DO ELECTIONS for governor matter in national politics? Not in the way they used to.

A gain in governorships by one party in the off-year is not a harbinger of victory in the next presidential year. Nor does any winner of a gubernatorial contest become the general of an army of patronage workers that can make the difference in the next election. Voters understand that different things are at stake in state and national elections and make their choices accordingly. So what should you look for in the 36 races for governor this year?

One thing not to be much concerned about is the partisan balance. Democrats currently hold 27 of these 36 governorships, and will surely lose some because of the retirement of popular Democratic incumbents in heavily Republican states. This may be importan to you if you're a strong partisan living in Arizona or Nebraska or Colorado. Otherwise it matters only to those who need to know how many chairs to set out at the next Republican and Democratic governors conferences.

There are, however, three races in which you might want to keep your eye on partisan trends. The three giant Sun Belt states of California, Texas and Florida account for more than half the nation's population gain in the 1980s and, in different ways, forecast the national future. Will conservative Republican George Deukmejian beat Los Angeles Mayor Tom Bradley as he did, narrowly, in 1982? Recent polls have gone both ways. Will moderate Democrat Mark White hold the office he won in an upset in Texas four years ago? Will Bob Graham, now running for the Senate in Florida, be succeeded by a like-minded moderate Democrat or by a Republican? The answers will provide interesting clues about where the nation's politics is headed.

Another interesting question is how many incumbent governors will win reelection. The big story in the 1985 gubernatorial elections was the strength of incumbents: Republican Thomas Kean won 70 percent in New Jersey and Charles Robb's fellow Democrats won impressively in Virginia. Their high job ratings are not unique. Even in large states that until recently had corrosively negative politics -- such as New York, Massachusetts and Michigan -- incumbent governors have 65-plus percent positive job ratings. Such positive feelings, about politicians and about their states' prospects, were uncommon in the cynical 1970s but have become something close to the norm today.

This is not just a mindless response to a few feel- good TV ads. You can hardly spend any time with a governor of either party without hearing an infectiously optimistic recital of the buoyant economic growth in his or her state plus an enthusiastic recounting of how state and local government have been working to make things better. While national politicians wrangle, local politicians have been doing things, and life at home has been improving. Or so at least voters have been saying in the pro-incumbent results of 1984 and 1985. Will this bias toward incumbents continue in the 1986 elections for governor? The answer will tell more about the future of our politics than any of the standard indicators.