Anticipating the Unintended is an excellent newsletter, and you should subscribe to it. This past Sunday, the authors came up with a lovely read on the Agnipath scheme. You may or may not agree with their analysis, but I would heavily recommend that you read it. There’s a lot that is important in it, but for today’s blogpost, I want to focus on this excerpt:
Considering the constraints, it is difficult to see what else the government could have done here. The need to reduce wage and pension costs to fund modernisation is real. And given the fiscally conservative instinct of this government, it won’t deficit fund the modernisation programme. As is its wont, it has chosen to put a bold announcement with emphasis on other benefits while trying to solve its key problems under cover. There’s this myth that a big bang approach to reform is the only model that works in India. That’s wrong. A lot of what has looked like big reforms in India have actually had a long runway that’s often invisible to people. A more comprehensive reading of the history of ‘91 reforms makes this clear.https://publicpolicy.substack.com/p/173-lathpath-lathpath-lathpath-agnipath#details
So, the usual template has been followed so far: minimal consultation, no plans to test it out at a smaller scale and instant big bang implementation. The results are unsurprising.
Let’s figure out they key questions at play:
- What is the problem?: Wage and pension costs for the military spiraling out of control. You could argue that this is an old, inevitable problem made worse by the implementation of the OROP scheme, but for the moment, look past the cause and consider the effect. And the effect is that the Indian government spends over half of it’s budget on pension benefits (24%) and on wages (28%).
- What is the proposed solution?: That’s the Agnipath scheme in its entirety. I invite you, once again, to read the whole thing, but the first two to three paragraphs in the newsletter summarize the scheme very well, if you are not familiar with it yet.
- Why is this solution important?: Their takeaway is that the main focus of the Agnipath program is to reduce wage and pension costs, and that this is necessary. I agree on both counts – no matter what else is being said, and no matter what else Agnipath might achieve (a younger military, among other things) it’s main aim is to reduce wage and pension costs. And even a cursory glance at our government’s finances should make clear that this is necessary.
- The How: That’s what the rest of today’s blogpost is about! Here’s the thing: there is a problem, and it needs a solution. That (to me, at any rate) is clear. But is this (Agnipath) the best possible solution? And even if it is, is the current method of implementation the best way of going about it?
I don’t mean to get into a discussion of whether Agnipath is the best possible solution for this specific problem, nor do I mean to definitively answer the question of whether the method of implementation is optimal.
Instead, I hope to help you build a framework to start to think about the answer to these two questions (is this the best solution | is this the best way to implement said solution) in general. And then, if you like, you might want to use said framework to judge for yourself the Agnipath solution. I do exactly that in what follows: outline the principle, and apply it to the Agnipath case.
Minimal Consultation | No Plans to Test It Out | Instant Big Bang Implementation
Of the three things that RSJ and Pranay have highlighted in their excerpt , I plan to focus on the latter two in terms of the how question. The answer to the question about whether consultations were done or not, and whether they were minimal or not is essentially a grotesque Rorschach test, and I’ll skip it entirely. But the latter two – no plans to test it out, and instant big bang implementation – don’t just ring true, but are also truly important, especially if you are a student of public policy.
Let’s focus on the word “It” in the second phrase, “No Plans To Test It Out”. What does “it” mean, in this context? That’s simple, you might say – test the solution out.
And what’s the solution? Agnipath, you might answer a tad impatiently. Ah, but how do we know that this is the best possible solution? And here’s the simple answer to this question: you don’t know that this is the best possible solution, because as with everything else in life, the proof of the pudding is in the eating of it. Your models might tell you that this solution is the best one, but all models work well, if at all, only in theory. To mix an apt metaphor, no battle plan survives contact with the enemy.
In other words, Agnipath is one of many different solutions to this problem. Moving to something like the NPS might be one, a modified version of Agnipath might be another, curtailing expenditures in other areas might be a third (to those who know their public finances in an Indian context, no I don’t think so either, but play along for the moment). May be the final solution that will be implemented at scale will be a mix of all these and more, who knows – but the point is, there are many possible solutions, of which at least some are worth trying out in an experimental sense.
The process should be seen as experimental, and probably involve acting on multiple potential solution ideas at a time (instead of just one). It can also be accelerated to ensure the change process gains and keeps momentum (to more or less degree, depending on where one is in the change process and whatAndrews, M., Pritchett, L., & Woolcock, M. (2017). Building state capability: Evidence, analysis, action (p. 288). Oxford University Press., pp 170
problems, causes or sub-causes are being addressed). Trying a number of small interventions in rapid “experiments” like this helps to assuage common risks in reform and policy processes, of either appearing too slow in responding to a problem or of leading a large and expensive capacity building failure. This is
because each step offers quick action that is relatively cheap and open to adjustment; and with multiple actions at any one time there is an enhanced prospect of early successes (commonly called “quick wins”).
So begin small, and begin with many different potential solutions. See which of these work and which don’t – which need tiny modifications to be better, and which need major surgery. Which might be usefully combined with some other solution(s)? Iterate towards a solution set that works “best” – and then implement this solution set at scale. But never presume, and especially without on the ground small scale implementation, that a proposed solution is necessarily the best one.
Pritchett and Woodcock have an excellent diagram for helping us think through this:
“A.” in this case is the status quo, and given that there are no overt protests about it, we can say that the status quo is administratively and politically possible. But it is, as we have discussed, not fiscally sustainable, and hence undesirable.
“D.” in this case is the Agnipath solution. Let’s assume that it is a technically correct solution, and let’s assume that it (or something very similar to it) has either solved a similar problem in other countries (empirically validated) or has solved this exact problem, but in a simulation (theoretically validated). But it is not, as we are observing, politically possible. Now, sure, you can talk about the political motivations of the organizers of the protests, you can (and you should!) condemn the use of arson and wanton violence, and you can bemoan the role of the media. But if you accept that there is at least an inconvenient iota of truth in the idea that some sections of society feel hard done by this decision (warranted or otherwise), then you do in effect accept that it is not altogether politically possible. Your assessment may well differ from mine (and that is fine) but hopefully only in magnitude and not in direction.
The point is that we are trying to move from “A” to “D” in one fell swoop. Not only is that not a good idea in public policy, but we don’t even know if Agnipath is “D”! The very definition of D (or Agnipath) has changed in the recent past, is changing as we speak, and is likely to further iterate in the days/weeks/months/years to come. Which is as it should be, of course – my point simply is that these iterations and experimentations should happen in the design stage, not the roll-out stage.
So the general principle is that iterating through potential solutions at restricted scales is better. Even better when you learn from these iterations, modify your solutions, and come up with a hybrid that stands a better chance of working at scale. This helps in building out buy-in for your proposed solution as well. Won’t work perfectly, because nothing ever does, but remember that in public policy utopia ain’t our aim, being better than the status quo is.
Also, if you’re wondering about the title of the blog post, it is a tribute of sorts to a paper that ‘started’ studies in this particular area. Look up that phrase and the the name “Charles Lindblom” to go down a very nice little rabbit hole.
Bottomline: Crossing the river by feeling the stones remains excellent advice.