Racial lines.

As of this writing, according to Nexis, "landslide" has been used to describe the 1994 elections 469 times, followed by "revolution" (267 mentions), "earthquake" (135), "tidal wave" (120), and "tsunami" (18). Those descriptions are overwrought and misleading. The earth did not actually move on November 8. Something even more interesting may have happened.

In 1994 Republicans won 51 per cent of all votes cast in contested House elections, which is not a landslide and indeed is the narrowest popular-vote victory for either party since 1984. But the Republicans' margin did win them 53 per cent of the available seats. This fact should be at least mildly surprising, because Democrats controlled more state legislatures than did Republicans following the 1990 census and so could engage in more gerrymandering. With a margin that narrow, a few competent gerrymanders should have allowed the Democrats either to retain their congressional majority or to come much closer than they did. What happened?

What happened was the Voting Rights Act (VRA). As amended and reinterpreted after 1982, the VRA often requires state legislatures to construct districts in which blacks or Hispanics constitute a numerical majority ("majority-minority districts"). These minorities are overwhelmingly Democratic in most of the country. "Packing" Democrats' core voters into just a few districts does help elect more black and Hispanic representatives. But it also elects more Republicans - 12 to 15 more, according to many observers, than would have been elected without the race-conscious redistricting following the 1990 census.

The VRA's pro-Republican bias is a nice irony, because no legislative program has been more important to the Democrats - the party of unintended consequences. It is difficult to overstate the amount of political capital that the Democratic Left has invested in racial gerrymandering. Indeed, Lani Guinier recently argued in The New York Times Magazine that within the existing (assertedly unfair) system of winner-take-all, single-member districts, stacking congressional districts by race is the only ticket to representation in Congress for African-American and Hispanic people.

Miss Guinier's appraisal is hardly compelling, because it explicitly assumes that elected officials of one race (or sex, religion, etc.) cannot legitimately, and will not in practice, represent the interests of constituents of a different race. Yet in representative systems, everyone must be represented by someone other than himself and, therefore, different from himself in important ways. "Representation" may be a legal fiction, but that does not make it an hallucination. Furthermore, it is not a good idea to encourage voters to insist on the political primacy of their racial identities rather than inspiring them to define themselves in more constructive, less incendiary ways.

But this rejoinder only rehearses the long-standing argument against the race-conscious classifications that the Left has been asserting against the rest of America for the past 25 years. Quotas are the cardinal issue of the Left; without racial politics there scarcely is a political Left in this country. So it is a barricade that will be defended to the end, whatever the cost to the Democratic Party.

The Democrats' Albatross 

This political reality is obvious, but it would be embarrassing for any of the players to acknowledge that race-conscious districting, indeed racial quotas in any form, have become an albatross for the Democratic Party. Hence the bouquets of denial about the connection between the VRA and the Democrats' late change of fortune. "Don't Scapegoat the Gerrymander," Miss Guinier implored in the Times Magazine. "The Democrats lost for many reasons," she argued - chiefly white flight, the lassitude of the party's traditional base, and low voter turnout. In other words, the Democrats should move left and forget about the tilt of the playing field.

Elaine Jones, head of the NAACP Legal Defense and Education Fund (NAACP-LDEF), likewise condemned the notion that the VRA had cost the Democrats dearly. In a detailed analysis reported by the New York Times on December 1, she invoked the tidalwave theory. Eric Schnapper, an NAACP-LDEF lawyer, agreed: "A thirty-foot tidal wave came along." One of their colleagues, Penda Hair, put it succinctly: "It's pure racism to say that because blacks are more loyal and you can identify them by skin color, you can use them as padding to save white Democrats," she said.

Squandering Democrats

Is there, then, no connection between the VRA and the Democrats' undoing? None, says the NAACP-LDEF. At most, the VRA cost Democrats one seat, in North Carolina; on balance, the statute probably resulted in a net gain for Democrats.

The NAACP-LDEF dismisses the possibility that the VRA had an impact on 15 other seats gained by Republicans because the districts involved were not contiguous with a majority-minority district. But this is a grievous non-sequitur. The impact of a gerrymander, as of any districting decision, is pervasive. Every district in a state, and not merely contiguous districts, is affected by a gerrymander. The electoral makeup of a district in downstate Illinois is affected by the design of the district next door, which is in turn affected by the design of the next district over, and so on; how one gerrymanders the suburbs north of Chicago can and does affect the district map four hundred miles to the south. It therefore makes no sense to look at electoral outcomes only in districts adjoining VRA gerrymanders. The VRA hurts the Democrats precisely because Democratic-leaning voters who are concentrated for the sake of a racial gerrymander are thereafter unavailable for use in a partisan gerrymander.

Partisan gerrymanderers have long understood the principles of rigging the electoral game. Though they used to have to do it by hand instead of with a computer and a CD-ROM from the Census Bureau, it wasn't too hard to tilt the playing field, so long as the only goals were maximizing a given party's electoral success and protecting certain incumbents. Once the constraint of VRA compliance is added, modern districting becomes a far more demanding business. Having to deal with this extra factor is what largely accounts for the contorted shapes of many majority-minority districts.

It is quite hopeless to "prove," in the mathematical sense, what effect the VRA has had on election outcomes, because results inevitably depend on any number of indissolubly mixed-up variables: the qualities of individual candidates, local issues, money, luck, national political trends, the Zeitgeist, etc. So the VRA's consequences will have to remain a matter of interpretation. But common sense makes up for the lack of apodictic proof.

In 1994, winning Democrats in majority-minority districts garnered, on average, 73 per cent of the vote. Winning white Democrats had a much harder time: their average vote was 60 per cent, and fully one-fifth won with 52 per cent or less. Ten per cent of majority-minority district races were uncontested, as compared to less than 4 per cent of the other districts with a Democratic incumbent.

A few examples from the South illustrate the gerrymander's unhappy effect on Democrats. In North Carolina, Republicans won an aggregate of 54 per cent of the vote in the 11 contested districts and carried 7 (or 64 per cent) of those seats. In the two majority-minority districts, the Democratic winners had a total of 50,000 votes to spare. Meanwhile, three fellow Democrats who were white lost their races by 17,000, 7,000, and 1,300 votes. Charlie Rose, 22-year incumbent and czar of the tobacco and peanut subcommittee, was re-elected with only a 5,000-vote margin. (Expect a strong challenger next year, Charlie.) In Georgia, the three majority-minority districts yielded a combined victory margin of 113,000 votes. Democratic incumbent Buddy Darden could have used some of those stalwarts. He lost to Republican Bob Barr by only 5,000 votes. Two other Georgia Democrats lost by margins of 24,000 and 36,000.

There are also examples outside of the South. Dan Rostenkowski used to win the old 8th District of Illinois by 4 to 1 margins with the help of several Hispanic enclaves whose residents by 1990 accounted for nearly one-third of the district's population. After the last census, Hispanics were given their own district, a deformed fractal horseshoe of territory, stripping Rostenkowski of one of his most reliable blocs of supporters. Instead he was left with tens of thousands of Lincoln Park yuppies, who were never enthusiastic about being represented by an old-school Chicago pol and who jumped off the fantail right after the chairman's indictment.

In addition to such direct consequences, we have to consider the opportunity costs of VRA gerrymanders. How many seats would Democrats have won if, at the time of the decennial redistricting, state legislatures and governors had been at liberty to gerrymander without having to comply with the VRA? Under scores of scenarios Democrats would have been able to cobble together 13 extra seats, and a majority, in the House.

Both Lani Guinier and American University history professor Allan Lichtman have tried to shield the VRA from these brickbats by pointing out that many of the same districts that elected Republicans last November had elected Democrats only two years earlier. But the real question is more subtle than, "Who won?" It is, "Who will tend to win?" In other words, whom do the ground rules favor?

The Swing Ratio

Political scientists use what is called the "swing ratio" to measure the political impact on legislatures of changes in voter preferences. The swing ratio tells how many additional seats a party can expect from a given percentage increase in its margin of victory. It has long been appreciated that a majority of the vote in American-style single-member congressional districts will lead to a somewhat inflated congressional majority. A party that typically wins 54 per cent of the votes in a fair election might expect to win 58 per cent of the seats. The gap widens as the majority increases. A party that wins 70 per cent of the popular vote might win more than 90 per cent of the seats.

Ceteris paribus, the ratio of votes to seats is symmetrical between parties, If Republicans get 58 per cent of the seats when they get 54 per cent of the votes, then Democrats should expect 58 per cent of the seats when they get 54 per cent of the votes. A run-of-the-mill partisan gerrymander alters a state's swing ratio, introducing a systemic bias in favor of one party. Thus, in California in 1984, Republicans actually received more votes than Democrats in House races, but Democrats won 60 per cent of the seats because of a gerrymander conceived by the late Representative Phil Burton, a consummate exhibition of how to steal elections legally. Had Democrats polled only 45 per cent of the vote, they probably still would have won a majority of the delegation; had they carried as much as 55 per cent of the popular vote, few California Republicans would have been left. Mr. Burton's virtuosity had changed the electoral equation across the entire

A change in the swing ratio caused by racial gerrymandering has the same multi-level effect. Close races formerly won by Democrats are now won by Republicans. Close races favoring Republicans become Republican walkovers. Formerly safe Democratic seats become contestable, which attracts better, more electable Republican candidates, whose prospects brighten as Democrats are obliged to spend money to defend formerly safe seats.

There is a secondary ripple effect that may be even more important in the long run. As John J. Miller points out in the February issue of Reason, racial gerrymandering inexorably pushes the Democratic Party as a whole to the left - which is to say, toward the margins of influence in political life. "Blacks now control almost 19 per cent of the seats held by Democrats in the House, up from about 15 per cent in the 103rd Congress," Miller writes. "Many are on the far-left fringes of the party, while most are just plain liberal." The country grows more conservative even as the relative influence of Jesse Jackson's rooting section increases within Democratic Party councils.

An Ambivalent Court

This is a grim prospect for Democrats, maybe even a recipe for a crack-up, but - another other irony - they may have their bacon saved for them, over their own squeals of protest, by the Supreme Court, seven of whose members are Republicans. The Court seems not to know what to think of the VRA and racial gerrymanders. In Thornburgh v. Gingles eight years ago, it approved race-conscious districting for the purpose pose of enhancing minority representation in legislatures. But in Shaw v. Reno, decided in 1993, the Court said bizarrely shaped districts drawn to comply with the VRA, at least if they were too bizarre, might be unconstitutional without a "compelling justification."

The Court obviously needs to compose its thoughts on this matter, and there is a case before it, Miller v. Johnson, that will allow it to do just that. A decision is expected this spring. The result may be a holding that race-conscious districting is out-and-out unconstitutional. A clean decision on this subject would be a relief, but there is no particular reason to expect anything so lucid. The Supreme Court is just a glorified committee, after all, one with a rich tradition of obfuscation and tap dancing where racial politics is concerned. So it may not decide anything coherent, in which case the onus would be on Congress to decide what, if anything, to do about the VRA.

The temptation will be great for the Republican majority to write the VRA down as the Democrats' problem and walk away from it. This they must not do. The VRA is the supreme instance of racial politics in American life. It is time for the era of color-consciousness to come to an end, even if that means helping the Democrats.

Still, the majority should not be overly hasty in attacking the VRA. Quotas, "goals and timetables," set-asides, race-norming, "statistical discrimination," and a massive apparatus of administrative procedure and practice have driven the poison of rent-seeking by race into every nook and cranny of American life. All this has to be dismantled. There is a great deal to do. There is no good reason to do the VRA first. And there is an excellent reason to do it last.