Gorodnichenkoの共著論文をもう一丁。以下はYuriy Gorodnichenko(UCバークレー)、Oleksandr Talavera(バーミンガム大)、Nam Hoai Vu(同)によるNBER論文「Quality and Price Setting of High-Tech Goods」(ungated版)の要旨。

This paper investigates the link between product quality and price setting for central processing units (CPUs). Using thousands of price quotes from a popular price-comparison website, we find that market fundamentals, such as the number of sellers, median price, share of convenient prices and level of seller stability, are important factors for explaining price stickiness and price dispersion. We demonstrate that calculations of price inflation require conditioning not only on CPU quality, but also on market fundamentals to ensure that CPU attributes are priced correctly. Failing to do so can result in an understatement of CPU price deflation in the sample period.


Quality improvement has long been a major challenge for measuring inflation, especially for goods with rapid technological advancements. Hedonic regression is a popular approach for ensuring the correct prices for upgrades of product attribute trends (see e.g., Gordon and Griliches, 1997; Shiratsuka, 1999; Bils, 2009). However, this approach ignores the fact that prices are also affected by other factors, such as market and product characteristics.
We follow the common approach in the existing literature (see e.g., Byrne et al., 2018; Aizcorbe et al., 2020) and run the hedonic regression for each pair of consecutive years in our sample. More specifically, we include product characteristics and market fundamentals as control variables. Our focus is on the quality-adjusted price movement which is reflected by the coefficient of the year dummy for each time period. We show that hedonic adjustment may be incomplete if pricing of product attributes varies across products. Thus, one should control, not only for product attributes, but also for market fundamentals.
品質の向上はインフレを測定する上で長年の間大きな課題になっており、技術進歩が急速な商品については特にそうである。ヘドニック回帰は製品の特性トレンドのアップグレードについて正しい価格を保証する上で人気のある手法である(例えばGordon and Griliches, 1997; Shiratsuka, 1999; Bils, 2009を参照)。しかし、この方法は、市場と製品の属性のような他の要因にも価格が左右されることを無視している。
我々は既存の研究の一般的な手法(例えばByrne et al., 2018; Aizcorbe et al., 2020を参照)を踏襲し、我々のサンプル期間の連続した年の各ペアについてヘドニック回帰を行った。より具体的に言うならば、製品の属性と市場ファンダメンタルズをコントロール変数として含めた。我々は、各期間の年ダミーの係数に反映される品質調整済みの価格変動に焦点を当てた。我々は、製品属性の値付けが製品間でばらついている場合、ヘドニック調整が不完全であろうことを示す。従って、製品の属性だけでなく市場ファンダメンタルズもコントロールする必要がある。

*1:ungated版の本文では「convenient prices (also called price points such $9.99 or $99)」と書かれており、price pointはwikipediaで「Price points are prices at which demand for a given product is supposed to stay relatively high.」と説明されている。