Running Scared

Painfully often the legislation our politicians pass is designed less to solve problems than to protect the politicians from defeat in our neverending election campaigns. They are, in short, too frightened of us to govern.

Scared politician

TO an extent that astonishes a foreigner, modern America is about the holding of elections. Americans do not merely have elections on the first Tuesday after the first Monday of November in every year divisible by four. They have elections on the first Tuesday after the first Monday of November in every year divisible by two. In addition, five states have elections in odd-numbered years. Indeed, there is no year in the United States -- ever -- when a major statewide election is not being held somewhere. To this catalogue of general elections has of course to be added an equally long catalogue of primary elections (for example, forty-three presidential primaries last year). Moreover, not only do elections occur very frequently in the United States but the number of jobs legally required to be filled by them is enormous -- from the presidency of the United States to the post of local consumer advocate in New York. It has been estimated that no fewer than half a million elective offices are filled or waiting to be filled in the United States today.

Americans take the existence of their never-ending election campaign for granted. Some like it, some dislike it, and most are simply bored by it. But they are all conscious of it, in the same way that they are conscious of Mobil, McDonald's, Larry King Live, Oprah Winfrey, the Dallas Cowboys, the Ford Motor Company, and all the other symbols and institutions that make up the rich tapestry of American life.

Political analysts Jonathan Rauch and Larry Sabato respond to points made in this article.

To a visitor to America's shores, however, the never-ending campaign presents a largely unfamiliar spectacle. In other countries election campaigns have both beginnings and ends, and there are even periods, often prolonged periods, when no campaigns take place at all. Other features of American elections are also unfamiliar. In few countries do elections and campaigns cost as much as they do in the United States. In no other country is the role of organized political parties so limited.

America's permanent election campaign, together with other aspects of American electoral politics, has one crucial consequence, little noticed but vitally important for the functioning of American democracy. Quite simply, the American electoral system places politicians in a highly vulnerable position. Individually and collectively they are more vulnerable, more of the time, to the vicissitudes of electoral politics than are the politicians of any other democratic country. Because they are more vulnerable, they devote more of their time to electioneering, and their conduct in office is more continuously governed by electoral considerations. I will argue that American politicians' constant and unremitting electoral preoccupations have deleterious consequences for the functioning of the American system. They consume time and scarce resources. Worse, they make it harder than it would otherwise be for the system as a whole to deal with some of America's most pressing problems. Americans often complain that their system is not sufficiently democratic. I will argue that, on the contrary, there is a sense in which the system is too democratic and ought to be made less so.

The Perpetual Campaign
Although this article is written by a foreigner, a Canadian citizen who happens to live in Great Britain, it is not written in any spirit of moral or intellectual superiority. Americans over the years have had quite enough of Brits and others telling them how to run their affairs. I have no wish to prolong their irritation. What follows is the reflections of a candid friend.

POLITICS and government in the United States are marked by the fact that U.S. elected officials in many cases have very short terms of office and face the prospect of being defeated in primary elections and have to run for office more as individuals than as standard-bearers for their party and have continually to raise large sums of money in order to finance their own election campaigns. Some of these factors operate in other countries. There is no other country, however, in which all of them operate, and operate simultaneously. The cumulative consequences, as we shall see, are both pervasive and profound.

The U.S. Constitution sets out in one of its very first sentences that "the House of Representatives shall be composed of members chosen every second year by the people of the several states." When the Founding Fathers decided on such

a short term of office for House members, they were setting a precedent that has been followed by no other major democratic country. In Great Britain, France, Italy, and Canada the constitutional or legal maximum for the duration of the lower house of the national legislature is five years. In Germany and Japan the equivalent term is four years. Only in Australia and New Zealand, whose institutions are in some limited respects modeled on those of the United States, are the legal maximums as short as three years. In having two-year terms the United States stands alone.

Members of the Senate are, of course, in a quite different position. Their constitutionally prescribed term of office, six years, is long by anyone's standards. But senators' six-year terms are not all they seem. In the first place, so pervasive is the electioneering atmosphere that even newly elected senators begin almost at once to lay plans for their re-election campaigns. Senator Daniel Patrick Moynihan, of New York, recalls that when he first came to the Senate, in 1977, his colleagues when they met over lunch or a drink usually talked about politics and policy. Now they talk about almost nothing but the latest opinion polls. In the second place, the fact that under the Constitution the terms of a third of the Senate end every two years means that even if individual senators do not feel themselves to be under continuing electoral pressure, the Senate as a whole does. Despite the Founders' intentions, the Senate's collective electoral sensibilities increasingly resemble those of the House.

Cowed politician Most Americans seem unaware of the fact, but the direct primary -- a government-organized popular election to nominate candidates for public office -- is, for better or worse, an institution peculiar to the United States. Neither primary elections nor their functional equivalents exist anywhere else in the democratic world. It goes without saying that their effect is to add a further dimension of uncertainty and unpredictability to the world of American elective politicians.

In most other countries the individual holder of public office, so long as he or she is reasonably conscientious and does not gratuitously offend local or regional party opinion, has no real need to worry about renomination. To be sure, cases of parties refusing to renominate incumbent legislators are not unknown in countries such as France, Germany, and Canada, but they are relatively rare and tend to occur under unusual circumstances. The victims are for the most part old, idle, or alcoholic.

The contrast between the rest of the world and the United States could hardly be more striking. In 1978 no fewer than 104 of the 382 incumbent members of the House of Representatives who sought re-election faced primary opposition. In the following three elections the figures were ninety-three out of 398 (1980), ninety-eight out of 393 (1982), and 130 out of 409 (1984). More recently, in 1994, nearly a third of all House incumbents seeking re-election, 121 out of 386, had to face primary opposition, and in the Senate the proportion was even higher: eleven out of twenty-six. Even those incumbents who did not face opposition could seldom be certain in advance that they were not going to. The influence -- and the possibility -- of primaries is pervasive. As we shall see, the fact that incumbents usually win is neither here nor there.

To frequent elections and primary elections must be added another factor that contributes powerfully to increasing the electoral vulnerability of U.S. politicians: the relative lack of what we might call "party cover." In most democratic countries the fate of most politicians depends not primarily on their own endeavors but on the fate -- locally, regionally, or nationally -- of their party. If their party does well in an election, so do they. If not, not. The individual politician's interests and those of his party are bound together.

In contrast, America's elective politicians are on their own -- not only in relation to politicians in most other countries but also in absolute terms. Party is still a factor in U.S. electoral politics, but it is less so than anywhere else in the democratic world. As a result, American legislators seeking re-election are forced to raise their own profiles, to make their own records, and to fight their own re-election campaigns.

Empty Promises If politicians are so vulnerable electorally, it may be protested, why aren't more of them defeated? In particular, why aren't more incumbent congressmen and senators defeated? The analysis here would seem to imply a very high rate of turnover in Congress, but in fact the rate -- at least among incumbents seeking re-election -- is notoriously low. How can this argument and the facts of congressional incumbents' electoral success be reconciled?

This objection has to be taken seriously, because the facts on which it is based are substantially correct. The number of incumbent congressmen and senators defeated in either primary or general elections is low. But to say that because incumbent members of Congress are seldom defeated, they are not really vulnerable electorally is to miss two crucial points. The first is that precisely because they are vulnerable, they go to prodigious lengths to protect themselves. Like workers in nuclear-power stations, they take the most extreme safety precautions, and the fact that the precautions are almost entirely successful does not make them any less necessary.

Second, congressmen and senators go to inordinate lengths to secure re-election because, although they may objectively be safe (in the view of journalists and academic political scientists), they do not know they are safe -- and even if they think they are, the price of being wrong is enormous. The probability that anything will go seriously wrong with a nuclear-power station may approach zero, but the stations tend nevertheless to be built away from the centers of large cities. A congressman or a senator may believe that he is reasonably safe, but if he wants to be re-elected, he would be a fool to act on that belief.

AMERICAN politicians run scared -- and are right to do so. And they run more scared than the politicians of any other democratic country -- again rightly. How did this come to be so?

The short answer is that the American people like it that way. They are, and have been for a very long time, the Western world's hyperdemocrats. They are keener on democracy than almost anyone else and are more determined that democratic norms and practices should pervade every aspect of national life. To explore the implications of this central fact about the United States, and to see how it came to be, we need to examine two different interpretations of the term "democracy." Both have been discussed from time to time by political philosophers, but they have never been codified and they certainly cannot be found written down in a constitution or any other formal statement of political principles. Nevertheless, one or the other underpins the political practice of every democratic countryeven if, inevitably, the abstract conception and the day- to-day practice are never perfectly matched.

One of these interpretations might be labeled "division of labor." In this view, there are in any democracy two classes of people -- the governors and the governed. The function of the governors is to take decisions on the basis of what they believe to be in the country's best interests and to act on those decisions. If public opinion broadly supports the decisions, that is a welcome bonus. If not, too bad. The views of the people at large are merely one datum among a large number of data that need to be considered. They are not accorded any special status. Politicians in countries that operate within this view can frequently be heard using phrases like "the need for strong leadership" and "the need to take tough decisions." They often take a certain pride in doing what they believe to be right even if the opinion of the majority is opposed to it.

The function of the governed in such a system, if it is a genuine democracy, is very important but strictly limited. It is not to determine public policy or to decide what is the right thing to do. Rather, it is to go to the polls from time to time to choose those who will determine public policy and decide what the right thing is: namely, the governors. The deciding of issues by the electorate is secondary to the election of the individuals who are to do the deciding. The analogy is with choosing a doctor. The patient certainly chooses which doctor to see but does not normally decide (or even try to decide) on the detailed course of treatment. The division of labor is informal but clearly understood.

It is probably fair to say that most of the world's major democracies -- Great Britain, France, Germany, Japan -- operate on this basis. The voters go to the polls every few years, and in between times it is up to the government of the day to get on with governing. Electing a government and governing are two different businesses. Electioneering is, if anything, to be deplored if it gets in the way of governing.

This is a simplified picture, of course. Democratically elected politicians are ultimately dependent on the electorate, and if at the end of the day the electorate does not like what they are doing, they are dead. Nevertheless, the central point remains. The existing division of labor is broadly accepted.

The other interpretation of democracy, the one dominant in America, might be called the "agency" view, and it is wholly different. According to this view, those who govern a country should function as no more than the agents of the people. The job of the governors is not to act independently and to take whatever decisions they believe to be in the national interest but, rather, to reflect in all their actions the views of the majority of the people, whatever those views may be. Governors are not really governors at all; they are representatives, in the very narrow sense of being in office solely to represent the views of those who sent them there.

In the agency view, representative government of the kind common throughout the democratic world can only be second- best. The ideal system would be one in which there were no politicians or middlemen of any kind and the people governed themselves directly; the political system would take the form of more or less continuous town meetings or referenda, perhaps conducted by means of interactive television. Most Americans, at bottom, would still like to see their country governed by a town meeting.

IN this political ethos, finding themselves inhabiting a turbulent and torrid electoral environment, most American elective officials respond as might be expected: in an almost Darwinian way. They adapt their behavior -- their roll-call votes, their introduction of bills, their committee assignments, their phone calls, their direct-mail letters, their speeches, their press releases, their sound bites, whom they see, how they spend their time, their trips abroad, their trips back home, and frequently their private and family lives -- to their environment: that is, to their primary and overriding need for electoral survival. The effects are felt not only in the lives of individual officeholders and their staffs but also in America's political institutions as a whole and the shape and content of U.S. public policy.

It all begins with officeholders' immediate physical environment: with bricks, mortar, leather, and wood paneling. The number of congressional buildings and the size of congressional staffs have ballooned in recent decades. At the start of the 1960s most members of the House of Representatives contented themselves with a small inner office and an outer office; senators' office suites were not significantly larger. Apart from the Capitol itself, Congress was reasonably comfortably housed in four buildings, known to Washington taxi drivers as the Old and New House and Senate Office Buildings. The designations Old and New cannot be used any longer, however, because there are now so many even newer congressional buildings.

Congressional staffs have grown at roughly the same rate, the new buildings having been built mainly to house the staffs. In 1957 the total number of people employed by members of the House and Senate as personal staff was 3,556. By 1991 the figure had grown to 11,572 -- a more than threefold increase within the political lifetime of many long-serving members. Last year the total number of people employed by Congress in all capacities, including committee staffs and the staffs of support agencies like the Congressional Research Service, was 32,820, making Congress by far the most heavily staffed legislative branch in the world.

Much of the growth of staff in recent decades has been in response to the growth of national government, to Congress's insistence on strengthening its policymaking role in the aftermath of Vietnam and Watergate, and to decentralization within Congress, which has led subcommittee chairmen and the subcommittees themselves to acquire their own staffs. But there is no doubt that the increase is also in response to congressional incumbents' ever-increasing electoral exposure. Congress itself has become an integral part of America's veritable "elections industry."

One useful measure of the changes that have taken place -- and also an important consequence of the changes -- is the increased proportion of staff and staff time devoted to constituent service. As recently as 1972 only 1,189 House employees -- 22.5 percent of House members' personal staffs -- were based in home-district offices. By 1992 the number had more than doubled, to 3,128, and the proportion had nearly doubled, to 42.1 percent. On the Senate side there were only 303 state-based staffers in 1972, making up 12.5 percent of senators' personal staffs, but the number had more than quadrupled by 1992 to 1,368, for fully 31.6 percent of the total. Since a significant proportion of the time of Washington-based congressional staffs is also devoted to constituent service, it is a fair guess that more than half of the time of all congressional staffs is now given over to nursing the district or state rather than to legislation and policymaking.

Much constituent service is undoubtedly altruistic, inspired by politicians' sense of duty (and constituents' understandable frustration with an unresponsive bureaucracy); but at the same time nobody doubts that a large proportion of it is aimed at securing re-election. The statistics on the outgoing mail of members of Congress and their use of the franking privilege point in that direction too. Congressional mailings grew enormously in volume from some 100 million pieces a year in the early 1960s to more than 900 million in 1984 -- nearly five pieces of congressional mail for every adult American.) New restrictions on franking introduced in the 1990s have made substantial inroads into that figure, but not surprisingly the volume of mail emanating from both houses of Congress is still invariably higher in election years.

The monetary costs of these increases in voter-oriented congressional activities are high: in addition to being the most heavily staffed legislative branch in the world, Congress is also the most expensive. But there is another, nonmonetary cost: the staffs themselves become one of the congressman's or senator's constituencies, requiring management, taking up time, and always being tempted to go into business for themselves. American scholars who have studied the burgeoning of congressional staffs express concern about their cumulative impact on Congress as a deliberative body in which face-to-face communication between members, and between members and their constituents, facilitates both mutual understanding and an understanding of the issues. Largely in response to the requirements of electioneering, more and more congressional business is conducted through dense networks of staffers.

One familiar effect of American politicians' vulnerability is the power it accords to lobbyists and special-interest groups, especially those that can muster large numbers of votes or have large amounts of money to spend on campaigns. Members of Congress walk the electoral world alone. They can be picked off one by one, they know it, and they adjust their behavior accordingly. The power of the American Association of Retired Persons, the National Rifle Association, the banking industry, and the various veterans' lobbies is well known. It derives partly from their routine contributions to campaign funds and the quality of their lobbying activities in Washington, but far more from the votes that the organizations may be able to deliver and from congressmen's and senators' calculations of how the positions they take in the present may affect their chances of re-election in the future -- a future that rarely is distant. Might a future challenger be able to use that speech against me? Might I be targeted for defeat by one of the powerful lobbying groups?

A second effect is that American politicians are even more likely than those in other countries to engage in symbolic politics: to use words masquerading as deeds, to take actions that purport to be instrumental but are in fact purely rhetorical. A problem exists; the people demand that it be solved; the politicians cannot solve it and know so; they engage in an elaborate pretense of trying to solve it nevertheless, often at great expense to the taxpayers and almost invariably at a high cost in terms of both the truth and the politicians' own reputations for integrity and effectiveness. The politicians lie in most cases not because they are liars or approve of lying but because the potential electoral costs of not lying are too great.

At one extreme, symbolic politics consists of speechmaking and public position-taking in the absence of any real action or any intention of taking action; casting the right vote is more important than achieving the right outcome. At the other extreme, symbolic politics consists of whole government programs that are ostensibly designed to achieve one set of objectives but are actually designed to achieve other objectives (in some cases simply the re-election of the politicians who can claim credit for them).

Take as an example the crime bills passed by Congress in the 1980s and 1990s, with their mandatory-minimum sentences, their three-strikes-and-you're-out provisions, and their extension of the federal death penalty to fifty new crimes. The anti-drug and anti-crime legislation, by the testimony of judges and legal scholars, has been at best useless and at worst wholly pernicious in its effects, in that it has filled prison cells not with violent criminals but with drug users and low-level drug pushers. As for the death penalty, a simple measure of its sheer irrelevance to the federal government's war on crime is easily provided. The last federal offender to be put to death, Victor H. Feguer, a convicted kidnapper, was hanged in March of 1963. By the end of 1995 no federal offender had been executed for more than thirty years, and hardly any offenders were awaiting execution on death row. The ferocious-seeming federal statutes were almost entirely for show.

The way in which the wars on drugs and crime were fought cannot be understood without taking into account the incessant pressure that elected officeholders felt they were under from the electorate. As one former congressman puts it, "Voters were afraid of criminals, and politicians were afraid of voters." This fear reached panic proportions in election years. Seven of the years from 1981 to 1994 were election years nationwide; seven were not. During those fourteen years Congress passed no fewer than seven major crime bills. Of those seven, six were passed in election years (usually late in the year). That is, there was only one election year in which a major crime bill was not passed, and only one non-election year in which a major crime bill was passed.

Another effect of the extreme vulnerability of American politicians is that it is even harder for them than for democratically elected politicians in other countries to take tough decisions: to court unpopularity, to ask for sacrifices, to impose losses, to fly in the face of conventional wisdom -- in short, to act in what they believe to be their constituents' interest and the national interest rather than in their own interest. Timothy J. Penny, a Democrat who left the House of Representatives in 1994, put the point starkly, perhaps even too harshly, in (1995).

Voters routinely punish lawmakers who try to do unpopular things, who challenge them to face unpleasant truths about the budget, crime, Social Security, or tax policy. Similarly, voters reward politicians for giving them what they want -- more spending for popular programs -- even if it means wounding the nation in the long run by creating more debt.

America's enduring budget deficit offers a vivid, almost textbook illustration. For nearly a generation -- ever since the early 1980s -- American politicians have bemoaned the deficit and exhorted themselves to do something about it. However, they have never done nearly enough, even in their own eyes. Why? Part of the answer undoubtedly lies in genuine ideological differences that make it hard for conservatives and liberals to compromise; but much of the answer also lies in the brute fact that every year in the United States is either an election year or a pre-election year, with primaries and threatened primaries intensifying politicians' electoral concerns. In 1985 Senator Warren Rudman, of New Hampshire, reckoned that he and other senators who had voted for a bold deficit-reduction package had flown a "kamikaze mission."One of his colleagues said they had "jumped off a cliff."Twelve years later, not surprisingly, the federal budget remains in deficit.

NUMEROUS opinion polls show that millions of Americans are profoundly dissatisfied with the functioning of their political system. Consequently, there is a widespread disposition in the United States -- at all levels of society, from the grass roots to the editorial conference and the company boardroom -- to want to make American democracy "work better," and concrete proposals abound for achieving this goal.

The proposed reforms can be grouped loosely under four headings. First come those that if implemented would amount to the creation of electronic town meetings, taking advantage of technological developments such as CD-ROM, interactive cable systems, electronic mail, and the Internet. The Wall Street Journal referred in this general connection to "arranging a marriage of de Tocqueville and technology."

Second, and related, are proposals for promoting democratic deliberation and citizen participation. The Kettering Foundation and the Public Agenda Foundation already organize National Issues Forums that embrace some 3,000 educational and civic groups across America. David Mathews, the president of the Kettering Foundation, considers these modern forums to be directly linked to America's ancient "town meeting tradition." Benjamin R. Barber, a political philosopher at Rutgers University, would go further and create a nationwide network of neighborhood assemblies that could take actual decisions on strictly local matters and also debate and lobby on broader national questions. James S.Fishkin, a political scientist at the University of Texas, likewise seeks to leap the modern barriers to face-to-face democracy by means of what he calls "deliberative opinion polls" (which have been tried, with considerable success, in England).

The third group of proposed reforms is equally radical but more old-fashioned. This group seeks to complete the work of Progressive Era reformers by extending to the federal level the characteristic state-level reforms that were introduced in that period: the referendum, the initiative, and the recall. The political analyst Kevin Phillips, for example, suggests that "the United States should propose and ratify an amendment to the Constitution setting up a mechanism for holding nationwide referendums to permit the citizenry to supplant Congress and the president in making certain categories of national decisions." He would also like to see congressmen and senators be subject to popular recall once they have been in office for a year. Certainly proposals of this kind have broad public support. Depending on the precise wording of the question, more than 50 percent of Americans support the idea of national referenda and more than 80 percent support both the initiative and the recall.

Finally, many commentators -- and the majority of the American public -- strongly back the newest and most fashionable item on the "making democracy work better" agenda: the imposition of term limits on both state and federal elected officials, notably members of Congress. But the great majority of those who favor term limits, true to the American democratic tradition, are less concerned with good government and the public interest as such than with the present generation of politicians' alleged lack of responsiveness to the mass of ordinary people. At the center of this argument is the idea that the United States is now governed by an unresponsive, self-perpetuating, and increasingly remote class of professional politicians, a class that ought to be replaced as soon as possible by "citizen legislators" -- men and women who will serve the people simply because they are the people. As one advocate of term limits puts it, ordinary people -- the proposed citizen legislators of the future -- "know things about life in America that people who have lived as very self-important figures in Washington for thirty years have no way of knowing or have forgotten."

Some of the items on this four-part shopping list of reforms are intrinsically attractive, or at least a good case can be made for them. Nevertheless, taken as a whole, the mainstream reformist agenda, with its traditional American emphasis on agency democracy and its view of politicians as mere servants of the people's will, rests on extremely tenuous conceptual foundations and, more important, is almost certainly inappropriate as a response to the practical needs of turn-of-the-century America. America's problem of governance is not insufficient responsiveness on the part of its elected leaders. On the contrary, America's problem is their hyper-responsiveness. Politicians do not need to be tied down still further, to be subjected to even more external pressures than they are already. Rather, they need to be given just a little more political leeway, just a little more room for policy maneuver. Reforms should seek to strengthen division-of-labor democracy, not to create a still purer form of American-style agency democracy.

IF the reformist prescriptions are bad ones, there may be something wrong with the reformist diagnoses on which they are based. What are the principal sources of dissatisfaction with the current state of American democracy?

Many commentators have gotten into the habit of blaming Americans' dissatisfaction, in an almost knee-jerk fashion, on "the Vietnam War and Watergate." It is certainly the case that evidence of widespread dissatisfaction began to appear during and shortly after Vietnam and Watergate. Post hoc, ergo propter hoc? Maybe. But in the first place, Vietnam and Watergate led to a flowering of idealism as well as cynicism (and to the election, in 1974, of the "Watergate babies," one of the most idealistic and public-spirited cohorts ever to be elected to Congress). And in the second place, it seems strange to attribute the dissatisfactions of the 1990s to events that took place in the 1960s and early 1970s.

That distance in time is roughly that between the two world wars; most of today's college students were not yet born when President Richard Nixon resigned. To be sure, subsequent scandals have undoubtedly (and deservedly) damaged the reputations of the White House and Congress, but at least some of the sleaze of recent years has come about because politicians need such enormous sums to finance their re-election campaigns.

Two other hypotheses can be dismissed, or at least assigned little importance. One is that politicians today are a poor lot compared with the intellectual and moral giants of the past. It probably is the case that having to run scared all the time has tended to drive some able people out of politics and to discourage others from coming in. But the phenomenon is a relatively recent one, and for the time being there is no reason to think that the average congressman or senator is in any way inferior to his or her predecessors. The quality of America's existing political class is at most a small part of the problem.

The same is almost certainly true of the idea that divided government -- in which one party controls one or both houses of Congress while the other controls the presidency -- is to be preferred. Divided government has characterized America for most of the past thirty years, and it has been associated with some of the more spectacular political and policy failures of that period -- the Iran-contra scandal of the 1980s (which arose out of a Republican Administration's desire to circumvent a Democratic Congress), and successive shutdowns of parts of the government as Presidents and Congress have failed to agree on timely taxing and spending measures. Other things being equal, divided government is probably to be regretted.

All the same, it is hard to credit the idea that Americans' disillusionment with their politics would be significantly less today if party control had been mainly undivided over the past thirty years. On the one hand, recent periods in which the government has not been divided (the Carter years, 1977-1980, and the first two Clinton years, 1993-1994)were not notably successful (Carter never surmounted the energy crisis, and Clinton failed to reform America's health-care system even though that reform had figured prominently in his campaign promises). On the other hand, as David R. Mayhew, a political scientist at Yale University, has shown, periods of divided government have often been extremely productive in legislative terms. On balance, divided government appears to be more of a nuisance and a distraction than a root cause of either the government's difficulties or the public's disillusionment.

The idea that the system suffers from the excessive power of interest groups, however, needs to be taken seriously. Jonathan Rauch, in his recent book , argues persuasively that America's interest groups have become larger, more numerous, and more powerful over the past three decades, to the point that they now have the capacity to prevent the government from doing almost anything that would disadvantage or offend any of the clients they represent -- taking in, as it happens, virtually the whole American population.

Rauch is probably right; but one needs to go on to ask, as he himself does, what the power of these pullulating and all-encompassing lobby groups is based on. The answer is straightforward: their power depends ultimately on their money, on their capacity to make trouble for elected officials, on the votes of their members (the AARP has more than 30 million members), and on elective politicians' fear of not being re-elected. The groups' power, in other words, depends on politicians' electoral vulnerability; and America's interest groups are peculiarly powerful in large measure because America's elective politicians are peculiarly vulnerable. It is not quite as simple as that -- but almost.

It is also important to note the precise timing of the developments described by Rauch and by almost everyone else who has written on this subject. Nearly all these developments date, almost uncannily, from the past thirty years: the rise in the number of interest groups, the growth in their membership and power, the decline in the public's trust in government officials, and the increased sense among voters that who they are and what they think do not matter to politicians and officials in Washington. In other words, the origins of the present era of democratic discontent can be traced to the end of the 1960s and the beginning of the 1970s. It was then that people began to think something was wrong not with this or that aspect of the system but with the system itself.

What happened at that time? It is hard to escape the conclusion that the crucial developments, largely provoked by the Vietnam War and Watergate, were the attempts from 1968 onward to open up the American system, to make it more transparent, to make it more accessible, to make it, in a word, more "democratic." These attempts led to an increase in the number of primary elections, to a further weakening of America's already weak political parties, to increases in the already high costs of electoral politics, and to the increasing isolation, in an increasingly hostile environment, of elective officials. In short, the post-Vietnam, post-Watergate reforms led, as they were meant to lead, to increased vulnerability to their electorates on the part of individual American officeholders.

The paradox that has resulted is obvious and easily stated. Recent history suggests that when large numbers of Americans become dissatisfied with the workings of their government, they call for more democracy. The more they call for more democracy, the more of it they get. The more of it they get, the more dissatisfied they become with the workings of their government. The more they become dissatisfied with the workings of their government, the more they call for more democracy. The cycle endlessly repeats itself.

PRECISELY because American politicians are so exposed electorally, they probably have to display -- and do display -- more political courage more often than the politicians of any other democratic country. The number of political saints and martyrs in the United States is unusually large.

There is, however, no special virtue in a political system that requires large numbers of politicians to run the risk of martyrdom in order to ensure that tough decisions can be taken in a timely manner in the national interest. The number of such decisions that need to be taken is always likely to be large; human nature being what it is, the supply of would-be martyrs is always likely to be small. On balance it would seem better not to try to eliminate the electoral risks (it can never be done in a democracy) but to reduce somewhat their scale and intensity. There is no reason why the risks run by American politicians should be so much greater than the risks run by elective politicians in other democratic countries.

How, then, might the risks be reduced? What can be done? A number of reforms to the existing system suggest themselves. It may be that none of them is politically feasible -- Americans hold tight to the idea of agency democracy -- but in principle there should be no bar to any of them. One of the simplest would also be the most radical: to lengthen the terms of members of the House of Representatives from two years to four. The proposal is by no means a new one: at least 123 resolutions bearing on the subject were introduced in Congress in the eighty years from 1885 to 1965, and President Lyndon B. Johnson advocated the change in his State of the Union address in January of 1966.

A congressman participating in a Brookings Institution round table held at about the time of Johnson's message supported the change, saying, "I think that the four years would help you to be a braver congressman, and I think what you need is bravery. I think you need courage." Another congressman on the same occasion cited the example of another bill that he believed had the support of a majority in the House. "That bill is not going to come up this year. You know why it is not coming up? . . . Because four hundred and thirty-five of us have to face election. . . . If we had a four-year term, I am as confident as I can be the bill would have come to the floor and passed."

A similar case could be made for extending the term of senators to eight years, with half the Senate retiring or running for re-election every four years. If the terms of members of both houses were thus extended and made to coincide, the effect in reducing America's never-ending election campaign would be dramatic.

There is much to be said, too, for all the reasons mentioned so far, for scaling down the number of primary elections. They absorb extravagant amounts of time, energy, and money; they serve little democratic purpose; few people bother to vote in them; and they place additional and unnecessary pressure on incumbent officeholders. Since the main disadvantage of primaries is the adverse effect they have on incumbents, any reforms probably ought to be concerned with protecting incumbents' interests.

At the moment, the primary laws make no distinction between situations in which a seat in the House or the Senate is already occupied and situations in which the incumbent is, for whatever reason, standing down. The current laws provide for a primary to be held in either case. An incumbent is therefore treated as though the seat in question were open and he or she were merely one of the candidates for it. A relatively simple reform would be to distinguish between the two situations. If a seat was open, primaries would be held in both parties, as now; but if the incumbent announced that he or she intended to run for re-election, then a primary in his or her party would be held only if large numbers of party supporters were determined to have one -- that is, were determined that the incumbent should be ousted. The obvious way to ascertain whether such determination existed would be by means of a petition supervised by the relevant state government and requiring a considerable number of signatures. The possibility of a primary would thus be left open, but those who wanted one would have to show that they were both numerous and serious. A primary would not be held simply because an ambitious, possibly demented, possibly wealthy individual decided to throw his or her hat into the ring.

Any steps to strengthen the parties as institutions would be desirable on the same grounds. Lack of party cover in the United States means that elective officeholders find it hard to take tough decisions partly because they lack safety in numbers. They can seldom, if ever, say to an aggrieved constituent or a political-action committee out for revenge, "I had to vote that way because my party told me to,"or even "I had to vote that way because we in my party all agreed that we would." Lack of party cohesion, together with American voters' disposition to vote for the individual rather than the party, means that congressmen and senators are always in danger of being picked off one by one.

WHAT might be done to give both parties more backbone? Clearly, the parties would be strengthened -- and elective officeholders would not need to raise so much money for their own campaigns -- if each party organization became a major source of campaign funding. In the unlikely event (against the background of chronic budget deficits) that Congress ever gets around to authorizing the federal funding of congressional election campaigns, a strong case could be made for channeling as much of the money as possible through the parties, and setting aside some of it to cover their administrative and other ongoing costs.

The party organizations and the nexus between parties and their candidates would also be strengthened if it were made easier for ordinary citizens to give money to the parties and for the parties to give money to their candidates. Until 1986, when the program was abolished, tax credits were available for taxpayers who contributed small sums to the political parties. These credits could be restored. Larry J. Sabato, a political scientist at the University of Virginia, has similarly suggested that citizens entitled to a tax refund could be allowed to divert a small part of their refund to the party of their choice. Such measures would not, however, reduce candidates' dependence on donations from wealthy individuals and PACs unless they were accompanied by measures enabling the parties to contribute more generously to their candidates' campaigns. At the moment there are strict legal limits on the amount of money that national or state party organizations can contribute to the campaigns of individual candidates. The limits should be raised (and indexed to inflation). There is even a case for abolishing them altogether.

All that said, there is an even more straightforward way of reducing incumbents' dependence on campaign contributors. At present incumbents have to spend so much time raising funds because the campaigns themselves are so expensive. They could be made cheaper. This, of course, would be one of the effects of making U.S. elections less numerous and less frequent than they are now. Another way to lower the cost of elections would be to provide candidates and parties with free air time on television and radio.

CLEARLY, the idea of term limits also needs to be taken seriously. After all, if American politicians are excessively vulnerable at the moment, one way of rendering them invulnerable would be to prevent them from running for re-election -- no impending election contest, no need to worry overmuch about the voters.

As is evident, much of the actual campaigning in favor of term limits takes the form of ranting -- against big government, against Washington, against "them," against taxes, against the deficit. Much of the rhetoric of term-limiters is sulfurous, and their principal motive often seems to be revenge. They claim that members of Congress are insufficiently responsive to their constituents, when the evidence suggests that, on the contrary, they are far too responsive. The term-limits movement is of a piece with previous outbursts of frustrated American populism, including the Know-Nothing movement of the 1850s -- an essay, as one historian has put it, in "the politics of impatience."

Nevertheless, there is an alternate case for term limits, based not on American politicians' alleged lack of responsiveness to the voters but on their alleged overresponsiveness to the voters and interest groups in order to secure their own re-election. The most persuasive and subtle advocate of this line of argument is the political commentator George F. Will. His goal, Will says partway through his book (1992), "is deliberative democracy through representatives who function at a constitutional distance from the people." He reiterates the point about distance in his final paragraphs: "Americans must be less demanding of government. They must give to government more constitutional space in which to think, more social distance to facilitate deliberation about the future."

The case for giving American politicians more space and distance is undoubtedly a strong one, but assuming these objectives are desirable, it is still not clear that term limits are a suitable means for achieving them. Three questions arise. Would term limits achieve the desired objectives? Would they do so at an acceptable cost in terms of other American goals and values? Might the desired objectives not be better achieved by other means? The first question is strictly empirical. The other two mix the empirical and the moral.

One way in which term limits might promote deliberation is by causing some incumbent legislators -- namely those serving out their final term under term limits -- to think, speak, and vote differently from the way they would have thought, spoken, and voted if they had been eligible and running for re-election. In addition, for term limits to affect the behavior not just of certain individuals but of Congress as a whole, it would be necessary for any given Congress to contain a significant number of these final-term members. In other words, congressional lame ducks would have to quack differently from other ducks, and there would have to be a fair number of them on the pond.

It is impossible to be sure, but it seems unlikely that term limits would have significant effects along these lines. In the first place, existing research (along with most human experience)suggests that a final-term congressman or senator, after eleven or twelve years on Capitol Hill, would be unlikely to alter his pattern of behavior in any radical way. He might send out fewer pieces of franked mail and make fewer trips back home, but he would probably not execute many U-turns in the way he spoke and voted. In the second place, although the proportion of senators who would be in their final term under term limits would normally be large (possibly half if senators were restricted to two terms), the proportion of lame-duck congressmen would normally be much smaller (an average of sixty to seventy out of 435 if House members were limited to six terms). The cumulative impact of the lame ducks would thus be much greater in the Senate than in the House, and in both houses it would probably be felt mainly at the margins (though of course the margins can, on occasion, be important).

But those who advocate term limits in fact build very little of their case on the expected future behavior of lame ducks. Rather, they are seeking to create a wholly new class of elected representatives. George Will holds out the prospect that mandatory term limits would have the effect of replacing today's political careerists with noncareerists -- in other words, of replacing today's ducks with creatures more closely resembling swans. The new legislators, because they were not careerists, would not be driven by the need to secure re-election, and for that reason they would be more likely to concern themselves with the national interest. Also because they were not political careerists, they would be more likely to have some personal, hands-on understanding of America and its real concerns.

The prospect is undoubtedly attractive. But is it realistic? Would term limits in fact diminish the number of careerists and produce legislators who were more national-minded and disinterested?

The most important difficulties with Will's hypothesis are twofold. One is that modern politics at all levels, local and state as well as national, is an immensely time-consuming, energy-consuming activity that demands enormous commitment from those who are attracted to it. Legislative sessions are long, constituents' demands are exigent, policy problems are increasingly complicated. As a result, politics all over the world, not just in the United States, is becoming professionalized. Men and women in all countries increasingly choose a political career at an early age and then stick with it. It seems likely that even under term limits the great majority of congressmen and senators would be drawn from this professional political class, which has not only the commitment to politics but also the requisite patience, skills, and contacts. To be sure, people's political careers would take a different shape; but they would still be political careers.

The other difficulty is the reverse of the first. Just as politics is becoming more professionalized, so is almost every other occupation. As many women in particular know to their cost,

it is becoming harder and harder to take career breaks -- those who jump off the ladder in any profession find it increasingly hard to jump back even to the level they were on when they left, let alone the level they would have attained had they stayed. For this reason it is hard to imagine that many upwardly mobile corporate executives or successful professionals or small-business owners would take time off to serve in Congress on a citizen-legislator basis. The citizens who sought to serve on this basis would probably be largely the rich and the old.

DESPITE their differences, term limits and the proposals offered here have in common the fact that they seek major changes in America's political institutions -- in some cases involving an amendment to the Constitution. But of course America's politicians are free to alter the way they behave in the context of the country's existing institutions. They can try to find alternative ways of insulating at least some aspects of policymaking from the intense campaigning and electioneering pressures they are now under.

Short of taking difficult issues out of electoral politics altogether, there are tactics that could be employed. Most of them are out of keeping with the contemporary American preferences for direct democracy, high levels of political participation, and the maximum exposure of all political processes to the public gaze; but that is precisely their strength. Bismarck is reputed to have said that there are two things one should never watch being made: sausages and laws. Both should be judged more by the end result than by the precise circumstances of their manufacture.

One available tactic might be called "the collusion of the elites." There may be occasions on which the great majority of America's politicians, in both the executive and legislative branches, are able to agree that an issue is of such overriding importance to the nation that it must be dealt with at almost any cost; that the politicians involved must therefore be prepared to set aside their ideological and other differences in the interests of finding a workable solution; and that having found a solution, they must stick together in presenting it to what may well be a disgruntled or even hostile electorate. In order to be successful, the collusion-of-elites tactic requires not only a substantial degree of bipartisanship (or, better still, nonpartisanship) but also unusually small teams of negotiators, complete secrecy (not a single ray of "sunshine" must penetrate the proceedings), and the presentation to Congress and the public of a comprehensive, all-or-nothing, take-it-or-leave-it proposal.

The number of occasions on which politicians will be prepared to set aside their ideological differences and pool their political risks in this fashion will inevitably be small. There were no signs that such a spirit might prevail when President Clinton and the Republican majorities in Congress wrangled over how to cut the budget deficit last winter. But there have been instances of the successful collusion of elites, even in relatively recent times.

One of them occurred in 1983, when representatives of President Reagan and the two party leaderships on Capitol Hill colluded to save the Social Security system, which at that time was in imminent danger of bankruptcy. Paul Light's classic account of the 1983 Social Security reform, (1985),is in effect a case study of how to conduct collusion-of-elites politics and of the circumstances in which it may succeed. The so-called Gang of Seventeen that was originally put together to hammer out a deal (and was later reduced to a Gang of Nine) excluded all the more-extreme ideologues and met in circumstances of great secrecy, even using, according to one participant, "unmarked limos."

Of the Gang of Seventeen's activities, Light writes,

The meetings seemed to inaugurate a new form of presidential-congressional government. The meetings were secret. There were no minutes or transcripts. All conversations were strictly off the record. The gang was free to discuss all of the options without fear of political retaliation. It . . . [existed] completely outside of the constitutional system.

Ultimately, as Light relates, the "secret gang built a compromise, wrapped it in a bipartisan flag, and rammed it through Congress. There was no other way to move. It was government by fait accompli." It was also successful government -- and none of the participants suffered electoral damage.

Another possible tactic, with many similarities to the collusion of elites, might be called "putting it into commission." If taking tough decisions is too risky politically, then get someone else to take them. If someone else cannot be found to take them, then make someone else appear to take them. The someone else need not be but usually will be a bipartisan or nonpartisan commission of some kind.

Such a commission, the National Commission on Social Security Reform, played a role in the passage of the 1983 act, but an even better example was the procedure adopted by Congress in 1990 for closing redundant military bases. Earlier practice had been almost a caricature of Congress's traditional decision-making process. The Secretary of Defense would propose a program of base closures. Senators and congressmen would immediately leap to the defense of targeted bases in their home states or districts. They of course had the support of their colleagues, who were threatened with or feared base closures in their home states or districts. Almost never did anyone manage to close any bases.

Realizing that the process was absurd and that huge sums of taxpayers' money were being wasted in keeping redundant bases open, Congress decided to protect itself from itself. It established the Defense Base Closure and Realignment Commission, which employed an extraordinarily simple formula. The Defense Secretary every two years published a list of the bases he proposed to close, together with a statement of criteria he had used in compiling his list. The commission then examined the list in light of the criteria, held public hearings, and recommended a modified list (with additions as well as deletions) to the President. The President was obliged to accept the commission's list as a whole or reject it as a whole. If, as invariably happened, he accepted it, Congress could intervene only if within forty-five legislative days it passed a bill overriding the President's decision and rejecting the whole list. This it never did.

The formula was a near miracle of voter-proofing. Members of Congress were left free to protest the closure of bases in their home districts or states, but the decision was ultimately taken by the President, who could nonetheless ascribe all blame to the commission, and all Congress had to do for the President's decision to take effect was to do nothing. In the event, hundreds of bases were closed and millions of dollars saved, but no member of Congress ever had to vote -- and be seen by his constituents to be voting -- in favor of closing a base near home. Beyond any question the results were in America's national interest.

It is not wholly fantastic to suppose that the President in odd-numbered years might, on the basis of advice received from a bipartisan commission, announce a list of "program eliminations," which Congress could countermand only by voting to reject the list as a whole. Presidents would probably prefer to put forward such lists at the beginning of their first term in office -- or at any time during their second term -- when they, at least, were not up for re-election.

A final tactic, which could also be adopted without major institutional change, might be described as "thinking big." Proposals that are put forward on a piecemeal basis can also be opposed, and in all probability defeated, on a piecemeal basis. In contrast, large-scale, broad-based proposals may have a better chance of success simply by virtue of their comprehensiveness. They can provide something for everyone -- conservatives as well as liberals, deficit cutters as well as program defenders, residents of the Sun Belt as well as of the Rust Belt. Gains as well as losses can be broadcast widely. The 1983 Social Security reform and the 1986 tax reform were certainly "big thoughts" of this general type. So, in its way, was the recent base-closure program.

Tactics like these -- the collusion of elites, putting issues into commission, and thinking big -- all have their virtues, but they also suffer from being tactics in the pejorative as well as the descriptive sense. At bottom they are somewhat cynical devices for getting around the real difficulty, which is the hyper-responsiveness of American politicians that is induced by their having to run scared so much of the time. Although it would be harder, it would be better over the long term to confront this problem directly and try to bring about at least some of the fundamental institutional changes proposed here. The American people cannot govern themselves. They therefore need to find appropriate means of choosing representatives who can do a decent job of governing on their behalf, and that means giving the people's representatives space, time, and freedom in which to take decisions, knowing that if they get them wrong, they will be punished by the voters. In twentieth-century America the airy myths of agency democracy are precisely that:myths. What America needs today, though it does not seem to know it, is a more realistic and down-to-earth form of division-of-labor democracy.

Illustrations by Richard McGuire

The Atlantic Monthly; January 1997; Running Scared; Volume 279, No. 1; pages 41-61