I was curious about what kind of timeline Prof Raghuram Rajan (representing the ministry of finance) envisioned in his dialogue with Prof Abhijit Banerjee (from the Massachusetts Institute of Technology) on ?Ideas for Indian Development? at the Annual South Asia Growth Conference, hosted by the International Growth Centre in New Delhi in July this year. When he responded ?close to a decade?, my first reaction was: ?That?s it?? The dialogue covered issues related to optimising tax collection to provide for infrastructure and manufacturing among others (including those on getting the macroeconomic fundamentals right) ? so clearly there was more to ask.
But despite my curiosity, I held back the temptation of asking more questions thinking it fit for a research scholar like me to absorb before questioning too much. But how much was too much?
As I have now found out, not questioning turned out to be the right decision. In a sense, the ?making? bit in ?policy-making? is a little misleading for it allows the fain mind to blindly assume that a policy expert can always make something up to set things right when the need arises. Contrary to the generally held belief, however, ?policy-making? is more akin to solving a jigsaw puzzle than manufacturing or making one.
The fundamental difference is that you know the complete picture when you manufacture a puzzle ? now try imagining the same when you just start solving one. This analogy is still an oversimplification. What makes policy-making more difficult is that you don?t even know if you have all the pieces, let alone the right ones. The very truth that we will never know all the truths (loosely speaking) is, in fact, a well-proven result, and one of central importance in mathematics due to Kurt G?del. In other words, with a limited tool-kit (axioms, if you will), there will remain statements that are true but cannot be conclusively proven even in the purer sciences (we social-scientists are mere mortals). Therefore, it follows that we will often have to live with hypotheses or conjectures.
In the context of my analogy, when I see a piece with the face of a tiger on it, I can only hypothesise that the complete picture (i.e. the one that will emerge if I got my hands on all the remaining pieces) is either of a forest, or that of a zoo. The challenge arises when we have to choose between the forest-hypothesis and the zoo-hypothesis above. Given that we will never know the truth, it only makes sense to ask: ?In which of the two pictures is my piece more likely to fit??
To offer a more relevant example, two firms may either compete against each other by revising their own prices downward (to attract more consumers) to the extent that eventually each offers a price deadly close to the cost of producing a unit, or they may collude at those low prices just to keep potential entrants away (I will be motivated to enter a market only if the price of what I sell is high enough for my business to survive). Collusion and competition are clearly different things, yet in both the cases, we make the exact same observation(s): two firms and very low prices. The authority will have to tell between collusion and competition beyond reasonable doubt to see if the policy against predatory pricing (collusion) is being compromised here.
This is where the role of empirical evidence or data comes in. The idea is to infer something we do not observe (i.e. collusion or competition?), from all that we can observe apart from the low prices (e.g. expenditure on advertisements, patent-ownership, wages and salaries paid to employees among other things). While this margin is too narrow to contain a detailed explanation, the basic idea is analogous to looking for traces of some form of construction (say, a fence) behind the tiger on our piece. Suppose we do observe a fence, then the correct language to convey the inference is: ?Given that a fence has been observed, we expect that our piece belongs to a picture of a zoo.?
A careful reader will immediately notice that in the process of conveying the correct inference above, one is in fact implicitly saying two different (but not opposite) things simultaneously. First, that the chances of inferring that the complete picture is that of a forest (given the fence), when in fact it is that of a zoo, are very dim. And second, that the chances of us inferring that it is a picture of a zoo (again, given the fence), when in fact it isn?t, are also very dim. To be able to say both the above statements still more confidently, it is clear that we would require more empirical evidence or data (like if we had many more pieces, preferably including one with the name of the zoo on it!). It should be clear by now that since we never know the reality, we researchers constantly face the risk of choosing the wrong hypothesis particularly in the absence of data, or ?pieces? of information, to be in line with our analogy in which I had to make do with (again loosely speaking) a model that links zoos with fences. We researchers work with models because we never know the complete reality; we forecast because we do not know the future; and we infer because we can?t conclude.
Now here is an inescapable truth?policy-makers are researchers first. To find out a policy-mix that is best suited to our scenario, we must first know the entire scenario. But since we never will, as I have established above, we must infer aspects of the same from what we are able to observe. Even Einstein?s relativity theory had to be empirically confirmed by a lensing experiment (due to Eddington), although Newton?s idea of gravity still explains the planets? orbits equally well.
To offer a final word in relation to policy-making, I will bring up Prof Rajan?s (now the Reserve Bank of India Governor) recent comment on inflationary tendencies that are likely to prevail till the end of 2013. A historical perspective is relevant here. Things looked great back in 2003-04, when the cash reserve ratio (or CRR?the proportion of deposits a bank has that it needs to park with the Reserve Bank of India) was reduced from 4.75% to 4.5% with the motive of gradually bringing it down to 3% by 2009. Banks now had more to lend to the public. Rise of liquidity increased consumer demand?more money in hand facilitated the willingness to spend. Business owners found opportunity and GDP kept increasing. A direct and steep drop in the CRR from 4.75% to 3% was out of question?too much liquidity at once would have immediately increased consumer demand relative to supply (which takes time to adjust to sudden changes in demand) resulting in steep price increases. But, there were things that we couldn?t foresee then. India was later competing against Chinese fuel demand (driven by internal developments) which internationally jacked up prices per barrel of oil. For India, the effect of this costly import trickled down to the domestic fuel prices, which in turn led to higher food prices and costs of transportation etc.
Finally, in order to nip this problem of expected inflation off the bud, the CRR had to be raised to levels around 6-7%. The ultimate motive of bringing it down to 3% by 2009 lost its value to this pre-emptive policy stance. Inflation persisted, not because policy-makers got it all wrong ? it?s just that the limitations to policy-making are primarily human. How much we can control, will always depend on how much we can know.
The author is a research scholar (PhD Economics) at the Indian Statistical Institute, New Delhi. He is also committee member, American Statistical Association; and member, Royal Economic Society of England. Views are personal