I believe that if there is a hope for humanity to approximate the direct democracy of ancient Athens, it is through decentralisation and its Protean manifestations.  

Whatever definition of decentralization you take, we should avoid to give an overall a priori opinion about it. In fact, much depends on: 1. the function/sector to which it is applied; 2. the way it is implemented. For instance, in certain sectors like large infrastructures and energy, decentralization does not work well, while in other sectors like education and agriculture it may work fine. Also, there should be a matching between functions/powers assigned and resource available to the decentralized authority, otherwise that is not decentralization but "de-responsabilization". Although decentralization and democracy are strictly related, the former may become a way to weaken and eventually dismantle central States.

It can be very practical for participatory monitoring and evaluation, which may contribute significantly to programme efficiency. Namely, local people can tell a local leader where things need to improve, and a decentralized government can act directly on this instead of just passing the info up the chain with the hope that some action will eventually flow back down.

However, in many countries there is major capacity deficit at local levels, so it can lead to lots of patronage and no results, even with the best of intentions.

Also, it can erode legitimacy of the central state unless it is billed in a very mutually cooperative manner. It takes a confident state to give powers to local authority.

I think capacity building at the local level should precede decentralization in most cases.

Consider the case of India for decentralized activities without good monitoring and evaluation systems. Basically, you get a lot of people running loads of mini programs which are not targeted towards specific outcomes, and which are designed on the basis of "holistic" approaches. I'm all for holistic approaches when it means that soft and touchy feelie stuff comes into the picture, but if you are not managing for results you are not likely to get them. (Sorry to use India as the whipping post, but this is a widely acknowledged problem and serves as a very good basis for a learning example.) Often, instead of building capacity of local leaders, you get some idealistic outsiders who manage to get many grants to make good things happen, but then these really cool projects, say, for sustainability, are implemented without specific objectives which can be evaluated, whether from the perspectives of outputs of specific project activities or from the perspective of final outcomes which could have been targeted (e.g., water table stops falling, 50 more farmers have access to irrigation, as opposed to "let's do this project then go back to the big city, then come back to do it again in five years after everything falls apart").

Consider the case of decentralized management in First Nations communities in Canada. It has often preceded significant capacity building, and occurred in the absence of significant checks and balances or monitoring mechanisms, largely in deference to traditional political structures which did not evolve in an era of management accounting or managing for results. In this case, autonomy may prove to be very expensive in the short run, and we can hope that these communities will evolve effective systems which will benefit all interested parties in the long run.

I would not deign to separate decentralization from accountability. Sometimes it will be better, sometimes it will be worse. In many cases, it almost seems like someone is planning on one or the other.

Does decentralization come with transferred money or local taxation capacity? This will be inextricably linked to almost every problematic involved.

Let me write from the experience of decentralisation in Kerala, India. It was introduced for deepening democracy in the state. The result however was the entry of a new set of local goons into politics. When local institutions decide over issues which are not confined to their geographical limits, it becomes a problem. For instance, a panchayat may wants to build a small dam in a river that would adversely affect people downstream. However the panchayat is not accountable to those people. Decentralisation has also led to the democratisation of corruption; a share of the cut would trickle down to the local elites. So the ordinary person has to give bribes from below. What matters is not centralisation or decentralisation but the principles that govern democratic functioning in society. If the governing principle of a democracy is market competition, that society will have more market than democracy.

@ Costas

Recall that the democracy of Athens excluded both women and slaves. "All citizens" was in fact a somewhat exclusive group. It reminds me more of early versions of democracy in Britain, where land ownership was a prerequisite to be able to access political positions.

But I do like the idea of selecting representatives by lots. The outcome is likely to be more representative of the population.

We get caught in 'binary reductive categories' (Kincheloe and Steinberg, 2008) of the social phenomenon. Decentralized and centralized system of decision making both are crucial. I feel discussion should be around identifying cases and scenario where decentralization is must and then facilitating it through an inductive process.

There are scholars who "fear" elite capture in when decentralization is in discourse (Dasgupta & Beard, 2007; Platteau, 2004; Iverson etal, 2006). But we need to examine who is behind elite capture at global level. The research titled "The Network of Global Corporate Control" by Vitali et al (2011) might throw some light on this (link to paper and TED talk appended. 

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0025995

http://www.ted.com/talks/james_b_glattfelder_who_controls_the_world   

i think instead of talking about the decentralization of a corporate or any organization, we should stick ourselves to the governance system in Developing Countries.

it has been widely felt in developing countries that ruthless implementation have caused great problems.

we need to work over 'a transitory framework' in which first atleast build the capacity at grass root level and public mobilization which would create a congenial and responsible atmosphere for the implementation of Decentralization.   

In my opinion decentralised organisations bear more performance potential than centralized ones, because of the multiplication of agents. If you allow me to be counter-intuitive, any organisation (say, a firm, for instance) is decentralised to some extent. However, the more decentralisation, the stronger the coherence which is required to avoid the explosion of the system.

Thus, as some (growing) effort has to be switched from final goals to coherence assurance, decentralisation both promises to gain effectiveness and loose efficiency.

As a consequence, the appropriate choice in a determined situation is a matter of trade offs

It seems that decentralisation can exist in all political systems. An example is Switzerland with canton system.   There are some Cantons that have a direct democracy. All the peple vote and there are no representatives. Ity is worth studying this kind of system. On the other hand in Greece decentralisation seems to be decentralisation of corruption! A successful decentralisation is a big and delicate problem, but it is woth trying .

Dear Ankur and Claudio,

thank you for your important posts!

Many people respond to the Switzerland case by suggesting that Switzerland is small and therefore direct democracy is not possible in other places.

Sounds like an excuse to me. The fact that they are highly educated and very aware of their political context, however, may be very relevant. I think many developing countries would not have very good macro results under direct democracy. Perhaps in a few generations?

Dear Nathan,

I agree with you. Decentralisation is a  kind of Protean concept. If you consider localities then the problem becomes more difficult. To reach something like Switzerland, we need a lot of stages. But we should sometime start. 

We should also note that decentralisation in Europe was in some sense implied by the strong atheling (feuds) together with the anode of the capitalist class, that requires "democracy".

One can glean lots of insight about the excellent question for this thread by reading Aristotle's Politics.   See, for example, Jowett's 1885 translation of the Politics of Artistotle, 355pp available at

http://tranductuan.files.wordpress.com/2013/04/the-politics-of-aristotle.pdf

This is an amazing work.   One of the thilngs experienced and thought quite a bit about was the community in a village as opposed to a social system in megapolis.  Life in a viiage is structured along the lines of a family of familes, where everyone knows everyone else.   In a viliage, it makes sense to have a patriarch who takes on the responsibility of governing (e.g., settling disputes about water rights).   In that context, centralized government makes sense.  With a megapolis, it makes more sense to consider a decentralised form of governing, where small political units as town councils assume the responsibility of governing the local community,

The tyranny of local elite is no way better than the tyranny of the distant, national elites. Decentralisation may lead to the consolidation of elite power at the local level. A society with ingrained oppressive structures must think twice before implementing policies of decentralisation. If it comes with radical programmes such as land reforms, it can be more effective.

@ James - I don't think you will find people arguing that the quality of democracy is improved, for example, when forcing many localities to agglomerate into a megopolis. However, the cost savings in not having "duplicit" town councils, or in being able to integrate services such as transportation, can make it worth it.

In Ontario, many municipalities around Toronto opposed amalgamation, but the province forced them to amalgamate because they (province) were ultimately finding themselves responsible for many costs, and also it made it a lot easier for the province to negotiate with a single municipality.

I think Aristotle may have been most inclined to borrow his argument on the need for middle class involvement in politics (perhaps they DONT know best, but who cares, they must be involved because they ultimately experience the impacts of decisions being made) to argue in favour of decentralization were he plopped directly into the modern context. I don't recall questions of administrative efficiency featuring heavily in his work. Rather, he was constructing a vision of a just society.

Dear Nathan,

For the most part, the societies we live in are just.   And laws are for the most part are upheld by members of each of the societies.   Aristotle becomes practical when we are willing to decentralize, shifting from the megapolis model to an approximation of the village model in Aristotle's Politics.

While dealing with recent developments in the new EU members I am wondering how far could be decentralization implemented in these countries if there is no "pressure" from the side of the EU and other international stakeholders. If for instance one compares recent statements of some national governments´ officials in these countries than it is clear that they evaluated decentralization in different ways, and some of them openly support centralization.

From my point of view any decentralization reform is a complex policy phenomenon, and decision on what we should prefer (centralization or decentralization) depends on the case and plenty of potential factors. In other words, it is obvious that decentralization associated with direct democracy - like experienced in Switzerland - could evoke significantly different outcomes and impacts in countries with different politico-administrative cultures, different average education levels of citizens, different economic states etc.

A persistent problem in various societies is oligarchy (rule of the wealthy).   Commerce has a grip on our lives out of necessity.   And usually commerce in a small community is more likely to reflect and serve the needs of the community.   And in a small community, the profit made by small businesses tend to be sufficient for survival but not at the expense of consumers in the community.    Problems (increased profit, decrease quality) tend to increase as the size of a community increases.   And then oligarchy tends to dominate in larger communities.   Because ot its separation from the larger community,  oligarchy tends to be characterized by insensitivity to the needs of the larger communicating.   

There are some interesting books and  articles about decentralisation:

• Decentralisation Advantages and Limitations of Decentralization        http://kalyan-city.blogspot.com/2010/07/decentralisation-advantages-and.html
• Centralized Vs. Decentralized Organizational Structure   http://smallbusiness.chron.com/centralized-vs-decentralized-organizational-structure-2785.html
• Centralisation / Decentralisation https://www.kbmanage.com/concept/centralisation-decentralisation
• Design Patterns for Decentralised Coordination in Self-organising Emergent Systems. In: Sven A. Brueckner, et al (Eds), Engineering Self-Organising Systems. Springer.
• [Gordan_Crawford,_Christof_Hartmann]_Decentralisation_in_africa
• Anirban Pal, Planning from the bottom up-Democratic decentralisation in action
• [Sten_Widmalm]_Decentralisation,_Corruption_and_Social_Capital-From_India_to_the_West.

Is there anything that has not been researched on decentralisation? if so what exactly?

Dear Jane,

It is difficult for me to know that! I suggest to choose one the existed books and try to find the state of the art on decentralisation!