In common usage, corruption is often used to refer to all types of immoral or harmful behaviour by public officials.
But in the social sciences and policy discussions, corruption refers specifically to the illegal use of power by politicians or bureaucrats for their own benefit.
The important point is that this definition does not presume that corruption is damaging, though it may be.
How damaging it is has to be established by theory and evidence, and here there is considerable debate.
Corruption involves two related activities. First, public power has to be acquired or purchased.
Resources are therefore spent in bribes or in efforts to directly capture political power.
These activities can waste resources which could have been more productively invested.
Secondly, public power is then used to create benefits for public officials, or those who have bribed them, or create obstacles for others.
The benefits are beneficial for those who get them, but can be damaging for society.
For instance, public power can be used to create monopolies to import goods, or to grant contracts at inflated prices.
In extreme cases of predation, public officials and their friends can simply loot resources.
Public officials can also create obstacles that citizens have to pay to avoid, like red tape and unnecessary restrictions.
The economic effect of the second set of activities can therefore also be negative and the total effect of corruption is then clearly negative.
International agencies like the World Bank and the IMF assume that corruption does have negative effects and also that it can be removed by reforms.
Therefore, they use their influence to persuade developing countries to spend time, effort and money to reduce corruption.
In this, they are often supported by civil society organisations and NGOs who are also against corruption for obvious reasons.
The policies they recommend include greater transparency and accountability, stricter prosecutions and punishments, and liberalisation to reduce the amount of discretion that public officials have to create privileges or allocate resources.
But much investment in these policies has generally not achieved significant reductions in corruption.
No one can be in favour of corruption. The question really is that, if corruption is so bad, why is it so pervasive? Why does every developing country suffer so greatly from corruption?
And why have all the resources spent on fighting corruption achieved so little in terms of sustained and lasting reductions in corruption, and what should we be doing about it?
To answer these questions we need to look at what the simple analysis of corruption is missing out.
First, it misses the fact that much of the corruption in developing countries is political corruption driven by the fact that political power is often based on the ability of politicians to deliver resources or privileges to their clients that they cannot offer through the budget.
Here the significant difference with advanced countries is that in the latter, the budget is big enough to allow competing parties to offer credible spending plans to voters that can potentially win one of them a majority.
In developing countries this is very difficult because the small budget cannot offer much to voters.
Rather, power is constructed through political networks where powerful faction leaders are rewarded with privileges to maintain political stability, mobilize voters and enable the state to function.
This is also corruption because resources are being spent, sometimes illegally, to construct these networks and the privileges created for the political organisers are often illegal as well.
But the problem is that in the absence of a fiscal base to allow social democratic politics, it is difficult to imagine how else politics can be organised.
In these contexts, the only feasible solution is to make politics more stable and developmental so that the budget can grow over time.
But attempts to immediately root out all corruption typically fail.
A second problem with the simplistic analysis is that what public officials ‘deliver’ varies greatly.
It is not always a monopoly or an obstacle. Sometimes citizens have to pay to get resources to which they are legally entitled and which are socially desirable, such as foodgrains for poor people.
Here corruption has a social cost, but it may be less than the cost of not having the programme at all. Another example is when states make resources available for investment in new or risky areas.
If the state has the capacity to ensure that these resources are not entirely wasted, economic development can take place even in the presence of corruption.
The corruption associated with support for industrial policy is often observed in East Asian countries.
In these cases the bribe is a bit like an illegal tax, which has a cost, but the net effect of intervention can be growth-enhancing for the economy.
These sorts of reasons explain why corruption can be associated with collapsing economies but also with some of the most dynamic economies in the developing world. Clearly developing countries have different mixes of corruption.
In poorly performing economies predatory types of corruption dominate as well as corruption that creates obstacles for investors.
In high-growth developing countries corruption is more like profit-sharing between business and public officials in a context where public officials facilitate and enable businesses to grow.
If we cannot get rid of all corruption immediately, we should certainly try to attack predatory behaviour and looting and try to create incentives for public officials to behave in developmental ways.
This is a very different strategy from the moralistic approach of much of global anti-corruption policies today.
And it has to be remembered that in advanced countries the rich do buy influence, but because of higher levels of institutionalisation, they usually buy influence legally, through lobbying, contributions to political parties, contributions to think-tanks and universities, and by employing ex-politicians on their boards.
This is another reason why corruption gradually disappears as a country becomes richer.
But if we are concerned about justice and democracy, we should be just as concerned with the legal forms of influence-buying in advanced countries.
Mushtaq Husain Khan is professor of economics at the School of Oriental and African Studies, University of London and his publications can be accessed at http://mercury.soas.ac.uk/users/mk17/