In early May 2015, U.S. Secretary of State John Kerry made a historic but little noticed visit to Somalia, a country no other U.S. secretary of state had ever visited. His trip symbolized both how far Somalia has come—from the blackest days of civil war, clan infighting, and famine in the 1990s; to the brutal rule of the jihadi group al Shabab in the late 2000s; to something getting closer to normal now—and how very far it still has to go.

The fact that a high U.S. official could enter the country at all speaks of real security improvements. During his visit, moreover, Kerry announced the reopening of a U.S. embassy in Somalia, which had been closed since 1991 when the government of long-term dictator Siad Barre collapsed. But the fact that Kerry’s visit was a brief few hours—during which he did not even leave the heavily-guarded Mogadishu airport—also points to deep and persistent security challenges. Moreover, his meeting with Somali President Hassan Sheikh Mohamoud and Prime Minister Omar Sharmarke comes at a time when the relationship between international donors and the Somali government has soured and the Somali people have grown increasingly weary of their government. The early optimism that the 2012 election of Mohamoud by appointed members of the Somali parliament would usher in badly needed changes in Somali politics, toward inclusiveness, effectiveness, and accountability, dissipated long ago.

Indeed, an observer’s bullishness about Somalia very much depends on his or her baseline. Compared to the early 1990s or 2011, when al Shabab controlled most of Mogadishu and most of central and southern Somalia, with only the semi-autonomous regions of Puntland and Somaliland escaping its grasp, Somalia is in much better shape. However, when compared to the spring of 2013, when I took a previous research trip there, the 2015 spring (my latest trip), and summer hardly look peppy. Security is tenuous, with al Shabab and the African Union Mission in Somalia (AMISOM) forces stuck in a draw, and politics has been regressing to many of the same old discouraging patterns.

The rest of 2015 and 2016 are important times for Somalia. They could either resurrect optimism about the country’s progress or reinforce disappointment. The current AMISOM mandate expires in November 2015. By 2016, as a compact between the international donors and Somalia government specifies, presidential elections are supposed to take place, a constitution redrafting is to be finished, and the transformation of a centralized state into a federal one with states formed is to be completed. From the perspective of the middle of 2015, this agenda looks daunting.

AL SHABAB’S BATTLEFIELD

After struggling against al Shabab for several years and hunkering down in a few blocks of Mogadishu, AMISOM forces, with the assistance of international private security companies and international funding, finally began to reverse the al Shabab tide in 2011. As clan militias defected from al Shabab, AMISOM succeeded in pushing the terrorist group out of Somalia’s major cities. U.S. air and Special Forces attacks against al Shabab leadership eliminated some key figures, such as the group’s amir, Ahmed Godane, in September 2014 and its previous leader, Aden Ayro, in May 2008.

That said, al Shabab is hardly defeated—even if its membership is thought to be down to around 6,000, with the most potent and hardcore Amniyat branch down to perhaps 1,500. (Such estimates, given by Somali government officials and international military advisors, need to be taken with a grain of salt, since the capacity of insurgent groups to replenish their ranks often outpaces the capacity of counterinsurgent forces to kill or arrest the groups’ members.) The group’s spectacular terrorist attacks in Kenya and Uganda, such as the one on Nairobi’s Westgate Mall in September 2013 and on a teaching college in the city of Garissa in April 2015, don’t necessarily mean that al Shabab has lost the capacity to operate in Somalia. In fact, if anything, al Shabab’s operations have become more targeted and more effective, and generate more casualties with the militant group losing fewer fighters. The fact that the group has deeply infiltrated Somali military and police forces helps it in that regard.

Although AMISOM still holds the major cities that it won back from al Shabab as part of the 2014 Operation Eagle and Operation Indian Ocean, al Shabab’s presence in supposedly liberated cities is often robust. The group extorts shopkeepers and intimidates the local population with threatening night letters that regularly appear in public spaces. People routinely receive cell phone texts such as “You forgot to pay your zakat (religious tax); tomorrow we cannot guarantee your security.” Such intimidation is prevalent even in Kismayo, a strategic port in the southern region of Juba that used to be a key source of revenue for al Shabab from customs and smuggling items like charcoal. Kismayo, and the newly-formed state of Jubaland, are controlled by Ahmed Madobe, who defected from his role as al Shabab commander several years ago and, with the support of Kenyan forces, took control of the area and declared himself president of the state.

Over the past year, al Shabab attacks have also escalated in Mogadishu. Assassinations are a daily occurrence. Many government officials have to live and work (often in the same room) in hotels close to the Mogadishu airport, a palpable symptom of the decline in confidence and sense of security since 2013. The fact that some assassinations are actually perpetrated by rival politicians, warlords, and businessmen, with al Shabab happily taking the credit, does not lessen the sense of insecurity.

Al Shabab also controls roads and limits AMISOM’s movement. Attacks on AMISOM convoys and IEDs are frequent. In fact, despite its two much-touted offensive operations last year, AMISOM is mostly in defensive garrison mode. Rarely does it actually fight al Shabab; in advance of AMISOM’s clearing operations, al Shabab often disperses. Usually, by the time AMISOM arrives, it finds a ghost village (sometimes destroyed by al Shabab). AMISOM leaves, and al Shabab comes back from the bush. Often locals, at best, sit on the fence and, at worst, continue to support al Shabab because of their calculation that al Shabab will ultimately be the dominant force in their area.

That does not mean that Somalis actually like al Shabab: Its brutality is still shocking; memories of the militant group’s aggravation of the 2011 famine are still vivid; and al Shabab has hardly been a competent ruler enabling local economic growth. Instead, the group often tried to suppress or undermine vital economic markets, such as in qat. And, thanks to its control of the roads, ordinary Somalis fear traveling on them. Those who are willing have to be prepared to pay bribes of about $30 dollars to travel to Mogadishu from Merka and over a $100 to travel from there to Kismayo. Only the wealthy can absorb such costs, increasing Somalis’ frustration and sense of insecurity. Likewise, urban Somalis are quick to point out that inflation, including the cost of basic food items, has significantly increased since deliveries must now either come by air, be smuggled in, or are levied with substantial extortion fees and illegal taxes.

STUCK IN THE SAND WITH AMISOM

On the other side of the fighting, AMISOM nominally numbers 22,000 soldiers from Burundi, Djibouti, Ethiopia, Kenya, and Uganda. It could and should be much more efficient in its fight against al Shabab. But it is not clear how many soldiers are actually on the ground at any one point. The capacity and training of the AMISOM deployments varies widely across the countries. Some of the forces, such as those from Burundi, do not speak English and have little training overall. Many of these militaries were built during their country’s own political revolutions and have had little deployment or battle experience since. Very few of the deployed troops have had any counterinsurgency training and they lack logistics, medevac, and intelligence and reconnaissance support. AMISOM was to be equipped with ten helicopters, with Uganda promising to provide four and the other United Nations member states the rest. Three, however, crashed into Mt. Kenya as they were flying from Uganda to Somalia, and Uganda is now in dispute with the international community over who will pay for the destroyed aircraft.

Moreover, the original expectation that a United Nations force would eventually replace AMISOM has long since died. Nor do the AMISOM forces necessarily want to get out of Somalia (or fully defeat al Shabab): The international funding they receive for their effort makes for good living for their soldiers and a substantial financial boost for their military institutions. Moreover, their presence in Somalia allows them to pursue their regional interests and enhance their importance with the broader international community.

AMISOM has weak headquarters to which few member countries pass on any information, let alone intelligence, or bother to coordinate. Some AMISOM commanders maintain highly personalized and sometimes outright subversive agendas: There are credible rumors that AMISOM units have sold fuel and arms to al Shabab or looted humanitarian convoys.

The fact that AMISOM is organized into five sectors operated mostly by one of the AMISOM member countries does not help with coordination and planning. The division of the sectors reflects the strategic interests of the intervening forces. Kenya and Ethiopia, although they have suspended some of their mutual rivalries, still mostly cultivate proxies in their sectors to create buffer areas, prevent the leakage of terrorism into their countries, disrupt support for separatists within their own countries, and project land and sea power. Offensive operations are decided mostly on a sector basis, with the forces in each area reporting and taking orders from their own capitals. Whether captured weapons are handed over to Somali forces varies by sector. So does how al Shabab terrorists are dealt with. There is little coordination among the sectors and little planning at AMISOM headquarters; in fact, they are generally only interested in working together when headquarters has something to offer to them, such as logistical support via the United Nations.

Not surprisingly, it has been hard for AMISOM to hold and build a “cleared” territory. At first, AMISOM forces exhibited little interest in providing any governance functions or even conducting stabilization operations, such as repairing bridges or providing clean water systems. They expected the Somali security forces and government to do so. But Somalia hasn’t been able to because local governance structures are frequently destroyed, blocked off by al Shabab, dominated by problematic powerbrokers, or lack resources. And so AMISOM has come under pressure from the United States and the international community to take over these stabilization functions.

Pushing AMISOM into stabilization operations is a difficult call. On the one hand, it should be the responsibility of the local and national government to administer its territory, and the credit for doing so should accrue to the Somali government, not to foreign forces. On the other hand, local communities are frustrated by the lack of security and services after AMISOM clears a territory. In either case, it isn’t clear that AMISOM militaries could do much better at governance, since they, too, lack resources and training. And the political sensitivities abound. Somalis do not see themselves as African, but rather as Arab; and al Shabab can easily label Burundi, Uganda, Kenya, and Ethiopia as Christian invaders. Although Somalis are deeply divided along scores of clan divisions, they also identify as nationalists, opposing foreign intervention.

If AMISOM does take on a stabilization role, it should be limited, discreet, and concrete, including short-term support for building water and other infrastructure. One of the current ideas is to deliver quick-impact projects only when some, even interim, local authority has been created and is supported by local peace committees consisting of clan elders, imams, women’s groups, businessmen, and civil society members. Even though the projects could still become fronts for graft, any accountability is better than none.

SOMALI NATIONAL FORCES IN TATTERS

Another major official combatant in the war is Somalia’s own forces, consisting of the army, police, and militarized intelligence service. They have not been able to provide stabilization operations on their own because, as still mostly a collection of disparate militias, they lack the capacity. They remain beholden to clans and powerbrokers, and lack both a national ethos and training. When pressure rises, they mostly fall apart or return to militia behavior. Underpaid and often not paid for months, they frequently resort to selling their equipment to obtain some income. They are also notoriously infiltrated by al Shabab. The paramilitary intelligence service run by the National Intelligence and Security Agency, and the preferred partner of U.S. and Ethiopian counterterrorism efforts, is somewhat better, but also rather brutal and beholden to clan politics.

Not surprisingly, the Somali people do not trust their national forces. Although the federal government nominally controls the national forces (while explicitly not controlling regional militia forces), its presence beyond Mogadishu is limited and it depends on AMISOM and international support for protection from al Shabab and rival powerbrokers. In order to wean itself off AMISOM, defeat al Shabab, and suppress regional conflicts, Somalia’s national forces would need to be significantly bigger than they are now at about 10,000 fighters. But donors, aware that a large percentage of foreign military aid disappears into personal pockets of Somali politicians, are reluctant to commit more money for larger Somali security forces.

The security forces of the semi-autonomous state of Puntland are somewhat more capable, but insecurity in Puntland, too, has been increasing since al Shabab was pushed into the state from central Somalia. Numbering perhaps about 4,000, the forces include a state-armed militia/police force known as darawish as well as other police forces and custodial forces. Many other unofficial entities also operate in Puntland, including the Puntland Security Force, which is paid by the United States to fight al Shabab and presumably reports to the Puntland president, and the Puntland Maritime Police Force, which is paid for by the United Arab Emirates. The latter was originally created to fight pirates, although recently it has also apparently been dispatched to fight al Shabab in the Galgadud area. The Puntland government has little interest in integrating these forces into the Somali national armed forces.

Somaliland remains the most secure part of Somalia with the best functioning government­—although, of course, the local leadership there continues to want to secede from the country and establish independence. Mediation talks in Ankara facilitated by Turkey collapsed in the spring of 2015. Since then, Somaliland has been preoccupied by presidential and parliamentary elections for the state government, which were to be held on June 26, 2015. But despite popular demand and strong pressure from international donors, the elections were delayed by at least 17 months due to a lack of preparedness, (as they had previously been in Puntland). This delay undermines governance and accountability in the state.

THE VICIOUS CIRCLE

It is not just security that has been sliding in Somalia for the past year and half. Equally, the sense of political momentum has dissipated. In 2013, there was a great deal of optimism among the Somalis whom I interviewed that Somalia hit rock bottom in 2011 and that the pernicious clan politics that plagued the country for the past three decades have ended. They placed a great deal of hope in their President, Mohamoud. A Somali professor and member of the country’s civil society, he was not a former warlord nor a member of the diaspora parachuted in. And although he was elected by a parliament of appointed (or self-appointed) clan elders and former warlords, he was not seen as beholden to any particular clan. The international community, including the United Kingdom and the United States, also embraced him.

But that was then. With little control over the country’s armed forces and budget, and unable to tackle pervasive and extensive corruption, the president fell back on one source of support: his Hawiye clan. And so the cycle of exclusionary politics began again, privileging access to business deals for his supporters and promoting clan backers for government positions.

Mohamoud’s government was soon paralyzed by the infighting between him and his prime ministers (a familiar story in Somalia over the past decade), whom he would repeatedly seek to replace. The Somali constitution makes the president the symbol of authority, but his role and relationship with the prime minister is not clearly defined. Ultimately, the constitution is generally interpreted as mandating a Hawiye president and a Darod prime minister. That design is meant to encourage inclusiveness. In truth, however, it mostly led to a struggle between the president and prime minister, mimicking the power fights between the two main clans.

The constant turnover of government officials at the federal and subnational levels is another major problem: With appointments often lasting only a few weeks, officials have far more interest in quickly making money and placing allies in other public sector positions than in governing effectively and building equitable and accountable state institutions—or any institutions for that matter.

To give itself legitimacy, the government has embraced a brand of conservative Islam that is not as far from al Shabab’s teachings as many would like. The president is reputed to have admiration for the Muslim Brotherhood in Egypt and is said to consider Mohamad Morsi, the imprisoned former Muslim Brotherhood–affiliated president of Egypt, a personal friend.

Indeed, the contest for political legitimacy in Somalia revolves around four elements: Who is more Islamic? Who is more nationalistic? Who delivers better security? And who is less corrupt and delivers better services? For years, the Somali federal government has struggled to win on any of these fronts. And it has exhibited little recognition of, or interest in, the problems of clan marginalization and poor governance, even though these grievances thrust Somalis into al Shabab’s hands.

To address some of these problems, under a 2013 compact between the international community and Somalia, Somalia was supposed to hit three milestones by 2016: hold presidential elections; adopt a new constitution; and form subnational states. All are important, and none is easy to do, much less do well, in the given timeframe. Yet international donors, not wanting to repeat their frequent sin of setting up conditions but still delivering aid after a Somali government fails to meet them, are loath to relax the 2016 timeline.

Pervasive insecurity makes holding national elections difficult. It also enables fraud and heightens feelings of purposeful exclusion. AMISOM has helped little when it comes to providing security for a vote. And the government has made few preparations itself. So far, there is not even a voter registry. In late May 2015, the Somali government launched a census effort (a step toward creating a voter registry). However, the census itself could lead to new conflict, particularly if the resulting counts of the Hawiye, Darod, and other clans and subclans make any one group unhappy—as is almost sure to happen. Meanwhile, the fact that the independent electoral commission is located within the presidential palace of Villa Somalia, even if for legitimate security reasons, makes it seem potentially biased and illegitimate.

But there is little alternative to holding a national election. Many Somalis want to see a change in government; and international donors are also increasingly frustrated with the current one. Perhaps the president could again be appointed by members of parliament, with all the legitimacy limits such a process brings. Ultimately, though, a vote and the creation of real political parties is important. It is the only way to realign Somali politics away from narrow clan parochialism and individual patronage networks and toward broader national representation and coalitions. But few Somali powerbrokers have an interest in allowing their formation; even under the best of circumstances, they will not materialize by the 2016 election.

It is also possible that the international community will agree to postpone the elections. It did so in Puntland, it now has to live with it in Somaliland; and it may do so again at the national level. Even if national elections do not take place, it is worth considering whether some subnational elections (such as for the mayor of Mogadishu) could be held to facilitate greater accountability.

The next task is revising the constitution in a way that increases inclusiveness. Donors do not want the redrafting process to drag on for years, as it has in Nepal for over a decade. Somalis are already disappointed with initial drafts, though: Quotas for women have disappeared from the constitution, and progressives have little faith that the current language—women should have a “meaningful representation” in all elected and appointed positions—will achieve progress. Moreover, the constitution drafters are still to tackle some of the most politically contentious issues, including how power (including arms, taxes, and other resources) will be distributed between the center and the newly forming states.

But the fact that Mogadishu has accepted federalism and power decentralization is perhaps the greatest political accomplishment in years. Competition over who controls Mogadishu and crucial resources has, for years, been a major source of conflict and corruption. Few outside of the capital, including Hawiye clans who dominate business there, want to be ruled by it.

However, there is as yet little agreement about the relative balance of power between the center and subnational states, including whether they will be allowed to retain their militia forces as some sort of paramilitary police. In the Jubaland State, Madobe, whose self-declared presidency was accepted by Mogadishu on an interim basis in 2013 for two years, has so far shown no inclination to give up control of any of his forces. In the Southwest State that has also been formed, local state officials decry the absence, incompetence, and untrustworthiness of national forces and clamor for their own armed services.

In both Jubaland and Southwest States, the state formation process was unable to avoid fighting between warlord and clan forces over which areas would be included in which state and under whose control. In Jubaland, the process ended with Madobe’s victory over the forces of Barre Hirale’s (who are still mostly hiding in the bush). In the Southwest State, the two local rivals created a coalition government, with over 60 ministers and plenty of built-in political dysfunction, nepotism, and paralysis. State formation still needs to be completed in other areas, such as the Shabelle. In April 2015, a state-formation conference was launched for the Central Regions State. Some representatives continue to question whether six states are enough and others are debating which state their territory should belong to.

How to generate revenues is another major challenge in the federalization process. Neither the state governments nor the national one trusts the other to share revenues: The states do not want to give up land taxes to the federal state; but the federal government strongly dislikes the idea of having to rely only on the tax revenues from fisheries and maritime routes. And the promise of potentially huge mineral resources under the Somali sand only makes the federal versus state competition more intense.

How control is devolved matters a lot. The biggest danger is that the exclusionary politics over spoils and war rents that have dominated Mogadishu for so long will be replicated at the local level. And given how the state formation processes have been going, there are reasons to fear that the clientalistic patronage networks that systematically discriminate against rivals will be reestablished at the state level. In some areas, especially in the Juba Valley, that is already underway, creating a significant number of internally displaced people and potentially allowing al Shabab to insert itself into the area on the side of the oppressed.

IT’S GOOD GOVERNANCE, STUPID

Over the past few decades, international actors have not paid enough attention to subnational governance in Somalia, and they are running that risk again. Many, including the United States, focus predominantly on the problem of al Shabab, even though al Shabab is merely the latest result of poor governance. Many of the crucial donors lack presence outside of Mogadishu, which limits their understanding of life at the regional, town, and village levels. Local peace committees of clan elders, imams, and representatives of civil society and the business community can be an important mechanism of better governance. But the international donors need to work with them, and to be aware of the politics behind the peace committees—such as, for example, of who is selected for them and who is excluded. Other international actors, such as Kenya and Ethiopia, often embrace problematic powerbrokers for the sake of their strategic and counterterrorism interests, even though these powerbrokers ultimately undermine stability.

Fundamentally, whether Somalia succeeds in breaking out of decades of conflict, famine, misery, corruption, and misgovernance depends on the Somali people. It depends on whether a sufficient constituency for better governance and less conflict eventually emerges or whether Somali businessmen and politicians continue to find the way to work around conflict or make money from it while the Somali people eke out survival amidst the harshest conditions without mobilizing for change. Since 2012, Somalia has had one of the best chances to pull off such transformation in years. It should not waste it.

This article was originally published by Foreign Affairs.