« Link from Twitter | Main | 280 Paper for Thursday »



Chandler Moody

I think one of the most important points that Kruger makes is that we need to always recognize underlying assumptions. We ought to approach any conclusion drawn from a model and see what has been omitted. Looking out for underlying assumptions is something that we did with Solow Growth model. Key assumptions here included the assumption that capital is subject to diminishing returns and that labor and capital are substitutable. While simplified models are useful, it’s crucial to be aware of their limitations. The power of simplification comes with recognizing what is being left out. I personally love models and think that they’re an aspect of economics that contributes to why I prefer it to other social sciences. I like the concrete-ness that they provide in understanding theories. Kruger’s story of Africa in the evolution of map making in the 20th century stands to support his argument that sometimes the less we know (or evaluate), the more we can know. Narrowing vision and using more simple models can be logical. It’s a paradox that I think he argues well, and in the end I agree with him.

Ferrell Carter

I agree with Wilson’s point that the very basis of economics is realizing the shortcomings of simplified models in order to understand their usefulness. While Hirschman and other high development economists of the time seemed to “fail” at relating these simple models to their more robust theories, such as economies of scale, Krugman explains that this made their ideas “unteachable.” I would disagree and say that complex economic theories are the reason for teaching economics. Though they may not come out perfect on paper, the simple models alone provide only a broad and unrealistic picture of the world. If by “unteachable” he means “imperfect,” then I suppose he is right.

Reading through some of the previous posts, it seems that a lot of the other bloggers agree that there are almost two sides: Krugman and the pro-modelers and Rodik and the case-by-case analyzers. I am under the impression that the two can go together quite nicely. For any science, as Krugman pointed out, there are models. Though some may be hesitant to embrace such rigid structures in a field such as social science, the subject is a science and thus requires some kind of scientific approach to looking at things. Each model is potentially a blank slate. It can tell the basic stories, but it cannot account for specific characteristics of countries and their politics or economies. For this reason, a case-by-case analysis is necessary for each and every development study that uses a basic economic model. As Krugman says, “[modeling] involves the evolution of ignorance as well as knowledge.” In other words, you have to give up specificity and detail to produce an economic model that tells a story. You have to start with something, an empty drawing of Africa, to get somewhere. In order to make up for this loss, Rodik’s point of taking everything case-by-case could fill in the gaps.

Callie Northrop

One of the most interesting points, and one that set the stage for the remainder of the paper, was the story about the African Maps. We live in a world that believes that with more knowledge, better technology and techniques, and easier access to information comes improvement. However, as the African map example shows, this is not always the case. Esther Duflo’s Women Empowerment and Economic Development also presents a case for this point when discussing the growing differences in sex ratios due to the technology improvements that make sex identification and abortion more accessible. Krugman uses this idea to show that progress in development economics actually reversed with the new standards for economics. It is interesting to see that because Marshall and other development economists published their material “smoothed over with parables and metaphors” instead of giving mathematical evidence or well-researched, nearly proven graphs, any insights and information they had gathered was considered worthless or never really looked into further, bring development economics back to zero. It is interesting to think of the advances we could have made if we had continued on the path of many of the early development economists.

Andrew Winter

After reading the posts above, I was immediately drawn to Curtis' first sentence where he mentions how it is most likely a mistake to look at this paper as making a stance and being either pro-models or anti-models. I believe it was our class last Thursday where we talked about how almost every conversation we have where we attempt to communicate a position stems from the use of models. Analogies and metaphors are in a sense just very basic models that we use to illustrate our thoughts. Building off of that, I don't think any rational economist would consider themselves anti-models. What Krugman and some of my peers above have alluded to is that there are times where certain models are unable to accurately illustrate what is truly taking place and that this is particularly true with models explaining economic development.

Krugman makes a concerted effort to express his admiration for Albert Hirschman but wants to make sure that he presents his argument against Hirschman in this specific context. In doing so, he states that Hirschman's flaw was not in his models, but rather in his methodology.

"The second theme is the problem of method in the social sciences. As I will argue, the crisis of high development theory in the late 1950s was neither empirical nor ideological: it was methodological. High development theorists were having a hard time expressing their ideas in the kind of tightly specified models that were increasingly becoming the unique language of discourse of economic analysis. They were faced with the choice of either adopting that increasingly dominant intellectual style, or finding themselves pushed into the intellectual periphery. They didn't make the transition, and as a result high development theory was largely purged from economics, even development economics."

Sorry for using such a long quote, but I thought this paragraph presented the best summary of Kirschman's mistake. It also gave Krugman an opportunity to point out that Kirschman's high development theory was perfectly reasonable. Krugman says that all he needed was a different perspective and attitude towards translating real life complexities into simple models, which is probably easier said than done.

Richard Nelson

I appreciated this article a lot. For one, he spelled out exactly what is was going to cover and clearly labeled his transitions. Second, he took the style of truly telling a strory, much like the folklore and other examples he offered. His analogy of the development of high development theory, which encountered an unnecessary time period of ignorance and neglect, is convincing. However, he seems to have a very negative tone in making this statement, almost as if we should be ashamed as a society that we have allowed ourselves to make this mistake again and again. This is where I began to disagree with a few points that Krugman makes about models. “We all think in simplified models, all the time. The sophisticated thing to do is not to pretend to stop, but to be self-conscious -- to be aware that your models are maps rather than reality.” I think this an excellent idea, but studies have shown us that even when we are made aware of our biases, we are still susceptible to them. It is our mode of rationalization – we want to believe what we are pre-positioned to agree with (confirmation bias). Krugman claims that the social sciences are uniquely inclined to face this issue, but I would argue that this is just human nature.

He then says, “When it comes to physical science, few people have problems with this idea.” He limits his physical science scope to physics and hands-on, model friendly studies. I took a course called the “History of Geology” and that situation was even worse. They argued for decades and centuries about rock formations, glacial deposits, geocentricity, etc…all things that they had to interpret from the evidence observable or left behind. Why did this happen? For a time, they were launching new studies and finding new evidence, but a majority of the back-and-forth debates were simply a popularity contest; which “lead” scientist had more people behind them and made a more convincing argument in his paper or conference decided the acceptable way of thought. If you couldn’t explain it through empirical evidence or a model, good luck. Krugman successfully told the progression of the high development theory, without question.

He analyzed it through similar situations, as in the mapping of Africa or the clues that the shape of a cloud can offer to weather patterns. But, I think he wrongly accuses the social science realm as having a chip on its shoulder towards new ways of thought. We often believe what we are told, and that is why we have to take a step back for every few steps forward.

Brian Lawler

As others have said, Rodrik does an excellent job of encouraging the use of simplified models as stories. However, I agree with Rodrik's statement at the end that we should not ignore either purely model based theories or story based theories. Rather, I believe we should strive to build theories that not only involve stories, but also refer back to some model whether it be advanced or not. A theory backed up with a sensible model more often times makes it to the classroom.

Mac McKee

Reading Krugman's paper, I was quickly reminded by a paper written by political scientist Giovanni Sartori, who argued the inverse relationship in social science between a theory's (or a model's) explanatory power and empirical validity. That is, as we attempt to describe real-world phenomena with more generalizations and universals, we can more easily explain these phenomena and infer greater insights. In order to gain these general insights however, we must distance ourselves from concrete, empirical evidence, resulting in the theory being invalid in some specific cases. On the other hand, if we craft a model supported strongly by empirical evidence, it will in all likelihood explain specific cases accurately. However, because the model focuses heavily on observation and specificity, it becomes more difficult to infer useful insights.

Krugman echos Sartori in an economics-specific context, the assumption necessary in modeling serving as an analogue to the generalization I mention above. Each assumption a model requires removes nuance and connection to the real world.

Models are of course useful - they allow us to explain, assuming certain conditions, real-world phenomena and to extract useful conclusions. Even better, if it's a good model, it can be applied with great breadth. I think they have their place; I'm actually of the opinion that generalized models are an effective and practical way of advancing the thought in a social science field. We only need to keep the assumptions in the front of our minds and recognize that sometimes the assumptions don't hold and - in that case - the model needs to go.

Griffin Cook

I thought this was a great companion piece to Rodrik’s article, as Krugman effectively promotes models as an imperfect but essential tool for understanding economic theory. The key is to avoid using economic models as the foundation for economic theory and development, and to instead use them as a tool for further personal exploration and experimenting rather than a rigid structure that limits the scope of one’s discovery. One of the inherent assumptions in any model is that is over-simplicity inevitably leads to it being wrong in some ways or inapplicable in certain situations. I think that Krugman’s map metaphor accurately illustrates this point. A map is not intended to be a perfect representation of reality, because its simplicity and design leave out almost all of the details that are important attributes of the real place being depicted. It’s impossible to say that someone “knows” a location simply by having observed a map of it, but the map can act as a guide for first-hand exploration. Furthermore, like with the example of the African maps losing their interior, the only way to build an accurate map of these locations is to first wander off the beaten path and explore. The same goes for economics: while others who cannot explore an area will need an accurate an effective map to understand the layout of an area, someone must first set-out on his or her own and do the dirty work of exploring the uncertain landscape first, and judgment will have to be used when deciding what to leave out and what will be assumed when capturing the principle idea in the form of a model. As several of the other bloggers have stated, the key is to not be “pro” or “anti” model, but to know when to what the limits of current models are as economic tools, but also that these tools are important for helping others understand a simplified version of a theory and verifying that the hypothetical argument behind a certain theory has some grounding in reality.

Austin Pierce

I think that Kate raises a valid concern in how models might seemingly work in one specific set of conditions but not across variations in circumstance. However, I think this is where it is important to remember the "ceteris paribus" and assumption rules of models and theories. Many assumptions are made that might not hold in reality. If the assumptions do not hold, then the theory itself says that the results might not be as the theory predicts when the assumptions hold. I think the problem here is that people have applied theories without remembering to check the assumptions.

Another possibility is, of course, the dreaded omitted variable bias. Religious and ethnic strife is of course likely to significantly and substantially affect growth, and one could easily imagine a situation in which the strife leads to stagnant growth due to divisions of skill, assets, and many other variables.

I overall would agree with Krugman that the models are still powerful and useful tools, and more importantly probably the proper tool by which to pursue economics. But as we go forward, we must look at ways to compound various models and theories to achieve more holistic pictures.


I always enjoy reading Paul Krugman articles for class. Furthermore, the fact that he argues for the necessity of using economic models in the paper made me think: at some point Professor Goldsmith read this piece and was undoubtedly thrilled. But I digress. The notion that models do not have to be extremely complex or entirely comprised of mathematical operations to be helpful and/or significant is an important one. Without the emergence of this thought, development economic methods and theory would perhaps have stalled.

Another interesting conclusion in the paper, which ties into the notion of modeling but can be expanded to include other areas of practice, is the idea that increases or changes in technology can lead to higher levels of ignorance or even the abandonment of previously gathered/determined data. The apparent obsession of some to create extremely complex and perfect models can minimize the importance or impact of data derived from some other means. As many of my classmates have already discussed, Krugman's example of Africa is fascinating. It made me wonder: what else has fallen victim to the persistent need from some to perfect the mathematical modeling process?

The comments to this entry are closed.