Now it seems to me that the opposite interpretation was indeed the correct one. The following is excerpted from Schwarz (2000):
As a large body of experimental research documents, individuals who are in a happy mood are more likely to adopt a heuristic processing strategy that is characterised by top-down processing, with high reliance on pre-existing knowledge structures and relatively little attention to the details at hand. In contrast, individuals who are in a sad mood are more likely to adopt a systematic processing strategy that is characterised by bottom-up processing, with little reliance on preexisting knowledge structures and considerable attention to the details at hand (for a review see Schwarz & Clore, 1996). [...] Consistent with the more detail-oriented processing style fostered by negative moods, Luce, Bettman, and Payne (1997, p. 384) observed that ``decision processing under increasing negative emotion both becomes more extensive and proceeds more by focusing on one attribute at a time’’ .
These differences in processing style presumably reflect that our thought processes are tuned to meet the requirements of the current situation, which are in part signalled by our affective states (Schwarz, 1990). In a nutshell, we usually feel bad when things go wrong and feel good when we face no particular problems. Hence, negative affective states may signal that the current situation is problematic and may hence elicit a processing style that pays close attention to the specifics of the apparently problematic situation. In contrast, a positive affective state may signal a benign environment that allows us to rely on our usual routines and preexisting knowledge structures. [...]
Hertel, Neuhof, Theuer, and Kerr (this issue) extend this line of research by addressing the impact of moods on individuals’ decision behaviour in a chicken dilemma game. Consistent with previous theorising, their findings suggest that individuals in a happy mood are likely to heuristically imitate the behaviour of other players, whereas individuals in a sad mood based their moves on a systematic analysis of the structure of the game.
The existence of the hedonic treadmill, I believe, supports this version of the hypothesis. Beyond a certain point, increases in e.g. income affect happiness only in relation to the income of others. Once your basic needs have been met, getting more money makes you happier only if it makes you better off as compared to others. People want to "keep up with the Joneses", which makes sense from an evolutionary perspective. For your genes to spread in a population, it is not enough to manage: your genes must spread more than those of your competitors.
Of course, the usual evolutionary psychology caveats must be kept in mind. The "keeping up with the Joneses" phenomenon is a very Western one, and Westerners have been shown not to be representative of the human race as a whole. It is my understanding that the Eastern cultures have traditionally been considerably less materialistic, which undermines the support that the Joneses provide to this hypothesis.
Curiously, Clore & Huntsinger (2007) summarize various studies as indicating that "participants in positive moods tend to rely more on stereotypes to guide their thinking about members of various social groups than do those in negative moods, who tend to rely on individuating information". This doesn't fit the usual intuitions of racism being caused by fear or adverse circumstances.
Contrary to most people’s intuitions, happy moods promote group stereotyping, whereas sad moods promote a focus on individuals [27,28]. One relevant study involved a mock trial in which a Latino student was accused of a stereotype-consistent offense. The results showed that individuals in happy moods were more likely than those in sad moods to have their verdicts influenced by the stereotype .
In this experiment, the stereotyping seems to reflect a general cognitive style rather than prejudice as such. Indeed, similar findings come from marketing and political science studies showing that happy moods promote reliance on brand names as opposed to product attributes among consumers , and a reliance on political party as opposed to candidate positions among voters .
In addition, a surprising result in the mock jury study  was that angry jurors responded like happy jurors, rather than like sad ones. This finding is consistent with affect-as-information logic, which always asks about the information inherent in affective states. Despite being a negative emotion, anger carries positive information about one’s own position. When angry, one believes oneself to be correct, which should increase confidence in one’s own cognitions. Thus, anger would be expected to show the same processing effects as happiness .
G.L. Clore & J.R. Huntsinger (2007). How emotions inform judgment and regulate thought. Trends in Cognitive Sciences, 11 (9), 393-399. doi:10.1016/j.tics.2007.08.005
N. Schwarz (2000). Emotion, cognition, and decision making. Cognition and Emotion, 14 (4), 433-440.