Abstract: We study sell-side analyst reports in China and ask why many reports forecast a wide set of non-EPS indicators and what this diversity implies for information quality and investment value. Prior research shows that analysts often forecast sales,operating cash flows,pre-tax income,cash dividends,and other metrics beyond earnings. A common interpretation is signaling mechanism:analysts who disclose more categories of forecasts are thought to hold superior private information or to exert greater research effort. Under that view,richer non-earnings disclosure should improve earnings-forecast accuracy and raise the investment value of stock recommendations. We question whether that mechanism holds in China. In the United States and Europe,analysts face broad discretion over whether and how to disclose non-EPS forecasts,so cross-sectional diversity plausibly tracks individual skill,effort,and strategic choices. By contrast,our industry interviews and institutional facts indicate that brokerage firms' compliance functions issue guidance on which financial indicators analyst reports should cover. This guidance likely drives much of the observed diversity. In this setting,the number of indicators forecasted may reflect the form and intensity of broker compliance rather than analysts’ private information. We therefore propose a compliance-regulation mechanism. Under broker compliance frameworks,reports must disclose specified forecast values for designated indicators. When compliance requires a broader set of forecasts,analysts must maintain self-consistency across indicators. This requirement compels more rigorous financial analysis and narrows the scope for arbitrary embellishment or optimistic bias in earnings forecasts,thereby reducing both forecast errors and optimism bias. However,because compliance regulation does not generate private information,greater indicator diversity should not increase the investment value of stock recommendations. We define forecast-indicator diversity as the number of distinct financial categories for which a report provides explicit forecasts. We first examine the relation between diversity and earnings-forecast errors and optimism bias.The results show that higher diversity is associated with lower errors and less optimism bias. We then test investment value by relating diversity to abnormal returns around recommendation revisions. However,we do not find a significant relation. Taken together,these results support the compliance-regulation mechanism and do not support the signaling mechanism. Next,we separate diversity into broker-level and analyst-level components. At the broker level,analysts employed by the same broker are expected to exhibit similar diversity for two reasons. On the one hand,a common research platform and shared resources could create correlated information advantages,in which case broker-level diversity might improve both earnings-forecast quality and recommendation value. On the other hand,if brokers’ formal compliance reviews and template requirements drive convergence in what indicators must be forecasted,broker-level diversity should primarily improve earnings-forecast quality while having little effect on recommendation value.At the analyst level,individual analysts retain discretion over which indicators to emphasize. If the signaling mechanism were correct,analyst-level diversity should positively affect both forecast quality and the investment value of recommendations. Operationally,we measure the broker-level component as the monthly average diversity across a broker's reports and define the residual as analyst-level.We find that 91.2% of the total variation in forecast-indicator diversity is explained by variation at the broker-by-year-month level,indicating that most variation arises from broker-level forces rather than individual analyst discretion. We then estimate the separate effects of broker-level and analyst-level diversity on earnings-forecast errors,optimism bias,and the investment value of recommendations. The broker-level component significantly reduces both forecast errors and optimism bias,whereas the analyst-level component has no significant impact. Neither component improves the abnormal-return payoff to recommendations revision. Overall,the evidence favors the compliance-regulation mechanism over the signaling mechanism. To probe how broker-level diversity improves earnings-forecast quality,we conduct two mechanism tests. First,in the six months after an administrative penalty imposed by the CSRC on a broker,the negative association between broker-level diversity and both forecast errors and optimism bias becomes more pronounced.This amplification suggests that external regulatory pressure is transmitted through internal compliance to discipline forecast production. Second,greater broker-level diversity increases the self-consistency of reported indicators,constraining arbitrary forecasting behavior and improving the rigor of analysis. These results provide additional support for the compliance-regulation mechanism. Finally,we study implications for market quality.Although offering more forecast indicators does not appear to reflect private information advantages,it generates positive externalities. Broker-level diversity significantly improves stock liquidity and price information efficiency while reducing both overall price volatility and noise-driven price volatility. However,analyst-level diversity does not significantly improve market quality. This study contributes to two strands of research.First,on non-earnings forecasts,we show that in China,forecast-indicator diversity is shaped more by formal broker compliance than by analysts' private information or incremental effort. Compliance enhances the self-consistency of disclosed indicators,thereby improving earnings-forecast quality without increasing the investment value of recommendations. Second,it complements work on the determinants of analysts’ earnings-forecast quality. While prior explanations emphasize information access and conflicts of interest,our findings indicate that brokers' form-focused compliance requirements—though not expanding analysts’ private information—induce greater self-consistency cross indicators and thus higher-quality earnings forecasts.