S}_i}{\text{d}{P}_i}=-d_i$

We rewrite:

$\left(r_i-1\right)=-\frac{\left(1-\hat{\chi }\right)\left(M_i\left(r_i\right)-d_i\right)+\frac{\chi -\hat{\chi }}{\chi }\cdot q_i}{\left(1-\hat{\chi }\right)\cdot {M^{\prime }}_i(\; r\; i\; )}$

The imports correspond to demand minus supply, thus we have:

$\left(r_i-1\right)=\frac{q_i\left(\hat{\chi }\left(1-\chi \right)\right)}{\chi \cdot \left(1-\hat{\chi }\right)\cdot {M^{\prime }}_i(\; r\; i\; )}$

The optimal tariff for the population is:

$\left(r_i^P-1\right)=\frac{q_i}{{M^{\prime }}_i\left(r_i\right)}\cdot \frac{\hat{\chi }}{\chi }\cdot \frac{1-\chi }{1-\hat{\chi }}$

$\left(r_i^P-1\right)$ is negative, which corresponds to an import subsidy or an export tax. In the absence of lobbies, not only will the population favor free trade (maximization of the consumer’s surplus), but it may also wish the import prices to be diminished through a subsidy.

Optimal tariff for the lobby

We have:

$W_i^l\left(r,\pi \right)=\hat{\chi }\cdot L_i+\frac{\hat{\chi }}{\chi }\cdot {\Pi }_i\left(r_i{\pi }_i\right)+\hat{\chi }\cdot \left[R\left(r,\pi \right)+S\left(r,\pi \right)\right]$

$\frac{\partial W_i^l}{\partial r_i}=\frac{\hat{\chi }}{\chi }\cdot {{\Pi }^{\prime }}_i\left(r_i{\pi }_i\right)+\hat{\chi }\cdot \left[\left(r_i-1\right){M^{\prime }}_i\left(r_i\right)+M_i\left(r_i\right)+{S^{\prime }}_i\left(r_i,{\pi }_i\right)\right]=0$

entailing:

$\left(r_i-1\right)=-\frac{\hat{\chi }\cdot \left(-q_i\right)+\frac{1}{\chi }\cdot q_i}{{M^{\prime }}_i\left(r_i\right)}$ , or: $\left(r_i^l-1\right)=\frac{q_i\left(1-\frac{1}{\chi }\right)}{{M^{\prime }}_i(\; r\; i\; )}$

NOTES

¹See here a recent analysis of multiple political alternations in Europe for the latest years: https://www.touteleurope.eu/actualite/quelles-evolutions-politiques-en-europe-ces-5-dernieres-annees.html

²Note that Ferejohn (1986) or Rogoff (1990) have built model on endogenous policies, including dynamic aspects, but where electoral support is not taken into consideration.

³Since 1856 the majority in the US House of representatives has changed 18 times, that is almost one-third of the time, but more than three-quarters of the House majority changes have occurred during a midterm. These political alternations, not connected to presidential campaigns, show the underlying trends in the erosion of political support for ruling parties. (https://history.house.gov/Institution/Majority-Changes/Majority-Changes/).

⁴In the model, the tariff is therefore an indicator of protectionism. Thus, for example, it can be considered that in the case of the 2017 US campaign, the tariff announced by Trump publicly was higher than that of Clinton, in that the candidate had a pronounced inclination for protectionist actions.

⁵For more details see also Lagadec (2014).

⁶ Helpman (1995) considers that an individual’s support of a tariff is an increasing function of the part of specific production he owns. One may conclude that it’s from the fraction (or concentration) of the factor of specific production owned that the benefit of the tariff is high enough to compensate the cost of collective action and to make the political organization of the owners of factors of specific production possible. It may be deduced in our framework that if the specific factor is unequally distributed, only those who own enough will see an interest in collective action ( $\hat{\chi }$ ).

The developments below demonstrate the calculation of three kinds of tariffs: the optimal tariff for the population (except the lobby), the optimal tariff for the lobby and the implemented actual tariff.

⁷ ${\Pi }_i\left(P_i\right)$ is the rent coming from the specific sector. With Hotelling’s lemma, we have: $\frac{\text{d}{\Pi }_i}{\text{d}{P}_i}=q_i$ ; with $q_i$ the domestic quantity produced by sector i. The slope of the profit function is so the industry supply curve.

Conflicts of Interest

The author declares no conflicts of interest regarding the publication of this paper.

References

[1]	Ansolabehere, S., de Figueiredo, J. M., & Snyder Jr, J. M. (2003). Why Is There So Little Money in US Politics. Journal of Economic Perspectives, 17, 105-130. https://doi.org/10.1257/089533003321164976
[2]	Aragonès, E., Thomas, P., & Postlewaite, A. (2007). Political Reputations and Electoral Promises. Journal of the European Economic Association, 5, 846-884. https://doi.org/10.1162/JEEA.2007.5.4.846
[3]	Asako, Y. (2015). Campaign Promises as an Imperfect Signal: How does an Extreme Candidate win against a Moderate Candidate? Journal of Theoretical Politics, 27, 613-649. https://doi.org/10.1177/0951629814559724
[4]	Baldersheim, H., & Bàtora, J. (2012). The Governance of Small States in Turbulent Times (267 p). Barbara Budrich Publishers.
[5]	Banks, J. S. (1990). A Model of Electoral Competition with Incomplete Information. Journal of Economic Theory, 50, 309-325. https://doi.org/10.1016/0022-0531(90)90005-5
[6]	Baron, D. (1994). Electoral Competition with Informed and Uninformed Voters. American Political Science Review, 88, 33-47. https://doi.org/10.2307/2944880
[7]	Benedictis-Kessner (de), J. (2018). Off-Cycle and Out of Office: Election Timing and the Incumbency Advantage. The Journal of Politics, 80, 119-132. https://doi.org/10.1086/694396
[8]	Black, D. (1958). The Theory of Committees and Elections. Cambridge University Press.
[9]	Bochsler, D., & Miriam, H. (2019). The Three Stages of the Anti-Incumbency Vote: Retrospective Economic Voting in Young and Established Democracies. European Journal of Political Research, 58, 30-55. https://doi.org/10.1111/1475-6765.12269
[10]	Brock, W., & Magee, S. (1978). The Economics of Special Interest Politics: The Case of the Tariff. American Economic Association Papers and Proceedings, 68, 246-250.
[11]	Cairns, R. D. (1989). Dynamic Rent Seeking. Journal of Public Economics, 39, 315-334. https://doi.org/10.1016/0047-2727(89)90032-7
[12]	Callander, S., & Wilkie, S. (2007). Lies, Damned Lies, and Political Campaign. Games and Economic Behavior, 60, 262-286. https://doi.org/10.1016/j.geb.2006.12.003
[13]	Chatterjee, S., & Burcu, E. (2017). Incumbency Disadvantage in U.S. National Politics: The Role of Policy Inertia and Prospective Voting (52 p.). Research Paper, Federal Reserve Bank of Philadelphia. https://doi.org/10.21799/frbp.wp.2017.43
[14]	Da Fonseca, M. L. (2017). Identifying the Source of Incumbency Advantage through a Constitutional Reform. American Journal of Political Science, 61, 657-670. https://doi.org/10.1111/ajps.12287
[15]	De Benedetto, M. A. (2019). Personal or Partisan Incumbency Advantage? Evidence from an Electoral Reform at the Local Level in Italy. The B.E. Journal of Economic Analysis & Policy, 20, 1-18. https://doi.org/10.1515/bejeap-2018-0125
[16]	Eggers, A., & Spirling, A. (2014). The Advantages and Disadvantages of Incumbency (59 p.). Harvard University.
[17]	Ferejohn, J. (1986). Incumbent Performance and Electoral Control. Public Choice, 50, 5-26. https://doi.org/10.1007/BF00124924
[18]	Freelon, D., & Wells, C. (2020). Disinformation as Political Communication. Political Communication, 37, 145-156. https://doi.org/10.1080/10584609.2020.1723755
[19]	Grewenig, E., Lergetporer, Ph., Katharina, W., & Woessmann, L. (2019). Do Party Positions Affect the Public’s Policy Preferences? CESifo Working Paper, No. 7579. https://doi.org/10.2139/ssrn.3367215
[20]	Grossman, G., & Helpman, E. (1994). Protection for Sale. American Economic Review, 84, 833-850.
[21]	Grossman, G., & Helpman, E. (1998). Intergenerational Redistribution with Short-Lived Governments. Economic Journal, 108, 1299-1329. https://doi.org/10.1111/1468-0297.00344
[22]	Grossman, G. M., & Helpman, E. (2022). Electoral Competition with Fake News. European Journal of Political Economy. https://doi.org/10.1016/j.ejpoleco.2022.102315
[23]	Helpman, E. (1995). Politics and Trade Policy. NBER Working Papers Series, No. 5309. https://doi.org/10.3386/w5309
[24]	Hillman, A. L. (1989). The Political Economy of Protection. Harwood Academic Publishers.
[25]	Lagadec, G. (2014). Optimal Endogenous Tariffs with Implicit Campaign Contributions. Theoretical Economics Letters, 4, 296-304. https://doi.org/10.4236/tel.[phone]
[26]	Magee, S. (1997). Endogenous Protection: The Empirical Evidence. In D. C. Mueller (Ed.), Perspectives on Public Choice (pp. 526-561). Cambridge University Press. https://doi.org/10.1017/CBO9780511664458.025
[27]	Olson, M. (1966). The Logic of Collective Action: Public Goods and the Theory of Groups. Harvard University Press.
[28]	Parvin, P. (2018). Democracy without Participation: A New Politics for a Disengaged Era. Res Publica, 24, 31-52. https://doi.org/10.1007/s11158-017-9382-1
[29]	Peltzman, S. (1976). Toward a More General Theory of Regulation. The Journal of Law and Economics, 19, 211-248. https://doi.org/10.1086/466865
[30]	Persson, T., & Tabellini, G. (2000). Political Economics: Explaining Economic Policy. MIT Press. https://doi.org/10.1016/S1574-0048(99)10035-1
[31]	Rogoff, K. (1990). Equilibrium Political Budget Cycles. American Economic Review, 80, 21-36.
[32]	Russell, D. (2009). Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Oxford University Press.
[33]	Soo, K. T. (2007). Endogenous Economic Policy and the Structure of Production: Theory and Evidence. Scottish Journal of Political Economy, 54, 220-253. https://doi.org/10.1111/j.1467-9485.2007.00412.x
[34]	Spenkuch, J., & Toniatti, D. (2018). Political Advertising and Election Results. The Quarterly Journal of Economics, 133, 1981-2036. https://doi.org/10.1093/qje/qjy010
[35]	Stigler, G. (1971). The Theory of Economic Regulation. Bell Journal of Economics, 2, 3-21. https://doi.org/10.2307/3003160
[36]	Stigler, G. (1973). General Economic Conditions and National Elections. American Economic Association Papers and Proceedings, 63, 160-167.
[37]	Stratmann, T. (1995). Campaign Contributions and Congressional Voting: Does the Timing of Contributions Matter? The Review of Economics and Statistics, 77, 127-136. https://doi.org/10.2307/2109998
[38]	Tombs, R. (2017). France’s Presidency Is Too Powerful to Work. Foreign Policy.
[39]	Wirl, F. (1994). The Dynamics of Lobbying—A Differential Game. Public Choice, 80, 307-323. https://doi.org/10.1007/BF01053223
[40]	Woon, J. (2017). Political Lie Detection. Working Paper, University of Pittsburgh.
[41]	Woon, J., & Kanthak, K. (2019). Elections, Ability, and Candidate Honesty. Journal of Economic Behavior and Organization, 157, 735-753. https://doi.org/10.1016/j.jebo.2018.11.003
[42]	Yasar, M., Rejesus, R. M., Chen, Y., & Chakravorty, U. (2011). Political Influence of Firms in the Tradables and Non-Tradables Sectors: A Cross-Country Analysis. Economics & Politics, 23, 297-312. https://doi.org/10.1111/j.1468-0343.2011.00386.x