!VAR HAGENDOORN

Daron Acemoglu: Introduction to Modern Economic Growth

With Introduction to Modern Economic Growth Daron Acemoglu has written what may well be the definitive introduction to theories of economic growth. More than an introduction it is also an excellent survey of all key theories and models. As such it is likely to become an indispensable reference book for anyone working in macroeconomics. And with a total of about 1,000 pages one may rightly wonder what might come after this introduction.

With such a large book, to get a sense of the material covered, it is often best to first read the contents, the introductory chapter and the conclusion or in this case the epilogue, in which Acemoglu reviews all of the lessons that can be learned from the preceding chapters. To review and summarize this material here would require a longer essay than I have time for, so I will confine myself to a few points. I must also confess that I haven’t read all chapters start to finish and didn’t check the mathematical proofs.

Some countries are rich, while others are poor. Some of today’s rich countries were once poor, some such as Hong Kong, Singapore and South-Korea less than a few decades ago, and some countries that were once rich are now poor. These simple observations give rise to several fundamental questions. Why is it that some countries are rich and others are poor? Why do the economies of some countries grow rapidly while those of other countries stagnate? What sustains economic growth over longer periods of time and why did sustained growth start about 200 years ago? And why did it start where it did and not somewhere else?

Over the years numerous economists have tried to answer these questions. The first and foremost to do so was Robert Solow, who was awarded the 1987 Nobel prize in economics for his contribution to the theory of economic growth. For simplicity Solow assumed that the economy consists of two entities, households and firms and produces only one good. He also assumed that households own the capital stock, which they rent out to the firms and save a fixed fraction of their income, that is of the output generated by the firms. Solow next assumed that capital depreciates at a fixed rate. He then assumed that the total output of the economy at any moment in time is a function of the capital stock at that moment, the amount of labour and the level of technology, which determines the total factor productivity. Thus an increase in output in the next period could result from an increase in either the capital stock, the amount of labour or the level of technology. To solve the model we need to further assume that firms try to maximize their profits and that at any moment demand for both capital and labour has to equal supply. Since savings equal investments capital is accumulated at the rate of savings minus depreciation. This is one of the key features of the model. To see how all of this works it is best to take a look at a simple example or to take the model further apart than I have done here.

The Solow growth model as it has become known is strikingly simple. Its major accomplishment is that it creates a link between the capital-output ratio and the investment-depreciation ratio and that it does so in a dynamic model. The main test for any model is how well it holds up against the data. Perhaps the biggest surprise is that despite its simplicity the Solow growth model can be applied to economic data. The results have been mixed though as Acemoglu writes in an extensive review of the literature. This is not necessarily bad news, for it points at some of the other factors that contribute to economic growth and differences across countries. One conclusion that can be drawn from the empirical literature is that cross-country differences in income per capita cannot be understood on the basis of differences in physical and human capital alone. If we are to gain a better understanding of the forces that drive economic growth we should therefore study differences in technology and/or total factor productivity.

The Solow growth model laid down a general framework for studying economic growth. Many subsequent models departed from this basic model by relaxing its explicit and implicit assumptions. The model does not explain how or why technological progress occurs, nor does it explain the rate of capital accumulation. It also assumes that technology is equally available to everyone, that households and firms are homogeneous and that both the amount of labour and the savings rate are constant. Acemoglu has structured his book around the various efforts to remediate the limitations of the Solow growth model. This is a good choice, not only because it shows the incremental nature of scientific research, but also because it serves as a benchmark to test whether a new model actually fares better against the data.

The Solow growth model only specifies the production function. To this neo-classical growth models added a model of consumer behavior thereby making the savings rate endogenous. From a theoretical point of view this was a major advance and the neoclassical growth model has become one of the cornerstones of macroeconomics. However, as Acemoglu observes, study of the neoclassical growth model did not generate any new insights about the sources of cross-country income differences and economic growth relative to the Solow growth model. Its main contribution is that it paves the way for further analysis of capital accumulation, human capital investments, and endogenous technological progress.

A large part of Introduction to Modern Economic Growth is devoted to models of endogenous technological change. In the Solow growth model technological change is exogenous. This is obviously an unrealistic assumption, since technological innovations are the product of investments in research and development. Modelling technological progress opens a Pandora’s box of possible caveats. To make technological progress endogenous we again have to start by making various assumptions.

Solow divided the world into capital and labor. To this Paul Romer, one of the intellectual fathers of endogenous growth theories, added a distinction between objects and ideas. Objects include capital and labor, ideas can be thought of as instructions for using or making objects. An important concept introduced by Paul Romer is the nonrivalry of ideas. The same unit of labour or capital can only be used by one producer at the same time. Ideas by contrast are nonrivalrous in that they can be used simultaneously by any number of producers without reducing their inherent usefulness. Once a new idea has been invented it does not have to be reinvented to be used by others. Nonrivalry of ideas therefore leads to increasing returns to scale. However, under pure competition there is no incentive for any firm to innovate, since any innovation will be instantly adopted by other firms and thus the firm will be unable to recover its investment. This explains the existence of patent laws which give the innovating firm temporary monopoly power and the reluctance of many companies to transfer high technology production to China or Vietnam.

It follows from these considerations that in order to model endogenous technological progress we should also model market structure and allow for monopolistic competition. This is achieved in models that build on the work of Joseph Stiglitz and Avinash Dixit.

An odd feature of both the Solow and the neoclassical growth model is that they both treat firms and households as constant. In reality of course people are born and die and new firms are founded while older firms are driven out of competition. The first criticism is captured by so called overlapping generation models, the second by models of Schumpeterian growth, named after the Austrian economist Joseph Schumpeter. Schumpeter coined the term creative destruction to describe the process whereby new firms replace incumbent ones and new products or machines replace older models. Examples of creative destruction are the invention of digital photography and mp3 players, which not only replace film and cd’s, but also overthrow the photography and music industry.

As Acemoglu notes models of Schumpeterian growth bring the macroeconomic theory of economic growth closer in line with the industrial organization of innovation. To give a sense of the challenges of modelling Schumpeterian growth, a first step might be to assume that all innovations are due to new entrants. This one can model, except that it is at odds with the empirical evidence. We can relax this assumption and assume that both existing firms and new entrants can make innovations. To make our model even more realistic we may assume that innovations are to some extent cumulative. When launching the iPhone Apple could build on previous research and development that had led to the iPod, the Mac OS X operating system etc. To capture all of this in a model and ensure that markets are in equilibrium requires advanced mathematics.

Although Acemoglu does at times pause to discuss the empirical validation of the models he discusses, the focus in Introduction to Modern Economic Growth is on the mathematical formulation of the various models and on providing proofs for the theorems on which they rely. Having said so it follows that this is NOT a book for the average reader looking for an introduction to economic growth. Taken together the epilogue and the final paragraphs of each chapter in which Acemoglu reviews what each new model has contributed to our understanding constitute an excellent summary of current research.

The task of modelling is a constant going back and forth between model and reality, between relaxing one assumption while making others. Sometimes a series of realistic assumptions are better captured by a single unrealistic assumption. Thus it might be assumed that aggregate technological progress is stochastic. This is the topic of stochastic growth models and also forms the backbone of Real Business Cycle theory.

In a chapter on the fundamental determinants of differences in economic performance Acemoglu distinguishes between proximate and fundamental causes. If cross-country differences in physical and human capital and technology explain differences in economic growth, then what explains these differences? Acemoglu divides the various fundamental causes which have been proposed throughout the ages into four main categories: luck, geography, culture and institutions. In recent years Jared Diamond in his book Guns, Germs and Steel, Paul Collier and Jeffrey Sachs have been the most prominent supporters of the importance of geography. It seems undeniable that landlocked countries such as Niger, Chad, Paraguay and Zambia are at a disadvantage. Yet being landlocked or an absence of natural resources need not stand in the way of economic prosperity, as the examples of Switzerland, Austria, Singapore and Taiwan show. Geography matters, but only to a certain extent. The same is true for culture. Acemoglu acknowledges that cultural factors do matter, but as he argues South Korean or Singaporean culture did not change much after the end of World War II and yet both experienced spectacular economic growth in the second half of the 20th century. Culture is therefore best viewed in combination with institutions.

Households and firms do not operate in an institutional vacuum. Indeed a fundamental question in institutional economics is why individuals organize themselves into companies in the first place. One of the questions we started with is why some countries are rich and others are poor. To answer this question we should not only look at economic factors but also at political institutions that define a country. As Acemoglu observes in the epilogue the relationship between political institutions and economic growth is one of the frontiers of current research. It is also highly topical. Creative destruction and reallocation are two of the key contributing factors to economic growth, but they can have drastic social consequences when entire industries are displaced. How political institutions deal with these consequences is therefore equally important. For many years Germany has subsidized its loss making black coal mines with several billion euros per year. The mines, which employ more than 30,000 people, won’t be closed until 2018. The replacement of production from either Europe or the U.S. to Asia continues to stir debate. As of this writing the U.S. car industry is facing collapse and so the question is whether government should rescue the car industry or instead create the conditions for the emergence of new industries.

Introduction to Modern Economic Growth is an exemplary book and a triumph of scholarship. It is impossible to do justice to its achievement in a simple blogpost such as this. I can only think of very few other books in economics with similar rigor and scope (The Theory of Corporate Finance by Jean Tirole springs to mind). With governments around the world considering multi-billion dollar stimulus programs in an effort to fight the current economic downturn, that is “negative economic growth”, Introduction to Modern Economic Growth is also highly topical.

My only quible is with the mathematical appendix and the chapters on dynamic programming and the theory of optimal control. While both are prerequisites for an understanding of current research in economic growth as Acemoglu himself concedes the material in these chapters is the topic of many excellent applied mathematics textbooks. I understand that for reasons of completion and self-sufficiency an appendix listing all mathematical theorems used throughout the book was once considered necessary. However, the intended audience of the book will or should be familiar with most of the mathematics or else they will know where to find a reference. For instance, when I was at university the theory of optimal control was the topic of one term course. If you are unfamiliar with Kakutani’s Fixed Point theorem, the Mean Value theorem, the Saddle Point theorem, differential equations and the theory of dynamic games you will be little wiser after reading the appendix. What’s more in that case this book isn’t for you.

Now the question some may ask is to what extent economic growth is predicated on the use of non-renewable natural resources and thereby ultimately finite. Economic growth and capitalism rely on profit and not so much on production. Therefore both capitalism and economic growth are, in theory at least, reconcilable with sustainability. Differentiating between models of sustainable and unsustainable economic growth may be one of the greatest modeling challenges of the future.

Links

The syllabus and slides for Acemoglu’s introductory and advanced course on economic growth at MIT.