Friday, 19 August 2016

Physics meets Economics: "Why Information Grows?" (or What I've been reading, vol. 7)

I haven't posted any book reviews for a while now. It's not that I haven't been reading the books, it's just that I failed to find the time to write the reviews. So now, in the next three, four blog posts, I will be writing a series of reviews about 8 or so books (with more to come in the following months). I will start with Cesar Hidalgo's Why Information Grows for today's single book review (it's a big one, since this is an exciting new theory), and then I move on to 4 economic history books (in two separate posts), and close with three inequality/cronyism books. After all of that, in September I will continue with my standard ramblings about economic stagnation, the role of technology and the third industrial revolution, inequality and social mobility, the endless austerity vs stimulus debates, etc. Enjoy! 

Hidalgo, Cesar (2015) Why Information Grows. The Evolution of Order, from Atoms to Economies. Basic Books

Cesar Hidalgo, a physicist from MIT, in one of the most lucid books I have read so far, presents a new fascinating theory of economic growth. One of the most fundamental issues in the field of economics – why some nations tend to be rich while others stay poor – has been given a huge amount of attention in the field over the past 60 years. A number of theories have arisen and have been tested to account for and explain economic growth and development. From Sollow to Romer and Mankiw, from Kuznets to Leontief, from Porter to Acemoglu and Robinson, many have given their unique contributions, each having its own merits and being probably true to some extent. Hidalgo joins this group by producing a synthesis of social and natural sciences in the study of networks and information. He moves beyond the simple factors like land, labor, and capital, and replaces them with matter, energy, and information. It may sound puzzling at first, but once you dwell into the argument, it’s actually very simple and quite intuitive. Basically, in order to understand how economies grow, we need to first understand how information grows. Hidalgo’s theory is where physics meets economics.

Before I descend into the detailed flow of his argument, here is a brief summary of the basic idea relating networks, energy and information to economic growth. His theory relies on the novel concept of economic complexity. Economic complexity, in simple terms, is the diversity, the number, and the ubiquity of products a country exports.  It is not surprising to notice a positive relationship between economic complexity and economic growth, but what’s even more important is that this measure of economic complexity allows us to predict long-run economic growth. So based on the structure of our economy, and which kinds of products we are specialized in today, we can quantify economic growth ten to fifteen years from now. This makes sense, as the total production structure of an economy determines its long-run equilibrium path – i.e. potential output. Countries which are evolving in their complexity are able to catch up with the countries which have already reached the same equilibrium path. This explains faster growth of developing nations, and slower growth of developed nations – or in other words, convergence theory. E.g. consider China and India - they are evolving the complexity of their economies which is why they are more likely to keep high levels of economic growth in the future.

The theory is still in its infancy, as the research on economic complexity has just started (here's the link to the Observatory of Economic Complexity). However, the ingenuity behind some of the concepts makes it a worthwhile effort and may bring us closer to understanding cross-country inequality. The only thing that the book fails to uncover in greater detail is the initial difference in trade specialization patterns. Yes, the current export and production patterns may be good predictors of future growth, and good explanations over why the US is more complex and hence richer than Brazil, and why is Brazil more complex and richer than Chad. But, there is no systematic answer to how and why some products and industries originated in one place, and not the other. This is beyond the scope of the book, intentionally so. Even though it isn’t big in volume it does attempt to do a difficult task of explaining the interconnection between matter, energy, information, and economic growth. So let's dive in. 

Information vol. 1 (the physics behind the idea)

The first part of the book, the one explaining the origin of order from atoms is pure physics. Fascinating, a bit complex, but given the intuitive explanations the ideas tend to stick as you read more and more.

When an average layman hears the word information he or she thinks of it in standard terms; it is digital, immaterial, weightless, easily transferable, a message of some sort – essentially, a man-made phenomenon. However, Hidalgo emphasizes that information is physical. It is not a thing per se, but rather an arrangement of several physical things. Basically, information implies order. The concept of information therefore needs to be detached from its colloquial meaning in order to understand its usage in the book. The capacity to process information and sustain order which involves computation (in addition to a few other conditions) is what he defines as the accumulation of knowledge and knowhow. In this sense information (knowledge and knowhow) is embodied in the products we produce: how complex they are, how widespread they are, and who knows how to produce them. It is this difference in the accumulation of knowledge and knowhow across countries that gives rise to global inequality between nations. In other words, our patterns of production and exports will determine our different levels of development and economic growth.

To understand how information represents physical order, the author gives an interesting example. Consider smashing a $2.5m Bugatti against a wall. It will still be a sum of all its parts, except now the atoms that made the car will be arranged differently. This different arrangement of the same atoms makes the car lose its value. And significantly so. Furthermore, destroying a Bugatti means increasing the multiplicity of its states (increasing entropy), which implies less information about what these parts were. Creating a Bugatti from scratch implies combining iron, aluminum, plastic, and rubber (with of course some engineering and designer knowledge and knowhow), where all of them individually convey less information that their collective final usage. Therefore, destroying a product means destroying information, whereas the creation of a product implies creating information. So regardless of all the inputs we have at our disposal as a nation, our ability to use our knowledge and knowhow to turn these inputs into tangible and valuable outputs (products – pockets of information) is what will define our success as a nation.

But how does information grow? How does it differ across nations?

Essentially, the author defines three conditions necessary for information to grow, necessary for it to ascend from disorder into order (to oppose the second law of thermodynamics of closed physical systems). The first condition is a non-equilibrium system in a steady-state (I’ll explain), the second is the existence of solids in order to store information in them, and the third is the ability of matter to compute.

The simplest example of a steady-state of a non-equilibrium system (used in the book) is the whirlpool that forms when you empty a bath tub. It is easily created (just remove the stopper and the water starts flowing), and easily stopped (do the opposite). So it’s dynamic, and it requires energy to persist (out of equilibrium). The point of such a state is that once in it, the production of entropy is minimized. Think of this as a bowl where if you put a marble it will eventually end up at the minimum of its potential energy (this is also known as a thermodynamic potential – I actually co-authored a paper explaining high and low corrupt systems through double-well thermodynamic potentials). So in an out-of-equilibrium steady state, order will spontaneously emerge and by doing so it will minimize the destruction of information.

Once we have the system flowing from disorder into order, to preserve the newly formed information, it has to be stored in packages. In other words, it has to be stored in solids. Why? Because solids, unlike gases and fluids, are much better at crystallizing information. Information lasts much longer in solids. Finally, in order to have information grow, matter (and its evolution into life forms) needs to be able to process information, it needs to be able to compute. Compute doesn’t necessarily mean making calculations. It can be limited to the very simple perception of animals and even plants, all of which adapt and adjust to the environment around them. Basically, life itself and where we are today is due to our ability to compute (adapt/evolve). So the combination/existence of these three factors is the key determinant of information growth.

(Gosh, I hope I explained this correctly and intuitively enough. If you find it too cumbersome, read the book)

Information vol. 2 (the economics behind the idea)

Ok, so now that we know the physics behind it, let’s see how this information paradigm fits into the economic life.

First and foremost, our capacity to compute isn’t limitless. If it were, there would be no differences between countries. In addition to our intrinsic individual limitations, there are also societal limitations, such as institutions and the availability of technology. These can prove to be essential in limiting the networking capacity necessary for knowledge and knowhow to expand. The economy itself is an amplification engine of knowledge and knowhow. It is within the complex socio-economic network through which we learn how to produce products and in essence – make information grow. We stand on the shoulders of giants – not only in ideas, but primarily in products. Writing this would not have been possible without the internet or the personal computer. Old products embody information to create new products out of them.

One way of showing the difference in knowledge and knowhow, or in other words the difference in accumulating and processing information, is a country’s export structure. As Hidalgo says, the export structure of a country “is a fingerprint that tells us about the ability of people in that country to create tangible instantiations of imaginary objects such as automobiles, espresso machines, subway cars… In fact the composition of a country’s exports informs us about the knowledge and knowhow that are embodied in that country’s population”.

This is the point – the product structure of a country is defined by the population’s capacity to imagine and produce a wide range of products. The greater the available knowledge and knowhow in a country (economists would call this human capital), the better and more complex the range of products, and hence the more advanced a country will be. The export structure also tells you something else. Countries that are well off but only export raw materials which they have due to their access to abundant natural resources are not economically complex. They are actually quite fragile in their structure, as the first sign of a resource price shock, their economies end up in shambles (there is no better contemporary example for this than Venezuela in the past two years; Taleb’s favorite example of this fragility based on natural resources is Saudi Arabia – if the world switches from oil to renewable energies, Saudi Arabia is destined to follow in Venezuela’s footsteps).

Hidalgo shows that the resource curse is due to a lack of imagination (knowledge) to use the resources at their disposal in better fashion. Resource curse leads to a rent-seeking society, which works well as long as the people can all be well-off from the money flowing from the resource abundance. Smarter countries (e.g. Qatar, UAE, Norway) will use the opportunity to invest the money from the natural resources into something else, that will ensure their high living standards even after they drain their resource. This way they improve their knowledge and knowhow, they build their imagination to design more complex products and thus more advanced economies.
However information growth will be prevented if people lack the opportunity to document knowledge and knowhow. If you have a brilliant, life-altering idea, it really matters where you will have the chance to develop it. This is where institutions come in. In many countries having the proper opportunity to even gain knowledge not to mention to use it, is unfortunately unimaginable. Poor institutions and poor governance of such countries makes it hard for people within them to improve them. It is a reinforcing negative cycle. Breaking it up from within is immensely difficult, and breaking them up from the outside is often unfeasible. Going back to thermodynamics, low-income countries are stuck at the bottom of a thermodynamic potential, and unfortunately the only way to break this up is purely random. And the chances of the right combination of factors to occur for a country to break out of a low-income equilibrium are decreasing with the size of the corrupt network within a country. This leads to a very pessimistic conclusion that some countries are doomed (this is the pessimistic conclusion of my paper).

But I digress. Hidalgo also discusses all the possible limitations of expanding knowledge and knowhow. These include intrinsic individual capabilities which then translate to geographically-biased accumulation of knowledge (as individuals spread their knowledge within their own social groups, narrow and broad). This is important to comprehend since complex products (like a car, machine, airplane, or computer), in order to be created, require much more knowledge than one person can have. Hidalgo defines the individual capacity of a single human as a personbyte. Complex products involve a lot of personbytes. In addition to personbytes he also defines a firmbyte – how knowledge and knowhow are distributed across a network of firms. Modern manufacturing is a global process – it involves the coexistence of many networks through which exchange takes place and a product is created (in the modern supply chain economy there is not a single complex product that is being produced in one country, not to mention one place, alone).

Firms, like humans, also have their limitations. Many fail to adapt and go bust. Others succeed if they uncover cheap links to expand their networks in order to accumulate more knowledge and knowhow. For example governments can be both helpful and burdensome in maintaining the costs of such links. They can introduce helpful standards and the proper infrastructure, but they can also impose unnecessary bureaucratic procedures and overreaching regulations that will make links and networks more expensive. This undermines the capacity of economies to accumulate knowledge and put it to good use in producing complex products.

Another important category that affects the accumulation of knowledge is a social network. In particular, culture and trust. In fact the level of trust in asociety will influence the organizational structure of its economy. Hidalgo draws this from Fukuyama who emphasizes the differences in low-trusting “familial” societies (you can only trust your family, not strangers) which tend to be dominated by many small family businesses and a few dominant family conglomerates, and high-trusting societies where professionally-run businesses give rise to large networks of firms and consequentially more complex industries and their products (pharmaceuticals, aircraft, high-tech, etc.).

Therefore, producing complex products implies a lot of personbytes and a lot of firmbytes who get assembled into large networks (which is to a large extent determined by a country’s social, or informal institutions), which consequentially implies a huge accumulation of knowledge at one place. The more complex products you tend to produce, the more knowledge you will accumulate, and vice versa. This is the example of a positive feedback loop. Not to mention that accumulating a lot of knowledge and knowhow at one single place constitutes a comparative advantage of a given country. It is now becoming clearer how this theory explains the development conundrum.

As knowledge and knowhow get geographically clustered there should be a way to visualize this. And here is where Hidalgo does his best. Along with several colleagues, including the famous Ricardo Hausmann from Harvard, they use trade data to show how large networks of accumulated knowledge grow by mapping them into something called a product space (pictured below) (all the visuals are really good, play with them here).

The product space is a network connecting similar products. The closer the products the more similar they are, and the more likely that they will be co-produced and co-exported from one country. For example, a country exporting knit t-shirts is also likely to export other garments and undergarments. Or if you export beef you’re also likely to export cheese. Car engines and motor engines. And so on.

The product space shows that while some industries, and hence products (like clothes) are present everywhere, others are limited to only a few places (like cars, airplanes, or computers). Obviously, the greater the distribution of such industries the less complex they are and the less knowledge they require. The opposite goes for the highly clustered (nested) industries which tend to produce the most complex products. It is therefore easy to see how developed, richer countries, with much larger networks, are able to attract and accumulate the necessary knowledge and knowhow to produce the most complex products. This further ensures their dominance in the global trade and specialization patterns and preserves their position of wealth (e.g. Germany, Japan, US, etc). As Hidalgo says, moving large networks from one country to another is almost inconceivable. Poor countries on the other hand are thus limited to producing only a small subset of products which don’t require a lot of information nor a large network to be produced. This basically means that economic growth is path-dependent. A country cannot just jump from growing crops to producing heavy machinery. It takes time to develop complexity, as countries can only move to products close to the initial ones on its product space.

The problem (and the only criticism I have) however is that the theory doesn’t provide a historical account as to why knowledge accumulated in one place, and not the other. We know why countries differ, this part the theory explains quite well, but we cannot explain its initial divergence – the initial critical point that led some countries to become rich, knowledge-accumulating large networks capable of producing the most complex products, and descended others to a low-equilibrium self-reinforcing negative cycle with a lack of imagination and the capacity to accumulate knowledge at all. On the other hand, I don’t think the author had this ambitious task in mind. Not quite yet I believe.

The answer here probably lies in the development of institutional patterns from the Industrial Revolution onward, and perhaps some historical antecedents we carried from as long as we were able to plow fields. But more on that in the next few books. 

No comments:

Post a Comment