Other papers (about 5 minutes)

Other papers (about 5 minutes)

Two chunks for the last five minutes. About two and a half minutes each. Do not call either paper a job-market paper.

If the clock is short, do chunk 1 through the 20 percent result, then chunk 2 through the price result, and stop.


Chunk 1 — Unsold books (about 2.5 minutes)

The next paper is with Kohei Kawaguchi and Kei Kawai. It is still in progress. The title is “Too Much Unsold Stock?: Environmental Consequences of Price Commitment.” The industry is Japanese book retail under resale price maintenance.

The publisher sets the retail price before demand is known. The store then chooses how many copies to hold. If demand is high, the title stocks out. If demand is low, the store cannot cut the price, so unsold copies go back upstream and are discarded. About 35 to 40 percent of Japanese books printed are discarded. The data are monthly point-of-sale records for more than 1,000 stores: deliveries, sales, and returns, linked by ISBN.

The same rule does two things. A price that cannot fall gives the store a reason to hold inventory, so shoppers still find the book when demand is low. That is the benefit. The cost is the waste when demand disappoints. We want both, which is why a static model is not enough.

What I will defend today is not the unfinished counterfactual. Stockouts do not send shoppers to another region. They switch to another title, or they wait and buy later. And a static model that ignores inventory and demand uncertainty overstates marginal cost by about 20 percent. Any account of the environmental cost that starts from that static cost will be off. Whether a deposit-refund tax can cut the waste without giving up the inventory is still open. I will not quote a number for it.

Chunk 2 — Search (about 2.5 minutes)

The other paper is my M.Phil. thesis with Pak Hung Au, “Product Design with Heterogeneous Search Costs.” This one is about uncertainty before the purchase, not after.

People search sequentially, as in Wolinsky (1986), but they do not all face the same search cost. Search costs decide not only whether someone searches, but which product they search. In equilibrium they sort. Shoppers with low search costs look for niche products. Shoppers with high search costs look for standardized products. Firms use product design to pull those groups apart.

The result that matters is what happens when search gets cheaper. Cheaper search has the usual force: people search more, and that pushes prices down. It also changes which product people search. If that shift in who looks at which product is strong enough, it dominates the search-encouraging effect, and both prices can rise. A fall in search costs is not automatically pro-competitive.

Where the cost falls in the distribution matters, and the two ends do not move together. A technology that helps people who already had low search costs tends to raise prices. One that helps people who found search expensive can lower them. I would not summarize the paper as “the internet makes markets more competitive.” It depends on whose search got cheaper.

That is the same theme as the return-policy paper, on the other side of the purchase. There, the costly adjustment is a return. Here, the costly adjustment is another search. In both, a change that looks like it only helps shoppers can move the price the other way.