The association of analysts’ cash flow forecasts with stock recommendation profitability

• Pages: 343-361
• DOI: https://doi.org/10.1108/IJAIM-05-2019-0055
• Published date: 03 March 2020
• Date: 03 March 2020
• Author: Shanshan Pan,Zhaohui Randall Xu
• Subject Matter: Accounting methods/systems,Accounting & Finance

The association of analysts’cash

f‌low forecasts with stock

recommendation prof‌itability

Shanshan Pan and Zhaohui Randall Xu

Department ofAccounting, University of Houston Clear Lake,Houston, Texas, USA

Abstract

Purpose –The purpose of this paper is to examine whether analysts’cash f‌low forecasts improve the

prof‌itability of their stock recommendations and whether the positive effect of cash f‌low forecasts on

analysts’stockrecommendation performance varieswith f‌irms’earnings quality.

Design/methodology/approach –To test the authors’predictions, they identify a sample of 161,673 stock

recommendations with contemporaneous earnings forecasts and/or cash f‌low forecasts and regress market-

adjusted stock returns on a binary variable that proxies for the issuance of cash f‌low forecasts while controlling

for contemporaneous earnings forecast accuracy, earnings quality, analysts’forecast experience and capability

and certain f‌irm characteristics. The authors’test results are robust to alternative measures of recommendation

prof‌itability, earnings quality and the use of recommendationrevisions instead of recommendation levels.

Findings –The authors f‌ind that when analysts issue cash f‌low forecasts concurrently with earnings

forecasts, their stock recommendations lead to higher prof‌itability than when they only issue earnings

forecasts, after controlling for analysts’forecast capability. Moreover, the authors document that the

contemporaneous positive relationship between cash f‌low forecasts and recommendations prof‌itability is

stronger for f‌irms with low earningsquality than for f‌irms with high earnings quality. The f‌indings suggest

that cash f‌low forecastsissued by analysts in response to marketdemand likely play a more important rolein

f‌irm valuationthan cash f‌low forecasts issued by analysts mainlybecause of supply-side considerations.

Research limitations/implications –Future research could buildon these f‌indings to conduct further

investigationon the alternative incentives for analysts’forecastsof sales growth and long-term growth rates.

Practical implications –These f‌indings may alsohelp investors to better assess the quality of analysts’

researchoutputs and to identify superior stock recommendations.

Originality/value –This study provides insight into the role of cash f‌low forecasts in f‌irmvaluation and adds

fresh evidence to the debate on the usefulness of cash f‌low forecasts. It extends the stream of research on the

characteristics of analyst forecasts and increases our knowledge about the role of analysts in the f‌inancial market.

Keywords Earnings quality, Earnings forecasts, Analysts’cash f‌low forecasts,

Stock recommendations prof‌itability

Paper type Research paper

1. Introduction

Financial analysts serve as key information intermediaries in the f‌inancial market by

generating useful research outputsabout f‌irm performance, such as earnings forecasts and

cash f‌low forecasts, and outputs about f‌irm valuation and investment, such as stock

JEL classif‌ication –G24, G29, M41

The authors would like to thank C.S. Agnes Cheng, faculty and doctoral students at the Louisiana

State University Accounting Department, and participants at the American Accounting Association

Southwest Region Meeting for comments and suggestions on an earlier version of the paper.

Data availability: Data are publicly available from sources identif‌ied in the text.

Analysts’cash

f‌low forecasts

343

Received18 May 2019

Revised19 August 2019

Accepted20 August 2019

InternationalJournal of

Accounting& Information

Management

Vol.28 No. 2, 2020

pp. 343-361

© Emerald Publishing Limited

1834-7649

DOI 10.1108/IJAIM-05-2019-0055

The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1834-7649.htm

recommendations and target prices (Bradshaw,2002, 2004;Huang and Boateng, 2016).

Compared to the extensive research on the characteristics of the other forms of analysts’

outputs and the interactions among those, the literature on cash f‌low forecasts is relatively

new, and there is an active debate on whether cash f‌low forecastscontain useful information

incremental to that of earnings forecasts aboutf‌irm performance. Recent studies (Call et al.,

2009;McInnis and Collins, 2011;Ahmedand Ali, 2013) provide evidence that analysts make

more accurate earningsforecasts when they also forecast cash f‌lows and that the issuance of

cash f‌low forecasts leads to improvementin the quality of reported earnings and cash f‌lows

and to declines in earnings management activities. However, Givoly et al. (2009) show that

cash f‌low forecasts are less accurate than earningsforecasts and seem to be derived from a

mechanical extension of earnings forecasts, thus raising questions about the usefulness of

cash f‌low forecasts. More importantly, Bilinski (2014) f‌inds that analysts are more likely to

provide cash f‌low forecasts for f‌irms with high earnings quality, and are unwilling to

disclose cash f‌low forecastswhen the quality of earnings is low, which casts doubt about the

positive effect of issuing cash f‌low forecasts on analysts’earnings forecasts accuracy as

documented by Call etal. (2009).

This study attempts to provide more evidence on the usefulness and characteristics of

cash f‌low forecasts by examining whether analysts who supplement their earnings

forecasts with cash f‌low forecasts generate more prof‌itable stock recommendations and

whether the association between cash f‌low forecasts and stock recommendation

prof‌itability varies with f‌irms’earning quality. Prior research documents a clear link

between earnings forecasts and stock recommendation prof‌itability (Bradshaw, 2004;Loh

and Mian, 2006;Zhou, 2013). Both earnings and cash f‌lows constitute the fundamental

inputs to f‌irm valuation models (Schipper, 1991). Moreover, analysts are more likely to

supplement their earnings forecasts with cash f‌low forecasts when cash f‌lows are

incrementally useful to f‌irm valuation beyond earnings (Ali, 1994;Dechow, 1994;Defond

and Hung, 2003;Ahmed and Ali, 2013). Therefore, analystswho supplement their earnings

forecasts with cash f‌low forecasts should enjoy a distinct advantage in assessing f‌irm

valuation and generatingprof‌itable stock recommendations.

There is limited evidence on how cash f‌low forecasts affect analysts’ultimate research

output on investments, the buy/sell recommendations that are implicitly based on analysts’

assessment of f‌irm value. An exception is Hashim and Strong (2018), who examine the

interaction between cash f‌low forecasts and target price accuracy and f‌ind that analysts’

cash f‌low forecasts lead to more accurate target price predictions. However, their f‌indings

are subject to several limitations. First, the dependent variable in their test model, the

analyst target prices, is arguably a noisy and potentially biased measure of fundamental

value (Da and Schaumburg, 2011)[1]. The substantial noise and optimistic bias in target

price forecasts may undermine the integrity of Hashim and Strong’s results. In contrast,

analysts demonstrate the differential ability to make stock recommendations, and analysts’

compensation and job tenureincrease with the prof‌itability of their stock recommendations

(Stickel, 1992)[2]. After all, earning excess returns is the ultimate goal of investment

decisions. Therefore, this study focuses on evaluating the usefulness and sophistication of

analysts’cash f‌low forecasts by assessing the prof‌itability of analysts’stock

recommendations.

Second, prior research (Defondand Hung, 2003;Pae and Yoon, 2012) documents that the

quality of individual analyst’scashf‌low forecasts differs because analysts possess

differential cash f‌low forecasting skills and capabilities. Specif‌ically, Pae and Yoon (2012)

f‌ind that analysts exhibit individual differences in their cash f‌low forecasting abilities.

Without controlling for variation in analysts’forecasts capability, it would be diff‌icult to

IJAIM

28,2

344