Cleaning The Closet: Why Do Gang Wars Last So Long?

Perhaps the length is a trade-off for low intensity.

If every single gang member got up and met at the local square, armed to the teeth with rifles & RPGs, they could expect that to provoke a massive reaction from non-gang members: the authorities will show up with paramilitary (or military) units, arrest & convict the lot, declare martial law and begin house-to-house searches, etc.

The meetup at the square would quickly resolve things in a matter of minutes, and maybe with relatively few deaths (iirc, from my anthropological readings about the Yanomano and New Guinea highlanders, tribal ‘battles’ are decided after a few injuries or deaths). But the consequences are untenable.

On the other hand, an attack every couple days for months may not provoke much of a reaction. (What *can* the authorities do about it? Where would the SWAT teams go?)

Gang violence generally has one of two purposes: To retaliate for perceived slights, or to “send a message” (usually the message being “fear me” or “I don’t fear you”).

Wars are extended samplings of both purposes. One guy does something someone in a different gang doesn’t like. The other guy “sends a message” back that he didn’t like it. The first one retaliates because he didn’t like the violence of the just received message. The second one retaliates against the retaliation.

It lasts a long time because there is no strategic objective, just a series of largely unplanned and inconclusive skirmishes that mostly serve to make both gangs angry.

Meanwhile, the leaders are relatively safe because being good at keeping themselves alive is one of the skills that makes them leaders.

“All you have to do is kill the leader of the rival gang.” That’s probably easier said than done. Just by having risen to the top of the organization, you can infer that he’s either more ruthless, intelligent or both than the typical member–a harder target. Also, he’s more insulated. The low-level soldiers are exposed on the corners, the lieutenants are shuttling between the corners, but the leader can be barricaded away. Once a leader knows a war has started, he’s first going to use whatever resources his organization has to protect himself.

The Michael Corleone example was mentioned above, but at the end of Godfather II, he’s actually VERY smart. By simultaneously taking out the heads of the other families, he prevents them from “going to the mattresses” and a war of attrition. Pulling it off, however, is easier for an author than an actual gang leader.

A blitz attack on the leader would be high-reward/high-risk, and even if successful would have longer term risks for later rounds.

If successful in Round 1 — Rival A is eliminated and attrition is avoided.
If unsuccessful in Round 1 — Attrition begins, but it will be especially severe because Rival A would be extremely unlikely to compromise or trust in a negotiated settlement.

Regardless of the success in Round 1, in Rounds 2+, the leader who opted for sniping a rival has significantly increased his risk because other gang leaders would have to assume that the same strategy would be played (on them) and so they would have strong incentive to pre-empt. In street parlance, “put the Mad Dog down.” In fact, the more often the strategy is used, the greater the incentive for other leaders to move against the player using it.

Two companies in equilibrium in an oligopoly where neither company has access to courts to enforce contracts or property rights. Presumably, they reached such an equilibrium through a long trial and error process of self monitoring and enforcement. This statistically speaking, could be viewed as constructing a strong Bayesian prior about their relationship through trial and error.

A random outlier event says the prior could be wrong. How do you test it? A whole lot of peripheral, low cost tests. Enough of these can confirm or reject the hypothesis, but it takes a long time. Probably still lower cost than killing the top rival, and constructing new relationship with second in command.

(Just for fun, think if GS and JPM had to operate with each other without recourse to enforceable contracts! Trader at JPM fails to post enough collateral, what happens? Does head of desk at GS kill trader at JPM, or does head of desk at JPM kill own trader to signal to GS that they still follow the rules? What if JPM thinks their guy is a great trader who made an honest mistake, so they won’t kill him, and won’t let GS kill him? GS cannot roll over on this breach. GS at least takes some shots at the protected guy…you can see how this goes on forever, but no one wants to kill Dimon and Blankfein, at least for a long time.)

Hi, nice post! I really like reading it.
Keep the good work!