In practice, theory is something else. I've already heard variants of this expression in several countries and languages. Very often from people referring to the gap between abstract, generic principles and the implementation of projects and policies.
As World Bank President Jim Kim has recently remarked , "Many countries have strong coherent development policies and programs - on paper. But they are not getting the results they want. In country after country, sector after sector, the greatest challenge is delivery."
A recent Economic Premise  note - "Public Sector Management Reform: Toward a Problem-Solving Approach " - by my colleagues Jurgen Blum, Nick Manning and Vivek Srivastava, points to Public Sector Management (PSM) as an area where improving delivery is essential for economic growth and social welfare. After all, "Government revenues and spending average above 30 percent [of GDP] in developing countries", and "public sector management shapes how the public sector machinery translates these resources into results.”
I couldn't agree more. The way taxes are collected, budgets and policies are set up, and public employees are managed and behave affect the results that governments can achieve. And as we have approached here before – "Not All That Glitters Is Gold " -- "improving the quality of public spending may be a way to improve living standards without even necessitating official GDP increases."
Why then is reforming public sector management so hard to obtain, since benefits are so obvious? Jurgen, Nick and Vivek single out four main challenges.
First, designing reforms is easier than implementing them. Public sector management reforms demand changes in de facto behaviors of agents, much beyond de jure institutional arrangements. While the latter can be outlined "on paper", the former requires a lot more.
Second, the results chain is long and it often suffices to have a non-performing link to drag down the whole chain. Third, as the unfolding of reform results usually takes time, there’s not always an alignment with the dynamics of politics. This is particularly the case when beneficiaries of a certain reform are dispersed and the opponents can organize themselves more easily.
Fourth, there is a dearth of evidence on what works effectively. Typical public sector reforms are not amenable to controlled experiments. As the authors illustrate, "unlike deworming pills, a Medium Term (Public) Expenditure Framework cannot be randomized."
They then note how experience has been increasingly favoring a problem-solving approach to reforming public sector management. Instead of relying on first-best, generic reforms designed on paper, the likelihood of success is highest when efforts are based on specific diagnostics, and there is learning and adapting in accordance with evidence as one goes.
If you think this is only an issue for developing countries, think again. This Economic Premise includes results of a recent evaluation of public sector management reforms introduced in the European Union in which a very significant number appears as failing or at least failing to improve results.
Follow the latest from Otaviano Canuto at twitter.com/OCanuto  and keep up with the World Bank's efforts to help countries fight poverty and close gaps in income and opportunity at twitter.com/WBPoverty . For more Economic Premise notes, go to worldbank.org/economicpremise .