By Philip Zelikow

A recent New York Times story and today's column by David Ignatius have surfaced a dispute between the DNI and the director of the CIA about responsibilities for the selection of the "chief of station" at overseas posts.  (Tradtionally the chief of station is the senior CIA official at the post and oversees intelligence liaison relationships and operations in that country, as part of the country team headed by the ambassador.)

The DNI wishes to be able to select the head intelligence community representative in the country. The CIA director wants to retain his traditional authority to appoint all station chiefs. Ignatius argues, defending the CIA position, that the CIA does overseas operations and liaison relationships, so the CIA ought to look after the station chiefs.

The White House will ultimately settle this dispute. I write only to help readers understand why this issue is more complicated than one might think after reading the Ignatius argument.

First, overseas operations are much more than the human intelligence (HUMINT) collection that CIA manages. In some countries, the main overseas work -- and staffing to support it -- may concern signals intelligence or other technical operations that are managed by the National Security Agency or the National Reconnaissance Office, among others. In some cases the HUMINT role may also be eclipsed by defense intelligence relationships that are usually looked after by the Defense Intelligence Agency and the embassy's Defense Attache (which involves a set of challenging coordination problems in its own right). Then there are FBI/Legatt issues, Treasury issues, DHS issues, and so on.

Second, even where the CIA's role is large, even dominant, for the intelligence mission being carried out in the country, the CIA's role is rarely the whole story. Thus the station chief needs to be an interagency manager on behalf of the intelligence community, or else that job will be bumped to the ambassador or, more likely, get bumped back to Washington. In Washington, the DNI has significant authority to resolve the turf battles, but may not have direct authority over the station chief who should have been part of the solution, not part of the problem. Usually interagency issues are worked out pretty effectively in the field. But there are problem cases. Since the station chief should really be the representative of the intelligence community, not just a representative of one component of it, the DNI has a legitimate concern to ensure that station chiefs are appointed and managed in a way that conforms with their responsibility.

Third, it is the DNI's responsibility to manage the foreign liaison relationships. These relationships are a huge -- and little known -- dimension of U.S. intelligence policy of every kind. Having learned some lessons from years of initial experience with the new DNI structure, the U.S. government sought to clarify responsibility for managing these foreign relationships. After months of arguments and drafting exercises (involving the current Secretary of Defense and former CIA director, among others), the result was codified last year in a presidential Executive Order 13470 (30 July 2008). 

That order states that the DNI "may enter into intelligence and counterintelligence arrangements and agreements with foreign governments and international organizations." The DNI "shall formulate policies concerning intelligence and counterintelligence arrangements and agreements with foreign governments and international organizations." The DNI "shall align and synchronize intelligence and counterintelligence foreign relationships among the elements of the Intelligence Community to further United States national security, policy, and intelligence objectives."

So one can see why the DNI might feel some responsibility for the appointment of the intelligence community's overseas representatives. President Obama is certainly free to rewrite this executive order. If he and his advisers think it is worth their time to revisit this Executive Order, they will learn a lot about why it was written this way.

The most worrisome aspect of the current dispute is not its substance. It is an interesting problem, but there are several ways to work it out. One solution might draw on the role of the Secretary of State to the selection of ambassadors. The Secretary always recommends, usually gets her wish, but does not have the final power of decision.

Instead, the troubling aspect of the case is what it appears to say about the relationship between the DNI and the CIA director. That relationship is vital to the effectiveness of the intelligence community. Even before Obama took office, experience had shown why a president should take the time to be sure his appointees in these two positions work well together and understand their respective roles in the broader system.

By Philip Zelikow

The U.S. government is now bankrolling two substantial enterprises -- both too important to fail.

One is the automobile company, General Motors. The U.S. government appears to be putting up a total approaching $50 billion to keep the reduced company afloat. It is set to own 60 percent of the company.

The other is the Islamic Republic of Afghanistan, a sovereign government in central Asia. Purely by financial measures, the U.S. investment there is not quite as large as in General Motors. But since the prospects of getting the money back are not as good, perhaps the level of federal spending in this case can be regarded as roughly comparable. The federal government does not have a formal ownership stake in Afghanistan, but foreign donors pay more than 90 percent of the costs of operating the government. The United States foots most of this bill -- especially in security spending. And there is no prospect of Afghanistan being able to afford the security forces it has now, much less those proposed for it, any time in the foreseeable future.

Ordinarily a common description for the first case is that the company has been nationalized by the United States government. And a common description for the second case is that the state has become a protectorate of the power that pays for the army. (The United States does this in Afghanistan to a much greater degree than was ever the case in Israel, and to a much greater degree than was even the case in Iraq in recent years. In this respect, the proportion of U.S. sponsorship is a bit reminiscent of South Vietnam.)

The United States government does not like to use terms like "nationalization" or "protectorate." Why? Because these descriptions would make it sound like the U.S. government is in charge. Well, isn't it? In both cases, the answer is: Not quite. And in both cases the question is worth pressing a bit harder.

In the first case, GM, the U.S. government will own the company. But it will not offer or defend a business model for GM's success, since we disclaim any desire to run this company. (Though the government will impose certain new constraints on how the company can be run.) The argument is that, with the massive subsidy and the freedom from old constraints that will come from bankruptcy reorganization, the company should do fine. But some big questions arise:

(1) How does the U.S. government judge the viability of the business model?

(2) How does the U.S. government play this role while running a two-level game: one in which the government runs an auto company and the other in which the government has a carefully limited role in directing the private sector of the economy that trades automobiles, with competitors that include private U.S. companies (like Ford) and foreign companies manufacturing in the United States.

(3) If the U.S. owners are passive investors (not the case so far), who on the board will provide guidance and accountability for management?

In the second case, Afghanistan, the United States (and other donors) will subsidize the Afghan government. But the United States cannot authoritatively select a successful "business model" for the enterprise, since we disclaim any desire to run this country. (Though the United States will impose certain constraints on how Afghanistan can be run.) Yet here too, some large questions arise:

(1) How does the U.S. government judge the viability of the foreign government it subsidizes?

(2) How does the U.S. government play this role while running a two-level game: one in which foreign donors effectively run a parallel governance/security system in Afghanistan, with its own separate budget, and the other in which the United States has a carefully limited role working "alongside" the Afghan government's own governance/security structures. (There is another two-level game involving the seams between Afghan central and local authorities, but it is too complicated to develop in this post.)

(3) If the U.S. funders are relatively passive investors (with a 90+ percent stake!), who provides guidance and accountability for management? When the U.S. government provides almost all the money for a government, to what extent does it become responsible or accountable for what that government does and how that government is managed?

This is an old problem. In the history of various kinds of outside interventions in troubled enterprises and countries, the outside government often gets involved because the local enterprise is in extremis. It is in extremis because the status quo isn't working. So, in these cases, the outside government naturally starts thinking about regime change. General Motors has already had one change of management. But of course Afghanistan is different. It is a sovereign state...

I know these two stories are more complicated than I've let on in this little thought experiment. Yes, there are ways of sharing power. Still, in puzzling over the prospects of the U.S. government's investment in these two enterprises, I want to understand the strategy for future success and assess whether the strategy is viable. And this little thought experiment does spotlight threshold questions, like: Whose strategy? And who's responsible?

And then there is one more issue: If the enterprise is too important to fail, how can the U.S. government limit its exposure? Isn't it the case that, to have enough influence, the United States and its allies have to be willing to take responsibility, or be willing to walk away? And if the government is willing to disinvest, then isn't the government also saying that, if they are not adequately managed, the enterprises can be allowed to fail, after all?

By Philip Zelikow

An appropriate policy toward North Korea should quarantine and limit the threat the state can pose to the United States and its allies. U.S. diplomacy, properly conceived, should always have had two goals. First, to offer -- in good faith -- a genuine opportunity for the North to make a constructive strategic choice for the future.  Second, to strengthen U.S. and allied ability (political as well as military) to defend themselves if the North made a different choice.

Some people tend to emphasize only the diplomatic track; others only emphasize the defensive measures. The key point, which former Secretary Rice and former Deputy Secretary Zoellick understood very well, was that the first track is a necessary enabler for the second one. So in 2005, the United States reinvigorated the Six Party process to make the first track real. And in 2005, the United States took steps that effectively destroyed a Chinese bank in Macau, the Banco Delta Asia, illustrating America's readiness to pursue the other track as well.

This dual strategy heightened tension, culminating in North Korea's nuclear test of 2006. Yet the international response in 2006 displayed unanimity and firmness that had not been seen since 1953, evident in UN Security Council resolution 1718. The result was a fresh diplomatic opening, a promising agreement in February 2007, and a further test of North Korean intentions, one so specific and unequivocal that the results were bound to be revealing.

North Korean behavior in 2007 was indeed revealing. Despite some great pictures for CNN, North Korea failed adequately to account for its past nuclear trade, including possible transfers of enriched uranium to Libya and possible transfers of nuclear fuel (as well as much other help) to Syria. Although the known plutonium production facility was temporarily disabled, possible uranium enrichment facilities remained. Of course, the possible Libyan and definite Syrian choices were made in the past. But it was (and is) essential for the United States and its allies to develop some reasonable understanding of how that proliferation path worked -- and was funded -- to have adequate confidence that the path is gone.

Thus, during 2007, the United States and its allies could conclude that they would not be able to achieve a critical, realistic objective: a verifiable cap on North Korea's capacity to build nuclear weapons and produce weapons-usable nuclear material. Such a concrete objective would have been worth the candle -- a good prelude to a further, comprehensive phase of Korean diplomacy that would include the attainment of complete denuclearization, as required by UNSC 1718 and as pledged by North Korea in 1992, 2005, and 2007. Attainment of even that preliminary objective was in even greater doubt, though, given the evidence of 2007.

Nonetheless, the United States helped construct a further agreement (Beijing, October 2007) to keep the diplomatic process afloat rather than move it to a new phase. Why? I don't know. Today's Wall Street Journal editorial listed me as first, ahead even of Chris Hill and Condi Rice, in persuading President Bush to make the October 2007 decision to keep that diplomatic track alive and take North Korea off the terror list. That rank ordering in supposed infamy is especially bizarre, since I had left the administration at the end of 2006. (Perhaps someone wanted to sling something at me because of my stance on terrorism issues, and this was the only available clod of mud.)

The pros and cons of the October 2007 decision are hard for me to judge. I'm certainly inclined to give President Bush and Secretary Rice the benefit of doubt. Perhaps the moves to destroy the plutonium facility seemed so encouraging; the uranium enrichment concerns seemed wispy; and forcing the North to admit a past it could not acknowledge would seem merely backward-looking and punitive, rather than future-oriented and constructive.

Yet there were large downsides of keeping the process afloat with the October 2007 Beijing agreement, and they grew, especially as the Beijing agreement proved hollow. The uranium enrichment issues had been spotlighted by the new evidence on Libya and Syria ties and did not seem to be getting addressed. The coalition-building benefits with South Korea were diminishing, especially as the South Korean people repudiated the policy direction of the late president Roh Moo-hyun. The already-strained relations with Japan had to carry a heavier burden of mistrust. The bonds with China remained strong, but there was a danger of short-sightedness. As China effectively took on more responsibility as North Korea's protector and guarantor in the diplomacy, Chinese action or inaction on this topic could become another potential issue in an utterly vital connection: Chinese relations with Japan.

In any case, the United States definitely went the extra mile in its diplomacy. Now Washington can credibly offer coalition leadership in developing appropriate defensive measures of all kinds.

1. Sanctions?  It would be nice to enforce fully the ones already on the books in UNSC 1718.

2. Instead what is needed is international action by interested parties to redress the violation of UNSC 1718 with suitable defensive measures under Chapter VII. Either the UN should expressly authorize that, or note that this will happen, or the Security Council should remain silent. It set the international norm in 2006 and did so under Chapter VII.  The norm has been violated. Unless a further resolution is suitably empowering, silence might be best. The Security Council should not limit what can be done by specifying it.

3. The United States must now treat the North Koreans as having crossed the "red line" of proliferating nuclear material and, based on our analysis of how they did this, do everything possible to disable this capability.

4. Also, as I wrote in this space a few months ago, the United States should take necessary preparations with its allies to limit North Korean development of the ballistic missiles they could marry with their nuclear (or biological or chemical) payloads.

5. Keep in mind that all of this is a curtain-raiser for the Obama administration's still too-be-determined policy on Iran.

Certainly any measure that confronts North Korea carries risks of escalation. The North Korean government made the decision to act beyond its borders.  The United States should prepare with its allies to address these risks. Evidence of that preparation is the best way to reduce the risk. And our Chinese and Russian friends can judge for themselves how best to manage the risks they see arising from this cancer across the Yalu.

By Philip Zelikow

The two speeches on counterterrorism today by President Obama and former Vice President Cheney were each, in their way, exceptionally well-crafted addresses showcasing each man's rhetorical gifts. The two addresses repay careful study and comparison. But everything I could say to contrast the style and substance of the two speeches can be gleaned from this pair of quotations from them:

Mr. Cheney: 

Behind the overwrought reaction to enhanced interrogations is a broader misconception about the threats that still face our country. You can sense the problem in the emergence of euphemisms that strive to put an imaginary distance between the American people and the terrorist enemy. Apparently using the term 'war' where terrorists are concerned is starting to feel a bit dated. So henceforth we're advised by the administration to think of the fight against terrorists as, quote, 'Overseas contingency operations.'

President Obama: 

After 9/11, we knew that we had entered a new era -- that enemies who did not abide by any law of war would present new challenges to the application of the law; that our government would need new tools to protect the American people, and that these tools would have to allow us to prevent attacks instead of simply prosecuting those who try to carry them out....

Now let me be clear: we are indeed at war with al Qaeda and its affiliates. We do need to update our institutions to deal with this threat. But we must do so with an abiding confidence in the rule of law and due process; in checks and balances and accountability.

By Philip Zelikow

One short supplement to Will Inboden's good post on GOP foreign policy futures. The tendency to divide foreign policies into "idealist" and "realpolitik" is a telltale warning that frothy, superficial thinking lies ahead. For example, reflect briefly on the recent portrayals of the George H.W. Bush administration as one that exemplified "realpolitik."

Of course officials of that administration regarded themselves as capable and practical people. Elderly folks like me well recall when Republicans like George Shultz or Brent Scowcroft or Bob Gates or Dick Cheney (back then) regarded themselves, whatever their other differences, as the party of competence. Back then it was the Democrats who seemed trapped by shibboleths, harried by zealots, and uncomfortable wielding power.

Competence, though, did not mean indifference to deep political convictions or a commitment to preserving the status quo. The Bush administration's push for German unification and the transformation of Europe was hardly a play it safe approach. Contrast, for instance, the Bush administration's course in its European strategy with the quite different path recommended at the beginning of 1990 by Henry Kissinger himself (or George Kennan, for that matter) -- and those two men represented cautions then found across the spectrum of editorial and public commentary.

Or, perhaps there are those who think the obvious "realist" path in response to the Iraqi invasion of Kuwait in 1990 was to put 500,000 U.S. troops in the Arabian desert where before there had been none. Remember, this was the war that passed the Senate, after a real cliff-hanger debate, by a mere five votes. (Authorization for the 2003 war, the one that is now regarded as so much more discretionary, passed the Senate by fifty votes.)

As a historian, I think one of the more remarkable things about the Nixon-Kissinger approach to great power relations and détente is actually how anomalous it was in comparison to the record of America's international rhetoric and goals. That administration's relative indifference to the character and governance of the other states in the international system has no equal in any other other U.S. administration of the last 120 years.

The reasons for the Nixon-Kissinger anomaly are probably to be found more in the Vietnam War challenge and the upheavals around the world bucking the general ossification of the cold war system. There was a global retrenchment among governing elites across the globe in the early 1970s (a thesis Jeremi Suri has introduced in the last chapters of his Power and Protest). These more particular explanations seem more useful than arguments finding in this period the recurrent flowering of some long-running but dormant "realist" strain in America's collective thought. And the domestic base for the "détente" policy of that era had eroded almost to the vanishing point even by the end of 1974, eaten away from both left and right.

For at least the last hundred years, most full-throated critiques of how America should approach the world regard their views as realistic, whatever their argument. They all regard their foes as naïve or venal, people who either bury their heads in the sand or exaggerate threats to chase imaginary monsters. Arthur Link wrote quite thoughtfully of the "higher realism" of Woodrow Wilson.

So as Republicans wonder where they will find a foreign policy, please don't think the problem will be solved if only Republicans will be "realists" once more. On the other hand, there is a certain nostalgia in recalling a team that took so much pride in professional competence ...

By Philip Zelikow

One other consequence of the declassification of the Justice Department opinions and the revelation of the CIA interrogation program is a debate on the effectiveness of these methods. Some of my friends are troubled, because they have been told so often that these methods work.

I will have more to say on the topic of effectiveness later. But for now, I urge those following this controversy to read the op-ed piece published in the New York Times today by former FBI special agent Ali Soufan. I met and interviewed Soufan in the course of my work at the 9/11 Commission, while he was still doing important work at the FBI. From my commission work, my fellow staffers and I had direct knowledge about several of the specific assertions Soufan makes in this piece: about Abu Zubaydah, Ramzi Binalshibh, and Khalid Sheikh Mohammed. My fellow staffers and I considered Soufan to be credible. Indeed, Soufan is fluent in Arabic, and he seemed to us to be one of the more impressive intelligence agents -- from any agency -- that we encountered in our work.

Soufan states, accurately, that some of the CIA claims reprinted in the Justice opinions (and recently repeated in the media) are demonstrably false or overstated. Some of the very claims that Soufan describes were also used, while I was in government, in CIA memos defending the program that were submitted to the White House. Therefore, the declassification of those memos, as Vice President Cheney and others have called for, would only raise questions that would have to be answered with still more disclosures. The Senate Select Committee on Intelligence appears to be trying to sort this out.

To be fair to the program's proponents, many successful intelligence cases are accompanied by rival claims about credit, arguments about which agent or which source provided the key information. That is natural and healthy. And sometimes both sides can be right: a successful stream can have many tributaries. But Soufan's piece is a reminder to be wary of such claims, and especially wary about claims on the key point: whether the extreme techniques were necessary to get satisfactory results.

Among the many hidden costs to intelligence collection are ones like this point, which Soufan makes: "An F.B.I. colleague of mine who knew more about Khalid Shaikh Mohammed than anyone in the government was not allowed to speak to him." At the 9/11 Commission, we noticed this too. At the time (late 2003), my commission colleagues and I could not understand why this agent was not helping to question KSM. This was part of broader concerns we had about the substantive quality of the interrogations. I raised the issue in writing with the CIA's General Counsel in our efforts to dig deeper into the way the interrogations were being conducted --  a point that I mentioned in this previously disclosed memo.

FBI Director Robert Mueller has said publicly that he does not believe that coercive techniques made the difference in averting any planned terrorist attacks against the United States. Mueller has sweated the details of innumerable leads and cases over the last eight years. So, though he may or may not be right, his opinion is entitled to some attention, and it is not self-serving.

By Philip Zelikow

I first gained access to the OLC memos and learned details about CIA's program for high-value detainees shortly after the set of opinions were issued in May 2005. I did so as Secretary of State Condoleezza Rice's policy representative to the NSC Deputies Committee on these and other intelligence/terrorism issues. In the State Department, Secretary Rice and her Legal Adviser, John Bellinger, were then the only other individuals briefed on these details. In compliance with the security agreements I have signed, I have never discussed or disclosed any substantive details about the program until the classified information has been released.

Having been the executive director of the 9/11 Commission, I'm aware of what some of these captives did. The Commission wondered how captives were questioned (for details on that, see this previously disclosed report), and the matter is now the subject of a federal criminal investigation by special prosecutor John Durham. Nonetheless, the evidence against most -- if not all -- of the high-value detainees remains damning. But the issue is not about who or what they are. It is about who or what we are.

Based on what had earlier been released, I have offered some general views on "Legal Policy for a Twilight War." With the release of these OLC memos, I can add three more sets of comments, each of which could be developed at much greater length.

1. The focus on water-boarding misses the main point of the program. 

Which is that it was a program. Unlike the image of using intense physical coercion as a quick, desperate expedient, the program developed "interrogation plans" to disorient, abuse, dehumanize, and torment individuals over time. 

The plan employed the combined, cumulative use of many techniques of medically-monitored physical coercion. Before getting to water-boarding, the captive had already been stripped naked, shackled to ceiling chains keeping him standing so he cannot fall asleep for extended periods, hosed periodically with cold water, slapped around, jammed into boxes, etc. etc. Sleep deprivation is most important. 

2. Measuring the value of such methods should be done professionally and morally before turning to lawyers.

A professional analysis would not simply ask: Did they tell us important information? Congress is apparently now preparing to parse the various claims on this score -- and that would be quite valuable. 

But the argument that they gave us vital information, which readers can see deployed in the memos just as they were deployed to reassure an uneasy president, is based on a fallacy. The real question is: What is the unique value of these methods? 

For this analysis, the administration had the benefit of past U.S. government treatment of high-value detainees in its own history (especially World War II and Vietnam) and substantial, painful lessons from sympathetic foreign governments. By 2005, the Bush administration also had the benefit of what amounted to a double-blind study it had inadvertently conducted, comparing methods that had evolved in Iraq (different Geneva-based rules, different kinds of teams) and the methods the CIA had developed, with both sets being used to against hardened killers.

Opponents should not overstate their side either. Had a serious analysis been conducted beforehand (it apparently was not), my rough guess is that it might have found that physical coercion can break people faster, with some tradeoff in degraded and less reliable results.

Which underscores the importance of moral analysis. There is an elementary distinction, too often lost, between the moral (and policy) question -- "What should we do?" -- and the legal question: "What can we do?" We live in a policy world too inclined to turn lawyers into surrogate priests granting a form of absolution. "The lawyers say it's OK."  Well, not really. They say it might be legal. They don't know about OK.

3. The legal opinions have grave weaknesses. 

Weakest of all is the May 30 opinion, just because it had to get over the lowest standard -- "cruel, inhuman, or degrading" in Article 16 of the Convention Against Torture. That standard was also being codified in the bill Senator John McCain was fighting to pass. It is also found in Common Article 3 of the Geneva Conventions, a standard that the Supreme Court ruled in 2006 does apply to these prisoners. Violation of Common Article 3 is a war crime under federal law (18 U.S.C. section 2441), a felony punishable by up to life imprisonment. (The OLC opinions do not discuss this law because in 2005 the administration also denied the applicability of Common Article 3.)

The OLC holds, rightly, that the United States complies with the international standard if it complies with the comparable body of constitutional prohibitions in U.S. law (the 5th, 8th, and 14th Amendments). Many years earlier, I had worked in that area of the law. I believed that the OLC opinions (especially the May 30 one) presented the U.S. government with a distorted rendering of relevant U.S. law. 

At the time, in 2005, I circulated an opposing view of the legal reasoning. My bureaucratic position, as counselor to the secretary of state, didn't entitle me to offer a legal opinion. But I felt obliged to put an alternative view in front of my colleagues at other agencies, warning them that other lawyers (and judges) might find the OLC views unsustainable. My colleagues were entitled to ignore my views. They did more than that:  The White House attempted to collect and destroy all copies of my memo. I expect that one or two are still at least in the State Department's archives. 

Stated in a shorthand way, mainly for the benefit of other specialists who work these issues, my main concerns were:

• the case law on the "shocks the conscience" standard for interrogations would proscribe the CIA's methods;

• the OLC memo basically ignored standard 8th Amendment "conditions of confinement" analysis (long incorporated into the 5th amendment as a matter of substantive due process and thus applicable to detentions like these). That case law would regard the conditions of confinement in the CIA facilities as unlawful.

• the use of a balancing test to measure constitutional validity (national security gain vs. harm to individuals) is lawful for some techniques, but other kinds of cruel treatment should be barred categorically under U.S. law -- whatever the alleged gain.

The underlying absurdity of the administration's position can be summarized this way. Once you get to a substantive compliance analysis for "cruel, inhuman, and degrading" you get the position that the substantive standard is the same as it is in analogous U.S. constitutional law. So the OLC must argue, in effect, that the methods and the conditions of confinement in the CIA program could constitutionally be inflicted on American citizens in a county jail. 

In other words, Americans in any town of this country could constitutionally be hung from the ceiling naked, sleep deprived, water-boarded, and all the rest -- if the alleged national security justification was compelling. I did not believe our federal courts could reasonably be expected to agree with such a reading of the Constitution.

By Phil Zelikow

I will have more to say soon about interrogation policies in the Bush administration and the recently renewed debate over them, but for now I'll just say this: I am not eager to see any government officials prosecuted for crimes because of their zeal to protect their country. But crimes committed for worthy motives are still crimes, and we have institutions to sort this out.

So has anyone beside me found it troubling that President Obama is making announcements on who should be prosecuted for possible crimes? Whatever one's view of the matter, didn't the administration ardently announce its dedication to depoliticizing the Department of Justice? So why is it proper for the president to tell Attorney General Eric Holder what he should conclude?

There seem to be four possibilities here:

1. No unlawful conduct occurred. That judgment should, at least initially, be made by the Attorney General, free from political influence.

2. Unlawful conduct occurred, but the suspects have a credible defense -- that before undertaking their unlawful conduct, they relied in good faith on authoritative (though in retrospect, mistaken) legal opinions that the planned conduct would be lawful, and these opinions were also issued in good faith. Again, that judgment should be made, at least initially, by the attorney general, free from political influence.

3. Unlawful conduct occurred, and the legal opinions are not an adequate defense. Federal prosecutors, regular or specially appointed, then go to work. Again, the prosecutorial judgments should be free from inappropriate political influence.

4. Unlawful conduct occurred, and the legal opinions might not be an adequate defense. But President Obama decides to issue a blanket pardon for any possible criminal activity.

Or you have option #5, in which the president does not exercise his pardon power but instead, in effect, tells his attorney general what conclusions he should reach about whether federal officials broke the law.

Can you imagine what folks would say if a Republican president exercised option #5? I wish President Obama would just play this straight. He also does no favor to suspects if he politicizes the question of their innocence.