# Bandwidth, latency and price

Can a market be seen as a set of nodes (market actors) connected by edges with bandwidth and latency characteristics?

Are these terms used in economics?

If so would an import tariff affect bandwidth and/or latency?

• Suppose you have a small economy with only two actors. What would the bandwidth and latency of the edge connecting them express? I suppose the answer to the second question is no. – Bayesian Jun 20 '19 at 10:33
• Two fishermen, F1 and F2 on a beach. The beach is divided into two by an outcrop that blocks the view from F1 to F2. F1’s fish are priced the same as yesterday. F2’s fish are cheaper today. A buyer, B1, wants to buy a fish and he approaches F1. As far as he knows the best price of fish is that of F1 because the bandwidth of the edge between B1 and F2 is very low (the outcrop is in the way and he can’t see the cheaper prices). Latency would be a different scenario where instead of volume of data, the timeliness of data would impact B1’s decision making. – 52d6c6af Jun 20 '19 at 10:47
• Such situations are modeled differently in (mainstream) economics. For instance, with search costs, information cascades or experimenation. I am not sure how much your approach would add, but it certainly is not the go-to model. – Bayesian Jun 20 '19 at 11:17

There is an active body of research on network economics. See, for example,

for an introduction to the topic.

These models occasionally use weighted edges ("bandwidth") to model economic phenomena.

I'm not aware of work making use of the concept of latency, except when literally modelling latency (as in models of network congestion). An example of such a paper (though without the notion of nodes or edges) is Net neutrality and investment incentives by Choi and Kim.

I’m an ex-electrical engineer, so I’m familiar with the concept of bandwidth. Within academic economics, I’ve never seen bandwidth used in standard economics contexts. (Latency is a time delay, and shows up. However, time lags in this context are on the order of months, and are obviously not the result of delays due to communications, rather decision-making.)

(Update: see the answer by Ubiquitous for some references. I am unfamiliar with them, but based on the description, it’s an analogy to bandwidth, and not the formal electrical engineering bandwidth concept, related to the limit for data transmission down a pipe.)

In finance, high frequency trading is literally a set of actors sending high frequency trading signals at each other. They are very concerned about these concepts. However, most people probably think of that as finance not economics, but the line between finance and economics is blurry at times.

There is the new field of agent-based models, which are a set of connected actors (nodes). Bandwidth is not really an issue in economics context, since the effective minimum time step is about one working day - it is extremely rare for people to change jobs within a single working day, so production (which needs workers) is a slow-moving process.

Outside of academia, there are non-economists who write about the subject (on blogs, etc., I am technically one of those people). One of them may have tried to work the concept in.

As for your example of a tariff, it shows the limited usefulness of those concepts. The decision to import a good versus buying locally is already a complex decision, and there is already paperwork, etc. The addition of a tariff adds just one more thing to be taken into account. The increment of time associated with the increase in decision complexity is very small when compared to the shipping time of goods.

• Assume tariff-rate quotas are changed slowly and infrequently because of organisational inertia and bureaucracy (presumably this is actually true?!). Assume that a good g has a tariff t applied to it on day d1. A market change (new tech/whatever) means that t is suddenly wholly inappropriate. Finally, months later ond2 the tariff-rate quota is modified. The tariff has both degraded the bandwidth and increased the latency of the price signal. – 52d6c6af Jun 20 '19 at 11:08
• Retailers typically order goods for the Christmas shopping season sometime in the summer, i.e., months in advance. The extra time added for the decision is at most on the order of seconds, since it is just changing the prices in the decision. The only time it might be of concern is where tariffs are first imposed, since procedures were not put in place to deal with tariff administration already. It’s not a concept that adds value to analysis here. – Brian Romanchuk Jun 20 '19 at 11:19
• In my scenario, for duration d2 - d1 consumers inside the tariff wall see a price that is potentially wildly distorted? – 52d6c6af Jun 20 '19 at 11:24
• That’s not how tariffs normally work; they are a fixed percentage of the price. Prices would be set based on existing and expected tariff values, so there’s no reason to argue that the prices between tariff changes are “distorted” beyond the effect of tariffs themselves. Furthermore, this has nothing to do with finite information transfer capacity. This is a forum to ask questions, not get peer review of new theories of economics. – Brian Romanchuk Jun 20 '19 at 11:38
• Haha 😆 I had no idea this was a new theory. I come from computing and was simply inquisitive. Tariff schedules often contain quotas. The quota limit could be triggered ond1 and prohibit the import of g until d2. – 52d6c6af Jun 20 '19 at 11:44

By definition, a market is just a set of agents endowed with a preference relation over a set of goods. In some fashion, agents are nodes to trades; that is, each agent is a base on which a trade can occur.

If you look at things this way, then yes, a market may be seen as a set of nodes. I've seen some marketing papers that apply the term "capillarity" to refer to the bandwidth that an agent can affect other agents, much more like a biology term than an engineering one.

These are unusual terms in economics literature, but they may be applied if you base yourself on the logical definition of the terms. Also, yes, taxes are a bandwidth reducing parameter, since they affect the price vector of the market and the supply and demand of a certain good. Since they reduce the total number of agents (or nodes) included in transactions, they reduce the bandwidth of the trades.

The concepts are used; the language usually is not. Nodes come from graph theory, so you will see nodes and edges. Latency is generally discussed using the term of art "lags." Bandwidth isn't used as a term, but the concept has long been present in the distinction between stock and flows. It also is covered in discussions of congestion and capacity.

Tariffs impact both information and capacity.

The information content is delayed because many contracts are fixed contracts. Because the idea of a US trade war was unthinkable, the increase in tariffs was left out of deals. New agreements will incorporate the added duties when they expire. It will also begin adding in a tariff risk fee as well.

To a lesser extent, that would also be true for Brexit, but the lag time between approval and implementation has allowed most contracts to expire and readjust. The information will not be reported contemporaneously, though. Some things, such as income taxes, have a long lag between incidence and payment. It is the payment that is of consequence.

Secondarily, the tariffs move the equilibrium, but no one knows in advance where that will be. A process of tatonnement has to happen to incorporate the impact of the added import duties.

A tariff shifts the supply curve, reducing the global capacity to deliver goods and services, other things constant. China, for example, has retaliated against the United States by lowering the tariffs of every country except the United States. In doing so, Trump won the trade war, for the rest of the world, and is losing it for the United States.

The lowering of tariffs is a signal of friendliness. It is a way for China to say, "we are not mean like the Americans, come trade with us instead." On its own, it increases the long-run capacity of other nations and the purchasing power of China's citizens. In the short-run, it is of little consequence.

There are certain capacity issues that will be permanent, even if Trump ended the trade war today. The United States threatened the food safety of China. It is building farms. Those farms will permanently replace U.S. farmers and add planetary capacity, permanently reducing food prices unless US farms declare bankruptcy and the land is reused for other purposes or there is enough of a population increase.