Fall 2006 Vol. 30, No. 3 Delta Dawning Gulf Coast Eff orts Look Beyond Recovery Toward Lasting Renewal —By John Godges

Steady Under Fire: All-Volunteer Force Proves Its Resilience, So Far —By Bernard D. Rostker

Road Less Traveled: Four Developing Countries Blaze New Trails to Better Health —By Thomas W. Croghan Get the Big Picture

Editor-in-Chief RAND Review covers the big issues with an John Godges Associate Editors eye for the important details. Gordon Lee Paul Steinberg Assistant Editor Miriam Polon Proofreaders Peter L. Hoffman Kelly Schwartz Art Director Ronald Miller Designer Eileen Delson La Russo Production Editor Todd Duft Circulation Christine Troncoso Web Producer Jason Walkowiak

Editorial Board Timothy Bonds, Cynthia Cook, David Egner, Jonathan Grant, James Hosek, James Kahan, Iao Katagiri, Kevin McCarthy, Elizabeth McGlynn, Michael Rich, K. Jack Riley, Shirley Ruhe, Mary Vaiana

RAND Board of Trustees Ann McLaughlin Korologos (chairman), Jerry I. Speyer (vice chairman), Carl Bildt, Harold Brown, Frank C. Carlucci, Lovida H. Coleman, Jr., Robert Curvin, Timothy F. Geithner, Pedro Jose Greer, Jr., Rita E. Hauser, Karen Elliott House, Jen-Hsun Huang, Paul G. Kaminski, Bruce Karatz, Philip Lader, Arthur Levitt, Bonnie McElveen-Hunter, Lloyd N. Morrisett, Ronald L. Olson, Paul H. O’Neill, Michael K. Powell, Donald B. Rice, James E. Rohr, James F. Rothenberg, Ratan N. Tata, James A. Thomson, Marta Tienda

RAND Review is published periodically by the RAND Corporation, a nonprofi t institution. The mission of the RAND Corporation is to help improve policy and decisionmaking through research and analysis. Opinions are those of the authors and do not refl ect positions taken by RAND, its board, or its clients.

LETTERS Send letters to: Editor, RAND Review, 1776 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138, or email [email protected].

PERMISSIONS For information about using mate- rial published in RAND Review, visit www.rand.org/ publications/permissions.html.

COPIES For hard copies of RAND Review, call (877) 584-8642 (toll free) or (310) 451-7002 (outside the United States); send a fax to (412) 802-4981; or email [email protected].

REPORTS Most reports cited in these pages may be purchased. To place credit card orders, contact RAND at the phone numbers or email address listed immediately above. National Book Network (NBN) also car ries selected RAND titles. Call NBN at (800) 462-6420.

ABSTRACTS Selected abstracts of RAND documents For previous editions and free online subscriptions, visit are available at www.rand.org/Abstracts. www.rand.org/publications/randreview. © Copyright 2006 RAND Corporation RAND: OBJECTIVE ANALYSIS. EFFECTIVE SOLUTIONS. Message from the Editor

One Eye on the Past, Fall 2006 Vol. 30, No. 3 One Eye on the Future 4 News Any organization that’s been around for nearly 60 years • “Other” war now key has a right to tout the enduring value of its work—and • “Backdoor” security at forefront a responsibility to do so when the lessons of the past • “Prohibited” buyers of ammunition buying could be overlooked to the detriment of the present. Any • “Comprehensive” school reform incomplete organization that plans to be around for another 60 years • “Burdensome” immigrants using little health care should harbor a doubly vigorous skepticism toward the • “Behavioral” treatment standards inappropriate present way of doing things, especially when clear-cut opportunities for improvement arise. This issue of RAND 10 Perspectives Review offers examples of both. Shifting Course A news item and an editorial recall lessons that have Th oughts on the Future of American Foreign Policy been corroborated throughout 45 years of counterinsur- gency research. When the Vietnam War ended 30 years Steady Under Fire ago, many policy experts believed that counterinsurgency All-Volunteer Force Proves Its Resilience, So Far theories and practices had passed their useful shelf lives 12 By Bernard D. Rostker and that the real action lay in understanding the Cold War face-off between the superpowers. Yet in Iraq and —Three Decades Afghanistan today, applying the lessons of counterinsur- Centerpiece gency is long overdue. 16 of Experience with the The 30-year history of the all-volunteer U.S. military All-Volunteer Force Have force is another example of how lessons often need to be Charted a Course to Success relearned. As outlined in a feature story and centerpiece, cutbacks in military pay and recruiting have often been imposed at the worst times. The confl icts in the Middle COVER STORY East today have placed an unprecedented burden on the 18 Delta Dawning all-volunteer force, leaving little margin for further rep- Gulf Coast Eff orts Look Beyond Recovery etition of the historical error. Toward Lasting Renewal As for opportunities for improvement, health research- By John Godges ers have found reason to doubt some long-held assump- tions about the best ways to improve health in developing countries. As described herein, four developing countries Road Less Traveled that have fl outed the conventional wisdom appear to have Four Developing Countries Blaze New Trails benefi ted tremendously because of doing just that. 26 to Better Health Vast opportunities to improve public policy on behalf By Th omas W. Croghan of the common good now exist in the Gulf Coast states as a result of Hurricane Katrina. The cover story reports on some ways in which that nightmare of a year ago can 30 Publisher’s Page Old Lessons, New Doctrine: give rise to a better morning—if community members Classic Insights into Counterinsurgency and policy professionals seize the day. In a case as all- Are Especially Trenchant Today encompassing as this, any organization that’s been around a while understands that seizing the “day” means By James A. Th omson committing to the work for decades—and reminding people of the lessons learned when they are forgotten.

—John Godges On the Cover The late winter sun rises over a catfi sh pond outside De Ridder, Louisiana. AP PHOTO/AMERICAN PRESS, SHAWN MARTIN

WWW.RAND.ORG RAND REVIEW / F A L L 2006 3 News

Counterinsurgency Lessons Can Inform Global War on Terror As the U.S. military struggles with Long, a doctoral candidate in program in Iraq is heavily focused ever-evolving insurgencies in Iraq securities studies at the Massa- on top leaders, with little funding and Afghanistan, it does not need chusetts Institute of Technology, to encourage reporting on rank- to reinvent the wheel, according reviewed dozens of past RAND and-fi le insurgents. to a new summary of the lessons studies about tactics used against A third recommendation is to learned from RAND’s fi ve decades insurgencies in such disparate improve border security systems of counterinsurgency research. places as Vietnam, Algiers, and El for each country. Th e use of tech- Tactics used to battle Cold War– Salvador to distill lessons for today’s nologies such as remote-piloted era insurgencies—such as off ering confl icts. His work included eff orts vehicles cued by ground-based sen- amnesty to combatants and secur- to declassify several reports so that sors could provide a cost-eff ective ing national borders—could help their lessons could be incorporated way to monitor border infi ltration, the United States in its current into his study. allowing troops to respond quickly, confrontations with insurgents in He found four lessons of par- said Long. Iraq and Afghanistan. ticular relevance to today’s ongo- Th e fourth recommendation Th e new report argues that ing battles: the proper organization is to focus on ways to pacify local while counterinsurgency details of counterinsurgency eff orts; the areas rather than an entire nation. may vary greatly between the Cold use of amnesty and reward pro- In addition, development eff orts War and today, many of the diff er- grams; the control of national bor- should target local areas that show ences are overstated. For example, ders; and pacifi cation of a nation’s a willingness to support their own some today view the fragmented population. defense, while development eff orts nature of the Iraqi insurgency as He concluded that the best way at the national level should empha- almost without precedent. Yet to organize against insurgencies size the improvement of infrastruc- “many insurgencies during the would be to bring together political ture and other shared resources. Cold War were fragmented, with and military components into one Th e report also highlights the rebel groups fi ghting each other as decisionmaking group so that the relevance of two RAND-formulated well as a central opposition force,” eff orts of one do not undercut the theories: “hearts and minds” and said Austin Long, author of the progress of the other. He suggested “cost/benefi t.” Th e former stresses RAND report. that the Provincial Reconstruction the need to win popular support by Teams currently operating in Iraq ameliorating some of the negative and Afghanistan be expanded to eff ects of economic development create similar eff orts at the regional while accelerating the delivery The best way to organize and national levels. of modernity’s benefi ts. Th e lat- against insurgencies would be Another proven strategy would ter points out that coercive meth- both extend amnesty to those ods could actually stimulate an to bring together political involved in an insurgency and off er insurgency, leading to a spiral of and military components into rewards to people who turn in fi ghting between insurgent and those involved in insurgent fi ght- counterinsurgent—a spiral that one decisionmaking group ing. Amnesty should be off ered might be unwinnable by mod- so that the efforts of one in Iraq even to insurgents who ern democracies with moral and have been involved in killings, political limitations on the use of do not undercut the progress said Long, but only if the combat- force. ■ of the other. ants agree to cooperate fully with the government and relocate away For more information: On “Other War”: Lessons from Five Decades of RAND from insurgent-dominated areas. Counterinsurgency Research, ISBN 978-0- He noted that the current reward 8330-3926-2, 2006.

4 RAND REVIEW / F A L L 2006 WWW.RAND.ORG News

Maritime Terrorism Risk Involves More Th an Just Container Shipping A new RAND report has found rorist requirements of disruption, Th e authors state that civil that the heightened perception of destruction, and publicity.” liability is the de facto, principal maritime terrorism risks does not Th e study off ers a number of approach to managing maritime conform to the reality of today’s recommendations. It argues that terrorism risks through private- threats and vulnerabilities. Th e policymakers should base maritime sector incentives. However, cur- study assessed whether the mari- terrorism policy on empirical risk rent liability standards in this time terrorism threat is increasing, analysis rather than on perceived context are so ambiguous that “the what maritime terrorist scenarios threats. In addition, because the civil justice system may not be appear to be the most worrisome, greatest risks to human life turn eff ective either as a compensation and whether current liability rules out to involve onboard bombs or mechanism or in generating clear can help the private sector manage biological attacks on cruise ships incentives for private-sector fi rms,” the risks. and ferries, the study stresses the according to coauthor Michael Th e study compared a series need for policymakers to audit the Greenberg. of attack scenarios within three soundness of vessel and facility Because current civil liability shipping categories: container ships, security procedures for passenger standards provide so little guidance cruise ships, and ferries. It assessed ships, to improve security measures to fi rms, the study argues that the magnitude of the human and at ports for passengers and lug- policymakers should both review economic consequences of the vari- gage, and to implement vigorous the rationale of third-party liability ous scenarios and the likelihood of procedures for documenting crew for terrorist attacks (by which fi rms their occurrence—as determined and staff . are held responsible for attacks per- by target vulnerability and terrorist “Focusing solely on securing petrated by terrorists) and consider intent and capability. the container supply chain without the pros and cons of civil liability as As shown in the fi gure, which defending other parts of the mari- a method for dealing with maritime plots the scenarios within the time environment is like bolting terrorism risks and injuries. ■ three groups, two classes of risk the front door and leaving the back emerge as most important. First, door wide open,” said Henry Willis, For more information: Maritime Terror- ism: Risk and Liability, ISBN 978-0-8330- although container shipping attacks a report coauthor. 4030-5, 2006. are less likely, some—such as the detonation of a nuclear bomb in a container within a port—merit Two Classes of Risk Emerge as Most Important: Unlikely but attention because of their extreme Catastrophic Scenarios and More Likely but Moderate Scenarios potential consequences. Second, High Ferry scenarios smaller attacks on cruise and ferry Onboard bombs (ferries and cruise ships) Cruise ship scenarios ships—such as onboard bombs or Container ship scenarios food and water contamination— Area of greatest concern are considered important because Food, water contamination (cruise ships) they are well within the capabilities of contemporary terrorists and have direct ramifi cations for human life, even if the overall consequences are Attack likelihood Detonation of relatively moderate. nuclear bomb (container ships) According to Peter Chalk, Low one of the report’s coauthors, Low High Human consequences both classes of risk would vari- ously “meet the interrelated ter- SOURCE: Maritime Terrorism: Risk and Liability, 2006.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 5 News

Felons Reload While No One Is Watching, Report Finds Federal and state laws prohibit Tobacco, Firearms, and Explosives criminal histories, but there are no certain criminals from possessing to see whether the buyers appeared mechanisms to prevent ammuni- ammunition, but that did not stop in federal or state criminal data- tion purchases. Los Angeles and a 52 such criminals from purchas- bases. Th e study covers the months few other cities require ammuni- ing more than 10,000 rounds of of April and May 2004. tion sellers to collect information ammunition in a two-month period, Over those two months, more about ammunition buyers, but in according to a new RAND report than 2,000 people bought ammu- the past those records have not published in the October 2006 issue nition from the ten retailers, and been routinely reviewed. of Injury Prevention. the vast majority had no criminal “If lawmakers want to prohibit Th e study is based on informa- records (see fi gure). Of the 133 the illegal sale of ammunition, they tion collected from ten ammuni- buyers with criminal records, 52 could extend the instant background tion retailers by the Los Angeles were explicitly prohibited by law checks required before guns are sold Police Department and screened from buying ammunition. Th ose to also cover the sale of ammu- by the federal Bureau of Alcohol, 52 people bought 2.3 percent of nition,” said report coauthor Greg the nearly 440,000 rounds of Ridgeway. “But unless such a step is ammunition sold, amounting to taken at the state level, a felon could more than 10,000 rounds. simply purchase ammunition in a Past studies have shown that nearby city to avoid the check.” Of 133 buyers with criminal guns and ammunition possessed by Another alternative is for law records, 52 were explicitly felons and others prohibited from enforcement offi cials to monitor prohibited by law from owning weapons are more likely to the sales records of ammunition to be used in violent crimes than are gather tips about felons who might buying ammunition. weapons bought by people with no illegally possess fi rearms. “Ammu- nition logs have been used by Los Angeles area law enforcement offi - Few Who Purchased Ammunition Were Legally Prohibited from Doing So, But Those Few Still Bought More Than 10,000 Rounds cials to obtain search warrants that in Two Months have led to the recovery of illegal fi rearms,” noted lead author and Criminal record (133) Prohibited from criminologist George Tita. purchasing ammunition 10,050 —felony record (30) rounds Felons might resort to under- bought Prohibited from (2.3%) ground dealers. However, a study purchasing ammunition —non-felony record (22) conducted in Chicago communi- ties with high levels of gun violence No criminal record (1,898) found that stricter enforcement of

Allowed to purchase ammunition policies did not nec- ammunition (81) essarily lead to a black market in the product. Although the study focused on one part of Los Angeles, the Ammunition Ammunition purchasers purchasers (2,031) with criminal record (133) researchers point out that the fi nd- ings have implications for other SOURCE: “The Criminal Purchase of Firearm Ammunition,”Injury Prevention, Vol. 12, No. 5, October 2006, pp. 308–311, G.E. Tita, A.A. Braga, G. Ridgeway, G.L. Pierce. states and nations that monitor NOTE: Ammunition was purchased over a two-month period (April and May 2004) from ten retailers fi rearm sales but not ammunition in Los Angeles. purchases. ■

6 RAND REVIEW / F A L L 2006 WWW.RAND.ORG News

As for Implementation, Comprehensive School Reform Is Anything But To improve student performance, As shown in the table, few models.” Only a handful of practices more than 8,000 schools across the CSR schools have implemented the diff ered either among the types United States have adopted compre- core components very highly. For of model schools or between the hensive school reform (CSR) mod- all four models, the mean (average) model schools and their matched els that are intended to align the scores for the schools surveyed do non-model schools. core components of education with a not approach 2.0 (full implementa- Th e fi ndings, if replicated in coherent vision. Th e schools have used tion) in any component. future studies, have several broad more than $2 billion in federal funds Regardless of the model cho- implications, with this particularly to implement various CSR strategies. sen, some core components, such salient one: At the current level of So far, evaluations have shown as curriculum and student assess- implementation, CSR models are mixed results. “Th e eff ectiveness ments, were implemented more likely to have modest or no eff ects of any intervention depends on fully than others, such as meth- on student achievement. whether it’s delivered as its designers ods of instruction and increasing Th ere is signifi cant room to intended,” said Georges Vernez, lead parental involvement. But the study raise the level of implementation. author of a new RAND study on found that implementation falls For instance, schools could ensure the topic. “Our research shows that short of the recommended levels. that teachers receive the necessary schools are not fully implementing In general, teachers reported initial training and are committed to the CSR models they adopted.” only a lukewarm commitment to implementing the adopted models. Th e study quantifi es the levels implementing a school’s model, Without such support, the imple- of implementation of four diff er- with most feeling that the training mentation of CSR models cannot ent CSR models in 250 schools in received did not adequately prepare be expected to help schools meet the two states (Florida and Texas). Th e them to start using the model. federal No Child Left Behind goal implementation scores are based on “One of the more surprising of 100-percent profi ciency in read- a survey of principals and teach- fi ndings,” said Vernez, “was that, ing and mathematics by 2014. ■ ers, supplemented with in-depth on average, all schools in the survey For more information: Evaluating Com- case studies in 12 schools, allowing engaged in many of the same activi- prehensive School Reform Models at researchers to compare a large sample ties at the same frequency and level Scale: Focus on Implementation, 2006, available online only, at www.rand. of CSR schools and to compare CSR of intensity, regardless of whether org/pubs/monographs/MG546/. schools with non-CSR schools. schools used one of the four CSR

Surveyed Schools Are Not Fully Implementing Their Chosen Reform Models Model 1 (36 schools) Model 2 (42 schools) Model 3 (93 schools) Model 4 (79 schools) Mean Percent Highly Mean Percent Highly Mean Percent Highly Mean Percent Highly Core Component Score Implemented Score Implemented Score Implemented Score Implemented Curriculum 1.46 42 1.43 31 1.53 43 1.65 67 Methods of instruction 1.19 0 N/A — 1.19 4 1.23 3 Governance 1.36 3 1.26 12 N/A — 1.07 3 Student assessments 1.30 22 1.47 37 0.90 1 1.14 4 Parental involvement 1.01 13 0.84 10 N/A — 1.07 4 SOURCE: Evaluating Comprehensive School Reform Models at Scale, 2006. NOTES: Mean (average) scores range from 0 (no implementation) to 2.0 (full implementation). Ranges are low (0–0.39), medium-low (0.40–0.79), medium (0.80–1.19), medium-high (1.20–1.59), and high (1.60–2.0). N/A = not applicable, meaning that a component is not part of a model’s prescribed set of practices.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 7 News

Public Spends Little to Provide Health Care for the Undocumented Much political attention has rare piece of information about services. In L.A. County, many focused on the public burden of immigrants,” said Smith. foreign-born residents had almost providing health care for non- After deriving estimates for no contact with the formal health elderly undocumented immi- L.A. county, the researchers extrap- care system. Whereas only about grants. But a new RAND study olated estimates to the national a tenth of native-born residents in the November/December 2006 level. As shown on the left side had never had a checkup, that issue of Health Aff airs by Dana P. of the fi gure, total national medi- fraction jumped to a quarter for Goldman, James P. Smith, and cal spending for nonelderly adults foreign-born residents and to a third Neeraj Sood shows that a relatively in 2000 was nearly $430 billion. for undocumented immigrants. small amount of U.S. tax money Native-born residents accounted for Moreover, because L.A. County is is spent on health services for this 87 percent of the population but for known as an immigrant-friendly population. 91.5 percent of the spending. location for services, the national Researchers analyzed informa- Foreign-born residents (includ- estimates may overstate undocu- tion collected by the Los Ange- ing U.S. citizens, permanent resi- mented immigrant service use and, les Family Neighborhood Survey, dents, undocumented immigrants, thus, overstate medical costs. which interviewed families in 65 and temporary legal immigrants) One key reason for the lower use L.A. County neighborhoods dur- accounted for 13 percent of the of services is that immigrants, espe- ing 2000 and 2001. Th e county has population but for only 8.5 per- cially the undocumented, appear the largest concentration of immi- cent of spending. Undocumented to be much healthier than native- grants in the nation. Nonelderly immigrants, who constituted 3.2 born residents. participants—those between the percent of the nation’s population, Th e right side of the fi gure ages of 18 and 64—were asked accounted for only about 1.5 per- shows that foreign-born residents about their health status, whether cent of medical costs, according to use less funding from public insur- they had health insurance, the the study. ers (such as Medicare and Medic- type and amount of care used, and Th e lower medical spending is aid) and pay more for health care their immigration status, a “very driven by lower utilization of health out of pocket than do native-born residents—a pattern even more pronounced for undocumented Undocumented Immigrants Account for a Small Amount of Medical Costs and Pay More for Health Care Out of Pocket Than immigrants. Do Other U.S. Residents Given these fi ndings, the study

Percentage of Public insurers argues, the policy debate over immi- population Undocumented $429,966 Private insurers 3.2 ($6,449 = 1.5%) gration should focus not on health Out of pocket 13 Foreign-born ($37,455 = 8.5%) care costs but on a fuller analysis

Undocumented 17 47 36 of the fi scal benefi ts and costs of immigrants. “A more comprehen- sive analysis would incorporate the Native-born Foreign-born 16 57 27 87 ($392,511 = 91.5% taxes paid by immigrants. But it of total spending) also should include other public Native-born 21 59 20 benefi ts received—in particular, public school costs, where the costs 020 40 60 80 100 Total national medical spending Sources of payment of all types of immigrants, includ- in 2000 (millions of dollars) (percentages) ing undocumented immigrants, SOURCE: “Immigrants and the Cost of Medical Care,” Health Affairs, Vol. 25, No. 6, are likely to be much larger,” said November/December 2006, pp. 1–12, Dana P. Goldman, James P. Smith, Neeraj Sood. Smith. ■

8 RAND REVIEW / F A L L 2006 WWW.RAND.ORG News

New Study Suggests Th ree Steps to Improve Mental Health Care According to the National Health- dards and minimum competency ity—its division into public and care Quality Report, published by requirements, according to the private spheres—while balancing the U.S. Agency for Healthcare study, which argues for harnessing the needs for access, effi ciency, and Research and Quality, only 4 of market forces to help enforce them. quality. 179 quality indicators relate to For example, insurance fi rms could Th e study concludes that tak- mental health care. off er fi nancial incentives for pro- ing these three steps would require “Mental health has not been viders to seek additional training. engaging the public in a meaning- well-represented in nationwide qual- Th ird, health care fi nancing ful fashion, and it outlines the ity improvement initiatives, and it should be linked directly to inno- roles that each group of stakehold- has not been clear how to acceler- vations in quality improvement ers could play in taking each step. ate progress,” said Kavita Patel, a and mental health care delivery. Th e table illustrates the potential RAND Corporation researcher and Th e study suggests one approach— roles for those involved in the fi rst lead author of a study published in pay-for-performance—which would step. ■ the May/June 2006 issue of Health link salary bonuses or other mon- Aff airs. etary rewards to quality indicators, Th e study makes a case for how such as patient satisfaction. Fully The mental health system to accelerate progress, identifying developing such an approach to needs a better set of three steps. Th e fi rst step involves mental health care would need to expanding the range of mental account for the system’s complex- accountability mechanisms. health disorders for which quality improvement programs exist and Involving Multiple Stakeholders Is Key in Expanding the Range of adapting those programs for more Quality Improvement Programs for Mental Health Care diverse ages, groups, and service Stakeholders Potential Roles settings. For example, the strongest Health care Develop local training and continuing mental health quality improvement providers education on how to implement the programs programs have centered on treating Insurance companies Support regional adaptations of evidence- depression in well-established pri- based practices in using quality indicators and information systems mary care settings (such as hospital- Researchers Design improvement toolkits and based outpatient clinics), but fur- demonstrations ther progress is needed to adapt Policymakers Grant legislative and other support for these programs to other current evidence-based practices practice settings (such as those of Professional health Collaborate with other stakeholders to develop, individual practitioners or private care organizations implement, and monitor program guidelines group practices). Exercise consumer leadership, as in rating any competing programs Second, the mental health sys- Consumers Participate in quality improvement initiatives tem needs a better set of account- Support consumer leadership (of professional ability mechanisms. Cognitive- health care organizations) in developing the behavioral treatment, for example, programs currently has no standards or cer- SOURCE: “What Is Necessary to Transform the Quality of Mental Health Care,” Health Affairs, Vol. 25, No. 3, May/June 2006, pp. 681–693, Kavita K. Patel, tifi cation requirements, and con- Brittany Butler, Kenneth B. Wells. sumers have no easy way to identify practitioners who deliver the best quality care. Th e profession needs strong leadership to develop stan-

WWW.RAND.ORG RAND REVIEW / F A L L 2006 9 Perspectives

Shifting Course Th oughts on the Future of American Foreign Policy

THE AMERICANLED intervention in Iraq—with First, he argued that the Bush administration its immediate goal of regime change and its ultimate applied a doctrine of preventive war, which was “emi- goal of spreading democracy throughout the Middle nently justifi able against al Qaeda,” to the rogue state East—has been the cornerstone of the Bush adminis- proliferation problem presented by Iraq. “Th e latter is tration’s foreign policy. But as the situation continues a very serious problem, but of a lower order of magni- to deteriorate in Iraq and the prospect of democracy tude than the prospect of an attack by a stateless, nihil- there fades away, diffi cult questions have arisen about istic terrorist group potentially armed with weapons of how history will judge the intervention and the ideas mass destruction.” underlying it. Even more important, the outcome of Unlike preemption, which deals with imminent the intervention could have serious ramifi cations for threats, preventive war tries to head off threats that are America’s future foreign policy. months or more often years in the future. Th e problem Francis Fukuyama, director of the international with such a strategy is that it requires the United States development program in the Paul H. Nitze School to be able to accurately predict the future, not just in of Advanced International Studies at Johns Hopkins terms of enemy capabilities but also in terms of the University, addressed these questions in a wide-ranging complex calculations that foreign leaders will make discussion at the RAND Corporation. Drawing from years hence. “In Iraq,” he said, “American knowledge his recent book, America at the Crossroads: Democracy, of enemy capabilities—even its near-term capabilities Power, and the Neoconservative Legacy, he focused on with respect to weapons of mass destruction—was what has led to the current situation in Iraq, what sorely defi cient.” eff ect that situation is having on U.S. foreign policy Th is misjudgment was compounded by the Amer- now, and what direction U.S. foreign policy should ican failure to account for the undercurrents of take in the future. anti-American sentiments at the time of the 2003 intervention—“sentiments that had been brewing A Series of Misjudgments long before the Bush administration took offi ce and Fukuyama noted that although he initially shared that had centered on U.S. support for globalization,” the neoconservative principles that motivated the according to Fukuyama. When administration policy- Bush administration’s foreign policy, he was never makers formulated their neoconservative doctrine of persuaded by the administration’s rationale for the “benevolent hegemony,” he continued, they naively war in Iraq. According to Fukuyama, America fi nds assumed that the U.S. assertion of power in Iraq would itself where it is today in Iraq because of three fateful be supported because it was so widely understood that misjudgments. American foreign policy was “simply more moral” than that of other countries. Th e fi nal misjudgment stemmed from the admin- istration’s failure to heed the manifold lessons of As the situation continues to deteriorate nation-building in attempting to bring democracy to in Iraq and the prospect of democracy Iraq. As Fukuyama noted, RAND and other organi- zations, including many within the U.S. government, there fades away, diffi cult questions had accumulated and disseminated a great deal of have arisen about how history will knowledge about what is necessary for successful judge the intervention and the ideas nation-building—in particular, knowledge about the levels of troops, money, and time required in a variety underlying it. of nation-building environments.

10 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Unfortunately, he noted, “the Bush administration made little use of any of this extensive body of knowl- edge.” Th e administration assumed that once Saddam Hussein was removed from power, there would be a relatively painless transition to democracy, as happened in Eastern Europe after 1989; when that did not hap- pen, the administration was taken by surprise.

Avoiding a Backlash Th e deepening failure in Iraq has yielded a bitter harvest of deleterious consequences for American foreign policy, according to Fukuyama. “One of the unfortunate legacies of the Bush administration is that its actions are being increasingly viewed as ‘toxic’”—something that has led to a growing move- ment in foreign policy circles toward what Fukuyama calls a Henry Kissinger–like “realism,” a more iso- lationist U.S. foreign policy, and a retreat from the world stage. Actors on both sides of the political aisle are now disparaging the promotion of democracy as an DIANE BALDWIN illegitimate activity and are increasingly phobic about any use of American power whatsoever. people doing it themselves,” he noted. Th e demand Francis Fukuyama, author of America Fukuyama views this stance as an overreaction to must come from within. at the Crossroads: the over-militarized means used by the Bush admin- But once democracy has taken root organically Democracy, Power, and the istration to promote democracy abroad, not the ends. in a country, there is much that outsiders can do to Neoconservative Th e failure of nation-building in one country, he support the process. Th is support includes monitoring Legacy, greets guests at his suggested, should not discredit the notion of nation- elections, promoting civil society groups, providing RAND Corporation building in every country. open access to media, and helping to develop political speaking engage- ment. Fukuyama Fukuyama believes that greater modernization parties. is a former RAND and democracy will actually increase the alienation “Ultimately, democracy is spread by the prestige analyst and a underlying terrorism in the short run. In the long and moral credibility of countries that are demo- current member of the Pardee run, however, democracy is a better bet than relying cratic,” Fukuyama said, pointing out that the United RAND Graduate on illegitimate authoritarians. “Th e only way to get at States was a beacon to Eastern European countries School Board of Governors. political and social actors who support terrorist move- throughout the Cold War because of what America ments,” he said, “is to give them some political space in represented, not because of the way it used its military which to operate, hopefully under democratic rules.” power. ■ As one example of how exercising power in a pluralistic setting can lead to a gradual process of political accul- turation, he cited the Islamist PJD (Justice and Devel- The failure of nation-building in one opment Party) in Morocco, which has steadily adopted more moderate rhetoric over the past few years. country . . . should not discredit the notion of nation-building in every The Limits of Democracy Promotion According to Fukuyama, the primary lesson that country. America should learn from its travails in Iraq is not that the United States should eschew promoting democracy abroad, but that there are limits to what any nation can do in promoting democracy abroad. “No country has ever been democratized without the

WWW.RAND.ORG RAND REVIEW / F A L L 2006 11 Steady Under Fire All-Volunteer Force Proves Its Resilience, So Far

By Bernard D. Rostker a total collapse of the system, there is no better way to, in the words of the U.S. Constitution, “raise and sup- Bernard Rostker is a senior RAND fellow who has held many port armies” and “provide and maintain a navy.” senior government posts. In 1979, he was appointed director Th e all-volunteer force has shown it can be suc- of the Selective Service System by President Jimmy Carter. cessful during periods of confl ict as well as during From 1994 to 1998, Rostker served as assistant secretary of peacetime. Nevertheless, the sustained confl icts in the the U.S. Navy for manpower and reserve affairs, and from Middle East pose obstacles not faced before. Reenlist- 1998 to 2000 as the 25th undersecretary of the U.S. Army. ment rates have remained high in spite of the wars in From 2000 to 2001, he served as undersecretary of defense Iraq and Afghanistan, sustaining the overall size of the for personnel and readiness, the offi cial who is the U.S. force. But the U.S. Army, Army National Guard, and Secretary of Defense’s senior policy adviser on recruitment, Army Reserve have faced diffi culty recruiting new per- career development, pay, and benefi ts for 1.4 million active- sonnel. Th ere are no guarantees with a volunteer force. duty military personnel, 1.3 million National Guard and Only time will tell if the current level of operations can National Reserve personnel, and 725,000 civilian employees be sustained into the future. of the U.S. Department of Defense. Th e volunteer force has proven much more resil- ient than we had any reason to hope it would be and has far exceeded the early estimates of those who put it into place. But logic tells us there is a limit. Just merica’s all-volunteer military has been an because we have not yet broken the force does not overwhelming success since its inception in mean it cannot be broken. A1973, but the force faces an unprecedented challenge posed by the wars in Iraq and Afghanistan. Evolution of the All-Volunteer Force Th e all-volunteer military has become the world’s Th e United States has conscripted its armed forces strongest fi ghting force, attracting recruits who are for only 35 of its 230 years—nearly all in the 20th better educated and more skilled than those who century—and the American people have been gen- served under the U.S. military draft. However, after erally willing to accept the practice of conscription four years of war with mounting casualties in Iraq, when service has been perceived to be universal. Th at continuing insurgent attacks in Iraq and Afghanistan, acceptance began to erode in the 1960s. Th ere were and multiple additional deployments throughout the fi ve major reasons: world, the all-volunteer force has experienced recruit- • Demographics. Th e size of the eligible population ing shortfalls for the fi rst time since the late 1970s. of young men reaching draft age each year was To date, the all-volunteer force has done the job. so large and the needs of the military so small Under the draft, people served because we in America in comparison that, in practice, the draft was no made them serve. Under the all-volunteer force, people longer universal. serve because they want to serve, and they are serving • Cost. Obtaining enough volunteers was possible at very well in the most trying of circumstances. Short of acceptable budget levels.

12 RAND REVIEW / F A L L 2006 WWW.RAND.ORG • Moral and economic rationale. Conservatives and Figure 1— Today’s Enlistees Are Better Educated Than They Were libertarians argued that the state had no right to at the End of the Draft impose military service on young men without 100 New enlistees, 1973 and 2006 99 their consent. Liberals asserted that the draft placed U.S. Marine Corps 96 95 U.S. Army 85 unfair burdens on the underprivileged members of 81 75 U.S. Navy society, who were less likely to get deferments. U.S. Air Force 71 • Opposition to the war in Vietnam. Th e growing 50 58 unpopularity of the Vietnam War meant the coun-

graduates 45 try was ripe for a change to a volunteer force. • Th e U.S. Army’s desire for change. Th e army had 25

lost confi dence in the draft as discipline problems of high school Percentage among draftees mounted in Vietnam. 0 1973 2006 SOURCE: Quality by Fiscal Year from 1973, Curt L. Gilroy, Washington, D.C.: Office of Accession Policy, In 1969, President Richard Nixon created the Office of the Under Secretary of Defense (Personnel and Readiness), 2006. President’s Commission on an All-Volunteer Armed Force, also known as the Gates Commission. It consid- ered key military-manpower issues, including person- Figure 2— Moving to a Volunteer Military Has Dramatically nel supply and demand, attrition and retention, and Increased the Professionalism of the Force the ideal mix of career and noncareer members in the 70 context of management effi ciency and personal equity. Members with more than four years of service U.S. Marine Corps 66 60 Th e commission concluded that the nation’s inter- U.S. Army U.S. Navy 54 ests would be better served by an all-volunteer force 50 51 U.S. Air Force 49 rather than by a combination of volunteers and con- 46 40 42 scripts. In 1971, Nixon signed a law to end the draft and 37 35 to put the selective service structure on standby. After 30 31 Percentage 26 a two-year extension of induction authority, the end of 20 18 the draft was formally announced in January 1973. 16 10

Changes for the Better 0 1969 1979 2004 Th e quality of military personnel has improved since SOURCE: I Want You! 2006. the end of the draft, as measured by scores on intel- ligence tests. Th e percentage of new recruits with high school diplomas has surged (see Figure 1). Th e propor- percent and the 2002 civilian graduation rate of tion of career personnel and the profi ciency and profes- 79 percent. In addition, 95 percent of active-duty sionalism of the force have also dramatically increased offi cers have baccalaureate degrees, and 38 percent (see Figure 2). A largely unexpected consequence of have advanced degrees. moving to a professional military with better pay was a • Mental aptitude. Today’s American military scores higher rate of reenlistment and a sharp rise in the size of well above the civilian population on standard the career force relative to the overall force. At the same tests of intelligence (see Figure 3). time, the all-volunteer force has made the military more • Marital status. Th e larger career force means that the representative of the nation as a whole. number of service members who are married has For 26 years, the U.S. Department of Defense has increased. Today, 49 percent of enlisted personnel reported annually on social representation in the U.S. are married, compared with 40 percent in 1973. military. Th e 2004 report noted the following: Among active-duty offi cers, 68 percent are married. • Education level. Th e most recent statistics show • Gender. Today, 15 percent of the active-duty enlisted that 92 percent of new enlistees to the active- force is female, compared with less than 2 percent duty force are high school graduates. Th e fi gure when the draft ended. Sixteen percent of the offi - for the reserve components is 87 percent. Th is cer corps is female. Despite these improvements, compares favorably with both the 1973 goal of 45 women are still underrepresented in the military.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 13 Figure 3— Today’s Military Personnel Score Well Above Average • Race and ethnicity. In fi scal year 2002, African on Standard Tests of Intelligence Americans were slightly overrepresented among 70 new enlistees relative to the civilian population: 60 U.S. military personnel, 2002* 66 16 percent compared with 14 percent. However, U.S. civilians 58 50 this is considerably more equitable than was the 40 1973 level of 28 percent. Latinos are currently 41 30 36 34 underrepresented, making up 16 percent of all 30

Percentage 20 civilians but only 11 percent of new enlistees (see 10 Figure 4). 1 0 • Socioeconomic status. Recruits come primarily from Scoring above 108 Scoring 92 to 107 Scoring 72 to 91 families in the middle or lower middle classes. Few on intelligence tests on intelligence tests on intelligence tests recruits come from upper-income families, leading SOURCE: I Want You! 2006. some to criticize the all-volunteer military. But historically, few people from elite backgrounds have ever served in the military except during Figure 4— Racial and Ethnic Composition of U.S. Military Enlistees times of mass mobilization such as occurred dur- Is Fairly Representative of the Nation at Large ing World War II. 70 • Geographic representation. Th e South continues to 69 U.S. military new enlistees, 2002* 60 65 be overrepresented, with more than 41 percent of U.S. civilians 18–24 years old, 2002* 50 enlistments compared with 35 percent of the civil- 40 ian population (see Figure 5). 30

Percentage 20 Pillars of Success 16 14 16 10 11 America’s experience with the all-volunteer force sug- 0 6 5 gests four principal reasons for its success: leadership, White African Latino Native American, American Asian/Pacific Islander analysis, targeted programs, and adequate budgets. SOURCE: Population Representation in the Military Services—Fiscal Year 2002, David S. C. Chu, Whenever one of these factors has been missing over Washington, D.C.: Office of the Under Secretary of Defense (Personnel and Readiness), 2004. the past 30 years, the force has faltered. Th e fi rst factor is leadership from top manage- ment. Th e all-volunteer force would not have come Figure 5— The South Continues to Be Overrepresented in into being when it did without the leadership of Comparison with Other Regions of the United States President Nixon, who began the planning process and 100 0.8 announced the formation of the Gates Commission 14 17 66 within weeks of taking offi ce in 1969. 75 20.7 23.4 Within the military, U.S. Army General Maxwell Territories or possessions Northeast Th urman is considered by many as the single most 23 50 24.6 North central important person in the history of the all-volunteer West force, because he taught the Pentagon how to recruit.

Percentage South 25 He often said that it may be called an all-volunteer 41.4 34.9 force, but it is really an “all-recruited force.” More

0 than any other uniformed leader, Th urman recognized U.S. military U.S. civilians throughout the 1980s that the military had to compete new enlistees, 2002* 18–24 years old, 2002* aggressively in the civilian labor market for American SOURCE: Population Representation in the Military Services—Fiscal Year 2002, David S. C. Chu, Washington, D.C.: Office of the Under Secretary of Defense (Personnel and Readiness), 2004. youth—and had to do so with the right tools based on NOTES: South includes AL, AR, DC, DE, FL, GA, KY, LA, MD, MS, NC, OK, SC, TN, TX, VA, WV. market research and statistical analysis. West includes AK, AZ, CA, CO, HI, ID, MT, NM, NV, OR, UT, WA, WY. North central includes IA, IL, IN, KS, MI, MN, MO, ND, NE, OH, SD, WI. Northeast includes CT, MA, ME, NH, NJ, NY, PA, RI, VT. Th e second factor, then, is the use of quantitative Territories or possessions include American Samoa, Federated States of Micronesia, Guam, Marshall Islands, analysis to test, adjust, and evaluate policies. Almost Northern Mariana Islands, Palau, Puerto Rico, U.S. minor outlying islands, Virgin Islands. every change to the all-volunteer force has been made *2002 statistics were the most recent data available at the time of publication. only after research demonstrated its likely eff ect.

14 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Th e research of the 1960s and early 1970s reassured Gates, Jr. “While there is a reasonable possibility that decisionmakers that such a force might be possible at a peacetime armed force could be entirely voluntary,” acceptable budget outlays. In the 1970s and 1980s, wrote Greenewalt, “I am certain that an armed force test programs demonstrated the value of advertising, involved in a major confl ict could not be voluntary.” So educational incentives, and bonuses in encouraging far, Greenewalt’s reservations have not been borne out. enlistment. Analytical evidence later helped reform the But given the ongoing war in Iraq—with casual- compensation system. ties rising, enlistments dropping, deployments being Studies of enlistee test scores and job performance extended, the situation on the ground deteriorating, proved what now seems logical but was once very con- and a majority of the American public no longer troversial: People who score higher on standardized tests believing that the war is worth fi ghting—the issue do better on the job than do those who score lower. of recruiting enough volunteers to maintain the U.S. Th e resulting emphasis on quality attracted capable military at required levels is again relevant. With people and led to the increasing professionalism of the nearly 150,000 U.S. troops in Iraq and Afghanistan military. And ever since the fall of the Soviet Union in today, military commanders point to the outstand- 1991, which eliminated the threat that had dominated ing job being done in these nontraditional confl icts. national security strategy for four decades, personnel Remarkably, while enlistments have fallen off , reten- research has helped defense managers make the adjust- tion remains at historically high levels. Th ere were ments necessary to transform the post–Cold War force initial fears that soldiers would not reenlist if they had into a smaller, more agile, and more engaged one. to deploy twice into combat zones. Yet some soldiers Th e third factor is the implementation of targeted are now completing their third and fourth tours. U.S. programs to attract the required types and numbers troops have demonstrated their continuing commit- of recruits. To appeal to high-quality youths, the ment and willingness to serve. services had to craft marketing strategies and advertis- Some critics say that one solution to the current ing campaigns that conveyed the benefi ts of military dilemma would be to reinstate the draft as part of a service; off er money for education, bonuses for enlist- universal system requiring all young people to commit ing in certain occupations, and enlistment tours of time to national service of some type. But creating a diff erent lengths; and provide career opportunities system to employ all four million young people who that had civilian relevance. Th e services also had to reach adulthood every year would be very costly, if train a professional and motivated recruiting staff . such a system were to be universal. And if not everyone Finally, the key to creating a truly outstanding force were to serve, we would be in the same quandary in was persuading the most capable members to reen- which we found ourselves during the 1960s. list. Careerists demanded not just good pay but also A fi nal judgment on the all-volunteer force has not quality-of-life benefi ts, such as good housing, child been made. Th e 40-year partnership between policy- care, health benefi ts, family advocacy programs, and makers and policy analysts has produced not only the military stores. It was crucial that the services become fi nest fi ghting force the United States has ever fi elded “family friendly.” but also one that is broadly representative of the Ameri- Th e fourth factor required for success is adequate can people. Th e past 30 years—particularly the experi- fi nancial resources. Th e defense budget must be large ences in Iraq and Afghanistan—have proven that an enough to accomplish three things at once: support pay all-volunteer force can be sustained in peace and during raises that keep pace with both infl ation and civilian- the initial periods of military confl ict. Whether or not sector pay increases; provide resources for advertising, an all-volunteer force can be sustained over longer peri- recruiters, bonuses, and educational benefi ts; and fund ods of ongoing confl icts and recurring deployments, as the military retirement program and quality-of-life in the current situation, has yet to be determined. ■ initiatives.

Old Questions Resurface Related Reading In 1969, Crawford Greenewalt, a member of the Gates I Want You! The Evolution of the All-Volunteer Force, Bernard Rostker, RAND/MG-265-RC, 2006, 832 pp., ISBN 978-0-8330- Commission, shared his misgivings in a memorandum 3895-1 (hardcover with DVD), $68.50, 978-0-8330-3896-8 (hard- to the chairman, former Secretary of Defense Th omas cover), $48.50.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 15 Centerpiece

Three Decades of Experience with the All-Volunteer

SEVERAL RECURRING THEMES characterize the evolu- 1970 tion of the all-volunteer force: leadership in overcoming institutional resistance to change, analysis in forecasting 1970—Critical Breakthrough by Gates Commission Staff the costs and benefi ts of alternative courses of action, Was to Estimate the Supply of New Recruits, Pointing determination in identifying the right people to recruit to the Pay Increases Needed for an All-Volunteer Force and how to recruit them, the difference that pay and Cumulative number who would enlist at expected wage, W1 benefi ts make in recruiting and retaining personnel, Additional number who would enlist at expected wage, W2 and the need for an ongoing budgetary commitment to sustain success. At times since 1973, policymakers either allowed military pay to fall too far behind civilian earnings or W1 W2 curtailed recruiting and personnel programs. The cut- Aggregate potential • backs often came at the worst times: when the civilian recruits at W1 • Aggregate potential economy was strong and youth unemployment was recruits at W2 Expected enlistments low. Each time, recruiting and retention suffered. It took a large infusion of resources in 1978 and 1979 to correct the oversight before recruiting turned around in W1 W2 1980, and it took even larger increases in 1998 and 1999 Expected wages before recruiting rebounded in 2000. Economists transformed the blue area beneath a normal frequency distribution As noted by former U.S. Army General Maxwell (top) into a point on an aggregate enlistment supply schedule (above), showing that small increases in expected military benefits would attract large increases in Thurman, a principal architect of the all-volunteer recruits—up to that point. force: It might be called an all-volunteer force, but it is really an “all-recruited force.” ■

NATIONAL ARCHIVES February 21, 1970—Gates Commission (above) informs President Richard Nixon (center right): “We unanimously believe that the nation’s interest will be better served by an all-volunteer force, supported by an effective stand-by draft, than by a mixed force of volunteers and conscripts.”

1976—Recruiting funds are cut sharply.

BILL GENTILE/CORBIS 1979—All four services fail to meet their recruiting goals. U.S. Army General Maxwell Thurman (1931–1995), shown here in 1989, was Army falls short of its mission by 17 percent. Quality of an ardent supporter of the all-volunteer force. During the diffi cult years of the late 1970s and early 1980s, he argued that investments in “human recruits falls below minimum levels established by the U.S. capital” early on would ultimately yield great returns in an effi cient and Department of Defense. well-trained fi ghting force. In this effort, he challenged both the U.S. Department of Defense and the U.S. Congress to allocate the resources 1979—Army General Maxwell Thurman institutes “Be All needed to recruit and to retain high-quality personnel. You Can Be” advertising campaign.

16 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Force Have Charted a Course to Success

2004

1986—One of Several Experiments to Test Recruiting 2002—Pay for Midcareer Enlisted Personnel with Some Programs Compared the Effects of Enlistment Bonuses College Was No Longer Competitive with Civilian Pay, of Varying Amounts and Required Years of Service Prompting Creation of New Pay Table

Enlistment bonuses 60 Plan A: $5,000 for 4 years Plan B: $8,000 for 4 years 55 Plan C: $8,000 for 4 years, $4,000 for 3 years Seattle 50 Spokane Portland Butte Fargo 45 Portland Manchester Minneapolis Boise Syracuse Boston 40 Sioux Falls Milwaukee Albany Detroit Buffalo Wilkes Des Barre Springfield 35 Oakland Moines Chicago Cleveland Harrisburg New Haven Omaha Salt Lake Denver Ft. Hamilton City Indianapolis Columbus Pittsburgh Newark 30 Fresno Cincinnati Philadelphia Kansas St. Louis Richmond Louisville Beckley Baltimore 70th percentile City 25 Raleigh Average Los Nashville Charlotte Angeles Knoxville 50th percentile Albuquerque Oklahoma Little 20 Phoenix City Memphis Amarillo Rock Ft. Jackson 30th percentile Atlanta Jackson 15 Average RMC (E-1 to E-7) Dallas Shreveport Montgomery El Paso Jacksonville

New Annual earnings (thousands of dollars) Houston 10 San Orleans Antonio 0 5 10 15 20 25 30 Coral Gables Years of experience

Plan C triggered a startling 178-percent increase in three-year terms among those NOTES: Data compare July 2000 regular military compensation (RMC) for who would have signed up for two years anyway. Overall, the up-front bonuses enlisted ranks E-1 through E-7 with predicted year 2000 earnings of civilian spurred a 4-percent increase in high-quality recruits and a 15-percent increase males with some college. 50th percentile is the median (half earn more; in four-year terms. half less).

1979—The armed services discover a “tragedy of errors” Late 1990s—Recruiting budgets reduced in aftermath of involving four years of infl ated enlistee test scores on the Cold War. Unemployment reaches a 30-year low. Services Armed Services Vocational Aptitude Battery. again fi nd it diffi cult to achieve recruiting goals.

1980—Soviet Union invades Afghanistan. “The Selective 1998 and 1999—Congress approves large increases in Service System must be revitalized,” declares President recruiting resources, including pay raises, bigger advertising Jimmy Carter. budgets, and a “college-fi rst” program.

1980 and 1981—Congress legislates two large pay raises Since 2003—With nearly 150,000 U.S. troops in Iraq and for the military: 11.7 and 14.3 percent, respectively. Afghanistan, the all-volunteer force is being tested as never before. 1984—Congress passes Montgomery GI Bill, providing a “contributory educational program” for all who enlist after 2004—U.S. House votes 400 to 2 to reject a return to June 30, 1985. Enlistees may draw benefi ts upon completion conscription. of service.

1991—Operation Desert Storm (Persian Gulf War) proves “the all-volunteer force worked,” according to a Pentagon report. “The enlisted force exhibited unprecedented skill, commitment, maturity, and professionalism. The entire SOURCES: I Want You! The Evolution of the All-Volunteer Force, Bernard offi cer corps . . . consistently demonstrated skill, excellence, Rostker, RAND/MG-265-RC, 2006, 832 pp., ISBN 978-0-8330-3895-1 (hard- leadership, and professionalism we have not seen in this cover with DVD), $68.50, 978-0-8330-3896-8 (hardcover), $48.50; “Impact century—if ever.” of Pay and Draft Policy on Army Enlistment Behavior,” Alan E. Fechter, in Gates Commission, ed., Studies Prepared for the President’s Commission on December 25, 1991—Soviet Union ceases to exist. The an All-Volunteer Armed Force, Washington, D.C.: U.S. Government Printing end of the Cold War prompts leaders of both major political Offi ce, 1970; The Enlistment Bonus Experiment, J. Michael Polich, James N. Dertouzos, S. James Press, RAND/R-3353-FMP, 1986, available online only, parties to champion a smaller, transformed force through- at http://www.rand.org/pubs/reports/R3353/; Report of the Ninth Quadren- out the 1990s despite the rapid spread of nontraditional nial Review of Military Compensation, Curtis L. Gilroy, Washington, D.C.: operations. U.S. Department of Defense, 2002.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 17 Delta Dawning Gulf Coast Eff orts Look Beyond Recovery Toward Lasting Renewal

By John Godges and watch. I had to try to go back and see if I could contribute somehow.” John Godges is a RAND communications analyst and editor- His bosses at RAND and UCLA supported him. in-chief of RAND Review. Two days later, he was home. “It just so happened that I’m in a fi eld where there are persistent needs as a result en Springgate never thought he’d go home of the disaster.” so soon after his medical residency. Nor did Since his return, Springgate’s work has broadened he foresee that his professional calling toward from emergency care to community planning. At the community-based health care for the poor outset, he triaged evacuees at Louis Armstrong New Bwould be put to such an early and grueling test. Orleans International Airport, recruited and deployed A native of New Orleans, he completed his residency volunteer physicians to work in evacuee shelters, and at Tulane University in his hometown in late June of helped coordinate health care for emergency respond- 2005. He then moved to Los Angeles to start work as a ers in a French Quarter hotel ballroom and outside Robert Wood Johnson Clinical Scholar at UCLA, which a Wal-Mart. In the ensuing months, he and his col- led him to work on RAND research teams as well. Trou- leagues have opened new nonprofi t clinics in neighbor- bled by the health disparities he had observed across the hoods that were left without any local health services. country, he was planning to conduct community-based “Th e community can learn two important lessons participatory research to improve health care services for from the experience,” said Springgate. “First, we can’t homeless and low-income people. assume we’re going to be spared from these storms. We Th en, on August 29, 2005, Hurricane Katrina need to plan ahead and recognize what we need to do struck. In one fell swoop, the storm shifted the nation’s to limit the potential damage and loss. And second, as attention to the reality of aching poverty deep in the terrible as the disaster has been for the community, I’m country’s heart and soul. hoping there’s an opportunity now to overcome the “I watched the television reports in horror,” inertia that existed previously and to improve on some recalled Springgate, 34. “I couldn’t just simply sit back components of the community that weren’t function- ing optimally before.” He cited a reformed school sys- tem, smarter building practices, and improved health The hope is that the deluge of care quality and access as a few areas of top priority. RAND’s commitment to the region has broadened disappointment and heartbreak that in analogous ways. From an initial focus on emergency came with Hurricane Katrina can response and recovery in the most catastrophically sow the seeds of renewal incubated inundated areas, the preponderance of RAND work has since shifted toward helping the entire region and rooted in the Gulf region. reshape itself into something better than it was before

18 RAND REVIEW / F A L L 2006 WWW.RAND.ORG the hurricane hit. Th e hope is that the deluge of disap- pointment and heartbreak that came with Hurricane Katrina can sow the seeds of renewal incubated and rooted in the Gulf region. “What we have here is an opportunity to make a strong, nationally credible, nonpartisan resource avail- able in this region,” said George Penick, director of the RAND Gulf States Policy Institute, which is based in Jackson, Mississippi. Th e purpose of the institute, created in the wake of the hurricane, is to provide a long-term capacity to develop informed public policy in Louisiana, Mississippi, and Alabama. RAND studies in the region initially have focused on the repopulation of New Orleans, aff ordable hous- ing in Mississippi, student displacement and mental health, water resource management, and health care. Studies that are under development aim to help regional leaders strengthen an assortment of public and private responses related to housing, health, edu- cation, infrastructure, the economy, the environment, disaster preparedness, and government performance. “Th ese are issues that needed a great deal of atten- AP PHOTO/CAROLYN KASTER tion before Katrina and need even more attention after,” said Penick, who had served as president of the include the largely poor and African American commu- A slide trombone sports a steadfast nonprofi t Foundation for the Mid South before joining nities of Gentilly and the Lower Ninth Ward, as well as sticker during the RAND. “Th ey are long-term, transformational issues, the mostly middle-class and white community of Lake- Krewe of Cork parade through and they cannot be solved overnight. Th ey are daunt- view—all of which were swamped with more than four the French Quarter ing. But now is our opportunity to confront them.” feet of water. Th e severely damaged areas accounted for of New Orleans on February 17, 2006. 52 percent of all pre-Katrina New Orleans households, Repopulation of New Orleans home to 55 percent of the city’s residents. Th e limits of local recovery eff orts are best embodied At the other extreme, about 25 percent of the pre- by the city of New Orleans itself. A RAND team Katrina population was not exposed to fl ooding, and their led by demographer Kevin McCarthy has projected housing likely suff ered little or no fl oodwater damage. that the population of New Orleans will rebound by Th ese neighborhoods—including the whiter, wealthier about half within three years of Hurricanes Katrina French Quarter and Garden District—are perched on the and Rita. By September 2008, the population of New natural levees on both sides of the Mississippi River. Orleans will reach an estimated 272,000—or 56 per- Th e repopulation of the “high and dry” areas has cent of the pre-Katrina population of 485,000. probably already exceeded more than 100 percent of Th e homes of 55 percent of the pre-Katrina popu- pre-Katrina levels, as more people have crowded into lation suff ered severe damage when the levees were the limited supply of safe and available housing. Future breached and deep fl oodwaters inundated parts of the housing redevelopment is likely to be concentrated in city. “Th e extent of rebuilding in the most severely these areas as well. In contrast, the RAND team has devastated areas of the city will largely determine the projected that repopulation in the severely damaged pace of repopulation and also the total future popula- areas will reach just 30 percent by 2008 (see Figure 2). tion,” said McCarthy. “New Orleans before Hurricane Katrina was a Figure 1 maps the neighborhoods of New Orleans poor city that was highly divided by race and class,” according to their fl ood depths and the consequent said McCarthy. “Many poor residents were tenants severity of housing damage experienced as a result of who paid extremely low rents.” Th ese people may no Hurricane Katrina. Areas in the severe damage category longer be able to aff ord the higher rents that could

WWW.RAND.ORG RAND REVIEW / F A L L 2006 19 Figure 1— Flooding Was Most Severe in Low-Lying Areas, While environments of communities statewide. Th e prospect Neighborhoods on Natural Levees Had No Flooding of building innovative kinds of aff ordable housing St. Tammany Parish typifi es the promise of going beyond “rebuilding” toward “reshaping” the Gulf Coast into something better than ever. Th e most devastated areas of Mississippi are its Lake Ponchartrain Village coastal counties of Hancock, Harrison, and Jackson, De L’Est Viavant/ Venetian Isles where the storm surge ruined stately waterfront homes as well as modest rental units. Generally, though, high New Orleans East levels of poverty and low levels of home ownership Gentilly Lakeview Lake Borgne characterize the counties. A substantial share of the Lower Ninth Mid-City Ward destroyed homes in Mississippi were also built before Mid-City Bywater

CBD St. Bernard Parish 1980 and thus did not adhere to updated building Uptown Algiers Garden safety codes that stipulate higher foundations and District New Aurora/ English Turn more wind-resistant walls and roofs. Jefferson Parish Plaquemines Parish “As in New Orleans, the challenge of recovering

Flood depth: from the hurricane has been exacerbated by the fact that No flooding < 2 feet (minor) 2–4 feet (serious) > 4 feet (severe) some of the poorest communities were also some of the SOURCE: Flood depths calculated by GCR and Associates, Inc., from Federal Emergency Management hardest hit,” said Jack Riley, who launched the study of Agency data on water levels for September 10, 2005. NOTE: CBD = French Quarter/Central Business District. aff ordable housing in Mississippi as associate director of RAND Infrastructure, Safety, and Environment. His RAND colleagues estimated that 81,000 Figure 2— Projected Repopulation of New Orleans Neighborhoods homes—more than half the housing stock in the Depends on the Depth of Flooding coastal Mississippi counties—were exposed to either 300 wind or fl ooding or both. Because households with 268 Pre-Katrina population (in thousands) income levels below the U.S. median occupied about a 250 Projected 2008 population (in thousands) third of these homes, the team suggested that 27,000 Projected 2008 repopulation rate (in percentages) 200 aff ordable housing units might need to be rebuilt within the three counties alone. 150 134 122 Housing is defi ned as “aff ordable,” said Riley, when 110% 100 100% a family does not need to spend more than 30 percent 80 68 of its income on rent or on mortgage payments, insur- 50 45% 27 27 31 30% ance, and property taxes. RAND has proposed three 0 additional guidelines for the Gulf Coast: Th ere should No flooding < 2 feet (minor) 2–4 feet (serious) > 4 feet (severe) be an adequate supply of aff ordable housing across mul- Flood category tiple income levels; new aff ordable housing units should

SOURCE: The Repopulation of New Orleans After Hurricane Katrina, 2006. meet sustainable building and design standards; and aff ordable housing should be located within a reason- able distance of suitable employment opportunities. be demanded for reconstructed dwellings. “Th ere are RAND researchers have presented Mississippi likely to be major demographic and socioeconomic elected offi cials with a menu of options to pursue these disparities in the repopulation of New Orleans,” guidelines. To oversee housing recovery eff orts, for McCarthy concluded. example, Mississippi could create a new state entity that would set spending priorities, educate residents Sustainable Affordable Housing about rebuilding, and coordinate housing redevelop- In Mississippi, three coastal counties have become a ment with other regional planning goals. Th e oversight laboratory for new types of aff ordable housing that and coordination would accelerate the rebuilding. could not only shelter displaced residents but also To increase the supply of aff ordable housing, enhance the housing stock, economies, and natural Mississippi could create housing trust funds (possibly

20 RAND REVIEW / F A L L 2006 WWW.RAND.ORG infused with cash from retail, tourism, gambling, or Mississippi State Senator oil revenues); encourage employers to off er housing Billy Hewes III assistance to employees; give developers incentives to (R-Gulfport) stands before a build more aff ordable units; adopt zoning codes that coalition of law- favor mixed-use development; work with community makers, architects, contractors, land trusts or nonprofi t groups to manage long-term public offi cials, aff ordable rental units; and off er low-interest mort- and public safety experts at the gages to qualifi ed low-income homeowners. capitol in Jackson, To make aff ordable housing safer, Mississippi Mississippi, on January 19, 2006, could mandate tougher minimum building codes in calling for stronger hurricane-prone areas. Th e state could also develop building code safety codes specifi cally for modular housing, and legislation to help communities modular-housing factories could be opened in commu- weather future natural disasters. nities that need lots of rebuilding—and employment. AP PHOTO/ROGELIO SOLIS To sustain aff ordability for the long term, Mississippi could target the notoriously high maintenance and utility to create dynamic, diverse, and economically vibrant costs of cheaply built, energy-ineffi cient aff ordable hous- communities,” said Penick enthusiastically. “Aff ord- ing units. Th e state could off er incentives for builders, able housing will be critical to rebuilding the economy lenders, and insurers to construct, fi nance, and reduce in coastal Mississippi. If these eff orts succeed along the premiums for homes that use less water and energy. coast, they could also be used to reduce the aff ordable- And to promote greater local involvement in housing gap in other areas of the region that weren’t so rebuilding eff orts, Mississippi could launch a program aff ected by Katrina.” for training aff ordable-housing contractors; hire and train more building code offi cials; and incorporate Student Displacement and Trauma community feedback into state and local housing deci- A RAND study of Louisiana public schools in the sions. Mississippi communities will benefi t greatly if wake of Hurricanes Katrina and Rita has found that the local workforce can play a larger role in rebuilding, the storms displaced nearly 200,000 students—more and the eff orts will be more successful if supported by than a quarter of the state’s pre-storm enrollment. community groups. More than 81 percent of the displaced students came “From the earliest days after Hurricane Katrina, it from the most populous parishes in the heavily hit was obvious that housing for displaced coast residents areas: Orleans and Jeff erson parishes in the east (from would be the dominant issue for a long time,” said Hurricane Katrina) and Calcasieu Parish in the west Mississippi Governor Haley Barbour. “I certainly wel- (from Hurricane Rita). come specifi c recommendations that enhance the stra- Th e displacement persisted throughout the entire tegic rebuilding and renewal plan that is under way.” 2005–06 school year, with 55 percent of the displaced In October 2006, the governor announced that students ending the year outside their original schools the National Association of Realtors and the Mis- (see Figure 3). Th e students who remained displaced sissippi Association of Realtors had donated money were disproportionately minority students and those to help RAND extend its study of housing needs who had been achieving poorly prior to the storms. along the Mississippi coast. Th e study will compare the damage done across various housing submarkets; assess which ones are likely to rebuild on their own or with existing assistance programs; and identify what The prospect of building innovative further initiatives might off er the greatest promise for kinds of affordable housing typifi es the assisting those Mississippians, particularly renters in search of aff ordable housing, for whom the available promise of going beyond “rebuilding” programs and resources are inadequate. toward “reshaping” the Gulf Coast into “Th e state of Mississippi and the cities and coun- ties along the Gulf Coast have a huge opportunity something better than ever.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 21 from untreated mental health problems or continued loss of schooling.” More than a year after the hurricanes, some school mental health professionals in Louisiana, Mississippi, Alabama, and Texas still perceive a need for mental health care for their students, while others think the stu- dents have settled into their new schools and are adjust- ing well. In either case, the majority of the professionals interviewed by RAND have expressed a limited ability to implement mental health programs for their students because of shortages in funding and in trained staff . To help both the students and the schools, RAND has designed a tool kit, or guidebook, that school administrators can use to fi nd mental health programs

AP PHOTO/GERALD HERBERT for students exposed to traumatic events, such as vio- lence or natural disasters. Th e tool kit compares 24 Marvinique Gray, But even among those students who have returned programs that have been used by schools across the 6, arrives for fi rst grade as children to their original schools, “a substantial amount of country and in nations beset by ongoing wars and return to the Alice schooling was lost, and the eff ects of the storms lin- acts of terrorism. Th e kit outlines which programs Harte Elementary School in the ger,” said John Pane, a RAND education researcher. serve which student needs. It also describes the time, Algiers section of From a survey of Louisiana school principals, he has training, and other resources needed by school offi cials New Orleans on discovered that the displaced students, both those who to implement each program and discusses potential December 14, 2005, for the have returned and those who have enrolled elsewhere sources of funding for implementation. fi rst time since in the state, exhibit several common symptoms of “Following Hurricanes Katrina and Rita, schools Hurricane Katrina. trauma. Th e displaced students are more likely than were in a unique position to help displaced students others to engage in fi ghting, arguing, bullying, eating but had limited information about how to help,” alone, playing in isolation, and violating school rules; said Lisa Jaycox, a RAND behavioral scientist who and less likely to engage in school clubs, activities, spearheaded the kit’s development. “It is essential that social events, or sports teams. schools know what is out there so that they can choose “Principals frequently reported that displaced the program that best fi ts their students’ needs and for students were more likely than preexisting students which they have the appropriate resources.” to need mental health counseling,” noted Pane. “It Among the programs described in the kit are the will be necessary over the coming years to help these Friends & New Places program, used by the Dallas students recover and to prevent any further damage Independent School District to help 1,000 students displaced by Hurricane Katrina; the UCLA Trauma/ Grief Program for Adolescents, a program used in Figure 3— More Than Half of Displaced Louisiana Students Did New York City schools following the terrorist attacks Not Return to Their Original Schools During 2005–06 of September 2001 and in schools in postwar Bosnia; Displaced Public School Students Returned to original schools without and Cognitive-Behavior Intervention for Trauma in enrolling elsewhere in Louisiana Schools, a program designed by RAND, UCLA, and 52,978 Returned to original schools after the Los Angeles Unifi ed School District to treat stu- 65,397 having enrolled elsewhere in Louisiana dents exposed to any form of trauma. Did not return to original schools but “Exposure to traumatic events can lead to anxiety enrolled in other Louisiana schools and depression and cause some students to act out in Did not enroll in any Louisiana public 41,870 school, at home, and among their peers,” said Jaycox. 11,863 school “Th ese programs, which have been developed specifi - SOURCE: Student Displacement in Louisiana After the Hurricanes of 2005: Experiences of Public Schools and Their Students, 2006. cally for use in schools, can reduce emotional and behav- NOTE: Pie chart accounts for the 172,108 displaced Louisiana public school students for whom ioral problems while also fostering resilience for future consistent records were available. events.”

22 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Flood Zone Management History shows that building “bigger and better” fl ood History shows that building “bigger and protection systems in the aftermath of disastrous storms is not always the best policy. Instead, the best way to better” fl ood protection systems in the reduce future losses of life and property is sometimes aftermath of disastrous storms is not to surrender land to the water. By forgoing develop- ment of fl oodplains or allowing wetlands to reclaim always the best policy. land, policymakers can protect the public, safeguard the economy, and manage natural resources. “Th e critical concept that fl ood damage control of 1948; the Dutch province of Zeeland, where high includes conceding land to the water from time to time tides and a huge storm overwhelmed the sea defenses is a psychologically diffi cult one,” acknowledged James in January of 1953; the upper Mississippi River region, Kahan, a RAND social psychologist whose team of where extensive river fl ooding submerged the sur- researchers chronicled the lessons learned from four rounding communities in the summer of 1993; and major fl oods around the world during the latter part the Yangtze River in China, where a similar inunda- of the 20th century. “Th ere is an inherent bias toward tion occurred fi ve years later (see the table). recreating what used to be. But there are times when Th e historical cases illustrate an evolution in an excess of cure can be worse than the disease.” thinking that has taken place over the past 60 For at least fi ve decades, the Gulf Coast wet- years, said Kahan. “All illustrate the evolution from lands that form a natural hurricane protection system fl ood control to a broader notion of integrated water around the Mississippi delta have been eroded by resource management.” Th e shift goes from a near- the expansion of levees, the dredging of canals and exclusive focus on structural means of controlling shipping channels, and the ongoing pumping of oil fl oods (such as building dams and levees) to the inclu- and gas. All these activities have caused the naturally sion of nonstructural means (such as zoning laws and restorative sediment of the Mississippi River to fl ow other regulations) in order to stretch water resources out to the deep waters of the Gulf of Mexico rather as far as possible, to allocate them equitably across dif- than to replenish the wetlands. ferent social and economic groups, and to protect the To glean lessons that might apply to the Gulf water resources and their associated ecosystems. Coast recovery eff ort, Kahan and his team focused on In Oregon, the part of Portland that was once Van- the process of recovery from the four historical fl oods. port has since become a park and a major recreational Th ey are Vanport, Oregon, where the Columbia River center for the city. Th e remarkable aspect of Vanport is crashed through a protective dike on Memorial Day that it was never rebuilt. Th is owes less to enlightened

Recovery Process from Four Major Floods Contains Lessons for U.S. Gulf Coast Economic Geographic Type of Population of Damage Case Date Location Catastrophe Affected Area Lives Lost (in U.S. dollars) Vanport May 30, 1948 Columbia River Failure of enclosing dike 3 million 15–32a $100 million basin near during river fl ood Portland, Oregon Zeeland January 31, Southwest part of Storm surge overwhelms 300,000 1,835 $800 million– 1953 the Netherlands sea defenses $1.1 billiona Mississippi June to August Upper Mississippi River fl ood overwhelms 64 million 47–52a $16 billion 1993 River levee system Yangtze June to August Yangtze River Severe river fl ood 71 million 1,562 $20.5 billion 1998 basin, China exceeds defenses Katrina August and U.S. Gulf Coast Hurricane storm surges 38 million 1,840–2,000a $85 billion– and Rita September, overwhelm levees and $91.2 billiona 2005 coasts a Number varies depending on source. SOURCES: From Flood Control to Integrated Water Resource Management, 2006; U.S. Census Bureau; Federal Emergency Management Agency.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 23 in U.S. fl oodplain policy, particularly less development “There’s an extraordinary commitment in fl oodplains and more individual liability for those living there. However, development in Midwestern among people who have returned fl oodplains has continued to grow since 1993. Alterna- to try to build something new. tives such as property buyouts, wetlands restoration, and insurance reforms have not been employed as There’s unprecedented collaboration extensively as recommended. “Apparently,” a 2004 and selfl essness.” editorial in the St. Louis Post-Dispatch groaned, “we have learned nothing from the fl ood.” In contrast, after the Yangtze fl ood and mudslides water management policy than to widespread ambiva- of 1998, the Chinese government explicitly stated that lence about this federally built city that was created what is important is not to control but to manage water to house out-of-state residents—many of whom were resources. In addition to reinforcing levees, dredging black—who had come to build Liberty ships during riverbeds, and completing the massive Th ree Gorges World War II. Nonetheless, Vanport proved that if the Dam, Chinese water management policy now also social, political, and cultural circumstances permit, emphasizes creating mountain forest preserves, return- recovery need not be an attempt to restore the status ing agricultural lands to forests and lakes, and demol- quo but can instead serve broader societal purposes. ishing dikes to create fl oodplain water catchment areas. In the Netherlands, the 1953 storm triggered a “From a water management point of view, the Viewed from rethinking of water management that continues to this Yangtze River provides close to an ideal case,” said the air, a section of the Chandeleur day. Th e centuries-old Dutch tradition of building ever- Kahan. “A multilevel integrated water management Islands ravaged higher dikes gave way to a combination of dikes, dams, system has been put into place that refl ects learn- by Hurricane Katrina appears and storm surge barriers that remain open except dur- ing from previous experience. Of course, all this has sculpted by wind ing major storms. In this way, the restoration preserved occurred within a culture that accepts top-down direc- and water off Bay St. Louis, the economic value of fi sheries and mussel beds as well tion and control to an extent that is unacceptable in Mississippi, on as the maritime ecology. Today, the region preserves most Western cultures, much less the American one.” October 25, 2005. The hurricane its agriculture and draws tourists, not least because of He hopes, though, that Gulf Coast leaders can see vastly reshaped the these striking technological innovations. the promise that lies in the historical lessons. “Out of storm-absorbing barrier islands in In a stunning revolution for the Dutch, lead- tragedy can come opportunity,” he affi rmed. “Disrup- the Gulf of Mexico. ing policymakers are currently proposing that some tion of the status quo can create political conditions land once wrested from the sea be for broader-based social and economic change that returned. Historically pioneers in might otherwise have been delayed or might not have claiming land from the sea, the happened at all.” Dutch are now again at the fron- tier of water management in their Community Health willingness to give land back to Numerous RAND projects are trying to help local, the sea in exchange for enhanced state, and federal leaders make the Gulf Coast better, safety, economic value, and envi- stronger, and healthier than it was before. RAND is ronmental preservation. helping the U.S. Army Corps of Engineers develop a Th e 1993 fl ood along the long-term plan to protect Louisiana from hurricanes upper Mississippi River exposed and to restore its coastal habitats and ecosystems. the limits of levees that had been Another RAND project is culling the lessons that can constructed by local and federal be derived from the responses to 13 natural disasters governments throughout the pre- in 11 countries. A pending RAND “roadmap” for the ceding 150 years. In response, Gulf South is charting how various regional factors the federal government asked for can help or hinder future recovery eff orts. advice from the Interagency Flood- Meanwhile, RAND health researchers in New plain Management Review Com- Orleans, Baton Rouge, Houston, and San Antonio mittee. It called for major changes continue their eff orts to understand the physical and AP PHOTO/ROBERT F. BUKATY

24 RAND REVIEW / F A L L 2006 WWW.RAND.ORG mental health consequences and needs of hurricane evacuees and returnees. Among them is physician Ben Springgate. “Overall,” he reported from New Orleans, “most evacuees have not returned to their original neighbor- hoods or communities. Many who have returned are experiencing signifi cant symptoms of mental distress. Th e prevalence of symptoms of depression, anxiety, and post-traumatic stress disorder has about doubled.” Most of his patients tell him they’re not sleeping at night, they’re crying unexpectedly all the time, or they’re drink- ing more. “Th e day-to-day stress is palpable,” he said. In interviewing Katrina survivors, he has found that they have less access to health care and medica- tions now than before the hurricane. But the com- munity has responded by creating new models of care, such as mobile medical units and church-based health services. Many local organizations are now working KENNETH WELLS side by side to serve the returning population. “Th ese clinical partnerships would not have existed with the displaced people who still hope to return to Ben Springgate, a physician prior to the storm,” said Springgate. “Th e good news is New Orleans,” he declared. “We are trying to make whose work has that the community is drawing on novel resources and it possible for them to return to the home that they led him to UCLA and RAND and pursuing innovative adaptations to fi ll the gaps in care.” would want to return to: one that is better than before. back home again, He has organized a partnership between local We’re not giving up. We’re here for the long haul.” ■ embraces the prevailing senti- community groups and national academic institutions, ment outside the including RAND, UCLA, and Tulane. Consequently, Common Ground Health Clinic, community members are beginning to use research to Related Reading a volunteer-run assess their health care challenges and to resolve them. free clinic that People have also found great support from religion, Expanding Coverage to the Uninsured of Louisiana, Kavita emerged in one Patel, Susan Marquis, Sai Ma, Ben Springgate, RAND/WR-311, of the poorest spirituality, and one another, he emphasized. “Th ere’s 2005, available online only, at www.rand.org/pubs/working_ neighborhoods an extraordinary commitment among people who papers/WR311/. of New Orleans in the aftermath of have returned to try to build something new. Th ere’s From Flood Control to Integrated Water Resource Manage- Hurricane Katrina. unprecedented collaboration and selfl essness among ment: Lessons for the Gulf Coast from Flooding in Other Places in the Last Sixty Years, James P. Kahan, Mengjie Wu, Sara Haji- local agencies trying to fi ll the gaps and develop services amiri, Debra Knopman, RAND/OP-164-RC, 2006, 66 pp., ISBN for the people who were aff ected. Despite the day-to-day 978-0-8330-3984-2, $20. frustrations and lows, there is hope and tremendous How Schools Can Help Students Recover from Traumatic Expe- civic pride. People really want to see the community riences: A Tool Kit for Supporting Long-Term Recovery, Lisa H. Jaycox, Lindsey K. Morse, Terri Tanielian, Bradley D. Stein, bounce back and succeed.” RAND/TR-413-RC, 2006, 74 pp., ISBN 978-0-8330-4037-4, $23. He fi nds inspiration from those in his midst. “A lot Rebuilding Housing Along the Mississippi Coast: Ideas for of people could leave or never come back. It’d be under- Ensuring an Adequate Supply of Affordable Housing, Mark standable if they did. But I honestly believe we have the A. Bernstein, Julie Kim, Paul Sorensen, Mark Hanson, Adrian Overton, Scott Hiromoto, RAND/OP-162-RC, 2006, 92 pp., ISBN ability to take advantage of this break in the momentum 978-0-8330-3949-1, $20. and become a community that has a better education The Repopulation of New Orleans After Hurricane Katrina, Kevin system, is protected from natural threats, has a robust F. McCarthy, D. J. Peterson, Narayan Sastry, Michael Pollard, economy, and is still loved as much as ever for off ering RAND/TR-369-RC, 2006, 60 pp., ISBN 978-0-8330-3940-8, $18. such unique music, architecture, culture, and cuisine.” Student Displacement in Louisiana After the Hurricanes of He foresees the work ahead as a “ten-year process,” 2005: Experiences of Public Schools and Their Students, John F. Pane, Daniel F. McCaffrey, Shannah Tharp-Taylor, Gary J. specifi cally with respect to building enough aff ordable Asmus, Billy R. Stokes, RAND/TR-430-RC, 2006, 156 pp., ISBN housing for low-income workers. “We are in solidarity 978-0-8330-4119-7, $30.

WWW.RAND.ORG RAND REVIEW / F A L L 2006 25 Road Less Traveled Four Developing Countries Blaze New Trails to Better Health

By Thomas W. Croghan We at RAND recently investigated the factors that appear to have ameliorated these conditions and Tom Croghan directed this study as a senior natural sci- improved the health status of children in Bangladesh, entist at RAND. He is now a senior fellow at Mathematica Ecuador, Egypt, and Indonesia, four developing coun- Policy Research, Inc., and a professor in the Departments tries that throughout the 1990s reduced child mortality of Medicine and Psychiatry at Georgetown University at rates exceeding what might be expected from their School of Medicine, both in Washington, D.C. Board certi- poor economic circumstances. We sought to identify fi ed in internal medicine and rheumatology, he practices the common characteristics that could be associated primary care medicine at the Washington Free Clinic. with the ability of these countries to overcome their economic constraints. Our conclusions cast doubt on some long-held assumptions about the best approaches orldwide, more than 10 million children to improving health in developing countries. A volunteer, right, administers polio die each year, with nearly three-quarters In each country, the so-called contextual factors vaccine to a boy of these deaths attributable to just six of economic development, good governance, and as his mother W looks on during a causes: pneumonia, diarrhea, malaria, neonatal sepsis, strong health care systems mattered less than did the polio eradication premature birth, and suff ocation during the birth pro- amount of foreign aid and technical assistance provided drive in Dhaka, Bangladesh, on cess. Poor sanitation and inadequate nutrition add to and the eff ective implementation of targeted health April 16, 2006. the risk for these and other dangers. interventions, such as vaccination programs and fam- ily planning. Th ese fi ndings contradict prevailing U.S. foreign policy regarding the circumstances in which progress toward health goals can be made.

Decades of Debate During his inaugural address on January 20, 1949, President Harry S. Truman declared: “We should make available to peace-loving peoples the benefi ts of our store of technical knowledge in order to help them realize their aspirations for a better life.” Ever since Truman recognized the importance to U.S. interests of improving health in developing nations, there has been a lively debate about the most eff ective means to accomplish that goal. To be sure, a great deal of theoretical and empirical research over the decades has confi rmed the importance AP PHOTO/PAVEL RAHMAN

26 RAND REVIEW / F A L L 2006 WWW.RAND.ORG of contextual factors, such as a strong economy, bet- oping countries. Some of this research suggests that ter education, adequate nutrition, poverty reduction, income growth is not only necessary but is sometimes equity, and eff ective government; a functioning public suffi cient to improve health, prompting an oft-heard health system that provides sanitation, clean water, and cry of “wealthier is healthier” in some U.S. foreign infection control; and a comprehensive primary health policy circles. care delivery system. However, some countries with Money does help purchase medical supplies, cleaner ample resources and few barriers to health improve- water, and better food, of course, and could contribute ment, especially many oil-rich nations, have lagged to well-being in other ways. But the reverse could also be behind. Meanwhile, other countries have made great true: Good health has been proposed strides despite resource limitations, tropical diseases, as a determinant of national wealth. internal strife, and other obstacles. Th us, although con- In particular, better health at the Our conclusions text helps, we reasoned that it might not be the decisive national level results in an increased cast doubt on factor associated with improved health in many cases. number of healthy workers, especially In our analysis, we focused on countries that saw for labor-intensive industries. Costa some long-held sharp declines in child mortality of at least 40 percent Rica, now among the ranks of upper- assumptions about between 1991 and 2000—and where health expendi- income nations, is a case in point. tures in 2000 totaled no more than $150 per person Th is pattern is especially apparent the best approaches (in U.S. dollars adjusted for purchasing power). We in sub-Saharan Africa, where dra- to improving health excluded countries with very small populations (such matic increases in life expectancy as Cape Verde and Vanuatu) and those with very large in the 1970s, followed by equally in developing populations (such as India and China). Bangladesh, dramatic declines accompanying the countries. Ecuador, Egypt, and Indonesia met our criteria. Each HIV/AIDS epidemic, appear to have achieved very signifi cant reductions in child mortal- precipitated parallel swings in labor force participation ity despite remaining quite poor and spending less and country-level gross domestic product. on health than did countries that saw much smaller Figures 1, 2, and 3 compare several measures of reductions in child mortality. income and child mortality in the four countries of All four countries face major obstacles on the road interest with four other countries from the same geo- to health improvement. Bangladesh is the world’s most graphical regions. Th e comparison countries enjoyed densely populated country and the eighth most popu- considerably higher incomes and spent more on health lous, with 136 million people. It is among the world’s yet experienced much lower reductions in child mor- poorest countries and has few resources to devote to tality. Specifi cally, we compare Bangladesh with Paki- health, nutrition, education, and other public services. stan, Ecuador with Guyana, Egypt with Algeria, and Ecuador, the least populous of the four, with 13 million Indonesia with Th ailand. people, has nearly 85 percent more income per person Th e contrasts are clear. Bangladesh had nearly and spends two-thirds more per capita on health than 15 percent less income per capita but three times the does Bangladesh. But poverty has been a challenge— reduction in child mortality compared with Pakistan. more than half of Ecuador’s population in 2000 lived Ecuador had 30 percent less income per capita but on less than $2 a day. Egypt, with 66 million people, almost 2.5 times the reduction in child mortality com- has a national income more than twice that of Bangla- pared with Guyana. Egypt had nearly 40 percent less desh, about half the poverty, and little malnutrition, income per capita than Algeria, where child mortality but Egypt’s illiteracy and other education measures increased during the 1990s. And Indonesia’s reduction are comparable to those of Bangladesh. In Indonesia, in child mortality was more than 50 percent greater despite rapid economic development, less than half of than that in Th ailand, despite having more than 50 all children under fi ve have been vaccinated against percent less income per capita. measles, and malnutrition is very common. We also considered the possibility that income growth, less poverty, or greater economic equity was Wealthier Means Healthier? associated with better health. Th e rapid economic Much research has emphasized the critical nature of growth throughout Southeast Asia during the 1980s economic development in promoting health in devel- and 1990s could account for much of the improvement

WWW.RAND.ORG RAND REVIEW / F A L L 2006 27 Figure 1— In Reducing Child Mortality Rates, Some Neighboring in health in Indonesia, but the economic growth there Countries Are More Successful Than Others . . . of 8.3 percent between 1975 and 2000 was no greater 120 than the concurrent 9.2 percent growth in Th ailand. Deaths per 1,000 children under 5 in 2000 100 110 Percentage change, from 1991 to 2000, in child mortality rate And while Egypt had higher growth per capita than did 80 82 Algeria, both Ecuador and Bangladesh had lower rates 74 60 65 of growth than did Guyana and Pakistan, respectively. 40 43 48 32 29 In terms of poverty, the higher-achieving country 20 23 0 in each case had a higher percentage of people living on –20 –14 –18 less than $2 per day than did the comparison country, –40 –28 often dramatically so. And if greater income disparities –43 –44 –47 –60 –59 were associated with higher child mortality rates, then –80 Bangladesh Pakistan Ecuador Guyana Egypt Algeria Indonesia Thailand the Gini coeffi cient—a common measure of economic South Asia South America North Africa Southeast Asia equity—would be lower for the higher-achieving coun- SOURCE: Routes to Better Health for Children in Four Developing Countries. tries. Nevertheless, the coeffi cients were nearly identical across each comparison, except for Indonesia, where inequalities were less severe than those in Th ailand. We also found political development to be as dubi- ous a determinant as economic development for improv- Figure 2— . . . Even Though the Comparatively Successful Countries Generate Less Income Per Capita . . . ing health. Case studies by the Rockefeller Foundation 7,000 in the 1980s examined three countries (Costa Rica, $6,230 $5,910 China, and Sri Lanka) and one state in India (Kerala) 6,000 that had achieved good health despite great economic 5,000 $4,280 obstacles. Th e studies concluded that the single most 4,000 $3,560 important element underlying the health achievements $2,960 3,000 $2,830 in these cases was a strong political commitment to good $1,860 health. Th e studies identifi ed fi ve elements of “political 2,000 $1,600 will”: historical commitment to health as a social goal, 1,000 a social welfare orientation to development, widespread for purchasing power parity for purchasing in U.S. 2000 dollars adjusted

Gross national income per capita Gross 0 Bangladesh Pakistan Ecuador Guyana Egypt Algeria Indonesia Thailand participation in political processes, equity in access and South Asia South America North Africa Southeast Asia use of social programs, and eff ective intersectoral link-

SOURCE: Routes to Better Health for Children in Four Developing Countries. ages allowing for coordination and cooperation. In spite of this prior research, we found little evidence of political will or eff ective governance in the countries we studied that had achieved better child health. We did not fi nd consistent evidence of Figure 3— . . . Endure Higher Poverty Rates, and Exhibit Similar Levels of Income Inequality political consensus on or a historical commitment to health, nor did we fi nd much evidence of the types of 90 Percentage of population living on less than $2 per day in 2000 social welfare orientation to development observed by 80 83 Equity index (Gini coefficient): a higher number means greater the Rockefeller Foundation. Bangladesh, Ecuador, and 70 inequality 66 60 Indonesia rank among the world’s most corrupt devel- oping nations, and we found little evidence for partici- 50 52 52 40 44 45 44 43 patory governance and decentralized decisionmaking. 30 33 34 35 33 For example, crackdowns on extremists in Egypt and 32 30 20 Indonesia had led to a tightening of central controls 10 15 and authoritarian processes. Th e contrast between the 6 0 four countries we studied and those highlighted in the Bangladesh Pakistan Ecuador Guyana Egypt Algeria Indonesia Thailand Rockefeller case studies is dramatic and suggests that South Asia South America North Africa Southeast Asia while the fundamental elements of political will might SOURCE: Routes to Better Health for Children in Four Developing Countries. help, their absence can be overcome.

28 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Remaining Clues will help us understand whether healthier could very We did, however, uncover certain characteristics that well mean wealthier. the four countries of interest shared in common with Th e second lesson is that the absence of politi- respect to reducing child mortality. First, in each case, cal will or good governance does not doom eff orts data that allowed for the identifi cation of the specifi c to reduce child mortality in developing countries. health conditions associated with child mortality were Violent political change, corruption, and ineff ective available and appear to have been used eff ectively to government were common features in each of the four target delivery of specifi c health interventions for those countries we studied, and participation in the political most in need of them. process was rare. Second, although each of the individual interven- Th e third lesson is that foreign aid and technical tions probably accounted for only a small portion of assistance appear to be important. Each of the four the overall health improvements, multiple interventions countries received considerable fi nancial assistance. were introduced in each country in a highly project- Technical assistance in the form of disease surveillance specifi c and targeted manner. Th e situation in Bangla- was especially important for targeting interventions to desh, where the World Bank supported 128 individual those most in need. health projects in 1998, is especially informative. Th e fourth lesson is that public health and medical Th ird, none of the countries we studied had well- services do not require the comprehensive framework, developed, accessible health care systems. In fact, the or the attendant level of coordination and bureaucracy, systems were not just underdeveloped and inaccessible, endorsed by the International Monetary Fund, the but also often unpopular. World Health Organization, and the World Bank. Fourth, the four countries in our study appeared Comprehensive services could even be counterproduc- to form more eff ective relationships with donors than tive. As a key informant from Pakistan told us, trying did less successful countries. Several key informants to accomplish everything dilutes resources and eff orts, for our study suggested that Bangladesh is especially making it less likely that anything will be done well. receptive to foreign aid and more willing and able than U.S. foreign policy, along with international other countries to endorse the conditions or stipula- trends in development assistance, emphasizes economic tions that often accompany such aid. development and limited direct fi nancial assistance Finally, development partners in each of the four in countries that meet certain criteria of government countries appear to have provided fi nancial assistance eff ectiveness, corruption control, and economic free- that mitigated the eff ects of weak economies and of dom. Much of the aid and debt relief from agencies low levels of internal public and private health spend- such as the World Bank and the International Mone- ing. It also appears that such fi nancial aid may have tary Fund are also now tied to development strategies increasingly bypassed government in favor of direct focusing on economic growth and poverty reduction. allocations to nongovernmental organizations. Th is is not the route followed by the exemplary countries studied here, for which generous amounts of Healthier Means Wealthier? direct foreign aid supported the targeted introduction With regard to improving children’s health in develop- of eff ective interventions despite great poverty, weak ing countries, our fi ndings suggest that targeted health economies, political turmoil, widespread corruption, interventions and foreign aid matter more than do and poor governance. Th e results favor an alterna- economic development, good governance, and strong tive approach that places a priority on setting explicit health care systems. Th e experiences of Bangladesh, health goals and using targeted interventions to reduce Ecuador, Egypt, and Indonesia off er several important child mortality. ■ lessons for health and foreign policy. Th e fi rst lesson is that economic growth, poverty reduction, and economic equity are not necessary to Related Reading reduce child mortality. But the reverse could very well “Routes to Better Health for Children in Four Developing be true and could explain the association between Countries,” The Milbank Quarterly, Vol. 84, No. 2, June 2006, health and wealth. Tracking the economic develop- pp. 333–358, Thomas W. Croghan, Amanda Beatty, Aviva Ron. ment of these four countries in the coming decades

WWW.RAND.ORG RAND REVIEW / F A L L 2006 29 Publisher’s Page

gency research being put into action. Th is work began Old Lessons, during the Vietnam era, continued into the 1980s and New Doctrine 1990s with work on international terrorism and on insurgencies in Nicaragua and El Salvador, and carries Classic Insights into on today as we advise senior U.S. civilian and military Counterinsurgency Are authorities about emerging global threats. Especially Trenchant Today Perhaps no researcher fi gures more prominently throughout this history than does RAND analyst By James A. Thomson Stephen Hosmer. He brought Galula to our attention James Thomson is president and chief executive offi cer of and authored numerous classic RAND reports that the RAND Corporation. continue—years, if not decades, after publication—to infl uence national security policy. Hosmer’s 1963 report on a RAND symposium, which brought together what n October, the New York Times reported that the U.S. is arguably the most distinguished group of counterin- IArmy and U.S. Marine Corps were drafting a new surgency and guerilla warfare experts ever assembled, counterinsurgency doctrine for use within an updated is being used by U.S. military and civilian offi cials in fi eld manual. Refl ecting lessons from Afghanistan and Washington and Baghdad today. Iraq, the doctrine places a high priority on protecting A report Hosmer wrote in 1986 was one of the civilians, restoring municipal services, and reconstituting fi rst to identify terrorism, subversion, and insurgency PHOTO: DIANE BALDWIN indigenous security forces. as three distinct forms of covert aggression. His 1990 In crafting the doctrine, the military services “drew report, years ahead of its time, called for the U.S. upon some of the classic texts on counterinsurgency military to develop counterinsurgency doctrine, to train by the likes of T. E. Lawrence of Arabia and David counterinsurgency specialists and units, and to create a Galula,” the Times noted. counterinsurgency institute. Hosmer continues to work Little known outside military circles during most on issues related to the ongoing counterinsurgency of his career, Galula was an important early contributor eff orts in Iraq and Afghanistan. to counterinsurgency research at RAND. As a consul- Others at RAND, notably Brian Jenkins and tant here in the early 1960s, he wrote a path-breaking Bruce Hoff man, have assessed ways to counter and to analysis, Pacifi cation in Algeria: 1956–1958, based on his infl uence al Qaeda and other emerging terrorist threats, command of French army troops during Algeria’s war examined past counterinsurgency campaigns for les- of independence. Under his command, French forces sons to apply to Iraq and Afghanistan, and explored cleared an insurgent-fi lled area near Algiers and restored how computer links and other networks are changing it to government control. Th e experience gave him a counterinsurgency theory and practice. ground-level view into what did and didn’t work. Although these studies have diff ered in tone, they In this RAND study and his later book, Counter- all echo Galula in calling for policymakers to build insurgency Warfare: Th eory and Practice, Galula argued counterinsurgency and counterterrorism eff orts on a that the key to counterinsurgency is support from the political as well as a military foundation. Th is mes- local population, requiring that military authorities sage has been a consistent theme running through 45 ensure the people’s safety and maintain eff ective police years of RAND research. Sadly, this message—and and other public services. Force alone will not succeed. similar ones from other experts in the United States Counterinsurgency succeeds when it melds the political and elsewhere—became lost after Vietnam, despite with the military. All these recommendations appear in the perseverance of those like Hosmer to keep it alive. today’s new counterinsurgency doctrine. Th at our message is now being heard and acted upon With the growing insurgencies in Iraq and Afghan- at the highest levels is a testament to the salience and istan, it is gratifying to see the lessons RAND has longevity of our eff orts in service of public welfare and learned in the course of fi ve decades of counterinsur- security—both at home and abroad. ■

30 RAND REVIEW / F A L L 2006 WWW.RAND.ORG Join RAND in Rebuilding the Gulf Coast

When Hurricane Katrina roared ashore on August 29, 2005, it laid bare many of the economic, environmental, and social justice issues that have long kept the Deep South from sharing in the progress enjoyed by much of the rest of the nation. Now, the region faces the long-term challenges of not only rebuilding the broken com- munities but also strengthening the governance systems for the people in those communities. Within a week of the hurricane, several leaders in the Gulf Coast region contacted RAND President Jim Thomson and asked him to bring RAND’s unique brand of research and analysis to bear on the region’s problems. Drawing on gifts from donors and other resources, RAND has since invested more than $2 million in assist-

AP PHOTO/BEN MARGOT ing Gulf Coast decisionmakers and establishing the RAND Gulf States Policy Institute, headquartered in Jackson, Mississippi. Although the nation has invested billions of dollars to assist the Gulf States, regional leaders need solid, objective, unbiased, non-partisan research to help them decide how to use the large amounts of available public and private resources most effectively. Donations to support the RAND Gulf States Policy Institute or other high-priority needs at RAND can be made by using the envelope inserted in this issue of RAND Review. If you would like additional information about the current work and activities of RAND in the Gulf Coast region, you can contact George Penick, director, RAND Gulf States Policy Institute, at (601) 979-2449 or e-mail [email protected].

I Want You! The Evolution of the All-Volunteer Force Bernard Rostker

“…A virtual archive of the many events, issues, facets, and fundamentals constituting the All-Volunteer Force, [which] exceeds by far, in my judgment, any prior attempts to explore this subject.” —From the foreword by Melvin R. Laird, former U.S. Secretary of Defense (1969–1973)

I Want You! reveals the inside story of how the American military transformed itself over the past fi ve decades, from a poorly disciplined force of conscripts and draft-motivated “volunteers” to a force of professionals revered throughout the world. Starting in the early 1960s, the account runs through the current war in Iraq. The author, who participated as an analyst and government policymaker in many of the events covered in this book, provides a behind-the-scenes look at Pentagon and White House decisionmaking. An accompanying DVD—available with the special DVD edition—contains more than 2,300 primary-source documents, including thou- sands of pages of government letters, staff papers, reports, and formerly classifi ed memoranda between presidents and secretaries of defense. More than 1,700 of these documents are linked from citations in the electronic version of the book, allowing readers to view in context the actual correspondence between government offi cials at the highest levels.

Send an email to [email protected] to receive a free DVD demo disc of I Want You!—or fi nd it online at www.rand.org/pubs/monographs/MG265/index2.html.

September 2006 • 832 pages | $68.50 • hardcover with DVD • ISBN 978-0-8330-3895-1 | $48.50 • hardcover • ISBN 978-0-8330-3896-8

Available at better bookstores, online stores, or directly from RAND. Join the RAND Policy Circle

THE RAND POLICY CIRCLE provides an opportunity for peo- ple throughout the community to experience RAND from the inside. Through direct interactions with RAND experts, solicited feedback, and philanthropic support, Policy Circle members contribute to RAND’s growth and the achievement of our mission.

Recently, RAND Policy Circle members had reserved seats at a standing-room-only Policy Forum on “A Vision for the Arts in Los Angeles: Opportunities and Challenges.” Other recent events included a roundtable discussion about unfolding events in the Middle East, an evening with best-selling author and political theorist Francis Fukuyama as part of RAND’s Distinguished Speaker Series, and a RAND Conversa- DIANE BALDWIN tion in Pittsburgh on “Music Lyrics and Sexual Activity Among Youth.” People who join the Policy Circle also receive selected policy briefs and newsletters to keep them apprised of RAND’s nationally acclaimed work on a wide range of national security and social policy topics.

For more information about how to join the RAND Policy Circle, please visit http://www.rand.org/giving/ about.html or contact us at (800) 757-4618.

The RAND Corporation is a nonprofi t institution that helps improve policy and decisionmaking through research and analysis. Contributions to RAND are tax-deductible to the extent allowed by law. Tax ID 95-1958142.

CP-22-0612

NONPROFIT ORG. R U.S. POSTAGE PAID Corporate Headquarters THE RAND 1776 Main Street, P.O. Box 2138 CORPORATION Santa Monica, CA 90407-2138

Washington Offi ce 1200 South Hayes Street Arlington, VA 22202-5050

Pittsburgh Offi ce 4570 Fifth Avenue, Suite 600 Pittsburgh, PA 15213-2665

RAND Gulf States Policy Institute P.O. Box 3788 Jackson, MS 39207-3788

RAND-Qatar Policy Institute P.O. Box 2364 4 Doha, Qatar

RAND Europe Westbrook Centre Milton Road Cambridge CB4 1YG United Kingdom