There is a sharp difference between theory and policy in economics. Theory aspires to be objective, insofar as a social science - the science of human beings studying themselves - can be; whereas policy is explicitly normative in the sense of being prescriptive - however much propounders of a specific slant may try to convince us all that their is the "fairest of them all."
In the natural sciences (such as physics, chemistry and biology), the methodology followed is "induction" where repeated experiments can be conducted to test the validity of a proposition in order to "get at the truth". But even induction has been famously known to have its own problem, as exemplified by the case of the black swans - a paradigm shift that can be as profound as Galileo's.
In trying to come to some kind of a decent conclusion for economic propositions, economic theorists have decided on "positivism" (a la A.J. Ayer) as the philosophical basis for its scientific foundation. The positivist argument is that while the premises may not be testable (as in "limitlessness of human wants"), the eventual conclusions or propositions derived should be subject to test and be testable. Herein lies the whole study of statistics and econometrics and modelling for the purpose of testing economic hypotheses.
We shall remind ourselves that reality is not neatly separated into economics and everything else. Reality is a mixture of economics, politics, religion, environment, etc. The economic data that we collect to test economic propositions are in themselves constructed out of theories themselves and therefore are bias towards the theoretical framework. There are many structural changes to the global economies as well as the local economies, such that the data points (for a young nation like Malaysia) are inadequate to take into account all the relevant factors. The other factors such as politics and environment are not as "quantitifiable" as economics which in turn is not as "quantifiable" as CO2, for example.
In the face of insufficient data (especially uptodate data) and poorly tested propositions (or "theories" as economists would like to call them), many economists have to resort to a particularly way of thinking on economic issues and to rely on economic insights drawn from years of keen observation to arrive at a judgement on what the economic reality actually is.
The economic way of thinking is the observation of how human beings behave to satisfy their wants and desires, and how they respond to stimuli. While economists focusing on the GDP will want us to believe everything boils down to goods and services, I wish to make the point that how the structure of the economy evolves to be what it is today and how it will evolve into tomorrow is a respond to the fears and uncertainties faced by human beings. At the meta level, i.e., the underlying fundamentals, we know that human beings are destroying the very ground on which they stand in order to have a firm grasp of their future (which we all know is elusive and never here).
In the end, the policy we pursue is a reflection of our own value judgement of how our society should be and become. Policy is normative, with no pretense on objectivity. We may argue that we should be market-oriented in order to be economically efficient, but our society may become more treacherous as a result. In policy, we wish to argue a position for our society to take, as we would wish to argue for our particular way of living. There are some basic economic relationships which we must recognise and take into consideration, if only to arrive at an ideal state - as ideal as anyone can tell us that ideal is necessarily good. But it does not mean that foolish do not exist, and that wrong policy actions are not taken. Errors occur everyday; it is the ability to undertake error-correction that keeps us from ultimate disaster.
In policy, we tend ultimately to argue for a particular way of life for our society or the degree of freedom to live the way we prefer.