To Be or Not to Be Conservative

Should conservatives support moderate Republicans? I provide here pragmatic arguments for supporting more conservative Republicans right now.

To be clear, I am talking about the value of voting for the more conservative candidates in Republican primaries. I am not talking about a third party or the general election. Such cases can be made, but I do not make them here.

The counterargument is that if we support the more conservative candidate, we risk losing to an extremely liberal Democrat. We make perfect the enemy of the good.

One supposed example of such a case was provided recently by Senator Orrin Hatch. According to the Salt Lake Tribune,

Hatch blamed extreme conservatives for the 2008 defeat of Sen. Gordon Smith, a politically moderate but fiscally conservative Republican from Oregon.  Hatch said if the Tea Party had not backed a constitutionalist candidate in that race, Smith wouldn't have lost to Democrat Jeff Merkley, whom Hatch described as "the most liberal senator," by 45,000 votes.

Merkley received 49% of the vote, Smith received 46%, and David Brownlow, the Constitution Party candidate, received 5%. The Hatch calculation was that if the third-party voters had gone for Smith, he would have won 51-49 over Merkley.

The result was that instead of a senator with an American Conservative Union score of 33, we have one with an ACU score of 4. (Smith's 2008 ACU score put him one point above Mary Landrieu [32] and nine points below Arlen Specter [42], both Democrats now.)

I now present my arguments, both specifically against this Gordon Smith example and also more generally.

First, the Gordon Smith example refers to support for a third-party candidate in the general election. That is a straw man. What most of us are talking about is whom conservatives should support in Republican primaries. That, at least, is the only case I'm making now.

Secondly, the third party in this case was the Constitution Party, not the Tea Party at all. To associate the Constitution Party with the Tea Party movement was an irresponsible leap of logic.

Third, the phrase "politically moderate but fiscally conservative" is simply inaccurate (as it almost always is). The National Journal breaks down vote scores into economic, social, and foreign categories. The NJ's economic score for Gordon Smith (54) was actually below his social score (60). His foreign score was even lower (46). The next four economic scores below him belonged to Susan Collins (53), Evan Bayh (52), Mary Landrieu (51) and Clair McCaskill (49), the last three being Democrats. Gordon Smith was moderate on everything and conservative on nothing.

Fourth, conservatives and moderates make up 75% of all of us, and even 70% of Oregonians. Why should we have to choose between candidates who are between 4% and 33% conservative?

The way to avoid a Gordon Smith type outcome is to remove the Gordon Smith candidate in the primary and replace him with someone more conservative. Yet the GOP's continued support for incumbents and favored candidates works precisely against such a thing.

If conservatives vote for Republicans no matter what, the GOP receives no signal that it must move right. Then the GOP learns nothing, drifts left, and keeps nominating ever-more-liberal candidates as long as they are at least a hair more conservative than the Democrats. Where does that lead?

It leads to 2001-2006, when Republicans had the presidency, the House, and the Senate, and they accomplished nothing on the domestic front except

  • No Child Left Behind,
  • Campaign Finance Reform,
  • Prescription coverage under the already-bankrupt Medicare system,
  • Jim Jeffords,
  • Lincoln Chafee,
  • Arlen Specter, and
  • Ignominious electoral defeat in 2006 and 2008.

You get more of what you consistently vote for.

Finally, there can be a short-term reason to vote for the more conservative Republican in a primary, even if you think he or she has a lower chance of winning the general election than the moderate Republican. There is a fundamental calculation in any one race that can be made in a cold, rational manner.

BEGIN MATH WARNING (Skip to end of MATH WARNING if you are math-phobic.)

The game-theory, or decision-theory, payoff matrix is given here. (MBAs learn how to do this in graduate school. Engineers learn it in one of their easy classes.)

Payoff to Conservatives For Four Possible Outcomes

Republican Wins

Democrat Wins

Conservative Republican



Moderate Republican



If we had our choice among all possible outcomes, we would of course choose the "Good" one: The conservative Republican wins. But we do not have that choice. The voting comes in two steps: the primary and the general election. In a primary, we choose only between Republicans. And which Republican we choose also affects the probability that that Republican will win the general election.

The new payoff matrix is four probabilities. In a primary, we get to choose only between the two rows. I define "Pc" as the probability that the conservative Republican would win the general election over the Democrat and "Pm" as the probability that the moderate one would.

Expected Payoff to Conservatives For Four Possible Outcomes

Republican Wins

Democrat Wins

Conservative Republican

Pc X Good

(1-Pc) X Bad

"Moderate" Republican

Pm X Fair

(1-Pm) X Bad

If we choose the Conservative Republican, our total expected payoff is the sum of the two possibilities in that row.

(Pc X Good) + [(1-Pc) X Bad]

We want to choose the row with the highest expected payoff: the best chance of having the most conservative officeholder after the election. If you trust my algebra, we should choose the moderate Republican in the primary if

Pm/Pc > (Good - Bad)/(Fair - Bad)

If you followed this so far, you are now wondering what numbers to put in. For Good, Bad, and Fair, we could use, say, ACU scores. Let's use a real primary race as an example: the McCain-Hayworth race.

McCain's lifetime ACU score is 82, but his most recent score was only 65. I'm going to give him a generous score of 70, which replaces "Fair" in the formula. Hayworth's lifetime score was 97, with his last two years being 96 and 100. So "Good" will be 97. I'll assign a typical Democrat ACU score of 10 for "Bad."

Our decision criterion is then

Pm/Pc > (97 - 10)/(70 - 10) = 1.45

We should vote for McCain if we think his chances of beating the Democrat are at least 1.45 times better than Hayworth's.

McCain has a good chance of winning the general election -- let's say 90%. Then by the above formula, if conservatives think Hayworth has at least a 62% chance of winning the general election, they should vote for him; otherwise, they should vote for McCain.

(Frankly, this analysis makes it a pretty close call, or even favors McCain. I can assure you that that was not my intention. But you don't have to use my numbers. Maybe ACU scores are not what you consider the best measures of "goodness," after all.)

END OF MATH WARNING (You may now proceed to the meaning of all this math.)

I admit that the mathematical approach was a bit tongue-in-cheek, but not totally. The idea is that this can be looked at rationally. We might not know all the numbers required for the rational calculation, but the lesson is this:

It is not necessarily stupid to vote for the more conservative candidate in a Republican primary, even knowing his or her chances of winning the general election are lower than the moderate's chances.

The "right" decision depends on how good, bad, or indifferent we think the candidates are and their relative chances of winning in the general election. But the calculation is not a no-brainer. It depends.

And even that formulaic approach is purely for a single race, trying to maximize the "conservativeness" of your next office holder. If we look at the longer term, there are stronger reasons to favor the more conservative candidates having to do with signals from the voters and learning by the GOP.

Essentially, voting for the more conservative candidates is the only way conservatives can send a signal to the GOP to move right. If the party does not learn this the easy way in the primaries, then it could very well learn it the hard way in the general elections. I would say that 2008 was a hard lesson for the GOP. Did they learn anything?

Randall Hoven can be contacted at or via his website,
If you experience technical problems, please write to