For all you readers that are questioning yourselves, your economics profession, or a convex combination between the two, I shall give a summary and interpretation of what Lionel Robbins, one of the great godfathers of economic thought, had in mind we (environmental) economists should be doing. And hopefully you find yourself mirrored in these views.

There are many times when I wonder whether our profession, namely economists, is of any use at all. This mostly happens after having received a rejection from a journal which leaves me questioning whether the referee has really read my paper at all, and whether it was my lack of economics or his/hers that led to the editor’s decision. If my job contract did not require me to publish then I would simply write for the sake of writing, research for the sake of researching, and leave the papers unpublished or wait until a journal comes to me and says: I want to publish your paper, please!! Since this seems unlikely, well, as they say: Gotta feed the beast!

Furthermore, there are so many cases where economists provide policy suggestions without acknowledging that there are more interpretations or possible ways to achieve similar outcomes. For example, there are quite a few (environmental) economists that only lobby for a worldwide, uniform carbon tax. However, if this carbon tax is politically infeasible and we do not inform policy makers that there are other options (like cap-and-trade programms, or differentiated carbon taxes which may be better than no taxes), then we may end up without any policy at all. Seeing this is something that I view as rather unfortunate.

So during those times it is good to lay back, return to one’s roots and remember why we do what we do. One of the roots of economic thinking is Lionel Robbins. For everyone who does not want to go through his main contribution to economic thinking, his book “An Essay on the Nature and Significance of Economic Science”, I recommend to read his earlier and much shorter Inaugural lecture delivered at the London School of Economics and Political Science on January 30th, 1930. This lecture is entitled “The Present Position of Economic Science”, has been published in Economica, Vol. 10 (March 1930), pp. 14-24, and contains quite a bit of his later book.

The first thing that is interesting to see is that Lionel Robbins’ views from 1930 are still as important and insightful as they were 86 years ago. Whether this implies that his understanding of what economic science is, and what it should do, was perfect, or that this implies that the state of economic science did not in any way improve from his times, I dare not answer and thus I shall leave this open to you, the free-spirited reader out there.

Basically, Lionel Robbins had three points about economic science:

(1) Firstly we have what are admittedly gaps in the existing body of theory.

I think everyone can fully acknowledge that this is as true as ever. With every new research question that I attempt to answer several new ones jump out. However, whether our research is fishing out the pool, and thus only closes gaps in the existing body of theory, or whether it is standing on the shoulder of giants and thus ever expanding, is an important distinction and open question. During my PhD I worked with a good friend of mine, Jakub Growiec, on what we (fortunately only) initially called the Amoeba Theory of R&D (building upon Ola Olsson’s work, thus we were at most innovating incrementally). Radical ideas increase the knowledge frontier, while incremental ideas fill the gaps in between. Whatever grows faster obviously determines whether our overall knowledge is stuck within a certain frame or potentially expanding outside of this mindset.

Lessons: I think most of our work is filling in the gaps. Acknowledging this should make us researchers more humble in how we view our own work as well as our contribution to society.

(2) Secondly come deficiencies due to defects in the logical structure of existing theory.

Specifically, Lionel Robbins refers to

… links in [an economist’s] arguments of whose weakness he was dimly conscious, although maybe he was convinced that in the interests of progress it was necessary to tolerate their imperfections.

I think this is where our mathematical modeling is holding us back and aiding us at the same time. It is holding us back because most journals require a well-defined mathematical model as support for one’s arguments and views. This, however, often requires very stringent assumptions that are either unrealistic or unduly restrictive and may subsequently lead us to wrong conclusions.

For example, there has been much focus on whether a variable, let’s say non-renewable resources, is extracted monotonically or not. Initial models claimed it is and this was an important result. But this result was due to strong restrictions on functional forms. Similar restrictions are found, for example, in the literature on coalition formation. The functional forms used lead one to conclude that, in climate change coalition games, no stable coalition larger than N=4 exists. This result has obvious important policy repercussions, and it would be interesting to know inhowfar this has impacted real policy. But the main lesson to take away is that if one were to introduce more general functional forms, then the coalition size can be large up to even the grand coalition.

The mathematical model also helps us since it requires us to clearly state our assumptions and conditions used, which then helps us in understanding where potential limitations are coming from. This sets us apart from other disciplines that use a less rigorous approach and consequently aids us in further developments. Although, it must be said, the immense number of papers that is being published recently on all topics throughout our discipline makes it extremely difficult to keep an overview.

Lessons: Thus, we have to be wary of our assumptions and understand their repercussions. Again, being more humble about our results given stringent assumptions would be helpful.

Our godfather had a particularly important insight here:

I should be inclined to say that what has happened, here as elsewhere in the theory of value, is that the old plan of analysing “one thing at a time”… seems to lead to inconsistencies which can only be eliminated by the abandonment of this method.

One of the great difficulties of combining theoretical speculation with statistical analysis in the past has been that theory has been preoccupied mainly with what happened when one element in a situation is varied, other things remaining the same. One of the most hopeful developments of the present time is the disposition of theoretical economists to release more than one variable from the pound of ceteris paribus. The theory becomes more complex, but its application becomes more practicable.

Ceteris paribus is useful to study certain conditions if the general case does not provide us with clear enough answers. But this also means that in the general case there are no clear enough answers. The focus, if possible, should therefore be more on the general equilibrium. This is obviously much more difficult than a partial equilibrium analysis. Thus, referees need to appreciate it if researchers try to go general equilibrium, even if it is sometimes not possible to get complete answers.

It may be agreed that it is highly desirable that economists should ask themselves more frequently than they do what is the ultimate significance of their conclusions, not only in the universe of discourse of their own sometimes narrow postulates, but also in that wider universe of discourse common to all the social sciences. There is not enough of this, and I think it is a task which is particularly incumbent.

This also implies that we should not punish those that come with new ideas and approaches, or those that question existing ones, since they may be the ones that lay the foundation for radical ideas and thereby help expand the frontier. While it may be difficult to see any direct application of this kind of work, as long as it helps to question, time and again, what we do, then it is already an endeavor worthwhile to undertake.

The third and final point raised by Lionel Robbins:

(3) Thirdly come deficiencies, due, not to logical defects, but to an oversimplicity in the assumptions of existing theory. A theory may be perfectly consistent in itself and yet not be applicable to the explanation of a given situation, because the assumptions from which it starts are either too simple or not in harmony with the facts. Undoubtedly the methods which have been devised for our use by statisticians are methods of great power and utility, but there is no reason to suppose that they will be any more fruitful than the more primitive methods of the past, unless guided by suitable theory.

I feel that much of the literature in, for example, empirical energy economics lacks precisely this suitable theory, this more foundational work. There are, of course, very good works out there, but there is also quite some data mining and throwing in variables into regressions until something comes out significant. While this may help under certain circumstances, just like the IPAT formula has helped us to e.g. acknowledge the importance of population for the environment, it should not be interpreted as the real relationship between the variables in question.

Lessons: We need to remember or develop the framework of analysis before trying to interpret whatever correlations or relationships we see.

Sometimes we shall go fastest by theoretical speculation. Sometimes it will be the study of particular concrete problems which suggest new generalisations. Sometimes we may do well to go back to the origins of existing doctrines and discover the precise intentions with which they were elaborated. Out of all these channels of inquiry the economics of the future will take shape.

A friend of mine once said: “If the data does not fit the theory, too bad for the data.” So it is of course true that the theory may not hold for a particular dataset, while it holds for another. But then one has to understand why that is the case!

Lessons: Especially in econometric works, we have to be sure about what we expect based on the theory and not just adjust the econometrics until it fits the theory, but adjust the theory as well until some kind of convergence is achieved. A mix of induction and deduction is the way to go.

The world of economic reality is a complicated thing, and it is not to be expected that as we come to understand it better if our generalisations should be less complicated. It is no service to knowledge to make things simpler than they are.

But of course it is important to know where precisely is the fine line between generalization, simplification, and thus making the theory more widely applicable; and ending up with saying next to nothing! This is maybe a question of experience and asking the right questions, too, looking at a problem not only from one modeling perspective but from a variety of angles.

I would close this with what I view as potentially Lionel Robbins’ most important point about our profession:

Economics cannot tell you whether you ought to do a certain thing. But it should be able to tell you what will happen if you do it. Lack of good will it cannot cure. But lack of that understanding without which good will cannot be effective, it can gradually learn to remedy.