Read responses to this article.

Why can’t Massachusetts elect a Democratic governor? This stands as one of the bigger political puzzles of the last decade. In national politics, Massachusetts is a bulwark of Democratic strength, and has been in most elections since 1928, when the state went for Al Smith. After two Bay State victories for Dwight D. Eisenhower, in 1960 favorite son John F. Kennedy put Massachusetts in the Democratic column so solidly it has hardly budged. In 1972, Massachusetts was the only state in the union to support Democratic nominee George McGovern, and although Ronald Reagan twice won the state by a narrow margin, since then the Bay State has been a Democratic stalwart.

In the three presidential elections of the last decade, Massachusetts consistently outperformed the national Democratic vote. Clinton ran five points ahead of his national vote here in 1992, and 12 points ahead of his national vote in 1996. In 2000, the outcome here was in so little doubt that Democratic nominee Al Gore aired no ads in the Boston television market, the sixth largest in the country. Massachusetts money raisers like Alan Solomont had to go to campaign headquarters in Nashville, Tenn., to see the commercials their dollars bought.

During this period Massachusetts also showed overwhelming Democratic strength at other levels. In spite of serious Republican challenges to Ted Kennedy in 1994 and John Kerry in 1996, Massachusetts retains two Democratic US senators. All 10 Massachusetts seats in the US House of Representatives are held by Democrats. In the state Legislature, the number of Republicans has dwindled to near-token levels. Since 1993, there have been insufficient Republican members of the 40-seat Senate to sustain a gubernatorial veto, and in the House, the Republicans have a margin of just three over the 20 needed to force a roll-call vote. In party registration, Democrats outnumber Republicans by more than 2-to-1.

This lopsided partisan advantage made the 2002 election of Gov. Mitt Romney a bitter disappointment for Massachusetts Democrats. But if the loss by Shannon O’Brien–a moderate who united all factions of the party, in addition to being the state’s first female nominee for governor from a major party–came as a shock to the Democratic faithful, it should not have been such a surprise. Indeed, the Democrats’ showing in the gubernatorial elections of the ’90s was the exact opposite of the party’s strength in presidential tallies. Whereas Democratic presidential contenders received 48 percent, 61 percent, and 60 percent, for an average well over 50 percent, Democratic gubernatorial candidates got 47 percent of the vote in 1990, 27 percent in 1994, and 47 percent in 1998. With O’Brien capturing 45 percent of the vote last November, it’s about time for Massachusetts Democrats to realize that four gubernatorial losses in a row may be a trend to be explained, rather than flukes to be explained away.

So what’s going on? Why can’t Massachusetts, which is so Democratic at every other level of politics, manage to elect a Democratic governor? A closer look at the electorate–and at those votes of the 1990s, presidential and gubernatorial–yields some answers.

Party, ideology, and the middle class

Though they remain by far the largest political party in Massachusetts, Democrats slipped from 41.8 percent of registered voters to 36.6 percent between 1990 and 2000. But their loss was not the Republicans’ gain–the GOP’s share of party registration remained almost identical over the decade, at roughly 13 percent. Instead, the percentage of independent, or “unenrolled,” voters increased by 7.5 points, from 42.1 percent to 49.6 percent. Virtually half the Massachusetts electorate now identifies itself with neither major party.

Among these independent voters, the Democratic candidates had varying degrees of success in the 1990s. Though gubernatorial nominee John Silber and presidential candidate Bill Clinton got roughly the same share of the independent vote (44 and 43 percent, in 1990 and 1992, respectively), in 1996 and 2000 the Democratic presidential candidates (Clinton, in re-election, and Al Gore) won solid majorities of the independents, while gubernatorial candidates Mark Roosevelt, in 1994, and Scott Harshbarger, in 1998, won over no more than 37 percent of the unaffiliated. The presidential hopefuls combined majority shares of independents with a solid Democratic base for comfortable victories in Massachusetts, but the gubernatorial candidates had just one-third of the independent vote to add to the party faithful–not enough for a win. The weakness of statewide Democratic candidates among this important and growing part of the electorate is a large part of the explanation for their failure to capitalize on the advantage Democrats hold in Massachusetts.

This shift in the electorate away from party identification does not signal any shift in ideology that would seem to disadvantage Democratic candidates for governor. If anything, the Massachusetts electorate grew slightly more liberal over the course of the 1990s, with Voter News Service exit polls showing a slight rise in self-declared liberals from 23 percent in 1990 to 26 percent in 1996 and ’98, and a jump to 35 percent in 2000. Self-described moderates made up roughly half of the electorate for most of the decade, then dropped somewhat, to 42 percent, in 2000.

But ideological appeal explains little about the voting patterns of the 1990s. Presidential candidates Clinton and Gore, who both identified themselves as “New Democrats” periodically at odds with the liberal base of the party, captured a far larger share of liberal votes in Massachusetts than did Roosevelt and Harshbarger (let alone Silber, who seemed to delight in irritating liberals). They won more moderate votes as well. This would seem to give the lie to typical ideological critiques of the gubernatorial candidates, both from the left (they were insufficiently liberal) and from the right (they were too liberal to appeal to moderates). Democratic candidates for governor in the 1990s neither mobilized the liberal base nor made as strong inroads among the moderates as the party’s presidential standard-bearers did. With the exception of Silber’s relatively strong showing among Massachusetts conservatives, the gubernatorial candidates of the ’90s failed to win as big a following as the presidential candidates did across the ideological spectrum.

The fact is, ideological labels matter less in state elections than they do in national politics. For instance, governors don’t appoint Supreme Court justices who could overturn Roe v. Wade, the preservation of which has been one of the rallying points for liberals and feminists for nearly three decades. In addition, Republican gubernatorial candidates in Massachusetts have managed to be sufficiently liberal, particularly on social issues, to pacify some liberal voters and, most importantly, not to scare off moderate ones the way GOP presidential candidates did throughout the ’90s. Finally, it has been the Republican candidates, not the Democrats, who have staked out the New Democratic territory–conservative on fiscal matters and liberal on social matters such as abortion and affirmative action.

Finally, there are the demographics of the electorate. Democrats have always fancied themselves the champions of “working people.” But the average working person of the 1990s bears little resemblance to the factory worker of the 1930s, around which so much Democratic Party rhetoric still swirls. The electorate has been changing in ways consistent with the emergence of an information-age economy. The industrial economy, which was centered around factories that were breeding grounds for the politics of class, are distant memories to many voters–although for most of the candidates at the June 2002 Democratic Convention in Worcester, that world seemed to be alive and well.

According to exit polls, during the 1990s the proportion of voters with a high school education or less dropped from 30 percent to roughly 20 percent, while the segment with post-graduate degrees increased from 20 percent to more than a quarter. At the beginning of the decade, college graduates constituted slightly less than half of all voters, but they were up to 55 percent by the end.

These rising education levels ought to be good news for Democrats. The most reliable Democratic vote in the ’90s came from those with post-graduate degrees–the fastest growing portion of the electorate. But at all other levels of education, Democratic presidential candidates held on to clear majorities, whereas their gubernatorial counterparts did not.

The information age has changed the economic structure of the electorate as well. In the economic boom of the 1990s, the growth in the electorate in Massachusetts came at the top of the income distribution. In 1992, 64 percent of those who went to the polls earned less than $50,000 a year; by 2000, only 37 percent earned less than that. The share of voters earning more than $100,000 a year grew from 9 percent in 1994 (it was not a reported category in 1992) to 21 percent in 2000.

Democratic candidates in the ’90s did well most reliably at the bottom of the income scale, where numbers are shrinking, and least well at the very top of the income distribution, which is small but growing. But the Democratic presidential candidates of 1996 and 2000 still managed to capture roughly two-thirds of the vote up to the $75,000 income level, and majority support even in the highest income brackets. In between, gubernatorial candidate Harshbarger lost almost as much ground in the broad middle class ($30,000 to $100,000) as he did among the most affluent, holding on to a slim majority only in the $50,000-to-$75,000 category.

These findings, on educational attainment and income, suggest that national Democrats have a strong appeal to Massachusetts voters across social class and education levels. Democratic presidential candidates have managed to maintain their strength among the poorest voters while still appealing to the broad middle class–even, to an extent, to the growing number of affluent voters. Democratic candidates for governor, however, hold sway principally among the poor; they lose political traction not only among the wealthy, but even among the moderately well-off, except for the most highly educated.

One demographic advantage Massachusetts Democrats have long counted on is the gender gap, but it has done more good for presidential candidates than for gubernatorial nominees. Republican candidates in Massachusetts have managed to avoid the right-wing positions that characterize the national Republican Party, and therefore have avoided the fate of Republican national candidates at the Massachusetts ballot box.

In recent presidential races other than 1992’s (when independent candidate H. Ross Perot siphoned off many men’s votes and left no gender gap between Republican George H.W. Bush and Democrat Bill Clinton), women’s Democratic votes exceeded men’s Republican votes by large enough margins to produce huge advantages for the Democratic candidates. In 1996, Clinton beat Republican Bob Dole in Massachusetts by 20 points among men and by 46 points among women, a “net” gender gap advantage of 26 points for the Democratic candidate. Al Gore came out of Massachusetts with a net gender advantage of 24 points. But in every governor’s race of the 1990s, the male vote for the Republican candidate exceeded the female vote for the Democratic candidate, thus producing a net GOP advantage. In 1990, Democrat Silber seemed to go out of his way to insult the working women who have, traditionally, given Democrats a net gender gap advantage. But that doesn’t explain why Roosevelt gave away a two-point gender advantage to Weld, and Harshbarger a 10-point gap to Paul Cellucci.

Unfortunately for the purposes of this article and our general understanding of politics, the Voter News Service failed to release exit polls from the 2002 election. As a result, we cannot fill out the story of the past 15 years with as precise a picture of the voters as we have obtained from previous election results.

But a University of Massachusetts same-day poll of people who voted on Election Day, conducted for The Boston Globe, confirms some of the patterns of the 1990s. By income, O’Brien won a majority only among those earning less than $50,000 a year, a group that constituted just 26 percent of respondents; in education, only voters at the two ends of the spectrum–those with a high school diploma or less and those with post-graduate degrees–favored the Democrat. Even for the first major-party female gubernatorial nominee, the gender gap went against the Democrat: Though women favored O’Brien by 9 percentage points, men preferred Romney by 11 points, for a net gender gap of 2 points in the Republican’s favor.

Although we can’t conclude anything specific about how ideological groupings (liberal, moderate, and conservative) voted in the last election, the UMass poll tells us that Romney beat O’Brien among independents by a margin of 49 to 45 percent. (O’Brien did seem to solidify the party base better than previous gubernatorial candidates had done: Only 24 percent of Democratic voters defected to Romney.) Geography tells a political tale as well. As reported in the Boston Phoenix last year, Sarah Kahan, a Princeton University senior, did a study of voting patterns in Massachusetts and found that the majority (52 percent) of unenrolled voters live in the 89 towns between Route 128 and Interstate 495–the Silicon Valley of Massachusetts. (Population growth in these areas undoubtedly contributed to the growing numbers of unenrolled voters statewide over the past decade.) This region was not kind to O’Brien. According to the UMass poll, Romney trounced O’Brien in the Route 128 area, 56 to 42 percent, and edged out a two-point win, 49 to 47 percent, in the vicinity of I-495.

A Democratic message for the New Economy

Looking at the patterns that emerge from the last seven state and national elections, the good news for Democrats is that many Massachusetts voters continue to identify with the Democratic Party and, more importantly, very few voters identify with the Republican Party, in spite of a decade of Republican governors. Republican successes at the gubernatorial level have yet to result in large-scale success in other offices in the state. But that’s cold comfort for the Democratic candidates who keep making futile runs at the governor’s office–and to those who think that a Democratic executive branch could make a difference in the lives of Massachusetts citizens.

Comparing Democratic voting strength in the governor’s races with Democratic voting strength in the presidential races, we find the following:

• Democratic presidential candidates managed to win independent voters and the Democratic gubernatorial candidates did not;

• Democratic presidential candidates managed to win among moderates and run up large margins among liberals, while Democratic gubernatorial candidates did not;

• Democratic presidential candidates did better than the gubernatorial candidates in the heart of the middle class and at the top of the income distribution, as measured by both education and income; and

• Democratic presidential candidates produced a gender gap that worked in their favor, while the gender gap worked against Democratic gubernatorial candidates.

So what should this mean for the Massachusetts Democratic Party going forward? Some answers can be found in a book from the early 1990s, authored by Morley Winograd and Dudley W. Buffa, called Taking Control: Politics in the Information Age. The book was little known and largely overlooked by the purveyors of conventional wisdom in Washington, DC. But it did have one enthusiastic reader: President Bill Clinton.

The book’s thesis was remarkably simple: The information-age economy will create a new politics that will replace the politics of the industrial age. “The new technology of the information age will change the American economy and the American government,” Winograd and Buffa wrote. “Knowledge workers will become the new majority in American politics. Whoever first offers them a new social contract for the information age will become the dominant political force in America in the twenty-first century.” Writing in 1994, Winograd and Buffa could have easily been writing about Massachusetts eight years later.

In the 1990s, Massachusetts found itself at the forefront of the information-age economy. But the Democratic gubernatorial candidates of these years, and since, often sounded as if they were stuck in the industrial-age economy. As recently as last year’s state Democratic Convention, candidates for governor went out of their way to pledge increases in the minimum wage, always to the cheers of delegates. Raising the minimum wage is not a bad policy, but it is a policy with direct relevance to less than 20 percent of the electorate. (Indeed, the minimum wage has so little punch in Massachusetts politics that Republicans don’t even bother to oppose it: Both Cellucci in 1998 and Romney in 2002 came out in support of an increase.) It is not a policy calculated to capture the imagination of the middle- and upper-income suburban voters who are increasingly critical to electoral success. At the same time, these Democrats did not manage– or, in most cases, even try–to identify themselves with making government more efficient or with protecting the tax dollars of the voters.

As we learn more about information-age voters, one of the gross generalizations that can be made is that they tend to be conservative on economic policy, liberal on social policy, and, increasingly, resistant to partisan attachments. In national politics, as demonstrated in the presidential votes of the ’90s, the indifference to safety-net needs and extremism on social issues associated with the GOP and its candidates alienate Massachusetts New Economy voters nearly across the board–independents as well as Democrats, moderates as well as liberals. But in governor’s races, where Bay State Republicans are moderate and hot-button social issues are less relevant, the issues of economic growth, fiscal prudence, and administrative efficiency become paramount. On these issues, Democrats lose some, if not all, of their built-in partisan advantage. And where their deep penetration of the Massachusetts political fabric becomes associated with an old-fashioned, machine-style politics, the Democrats’ very dominance can become an electoral liability.

These factors–overemphasis on traditional appeals to a stagnant, if not shrinking, Democratic base; feeble courtship of political independents in the high-tech belt; association with an insular and self-serving political establishment–could be seen in the waning days of the 2002 gubernatorial campaign. Take the four-day O’Brien campaign bus tour just before Election Day. Beginning with a rally in New Bedford featuring former President Clinton, the bus trip made 22 stops, most of them in the working-class enclaves and black churches of the Democratic base. Only one stop, on the Natick Common, was in a suburb. It is no wonder that suburban voters felt that the Democratic candidate had nothing to say to them.

Meanwhile, the Romney campaign was not only courting those suburban voters, it was playing up O’Brien’s ties to the insider politics of Beacon Hill. In a very effective television commercial called “The Gang of Three,” Romney depicted the triumvirate that would rule state government if O’Brien were to become governor: O’Brien, House Speaker Thomas Finneran, and Robert Travaglini, who had recently announced that he had the votes to succeed Thomas Birmingham as Senate president. (Birmingham’s failed gubernatorial run should have been a warning to Democrats about the image the Legislature gives to the party.) The implication–and the impact–was clear: O’Brien should be viewed not as a new, managerially minded reformer who could run the state but as part of the machine politics of the past that runs Beacon Hill.

“When it comes to the top job, voters showed that they are willing to embrace a man almost alien to the state’s governmental culture rather than opt for a woman seen as too much a part of that culture,” observed Boston Globe columnist Scot Lehigh in an election postmortem. “Once you get outside of Route 128, you find communities who look askance at the things occurring in the State House,” Gerard Desilets, who headed O’Brien campaign operations in the western suburbs, told the Globe. “Romney’s campaign was very sophisticated in playing that message.”

Since the election, Democrats in Massachusetts have begun a healthy debate over what went wrong. Some have accused O’Brien, whose credentials as a moderate were part of her strength in the Democratic primary, of a “leftward lurch” at the end of the campaign. Her response to a question on abortion and parental rights in the October 29 debate struck many as flip and disrespectful. And other issues, such as gay marriage, which she impulsively came out in support of, got O’Brien stuck in the morass of social issues that do Democrats little good on the state level, distinguishing them from live-and-let-live Republicans only through liberal stands that offend traditionalists even within the Democratic base.

But if O’Brien allowed herself to get caught out on a (left) limb on social issues, it may have been just as harmful, if not more harmful, that she was so easily portrayed as a creature of the Beacon Hill establishment, not as an agent of change. This is where the negative image of the Democratic Legislature spilled over to hurt the Democratic candidate for governor. It was not that O’Brien ran a bad campaign. She was the only candidate at the state party convention to reach out beyond the activists in the Worcester Centrum to the independents in the electorate. She tried to portray herself as a fiscal watchdog, playing up her role in exposing Big Dig overruns, etc. But these campaign tactics did not fit her biography, which was really an insider’s biography. In the end, the first female nominee for governor was too easily cast as one of the Good Ole Boys–a disaster for the Democrats.

The negative image of the Democratic Party, despite the Democratic reflex of Massachusetts voters, has swamped Democratic nominees before. One of the great unnoticed ironies of Massachusetts politics occurred in the 1998 gubernatorial race. As attorney general, Scott Harshbarger had targeted corruption within the Democratic Party establishment. But this did him little good with the voters when he became the Democratic nominee for governor. According to exit polls from that race, more than 20 percent of the voters cited “ethics” and “moral concerns” as the issues that mattered most in their choice for governor. Even among those ethics-minded voters, Harshbarger the crusading prosecutor lost by 2 percentage points to Cellucci! This curiosity should have been a warning to Democrats that their image as a party had become an albatross around the necks of their gubernatorial nominees.

In the “Gang of Three” attack, Romney found the real Achilles’ heel of Massachusetts Democrats: They are viewed as a party too entrenched to be trusted to run government, at least all by themselves. For Democrats to win back the governorship, they will have to handle not only the balance of left-versus-right on issues, but also new-versus-old in governance.

Nevertheless, if Democrats can come back in Washington, as they did in the 1990s, they can come back in Massachusetts. On the state level as well as the national, when Democrats convincingly present themselves as competent managers of the economy and the government, they win. That’s the way to send a powerful message to the heart of the middle class–to the people who supported Bill Clinton and Al Gore overwhelmingly in three consecutive presidential elections but abandoned Democrats in recent gubernatorial races. Many people in the middle of the income distribution and the middle of the educational distribution are on the fringes of the New Economy –which is the future of prosperity in Massachusetts, the current economic doldrums notwithstanding. They want help in getting into the middle of it, and they want security so that they can rise within it.

The next Democrat to be elected governor of Massachusetts will be someone who shows that he or she can manage a modern, information-age economy and a modern, information-age government. Whether that is in 2006, or not until long thereafter, depends on the party–and its candidates.

Elaine C. Kamarck is a lecturer in public policy at Harvard’s Kennedy School of Government. She served as senior policy advisor to the vice president in the Clinton-Gore Administration, where she led the National Performance Review.