TÉCNICAS E INSTRUMENTOS PARA LA RECOLECCIÓN DE DATOS

The voter-oriented government maximizes its chance of winning the elections. We assume that there are two candidates running for office. We use a model of pre-election politics where the candidates are able to commit fully to their announced policy platforms and the citizens vote on the basis of these an- nouncements. Ex ante, all voters are identical as they all face the same prob- abilities of being employed or unemployed. The median voter is thus equal to the representative voter. There are no veto players or special interest groups. Thus, the voters elect the candidate who promises them the highest expected income. We assume that elections happen just before period 1.

Both candidates will choose the political program, that is, privatization or restructuring with the respective choices of the privatization price or the employment level in the state-owned firm, that gives the identical voters the highest expected income. Each candidate then has a 50% chance of winning. As the preferences of the voters are strictly increasing in their expected income, both candidates in equilibrium run on the same policy platform. Thus, we do not have to consider the electoral process further. We merely assume that the voter-oriented government maximizes the expected income of the voters.20

Privatization

If the voter-oriented government privatizes the firm, the investors decide on employment as described in section 3.2.2. The government’s objective function is then given by:

VP =pPπP(LP)−wmP −aX+pPwLP +

1

1 +λ[E+aX] (3.10)

Voters receive the firm’s profits after privatization net of the privatization price. They also get their expected wage. In addition, the government redis- 20_{There are many ways to model electoral competition. Our assumption that voters vote}

on the basis of their expected income excludes the possibility that voters have heteroge- neous policy preferences because some groups suffer from a policy ex post. An inclusion of opponents of a policy would imply changed incentives both for restructuring and privatiza- tion: For example, if voters were to vote by retrospection, the unemployed could suffer from privatization policies. They might then punish the government by not reelecting it. On the other hand, when they become shareholders, workers might support governments that privatize. Biais and Perotti (2002) and Schmidt (2000) use setups of this kind. However, also our simpler setup does not preclude the choice of either policy.

3.4. THE VOTER-ORIENTED GOVERNMENT 63 tributes its entire revenue. This is captured in the last term: The government gives its endowment and its privatization revenue to the voters in order to maximize their payoff. Note that the unemployment benefits are a part of that amount. The redistribution process leaves the fraction ₁₊1_λ of this amount to the voters, the rest is lost. When it has decided to privatize, the voter-oriented government chooses the privatization prize aX in order to maximize VP.

Lemma 3.2 The voter-oriented government uses underpricing. It chooses the lowest possible privatization price a= 0.

Proof. After simplification, a enters VP with −₍₁₊λ_λ)aX. That is, any reduc-

tion of a increases VP. Thus, a is chosen as low as possible.

Any redistribution of government revenue entails the loss of a fraction λof the amount that reaches the citizens. These efficiency losses give the govern- ment the incentive to use underpricing to increase the revenue of its voters.21

The value of the government’s objective function under privatization thus is:

VP =pPπ(LP)−wmP +pPwLP +

1

1 +λE (3.11)

Restructuring

The voter-oriented government’s objective function under restructuring is:

VR = p(e)wL−wm+

1

(1 +λ)[E+p(e)π(L)] (3.12) Voters receive the expected wage. Moreover, under restructuring, the gov- ernment distributes its endowment and all profits from the firm to the vot- ers. Unemployment payments are part of this amount. Again, the fraction

1− 1 1+λ

of the redistributed amount is lost. When restructuring is suc- cessful, the voter-oriented government chooses employment to maximize its objective functionVR: LV =argmax[VR(L, γg)] =argmax[wL+ ₍₁₊1_λ)π(L)].

The employment level under restructuring, LV, is then given implicitly by:

∂Y(L)

∂L =γg−λw (3.13)

21_{It is not crucial for this result that privatization does not entail efficiency losses at all.}

64 CHAPTER 3. PRIVATIZATION

Compare this condition to condition 3.6: With w > w, we have that

γg − λw > γg − λw. Thus, with Y(L) concave, we get LV > LR. The

employment level chosen by the voter-oriented government is higher than the socially optimal employment level. The reason is that the voter-oriented gov- ernment uses employment as a way to distribute revenue to the voters. For the social planner, wage payments are welfare-neutral. As the wage is paid out directly by the firm, there are no redistribution losses.22 _{The voter-oriented}

government overvalues the positive aspect of wage payments: For the utility of the wage recipients, the full wage is taken into account. On the other hand, the redistribution of profits of the firm entails losses of λ. The wage costs as part of the production costs are thus discounted by λ.

In period 1, the government hires a manager who invests in reorganizing the production process. The manager’s optimal effort choice is defined by:

∂p(e)

∂e =

1

u(π(LV))

(3.14) This also uniquely gives us the success probability of a reorganization after restructuring for the voter-oriented government P rob(γg) = pV. The fixed

wage paid by the government to the manager then is w_Vm =eV −pVu(π(LV)).

Lemma 3.3 The effort the manager exerts and the probability for the low- cost state of the world γg is lower when he is employed by the voter-oriented

government than when he is employed by the social planner: eP > eR > eV

and pP > pR> pV.

Proof. The voter-oriented government chooses an employment level that is higher than the socially optimal employment level under restructuring, LV >

LR. Therefore, the voter-oriented government receives less profits of the firm:

π(LV) < π(LR). This leads to u(π(LV)) < u(π(LR)). Thus, the first order

condition for the manager’s effort choice is fulfilled by a smaller e in the case of restructuring by the voter-oriented government. As p(e) is increasing in e, we also get a lower probability of the low-cost state of the world.

22_{We assume that the redistribution losses occur only when government income is re-}

distributed. The firm’s revenue that is paid out as wages stays inside the firm and is not administrated by the state bureaucracy. If we assumed efficiency losses in wage payments also, this would lower the net wages under restructuring. We would create additional advan- tages for privatization. The assumption that wage payments are without such losses thus strengthens our point that governments can have inefficiently high incentives to privatize.

3.4. THE VOTER-ORIENTED GOVERNMENT 65 Given the choice of LV, the payoff for the voter-oriented government is:

VR =pVwLV −wVm+

1

1 +λ[E+pVπV(LV)] (3.15) Policy Choice of the Voter-Oriented Government

We can now look at the policy choice of the voter-oriented government.

Proposition 3.2 The voter-oriented government privatizes if and only if

VP > VR, i.e.,

pPπ(LP)−

1

1 +λpVπ(LV)> w(pVLV −pPLP). (3.16) Proof. Condition 3.16 is derived directly from equations 3.11 and 3.15 where, according to our assumption,wm

P =wmV.

Privatization is attractive because it leads to higher profits. Privatization at the price of zero gives the firm’s profits directly to the citizens. Because of the higher employment level, expected profits under restructuring are lower than under privatization. Moreover, as the government has the incentive to distribute the firm’s profits to the voters, it incurs the redistribution losses. This means that in the calculus of the voter-oriented government, the expected profit from restructuring pVπ(LV) is further reduced by ₁₊1_λ. For the voter-

oriented government, the advantage of restructuring is the higher employment level with respect to privatization, given by the right hand side of condition 3.16. As the voter-oriented government strives to maximize the voters’ ex- pected income, it is interested in increasing expected wage payments.

These results imply that voters value their own income more than the money the government spends, e.g., on public goods or to repay foreign debt. The social planner does not make this distinction. We can motivate this as- sumption by considering short-term versus long-term expenditures. The time- horizon of the social planner is not restricted. In contrast, the voter-oriented government maximizes the voters’ income in the short term, in spite of the redistribution losses. Legislative periods are not modelled. Yet, the voter- oriented government only cares about being reelected. It has no incentive to engage in long-term considerations or to keep revenue for later investments as this means that it might loose the next elections.23

23_{E.g., an option would be to restructure a firm and then sell it as a “cash cow” for a}

66 CHAPTER 3. PRIVATIZATION