the stable component of happiness (in ten year terms), is probably 80% genetic. This is the results of research done on identical twins (Lykken & Tellegan 1996), it accords well with research on the stable component of personality whihc is also 80%. And remember that happiness is strongly correlated with personality traits. (I do not know on any study contradicting the above data).
There maybe sophisticated ways to get around these thing. I will ignore them now.
Most of the variation in personal happiness is in the element changing over time.
Practical implication: Only invest in improving happiness at this second, this minute, or maybe this day etc. long term improvement, are on the face of it, impractical.
Intuitively speaking, improving next day mood is much easier. While "long term" stuff is kind of fantasy.
Aside from the above, there are other reasons to be short term.
1) The longer the horizon the less reliable our opinions (see Gilbert' Stumbling on Hapiness for a great read AND much interesting stuff)
2) Uncertaincy grows with time. We know so little about ten years from now, that it is simply stupid to assume knowledge and even suffer now for then.
3) Usually, the longer the horizon our expectation plans etc. are more complicated. Reliability of a chain is practically the product of the weakness of its chains, implying that long chains of causation are worthless. Reminds me the teacher telling us at age 11 "You be good boys now get accepted to a good school, be good boys then, get accepted to a good college, then you will get a good wife and have a happy life". I was so young, but felt that going throught life that way is utterly stupid. Not just stupid, uttrly dtupid!