Following the introduction of Regulation Fair Disclosure by the SEC in 2000, most of the recent empirical research on the impact of disclosure and dissemination on information asymmetry for investors and companies has come out of the United States.  As a result, we were intrigued to hear of a paper recently published by a team of German researchers from Goethe University Frankfurt entitled “Do Analyst Conferences Provide Informational Benefits? Evidence from Analyst Forecast Properties in a Non-US Forecasting Environment”.
The study is apparently the first to analyse the informational role of analyst conferences in a non-US type forecasting environment. The work examined the impact of conferences / calls with listed German firms on upcoming annual earning forecasts by analysts, specifically looking at the impacton forecast error and forecast dispersion. In Germany, unlike the US following the implementation of Reg FD, most analyst conferences are conducted as “closed calls”, i.e. access is restricted to invited participants only. Interestingly, the research found that these conferences and calls improve analysts' ability to forecast future earnings accurately, suggesting that material information is being released during analyst conferences.
We spoke earlier with the team, Julian Pachta, Zoltan Novotny-Farkas and Moritz Bassemir, to find out more about their recent work and its possible implications.
Thanks for speaking to Stockopedia. Could you please talk a bit about the background/profile of the research team?
Moritz Bassemir: We are all from Goethe University Frankfurt. Zoltan and I are fourth-year accounting Phd students at the chair of Prof. Günther Gebhardt. Zoltan is also a member of the INTACCT research network. INTACCT is dedicated to accounting and capital markets research at the top international level and is a major collaboration between some of the leading universities in Europe like Lancaster University, HEC Paris and Goethe University Frankfurt. Julian graduated from Goethe University and is currently working for an international investment-banking firm in Frankfurt. Our research interests are mainly concerned with the interaction of firm disclosure and capital markets.
What inspired this research project? Why has European research in this area lagged behind the US?
Julian Pachta: Analyst conferences and calls are an interesting disclosure event. They differ from other forms of disclosure due to their unscripted, interactive nature which allows analysts to pose open questions to management and to potentially elicit information that is not disclosed otherwise. Over the last years we were observing that an increasing number of firms in Germany has started to host analyst conferences on a regular basis. At the same time, overall disclosure levels in Germany have been characterised as being low. We were interested in seeing how the establishment of analyst conferences as disclosure channel affects analysts’ forecasts in such an information environment.
I think the main reason why European research in this particular area lags behind the US is that analyst conferences have only recently been established in Continental Europe while US companies have already used analyst conferences on a broader scale since the 90s. However, it is also a general observation that, compared to the US forecasting environment, work on analysts’ forecasts in Europe is at a more nascent stage.
What would you describe as the key finding from the report?
Zoltan Novotny-Farkas: In my view, the key finding of our study is that analyst conferences provide valuable new information to participants that cannot be accessed through other information channels like earnings announcements or financial statements. Specifically, our results show that both consensus forecast error and individual forecast errors decrease more in quarters where firms host analyst conferences in conjunction with earnings announcements than in quarters where firms do not host analyst conferences concurrent with the earnings announcement. The relatively greater decrease in the mean forecast error persists over time which indicates a sustained effect of conference calls on the information set of analysts.
We also shed light on the question whether all financial analysts benefit equally from the information disseminated during analyst conferences. To this end we partition analysts into high- and low-ability analysts based on their prior forecasting ability. We show that neither of those groups benefits more from analyst conferences. Further, we find that forecast dispersion is largely unaffected by analyst conferences.
What do you see as the main differences between the US forecasting environment and Germany?
Moritz Bassemir: The German forecasting environment differs along various dimensions from the US setting. To give you some examples, prior research has shown that German analysts' forecasting accuracy is relatively poor compared to that of analysts from Anglo-Saxon countries, their recommendations add less value than the recommendations of US analysts, quarterly forecasts are less common, and a smaller number of public firms are covered by analysts. Also, German analysts have less experience than their US counterparts, and German brokerage houses are smaller.
The differences in the forecasting environment can, at least partly, be explained by the stronger insider-orientation of the German disclosure system. That is, information asymmetries are largely resolved via private information channels and therefore analysts’ forecasts play a relatively minor role as a source of information. Let me also add that given these differences we think it might not be appropriate to assume that results documented in the US equally apply to other institutional settings. This further underlines the need for research on this topic in a non-US forecasting environment.
Can you describe the underlying methodology for the work in layman’s terms?
Julian Pachta: In simple terms, we apply the following approach: In the first step, we measure forecast error and forecast dispersion before and after a firm’s earnings announcement. Then we isolate the effect of analyst conferences on analysts’ forecasts by comparing the change in forecast error and forecast dispersion between quarters where firms host earnings-related analyst conferences (conference call quarters) and quarters where firms do not host analyst conferences concurrent with earnings announcements (non-conference call quarters). For a sample of 352 German firms we use 3673 earnings announcements including 1875 conference call and 1798 non-conference call quarters.
Your work suggested a sustained impact from analyst conferences/calls. Is it just a more efficient means of communicating the same information or is there evidence that these events provide new material information that is not released through other types of disclosure?
Moritz Bassemir: This is an excellent question and a key issue we had to deal with in our study. You are right in that analyst conferences may merely contain the information already provided through other types of disclosure like earnings releases, financial statements, etc. Actually, a frequently cited motivation for hosting conference calls is their use as a substitute for direct contact with analysts. Then, analyst conferences serve as a more efficient means to communicate the same information that otherwise would be disclosed through other channels. If this is the case, however, we should not see a greater decrease in forecast error in conference call-quarters than in non-conference call quarters. Our result that analysts do infer valuable new information from conference calls, however, is plausible for several reasons. In our discussions with analysts they particularly emphasized the advantage of the open, interactive nature of the questions-and-answers section where analysts have the opportunity to elicit additional information that was not discussed in the earnings release. They also pointed out that conference calls convey valuable information beyond the earnings numbers, for instance, through managers’ body language and managers’ linguistic style or tone. In addition to this anecdotal evidence we reviewed several conference call transcripts and found that firms disclosed specific piece of information during conference calls that was not contained in the respective earnings announcements.
Comparing your results to results for the US markets, you note that the observed decrease in forecast error is larger. What does this suggest about the impact of policy commitments to increased disclosure in Europe?
Zoltan Novotny-Farkas: This is a very complex question and there is not a clear-cut answer. Given that we find an increase in forecast accuracy for conference call firms, this commitment to greater disclosure may ultimately result in a decrease in the cost of capital of these firms. And these benefits of additional disclosure may be even larger than in the US where the disclosure environment is already rich. However, our study only provides one piece of evidence to this broader issue. We should also note that the mere existence of benefits to disclosure is not sufficient to justify mandatory disclosure, because firms already have incentives to provide voluntarily information. One has to bear in mind that disclosure regulation has both costs and benefits.
The direct costs of mandating analyst conferences may include the resources, e.g. manpower, maintenance of a large investor relations department, that are necessary for the preparation and hosting of analyst conferences. Such fixed disclosure costs can make certain disclosures particularly burdensome for smaller firms. Mandatory disclosures might also have indirect costs because information provided to capital market participants can also be used by other parties (e.g. competitors). If the costs of additional mandatory disclosure exceed the expected benefits, affected firms might pursue avoidance strategies like a delisting from a specific segment of the stock exchange.
In more general terms, a regulator also has to consider the market wide effects of disclosure. An individual firm’s disclosure may have externalities that might benefit other non-competing firms from different industries. For example, information provided by a firm about e.g. new technologies, consumer trends etc. may provide a useful decision base for other non-competing firms. In fact, such externalities and market-wide costs savings are two important arguments to justify the regulation of disclosure activities. However, mandating disclosure might also generate substantial costs related with the implementation and enforcement of such a mandatory disclosure regime. To sum up, the potential firm-specific and market-wide costs and benefits of disclosure regulation are complex.
Can you see any biases that might skew the results of your work? You say that these findings are robust to a number of control variables and sensitivity tests. What did you control for?
Zoltan Novotny-Farkas: This is a good point and one we had to think about a lot. The biggest issue we had was to ensure that our results were not driven by differences in the information environment of the firms. To put it in other words, it could be that firms that host more frequently analyst conferences have generally a richer information environment. In such a case our results might be biased in favour of finding an impact of analyst conferences on analysts’ forecasts. However, several aspects of our research design control for differences in information environments. First of all, we use changes in forecast error and dispersion instead of levels which mitigates the effect of different information environments. Further, we include variables like size and number of analysts following in our regressions which also capture information environment. Besides cross-sectional analyses we also employ a within-firm design where each firm serves as its own control.
Interestingly, your work did not show a decrease in dispersion whereas, perhaps one might have expected greater convergence of analyst views following shared participation in the analyst conference - why do you think this was?
Moritz Bassemir: This is an interesting point. You are suggesting that the consensus among analysts should increase after a public disclosure event such that forecast dispersion decreases. Despite being public, analyst conferences, however, may also prompt analysts to trigger new private information that is not shared with other analysts. During a conference call, each analyst is typically only allowed to ask a few questions. Therefore, the analyst can be expected to seek for pieces of information which are missing in its own information set. And although the answer to conference call questions will be heard by all analysts, the asking analyst may uniquely complement his information set and develop private information relative to other analysts. Our analysis suggests that, on average, analysts develop new private information in conjunction with the call. This in turn increases information asymmetry among analysts and does not lead to a decrease in forecast dispersion.
Do you see any policy implications to the work?
Zoltan Novotny-Farkas: In our view there are some policy implications. In Germany, as in other European countries like the UK, Spain or Italy, there is so far no legal requirement for firms to grant open access to analyst conferences for all interested parties. And indeed, what we observed is that the vast majority of analyst conferences in Germany are conducted as “closed calls”, that is real-time access to the call is restricted to invited participants only. Since our research suggests that analyst conferences provide material information to analysts, restricting access to these events places non-invited market participants at an informational disadvantage. This conduct is not consistent with the goal of European regulators to provide a “level playing field” among investors with regard to corporate disclosures.
In the UK, there is some discussion about opening access to archived analyst research in order to level the information playing field. Is that a good idea, in your view?
Julian Pachta: In general, making analyst reports publicly available can be interesting, because it potentially makes analysts’ earnings forecasts and stock recommendations less of a black box. This may be useful, for example, to investors who use analysts’ earnings forecasts as an input for their valuation model or stock recommendations for their investment decisions.
With regard to levelling the informational playing field, however, open access to analyst reports may not be very helpful. In a situation where access to analyst conferences is restricted, merely providing access to archived analyst research still puts non-invited parties at a temporary informational disadvantage. They receive, at best, delayed information filtered through analyst reports. Rather than focus on secondary/derivative material, our research implies that it might be better for regulators to mandate open access to analyst conference calls as the primary information source.
Following this work, what do you see as potential fruitful avenues for future research?
Julian Pachta: I think focusing on the actual content of analyst conferences could be an exciting field of research. For us as accounting researchers, it would be very interesting to gain insight into how analysts’ questions during the conference call evolve around accounting related issues like the use of fair value accounting. This may be particularly interesting during a period of financial crisis. Furthermore, it would be interesting to examine analysts’ forecast accuracy with respect to specific earnings components. For example, loan loss provisions represent the largest accrual component of banks’ earnings, and thus, a main determinant of their earnings quality. Future research might evaluate analysts’ accuracy in forecasting loan loss provisions and how accuracy varies across different loan loss provisioning regimes.
Thanks for your time. For those that are interested, the research paper is available for download here.