The Interconnectivity of Production Networks

November 06, 2019

This 8-minute podcast was released Nov. 6, 2019.

Sungki Hong in studio with Christine Smith | St. Louis Fed

“Traditionally, economists have seen the economy as a whole body, as one. But if we look into the details underlying the economy, we see a very complicated production network, that each industry’s production is actually relying on other industries’ output,” says Sungki Hong, an economist at the Federal Reserve Bank of St. Louis. He talks with Christine Smith, digital content editor, about hub industries, economic downturns and why he’s curious about the technology sector today.


Christine Smith: Welcome to Timely Topics, a podcast from the Federal Reserve Bank of St. Louis. My name’s Christine Smith, and I will be your host for this episode. So joining us in the studio today is Sungki Hong, who is a research economist here at the St. Louis Fed. Thanks for being here, Sungki.

Sungki Hong: Thanks a lot for your invitation.

Smith: So you’ve done a lot of work looking at how industries, especially in advanced economies like that here in the U.S., are “interconnected.” So tell me about some of your research in this area. And what makes it such an interesting subject to study?

Hong: Traditionally, economists have seen the economy as a whole body, as one. But, if we look into the details underlying the economy, we see a very complicated production network, that each industry’s production is actually relying on other industries’ output.

Smith: So why might it be important for economists to think about the interconnectivity or the linkages between different sectors?

Hong: It’s because, in the production network, there are actually some “hub” industries that are highly interconnected with the rest of the economy. One example is the auto industry. To assemble a car, it needs to purchase a large amount of materials from other industries. And, at the same time, it can sell its cars to other downstream industries, like the transportation and wholesale and retail industries.

So, if the auto industry—whether its industry goes up or down, it would highly affect the industries in the upstream and also in the downstream.

Smith: So you mentioned some terms there. Some people may be familiar with them. Others may not. So, first you talked about “hub” industries and gave the auto industry as an example of that. So, that’s one that would be highly interconnected. Using the auto industry as an example, you mentioned there are some “upstream” connections and some “downstream” connections.

Hong: Yeah.

Smith: So, kind of refresh my memory on what those are again.

Hong: So, for the auto industry, the upstream industry could be the leather industry or the metal industry, that the auto industry could buy leather and metal from these two industries to assemble the car. So, for the auto industry, the leather and the metal industry would be an upstream industry.

Downstream industry is the industry that the auto industry can sell its products to. So, the examples are like transportation.

Smith: You’ve done a lot of analysis on production networks, kind of the web of product flows between firms and industries. In one of your papers, you dug into 30 different industries, from agriculture and mining, to construction, to oil and gas extraction. Which industries did you find were important hub industries?

Hong: So we used the data for 30 different industries to see what is the output for all the goods flow between the industries. So what is Industry A purchasing? If it’s purchasing, from which industry is it purchasing from? So we can look at the size of the flow between the industries, and from this dataset we can see what are the important suppliers. At the same time, what are the important buyers in the production networks?

Based on our analysis, we used the data that the material flows between—among the industries. So material flow manages how much one industry is buying from the rest of the industries, and it also measures how the rest of the economy is buying the materials from that one specific industry.

And, based on the analysis, we find a couple of hub industries based on the importance as an important supplier and buyer in the economy. And we find that the important hub industries are the warehousing industry, wholesale and retail, construction industry, motor vehicles, and other services.

Smith: OK. So those would be considered hub industries in the economy. Can identifying those key suppliers and buyers—those hub industries—help economists to, let’s say, analyze an economic downturn? Right? So, if we see trouble in a hub industry, does that then indicate that there might be a spillover effect to other industries that are closely connected with it?

Hong: Yes, definitely. So, in one of the studies that we have done, we actually studied the construction industry, which is one of the hub industries that I mentioned. So we know that during the recent financial crisis in 2008-2009 that the house prices has dropped a lot. So that has cost the plumbing and the construction industry.

We want to look at, like, what is the spillover effect from the construction industry to the rest of the economy? So we have ranked the industries that are most connected to the construction industry and also the industries that are least connected to the industries.

And we looked at the gross output of the most-connected industries versus the output of the least-connected industries.

So, based on the data, we see that in 2008-2009, the gross output or the total revenues of the construction industry has decreased by 13%. What has happened to the top 10 industries that are closely connected to the construction industry? We see that these mostly connected industries, their gross output has dropped by 15%.

How about the least connected industries? We see that their output has decreased, but it has only decreased by 5%, compared to the 15% of the most connected industries.

Smith: So the gross output noticeably decreased for those industries which were more closely tied or connected to the construction industry, then, after the Great Recession or amid the Great Recession.

So, can you tell me, what are some examples of the industries that were most closely connected to construction?

Hong: Oh, so some examples would be like the lumber industry, the furniture industry, and also the “FIRE” industry, which are the finance, insurance, and real estate industry.

Smith: So, amid the Great Recession then, the industries that were most closely connected to the construction industry, like did you see difference in how they got hit?

Hong: So, in terms of timing, I think like, whether it’s like a most-related or least-related industry to the construction, all these industries take the hit at the same time. But they are different in terms of the magnitude of the hit that they take.

So, the top 10 industry that closely related to construction, their gross output or their total revenue have dropped by 15%, compared to the least—that the top 10 least-related industries to the construction, their total revenue only dropped by around 5% during the financial crisis.

Smith: So, that was during the Great Recession then, right? The construction industry didn’t do so hot, and then closely connected industries also didn’t do so hot as far as measuring gross output goes.

Are there any industries that you’re keeping an eye on now, interesting hub industries that you’re watching?

Hong: So, one particular industry that I’m interested in is the tech industry. So, even though the total size of the tech industry is small compared to other, traditional industries that we have talked about, but we know that it’s growing with a very fast pace, like the “FANG” companies: like the Facebook, Amazon, Netflix and Google. They’re growing really fast.

But, as an economist or the traditional economist, we are still not quite sure how these tech industries are connected to the traditional sectors, like the auto industry or the wholesale sectors. So I’m very interested in looking at the interaction between the tech and the rest of the economy and see how that interaction will play out.

Smith: So it sounds like part of the fun or part of the challenge is first identifying within the tech industry, who is it mostly closely connected to? And then after that, looking at the sector itself and starting to find some economic benchmarks to determine how it’s doing?

Hong: That’s correct. That’s correct.

Smith: Sungki, thanks for talking with us today about production networks.

Hong: Thank you for inviting me.

Smith: Our listeners can find more of Sungki’s research on production networks and other subjects at And for more podcasts, you can visit You can also stream this series on Apple Podcasts, Spotify and Stitcher, or ask your Amazon device, “Alexa, play Timely Topics from TuneIn.”

Related Topics

Economists and experts talk about their research, topics in the news and issues related to the Fed. Views expressed are not necessarily those of the St. Louis Fed or the Federal Reserve System.

Back to Top