Posted by: Christian | November 5, 2009

A Hybrid Rumsfeld/Soviet Strategy for Afghanistan

Two failed strategies from the past, combined into one new strategy?

Is this negative × negative = positive? Or negative + negative = negative? And have past failures (i.e., in the title of this blog entry) been ignored? First a brief survey of what may be coming.

If you already know this stuff pretty well and would like to skip the re-read, scroll down to the picture of the underpants gnome diagram below (blatantly stolen from Registan).

I’ll start with the NYT, from October 28th:

President Obama’s advisers are focusing on a strategy for Afghanistan aimed at protecting about 10 top population centers, administration officials said Tuesday, describing an approach that would stop short of an all-out assault on the Taliban while still seeking to nurture long-term stability. […]

At the moment, the administration is looking at protecting Kabul, Kandahar, Mazar-i-Sharif, Kunduz, Herat, Jalalabad and a few other village clusters, officials said. The first of any new troops sent to Afghanistan would be assigned to Kandahar […]

But military planners are also pressing for enough troops to safeguard major agricultural areas, like the hotly contested Helmand River valley, as well as regional highways essential to the economy… […]

“We are not talking about surrendering the rest of the country to the Taliban,” a senior administration official said. Military officers said that they would maintain pressure on insurgents in remote regions by using surveillance drones and reports from people in the field to find pockets of Taliban fighters and to guide attacks, in particular by Special Operations forces.

Next up is an opinion piece from the Weekly Standard:

In its continuing search for an alternative to General Stanley McChrystal’s comprehensive counterinsurgency approach to the war in Afghanistan, and with President Obama having eliminated the minimalist counterterrorism plan of Vice President Joe Biden, the White House has lately been floating a split-the-difference trial balloon: “McChrystal Lite” or, to give the veep his due, “McChrystal for the cities, Biden for the countryside.”

The authors (Tom Donnelly and Tim Sullivan from AEI) then get into technical details and posit probabilities. And they seem to want a full surge, as one would expect from people who work at the American Enterprise Institute.

Back to  the newspapers, from WaPo on October 31st:

The military chiefs have been largely supportive of a resource request by Gen. Stanley A. McChrystal, the top U.S. and NATO commander in Afghanistan, that would by one Pentagon estimate require the deployment of 44,000 additional troops. But opinion among members of Obama’s national security team is divided, and he now appears to be seeking a compromise solution that would satisfy both his military and civilian advisers. [..]

Before he can determine troop levels, his advisers have said, he must decide whether to embrace a strategy focused heavily on counterinsurgency, which would require additional forces to protect population centers, or one that makes counterterrorism the main focus of U.S. efforts in the country, which would rely on relatively fewer American troops. One option under review involves a blend of the two approaches, featuring an emphasis on counterterrorism in the north and some parts of western Afghanistan as well as an expanded counterinsurgency effort in the south and east, one of the officials said.

Staying with WaPo, two days later David Ignatius wrote:

As Obama has deliberated Afghanistan strategy, the debate has tended to polarize between “CI” and “CT” advocates. But this is a false argument. What the United States actually has in Afghanistan is a mixture. Obama must now decide whether to provide the resources — and take the risks — to test whether this combined strategy can succeed.

Speaking with Farid Zakaria and Newsweek, Tom Ricks commented:

In fact, the crucial judgments that have to be made involve what the troops will do and how much of Afghanistan to cover. Ricks said to me, “Why not do the Petraeus plan [counterinsurgency] for the major population centers and the Biden plan [counterterrorism] for the rest of the country?” That sounds like a middle course that is smart and practical, which might need some more forces or perhaps can make do with the almost 100,000 already there. Obama should carefully consider these and other options before racing out to demonstrate how tough he is.

And once again, back to WaPo with some behind the scenes info:

The Pentagon’s top military officer oversaw a secret war game this month to evaluate the two primary military options that have been put forward by the Pentagon and are being weighed by the Obama administration as part of a broad-based review of the faltering Afghanistan war, senior military officials said.

The exercise, led by Adm. Mike Mullen, chairman of the Joint Chiefs of Staff, examined the likely outcome of inserting 44,000 more troops into the country to conduct a full-scale counterinsurgency effort aimed at building a stable Afghan government that can control most of the country. It also examined adding 10,000 to 15,000 more soldiers and Marines as part of an approach that the military has dubbed “counterterrorism plus.”

And most importantly, WaPo gives a possible time line for a decision:

Administration officials say Obama might settle on a plan but delay announcing it until after a runoff in the Afghan national elections, scheduled for Nov. 7. The president is to begin a 10-day trip to Asia on Nov. 11.

And recently, from the CS Monitor (h/t Ink Spots):

Put simply, there are two poles in Washington: the counterinsurgency experts, or COIN-istas [sic], who believe Afghanistan’s deteriorating security can only be reversed by adding tens of thousands of troops – perhaps as many as 80,000; and those who believe US interests in Afghanistan are few, and the best way to keep it on a low simmer is to employ a counterterrorism-like model – using drones, bombs, and special forces teams to keep Al Qaeda at bay. The debate has become protracted, with military commanders like Gen. Stanley McChrystal politely urging the commander in chief to make a decision soon.


So yeah. What about that intermediary phase/step? What happens there?

What sayeth the blogosphere? Gulliver at the group blog Ink Spots ( a must read from the moment it started, and honi soit qui mal y pense to anyone that thinks otherwise) – a blog that has lately been taking a hiatus from copious use of French in America – comments in regards to the NYT article above:

Might this be the best possible middle road? It sure seems to beat the hell out of “muddling through,” the other non-drawdown, non-escalation path currently being advocated by some people, in which we’d, uh, just keep doing what we’re doing (because that seems effective). The best line in the whole story cites one official calling the plan “McChrystal for the city, Biden for the country.”

Of course, if this approach is adopted, there will be critics who object to the “abandonment” of parts of the country, and others who question the effectiveness of a sequential, oil spots approach. Bernard Finel has raised some reasonable questions on this point here, and I hope to discuss them a bit more in the near future.

So what does Bernard Finel say? A lot. The friendly back and forth that start at taches d’huile stirs no mauvais quart d’heure. Finel starts with:

But while you’re consolidating government control inside the ink spots, they are consolidating insurgent control in their “ink spots.” Why assume our ink spots will spread, but theirs will shrink?

And once you have moved on, why can’t the insurgents resume operations in their ink spots? And how do you deal with the massive legitimacy loss from the inital abandonment?

…and ends with:

If that is the case, isn’t it obvious that the choice isn’t between well-crafted COIN and imperfect (due to intel and logistical limitations) CT, but between a genuinely half-assed COIN campaign and an imperfect CT approach? Which is why I worry that we are stacking ill-considered assumptions about “rolling COIN” and ink spots on top of the already heroic assumptions embedded in 3-24.

I am really try to understand all of this… but I just don’t get it. What am I missing?

Finel says a lot that leads up to that, mostly to do with the problems of implementation, sustained effort and resourcing. The article is essential reading and I encourage you to give it a once-over. Gulliver, of course, responds. The exchange is the best public debate I can find, and is thankfully free of the overly-politicized language found elsewhere.

That’s the present and the future. But what about the past? The strategies discussed sound familiar. The focus on counter-terrorism in rural areas sounds like Donald Rumsfeld’s “light footprint strategy” and the focus on cities sounds like the later Soviet strategy. First, a graphic illustration of the later Soviet “strategy.”

Afghanistan Communist map

The shaded areas (the cities and the roads) are the level of control as the Soviets were heading out the door. Click here for a larger, better detailed, look at the map. The map is from page 186 of Gilles Dorronsoro’s Revolution Unending (it’s in my top 5 books). I use strategy in quotation marks above as this is strategy as much as any desperate last resort is a strategy. I’ll get back to Dorronsoro. But first, from a 1987 article in the LA Times:

Kabul remains a beautiful and wild city. […] But it is also a deeply troubled place. Since the war began 7 1/2 years ago, its population has jumped from 500,000 to more than 2 million. Since it is one of the few relatively safe places left in the country, people have flocked here for protection. So much of the countryside has been devastated by the war that Afghanistan has folded its population into this city-state– Kabulistan , the natives call it. With more than 4 million Afghans living as refugees in Pakistan and Iran, and up to 1 million killed in the war, Kabul’s current population may amount to one-third of Afghanistan’s remaining people.

Soviet armor and artillery, the so-called “rings of steel” defense, are positioned in concentric circles around the city and its immediate suburbs. Trips to other cities by road take days and are very dangerous. So Kabul has become a fortress town, connected to the rest of the world by infrequent commercial airline flights and to the Soviet Union by the military air umbilical cord. Diplomats have counted more than 50 Soviet military cargo flights in one day. Kabul has become like blockaded Berlin during the 1948 airlift.

Clear BS on the population guess. But what about those “concentric circles”? A US government publication writes:

Western diplomatic sources reported in early 1985 that guerrilla rocket attacks into Soviet military camps in and near Kabul reportedly had killed 60 and wounded 34. The Soviet response to these attacks included setting up a security network of eight concentric circles around Kabul: at each post an estimated 100 soldiers, plus eight to 10 tanks or APCs, were posted.

And in War in Afghanistan, Mark Urban notes that once the rocketing started increasing in range (up to 26km by 1988), defending Kabul with these rings became unfeasible with the number of troops available. However, he does note that Kabul experienced rocketing since 1984. [pp.64-5, 248-50]

Now back to Dorronsoro. On page 189 he notes that

Between 1980 and 1983 the Soviets adopted a defensive position, avoiding direct involvement and relying largely on the Afghan forces. [in 1982]… the Soviets attempted to improve the defenses of major towns.

That first sentence sounds like Rumsfeld’s Light Footprint strategy (but minus his reliance on shady local militia types in favor of the shady old ANA), doesn’t it?

Dorronsoro continues and describes the second phase (1984 and 1985) that resulted from the failure of the first phase (1980 to 1983). In the second phase the Soviets undertook large offensives throughout the country while again attempting to improve the defenses of urban areas. Basically, a dual surge that was stymied by the inability to hold territory due to lack of troop numbers. By 1986 the Soviets “were getting ready to pull out and were launching fewer offensives.” (p. 190).

It was this later phase that resulted in the map you see above.

Mark Urban describes a similar time frame. He writes (pp. 287-8):

From early 1986 the Soviet Army switched to a more defensive strategy. Rural operations were reduced and defences around towns increased while Mikhail Gorbachev prepared the army for withdrawal. During 1984-5 there were probably six offensives involving more that 5000 Soviet troops each. In 1986-9, if one excludes the actual withdrawals of Soviet garrisons, there was just one – Operation Highway.

Going off of formerly  secret archives, during the Soviet Afghan War there was a strategy that also sounds very familiar. From page 123 of Mitrokhin:

The Soviet nomenklatura was forced to change its tactics in the war and to abandon the idea of conquering the whole country at once. A plan was devised to keep firm control over the regions which could be effectively controlled and to introduce national, social and economic changes in them. They reckoned that this would allow them gradually to gain control of the whole country. The aim was to achieve a decisive victory in the northern zones bordering the Soviet Union first. Territorial committees for the defense of the revolution were set up in villages, streets and groups of houses. Surveillance of the population and its mood was stepped up and any movements and new people were noted.

Some translated Soviet documents very recently released by the National Security Archive have resulted in some media attention by various authors who noted the parallels (i.e. in The Guardian and the NYT). Nothing very original in those two articles. As far as lessons learned from the Soviet experience… well, Artemy Kalinovsky says it best while writing in Foreign Policy:

It’s a failure the United States apparently has no intention of repeating — to the extent that it doesn’t even seem to study it. The U.S. Army/Marine Corps Counterinsurgency Field Manual does not mention the Soviet experience once. One analyst told me that when she suggested including the conflict as a way to inform current policy, Pentagon officials seemed to have little awareness about what Moscow had been trying to do there or for how long.

So chto delat‘? What is to be done? Dorronsoro, who is very aware of history, suggests in his WaPo article:

To succeed, the coalition must control Afghanistan’s cities, where institution-building can take place and where the population is more neutral or even favorable to the coalition. The Afghan army and, in certain cases, small militias must protect cities, towns and the roads linking them. Fewer casualties and the improvement of the Afghan security forces — Afghanization — will allow the coalition to focus more resources in the north, where the situation is becoming extremely unstable. Stabilizing the country will allow the coalition to focus on al-Qaeda, the enemy that attacked the United States on Sept. 11.

He provides a similar, but more in-depth argument, in this paper.

But how about I posit the same failure as the Soviets? A weak partner in the form of the Afghan government, a dysfunctional local army and police, a loss of goodwill, untrustworthy potential militia clients, and insurgents that are increasingly able to strike the cities and roads. Will those Soviet concentric circle around Kabul be recreated? Can they?

Getting back to the bigger picture, who is influencing the current debate (in a diffuse intellectual manner)? As far as pushing for the strategies in the title of this blog entry, it sounds like Rory Stewart when I hear about how the cities and the north need to be prioritized. He’s been saying that for a while. But he also complains that no one in government listens to him. He may be wrong about being ignored. His ideas are being considered, even if he is not the source. And since Gilles Dorronsoro came to Washington the views that he has expressed above have become much more prevalent. But more likely, one should look to the September Senate testimony of John Nagl to get a small glimpse of what is being told to the big people upstairs.

It seems some of the current proposals being floated are essentially a return to the Rumsfeld days of counter-terrorism mated with the late Soviet strategy that allowed the government in Kabul to survive in a form that never did extend its authority from its more secure areas. So the Soviet strategy actually stabilized/stalemated the situation for a few years. I accept that. Of course, when it attempted to extend central authority with its unsuccessful try to take over of Rashid Dostum’s militia the civil war got fully started. Is this a move towards the least worst option? Or just a bad option?

Are “we” better at counter-terrorism in Afghanistan? Probably, since it was so terrible back in the Rumsfeld years. But it is still bad. And can “we” do a counter-insurgency strategy in Afghanistan that cedes an enemy’s favorite territory? Does that strategy essentially open up a “highway”  straight from Pakistan to the “Urban shooting/bombing galleries”? Will we see an endless stream of attacks against coalition forces and the Afghan government as they hunker down in their urban fortresses and attempt to run the gauntlet of the main roads? So yes, more debate is needed. I suppose that is happening behind closed doors as I write. We should find out in a couple of weeks what the president’s decision is.

It took several years to determine that Rumsfeld’s Light Footprint strategy was a failure. It is in a similar time frame that the Soviet’s “Cities, Roads and the North” focus was a failure (after it was inherited). There won’t likely be such a long time frame in the present case.

Is this obsessive analogizing with the past? Maybe. Afghanistan is not condemned to follow the fate of its 1979-2001 trajectory. Foreign forces are not the Anglo-Afghan expeditionary force. As the title of Kalinovsky’s article reads, “Afghanistan is the new Afghanistan.”

So what do I think? I’m on board with Bernard Finel‘s comments. I think his points are good enough to warrant a repeat:

But while you’re consolidating government control inside the ink spots, they are consolidating insurgent control in their “ink spots.” Why assume our ink spots will spread, but theirs will shrink?

And once you have moved on, why can’t the insurgents resume operations in their ink spots? And how do you deal with the massive legitimacy loss from the inital abandonment?


If that is the case, isn’t it obvious that the choice isn’t between well-crafted COIN and imperfect (due to intel and logistical limitations) CT, but between a genuinely half-assed COIN campaign and an imperfect CT approach? Which is why I worry that we are stacking ill-considered assumptions about “rolling COIN” and ink spots on top of the already heroic assumptions embedded in 3-24.

I am really try to understand all of this… but I just don’t get it. What am I missing?

I concur. More debate is needed. More information is needed.


  1. Of course the huge difference between the Soviet experience and today, which you don’t mention, is that the Soviets were fighting pretty much everyone in Afghanistan while the insurgency today is almost completely Pashtun. The “McChrystal for the cities, Biden for the countryside” would therefore be quite a bit different in implementation than anything the Soviets did. With the exception of some Pashtun enclaves in the north there isn’t a huge need for a huge security presence in the non-Pashtun areas. So I think we are talking more along the lines of one large oil spot in the north, some cities in the south, east and west, and the ring road. What’s left is all the rural Pashtun areas which would be subject to “CT”. I’m wondering what “CT” would actually mean since there aren’t very many bona-fide foreign terrorist types running around Afghanistan. Maybe “CT” means C(k)ill Taliban.

    It’s a strategy, but what is the point? Is it simply a political compromise? What is the goal? Is it to de-facto split Aghanistan along sectarian lines; is it a real effort to eventually unite the country; or is it simply a planned kind of muddle-through until something else becomes politically possible? I don’t know, but I suspect it’s the latter.

  2. Andy,

    I don’t see the Soviet-Afghan War as the Soviets versus “pretty much everyone.” That view is no longer in favour with the historians.

    And those pockets in the north are not isolated pockets. The Pashtun communities there were strategically placed in prime economic, commercial, agricultural and strategic areas. This occurred from the beginning of Abdurrahman’s reign until at least as late as the 1930s. And into the 1970s if smaller land expropriations are counted. Pashtuns come and go as they please in every city in the north. Even if one wants to look at it as a “almost completely Pashtun” insurgency (which is changing), one must take into account that Pashtuns are in all the strategic spots in the north. [And I really hope that any strategy doesn’t lead to a broad, clumsy harassment of Pashtuns in the north]

    As far as what those commentators above are calling CT, some are qualifying it, others not. I think some of them just mean SF, drones and airstrikes.

    And I also lean towards the “muddle through for now” comment

  3. Christian,

    My point on the Soviets vs “everyone else” was a simplification but I don’t think it’s that inaccurate when compared to today. We’re not fighting the Hazara, Tajiks and Uzbeks in any meaningful sense. There’s a reason all the NATO forces who don’t want to do any fighting are assigned to these areas. We don’t need forces spread out in rural villages to defend the population because there isn’t an insurgent threat (though there certainly is a criminal threat).

    I agree that the “pockets” of Pashtuns in the north are in very important areas, but that doesn’t change the reality that they are still pockets with all the limitations that brings. They are easier to control than cities in the south.

  4. The Thunder Run has linked to this post in the blog post From the Front: 11/05/2009 News and Personal dispatches from the front and the home front.

  5. Thanks for replying Andy,

    Your point on the Tajiks is taken. The handful of Tajiks fighting the government today is insignificant (for now). Although in the 80s the Hazaras effectively neutralized themselves as an insurgent force by engaging in an inter-Hazara civil war (which the Soviets gladly watched from the sidelines). And by the end Rashid Dostum’s militia greatly reduced the significance of ethnic Uzbeks in the insurgency.

    But yes, I do accept that at present the coalition forces do not face as wide an array of insurgents. I’m just always worried that people who may skim the comments might get the impression that the Afghan communist government was 600 guys and the soviet army against everyone else. So I think we’ve effectively qualified the situation.

    But again on the northern Pashtuns. Olivier Roy remarked that all the ethnic maps of Afghanistan are inaccurate. They are all actually based on a Soviet map from the 1950s. There are no neat lines around ethnic enclaves that can be “contained,” certainly not with the lack of knowledge we have on the subject. For example, Kunduz is a particularly difficult spot.

    But of course I do accept that most of the north can still be handled by some of the europeans.

  6. Something I can’t understand about the idea of focusing efforts on protecting the cities is how it could possibly harm a rural insurgency like the Taliban’s. This is not Baghdad and the insurgents are not stupid enough to try to fight inside the cities. My best guess is that the real focus is on preventing the bad publicity of more terrorist attacks in Kabul destroying what little U.S support remains. Also, to me the idea of combining two military strategies is very foolish. This isn’t a bipartisan bill where both sides agree to walk away with something, this is a war where compromise on strategy will probably weaken both efforts.

  7. […] Ghosts of Alexander – A Hybrid Rumsfeld/Soviet Strategy for Afghanistan […]

  8. I think that the problem of there being no firm definition of victory is a key issue. Granted, I don’t hear much about long-range artillery rockets by Taliban forces in the current conflict – I imagine American money made much of that possible against the Soviets – but nonetheless, this is “containment” by trying to keep Taliban out of particular areas… as if the Taliban do not have a vote and cannot try to follow NATO forces to these areas. Add that the Taliban is one spectacular killing away from driving the UN out of the whole country and building with “soft power” would appear in danger of complete checkmate. It’s not a pretty strategic picture.

  9. […] don’t think I agree with the analytical conclusions of this item from the Ghosts of Alexander blog assessing the prospects for “A Hybrid Rumsfeld/Soviet […]

  10. Does the fact that there is no rival superpower pushing the insurgents mean anything? Or was that not a factor toward the end of the situation?

  11. […] is now needed is for Predator drones to drop multiple copies of the Biden Plan onto their heads. The Biden plan: drone strikes and SF…. coming soon? Or not at all? Because seriously, that is what was […]

  12. […] to the fascinating Ghosts of Alexander blog, which the Parallax Brief found via Matthew Yglesias, the major policy debate in Washington is now […]

  13. Christian — I’m eventually going to get back to this subject, so I don’t want you to think that I’ve ignored your outstanding and comprehensive post.

    In other (sort of awesome news): last night in a talk at SAIS in DC, Dave Kilcullen cited the Underpants Gnomes method for operational success in Afghanistan. I’m sorry to say that neither you nor Josh was credited.

  14. […] bien las cuentas; 100.000 americanos. En contra de lo que dice el tópico, por cierto, a los rusos no les fue tan mal; a finales de 1985 controlaban una cantidad decente del país. Su retirada fue más política que […]


%d bloggers like this: