Baron Diermeier (2006) - Strategic Activism And Nonmarket Strategy
Article | |
---|---|
Has bibtex key | |
Has article title | Strategic Activism And Nonmarket Strategy |
Has author | Baron Diermeier |
Has year | 2006 |
In journal | |
In volume | |
In number | |
Has pages | |
Has publisher | |
© edegan.com, 2016 |
- This page is referenced in BPP Field Exam Papers
Reference(s)
Baron, D. and D. Diermeier (2007), Strategic Activism and Nonmarket Strategy, Journal of Economics and Management Strategy 16, 599-634. pdf (2006 Draft Paper pdf)
c.f. Dal Bó, Ernesto, Pedro Dal Bó, and Rafael di Tella (2006), "Plata o Plomo? Bribe and Punishment in a Theory of Political Influence", American Political Science Review, Vol 100, No. 1 Feb., pp. 41-53. pdf
Abstract
Activist NGOs have increasingly foregone public politics and turned to private politics to change the practices of firms and industries. This paper focuses on private politics, activist strategies, and nonmarket strategies of targets. A formal theory of an encounter between an activist organization and a target is presented to examine strategies for lessening the chance of being a target and for addressing an activist challenge once it has occurred. The encounter between the activist and the target is viewed as competition. At the heart of that competition is an activist campaign, which is represented by a demand, a promised reward if the target meets the demand, and a threat of harm if the target rejects the demand. The model incorporates target selection by the activist, proactive measures and reputation building by a potential target to reduce the likelihood of being selected as a target, fighting a campaign, and credible commitment.
The Model
There are two (types of) players:
- A firm (target) with a concave profit function [math]\pi(\cdot)\,[/math] that has negative slope (i.e. [math]\pi'(\cdot)\lt 0\,[/math]), that has a default activity [math]x_0\,[/math]
- Firms may be Strategic (and responsive to demands) with probability [math]p\,[/math], or Recalcitrant with probability [math]1-p\,[/math]
- An activist with a strictly increasing utility function [math]v(\cdot)\,[/math]
There is a campaign consisting of:
- [math]x_D\,[/math] a demand
- [math]r\,[/math] a reward
- [math]h\,[/math] a harm
The simplest model
An activist gives a firm a demand, with a promise of reward if it accepts and of harm if it rejects. The offer is TIOLI.
The firm concedes iff:
- [math]\pi(x_D) + r \ge \pi(x_0) -h\,[/math]
The activist has utility:
- [math]u(x_D,r,h) = p(v(x_D) - c(r)) + (1-p)(v(x_0) - g(h))\,[/math]
where [math]c(\cdot)\,[/math] and [math]g(\cdot)\,[/math] are both cost functions that satisfy Inada conditions.
Rearranging the firms concession function (when it binds with equality) gives:
- [math]r = \pi(x_0) - \pi(x_D) -h\,[/math]
Substituting this into the activitist's utility function and maximizing with respect to [math]x_D\,[/math] and [math]h\,[/math] give two first order conditions that the optimal campaign must satisfy, as well as [math]r^* = \pi(x_0) - \pi(x_D^*) -h^*\,[/math], providing the gain to the campaign is positive:
- [math]G = u(x_D^*,r^*,h^*) - v(x_0)\ge 0\,[/math]
which gives:
- [math]G = p(v(x_D^*) - c(r^*) - v(x_0)) - (1-p)(g(h^*)) \ge 0\,[/math]
If this condition is not satisfied then the activist doesn't conduct the campaign and target does not change practices.
Comparative statics
The campaign ([math]x_D^*,r^*,h^*\,[/math]) has the following comparative statics:
- The higher [math]p\,[/math] (i.e. more responsive the target) the more aggresive the demands [math]x_D^*\,[/math]
- The higher [math]p\,[/math] (i.e. more responsive the target) the more aggresive the harm [math]h^*\,[/math]
- The effect on rewards is ambiguous
- The reward increases when the demand goes up
- The reward decreases when the threat of harm goes up
- The demand is strictly increasing in [math]x_0\,[/math], so better current practices result in higher demands
- The optimal harm may be increasing or decreasing in [math]x_0\,[/math] depending on the functional forms of [math]\pi\,[/math] and [math]v\,[/math]
A Specific Functional Form
From here on we will assume that:
- [math]v(x)=\gamma x\,[/math] : [math]\gamma\,[/math] is the marginal valuation of the target's practices
- [math]\pi(x) = \overline\pi - \eta x\,[/math] : [math]\eta\,[/math] acts as a marginal cost of conceding
- [math]c(r) = \alpha r^2\,[/math] : When rewards are difficult to provide [math]\alpha\,[/math] is high
- [math]g(h) = \beta h^2\,[/math] : When harm is expensive [math]\beta\,[/math] is high
This form gives a demand that is linear and increasing in harm, and the optimal campaign is given by:
- [math]x_D^* = x_0 + \frac{\gamma(p\alpha + (1-p)\beta)}{2\eta^2 \alpha \beta (1-p)}\,[/math]
- [math]r^* = \frac{\gamma}{2 \eta \alpha}\,[/math]
- [math]h^* = \frac{p \gamma}{2 \eta \beta (1-p)}\,[/math]
Which has the following comparative statics:
- The demand [math]x_D^*\,[/math] are strictly increasing in [math]\gamma ,p, x_0\,[/math] (the greater the marginal benefit [math]\gamma\,[/math], the more responsive a target [math]p,\,[/math], and the better current practices [math]x_0\,[/math], the higher is demand)
- The demand [math]x_D^*\,[/math] are strictly decreasing in [math]\alpha, \beta, \eta\,[/math] (the higher costs of a campaign lead to lower demands)
- Reward is increasing in [math]\gamma\,[/math] and decreasing in [math]\alpha, \eta\,[/math]
- Harm is increasing in [math]\gamma,p\,[/math] and decreasing in [math]\eta,\beta\,[/math]
You can derive a reward to harm ratio,
- [math]\frac{r^*}{h^*} = \frac{(1-p)\beta}{p\alpha}\,[/math]
This suggests that more responsive firms will be threatened with harm over rewards, and that harm will be preferred when rewards are costly. If the internet reduces [math]\beta\,[/math], then we would expect to see campaigns with higher demands that emphasize more harm after its adoption.
Why Are Campaigns Negative?
Campaigns are most likely to be negative because;
- There is an industry effect - rewards alone would increase the profit of the firm, whereas harm decreases them. Therefore harm acts to discourage investment in the industry, which reduces the scale of the industry and hence the objectional practises
- There is an endogenous selection effect - targets are selected by the responsiveness [math]p\,[/math], and both the demand and the harm are increasing in [math]p\,[/math], whereas effect on [math]r^*\,[/math] is ambiguous
- Rewards may be costly to provide (i.e. [math]\alpha\,[/math] maybe high)
- Harm can induce proactive self-regulation (see later)
Target Selection
Targets are selected, all else equal, when:
- [math]v(x_D) - v(x_0)\,[/math] is high - the campaign gives high utility to the activist
- Likewise if [math]\gamma\,[/math] is high
- If [math]p\,[/math] is high
- If there are low costs to the campaign ([math]\eta, \alpha, \beta\,[/math])
The Market for Activists
Activists must be supported by citizens and could adopt strategies or either rewards, harm, or both.
The expected cost to only rewards (in equilibrium) are:
- [math]C^r = \frac{p \gamma^2}{4 \eta^2 \alpha}\,[/math]
The gains to rewards can likewise be calculated and are [math]G^r = C^r\,[/math]. In fact the ratio of gains to costs for any strategy are the same:
- [math]\frac{G^r}{C^r} = \frac{G^h}{C^h} = \frac{G^*}{C^*} = 1\,[/math]
Where [math]C^h\,[/math] and [math]C^*\,[/math] are the expected costs for activists that only use harm and that use both, respectively.
However, the demands are as follows:
- [math]x_D^* = x_D^r + \frac{1}{\eta}h^* = x_D^h + \frac{1}{\eta}r^*\,[/math]
So the activist that uses both harm and rewards accomplishes more.
Self-Regulation
With a Single Target
Assume that the activist can commit not to conduct a campaign once a concession is made. Such a committment could be credible if the activist had a reputation.
The activist will not conduct a campaign if:
- [math]v(\hat{x}) \ge u(x_D^*, r^*, h^*)\,[/math], or [math]\hat{x}-x_0 \ge \frac{p}{2}(x_D^*-x_0)\,[/math] in example
Likewise the target will adopt if:
- [math]\pi(\hat{x}) \ge \pi(x_0) - h^*\,[/math], or [math]\frac{h^*}{\eta} \ge \hat{x}-x_0\,[/math] in example
Putting these together (noting that [math]\hat{x} \lt x_D^*\,[/math]) [math]\hat{x}\,[/math] exists in example iff:
- [math]\frac{2-p}{1-p} \ge \frac{\beta}{\alpha}\,[/math]
Therefore if harm is emphasized over reward then pro-active measures can be observed. Note that if the activist can not commit then there is a hold up problem that prevents pro-active measures: After the firm had implemented [math]\hat{x}\,[/math] the activist would consider it the new [math]x_0\,[/math] and begin the cycle again. This is where a reputation may be necessary to provide the commitment.
With Multiple Targets
If there are no strategic interactions between the firms then a firm has an incentive to adopt a pro-active measure to shift the activist's focus elsewhere. Thus there is a multiplier effect - the activist need only target one firm to make the entire industry shift pro-actively. However, without strategic interactions the shift is identical to that in single firm example above.
With strategic interactions, a competition in pro-active measures ensues. The result is equivalent to the activist conducting a second-price auction for the opportunity to avoid a campaign. A sufficient condition (but not necessary) for this is that harm is emphasized over rewards. The race to the top leads to greater aggregate change than targeting a single firm. This is the case regardless of commitment.
Target Reputation
Suppose that targets can be either Hard or Soft, with probability [math]p_j\,[/math] of being responsive, where [math]0 \lt p_H \lt p_S \lt 1\,[/math]. The prior probability that a target is H is [math]p_0\,[/math].
The ex ante probability that a target will concede is:
- [math]p_0 = \rho_0 p_H + (1-\rho_0)p_S\,[/math]
The target can send a message [math]m_H\,[/math] or [math]m_S\,[/math] to the activist.
Let [math]\sigma_H(j)\,[/math] be the probability that type [math]j\,[/math] sends a hard message, and assume that:
- [math]\sigma_H(H) = 1\,[/math]
- [math]\sigma_H(S) \lt 1\,[/math]
The posterior probability that the target is Hard given a signal [math]m_H\,[/math] is:
- [math]\rho(m_H) = \frac{\rho_0}{\rho_0 + (1-\rho_0)\sigma_H(S)'}\,[/math]
and likewise:
- [math]\rho(m_S) = 0\,[/math]
This gives us the activist's belief that a target will concede:
- [math]p(m_S) = p_S\,[/math]
- [math]p(m_H) = \rho(m_H)p_H + (1-\rho(m_H))p_S \in [p_H, p_S]\,[/math]
This results in the activist pursuing a campaign that:
- Is more aggressive when [math]m_S\,[/math] is received
- Is less aggressive when [math]m_H\,[/math] is received
This leads to a signalling strategy by the Soft firms such that sending the Hard type message to avoid the more aggressive campaign is increasing in [math]\gamma\,[/math] and [math]p_S\,[/math], and decreasing in [math]\eta\,[/math] and [math]\beta\,[/math].
Less aggressive campaign result for Soft firms if it sends [math]m_H\,[/math] than if it sends [math]m_S\,[/math], but the campaign given [math]m_H\,[/math] is more aggressive than it would have been based on prior information.
Contesting the Campaign
Suppose that the firm can fight back with intensity [math]f \ge 0\,[/math], where [math]k(f)\,[/math] is the cost of fighting (increasing and convex), and that [math]\theta \in (0,\infty)\,[/math] is the public's support for the campaign. Then the probability of success is defined as:
- [math]q = \frac{\theta h}{\theta h + f}\,[/math]
Assume the campaign lasts for a duration 1, and the target chooses to fight or not at time 0, A fight lasts for [math]\delta \in \left[0,1\right)\,[/math].
Extension to this Sub-Model
Extensions include:
- Opportunistic Behavior - the activist makes the most aggressive demand it can if it wins, so in equilibrium, target will fight
- Commitment not to act opportunistically - the activist commits to not increase demand if it wins, so campaign will be less aggressive, thus inducing responsive target to accept and recalcitrant target to fight
Summary
The following is taken directly from the paper, but is a very useful summary of principal results:
- When its campaign is credible, an activist prefers an issue with high value and strong public support and a target that is responsive to a campaign, for which the cost of a campaign is low, and the cost of fighting is high.
- The campaign is more aggressive and more negative the weaker (more responsive) the target. For the example, the activist’s demand is more aggressive the more important is the issue, the more responsive is the target, and the lower the marginal costs of conducting the campaign.
- An activist prefers harm to reward because harm decreases investment in the targeted activity, whereas rewards alone can increase investment. Selection among potential targets leads to more negative campaigns, and harm is emphasized when rewards are costly to deliver.
- Activists that only provide rewards, only provide harm, and provide both can be present in the market for activism.
- An activist has an incentive through repetition to follow through on its campaign promises of reward and harm and for not exploiting targets that accept its demands.
- A potential target can forestall a campaign through self-regulation by changing its practices proactively but only if the activist can commit not to subsequently launch a campaign or if the proactive change shifts the activist to an alternative target. Self-regulation is plagued by a hold-up problem.
- With multiple potential targets the activist can generate a race to the top in proactive measures. This creates an incentive for an industry to act collectively.
- A potential target may develop a reputation for toughness to forestall a campaign, and the incentive to do so is strengthened by a moral hazard problem associated with revelation of its type. Conversely, a potential target that reveals itself as responsive or soft will be a more attractive target and campaigns will be more aggressive in their demands and threats. Potential targets thus have an incentive to signal that they are tough using both public and private politics strategies.
- In an infinitely-repeated game the activist can implement the optimal single-period campaign and has no incentive to shirk on the delivery of rewards and harm if its horizon is sufficiently long. For any given discount factor, however, the activist has an incentive to shirk on the delivery of harm in the optimal single period campaign if the probability of responsiveness is sufficiently high. Consequently, firms that are highly likely to be targets will not incur the single-period optimal campaign.
- If a campaign can be contested and the activist cannot commit to exploit a successful campaign, the target fights on the equilibrium path of play. If the activist can commit not to exploit a successful campaign,a responsive target concedes immediately and a recalcitrant target fights. When the cost of fighting is linear, the campaign is less aggressive when the activist can commit not to exploit a successful campaign.