Microsoft Word links draft1.doc

Transcript

1 Economic Links and Predictable Returns* Lauren Cohen Yale School of Management Andrea Frazzini University of Chicago Graduate School of Business This draft: February 23, 2006 First draft: January 30, 2006 * Lauren Cohen, Yale School of Management, 135 Prospect Street, New Haven, CT [email protected] 06511. E-mail: ity of Chicago Graduate . Andrea Frazzini, Univers School of Business, 5807 S Woodlawn avenue, Chicago, IL, 60637. E-mail: . We would like to thank Nick Barberis, Judy Chevalier, [email protected] Doug Diamond, Will Goetzmann, Anil Ka Toby Moskowitz, shyap, Owen Lamont, Monika Piazzesi, Joseph Piotroski, Doug S kinner, Matt Spiegel, Amir Sufi, Josh Rauh and seminar participants at the Universit y of Chicago and Yale University for helpful comments. We also thank Vladimir Vladimir ov for excellent research assistance and Husayn Shahrur and Jayant Kale for provid ing us with some of the customer-supplier data.

2 ABSTRACT This paper finds evidence of return predictability across economically linked firms. We test the hypothesis that in th e presence of investors subject to attention constraints, stock prices do not promptly incorporate news about economically related firms, generating return predictability across assets. We use a dataset of firms’ principal customers to identify a set of economically re lated firms, and show that stock prices do not incorporate news involving related firms, generating predictable subsequent price moves. A long/short equity st rategy based on this effect yields monthly alphas of over 150 basis points, or over 18 percent per year. JEL Classification: G10, G11, G14 Key words: Economic links, customers, suppliers, inattention, momentum

3 Introduction I. Firms do not exist as independent entiti es, but are linked to each other through many types of relationships. Some of these links are clear and contractual, while others are implicit and less transparent. We use the fo rmer of these, clear economic links, as an well defined customer- ecifically, we focus on instrument to test investor inattention. Sp supplier links between firms. In these case s, it’s clear that the partner firms are stakeholders in each others’ operations. Thus, any shock to one of the firms has a resulting effect on its linked partner. From this starting point, we examine how shocks to one firm in the relationship translate into shocks to the linked firm in both real quantities (i.e. profits) and stock prices. If investors take into account the ex-ante 1 publicly available pplier links, prices of the partner and often longstanding customer-su firm is released into the market. If, in firm will adjust when news about its linked contrast, investors ignore publicly available lin ks, stock prices of related firms will have a predictable lag in updating to new information about firms’ trading partners. Thus, the asset pricing implications of investors with limited attention is that price prices will adjust with a lag to shocks of movements across related firms are predictable: related firms, inducing predictable returns. There are two conditions that need to be met to test for investor limited be overlooked by investors needs to be attention. First, any information thought to available to the investing public before pri ces evolve. Second, the information needs to be, in fact, salient information that investors should be reasonably expected to gather. The latter of the two conditions is clearly less objective and a more difficult condition to satisfy. We believe that customer-supplier links do satisfy both requirements, and provide a natural setting for testing investor limited attention. First, information on the customer-supplier link is publicly available in that firms are required to disclose information about operating segments in their financial 1 The customer-supplier links we examin e in the paper are those sufficiently material as to be required by e reporting standard in Section III. SFAS 131 to be reported in public financial statements. We discuss th Page 3

4 statements issued to the shareholder. Regula tion SFAS No. 131 requires firms to report than 10% of the total sales in interim the identity of customers representing more financial reports issued to shareholders. In our linked sample, the average customer accounts for 20 percent of the sales of the s upplier firm. Therefore, customers represent substantive stakeholders in the supplier fi rms. Furthermore, many of the customer- and well defined contractual ties. Second, supplier links are longstanding relationships and more importantly, as we do examine ma terial customer-supplier links, the link is in fact salient information when forming expectations about future cash flows, and so s should take this relationship into account, prices. Not only is it intuitive that investor firms depend on the customer-supplier link. we provide evidence that real activities of oup stocks in different classes for which To test for return predictability, we gr news about linked firms has been released into the market, and construct a long/short equity strategy. The central prediction is th at returns of linked firms should forecast in future returns of the partner firms’ portfolios. cross sectional differences nsider the custom er-supplier link of To better understand our approach, co Coastcast and Callaway, which is shown in accompanying Figure 1. In 2001, Coastcast Corporation was a leading manufacturer of golf club heads. Since 1993 Coastcast’s oration, a retail co major customer had been Callaway Golf Corp mpany specialized in 2 golf equipment . As of 2001, Callaway accounted for 50% of Coastcast total sales. On July 7 at 11:37 am, Callaway was downgraded by one of the analysts covering it. In a press release on the next day (June 8, 6 am) Callaway lowered second quarter revenue projections to $250 million, down from a previous revenue of $300 million. The announcement brought the expected second quarter earnings per share (EPS) down to between 35 cents and 38 cents, about half of the current mean forecast of 70 cents a y shares were down by $6.23 to close at share. By market close on July 8, Callawa $15.03, a 30% drop since June 6. In the following week the fraction of analysts issuing “buy” recommendation dropped from 77% to 50%. Going forward, nearly two months later, when Callaway announced earnings on July 19, they hit the revised mean analyst estimate exactly with 36 cents per share. Surprisingly, the negative news in early June about Callaway future earnings did 2 Both firms traded on the NYSE and had analyst coverage. Page 4

5 not impact at all Coastcast’s share price. Coastcast’s stock price was unaffected, despite for half of Coascast’s total sales dropped the fact that the single customer accounting 30% of market value in two days. Both EP S forecast ($2) and stock recommendations (100% buy) were not revised. Furthermore, a LexisNexis search of newswires and financial publication returned no news me ntions for Coastcast at all during the two- nouncement. Coastcast announced EPS at -4 month period subsequent to Callaways’s an cents on July 19, and Coastcast experienced negative returns over the subsequent two months. any salient news release about Coastcast In this example, we were unable to find venue of its major customer. However, it other than the announcement of a drop in re was not until two months later that the price of Coastcast adjusted to the new information. A strategy that would have shorted Coastcast on news of Callaways’s of 20% over the subsequent two months. slowing demand would have generated a return The above example represents in fact a much more systematic pattern across the universe of US common stocks: consistent with investors’ inattention to company links, there are significantly predictable returns across customer-supplier linked firms. Our main result is that the monthly strategy of buying firms whose customers had the most positive returns (highest quintile) in the pr evious month, and selling short firms whose t quintile), yields abnormal returns of customers had the most negative returns (lowes 1.55% per month, or an annualized return of 18.6 per year. We refer to this return reover, return the customer momentum predictability as “customer momentum”. Mo strategy has little or no exposure to the st andard traded risk factors, including the firm’s own momentum in stock returns. We test for a number of alternative ex planations of the customer momentum result. It could be that unrelated to invest the customer-supplier or limited attention of link, the effect could be driven by the s upplier’s own past returns, which may be contemporaneously correlated with the customers. In this case customer return is simply a noisy proxy for own past return of the supplier. Thus, we control for the firm’s own past returns, and find that controlling fo r own firm momentum does not affect the magnitude or significance of the customer momentum result. Alternatively, the result nblatt and Moskowitz (1999)) or by a lead- could be driven by industry momentum (Gri Page 5

6 lag relationship within industries (Mosko witz and Hou (2005) and Hou (2005)). As result, 78% of the customer-supplier link evidence against these explanations driving the momentum is unlikely to be relationships are in fact ac ross industries, so industry rolling for both of these effects does not driving the results. Not surprisingly then, cont have an impact on the magnitude or the sign ificance of the customer momentum result. Menzly and Ozbas (2005) uses upstream and downstream Finally, a recent paper by definitions of industries to define cross-industry momentum . We find that controlling for cross-industry momentum does not aff ect the customer momentum result. If limited investor attention is driving th is return predictability result from the at varying inattention varies the magnitude customer-supplier link, it should be true th funds holding to identify a subset of firms and significance of the result. We use mutual re or less likely to collect information on both customer where investors are, a priori, mo bility is indeed more (less) severe where and supplier. We show that return predicta inattention constraints are more (less) likely to be binding. Finally, we turn to measures of real act ivity and show that the customer-supplier link does matter for the correlation of real ac tivities between the two firms. We do this same firms being linked by exploiting time series variation in the and not linked over the sample. We look at real activity of lin ked firms and find that during years when the come are significantly more correlated than firms are linked, both sales and operating in that when two given firms are linked, during non-linked years. We then also show customer shocks today have significant predic tability over future supp lier real activities, while when they are not linked, there is no predictable relationship. Also, the sensitivity of suppliers’ future returns to customer sh ocks today doubles when customer-supplier are linked as oppose d to not linked. The remainder of the paper is organized as follows. Section II provides a brief background and literature review. Section III de scribes the data, while Section IV details the predictions of the limited investor attention hypothesis. Section V establishes the main customer momentum result. Section VI provides robustness checks and considers alternative explanations. Section VII examines the real effects of the customer-supplier link. Section VIII concludes. Page 6

7 II. Background and literature review There is a large body of literature in psychol ogy regarding individuals’ ability to allocate ve a difficult time attention between tasks. This literature sugge sts that individuals ha 3 processing many tasks at once ive resource and attention to . Attention is a scarce cognit one task necessarily requires a substitution of cognitive resources from other tasks information available and their limited (Kahneman (1973)). Given the vast amount of cognitive capacity, investors may choose to select only a few sources of salient information. One of the first theoretical approaches to segmented markets and investor l, investors only obtain information on a inattention is Merton’s (1987). In his mode small number of stocks. Investors then only trade on those stocks about which they are informed, so that stocks with less information and fewer traders sell at a discount stemming from the inability of these investor s to share the risks of their holdings in these stocks. Hong and Stein (1999) develop a model with multiple investor types, in which information diffuses slowly across ma rkets and agents do not extract information from prices, generating return predictabilit y. Hirshleifer and Teoh (2003) and Xiong and Peng (2005) also model investor inattentio n and derive empirical implications for 2003) focus on the presentation of firm security prices. Hirshleifer and Teoh ( ect on prices and misvaluation. Xiong and information in accounting reports and the eff Peng (2005) concentrate on investors lear ning behavior given limited attention. An empirical literature is also beginni ng to build regarding investor limited attention. Huberman and Regev (2001) study investor inattention to salient news about New a firm. In their study, a firm’s stock price soars on re-release of information in the York Times that had been published in Nature five months earlier. Turning to return predictability, Ramnath (2002) examines how ea rnings surprises of firms within in the same industry are correlated. He finds that th e first earnings surprise within an industry has information for both the earnings surprises of firms within the industry, and of returns of other firms within the industry. Hou and Moskowit z (2005) study measures of firm price delay and find that these measures help to explain (or cause variation) in 3 see Pashler and Johnston (1998). For a summary of the literature, Page 7

8 many return factors and anomalies. Furtherm ore, they find that the measure of firm ntial proxies for investor recognition. Hou price delay seems related to a number of pote (2005) find evidence that such lead-lag ef fects are predominantly an intra-industry returns on small firms within the same phenomenon: returns on large firms lead industry. DellaVigna and Pollet (2005) use dem ographic information to provide evidence that demographic shifts can be used to predict future stock returns. They interpret this as the market not fully taking into account the information contained in demographic shifts. Hong, Lim, and Stein (2000), look at price momentum to test the model of Hong and Stein (1999) and find that information, a information, diffuses nd especially negative gradually into prices. rs are Hong, Tourus, and Valkanov (2005) Two recent papers closely related to ou and Menzly and Ozbas (2005). Hong, Tourus, and Valkanov (2005) look at investor inattention in ignoring lagged industry return s to predict total equity market returns. They find that certain industries do have pr edictive power over future market returns, with the same holding true in internat ional markets. Menzly and Ozbas (2005) use upstream and downstream definitions of i ndustries and present evidence of cross- industry momentum. While both papers provide valuable evid ence on slow diffusion of do not restrict the analysis to specific information, our approach is different. We industries or specific link with in or across industries. On the other hand we focus on what we believe from the investors’ standpoint may be the more intuitive links of customer and supplier. We do not impose any structure on the relation, but simply follow the evolution of customer/supplier fir m-specific relations over time. Thus, our data allows us to test for return predic tability of individual stocks stemming from company-specific linkages when firm-specific i nformation is released into the market and generates large price movements. Not surprisi ngly, our results are robust to controls for both intra and inter i ndustry effects. III. Customer data The data is obtained from several sources. Regulation SFAS No. 131 require firms to gments in interim financial reports issued report selected information about operating se Page 8

9 to shareholders. In particular, firms are requ ired disclose certain financial information than 10% of consolidated yearly sales, for any industry segment that comprised more assets or profits, and the identity of any customer representing more than 10% of the 4 . Our sample consist of all firms listed in the CRSP/Compustat total reported sales uity (BE) and market equity (ME) at the database with non missing value of book eq fiscal-year end, for which we can iden tify the customer as another traded 5 alysis on common stocks only. CRSP/Compustat firm. We focus the an principal customers from the Compustat We extract the identity of the firm’s 6 . Our customer data cover the period between 1980 and 2004. For each segment files is another company listed on the CRSP/ firm we determine whether the customer rresponding CRSP permno number. Prior to Compustat tape and we assign it the co abbreviation of the customer name, which 1998, most firms’ customers are listed as an may vary across firms or over time. For thes e firms, we use a phonetic string matching algorithm to generate a list of potential ma tches to the customer name and subsequently we hand-matched the customer to the corre sponding permno number by inspecting the 7 firm’s name, segments and industry information . We are deliberately conservative in s to make sure that customer are matched assigning customer names and firm identifier ial information. Customers for which we to the appropriate stock returns and financ could not identify a unique match are excluded from the sample. s are known before the returns they are To ensure that the firm-customer relation p between fiscal yearend dates and stocks used to explain, we impose a six month ga returns. This mimics the standard gap imposed to match accounting variables to 8 . The final sample includes 30,622 distinct firm-year- subsequent price and return data relationships, representing a total of 11, 484 unique supplier—customer relationships between 1980 to 2004. Table I shows summary statistics for our sample. In Panel A we report the coverage of the firms in our data as a fraction of the universe of CRSP common stocks. 4 Prior to 1997, Regulation SFAS No. 14 governed se gment disclosure. SFAS No . 131, issue by the FASB in June 1997, was effective for fiscal years beginning after December 15, 1997. 5 CRSP share code 10 and 11. 6 We would like to thank Husayn Shahrur and Jayant Kale for making some of the customer data available to us. 7 We use a “soundex” algorithm to generate a list of potential matches. Page 9

10 One important feature of the sample of stocks we analyze is the relative size between distribution of firms in our sample closely firms and their principal customers. The size mimics the size distribution of the CRSP un iverse. On the other hand, the sample of rge cap securities: the average customer size firm’s principal customers is tilted toward la th of above the 90 size percentile of CRSP firms. This difference partially reflects the data generating process. Firms are required to disclose the identity of any customer representing more than 10% of the total reported sales, thus we are more likely to identify larger firms as customers, since larg er firms are more likely to be above the 10% sale cutoff. We plot the distri of our sample in Figure 2. bution of market capitalization On average the universe of stock in th is study comprises 50.6% of the total l number of common stocks traded on the market capitalization and 20.25% of the tota NYSE, AMEX and NASDAQ. The last row of Panel A shows that on average 78% of 9 firm-customer relations are between firms in different industries . This is not surprising given that inputs provided by the firms in ou r sample are often quite different from the final outputs sold by their principal customer s. Thus, the stock return predictability we analyze is mostly related to assets in differe nt industries as opposed to securities within the same industry. Limited attention hypothesis and under-reaction IV. In this section we describe the main hy pothesis and design a related investment ecture that in the presence of investors that rule to construct the test portfolios. We conj k prices do not promptly incorporate news are subject to attention constraints, stoc about related firms, and thereby gene rate price drift across securities. Stock prices underreact to firm-specific HYPOTHESIS LA (LIMITED ATTENTION): information that induces changes in valuation of related firms, generating return derreact to negative news involving related predictability across assets. Stock prices un firms, and in turn generate negative subseq uent price drift. Simi larly, stock prices 8 See, for example Fama ands French (1993). 9 The assign stocks to 48 industries based on thei r SIC code. The industry definitions are from Ken French’s website. Page 10

11 underreact to positive news involving relate d firms, and in turn generate positive subsequent price drift In a world there investors have limited ability to collect and gather information, orm the rational expectations exercise to and market participants are unable to perf extract information from prices, returns acro ss securities are predictable. News travels slowly across assets as investors with limite d attention overlook the impact of specific information on economically related firms. These investors tend to hamper the transmission of information, generating return predictability across related assets. Hypothesis LA implies that a long-short portfolio, in which a long position in stocks whose related firms recently experienced good news is offset by a short position in news, should yield positive subsequent stocks whose related firms experienced bad returns. We refer to this strategy as the customer momentum portfolio. The customer momentum portfolio is the main test portfolio in our analysis. Since some firms in our sa mple have multiple princi pal customers over many rtfolio of the corresponding customers using periods, we construct an equally weighted po stomer link. We rebalance th ese portfolios every calendar the last available supplier-cu month. Hereafter, we refer to the monthly return of this portfolio as the customer 10 return . In our base specification, we use the monthly customer return as a proxy for news about customers. We believe that a retu rn-driven news sort is appropriate because action hypothesis at hand. it closely mimics the underre To test for return predictability, we examine monthly returns on calendar time portfolios formed by sorting stocks on their lagged customer return. At the beginning of t , we rank stocks in ascending order based on the customer returns in calendar month month t -1 and we assign them to one of five qu intile portfolios. All stocks are value 10 Using different weighting scheme to co mpute customer returns does not a ffect the results. We replicated all our results using customer retu rns computed by setting weights eq ual to the percent of total sales going to each customer. For most of the paper, we ch ose to focus on equally weighted customer returns to maximize the number of firms in our sample, since unfo rtunately the dollar amount of total sales going to each customer is missing in about 19% of firm-year observations of our linked data. Page 11

12 (equally) weighted within a given portfolio , and the portfolios are rebalanced every calendar month to maintain value (equal) weights. The time series of returns of thes e portfolios tracks the calendar month performance of a portfolio strategy that is based entirely on observables (lagged customer returns). This investment rule should earn zero abnormal returns in an efficient market. We compute abnormal retu rns from a time-series regression of the 11 portfolio excess returns on traded factors in calendar time. Positive abnormal returns cate the presence of customer momentum, following positive customer returns indi a sluggish stock price resp consistent with underreaction or onse to news about related ws. Under the Hypothesis LA, controlling for firms. The opposite is true for negative ne other characteristics associated with exp ected returns, bad customer news stocks stocks, generating positive returns of our consistently underperform good customer news zero cost long/short investment rule. Finally, note that since we are interested in testing whether investors in fact do take into account the customer-supplier link when forming and updating prices, in principle there is no reason to restrict the analysis to a customer momentum strategy. A natural extension would be look at predictability from supplier to customer as well. Unfortunately, the current financial regulation requires firms to report major customers of the 10% cutoff, our sample has more (and not major suppliers). Given the presence akeholders, and not the reverse. Thus, our information about customers who are major st k price response to cu stomers’ shocks. tests are in the direction of suppliers’ stoc V. Results Table II reports correlations between the va riables we use to group stocks into portfolios. The correlations are based on monthly observation pooled across stocks. Not e associated with each other. Customer surprisingly, returns and customer returns ar returns tend to be uncorrelated with firm size, defined as the logarithm of market capitalization at the end of the previous mo nth, marke to book ratios (market value of 11 We obtain the monthly factors and the risk -free rate from Ken French’s website: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french. Page 12

13 equity divided by Compustat book value of equity) and th e stock’s return over the previous calendar year. There is a distinctive characteristic of the data that should be emphasized. A caveat that arises when sorting stocks using customer returns, is that, given the large is likely for customer returns to be highly average size of the customers in our sample, it correlated with the return of the corresponding industry. The highest correlations (0.29 and 0.26) are between customer returns, the firm’s industry returns and the customer’s industry returns. Ideally, we would like our test portfolios to contain stocks with similar industry exposure (both to the underlying industry and to the corresponding customer industry) but a large spread in customer returns. In section VI we specifically address this issue by calculating abnormal returns of our test portfolios after hedging out inter and intra industry exposure. Table III shows the basic results of this t of paper. We report returns in month portfolios formed by sorting on customer returns in month t − . The rightmost column 1 shows the returns of a zero cost portfolio th at holds the top 20 percent high customer return stocks and sells short the bottom 20 percent low customer return stocks. To be included in the portfolio, a firm must have a non missing customer return and non us month. Also, we set a minimum liquidity missing stock price at the end of the previo threshold by not allowing trading in stocks with a closing price at the end of the portfolio returns are not driven by micro- previous month below $5. This ensures that capitalization illiquid securities. Separating stocks according to the lagge d return of related firms induces large differences in subsequent returns. Looking at the difference between high customer return and low customer return stocks, it is striking that high (low) customer return today predicts high (low) subsequent stoc k returns of a related firm. The customer momentum strategy that is long the top 20% good customer news stocks and short the bottom 20% bad customer news stocks de livers Fama and French (1993) abnormal returns of 1.45% per month (t-statistic = 3.61), approximately a staggering 18.4 % per year. Adjusting returns for the stock’s own price momentum by augmenting the factor model with Carhart’s (1997) momentum factor has a negligible effect on the results. line long short portfolio earn abnormal Subsequent to portfolio formation, the base Page 13

14 returns of 1.37% per months (t-statistic = 3.12). The results show that even after stomer momentum stocks earn high (low) controlling for past returns, high (low) cu subsequent returns. We return on this issue in section V where we use a regression approach to allow for a number of control variables. The alphas rise monotonically across the quintile portfolios as the customer return goes from low (negative) in portfolio #1 to high (positive) in portfolio #5. Although abnormal returns are large and sign ificant for both legs of the long/short ymmetric: the returns of the long short strategy, customer momentum returns are as portfolio are largely driven by slow diffusion of negative news. This pattern is consistent nstraints) exacerbating the delayed response with market frictions (such as short sale co d news arrives. Using equal weight rather of stocks prices to new information when ba than value weights delivers similar results: the baseline customer momentum portfolio earns a monthly alpha of 1. 3% (t-statistic = 4.93). Table IV reports factors loadings for the calendar time portfolios. Consistent with the results in Table II, the portfolios have si milar exposure to traded factors. None of the factor loadings is significant for the long/short customer momentum portfolio, which is consistent with returns being driven by under-reaction to the initial news content, rather than reflecting systematic risk. These results are consistent with the hypothesis LA: stocks prices drift after large price movements of related firms. Furthermore, the subsequent price drift is related to the m agnitude of the initial customer return. Figure 3 better illustrates the result by reporting how customer returns predict individual stock returns at different horizons . We show the cumulative average returns in month t+k on the long/short customer momentum portfolios formed on customer returns in month t. We also plot the cumu lative abnormal return of the customer portfolio (the sorting variable). To allow for comparisons, we show returns of the customer portfolio times the total fraction of the supplier firm’s sales accounted for by the principal customers. Figure 3 shows that supplier stock prices react to information that causes large swings in the stock price of their principal customers. Looking at the long/short portfolio, supplier stock prices raise by 3.9% in month zero, where the customer portfolio jumps by 7.8%. Nevertheless , stock prices drift in the same direction The customer momentum portfolio earns a subsequent to the initial price response. Page 14

15 cumulative 4.73 percent over the subsequent year. The predictable positive returns persist for about a year. een the customer returns, the initial In Table V we explore the relation betw the subsequent price drift on both customer stocks price reaction of related firms, and and supplier. We compute customer returns using weights equal to the percent of total sales going to each customer, and form calend ar time portfolios as before. In Panel A we report the average cumulative returns on a long/short portfolio formed on the firm’s (sales-weighted) customer return in month t. is the (sale-weighted) customer CRET CCAR is the customer cumulative returns over the subsequent six return in month t, months. RET is the supplier stock return in month t. CAR is its cumulative return over Panel B we report the “U nder-Reaction” coefficients the subsequent six months. In (URC) for both the customer and the supplie rs. URC is a measure of the initial price response to a given shock as a fraction of the subsequent abnormal return. URC is defined as the fraction of total return from month t to month t+6 that occurs in month t, URC = RET / (RET + CAR), and is designed to proxy for the amount of under- incorporates new information, this fraction reaction of a stock. If the market efficiently should on average be equal to one. Values of URC less than one indicate the presence of under-reaction or a sluggish stock price resp onse to news about customers. Conversely, values of URC greater then one indicate the presence of overreaction to the initial news 12 content embedded in the customer return . The results in Table V show than on average stock prices under-react to roughly 40%. That is, when customers information about related customers by t , the stock price of a related supplier reacts experience large returns in a given month , and it subsequently closes the t by covering about 60% of the initial price gap in month is can also be seen in the significant remaining 40% over the next six months. Th positive CAR of the supplier portfolio of 2.8 % (t-statistic = 3.74) following the initial price movement of the customer. Note from Pa nel B that the URC for customers is 0.94 and not statistically different from one. Anot her way to see this, from Panel A of Table V, is that customers do not have a signif icant CCAR following the initial price jump. That is, while information that generates large price movement for the customer is 12 for suggesting this measure. We would like to thank Owen Lamont Page 15

16 quickly impounded into the customer’s stocks price, only a fraction of the initial price k price, generating the profitability of the response (60%) spills over to supplier’s stoc r firms (defined as customer momentum portfolio. Looking at larger firms versus smalle lization of all CRSP stocks that month) firms below or above the median market capita tends to be negatively related to size. Larger reveals that the under-reaction coefficients firms cover 69% of the abnormal drift in the initial month, closing the remaining 31% gap in the subsequent six months. Smaller firms cover only 35% of the gap in the initial uent six months. We re month, closing the remaining 65% in the subseq turn to this issue in the Section VI. Although the customer mo mentum total abnormal return is roughly ices tend to converge faster for large cap the same in large and small cap securities, pr stocks. nd Figure 3 support Hypothesis LA: news The results in Table III to Table V a travels slowly across stocks th at are economically related, generating large subsequent returns on a customer momentum portfolio. When positive news hi ts a portfolio of a firm’s customers, it generates a large positive subsequent drift, as initially the firm’s when a portfolio of customers experience stocks price adjusts only partially. Conversely, large negative returns in a given month, st ocks prices have (predictable) negative subsequent returns. This effect generate s the profitability of customer momentum portfolio strategies. These findings are consis tent with firms adjusting only gradually to news about economically linked firms. VI. Robustness Tests Nonsynchronous trading, characteristics and size A. Although the results are consistent with the LA hypothesis, there are a number of other plausible explanations of the data . Table VI shows results for a series of robustness test. In the Table we show av erage monthly return of the long/short customer momentum portfolio. In column 1 to 4 we report return of portfolios sorted on lagged 1-month customer return. Page 16

17 13 Nonsynchronous trading can generate positive autocorrelation across stocks. In low priced stocks when constricting the the analysis, we use monthly data and exclude test assets, hence; nonsynchronous trading is unlikely to be driving the results. between portfolio Confirming this intuition, Table V shows that skipping a week formation and investment has little effect on the return of the customer momentum portfolio. Daniel and Titman (1998a, 1998b) suggest that characteristics can be better oadings. Following Daniel, Grinblatt, Titman, predictors of future returns than factor l and Wermers (1997), we subtract from each stock return the return on a portfolio of firms matched on market equi ty, market-book, and prior one-year return quintiles (a 14 total of 125 matching portfolios) . We industry-adjust returns in a fashion similar using 15 the 48 matching industry portfolios . The results in Table VI show that firms whose customer experienced good (bad) news ou t (under) perform their corresponding characteristic portfolios or industry benchm ark. Splitting the sample into smaller and larges firms (defined as firms below or ab ove the median market capitalization of all CRSP stocks that month) or splitting the sample in halves by time period has also no effect on the results. Columns 5 and 6 report results for portfolio sorted on one year customer returns. We skip a month between the sorting period and portfolio formation. Looking at one year customer momentum, the results do vary by firms’ size. For equally weighted portfolios (or for smaller firms) the one year customer momentum is large and highly rns returns of 1.13 % a month (t-statistic = significant. The baselines rolling strategy ea value weighted strategies (or larger cap 4.16). On the other hand, although returns of stocks) are large in magnitude (the average return of the value weighted one-year customer momentum is about 70 basis po int per month), we cannot reject the hypothesis of no predictability at conventional significance levels. Table VI reports additional robustness ch ecks. All the results tell a consistent story: lagged customer stock returns predict subsequent stock returns of related firms. 13 Lo and MacKinlay (1990). 14 These 125 portfolios are reformed every month based on the market equity, M/B ratio, and prior year return from the previous month. The portfolios ar e equal weighted and the quintiles are defined with respect to the entire CRSP universe in that month. Page 17

18 Prices react to news about firms’ princi pal customers but late r drift in the same direction. The drift is equally large (on av erage about 100 basis points per month) for both smaller and large cap securities, but its persistence is correlated with size: prices converge faster in large cap securities. For smaller firms or equally weighted portfolios, the predictable returns persist for over a year. Fama MacBeth regressions: hedged returns B. In this section we use a Fama and MacBeth (1973) cross sectional regression y due to customer-supplier links by hedging approach to isolate the return predictabilit out exposure to a series of variables know to have forecasting power for the cross section of returns. We are interested in testing return predictability of individual stocks generated by firm specific news about linked fi rms, hence it is important to control for alities across asset returns. variables that would cause common regressions of individual stock returns We use Fama-MacBeth (1973) forecasting on a series of controls. The dependent vari able is this month’s supplier stock return. The independent variables of interest ar e the one-month and one-year lagged stock returns of the firm’s principal customer. We also include as controls the supplier firm’s ar lagged stock return. These variables own one-month lagged stock return and one-ye control for the reversal effect of Jegadeesh (1990) and for the price momentum effect of Jegadeesh and Titman (1993). We control for the industry momentum effect of Grinblatt and Moskowitz (1999) and the intra-industry lead-lag effect of Hou(2005) by using lagged returns of the firm’s industry portfolio. Finally, we use lagged returns of the customer’s industry portfolio to control for the cross industry momentum of Menzly and Ozbas (2005). We include (but we do no t report) firms’ size as an additional control. s of the coefficients. We weight the Table VII reports the time series average estimates by the cross sectional statistica l precision, defined as the inverse of the standard errors the coefficients in the cros s sectional regressions. Table VIII reports the risk-adjusted returns of the portfolios implicit in the Fama MacBeth analysis. Since we 15 Industry are defined as in Fama and French (1997). Page 18

19 are running one-month ahead forecasting regre ssions, the time series of the regression ly return of zero cost portfolio that hedges coefficients can be interpreted as the month 16 out the risk exposure of the remaining variables . Nevertheless, achieving these returns weights of the long short portfolio sum up to is likely to be difficult since, although the nce the regression could call for extreme zero, the single weights are unconstrained, he overweighting of some securities. To obtain feasible returns, we follow Daniel and and negative portfolio weights so that the Titman (2005) and we rescale the positive coefficients correspond to the profit of going long $1 and short $1 (either equally 17 . Table VIII reports 4-factor alphas of each of these weighted or value weighted) portfolios. The returns in the table have the following interpretation: the profit of going long $1 and short $1 in a customer momentum strategy, after hedging out exposure to size, book to market, one-month reversal s, price momentum, industry momentum and cross industry momentum. In other words, they quantify the customer return predictability that is unrelated to these fa ctors. A major difference between the returns in Table VIII and the returns in Table III is th at we now include all the available stocks in one portfolio. The results in Table VII and VIII give an unambiguous answer: past customer turns. The effect is large, robust and is returns forecast subsequent supplier stock re almost unrelated to other documented predictability effects. Using the full set of e average net effect in Table VIII (after controls and value weighted portfolios, th hedging) is around 100 basis points per month. C. Variation In Inattention If limited investor attention is driving the return predictability results we find, it should be true that varying inattention va ries the magnitude and significance of the result. In this section, we use a proxy to identify subsets of firms where attention constraints are more (less) likely to be bi nding. We test the hypo thesis that return predictability is more (less) severe for those firms in which it is more (less) likely that 16 See Fama (1976). 17 See Daniel and Titman (2005). Page 19

20 information is simultaneously collected abou t both linked firms, reducing the inattention to the customer-supplier link. The proxy we use is COMMON. For every link relation, we use data on mutual equal to the number of fund holdings fund holdings to compute COMMON, which is both securities in their portfolio in that calendar month. The idea behind COMMON is that mutual funds managers holding both securities in their portfolios are more likely to gather information or monitor more closely both the customer and the suppliers, and related firms to be impounded quicker into their link. Thus we expect information about mber of common fund ownership. prices for stocks with a high nu To construct COMMON, we extract quarterly mutual fund holdings from the CDA/Spectrum mutual funds database and match calendar month and quarter end dates of the holdings assuming that mutual funds do not change holdings between reports. Table IX reports results of these tests. Every calendar month, we use independent sorts to ranks stocks in three groups (low 30%, mid 40% and high 30%) in terms of COMMON, and we then compute lo ng-short customer momentum portfolios within each of the three categories. By constructing long-short portfolios within breadth of ownership issues, as our long- ownership categories we sidestep liquidity and short portfolios have roughly a zero loading on these. The results are in Table IX. Consistent with the customer momentum returns g inattention significantly affects returns. being driven by investor inattention, varyin For stocks with a low (or zero) overlap of common mutual fund managers (high inattention) the customer momentum return s are 3.02 % per month (t-statistic = 2.70) (value weighted), while for stocks with a large amount of common ownership across funds (low inattention) the returns are on ly 0.55 % per month and not statistically different from zero (t-statisti c = 0.58). The spread in inattention causes a significant spread in the returns to customer momentum (high inattention — low inattention) of 2.48 % per month (t-statistic = 1.98). These results lend support to the significant customer momentum returns documented in Section V and Section VI being driven by investor inattention. Page 20

21 VII. Real Effects significant and predictable return in We have thus far in the paper shown a ignoring material and publicly available supplier firms, consistent with investors customer-supplier links. The investor limited attention explanation we have conjectured is based on the assumption that investors sh ould give attention to the customer-supplier support this assumption. We exploit time link. In this section we provide evidence to variation in our link data and we show that fi rms’ real operations are significantly more to periods when they are not linked. We correlated when they are linked, relative restrict the sample for these tests to those firms that are linked at some point in the sample period. This should allow us to get a less noisy estimate of the effects of solely the same firm pairs being linked or no t linked, abstracting from other firm characteristics that determine the likelihood of being linked at all. The real quantities Panel A of Table X gives the correlations we examine are sales and operating income. 18 between customer and supplier sales and operating income , both when the pair are linked and not linked. From Panel B, correlations and cross-correlations of all real 19 quantities rise substantially when the customer and supplier are linked . The correlation of customer to supplier operating income, for example, increases by 38.7 % (t-statistic = supplier sales increases by 51.4 % (t-statistic 3.88), while the correlation of customer to = 8.55) when linked. Panel C tests the ability of customer shocks today to predict future real shocks and return shocks in supplier firms, both when customer-supplier are linked and not linked. We test this relation in a regression fram ework where industry and time effects can be controlled for. The regressions now use the re al quantities scaled by assets to also alleviate any industry specific relation be tween customer and supplier assets. The dependent variables are suppliers’ future scaled real quantities of operating income and sales, and future monthly returns. The independent variable, CRET(t), in each regression is then today’s customer return. The categorical variable LINK is equal to 1 18 Both of the real quantities are winsorized at the .01 level in the Table. The results are not sensitive to logging the variables or using another winsorizing level. 19 The t-statistics of the correlations are not show n for space, but all correlations in Panel A are significant at the 99% level. Page 21

22 when two firms are linked as customer-supp lier, and zero otherwise. The interaction of the customer-supplier being linked on term LINK*CRET(t) then measures the effect supplier shocks. We include the ability of customer shoc ks today to predict future pair is defined as the distinct (Cus. Ind, industry-pair by date fixed effects. Industry- Supp. Ind.) pair that exists between customer and supplier firms. This is then interacted with date (year or month) to get the indust ry-pair-date fixed eff ect. This fixed effect should capture any relation specific to a certain industry pairing (ex. steel and automobiles), and any date specific shock that occurred to the pairing. It thus controls for any within industry, or upstream-downstream shocks that occurred in any given pair of industries at any given time. The coeffici ent on the interaction of CRET(t) and LINK can therefore be interpreted as the increased ability of customer shocks to have antities and supplier returns within that predictive power of future supplier real qu industry-pair (ex. steel and automobiles) and year (ex. 1981 ), solely because the given set of firms were linked as o pposed to not being linked. The results in Column 1 and Column 2 of Panel C suggest that controlling for industry-pair-date effects when customer-supplier are not linked, shocks to the customer quantities of suppliers. In contrast, when do not have predictive power over future real the two firms are linked (LINK*CRET(t)), cust omer shocks today do have a significant upplier firms. Column 3 presents similar ability to predict future real shocks in s evidence for returns. Customer shocks have a significantly larger effect on the future returns of suppliers when customer-supplie r are linked (LINK*CRET(t)). In fact, the sensitivity of future suppliers’ returns to to day’s customer returns over doubles when the two firms are linked as opposed to not linked. s real operations and returns are significantly This section has thus given evidence that firm more related when the two firms are linked as customer and supplier as opposed to not linked. This lends support to the assumption, and affirms the intuition, that material customer-supplier relationships do have significant impacts on the relation between the linked firms, and thus should be given attention by investors. Page 22

23 VIII. Conclusion This paper suggests that investor limited attention can lead to return predictability across assets. We provide evid ence consistent with investors displaying limited attention, with this limited attentio n having a substantive effect on asset prices. The customer-supplier links in the paper are publicly available and often longstanding relationships between firms, with the given customer on average accounting for 20 % of the supplier’s sales. Investors, however, fail to take these lin ks into account, resulting in selling the supplier firm follo predictable returns by buying or wing a positive or negative shock, respectively, to its customer. This cu stomer momentum strategy yields over a 20% return per year and is largely unaffected in both magnitude and significance by controlling for the 3 factor model, own firm momentum, industry momentum, within industry lead-lag relationships, and across industry momentum. As well, we focus on g monthly data, hence market microstructure noise typical short term predictability usin of studies with daily or intra-daily data and asset pricing model misspecification problems related to long term stud ies are less likely to be an issue. We believe the customer-supplier link pr ovides a natural framework to test licly available to all investors, but given investor inattention. Not only is the link pub difficult to argue that this important link our results on real effects of the link, it is rming expectations about suppliers’ future should be not taken into account when fo cash flows. More generally, customer-supplie r limited attention poses a large roadblock for standard asset pricing models. What we document is not an isolated situation or constrained to a few firms, but instead a systematic violation across firms having a material effect on asset prices. If it’s true that investors ignore even these blatant links, then the informational efficiency of prices to more complex pieces of information is potentially less likely. We believe the avenue of future research in limited attention should examine to what extent different types of informat ion and different information delivery paths affect investors’ attention. As well, whether attention to types of information varies across other financial instrument and product markets. The combination of these could give a better idea of how investors process information, and so given the information environment, allow us to make richer empirical predictions about asset prices. Page 23

24 References al Fund Performance, Journal of Finance, Carhart, M.M., 1997, On Persistence in Mutu 52, 57-82 Daniel, K., Hirshleifer, D., Subrahmanyam, A ., 1998. Investor psychology and security market under- and overreactions. Journal of Finance 53, 1839-1885 Daniel, K., Titman, S., 1998a, Characteristic s or Covariances?, Journal of Portfolio Management 24(4), 24-33. Daniel, K., Titman, S., 1998b, Evidence on the Characteristics of Cross-Sectional Variation in Common Stock Returns, Journal of Finance 52, 1-33. to Tangible and Intangible Information, Daniel, K., Titman, S., 2005 , Market Reactions , Forthcoming. Journal of Finance rmers, R., 1997. Measuring mutual fund Daniel, K., Grinblatt, M., Titman, S., We . Journal of Finance 52, 1035-1058. performance with characteristic-based benchmarks Della Vigna, S., Pollett, J., 2003. Attent ion, demographics, and the stock market. Unpublished working paper. UC Be rkeley and Harvard University. Fama, E., MacBeth, J., 1973. Risk, return and equilibrium: empirical tests. Journal of Political Economy 81, 607-636. Fama, E., 1976, Foundations of Finance: Portfolio Decisions and Securities Prices (Basic BooksInc., New York). Fama, E., French, K., 1993, Common Risk Fact ors in the Returns on Stocks and Bonds, Journal of Financial Economics 33, 3–56. Industry Costs of Equity, Fama, E., French, K., 1997, Journal of Financial Economics, 43, 153-193. Gabaix, X., Laibson, D., Moloche, G., Weinberg, S., 2003. The allocation of attention: theory and evidence. Unpub lished working paper. MIT and Harvard University. Grossman, S., Stiglitz, J., 1980. On the impossibility of informationally efficient markets. American Econ omic Review 70, 393-408. Hirshleifer, D., Teoh, S., 2003. Limited atte ntion, financial reporting and disclosure. Journal of Accounting and Economics 36, 337-386. Hirshleifer, D., Teoh, S., 2004. Limited investor attention and earnings-related under- and over-reactions. Unpublished workin g paper. Ohio State University. Hirshleifer, D., Lim S., Teoh, S., 2003. Disclo sure to a credulous audience: The role of limited attention. Unpublished working paper. Ohio State University. Page 24

25 Hirshleifer, D., Hou, K., Teoh, S., Zhang, Y., 2004. Do investors overvalue firms with Accounting and Economics 38, 297-331. bloated balance sheets? Journal of Hong, H., Stein, J., 1999. A unified theory of underreaction, momentum trading, and Journal of Finance 54, 2143-2184. overreaction in asset markets. Hong, H., Lim, T., Stein, J., 2000. Bad news travels slowly: Size, analyst coverage, and the profitability of momentum strategies. Journal of Finance. LV-1, 265-295. Hong, H., Torous, W., Valkanov, R., 2005. Do industries lead the stock market? Journal of Financial Economics, forthcoming. s, price delay, and the cross-section of Hou, K., Moskowitz, T., 2005. Market friction ancial Studies. 18-3, 981-1020. expected returns. Review of Fin ff ff ect in stock returns. Hou, K., 2005. Industry information di usion and the lead-lag e Unpublished manuscript, Oh io State University. Huberman, G., Regev, T., 2001. Contagious sp eculation and a cure for cancer: a non- event that made stock prices s oar. Journal of Finance 56, 387-396. Jegadeesh, N., 1990, Evidence of Predictable Behavior of Security Returns," Journal of Finance, 45, 881-98 Jegadeesh, N., Titman, S., 1993. Returns to buying winners and selling losers: ency. Journal of Finance 48, 65-91. Implications for stock market effici Effort. Prentice Hall, New Jersey. Kahneman, D., 1973. Attention and ices do not follow random walks: evidence Lo, A., MacKinlay, C., 1988. Stock market pr view of Financial Studies 1, 41-66. from a simple specification test. Re Lo, A., MacKinlay, C., 1990a, When are cont rarian profits due to stock market overreaction?, Review of Financial Studies 3, 175-205. Menly, L., Ozbas, O., 2005. Cross-industr y momentum. Unpublished manuscript. University of Southern California. Merton, R., 1971. Optimum consumption and portfolio rules in a continuous-time model. Journal of Economic Theory 3, 373-413. Moskowitz, T., Grinblatt, M., 1999. Do industries explain mo mentum? Journal of Finance 54, 1249-1290. Pashler, H., Johnston J., 1998. Attentional lim itations in dual-task performance. In: Pashler, H., (Ed.), Attention, Psychology Press. Peng, L., Xiong, W., 2005. Investor attentio n, overconfidence, and category learning. Journal of Financial Economics, forthcoming. Ramnath, S., 2002. Investor and Analyst Reactions to Earnings Announcements of Page 25

26 Related Firms: An Empirical Analysis. Jour nal of Accounting Research. 40-5, 1351-1376. Page 26

27 Table I: Summary statistics This table shows summary statistics as of December of each year. Percent coverage of stock universe (EW) is the number of stocks with a valid custom er-supplier link divided by total number of CRSP stocks. Percent coverage of stock universe (VW) is the total market capitalization of stocks with a valid lue of the CRSP stock universe. Market to book is customer-supplier link, divided by the total market va the market value of equity divided by Compustat book value of equity. Size is the firm’s market value of equity. Min Max Mean Std Dev median Panel A: Time series (24 a nnual observations, 1981 — 2004) per year 390.0 1470.0 917.16 290.94 888.00 Number of firms in the sample Number of customer in the samp le per year 208.0 650.0 432.45 115.98 410.50 k universe (EW) 13.2 31.3 20.25 5.02 Full sample % coverage of stoc 19.76 Full sample % coverage of stoc 29.1 70.7 50.62 11.83 48.31 k universe (VW) Firm % coverage of stock universe (EW) 22.8 12.79 4.03 13.14 8.5 3.28 20.0 9.03 Firm % coverage of stock universe (VW) 9.15 4.49 Customer % coverage of stock universe (EW) 4.9 11.5 7.56 1.77 7.37 Customer % coverage of stock universe (VW) 26.34 66.42 46.44 11.27 43.50 % of firm-customer in the same industry 27.22 22.91 1.85 22.64 20.6 rm—year observations) Panel B: Firms (Pooled fi Firm size percentile 0.01 0.99 0.48 0.27 0.48 Customer size percentile 0.01 0.99 0.91 0.15 0.98 Firm book to market percen tile 0.01 0.99 0.51 0.28 0.52 0.01 0.99 0.47 Customer book to market percentile 0.49 0.26 Number of customers per firm 1.00 20.00 1.60 1.09 1.00 Percent of sales to customer 0.00 100 19.80 17.05 14.68 Page 27

28 Table II: Correlation between customer returns and supplier returns, 1981—2004 ed over all months and over all available stocks for the following Correlation coefficients are calculat variables. CXRET is the monthly return of a portfolio of a firm’s principal customers minus the CRSP value weighted market return. over the prior twelve months. Size MOM is the stock’s compounded return the previous calendar month. B/M is book-market is the log of market capitalization as of the end of ratio, which is the market value of equity divided by Compustat book value of equity. The timing of B/M follows Fama and French (1993) and is as of the previous December year-end. IXRET is the (value lue weighted market return. CXIRET is the (value weighted) stock’s industry return minus the CRSP va e CRSP value weighted market return. We assign weighted) stock’s customer industry returns minus th each CRSP stock to one of 48 industry portfolio at the end of June of each year based on its four-digit SIC code. Panel A: correlation coefficients CXRET XRET MOM SIZE B/M IXRET CXIRET 0.008 -0.004 0.008 0.210 0.288 CXRET 1.000 0.115 1.000 -0.010 -0.037 0.029 0.162 0.259 RET 1.000 MOM 0.162 0.030 0.002 0.016 SIZE 1.000 -0.226 0.000 0.039 0.015 0.028 B/M 1.000 IXRET 1.000 0.320 1.000 Panel B: Spearman Rank correlation CXRET XRET MOM SIZE B/M IXRET CXIRET 0.016 0.000 0.023 0.218 0.282 CXRET 1.000 0.122 RET 1.000 0.037 0.031 0.045 0.168 0.254 MOM 1.000 0.267 0.075 0.008 0.046 SIZE 1.000 -0.264 0.005 0.043 B/M 1.000 0.022 0.042 IXRET 1.000 0.291 1.000 Page 28

29 Table III: Customer Momentum St rategy, abnormal returns 1981—2004 This table shows calendar time portfolio abnormal re turns. At the beginning of every calendar month stocks are ranked in ascending order on the basis of the return of a portfolio of its principal customers at assigned to one of 5 quintile portfolios. All stocks the end of the previous month. The ranked stocks are , and the portfolios are re balanced every calendar are value (equally) weighted within a given portfolio month to maintain value (equal) weights. This table includes all available stocks with stock price greater than 5$ at portfolio formation. Alpha is the intercept on a regression of monthly excess return from the monthly returns from Fama and French (1993) rolling strategy. The explanatory variables are the mimicking portfolios and Carhart (1997 ) momentum factor. L/S is the alph a of a zero-cost portfolio that holds the top 20% high customer return stocks a 20% low customer return nd sells short the bottom stocks. Returns and alphas are in monthly percent, t- the coefficient estimates, statistics are shown below and 5% statistical significance is indicated in bold. Q1(low) Q2 Q3 Q4 Q5(high) L/S Panel A: value weights xret -0.596 -0.157 0.125 0.313 0.982 1.578 [-1.42] [-0.41] [0.32] [0.79] [2.14] [3.79] 3-factor alpha -1.062 -0.796 -0.541 -0.227 0.493 1.555 [-3.78] [-3.61] [-2.15] [-0.87] [1.98] [3.60] -0.821 -0.741 0.556 1.376 -0.488 -0.193 4-factor alpha [-2.93] [-3.28] [-1.89] [-0.72] [1.99] [3.13] Panel A: equal weights xret -0.457 0.148 0.385 0.391 0.854 1.311 [-1.03] [0.38] [1.01] [1.01] [2.04] [4.93] 3-factor alpha -1.166 -0.661 -0.446 -0.304 0.140 1.306 [-5.27] [-3.89] [-2.74] [-1.76] [0.71] [4.67] 4-factor alpha -0.897 -0.482 -0.272 -0.224 0.315 1.212 [-4.20] [-2.89] [-1.70] [-1.28] [1.61] [4.24] Page 29

30 folio, factor loadings 1981 — 2004 Table IV: Customer Momentum port This table shows calendar time portfolio abnormal re turns. At the beginning of every calendar month stocks are ranked in ascending order on the basis of the return of a portfolio of its principal customers at assigned to one of 5 quintile portfolios. All stocks the end of the previous month. The ranked stocks are balanced every calendar are value (equally) weighted within a given portfolio , and the portfolios are re includes all available stocks with stock price greater month to maintain value (equal) weights. This table than 5$ at portfolio formation. Alpha is the intercept on a regression of monthly excess return from the rolling strategy. The explanatory variables are the monthly returns from Fama and French (1993) ) momentum factor. L/S is the alph mimicking portfolios and Carhart (1997 a of a zero-cost portfolio that holds the top 20% high customer return stocks a 20% low customer return nd sells short the bottom statistics are shown below stocks. Returns and alphas are in monthly percent, t- the coefficient estimates, and 5% statistical significance is indicated in bold. Factor loadings (Value weights) 2 xret alpha mkt smb hml pry1 R Q1 (low) -0.596 -0.821 0.989 0.384 -0.318 -0.235 0.626 [-1.42] [-2.93] [14.31] [4.47] [-3.10] [-3.88] -0.741 1.057 0.307 Q2 -0.157 -0.115 -0.022 0.658 [-0.41] [-3.28] [17.57] [4.10] [-1.28] [-0.42] 1.063 0.309 -0.09 -0.029 0.633 Q3 0.125 -0.488 [0.32] [-1.89] [16.81] [3.92] [-0.96] [-0.52] Q4 0.313 -0.193 1.039 0.217 -0.15 -0.076 0.564 [0.79] [-0.72] [14.43] [2.42] [-1.40] [-1.20] Q5 (high) 0.982 0.556 0.982 0.681 -0.363 -0.056 0.650 [2.14] [1.99] [13.80] [7.69] [-3.43] [-0.90] L/S 1.578 1.376 -0.007 0.296 -0.045 0.179 0.041 [3.79] [3.13] [-0.07] [1.26] [-0.28] [1.93] Page 30

31 1.123 -1.521 -0.148 incipal customer. T- of every calendar month stocks ent estimates, and 5% statistical e return of the customer portfolio includes all available stocks with ile portfolios. All stocks are value rmed on the firm customer return in as the fraction of total returns from rom 1, which is the case of no under- 0.548 0.539 6.170 9.600 5.620 3.163 3.892 equent six months [t+1,t+6]. RET is the supplier’s PERCSALES quintiles tics are shown below the coeffici 5.350 4.715 3.842 4.555 r-reaction coefficients. At the beginning 086 0.132 0.199 0.313 0.615 0.529 [1.40] [1.78] [0.70] [1.30] [0.98] [0.91] [0.92] [1.58] [1.81] [4.52] [5.76] [-0.42] 2.769 2.457 1.929 0.502 0.460 0.183 0.337 0.391 -0.111 0.687 0.685 0.710 3.979 4.710 5.035 Page 31 1(low) 2 3 4 5(high) 5-1 [1.12] [1.24] [1.50] [0.63] [1.13] [0.88] [-1.17] [3.55] [0.64] [1.12] [1.29] [2.64] [3.22] [0.02] principal customers. Stocks are assigned to one of five quint firms [41.55] [30.26] [28.78] [42.43] [41.52] [43.99] [3.42] of its major customers at the end of the previous month. We us s the average cumulative returns on a long/short portfolios fo ced every calendar month to maintain value weights. This table Smaller e subsequent six months. t-statis B, the t-statistics represent the distance of the coefficient f is the customer cumulative returns over the subs Table V: Under-reaction coefficients CCAR firms Larger momentum portfolio and the corresponding unde B reports the under-reaction coefficients. URC (Under-reaction Coefficient) is defined All [1.59] [1.72] [3.74] [2.91] firms 6.791 6.795 7.026 4.192 5.270 2.055 6.076 2.799 2.383 3.854 0.600 0.689 0.348 0.939 0.932 0.954 0.888 0.911 0.965 0.948 0.961 0.073 [42.51] [41.74] [13.17] [14.57] [5.09] [3.89] [6.80] [7.56] [6.98] [9.42] [-1.09] is the customer return in month t. CRET sup cust significance is indicated in bold. Panel month t to month t+6 that occurs in month t (URC = RET / (RET + CAR)). PERCSALE is the % of firms sales accounted for by the pr reaction. 5% statistical significance is indicated in bold. stock return in month t. CAR is the cumulative return over th statistics are shown below the coefficient estimates. In Panel are ranked in ascending order based on the return of a portfolio month t. weighted within a given portfolio, and the portfolios are rebalan times the total fraction of the firm’s sales accounted for by the stock price greater than 5$ at portfolio formation Panel A report This table shows returns on the customer URC URC (sales weighted) PERCSALES 0.351 0.351 0.363 0. [1.53] [1.70] [1.15] RET CCAR[t+1,t+6] 0.442 0.495 0.336 CAR[t+1,t+6] Panel B: Under-reaction coefficients CRET [5.71] [3.89] [8.15]

32 Table VI: Robustness tests This table shows calendar time portfolio return. At the beginning of every calendar month stocks are l customers in at the end a portfolio of its principa ranked in ascending order on the basis of the return of of the previous month. The ranked stocks are assigned to one of 5 quintile portfolios. All stocks are value (equally) weighted within a given portfolio, and the overlapping portfolios are rebalanced every calendar month to maintain value (equal) weights. We report excess returns of a value (VW) and equally weighed (EW) zero-cost portfolio that hold stocks and sells short the bottom s the top 20% high customer return 20% low customer return stocks. “Larger cap stocks” are all stocks with market capitalization above the median of the CRSP universe that month, smaller stocks are below median. DGTW characteristic- nus the returns on an equally weighted portfolio of adjusted returns are defined as raw monthly returns mi all CRSP firms in the same size, market-book, an d one year momentum quintile. Industry adjusted returns are defined as raw monthly returns minus th e returns of the corresponding industry portfolio. Returns are in monthly percent, t-st atistics are shown below the coeffi cient estimates an d 5% statistical significance is indicated in bold. 1 year customer 1 month customer return return Skip a week Skip a month # months VW EW VW EW VW EW 1 2 3 4 6 5 xret 288 1.578 1.311 1.464 0.932 0.694 1.13 [3.55] [3.28] [1.85] [4.16] [3.79] [4.93] DGTW 288 1.121 0.839 1.061 0.634 0.616 0.737 [3.23] [3.23] [3.05] [2.53] [1.78] [2.90] Smaller firms 1.487 1.071 1.266 0. 879 1.093 1.216 288 [3.95] [3.06] [3.69] [2.45] [3.13] [3.66] Larger firms 1.475 1.336 1.375 1.243 0.524 0.987 288 [3.70] [4.21] [3.29] [3.87] [1.41] [3.19] 144 1.963 1.391 1.763 0.943 0.237 1.137 1981 — 1992 [4.39] [4.28] [4.08] [2.95] [0.67] [3.63] 1993 - 2004 144 1.266 0.698 1.161 0.871 1.081 1.153 [1.99] [1.66] [1.72] [1.96] [1.77] [2.75] 288 0.975 0.508 0.882 0.529 0.50 0.698 Industry adjusted [2.89] [2.14] [2.55] [2.25] [1.41] [3.24] Different industry 288 1.157 1.162 1.023 0. 883 0.817 0.945 [4.83] [2.84] [3.43] [3.01] [2.03] [3.97] Same industry 288 1.288 1.192 1.173 0.901 0.705 0.349 [2.49] [2.90] [2.34] [2.90] [1.42] [0.90] Page 32

33 Table VII: Cross sectional regressions ons of individual stocks returns. The dependent This table reports Fama-MacBeth forecasting regressi variable is the monthly stock return. The explanatory variables are the lagged customer return ( cret ), the stock’s own lagged return ( ), lagged return of the corresponding industry portfolio ( indret ), and lag ret return of the corresponding customer industry portfolio ( cindret ). Cross sectional regressions are run every l statistical precision, defined as the calendar month and the esti mates are weighted by the cross sectiona efficients in the cross sectional regressions. Cross sectional standard inverse of the standard error the co errors are adjusted for heteroskedas ticity. Fama MacBeth t-statistics are reported below the coefficient estimates and 5% statistical sign ificance is indicated in bold. (1) (2) (3) (4) (5) (6) (7) (8) cret 0.043 0.042 0.042 0.035 0.037 0. 037 0.023 0.026 t − 1 [4.96] [4.88] [5.15] [3.41] [3.82] [3.84] [2.30] [2.64] cret 0.011 0.010 0.010 0.010 0.009 12, −− 2 tt [4.68] [4.54] [3.62] [3.61] [2.74] ret -0.018 -0.021 -0.018 − t 1 [-2.92] [-3.26] [-2.96] ret 0.004 0.004 tt 2 12, −− [2.14] [2.13] indret 0.105 0.091 0.098 0.071 0.067 − t 1 [3.44] [3.06] [3.40] [2.42] [2.40] indret 0.011 0.011 0.008 tt 12, 1 −− [1.83] [1.76] [1.40] cindret 0.213 0.208 − 1 t [6.01] [6.20] -0.008 cindret −− tt 12, 1 [-1.27] 2 0.010 0.013 0.028 0.017 0.023 0.037 0.022 0.042 R Page 33

34 Table VIII: Cross sectional regressions, hedged returns rtfolio constructed using Fama-MacBeth forecasting This table reports monthly abnormal returns of po nthly stocks return. The regressions of individual stock returns. The depe ndent variable is the mo explanatory variables are the lagged customer returns ( ), the stock’s own lagged return ( ), lagged cret ret return of the corresponding industry portfolio ( indret ), and lag return of the corresponding customer industry portfolio ( every calendar month. We rescale the cindret ). Cross sectional regressions are run portfolio weights to correspond to the profit of going long $1 and short $1 (either equally weighted EW or cept on a regression of monthly excess return from value weighted VW). Abnormal returns are the inter the rolling strategy. The explanatory variables are the monthly returns from Fama and French (1993) mimicking portfolios and Carhart (199 7) momentum factor. Returns are in monthly percent, t-statistics are shown below the coefficient estimates, and 5% statistical significance is indicated in bold. VW EW cret 0.895 0.691 0.730 0.724 0.445 1.170 0.906 1.151 1.178 0.855 t − 1 [4.03] [2.59] [2.99] [3.01] [1.83] [3.57] [2.44] [3.10] [3.26] [2.26] cret 0.529 0.598 0.604 0.529 -0.136 -0.029 -0.043 -0.102 −− 12, 2 tt [2.88] [2.80] [2.83] [2.44] [-0.43] [-0.08] [-0.12] [-0.29] ret -0.862 -0.119 0.026 -0.079 -0.866 -1.005 − t 1 [-2.69] [-3.22] [0.07] [-0.22] [-2.69] [-0.32] ret 0.167 0.194 0.373 0.283 tt 12, 2 −− [0.53] [0.62] [0.86] [0.66] indret 1.013 0.791 0.819 0.518 0.563 0.297 0.243 0.098 − t 1 [3.51] [3.04] [3.32] [2.33] [1.52] [0.87] [0.74] [0.30] indret 0.208 0.219 0.180 -0.286 -0.271 -0.280 −− 12, 1 tt [0.92] [0.97] [0.85] [-0.79] [-0.73] [-0.80] cindret 1.407 1.096 − 1 t [4.92] [3.35] cindret -0.380 0.202 −− tt 12, 1 [-1.79] [0.62] Page 34

35 measured ue (equally) 5% statistical and sells short the on the basis of the weights. We report 5 quintile portfolios. low the coefficient estimates. stocks are assigned to one of that calendar month. All stocks are val 2.057 -1.722 the top 20% high customer return stocks nth stocks are ranked in ascending order , mid 40% (P2), top 30% (P3) based on COMMON, which is ntion: mutual fund holdings [0.58] [0.46] [2.70] [1.95] 3.032 -2.479 lanced every calendar month to maintain value (equal) rcent, t-statistics are shown be Page 35 the previous month. The ranked customer and supplier in their portfolio in ighed (EW) zero-cost portfolio that holds rn. At the beginning of every calendar mo [-1.98] [-1.44] P3 minus P1 (High COMMON) COMMON VW EW P1 [1.30] [0.57] P2 1.314 0.469 (Low COMMON) P3 0.554 0.335 three groups, bottom 30% (P1) overlapping portfolios are reba Table IX: Variation in inatte as the number of mutual funds holdings both Stocks are further independently ranked in significance is indicated in bold. weighted within a given portfolio, and the This table shows calendar time portfolio retu bottom 20% low customer return stocks. Returns are in monthly pe excess returns of a value (VW) and equally we return of a portfolio of its principal customers in at the end of

36 51.4% 38.7% RET is the not reported. efined in text). d non-link year ressions include es and operating nk year is defined andard errors are rrelation matrices of able are winsorized at % Increase When Linked hich the customer and ons (Linked — Not Linked) [8.55] [3.88] 0.145 0.077 ability to previous tables. C Linked Sales of Customer Not Linked Sales of Customer ing income. Panel A presents co (Linked - Not Linked) = = ) rn regressions. All variables in the t Cus Cus NL L or using other winsorizing cutoffs. All reg s are estimated with constants, which are Panel B reports differences between link an ) Cus S S — Differences In Correlati mer as a major customer (major customer is d Cus OI , S , Sup Sup S OI ( ( Panel B Correlation Page 36 turns are monthly to keep compar real quantities and returns on past customer shocks. Both sal s. Industry-pair is defined as the pairing of industries to w tities of firm sales and operat Sup NL S Table X: Real Effects of Company Links Operating Income of Supplier /Assets Operating Income of Supplier / Assets Not Linked Linked The results are not sensitive to logging at the year or monthly level. T-statistics calculated using the robust clustered st and supplier are not linked in the data. Sup NL omer-supplier relationship. The regression annual variables and prior month for the retu = Not Linked = OI 0.237 0.283 0.199 0.222 NL l significance is indicated in bold. Sup Sup L a year when the supplier reports the given custo Cus OL NL OI / Cus NL S OI Sup L S Linked Sup L OI Panel A — Correlations of Real Quantities 0.315 0.428 0.275 0.358 for each customer-supplier pair as annual sales and operating incomes of customers and suppliers, along with lagged year customers’ sales and operating income. Li Non-link year is a year when the customer This table presents the effect of company links on the real quan correlations. Panel C reports predictive regressions of supplier income are scaled by firm assets and are annual figures, while re customer returns in the prior year for the the 1 percent level throughout the table. supplier, respectively, belong in the cust industry-pair by date (year and month, respectively) fixed effect Standard errors are adjusted for clustering reported in parentheses. 5% statistica Cus L Cus L S OI

37 Yes [2.22] [2.20] 0.339 0.016 0.012 Returns(t+1) LINK* CRET(t) Yes [2.91] [-0.84] 0.540 0.072 Sales/Assets (t+1) LINK* CRET(t-1) Company Links (continued) Page 37 Yes [3.00] [-0.77] 0.024 0.422 Table X: Real Effects of Operating Income/Assets (t+1) LINK* CRET(t) CRET(t) -0.004 CRET(t) -0.011 CRET(t) 2 Ind-Pair-Date Fixed Effects (1) (2) (3) R Dependent variable Panel C — Real Effects of Customer Shocks — Linked and Not Linked

38 May r = ELY) between 20010726 36 cents rporation (ticke July 25. Company announces EPS at -4 cents July 19.Company announces EPS at 20010712 Coastcast PAR 20010627 July 5 CEO and Funder of Callawy dies. Page 38 20010613 icker = PAR) and Callaway golf co No revision in Annual EPS forecast (2$) Callaway ELY (customer) 20010530 Figure 1: Coastcast Corporation and Callaway Golf 20010515 malized (05/01/2001 = 1). ces of Coastcast Corporation (t revised from $0.70 to $0.36 30% from June 6. Quarterly EPS forecast June 8. At close Callaway's price dropped (market closed) June 8, 6am. Callaway announces earnings will be lower than expected June 7 , 11.37 am: Callawy is downgraded 20010501 0.5 1.0 0.6 1.1 0.7 0.9 0.8 and August 2001. Prices are nor This figure plots the stock pri

39 Figure 2: Size distribution This figure plots the distribution of mark et capitalization of the customer/supplier sample. Every calendar month we eciles using NYSE breakpoint. assign socks to size d We plot the % of stocks in each size bin. Th is figure includes all available stocks with price greater than 5$ between 1981 and 2004. 0.44 0.30 0.25 0.20 % of firms 0.15 0.10 0.05 0.00 P5 P4 P10 P9 P8 P7 P6 P1 P2 P3 Size decile (NYSE breakpoints) All CRSP firms Sample of supplier firms Sample of customer firms Page 39

40 Figure 3: Customer mome ntum, event-time CAR t+k on a long/short This figure shows the average cumulative return in month portfolios formed on the firm customer retu rn in month t. At the beginning of every calendar month stocks are ranked in ascending order based on the return of a portfolio vious month. Stocks are assigned to one of of its major customers at the end of the pre five quintile portfolios. The figure shows aver age cumulative returns (in %) over time of gh customer return stocks and sells short a zero cost portfolio that holds the top 20% hi the bottom 20% low customer returns stocks 12 10 8 sorting variable L/S returns 6 ret % 4 2 0 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 22 24 -12 -10 -8 month t+k CAR of customer portfolio CAR of supplier portfolio Page 40

41 Figure 4: Annual returns of customer momentum strategy This figure shows annual buy-and-hold return s on a long/short portfolios formed on customer return in month t. At the begi nning of every calendar month stocks are rn of a portfolio comprised of its major ranked in ascending order based on the retu customers in at the end of the previous mo nth. Stocks are assigned to one of five quintile portfolios. Portfolios are rebalanced monthly to maintain value weights. The figure shows annual returns of a zero cost portfolio that holds the top 20% high customer return stocks and sells short the bottom 20% low customer returns stocks Buy and Hold return 80 1988 1997 60 1992 1995 1982 1990 1981 40 1993 1986 2002 1985 1984 1987 2003 20 1989 Percent 2004 1994 2000 1991 1996 0 2001 1983 -20 1998 1999 -40 year Buy and Hold return Page 41

Related documents

CalCOFI Atlas 33

CalCOFI Atlas 33

THE EARLY STAGES IN OF THE FISHES CALIFORNIA CURRENT REGION CALIFORNIA FISHERIES COOPERATIVE OCEANIC INVESTIGATIONS ATLAS NO. 33 BY THE SPONSORED STATES OF COMMERCE DEPARTMENT UNITED OCEANIC AND ATMOS...

More info »
Computer Vision: Algorithms and Applications

Computer Vision: Algorithms and Applications

Computer Vision: Algorithms and Applications Richard Szeliski September 3, 2010 draft c © 2010 Springer This electronic draft is for non-commercial personal use only, and may not be posted or re-distr...

More info »
CDIR 2018 07 27

CDIR 2018 07 27

S. Pub. 115-7 2017-2018 Official Congressional Directory 115th Congress Convened January 3, 2017 JOINT COMMITTEE ON PRINTING UNITED STATES CONGRESS UNITED STATES GOVERNMENT PUBLISHING OFFICE WASHINGTO...

More info »
Plant Yourself E Copy

Plant Yourself E Copy

National Park Service National Park Service U.S. Department of the Interior U.S. Department of the Interior Ethnography Program Ethnography Program Northeast Region Northeast Region Y Y l a n t M o u ...

More info »
BUL119030488.tif

BUL119030488.tif

PsychoLogical 1996 the American Psychological Association, Inc. Copyright Bulletin by No. Vol. 0033-2909/96/$3.00 1996. 3,488-531 119. • Support Between Processes: Relationship The and Social Physiolo...

More info »
13128

13128

This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?record_id=13128 The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better...

More info »
‘You just believe that because…’

‘You just believe that because…’

Forthcoming in Philosophical Perspectives 1 You just believe that because... Roger White MIT 1. Introduction I believe that Tom is the proud father of a baby boy. Why do I think his child is a boy? A ...

More info »
The 9/11 Commission Report

The 9/11 Commission Report

Final FM.1pp 7/17/04 5:25 PM Page i THE 9/11 COMMISSION REPORT

More info »
u7112 connectplus directories 2019

u7112 connectplus directories 2019

UCare Connect + Medicare Provider and Pharmacy Directory Introduction This Provider and Pharmacy Directory includes information about the provider and pharmacy types in UCare Connect + Medicare and li...

More info »
U7112 UCARE CONNECT + MEDICARE PROVIDERDIR MAY 2019 DATA.sv

U7112 UCARE CONNECT + MEDICARE PROVIDERDIR MAY 2019 DATA.sv

UCare Connect + Medicare Provider and Pharmacy Directory Introduction This Provider and Pharmacy Directory includes information about the provider and pharmacy types in UCare Connect + Medicare and li...

More info »
quality cancer care

quality cancer care

This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?record_id=18359 Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis Laura Levit,...

More info »
NAS report on incarceration

NAS report on incarceration

This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?record_id=18613 The Growth of Incarceration in the United States: Exploring Causes and Consequences Jeremy Tra...

More info »
The Digital Person: Technology and Privacy in the Information Age

The Digital Person: Technology and Privacy in the Information Age

GW Law Faculty Publications & Othe r Works Faculty Scholarship 2004 y in the g it al P e rson: T echnolog y a nd P r i v The D c i a I m a t ion A ge nfor a niel J . S olo ve D Ge orge W ashington U n...

More info »
Vaginal Birth After Cesarean: New Insights. Evidence Report/Technology Assessment, No. 191

Vaginal Birth After Cesarean: New Insights. Evidence Report/Technology Assessment, No. 191

Evidence Report/Technology Assessment Number 191 Vaginal Birth After Cesarean: New Insights Prepared for: Agency for Healthcare Research and Quality U.S. Department of Health and Human Services 540 Ga...

More info »
i am us

i am us

Copyright 2005 by the American Psychological Association Journal of Personality and Social Psychology 0022-3514/05/$12.00 DOI: 10.1037/0022-3514.89.4.566 2005, Vol. 89, No. 4, 566–582 “I Am Us”: Negat...

More info »
Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation

Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation

MANAGING THE RISKS OF EXTREME EVENTS AND DISASTERS TO ADVANCE CLIMATE CHANGE ADAPTATION SPECIAL REPORT OF THE INTERGOVERNMENTAL PANEL ON CLIMATE CHANGE

More info »
more algebra

more algebra

MORE ON ALGEBRA 05E3 Contents 1. Introduction 3 2. Advice for the reader 3 3. Stably free modules 3 4. A comment on the Artin-Rees property 5 5. Fibre products of rings, I 6 6. Fibre products of rings...

More info »
COMM 750 Network S Directory   508

COMM 750 Network S Directory 508

SM Provider Directory – Blue Network S UPDATED SEPTEMBER 2016

More info »
Dialogue on Private Property

Dialogue on Private Property

Yale L aw Sc hool aw Sc hool L hol ars hi p R eposit ory egal Sc Yale L Yale L aw Schoo l Faculty Scholarship eries Faculty Scholarship S 1954 ialog ue on P D rop erty rivate P Feli . C ohe n x S aw S...

More info »