Discussions about the role of the state are invariably contaminated by data misrepresentation and political demagogy.
The Portuguese Government in 2013, called a four billion Euros cut in public spending agreed with the IMF a re-foundation of the state. Like in many similar examples of faux liberalism, it was simply a matter of "get out so that I can sit there". Indeed, the government simultaneously borrowed an equivalent amount to create a new development bank. In a country where the state already controls more than 50% of the banking sector, to create one more state-owned bank, in a model that has already failed in the past in Portugal and the rest of the world, can only be deceit or irresponsibility.
In short, the Portuguese government's proposal for a debate on the role of government is not serious and would not deserve commenting. But a real debate about the role of the state is important and should be always present in political discussions.
As Martin Wolf wrote in a recent article on the topic "this is the most important issue of political economy" and has been debated since antiquity by Plato and other philosophers. As he points out, one of the first questions to ask is about the limits and extent of the protective function of the state. The theme is handled brilliantly by the author so I strongly recommend you read his article.
However, to have an intelligent discussion about such limits it is important to know more or less in detail the current role of the state in terms of function, cost and contribution to national wealth.
For example, public spending is usually broken down into ten categories, as shown in the following table with data for Portugal:
However, this breakdown is not the most appropriate to understand the functions of the state. The functions of the state should be divided into five major activities: sovereignty, regulatory, insurance, production and distribution.
The importance (for good and for bad) of each of these activities has an impact on the various categories of public expenditure listed in the table above but are not necessarily fully reflected in budgetary terms. For instance, regulation may have little or no budgetary costs but can have huge economic costs. Moreover, they may be budgeted or not (e.g. in Portugal the electricity rates paid to regulators are not fully budgeted). Equally, the redistribution function of the state can be undertaken on the revenue or the expenditure side.
We can also question whether insurance, production and distributive activities must be carried out by public or private entities subcontracted by the administration. For example, why contract construction services in public works and not education services. That is, the debate on state production and state provision must be clearly separate from the debate on the functions of the state.
Most importantly, the insurance function must be clearly separated from the redistribution function. For example, whether we are talking of health, unemployment or weather insurance, one must make a clear distinction between a component of compulsory insurance (subsidized or not by the state) and a discretionary component funded by taxes to deal with exceptional situations (epidemics, natural disasters, etc.).
Finally, with regard to life insurance and pensions, in addition to its subsidization, the fundamental debate should be on the minimum and maximum levels. For example, does it makes sense that the State offers pension insurance to millionaires? Given the weight that pensions have in public spending this is perhaps the most critical aspect in any debate about the role of the state.
In summary, a deep and thoughtful debate about the limits to the role of the state needs a breakdown of public expenditure and its financing by each of the five categories of state activity. The creation of an accounting system that allows this analysis should be the first step of any government that wants to make a serious debate about the functions of the state.
Only after agreeing the limits in each state activity should we discuss separately the issues related to the relative effectiveness of direct or delegated administration, as well as how to solve the free-riding, theft and nepotism inherent in any political system.