This is also the conclusion of Rocco and Warglien (1995), who report a study showing that it is the communication face to face that makes the big difference (on this point, see also Frey and Bohnet, 1995; Bohnet and Frey, 1999a; Bohnet and Frey, 1999b; and Ostrom, 1998).
So far we have analyzed common-pool resource games. However, many of the arguments apply also to public goods games. In fact, public goods games and common-pool resource games are very similar. Whereas in a common-pool resource game, subjects’ decisions impose negative externalities on other subjects, subjects in a public goods game produce positive externalities. In a common-pool resource game, it is nice or kind not to appropriate too much, while in a public goods game it is kind not to contribute too little to the public good. Public goods situations are very important and very frequent in reality.18 Moreover, there exists a huge experimental literature on public goods games. As we will show, many of the findings reported on common-pool resource problems carry over to those of public goods. In this section we discuss a one-stage public goods game (similar to the standard common-pool resource game) and a two-stage public goods game, where after the first stage, subjects have a sanctioning opportunity (similar to the common-pool resource with sanctioning opportunities).
We start with the following linear public goods game. There are n ≥ 2 players who decide simultaneously on their contribution levels to the public good. Each player has an endowment of y. The monetary payoff of player i is given by where 1 / n < a < 1. Because a < 1, a marginal investment into G causes a monetary loss of (1 – a), that is, the dominant strategy of a completely selfish player is to choose gi = 0. However, because a > 1/n, the aggregate monetary payoff is maximized if each player chooses gi= y.
Consider now a slightly different public goods game that consists of two stages. At stage 1 the game is identical to the previous game. At stage 2 each player i is informed about the contributions of all other players and can simultaneously impose a costly punishment on the other players, just as in the sanctioning common-pool resource game discussed.
What does the standard model predict for the two-stage game? Because punishments are costly, players’ dominant strategy at stage 2 is to not punish. Therefore, if selfishness and rationality are common knowledge, each player knows that the second stage is completely irrelevant. As a consequence, players have exactly the same incentives at stage I as they have in the one-stage game without punishments, that is, each player’s optimal strategy is to contribute nothing.
To what extent are these predictions of the standard model consistent with the data from public goods experiments? For the one-stage game there are, fortunately, a large number of experimental studies. In a meta-study of 12 experimen-