While voters rarely reward a governor for making an inspired Senate choice, any decision can backfire and taint his own future. In Alaska, former Gov. Frank Murkowski picked his daughter Lisa for his Senate seat. She won re-election two years later, but her father lost his own re-election bid to Sarah Palin.
The current system persists not because governors crave this power but because whenever the issue arises, the incumbent party is reluctant to give it up. Just this week, Democrats in Springfield cooled on the idea of a special election, for fear that Republicans might have a chance to pick up the seat.
Ironically, if history is any guide, partisan self-interest argues exactly the opposite. In the century since the 17th Amendment provided for direct election to the Senate, about 180 senators have been appointed to fill vacancies. When their appointed terms ran out, those senators met with three fates in equal measure. One-third of them ran for election and lost. One-third ran and won. One-third decided not to run at all. So as a general rule, only one-third of all appointed senators win the voters' blessing in their own right, and half of appointed senators who run for election lose their seats.
By congressional standards, that's an astonishingly low survival rate. By contrast, over the last quarter-century, incumbent senators have won re-election about 87 percent of the time. That's true even in a wave election like 2008, when only four or five incumbents out of 30 lost (pending the outcome in Minnesota). As far as I can tell, there has never been a Senate cycle where incumbents lost as many races as they won—not even in the historic wipeouts of 1932, 1958, and 1980. In other words, in an entire century of direct elections, no Senate class has ever done as badly in a single disastrous year as appointed Senators do on average.
Even those numbers may understate the odds against replacement players in the Senate. Since World War II, appointed senators who sought election have done still worse, losing 56 percent of the time. At that rate, a political party would actually stand a better chance of winning a Senate race if the governor appointed someone from the other party to fill it.
Well-aware of those odds, governors must choose between picking an interim senator who will not seek re-election (as outgoing Delaware Gov. Ruth Ann Minner did in selecting Ted Kaufman) or trying to figure out which candidate stands the best chance of holding on to the seat by winning statewide (as New York Gov. David Paterson and Colorado Gov. Bill Ritter have pledged to do). Of course, the most effective way to tell who can win statewide is to do exactly what the current system in most states is designed to avoid: Hold an election.
As an electoral matter, then, Senate appointments are all too often self-defeating. The other arguments sometimes made for the status quo are convenience and national emergency. Special elections cost money, although any state crass enough to put a price tag on democracy might want to take into account how much it stands to lose by taking a 2-in-3 chance that appointing a replacement will cost it two years of Senate seniority. Norm Ornstein of the American Enterprise Institute argues that in the event of a terrorist attack, we need the ability to name replacements quickly. But that's hardly an argument to rely on appointed replacements in normal times.