Chapter 3

Application to the Chinese Automobile Market

Estimating the spectacle premium in China versus a European reference market using the extended BLP framework

Back to research overview
🚧

Data not yet collected. This interface is a preview. Automobile sales data, vehicle characteristics, and advertising content have not been assembled yet. Data collection and estimation are planned for Year 3 of the doctoral programme (2028–2029).

Objective

Apply the extended BLP model developed in Chapter 2 to the Chinese automobile market and compare results with a reference market (France or Germany). The automobile sector is chosen for four reasons: data availability at the model level, detailed price and characteristics data, the centrality of brand narratives in automotive marketing, and the existence of a well-established BLP literature on automobiles since the original 1995 paper.

Data sources

CN

China — Sales by model from best-selling-cars.com; brand-level data from CAAM (China Association of Automobile Manufacturers); technical specifications and prices from Autohome (汽车之家); advertising data from IAI Yearbook.

FR

France (reference) — New car sales data by model; vehicle specifications from manufacturer catalogues; advertising content from publicly available campaigns.

AD

Advertising coding — For each model, advertisements will be classified by the ratio of narrative content (lifestyle, celebrity, brand story) to objective content (specs, performance, price). NLP tools may assist the coding process.

Core empirical test

The chapter estimates the spectacle coefficient γ on both markets and tests whether the cross-country difference is statistically significant:

H₀: γChina = γFrance
H₁: γChina < γFrance

A lower γ in China implies that Chinese consumers are less sensitive to the spectacle premium, consistent with lower image-driven consumption and higher savings rates.

Cross-Country Comparison Interface

Run the extended BLP on automobile data for China and a reference market. Currently inactive — awaiting data assembly.

Expected results

Estimated γ for China and the reference market, with confidence intervals and a formal test of the cross-country difference. Counterfactual simulations showing how Chinese automobile demand and implied savings would change if spectacle sensitivity were at European levels. These results will quantify the "spectacle premium" gap and its contribution to the savings puzzle.

Timeline

This applied chapter is the final piece of the dissertation, planned for Year 3 (2028–2029). It depends on the successful construction of the spectacle index in Chapter 2 and the availability of sufficiently granular automobile data for both markets.