Price informativeness and institutional ownership: evidence from Japan

See Mehran (1995), Denis et al. (1997), Ang et al. (2000), and Cremers and Nair (2005).

According to the literature on WPC, it makes little difference whether returns or price changes are used in the definition of WPC. For studies that use price changes as the case in this paper, see Huang (2002). In our analysis, results are qualitatively the same if returns are used in the definition of WPC.

However, empirical studies in the existing literature find that β is often less than one. Barclay and Hendershott (2003) argue that this can be a result of the errors-in-variables problem due to bid-ask spreads or pricing errors. If the observed price are contaminated by noise, regressing r _cc,t on r _co,t using ordinary least squares could produce an estimator which is biased downward as discussed below.

Assuming that the returns are measured with errors such that r _cc,t  = R _cc,t  + ν _1,t and r _co,t  = R _co,t  + ν _2,t , where R _cc,t and R _co,t stand for the true return, and ν _1,t and ν _2,t have zero means with variances σ _v1,t ² and σ _v2,t ² , respectively, then the biased β estimate should satisfy the following relation:\( p\lim \, \widehat{\beta } = \beta (\frac{{\sigma_{{R_{co,t} }}^{2} }}{{\sigma_{{R_{co,t} }}^{2} + \sigma_{{v_{2,t} }}^{2} }}) \) where σ _Rco,t ² denotes the information released during the close-to-open time slot, and σ _v2,T ² captures the noise in prices attributable to the non-trading period. The estimate \( \widehat{\beta } \) thus allows us to infer the signal-to-noise ratio.

Details about the sources of all variables are provided in Panel A of Table 1.

According to the Nikkei Net Interactive, the Nikkei reviews the component stocks of the NIKKEI 500 annually, usually at the end of March. Our sample is based on the component stocks of the NIKKEI 500 updated in 2004 which have a full-year quote and trade data. For information of the component stocks, see http://​www.​nni.​nikkei.​co.​jp/​CF/​FR/​MKJ/​list_​nikkei_​constituents.​cfm.

There is a lunch break between 11:00 A.M. and 12:30 P.M. on the Tokyo Stock Exchange (TSE).

Even though the Tokyo Stock Exchange differs from the US exchanges in the sense that the former is driven by orders without any market makers while the latter are quote-driven, Lee and Ready (1991)’s algorithm can still be applied to Japan in view of Ahn et al. (2001)’s study which applies the same rule to Hong Kong’s Stock Exchange which is also an order-driven market.

Three firms’ institutional ownership data is missing when the intraday transaction data is matched with the institutional ownership data.

The test results are not reported, but are available upon request.

Definitions of these control variables are provided in Panel A of Table 1.