Please excuse this very very wonky post. I’m going to be thinking out loud here about import penetration, and working out an example.
Here’s the starting question: When consumers start buying and retail sales goes up, how much of that added purchasing power stays in the United States and how much goes overseas as imports?
The answer is not obvious. But even the right question is not obvious either.
Let’s look at a simple example. Suppose that American consumers are buying 10,000 dining room tables annually from American furniture factories. Each dining room table leaves the factory door at a producer cost of $1000. Along the way, the costs of trucking and distribution add another $1000, so the price to the consumer is $2000 per table.
Now let’s suppose that after a few years, half the dining room tables sold in the U.S. are made in China (5000 out of 10000). The reason why? Chinese factories can produce and deliver the tables cheaper—it costs $700 per table to make the table, ship it across the Pacific, and put it on a truck at the Port of Los Angeles. After that, the same $1000 in domestic trucking and distribution costs apply, so that the table is sold for $1700.
Question: How much of the U.S. market have the imported tables taken? (This is another way of asking: If retail sales of tables goes up, how much of the extra spending will leak abroad).
e) all of the above
The answer, of course, is all of the above. Let me explain each of them. The first number, 50%, just comes from just counting tables. If half the dining tables bought in the U.S. are imported, then in some intuitive sense imports have 50% of the market.
The second number, 19%, is the cost of imported tables as a share of total consumer spending on tables.
Cost of imported tables = 5000*$700 =$3,500,000
Total consumer spending on tables = 5000*$1700 + 5000*$2000 = $18,500,000
19% ≈ 3,500,000/18,500,000
For the whole economy, this figure is precisely analogous to calculating imports as a share of gross domestic purchases. (In the fourth quarter of 2009, this number for the whole economy was 14.2%). 19% is much lower than 50% because the consumer dollar pays for a lot of domestic trucking, warehouses, wholesaling and retail employment and capital investment, which is there whether or not the table is imported.
The third number, 46%, represents the retail value of imported tables as a share of total consumer spending on tables. It’s lower than 50% because the imported tables cost less. If the retailers charge the same price for imported tables as domestic ones (allowing a higher profit margin on imported tables), then this number could be 50% also.
The fourth number, 41%, represents the cost of imported tables as a share of the production/acquisition costs of tables
Cost of producing domestic tables = 5000 *$1000 = $5,000,000
Cost of imported tables = 5000*$700 =$3,500,000
Production/acquistion costs of tables = $8,500,000
41% ≈ 3,500,000/8,500,000
Several questions here. First, which of these numbers can be calculated using government data? (a) requires us to know the relative production/acquisition price of imported versus domestic tables. This data does not exist (see, for example, Bill Alterman’s presentation at the “Measurement Issues Arising from the Growth of Globalization” Conference).
(b) can be calculated using existing data. (c) cannot be calculated using existing data, because the government does not distinguish between imports and domestic products once the imports have entered the country–imports are assumed to be indistinguishable from the domestic products that they compete with. (d) in theory can be calculated using Census data on individual industries to measure shipments, and then link that with the import and export data. In practice, however, we don’t know whether the imports are competing with domestic products, or used in the production of the domestic product, which makes a big difference.
To get back to the original question. How much of retail sales flows through to imports? I can use the BEA’s input-out tables to take a stab at (b). And by making some assumptions about the relative price of imports and domestic goods, I can make a stab at (a). Here are my sentences about the calculations. I would appreciate anyone who cite these numbers quote whole sentences.
**My preliminary and tentative calculations–which I reserve the right to revise or completely repudiate at any moment–suggest that roughly 15% of every retail spending dollar flows overseas. However, my preliminary and tentative calculations also suggest that 25-30% of the goods we buy are made overseas. **
The first number corresponds to concept (b), and the second range corresponds to concept (a). They answer different questions. (b) answers the question about increased retail spending and overall job creation. For some products–notably consumer electronics–it may be that nearly 100% of the products we buy are made overseas, but the cost of the good only accounts for 60% of the final price. So even complete penetration means that 40% of the consumer spending dollar on a big-screen television still goes to support retail clerks and truckers.
On the other hand, (a) answers the question about retail spending and manufacturing production. How much does an extra retail dollar go to get factories running again…if we spend more on big screen televisions, it has very little direct impact on manufacturing jobs in the U.S.
I thank any of you kind people who got down to the bottom of this long and baroque post. I’m getting closer to understanding what’s going on.