First Principle Incentives In Academia
Editor | On 12, Jan 2018
Back in August, we introduced a case study on ‘Perverse Incentives in Academia’ (Issue 185). The point of that article was to try and help those caught in the problem to escape from the usual either/or pendulum-swing solutions that academia often finds itself prone to. Our job in the article was primarily to step back and look at the dominating contradiction and its potential resolution at a conceptual level. We didn’t dig deeper because the point was to make the connection between oscillatory systems and contradictions. A deeper level solution, we knew, would require a more granular look at the behaviours within the system as they relate to the ‘first principles’ upon which the outcomes from the system emerge. The premise when we’re looking at this ‘first principle’ level is that, only if we affect the system at this level do we have any kind of hope of achieving sustainable changes in the outcomes the system produces.
‘First principles’ as far as human behavior are concerned, as at least mentioned in the August article, are the Autonomy, Belonging, Competence and Meaning elements in our ABC-M model. When we look at the incentives that had been created by the well-intentioned educators behind the original paper through the ABC-M lens, we see an immediate problem:
Looking at the four columns on the right-hand side of Figure 1 reveals that the effect of the designed incentives on Autonomy, Belonging, Competence and Meaning are largely negative – i.e. they disobey the ‘ABC-M all get better’ heuristic found in all successful solutions. In regard to the Meaning element, every single one of the incentives put in place within the system serve to destroy meaning. Two of the incentives – rewarding researchers for increased grant funding and rewarding Departments for US News ranking – serve to destroy all four of the ABC-M elements.
We know from the Figure 1 image what the ‘first principle’ problems are – we need to design a system in which, if there is a need for incentives (in an ideal solution there wouldn’t be) they move the ABC-M values in the right direction.
So far so good. The only problem with this kind of first principle solution is that we’re not starting from a blank piece of paper. This is a difficulty when we’re dealing with any kind of complex system: nothing exists in a vacuum. It’s the ‘can’t get there from here’ problem. Knowing where you’re trying to reach doesn’t mean you can ever reach it. This is why we talk about ‘sense or progress’ a lot with our clients… people will stay on the journey if they feel they’re moving in the right direction.
What we’re far less likely to talk about is that ‘sense of progress’, if you’re not careful, can become a journey that takes people further and further along a dead-end street. ‘Sense of progress’ can very easily be interpreted as a continuous improvement job. But all that means is you’re climbing to the top of your current s-curve. Real ‘sense of progress’ recognizes that you always need to be looking for the next contradiction and solving that.
So, how do I know what the right ‘next contradiction’ is?
We can only ‘make progress’ towards our ideal if we choose the current contradiction causing us to become stuck at the current dead-end. This is perhaps one of the reasons the current incentive system appears to be stuck. The other is that, as far as I can see, no-one working in the further education system appears to understand the word ‘system’. Systems – particularly complex ones – operate on the basis of feedback loops. If we don’t design the right feedback loops, we get the ones that naturally emerge. Which rarely means we get the outcomes we want.
If we’re to genuinely solve the perverse incentives problem, we also need to understand the basics of how systems and feedback loops work. Figure 3 offers up an attempt to create this understanding, in relation to the ‘first principles’ idea. The ‘input’ in this model is the outcome we desire. This might be our ‘ideally’ statement for example. Next comes the Controller. This is essentially the dials that managers and leaders are able to alter. In this case study it is the incentives. Next comes the process itself. In this case it represents all the formal and informal protocols and methods through which stuff gets done. Out of the process comes our first lot of results. The feedback loop then looks at these results in order to compare them with the desired input. If the two things are different, i.e. we’re not achieving the results that we set out to achieve, then we’re supposed to change the system in some way. If we’re working at the ‘first principles’ level, that means we’re trying to alter one or more of the incentive dials in the Controller. These are effectively the only things that managers and leaders can directly affect. We can re-design the process of course, at least in theory, but if we consider the ‘process’ to be the emergent combination of written and unwritten protocols and behaviours, its very easy for managers to re-design the formal process only to discover that the informal processes evolve in a manner that detracts from the design intent. Not to mention the fact that in complex systems the links between cause and effect are in any event often quite tenuous, if we’re thinking about systems at the first principles level, it is safest to assume that the ‘process’ is an entity that is emergent rather than something that is design-able. The only realistic way to re-design the process in other words, is to re-design the (first principle) incentives in the Controller.
Probably a lot more needs to be said on the subject. Just not here. The reason for mentioning the issue at all in this article, is that it seems the perverse incentives problem in further education doesn’t appear to have any feedback loops other than the fact that the authors found the energy to write the paper they did. It is difficult to see anything beyond what they’ve done in terms of anybody looking at the results, comparing them with the desired outcomes and changing the system in any way.
Maybe – if we’re being kind to the actors within the system – they’re stuck at the ‘next contradiction’ problem too. When cause and effect links are tenous and inter-connected, it is difficult to be confident that any change to the Controller will deliver outputs that are closer to the desired outcomes.
So, now back to the ‘right contradiction’ question. Here’s where a bit more first principles theory might help: what causes the top of an s-curve to get flat? Answer: the next contradiction. And what causes the contradiction in a complex system? Answer: the vicious cycle (or cycles) that have emerged.
So, our job becomes tracking down vicious cycles. Which is one of the things that the Perception Mapping process is designed to do: we need to draw a perception map of the current realities present within the system. Something, if we were smarter at naming tools and processes, we might call a ‘Current Reality Map’…
Fortunately, in this case, we have some considerable assistance from the original authors, since they did what appears to be a comprehensive analysis of what’s actually happening in their world. All we need to do is translate those Current Reality statements into a map. First up, here’s the how we create the list of Realities:
And then, once we’ve done this we can start mapping how each of the Current Realities ‘leads to’ the others. In an ideal world, we’d do this job with the stakeholders in the room with us, but in their absence, Figure 5 represents the output after we did the ‘leads-to’ analysis ourselves:
The map reveals a single vicious cycle. Zooming in to look at it in greater detail reveals the loop shown in Figure 6:
This in turn gives us a pretty good idea what the current limiting contradiction is. The thing we’re trying to improve is, per the outcomes denoting the premise of the original paper, ‘true scientific productivity’, and the reason we’re not getting it is the Figure 6 Vicious Cycle. Fortunately, TRIZ tells us, we’re not the only people to have had this problem. Figure 7 shows how we might use our Contradiction Template to map the problem, abstract it, and then use the Management version of the Contradiction Matrix to tap in to the Principles used by others to solve the problem.
And then, finally, if we’re going to get the best out of using these Inventive Principles, we need to go back to the Incentives shown in Figure 1 and use the Principles to get ABC-M moving in the right direction:
And now it’s your turn. We’ll explore some of our answers to the problem in next month’s follow-up article. In the meantime, your job over the holiday season (those that manage to get one) is to see what you can come up with. Remember: Autonomy, Belonging, Competence and Meaning all have to get better.