AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
The non-rival nature of information and the resulting price complementarity naturally create the conditions for information herds and frenzies to arise

Media Frenzies in Markets for Financial Information

AMERICAN ECONOMIC REVIEW, no. 3 (2006): 577-601

被引用271|浏览22
下载 PDF 全文
引用
微博一下
关键词

摘要

Emerging equity markets witness occasional surges in prices (frenzies) and cross-market price dispersion (herds), accompanied by abundant media coverage. An information market complementarity can explain these anomalies. Because information has high fixed costs, high volume makes it inexpensive. Low prices induce investors to buy informat...更多

代码

数据

0
简介
  • The search for the source of frenzies and herds is a search for complementarities in asset demand: How can one person buying an asset make the asset more attractive to other investors? Perhaps the complementarities lie not in the demand for assets, but in the demand for information used to price the asset.
  • There are large shifts in information demand as investors coordinate on a new asset to learn about
  • Such a shift in information suddenly raises an asset’s price above what a model without information would predict and generates excess price volatility.
  • The added unconditional variance prompts a switch to the positive-information equilibrium
  • Because it reduces conditional variance, abundant information raises the asset’s price above what the improvement in fundamentals would predict
重点内容
  • The search for the source of frenzies and herds is a search for complementarities in asset demand: How can one person buying an asset make the asset more attractive to other investors? Perhaps the complementarities lie not in the demand for assets, but in the demand for information used to price the asset
  • Why look to information markets as the source of frenzies and herds, rather than to asset markets themselves? The reason lies in the non-rival nature of information
  • Because information reduces the conditional risk of the asset payoff, it increases the asset’s price
  • Media frenzies are an abundance of information provided to investors about an asset or class of assets by a competitive market for information
  • The non-rival nature of information and the resulting price complementarity naturally create the conditions for information herds and frenzies to arise
  • The resulting price path looks like herds of investors stampeding from one market to the
结果
  • Because information reduces the conditional risk of the asset payoff, it increases the asset’s price.
  • Changes in the amount of information sold to investors increases price volatility.
  • Information is demanded only when asset payoffs and prices are high.
  • Provision of information pushes already high prices up higher.
  • Movements in extreme price levels have large consequences for the unconditional price variance.
  • Changes in information provision are large.
  • A switch from a no-information equilibrium to a positive information equilibrium can increase the fraction of informed agents by 80% in a single period
结论
  • Media frenzies are an abundance of information provided to investors about an asset or class of assets by a competitive market for information.
  • The non-rival nature of information and the resulting price complementarity naturally create the conditions for information herds and frenzies to arise.
  • Media frenzies raise asset prices by reducing the uncertainty about the asset’s payoff.
  • In multiple markets with trade-offs in information demand, a media frenzy in one market raises the market price and increases the cross-sectional price variance.
  • The resulting price path looks like herds of investors stampeding from one market to the
表格
  • Table1: Standard deviations of simulated variables and log price indices for 23 emerging markets
  • Table2: News and payoff volatility by country. Cross-section results from regressing a market’s average number of news stories on the standard deviation of its asset payoffs. All tables show standard errors in parentheses
  • Table3: News and payoff volatility over time. Panel estimation results from regressing number of news stories on asset payoff (R) volatility and country fixed-effects. Price indices are divided by their country mean. Seemingly unrelated regression estimation allows for contemporaneous correlation across countries and heteroscedasticity. Instrumental variables are the news volatility variable, dividends, in levels, changes and squared changes, and 8 lags of the volatility variable. (23 countries, 12,194 obs.)
  • Table4: Granger test of news exogeneity with respect to asset payoff volatility. Test includes 25 lags of each variable
  • Table5: News and Price Level. Time-series results from regressing price, or price-dividend ratio, on the number of news stories and country fixed-effects. Seemingly unrelated regression estimation allows for contemporaneous correlation across countries and heteroscedasticity. (12,194 observations)
  • Table6: Panel estimation results from simultaneously regressing price on the number of news stories, and news on payoff volatility. Payoff volatility is instrumented with contemporaneous and 1-period lagged dividend levels and dividend volatility, and one lag of payoff volatility. All equations include country fixed effects. Three stage least squares estimation allows for heteroscedasticity and contemporaneous correlation across equations and countries. (12,194 observations for each equation)
  • Table7: Test of the herding hypothesis: Price dispersion increases with news. Price dispersion is a time series of the standard deviation of prices across the 23 emerging markets. News is the total number of news stories for all emerging markets per week. Price dispersion increases when news is abundant. (704 observations)
  • Table8: Descriptive statistics for data set
  • Table9: Panel estimation results from regressing number of news stories on asset price (P) or payoff (R) volatility and country fixed-effects. Price indices are divided by their country mean. Seemingly unrelated regression estimation allows for contemporaneous correlation across countries and heteroscedasticity. Controlling for news variance isolates the effect of volatility on news. Instrumental variables are the news volatility variable, dividends, in levels, changes and squared changes, and 8 lags of the volatility variable
  • Table10: Prices increase when news is abundant, controlling for payoff volatility. Row 1 is a cross-sectional regression (704 observations). Rows 2 and 3 use the panel dimension to estimate a seemingly unrelated regressions system (12,194 observations).Price is the average price index for all countries in the sample, each week. News is the average number of news stories for all emerging markets per week. OLS estimation. All estimations include a constant. White standard errors in parentheses
  • Table11: Panel estimation results from simultaneously regressing price on the number of news stories, controlling for payoff volatility, and news on payoff volatility, controlling for news volatility. All equations include country fixed effects. Seemingly unrelated regression estimation allows for heteroscedasticity and contemporaneous correlation across equations and countries. (12,194 observations for each equation)
Download tables as Excel
基金
  • Simulations (10,000 iterations) show that an endogenously varying information price causes the asset price volatility to be 40% higher than in a fixed-information-price setting: vvaarr((PPGin−fSo)) = 1.4
研究对象与分析
observations: 12217
Data To assess the magnitude of the model’s frenzies and test the model’s predictions requires data on equity markets and financial news. The data is a panel consisting of weekly observations of a price index, total return index and the number of news stories pertaining to 23 emerging markets between 1989 and 2002. (12,217 observations) Table 8 contains descriptive statistics for the news and price data. The price and return indices are from the S&P/IFCI Emerging Markets Database

observations: 282
The results are for the unbalanced panel. Restricting the data to the period in which all countries have price data (February 1997 - June 2002, 282 observations) results in a smaller coefficient on news. (2.82) that is still statistically significant at the 99% level

引用论文
  • Avery, Christopher and Zemsky, Peter. “Multidimensional Uncertainty and Herd Behavior in Financial Markets.” The American Economic Review, September 1998, 88(4), pp.724-748.
    Google ScholarLocate open access versionFindings
  • Admati, Anat and Pfleiderer, Paul. “A Monopolistic Market for Information.” Journal of Economic Theory, 1986, v.39, pp.400-438.
    Google ScholarLocate open access versionFindings
  • Banerjee, Abhijit. “A Simple Model of Herd Behavior.” The Quarterly Journal of Economics, August 1992, 107(3), pp. 797-817.
    Google ScholarLocate open access versionFindings
  • Barlevy, Gadi and Veronesi, Pietro. “Information Acquisition in Financial Markets.” Review of Economic Studies January, 2000.
    Google ScholarLocate open access versionFindings
  • Bikhchandani, S., Hirshleifer, D., and Welch, Ivo. “A Theory of Fads, Fashion, Custom, and Cultural Change as Information Cascades.” Journal of Political Economy, 1992, 100, pp. 9921026.
    Google ScholarLocate open access versionFindings
  • Blanchard, Olivier and Watson, Mark. “Bubbles, Rational Expectations, and Financial Markets.” in Paul Wachtel ed.: Crises in the Economic and Financial Structure. Lexington Books, Lexington, MA, 1982.
    Google ScholarFindings
  • Chari, V.V. and Kehoe, Patrick. “Hot Money.” Journal of Political Economy forthcoming, 2003.
    Google ScholarLocate open access versionFindings
  • Duffie, Darrell. Dynamic Asset Pricing Theory. Princeton University Press, 1996.
    Google ScholarFindings
  • Froot, Kenneth; Scharfstein, David; Stein, Jeremy. ”Herd on the Street: Informational Inefficiencies in a Market with Short-Term Speculation.” Journal of Finance, 1992, v.47(4), pp.1461-84.
    Google ScholarLocate open access versionFindings
  • Grossman, Sanford and Stiglitz, Joeseph. “On the Impossibility of Informationally Efficient Markets.” American Economic Review, 1980, vol.70 (3), pp.393-408.
    Google ScholarLocate open access versionFindings
  • Lee, In-Ho. “On the Convergence of Informational Cascades.” Journal of Economic Theory, 1993, v.61, pp.395-411.
    Google ScholarLocate open access versionFindings
  • Levin, Andrew; Lin, Chien-Fu and Chu, James. “Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties.” Journal of Econometrics, 2002, v.108, pp.1-24.
    Google ScholarLocate open access versionFindings
  • Mankiw, Gregory; Romer, David and Shapiro, Matthew. “Stock Market Forecastability and Volatility: A Statistical Appraisal.” Review of Economic Studies, 1991, v.58, pp.455-77.
    Google ScholarLocate open access versionFindings
  • Mitchell, Mark and Mulherin, Harold. “The Impact of Public Information on the Stock Market.” The Journal of Finance, 1994, v.49 (3), pp.923-950.
    Google ScholarLocate open access versionFindings
  • Perloff, J.M. and Salop, S.C. “Equilibrium with Product Differentiation.” Review of Economic Studies, v.52, pp.107-120.
    Google ScholarLocate open access versionFindings
  • Roll, Richard. “R2” The Journal of Finance, 1988, vol.43 (2), pp.541-566.
    Google ScholarLocate open access versionFindings
  • Romer, Paul. “Endogenous Technological Change.” Journal of Political Economy, 98(5), 1990, pp.S71-S102.
    Google ScholarLocate open access versionFindings
  • Shapiro, Carl and Varian, Hal. Information Rules. Harvard Business School Press, 1998.
    Google ScholarFindings
  • Sims, Christopher. “Implications of Rational Inattention.” mimeo, 2001.
    Google ScholarLocate open access versionFindings
  • Van Nieuwerburgh, Stijn and Veldkamp, Laura. “Learning Asymmetries in Real Business Cycles.” NYU Stern Working Paper, 2003.
    Google ScholarLocate open access versionFindings
  • Veldkamp, Laura. “Slow Boom, Sudden Crash.” SIEPR Discussion Paper, 2000.
    Google ScholarLocate open access versionFindings
  • Vives, Xavier. “How Fast Do Rational Agents Learn?” Review of Economic Studies, 1993, v.60, pp.329-47.
    Google ScholarLocate open access versionFindings
  • Vives, Xavier. Oligopoly Pricing. MIT Press: Cambridge, MA. 1999.
    Google ScholarFindings
  • Welch, Ivo. “Sequential Sales, Learning and Cascades.” The Journal of Finance, June 1992, 47(2), pp. 695-732. For a derivation of this expression, see appendix B of Grossman and Stiglitz (1980). This equilibrium condition ignores integer constraints on the number of agents.
    Google ScholarLocate open access versionFindings
  • Following Perloff and Salop (1985), free-entry in the information market implies that equilibrium information price c(λt), times the number of agents who buy information N λt, times the probability an agent most prefers information of type j, Hj, must be equal to the fixed cost of entry.
    Google ScholarFindings
  • B(λ; θt) is the net benefit of buying a signal. Grossman and Stiglitz (1980, appendix B) show that equation 1 can be expressed as
    Google ScholarFindings
您的评分 :
0

 

标签
评论
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科