Knowledge that Transforms

An important dimension of audit committee (AC) effectiveness that has gained the attention of regulators and academics is the financial expertise of AC members (General Accounting Office 1991; Public Oversight Board 1993; Kalbers and Fogarty 1993; DeZoort 1997, 1998; Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees 1999; DeZoort, Hermanson, Archambeault, and Reed 2002; Sarbanes-Oxley Act of 2002 [SOX] 2002; Cohen, Krishnamoorthy, and Wright 2004). Section 407 of SOX requires the Securities and Exchange Commission (SEC) to adopt rules mandating that the AC of public firms include at least one member who is a financial expert or disclose reasons for not adopting this requirement. While SOX proposes a narrow definition of financial expertise, to include individuals with experience in accounting or auditing, the SEC controversially adopted a broader definition of financial expertise that includes accounting and certain types of nonaccounting (finance and supervisory) financial expertise ....

In the aftermath of recent Wall Street scandals, the efficacy of a self-regulatory model in the brokerage industry has been called into question (Boni and Womack 2002). This questioning is not surprising, considering the conflicts of interest faced by the National Association of Securities Dealers (NASD) in its dual roles as primary industry regulator and promoter of its constituents’ interests.1 Our study contributes to recent literature investigating the relevance of brokerage industry regulation by focusing on a disclosure initiative which informs investors of investment professionals’ backgrounds.2 We focus on the following question: Can ex ante uninformed investors seeking earnings research gain knowledge pertaining to two important analyst forecasting characteristics — forecast accuracy and market credibility — by using analysts’ background disclosures? We focus on earnings forecasts because of their importance to capital markets in forming earnings expectations (Fried and Givoly 1982; O’Brien 1988) and their value as inputs to other research outputs such as stock recommendations (Loh and Mian 2006), target price forecasts (Bandyopadhyay, Brown, and Richardson 1995), valuation models (Frankel and Lee 1998), and growth and return on equity investment models (Easton et al. 2002). We focus on the accuracy of earnings forecasts because of its importance to investors (O’Brien 1991; Stickel 1992; Mikhail, Walther, and Willis 1997; Clement 1999; Loh and Mian 2006), and on the credibility of analysts’ earnings forecast revisions because prior research reveals that investors consider a number of forecast factors in addition to past accuracy when evaluating the expected accuracy of an analyst’s forecast (Stickel 1992; Clement and Tse 2003; Bonner, Walther, and Young 2003). We hand-collect a sample of financial analysts with background disclosure events (hereafter “disclosed analysts”). There are eight types of background disclosures: criminal actions, customer complaints, bankruptcies, regulatory actions, terminations, civil judicial actions, investigations, and judgments/liens.3 Because the primary focus of our study is to examine the potential benefits of using background disclosures for ex ante uninformed investors seeking earnings research, we require disclosed analysts’ research outputs to be available in the period after the date of the original incident — that is, when an uninformed investor could potentially benefit from conducting a search of public disclosures on the analyst of interest. Hereafter, we refer to this date as the “incident date”.4 Because we are the first to study disclosed analysts, we provide descriptive statistics, beginning with the frequency of analysts’ earnings forecasts and the types of firms that they follow. We find that about 16 percent of firm-quarters at the intersection of the I/B/E/S and our hand-collected NASD databases have at least one disclosed analyst making an earnings forecast. Relative to nondisclosed analysts (i.e., analysts in the NASD database without disclosure events), disclosed analysts tend to follow firms with larger sales, earnings, and market capitalizations; smaller profitability; lower book-to-market ratios; and less debt in their capital structures. Because firm characteristics differ between firms followed by disclosed versus nondisclosed analysts, and because these characteristics may affect forecasting difficulty and stock returns, we match disclosed analysts to a control group of nondisclosed analysts by firm-quarter for both our accuracy and market tests. Relying on firm-quarter relative measures, descriptive analyses of forecasting characteristics indicate that, compared to nondisclosed analysts, disclosed analysts forecast less accurately in both the current and prior four quarters, later in the quarter, and more frequently. They make their forecasts with greater delay relative to other analysts, follow more firms, work for brokerage firms of similar size, have more experience, have more job turnover, and are more consistent in the specific firms they follow. Past research has shown that many of these forecast characteristics are related to both forecast accuracy and market reaction, consequently we include them in our accuracy and market models. Our forecasting performance model evaluates the quarterly earnings forecast accuracy of disclosed analysts by regressing earnings forecast accuracy on the disclosed analyst indicator variable and controls. Consistent with the notion that disclosed analysts’ earnings forecasts are less reliable, we find that disclosed analysts’ earnings forecasts are less accurate than those of nondisclosed analysts following the same firm-quarters. In supplemental analyses, we investigate whether this accuracy result is due to our disclosure variable capturing some persistent, unmeasured analyst characteristic versus being a product of the disclosure event per se. Based on several analyses, including examining disclosed analysts’ forecasts in the predisclosure period, the impact of multiple disclosure events, and the effect of elapsed time since the disclosure event, the totality of the evidence suggests the disclosure event signals a persistent analyst characteristic. To determine the market credibility of disclosed analysts’ earnings research, we examine the short-window market to analysts’ earnings forecast for and analyst forecasting we find a market to forecast revisions by disclosed analysts relative to those by nondisclosed We this as investors consider forecast revisions by analysts with disclosures to be less We not find a market to disclosed analysts’ forecasts in the period, the market is of the analyst characteristic by the disclosure prior to its To gain on whether market in to disclosed versus nondisclosed analysts’ forecast we in forecast Relative to we find by larger to analyst forecasts in we not find investors disclosed analysts’ forecasts to a greater than question from our is analysts with disclosure events at brokerage We several for their their performance with to forecasting and market credibility not the that they are to analyst are not at brokerage we focus on earnings research, into that analysts other important such as making stock and forecasting target is that disclosed analysts other analysts in these research disclosed analysts may less value than nondisclosed analysts, a and are are less in Our study is important for several we to the evidence that investors as past accuracy is not that when to analyst earnings forecast Stickel that investors more to forecast revisions of analysts are of the and Clement and Tse that investors to a of analyst characteristics past We analyst background as important characteristic that is with both forecasting accuracy and investor to analyst in the aftermath of recent Wall Street analyst scandals, have been for the research analyst the the in the efficacy of the current model of brokerage industry called into question (Boni and Womack 2002). a of the efficacy of brokerage industry is the of our in the of et al. we to the knowledge of the current of industry regulation by examining the between this self-regulatory disclosure and analysts’ accuracy and investors and as one of the important of an equity research firm this the in importance analyst between and some of the eight types of NASD disclosures with and criminal are more To the that, these disclosures provide a — — for and we provide evidence these characteristics with research performance and We our as the background and related the of our tests. our sample and the of several analyses and by the NASD in the public disclosure investors with a for knowledge of the background and of investment to the their public disclosure database more than by more than are in of of disclosure events, a past criminal NASD disclosure events are on in the and include from investment of the for Securities brokerage firms of the for and and such as the one is by a brokerage firm time an investment a its is to the for the of the NASD disclosure that, because the NASD is by the firms to this to conflicts of interest with to investors with a and of disclosures that NASD disclosures provide benefits to investors because they important events such as customer of In addition to the per some have the NASD of events and of firms to their literature has shown in earnings forecasting accuracy on a number of including forecast and O’Brien forecast frequency and since the forecast Clement and Tse number of and firms followed (Stickel Clement and et al. 1997; Clement 1999; et al. We consider analyst — those with and without background disclosures — as an forecast Because an analyst’s prior accuracy is related to forecast accuracy Brown, and 1997; investors benefit ex ante from knowledge of an analyst’s There are two a between background disclosures and analyst as an industry the NASD has both the knowledge and to the brokerage and to by disclosures to market We that the primary for these disclosures to be as is that they have performance one of a disclosure the NASD reveals past brokerage industry terminations, analyst job has been to earnings forecasting accuracy Walther, and Willis the literature knowledge to be an important of performance and and with disclosures may have regulatory and industry in to their of and the knowledge of the of in the financial knowledge may to their performance relative to their the other are two the of a between background disclosures and analyst is the of interest in a dual as promoter of its constituents’ and as a in to the efficacy of the public disclosure the disclosed analysts in the NASD database are their earnings research outputs may not be less the accuracy of disclosed analysts’ forecasts is an we industry to and conflicts of earnings forecast accuracy to for analyst earnings research our first with background disclosures forecast earnings less accurately than analysts without background research on analyst earnings forecast revisions reveals that forecast revisions are price in the that they are with short-window market on forecast revisions Givoly and and Stickel 1992; and Clement and Tse 2003; and Lee 2003; et al. 2003; et al. Consistent with research has that past forecast performance is by the market and is an important of performance because earnings forecast accuracy is an ex factors to the accuracy of an analyst’s current earnings forecast Clement 1999; et al. 1999; Clement and Tse 2003; et al. market credibility of an earnings a short-window market reaction, be as a of the of factors with ex Consistent with the market analyst with Stickel that the market more to earnings forecast revisions by of the We that market the credibility of analysts of these and their earnings expectations Because disclosed analysts have in their less of our to the market of their short-window market to disclosed analysts’ earnings forecast revisions is than that to nondisclosed analysts’ In the market we examine the potential benefits of using background disclosures for investors are uninformed as to an analyst’s we the question: Can uninformed investors gain the knowledge of investors on a of analyst credibility these public disclosures? We more investors to less to disclosed analysts’ earnings research than less of the important between more and less investors is the and related that investors such as research that investors consider the of their when making investment and and we that investors be less to on earnings forecasts by analysts with public disclosures of their to these credibility investors are more at evaluating of brokerage research credibility et al. 2003; and and This to our investors less on disclosed analysts’ earnings forecast revisions than less model are as We to that disclosed analysts forecast less accurately than nondisclosed on the variable quarterly earnings forecast accuracy for disclosed analysts relative to that of the sample of nondisclosed analysts, and the on to be not We to for the investors more to forecast revisions by disclosed analysts than to those of the sample of nondisclosed that the for the be not are as We for and by the using a to for between the two models. In both the and the on the the to nondisclosed analysts’ revisions after for the whether the to forecast revisions when on the disclosed analyst and the of and the to disclosed analysts’ consistent with include a greater market to disclosed analysts’ revisions for more versus less as as a to disclosed result is to between to nondisclosed versus disclosed the sample Our sample with quarterly forecast in the I/B/E/S database the period We as a because in this database a time in forecasts and We for and both forecasts greater than in and analysts with Consistent with prior research (O’Brien 1991; Clement 1999; et al. 1999; et al. we the earnings forecast by an analyst for firm-quarter quarterly earnings We the quarterly forecast sample with our hand-collected database of analyst disclosure are in to this are by an analyst in the I/B/E/S earnings forecast to their we a which analyst with their first and Because the not a first we with of analyst first we and search the and and We our of analyst that be to analyst to the NASD the database analyst’s first and brokerage into search in the event of a background disclosure event, a research To control for we include analysts be in the NASD There are of a search in the NASD database on our the analyst in our to an in the NASD database and disclosure event, the analyst in our to an in the NASD database and a disclosure event, and an to the analyst’s in our from the NASD To in our sample from analysts with past disclosure events that are not in the NASD we the first two search This analysts without and analysts with disclosure the I/B/E/S sample with the hand-collected database of NASD analyst disclosure events, our sample of firm-quarters. Because descriptive analyses a number of between the financial characteristics of firms that disclosed analysts relative to those not we our sample to firms by at least one disclosed and one nondisclosed to control for the firm This our sample to firm-quarters. Because we focus on benefits of disclosures to those investors are uninformed as to an analyst’s we our sample to analysts’ forecasts after the date of the disclosed in firm-quarters. We this sample for our primary earnings forecast accuracy to the market and for the earnings forecast we require prior earnings forecast the and and the Our for the market and of and descriptive on disclosed analysts and disclosure frequency by event number of disclosure events the number of analysts with disclosures because several analysts have multiple disclosure disclosed analysts have the following of disclosed criminal customer regulatory civil judicial and an of of the eight disclosure firm characteristics between firms followed and those not followed by disclosed Relative to nondisclosed analysts, disclosed analysts follow firms with greater and earnings smaller as by return on analysts follow larger firms by market firms with lower book-to-market and firms with less financial as by the in the are on firm-quarters with at least one disclosed and one nondisclosed disclosed to nondisclosed analysts with to their forecast sample Relying on firm-quarter relative measures, descriptive analyses of forecasting characteristics indicate that, compared to nondisclosed analysts, disclosed analysts forecast less accurately in the current and prior same quarter, later in the quarter, and more analysts make their forecasts with greater delay relative to other analysts’ prior follow more firms, have more experience, have greater job turnover, and more in the specific firms they follow. Past research on analyst forecast accuracy has shown that these forecast characteristics are related to forecast accuracy and market we include them in our accuracy and market tests. and for quarterly accuracy and forecasting control Consistent with our first firm-quarter relative forecasting accuracy is with the disclosed analyst indicator Consistent with past research, current forecasting accuracy is related to past forecasting forecast experience, size, job turnover, and in the of firms by an is related to forecast and the since the prior forecast of forecasting accuracy is not related to the number of firms from to the earnings forecast accuracy of disclosed model reveals a and on the disclosed analyst our first that disclosed analysts forecast earnings less accurately than analysts without Because we that uninformed investors benefit from the NASD analyst we examine a model which analysts’ past forecast This to the disclosures provide that in past Consistent with past research, the control variable is and and to our first and the on the of the disclosed variable is to lower the relative accuracy by We make several to the of our we examine the model are to of the we a of accuracy for accuracy and we similar Because of in the forecasting we examine the of using forecasting as our control variable in models and In we both forecasting and industry forecasting for and similar Because the literature has number of followed as as number of firms we the for the in models and and similar et al. the relevance of when examining the of relative forecast To for the of analyst with a disclosure indicator we one at a without of a disclosed and a nondisclosed analyst indicator and the number of disclosed on after analyst specific and to this accuracy in supplemental analyses in we investigate whether the result is due to our disclosure variable capturing some persistent, unmeasured analyst characteristic versus the disclosure event per se. Based on several analyses, including examining disclosed analysts’ forecasts in the predisclosure period, the impact of multiple disclosure events, and the effect of elapsed time since the disclosure event, we our evidence is consistent with the disclosure event a persistent analyst characteristic. from to short-window market to disclosed versus nondisclosed analysts’ forecast We several of the market model including a model control a model with forecast control a model with firm control and a model with both forecast and firm control we on model in the Consistent with prior the on the forecast is and that the short-window market on the of the is with the of the Consistent with our the is and that market less to the forecast revisions of analysts with disclosure the of with to and using the in the and and we the effect of by of versus a effect of a effect of the effect of disclosure on the forecast is which be as being We make several to the of this a return the market is value in we an and a to of both and with similar literature has that other disclosures events may affect price and analysts to their earnings forecasts and To this we the return with both and and which are consistent with those of our original prior for et al. and et al. we investor in to forecast revisions by disclosed Consistent with the short-window market we for to forecast our analyst disclosure and controls. is following et al. as the for firm in investor group the on the analyst earnings forecast the for firm in investor group the by the for firm in investor group the We for and using a which for between the of interest the two models. In both the and the on the the to nondisclosed analysts’ revisions for the whether the to forecast revisions when on the disclosed analyst and the of and the to disclosed analysts’ a to forecast revisions by nondisclosed analysts after for other the of and is a to revisions by disclosed analysts, because to disclosed analysts, the totality of the reveals that more to analysts in This evidence not our that more investors less on disclosed analysts’ earnings forecast revisions than less Because our study to determine the benefits of using disclosures for ex ante uninformed we require disclosed analysts to be in our sample after the date of the related incident — that is, when an uninformed investor could benefit from seeking background knowledge of the analyst of interest. an of disclosed analysts’ forecasting in the period on the of the disclosed analysts forecasting accuracy in the period, the disclosure incident may have the accuracy is, disclosed analysts’ accuracy may be due to available of with of the they follow. In analysts are in the period, this is consistent with the disclosure capturing a persistent unmeasured analyst characteristic that is with analyst performance the market the market in the period, this suggests that the market is by the public disclosure of the disclosed the market both in the and this suggests the market its to disclosed analysts’ earnings forecast revisions the disclosure To which is more we examine the period from on the sample a and on from to the short-window market on the period reveals an We the accuracy result as that the disclosure variable a persistent analyst characteristic. of a market in the period suggests that the disclosure is to capital To the that analysts with multiple disclosure events forecasting performance market to their research, we an indicator variable to for analysts with multiple disclosure events and from to forecast accuracy with the addition of reveals a and on our variable of with effect related to analysts with multiple disclosure events, from to the short-window market with the addition reveals a and on our variable of with effect related to analysts with multiple disclosure events, In not to be effect for analysts one versus multiple disclosure events in the accuracy and short-window market NASD database analysts with background disclosure events for analysts by a brokerage firm and those by a brokerage firm the two is of such that an event a time is We investigate is in the of the disclosed variable with relative accuracy market credibility on the of time that has elapsed since the We examine the of elapsed time in are we time in as an by at the elapsed time since the disclosed incident using an indicator to the is greater than to the and from to the earnings forecast accuracy with the addition of the variable a and on our variable of of the accuracy effect for analysts with from to the short-window market a and on our variable of not a of the market effect for analysts with disclosure In the accuracy and market not with to the notion that the disclosure variable an analyst Because the types of disclosures may have for forecasting performance and market we disclosure event types into those of a and those of a We customer complaints, regulatory actions, terminations, and as in and and as in Because criminal and civil judicial a of and we them from analysts have more than one disclosure for these we analysts as their disclosure events are of a a disclosure event is by a disclosure event, for a criminal a civil judicial we of a disclosure with a event in the primary and a such in a from to earnings forecast accuracy on the and a and on and an on using the of and reveals a and on and an on from to the short-window market on the and reveals on the the are not at in In our that earnings forecast accuracy is more by than by disclosure the market not to price these types of disclosure In the earnings accuracy and market is that brokerage is not by a brokerage To we with for the brokerage and for the firms with the analysts and and for the firms on a of brokerage on relative in stock and and and from to earnings forecast accuracy on these brokerage to a and on our variable of from to the short-window market on the a and on our variable of In the brokerage that are similar to our Our primary analyses require at least two analysts following a one disclosed analyst and one nondisclosed To the on followed firms, we analyses a of at least analysts following the firm in both the forecast accuracy and market tests. from evaluating forecast accuracy on a of at least analysts to a and on our variable of and evaluating the short-window market on the same reveals a and on our variable of our are not by the of followed analyst a disclosed the NASD the an incident an an and a the to the that are to to an include the the that the disclosed incident is the is the is are the of those are on customer complaints, which are the of the eight types of disclosure a we our after customer disclosure from evaluating earnings forecast accuracy after customer to a and on our variable of and evaluating short-window market after customer to a and on et al. a larger market to analysts, is by the of the disclosed analysts are not in the due to the financial to less analysts, may be the the market than the disclosed incident Because a of of analysts is the of this as an we model the to be related to in et al. as control in our market model as a include number of firms forecast experience, forecast and forecast these the not as control are forecast and these and related from to a and on our variable of not that including the of a market to disclosed analysts’ forecast brokerage industry is not due to conflicts of interest faced by the the regulatory in its as both industry regulator and promoter of its constituents’ et al. find to investors of an industry initiative brokerage stock disclosures in their analyst We this of research on brokerage by examining an industry initiative to investors investment professionals’ Our is to investors are uninformed as to an analyst’s background benefit from such Consistent with the notion that disclosed analysts’ forecasts are less reliable, we find that disclosed analysts forecast earnings less accurately a firm-quarter sample of nondisclosed analysts and that this effect is both prior to and after their disclosure Based on this and supplemental analyses examining the impact of multiple disclosure events and the effect of the of the disclosure event, we that their accuracy is consistent with the disclosure capturing a persistent analyst characteristic than as a result of the disclosure per We find a market to revisions by disclosed analysts in the period, consistent with the disclosure to equity In addition to these we two first such result is that we not find the market to forecast revisions from disclosed analysts in the predisclosure This is because we find that disclosed analysts less accurate than other analysts this time period, and past research has shown that the market past accuracy and et al. 2003; Clement and Tse 2003; et al. We are at least two for we to such an prior work that the market to forecast revisions is consistent with investors their knowledge of analyst time et al. the disclosed analysts in the predisclosure period have forecast available to the market than the same analysts later in time (i.e., after the disclosure a less forecasting may the market from forming performance such as this may the to the disclosure initiative in the first is, the NASD available forecasting as to analysts, be for such a our research a number of sample the is of analysts forecasting in both the and we this as to between two of disclosed analysts, the sample available for our market in to that we not find that investors forecasts from disclosed We are at least two for we to such an this the from the the of our in to our in the — the to of the NASD database — may not benefit from disclosure because they are more to be ex ante an analyst’s We the following from the totality of our ex ante uninformed investors seeking more accurate earnings research by the market in benefit from public disclosures of analyst the market as this disclosure a persistent analyst credibility characteristic with analysts’ We make several to the in an of the model of the financial we to the literature on regulatory by the relevance to market of a disclosure initiative investment professionals’ our that disclosed analysts forecast less accurately than other analysts is to research on forecasting accuracy per because the market on forecasts to its earnings expectations and as inputs to other important research investors and as one of the important of an equity research firm To the that the NASD disclosures provide — — for analyst and we that such analysts are less accurate earnings and their forecast revisions are less of specific disclosure events of with — an investment in stock in a in a and that the investment for due to investment in to in of of the of in that for a for which to be a the of which and to and without and to be a when the customer they as an of the and that into a and as an of the firm without and from the firm and into a as the this to of civil and on of of stock the period of with is the of a the between the investment and and of of of the is investigating in with research a in of is of is

The purpose of this paper is to advance research in internal audit (IA) evaluation by developing an IA assessment model that considers interrelationships among specific factors used by external auditors to evaluate the strength of the IA function. The model is based on three factors identified by auditing standards and by prior academic research: Competence, Work Performance, and Objectivity. We develop an analytical expression of the model using the belief function framework in order to overcome limitations of prior research. Our results reveal that modeling the And relationship is essential for assessing the strength of the IA function. As far as interrelationships are concerned, the analysis shows that, when the three factors have a strong or a perfect relationship, the strength of the IA function remains high even if there is positive or negative evidence about one of the factors. This result holds as long as there are high levels of belief about the other two factors. Further, we demonstrate how the quality of corporate governance affects the evaluation of the IA function and how a costbenefit analysis can be applied to this framework to help determine the amount of external audit work needed to comply with standards. Our analysis reveals that the extent of external audit work to be carried out by the external auditor depends on the strength of the IA function and the amount of litigation and regulatory costs likely to be faced by the external auditor.

Open source collaborations are increasingly among commercial rms whose interest is pro t. Why would pro t-motivated rms voluntarily share code? One reason is that cost reductions can outweigh increases in rivalry. This is especially persuasive when the contributors make complementary products. However, cost reductions do not explain why open source is a more pro table way of sharing than other forms of licensing. Why would rms use an in exible contract like the GPL? I present a model that shows how open source licensing can lead to higher industrywide pro t than would result if a rst innovator could choose the most pro table license once it nds itself in the position of rst innovator. From behind a veil of ignorance, that is, not knowing which rm will be rst, open source licensing creates higher expected pro t for the industry as a whole, and thus for each rm, than if rst innovators were allowed to choose. In the 1990's, open-source collaborations emerged as a new way of organizing software development (Eric S. Raymond, 1999). In an open-source collaboration, members disclose

Inputs and Impacts in Charter Schools: KIPP Lynn by Joshua D. Angrist, Susan M. Dynarski, Thomas J. Kane, Parag A. Pathak and Christopher R. Walters. Published in volume 100, issue 2, pages 239-43 of American Economic Review, May 2010

We study the dynamic selection of governments. A government consists of a subset of the individuals in the society. The competence level of the government in o ce determines collective utilities (e.g., by determining the amount and quality of public goods), and each individual derives additional utility from being part of the government (e.g., corruption or rents from holding o ce). We characterize the dynamic evolution of governments and determine structure of stable governments, which arise and persist in equilibrium. Our main focus is on the impact of di erent political institutions on the selection of governments. Perfect democracy, where current members of the government do not have an incumbency advantage or special powers, always leads to the emergence of the most competent government. However, any deviation from perfect democracy destroys this result. There is always at least one other, less competent government that is also stable and can persist forever. In addition, even the least competent government can persist forever in o ce. When there are stochastic shocks to the competence levels of di erent governments or to the rules determining the election of new governments, political institutions with a greater degree of democracy (less power for incumbents) are shown to perform better, because they can adapt to changes more successfully. This suggests that a particular advantage of democratic regimes is their relative exibility. We also show that, in the presence of stochastic shocks, \\royalty-like" dictatorships may be more successful than \\junta-like " dictatorships, because they might also be more adaptable to change.

Intrafirm Trade and Product Contractibility by Andrew B. Bernard, J. Bradford Jensen, Stephen J. Redding and Peter K. Schott. Published in volume 100, issue 2, pages 444-48 of American Economic Review, May 2010

This paper considers the relationship between adult child international migration and the health outcomes of elderly parents left behind in Mexico. Overall, the evidence suggests that having a migrant child is associated with a higher probability that the elderly parent in Mexico will be in poor physical and mental health. These results call into question the popular view that family members left behind in source countries undoubtedly benefit from the international migration of their relatives.

One lingering puzzle is why voluntary contributions to public goods decline over time in experimental and real-world settings. We show that the decline of cooperation is driven by individual preferences for imperfect conditional cooperation. Many people's desire to contribute less than others, rather than changing beliefs of what others will contribute over time or people's heterogeneity in preferences makes voluntary cooperation fragile. Universal free riding thus eventually emerges, despite the fact that most people are not selfish. (D12, D 83, H41, Z13)