Retrospective on American Economic Policy in the 1990s

Jeffrey Frankel and Peter R. Orszag
Peter R. Orszag Vice Chairman of Investment Banking, Managing Director, and Global Co-Head of Healthcare - Lazard

November 2, 2001

This paper is based upon the Introduction to American Economic Policy in the 1990s, which will be published in the Spring of 2002 by the MIT Press. The book is the outcome of a conference held at Harvard’s Kennedy School of Government in June 2001, which brought together leading policy-makers and economists with the goal of providing a preliminary history of U.S. economic policy-making during the past decade. The book is divided into 14 chapters, each examining a different area of economic policy: Monetary policy, fiscal policy, tax policy, international finance and crises in emerging markets, trade policy, information technology, industrial organization, energy policy, environmental policy, labor and education, poverty and welfare, health care and tobacco policy, Medicare policy, and the policy-making process. This paper, which is based upon the Introduction to the volume, includes a discussion of distinguishing characteristics of “Clintonomics” in historical perspective, a cataloguing of reasons for the good economic performance enjoyed by the country during the years 1993-2000, an explanation of the difficulties of apportioning credit for such outcomes, and some thoughts on the under-appreciated perils of excessive transparency in government.

Decades take on characteristics in our minds. The 1970s make us think of oil shocks and Watergate, and the 1980s of supply-side economics and the end of the Cold War. How will historians—or economic historians, at any rate—remember the 1990s?

It is not too soon to predict. The New Economy and Greenspan. The Internet, dotcoms, and IPOs. Monica and the West Wing. The cell phone and the Palm Pilot. Policy-wonk slogans, some carefully planned and some unplanned: “The Economy, Stupid”; “The era of big government is over”; and “Save Social Security First.” The Contract with America and the Gingrich revolution. NAFTA and the “giant sucking sound.” The Asian crisis. Tobacco litigation and the Microsoft trial.

The two Clinton terms occupied most of the 1990s. Unquestionably, history will remember this period as a time of strong economic performance. Between 1993 and 2000, the United States exhibited the best economic performance of the past three decades. In 2000, the U.S. economic expansion surpassed in length the expansion of the 1960s, and thus became the longest on record. During Clinton’s second term, real economic growth averaged 4 ½ percent per year, and unemployment fell to four percent, the level that had been specified by the Humphrey-Hawkins legislation three decades earlier to be the goal of national policy. During the early 1990s, economists would have considered these outcomes wildly unattainable.

Strong growth and low unemployment were particularly remarkable because they were accompanied by structural budget surpluses and low inflation. Long expansions have historically been fueled in large part by expansionary fiscal or monetary policies, with the result that, by their six-year mark, debt ratios and inflation rates had risen to high levels, sowing the seeds of a subsequent contraction. Furthermore, productivity growth typically slows as an expansion matures. The 1990s American boom, to the contrary, was led by private-sector spending and private-sector employment. And for the first time in three decades, productivity growth rose substantially in the late 1990s, despite the length of the expansion at that point. The cause of this acceleration in productivity is still the subject of debate. (Section II of this paper briefly catalogs the likely contributing factors to the strong performance of the U.S. economy during the 1990s.)

To be sure, some observers found cause for concern in various aspects of economic performance during the 1990s. Personal bankruptcies climbed, the personal saving rate plummeted (as measured in the national income statistics), the trade deficit expanded dramatically, and the stock market may well have become substantially overvalued. Overall, however, U.S. economic performance during the 1990s was outstanding.

The reader may wonder if the subject of American Economic Policy in the 1990s is the years 1991-2000, or the Clinton Administration per se (1993-2000). The answer is that it examines both. The period 1990-2000 approximates a complete business cycle. The economy was close to potential in early 1990, before Iraq invaded Kuwait. The recession of 1990-91 and the continued high unemployment rates over the two years after the recession had technically ended—which initially tarred the expansion with the label “jobless recovery”—are widely thought to have cost George Bush the 1992 election. The years 1990-92 thus set the stage for the years 1993-2000. This is true in macroeconomic terms, and as well in many of the specific areas of policy. By the year 2000, the economy had returned to approximately the same point in the business cycle it had occupied in 1990.

Incidentally, the pattern of recession in the first few years of the decade, followed by steady growth in the remainder, repeated the pattern of the 1960s, 1970s, and 1980s.

I. Clintonomics

Arguably, the Clinton Administration had an overarching vision. It sought to adopt some of the pro-market orientation associated with the ascendancy of the Republicans in the 1980s, and marry it with traditional Democratic values such as concern for the environment and a more progressive income distribution. This combination made sense from an economic viewpoint. Standard textbook economics says that environmental externalities, anti-competitive practices, and other market failures or social goals require government involvement, while everything else should be left to the market.

Yet it produced tension within the Administration and with traditional Democratic constituencies. The “soft-hearted” and “hard-headed” components of the strategy were often in conflict, as evidenced by the debate over welfare reform, corporate responsibility, trade and the environment, and deficit reduction instead of aggressive public investments.

The 1993 budget package offers perhaps the most vivid example of the underlying vision of—but also the tensions inherent in—the Clintonomics approach. The 1993 budget agreement was politically difficult. Democrats feared that it might cost them their positions in Congress, a fear that eventually turned out be accurate in some cases. By the end of the 1990s, however, Democrats had largely embraced what one pundit called the “progressive fiscal conservatism” of the 1993 package. Such progressive fiscal conservatism combines modest attempts at redistribution (the progressive component) and budget discipline (the fiscal conservative component). Thus the 1993 package included significant spending reductions and tax increases. But it concentrated the tax increases on upper-income taxpayers, while substantially expanding the Earned Income Tax Credit, Head Start, and other government programs aimed at lower earners.

The experience of the 1990s eventually convinced even many liberal-leaning policy-makers that the progressive fiscal conservatism at the heart of Clintonomics offered an auspicious alternative to more traditional Democratic fiscal policies, which had emphasized government spending and public investment rather than fiscal discipline and private investment. One explanation for the shift may be that fiscal discipline had traditionally been presented as offering long-term benefits in exchange for short-term costs, a calculus that is not usually appealing in a political context. But fiscal discipline appeared to promote strong economic performance even in the short run. The short-run stimulative effect of long-run deficit reduction packages with accommodating monetary policy brought potential benefits for lower-income workers and others disproportionately affected by cyclical fluctuations in the economy. It thus relieved some of the tension inherent in progressive fiscal conservatism. Since economic benefits of fiscal discipline appeared to manifest themselves even in the short run, sound economic strategy was also a viable political strategy. (President Clinton urged supporters to “vote Democratic so that they could live like Republicans.”)

Despite the conflicts in Clintonomics, or perhaps partly as a reflection of them, the combination of fiscal discipline and progressive policies made sense from a political viewpoint. If Democratic leaders in the 1970s and 1980s had often lost office by deviating too far to the “left” of the median American voter, and if Republican leaders since 1996 have often lost by deviating too far to the “right,” Bill Clinton’s policies were more in tune with the preferences of the median voter. This positioning undoubtedly contributed to his generally high popularity ratings and electoral success.

Indeed, critics of President Clinton often argued that he was too much in tune with the median voter—that he allowed polls and public opinion to determine policy positions. This is not to say that he never took positions that differed from what he thought the public wanted. The Mexican bailout of January 1995 provides perhaps the most vivid example. Management of this and subsequent financial crises in emerging markets may have turned out, quite unexpectedly, to be the second most important accomplishment of the Administration, after elimination of the budget deficit.

Critics of Clintonomics also argued that it was consumed with “nano”-level or small-bore policies, such as targeted tax cuts or minor changes to existing programs. The emphasis on small-bore policies reflected several fundamental forces.

First, the defining events in the Administration’s first two years were the enactment of the 1993 budget agreement and the failure to enact the Administration’s Health Security Act. The fiscal discipline created by the 1990 budget agreement and extended by the 1993 budget deal substantially reduced the scope for policy-makers to create or expand programs in a dramatic fashion. As a result, the natural tendency of policy-makers to “do” something, or at least to be viewed as doing something, was subsequently channeled into small-bore activities. The policies of the later years are therefore partly a reflection of the successful macro-policy of 1990 and 1993. Furthermore, the budget agreements also created pressure for emphasizing targeted tax cuts rather than spending programs within the sphere of small-bore policies. The budget agreements of 1990, 1993, and 1997 imposed tight limits on discretionary spending, but allowed tax cuts as long as they were coupled with “offsets” (reductions in entitlement spending or increases in other taxes). The budget rules, combined with the political economy favoring tax cuts rather than spending, biased policy-makers toward narrowly targeted tax cuts. Targeted tax credits were thus used in areas such as higher education (the Hope scholarship and the lifelong learning credit) and environmental research and development efforts (the proposed tax credit for high-mileage vehicles) that would have traditionally been addressed through spending programs.)

Second, the failure of the Health Security Act in 1994 convinced policy-makers in the Clinton Administration that it was wiser to attempt smaller changes than larger ones, since the larger ones were unlikely to be passed. Within health care, the upshot was a focus on providing health insurance coverage to women and children first. This approach continued a trend that preceded the Clinton Administration: Throughout the mid- and late-1980s, policy-makers had successively loosened access to the Medicaid program for poor children and mothers.

The basic continuity of policy across Administrations extends into many other areas. Often a new president, notwithstanding attacks on his predecessor during the election campaign, discovers not long after taking office that there turn out to have been good reasons for the way some things were done.

Third, the Republican takeover of Congress in 1994, the so-called Gingrich revolution, affected Administration strategy. In the face of Republican opposition to major Democratic initiatives, the Administration was forced to adopt more modest goals in many areas. The major domestic exception was welfare reform. Welfare reform was opposed by many traditional Democrats and may therefore support the general point that the political split between the executive and legislative branches precluded dramatic policy shifts toward “Democratic” policies; given Republican control of Congress, enactment of significant changes seemed possible only when the executive branch was willing to adopt a “Republican” policy. Some Republicans complained that Clinton was co-opting their issues. Moving to the political center on particular issues is, however, an option that is available to anyone.

This paper focuses primarily on the executive branch. The executive branch, however, does not exist in a vacuum. The constraints imposed by, and policies favored by, the Congress strongly condition executive-branch policy-making.

Fourth, small-bore initiatives often represented sound political strategy in the context of larger budget and programmatic reductions. For example, it was estimated that the Balanced Budget Agreement of 1997 would reduce Medicare spending by $385 billion over 10 years. Yet it also included small expansions in certain benefits—such as coverage for prostate cancer screening, annual rather than bi-annual mammograms, and annual pap smears—that helped to generate support for the overall package. The Administration’s presentations tended to emphasize the expansions almost as much as the overall reductions, despite the dramatically different scales of the two policies. Whether the small-bore policies were necessary in order to make the overall package politically acceptable, and therefore whether the small-bore policies played a critical economic as well as political role, is an important question. In some cases, the complexity created by the small-bore policies was arguably the cost necessary for enacting a broader set of economically beneficial policies.

Finally, it is important to emphasize that many individual policy questions properly merit at most a small-scale government response. The cumulative effect of even small-bore policies can be significant. In information technology, for example, the Clinton Administration adopted an array of initiatives. The cumulative effect of those initiatives represented sound policy with regard to the Internet and other technological developments. In other areas, similarly, relatively small annual expansions in programs cumulated to a significant increase over the eight years of the Administration. For example, funding for Head Start more than doubled in constant dollars between 1992 and 2000, while enrollment rose by 38 percent.

Leaving aside the question of the Clinton Administration’s focus on small-bore policies, some analysts argue that the Administration could not take credit for the success of the U.S. economy during the 1990s because that success was due to factors other than the Administration’s own policies. We turn now to the various factors affecting U.S. performance during the 1990s.

II. Explaining U.S. Economic Performance in the 1990s

What are the reasons for the outstanding U.S. economic performance of the 1990s? A variety of short-term, medium term and long-term factors played a role. The short-term factors were luck, the medium term factors were skill, and the long-term factors are ongoing favorable structural trends in the U.S. economy.

Short-term factors: Temporary good luck on prices

In the 1990s, relative prices for computers and health care fell significantly. Until 1999, world prices for oil also remained low. And U.S. import prices were low generally, due both to the appreciation of the dollar in the second half of the 1990s, and to deflation in some partner countries, particularly in East Asia. All these factors put downward pressure on inflation, and thus prevented overheating despite rapid growth in the economy. Measured consumer price inflation was also temporarily restrained by revisions to the price index. Thoughtful observers knew these trends were unlikely to continue for long, and indeed some of the trends did come to an end at the close of the decade. But even if one adjusted the inflation rate for such short-term factors, the record of price stability at a time of high employment and growth was still impressive.

Medium-term factors: Good macroeconomic policy

The skillful exercise of macroeconomic policy, both fiscal and monetary, contributed significantly to the strong economic performance of the 1990s. Three key fiscal policy turning points included the 1990 budget agreement, the 1993 budget agreement, and the 1998-2000 preservation of the emerging unified budget surpluses for debt reduction. The 1990 budget agreement represented an important step toward fiscal discipline, and included marginal tax rate increases that imposed substantial political costs on President George Bush. In 1993, the Clinton Administration built upon and strengthened the 1990 deficit reduction effort, by further raising taxes on the highest-income taxpayers and reducing spending relative to the budget baseline. These efforts, along with strong growth in income and therefore tax revenue, replaced intractable budget deficits with substantial surpluses by the end of the decade. Indeed, in the last years of the decade, the principal fiscal accomplishment of policy-makers was what they didn’t do—dissipate the accruing surpluses in the form of substantial tax cuts or program expansions—rather than what they did do.

Whatever the explanation, the Federal government’s movement from deficit to surplus accounted for all of the improvement in net national saving between 1993 and 2000. This additional saving restrained long-term interest rates, thereby boosting private-sector domestic investment. As noted above, the experience of the 1990s highlights the economic stimulus that can be provided by long-run deficit reduction accompanied by monetary ease—a phenomenon that had been discounted by many macroeconomists before the 1990s.

Fiscal responsibility ultimately proved to be more politically popular than could have been expected from President Bush’s experience in 1990. The surprising popularity that Ross Perot achieved in 1992 with the budget issue pointed the way. The public had apparently learned from the experience of the 1980s to be skeptical of politicians selling snake-oil tax cuts. A decade ago it looked as if the country might be doomed to a pro-cyclical fiscal policy of pursuing fiscal expansion when times were good and finding discipline only in time of recession, the point in the business cycle when raising taxes was least appropriate. The Clinton Administration’s clever “Save Social Security First” strategy in 1998 underscored the political viability of fiscal discipline even in an expansion, transforming what had previously appeared to be politically unattractive—the preservation of the unified budget surplus for debt reduction—into a politically successful strategy.

True, the long-term fiscal positions of Social Security and, especially, Medicare have not yet been fully addressed. Furthermore, it should be noted that the new Bush Administration succeeded in passing a significant tax cut in 2001. But despite an active Bush Administration effort to convince the public that the tax cut was needed, polls indicated that many Americans would rather have spent the money on paying off the debt and putting Social Security and Medicare on a sound footing than on a tax cut.

We turn now from fiscal policy to monetary policy. The Clinton Administration made two contributions to monetary policy. First, the elimination of the budget deficit allowed the Fed to lower interest rates. Second, the Clinton Administration’s monetary policy otherwise was simple to state: leave it to the Fed. Adhering to this policy is more difficult than it sounds. The political temptation is always strong to nudge the central bank toward an easier monetary policy: even if the monetary authorities don’t respond, the complaints give the Administration someone to blame in the future if the economy slows down. In addition, officials are naturally tempted to respond to press inquiries with statements that, while not intended to be critical, or even substantive, are nevertheless inevitably interpreted as second-guessing the Fed. With remarkably few exceptions, the Administration adhered to its self-imposed rule of silence.

The lack of Administration interference worked well because the Fed was skillful. Even if, in retrospect, the tightening of 1999-2000 may have gone one step too far, Chairman Greenspan’s record overall during the decade was quite impressive. The truly remarkable feature of the 1990s was not just its low inflation, but its low and steady inflation. At least some of the credit for this stability must belong to Greenspan. Like Paul Volcker before him, Greenspan followed a tight monetary policy early in his term, established a reputation for discipline, and was thereby able to take a more moderate stance during the remainder of his term. His forbearance during 1995-1998, even as growth and employment exceeded levels previously considered inflationary, was a gamble; but from many perspectives it appears to have been a wise gamble and an important component of the expansion’s longevity.

Perhaps the simplest overall conclusion one can draw from the 1990s is that the U.S. economy runs relatively well given a little luck and the avoidance of major macroeconomic policy mistakes.

Long-term factors

Many of the most fundamental factors in explaining U.S. economic performance during the 1990s stretch back over two decades or more:

  • Deregulation. The U.S. economy has long been less regulated than most other industrialized economies. But the past 25 years have witnessed important further steps toward deregulation. The deregulation trend began during the Carter Administration, in trucking, airlines, natural gas, and banking. During the Reagan Administration, deregulation was extended to the telecommunications sector. More recently, further deregulation has occurred in the electricity market, and market-friendly environmental regulation, such as in the sulfur dioxide permit program, has been expanded. Some of these deregulation efforts have faced bumps in the road, particularly banking and electricity. Nevertheless, the overall effect of deregulation has been to make the U.S. economy more efficient in the long run. The fundamental continuity of policy across Administrations in these areas also highlights a theme mentioned above: despite the drama of changes in Administrations, policy does not shift nearly as much as one would imagine.

  • Globalization. With important exceptions, the U.S. has had a basic free trade orientation since World War II. The ratio of trade to GDP has more than tripled since the middle of the 20th century and now stands at 26 percent (imports plus exports, including services, which have increased especially rapidly). Economic theory tells us that trade improves economic performance. This is true both of old trade theory (classical “comparative advantage”) and of new trade theory (which allows for changing technology, increasing returns to scale, and imperfect competition). Perhaps more convincingly, the statistical evidence also tells us that openness contributes to growth. Exports grew rapidly, a major selling point for free trade policy. Even the increases in imports and in the trade deficit during the 1990s, though politically unpopular, were a useful safety valve during the strongest phase of the U.S. expansion. They released pressure from rapidly growing domestic demand, pressure that would otherwise have shown up as higher inflation and interest rates.

  • Innovation. The third category of favorable long-run structural factors is innovation. Innovation can be further divided into three types: technology, competition and flexibility in goods and labor markets, and the public sector. Technological innovation, especially information technology, received much attention during the 1990s. Although the IT revolution was not the sole reason for strong U.S. economic performance in the 1990s, it certainly was a positive factor. The second type of innovation involves competitive markets. The United States has always had relatively competitive goods and labor markets, compared with Europe for example. But the last two decades have seen a further movement in this direction, including the initially unpopular corporate restructuring of the 1980s and the initially popular dot-com start-up firms of the 1990s. Anti-trust enforcement became more active during the Clinton Administration. The final category is innovation in the public sector. Public-sector reforms include reinventing government and defense reconversion, which have allowed previously low-productivity resources to be shifted to more productive uses, and welfare reform.

Apportioning credit

Short-term luck on the supply side, medium-term skill in macroeconomic policy management, and long-term favorable structural trends that were called by some a “New Economy”—this is a long list of factors. After compiling the list, is it possible to apportion the credit among the various factors? In particular, many observers and pundits are interested in how much credit for the strong economic performance of the 1990s should be given to the Clinton Administration’s policies. Unfortunately, providing a specific answer to this question is not possible, at least not in this book.

On the one hand, we often speak as if the perceived performance of the economy during a given period shows directly the virtues of the president and his team during that period. On the other hand, observers often note that the course of the economy in fact reflects exogenous factors to a greater extent than it reflects the actual personal strengths and weaknesses of the man who was president, with the implication that the contribution of the latter should be dismissed. It should be evident that the president in fact generally deserves a share of credit for what happens on his watch that is neither zero nor 100 percent. But it is more complicated than that.

A metaphor may offer insight. Imagine the president as the captain of a large ship. Whether the passengers in the hold experience rough seas or a calm ride depends on factors such as the weather and the construction of the ship to a greater extent than the skills of the captain. Nevertheless, helmsmen can be skilled or incompetent, and the difference affects the ride.

Disentangling causality, credit, and blame for either a smooth sail or a rough one is exceedingly complex. Things can go wrong through no fault of the leadership, or can go right for the wrong reason. If an unpredictable squall arises on a particular navigational route, it does not mean that the captain erred in choosing that route over another. Conversely, smooth sailing does not mean that he picked the best route. Thus it is not meaningful to parse out fractions of the total credit for the economy’s good economic performance in the 1990s to Bill Clinton, Alan Greenspan, information technology, or other forces.

An example illustrates the problems in parsing out credit for what happened (or did not happen). The White House sent Zhu Rhong Zhi away in the spring of 1999 without an agreement on Chinese accession to the World Trade Organization (WTO), even though China had made many important concessions (relative to what other developing countries have agreed to when joining the WTO, or what had been expected). Was this a mistake on the part of the White House? It seemed like a mistake over the subsequent months, as the U.S. business community reacted adversely, Zhu and other reformers were undercut politically in China, and the accidental bombing of the Chinese Embassy in Belgrade further worsened U.S.-China relations. In judging the wisdom of the spring 1999 decision, however, the key question is: Did the decision reflect a sensible balancing of the pros and cons at the time, given the information that was available? Such an evaluation would have to include the political constraints on the White House, which are often conveniently forgotten in discussions of this type. For example, the White House had to consider the reactions an agreement on Chinese accession would elicit from Senators opposed to such an agreement, and the reaction (or absence thereof) from businessmen favoring it. It would have been of little use to sign a good agreement that could not have been ratified by the Senate. Complaints from the business community about the lack of an agreement were in the end important in persuading Congress to vote for it.

Too often, we evaluate the captain’s performance only by conducting polls among the passengers in the hold. It is indeed worth knowing whether the passengers feel they are experiencing a smooth ride. But journalists, academics, and other opinion leaders have the ability to go on deck and observe for themselves what sort of job the captain is doing. It is their responsibility to do so, and they owe it to the passengers to report on what they see. In undertaking such evaluations, five different levels of performance are relevant:

  • Personal characteristics: How do the president’s abilities compare to others who have held the office or could have held the office? Relative criteria include stamina, capacity to absorb and synthesize information, ability to make decisions, speaking ability, ability to communicate warmth and inspire confidence in common people, honesty, skill at picking good staff, and so on. In the specific case of economic policy, another relevant attribute is whether the president is economically literate and numerate.

  • Process: How well does the president’s administration function? When it is reported in the press that the president has done something, often it is in fact not he but his representatives who have done it. Of course it is appropriate to hold him, as a manager, responsible. But analytically we must realize that other random factors begin to enter at this second stage. To take but one example, if many of the president’s nominees have not been confirmed by the Senate, nor even considered by the Senate, the president may not be wholly to blame if the policy process then does not run smoothly.

  • Quality of policies implemented: How good are the spending programs, tax policies, regulatory decisions, international negotiations, and so forth that are implemented? In the face of Congressional opposition, it may be foolish to blame the White House for a budget or a trade measure if it is clear that it did its best to argue for some other better outcome but was stopped by the Congress.

  • Outcomes: What are the actual outcomes in the economy? What are the rates of growth, unemployment, and inflation? How strong is the budget position? The government may function well, and the policies chosen may be good, but exogenous shocks, like the oil price increases of the 1970s or some favorable developments in the 1990s, may be beyond its control.

  • Perceptions: What are the public’s perceptions of the president’s performance? Although economists are primarily interested in actual outcomes, much of the rest of society is interested in this final level. What are the president’s popularity ratings, what do polls say on aspects of his performance, what are the editorial writers and other opinion leaders saying, how well do he and his party do at the next elections, and what is the judgment of history?

to be sure, causality runs in all directions among these five levels of evaluation. For example, confidence about the direction in which the country is headed is often reflected in consumer confidence and stock market prices, which in turn can affect consumption, investment, and overall growth. But the central point is that, even though only the first level concerns the intrinsic abilities of the president, additional exogenous random factors can enter at each stage as we move down the chain to lower levels. This makes it hard to judge a president solely by perceived or actual outcomes: we can’t trace our way back up the chain from the perceived or actual outcomes to the president’s own characteristics.

Thus it is not the task of this paper or the forthcoming volume to evaluate the quality of presidential leadership in the 1990s. Rather the point is primarily to examine in detail the many policy decisions that were made—the third stage above. This examination will shed light on the links back to the first and second levels, as well forward to the fourth and fifth levels. But it does not answer, nor even attempt to answer, what specific share of the credit for good economic performance during this period should be given to the Clinton Administration.

III. A Note of Warning

We close this paper with a note of caution: future endeavors of the type embodied in American Economic Policy in the 1990s may be undermined by the rules governing disclosure of internal information. Advocating transparency has become akin to advocating motherhood and apple pie. For example, it is an important part of reforms urged on governments in developing countries, as well as on the International Monetary Fund itself, in the aftermath of the international currency crises of 1997-98. But there can be such a thing as too much transparency. To take an example, sunshine laws essentially forbid a group of governors of the Federal Reserve Board from informally discussing monetary policy with each other outside official meetings. What is gained by this?

Our experience in the Clinton Administration suggests that despite the general benefits of disclosure in promoting good government, the formal disclosure rules have become counter-productive. In particular, the Freedom of Information Act (FOIA) and the rules governing Congressional subpoena powers sometimes discourage executive-branch policy-makers from writing things down. Such written material is too often subject to Congressional inquiry, discovered through the FOIA process, or otherwise proven to be politically problematic.

The intent of these rules is to expose the decision-making process to scrutiny. Such scrutiny is generally beneficial. But the ultimate effect of the rules may be to expose decisions to less scrutiny, because of the response of policy-makers to the incentives posed by the rules. It is not only the policy-making process that suffers as a result; historians will also find it much more difficult to discern what actually happened and why without the benefit of an accurate paper trail. Furthermore, compliance with the current system consumes an inordinate amount of policy-makers’ time.

The balance between public disclosure and internal candor has thus, in our opinion, tipped too far toward the former in the United States. For example, top Administration economists spent substantial amounts of time combing through files in response to Congressional subpoenas, looking for internal background memoranda on climate change. Such internal memoranda, which explored potential policy options and evaluated the pros and cons of different approaches, were essential to an honest internal policy process. But some were also politically sensitive. The benefits of transparency in such cases are not entirely clear: What specific public benefit (as opposed to political benefit to the Administration’s opponents) arises from publicly releasing the CEA’s internal analyses of potential Administration policies, including those not chosen by the Administration? How should one weigh any such benefits against the costs of the associated incentive to avoid writing things down?

In our view, it would be beneficial to provide more protection for pre-decision internal analyses. Put simply, the Administration should be expected to provide rigorous analysis of the option it finally chose, but should be able to protect—at least temporarily—materials regarding the options not chosen as well as internal views of the pros and cons regarding the chosen option. For such politically sensitive internal analyses, disclosure could be undertaken after some suitable period of time had elapsed, thereby limiting the potential political costs but still imposing discipline on policy-makers (who know that their internal analyses may eventually see the light of day).

This model would mimic the policy of the Federal Reserve Board, which releases the minutes of its confidential meetings only after some delay.

This volume benefited from the policy-making paper trail on numerous issues. The danger is that future volumes of this type may not benefit from such access, if disclosure rules discourage policy-makers from committing their thoughts and opinions to paper. The ultimate cost will then not be borne by future Administrations, but rather by those seeking lessons to be learned but finding only incomplete or inaccurate histories.