Introduction

We update our analysis that considers the market reaction of companies that beat and miss earnings. After showing an historically large difference during the Q3 earnings season, the market assigned a much smaller difference to companies that beat and missed earnings during the Q4 2022 earnings season.

Despite this shrinkage in the difference of the returns of stocks during earnings season, our NLP-based signal that extracts management sentiment from corporate earnings calls has continued to identify outperformers. The list of stocks with the highest overall sentiment has now outperformed the market for five consecutive quarters. We provide a list of stocks that rate favorably on this sentiment measure. We expect these stocks to outperform over the next quarter.

Market Dialed Back the Reward for Beating

During the past three years, we have noted the tendency for companies that beat earnings expectations to subsequently outperform, while companies that miss estimates tend to underperform (see Fig. 1). During the Q4 2022 earnings season, this trend continued. That said, the spread between the “winners” (companies that beat) and “losers” (companies that miss) fell substantially compared to the Q3 earnings season.

Note: Shows the 3-day relative return for stocks beating (dark blue bars), in line with (gray bars) and missing (light blue bars) earnings estimates. An earnings beat (miss) is defined as the stock reporting earnings at least 2% greater (less) than consensus estimates. Period of analysis from December 16, 2019 through March 3, 2023. Performance is relative to the S&P 500. Transaction costs are not considered.

The reduction in the spread between the “reward for beating” and “penalty for missing” indicates that investors have been relying less on fundamental data when valuing stocks during the just-concluded earnings season. Given the apparent decrease in importance of fundamentals, we focus the remainder of this note on a sentiment signal derived from analysis of corporate earnings call transcripts.

NLP-Based Sentiment Signal Continues to Generate Outperformance

Partnering with ProntoNLP, experts in applying NLP techniques to financial data, we developed a signal that evaluates the relative level of optimism or pessimism that managements express when discussing earnings. We instead focus on how sentiment embedded in an earnings call compares to the level set by management in prior quarters.

In our initial work[1], we found that companies for which management sentiment improves tend to outperform their peers; likewise, companies where management sentiment had soured tend to underperform. We also found that this signal was distinct from traditional sources of return; as a result, it represents a unique source of alpha. We update the performance of the NLP-based sentiment signal to reflect the latest batch of earnings results through February (see Fig. 2).

Note: Shows quarterly aggregated return of ProntoNLP sentiment-based favorable baskets relative to the S&P 500 index. Basket consists of stocks in the top decile of adjusted sentiment score. Baskets is equally-weighted and rebalanced monthly. Sentiment is based on ProntoNLP scores for stocks within the S&P 500. Period of analysis is from January 2017 through February 2023. Transaction costs are not considered.

From Fig. 2, we see that the basket of stocks for which management sentiment improved the most has generated outperformance relative to the S&P 500 for five consecutive quarters. Fig. 3 shows the breakdown of the constituents of the ProntoNLP-based sentiment list by sector. While industrials, consumer staples and health care are best represented, there is representation across the 11 GICS sectors. In the Appendix, we show the complete list of stocks for which management sentiment has seen the biggest improvement – we expect this list of stocks to outperform the S&P 500 going forward.

Note: Shows sector weights of current ProntoNLP sentiment-based long basket. Basket consists of stocks in the top decile of adjusted sentiment score. Sentiment is based on ProntoNLP scores for stocks within the S&P 500. Basket is as of February 28, 2023.

Conclusion

During the just-concluded earnings season, the difference between the reward for beating estimates and the penalty for missing fell substantially following an historically large reading during the Q3 earnings season. Investors relied less on fundamentals to distinguish “winners” from “losers” during the Q4 earnings season.

In such an environment, strategies that rely less on fundamental data should be useful. We update one such strategy, which relies on the extraction of management sentiment from earnings call transcripts to identify outperforming stocks. The basket of favored stocks from this strategy has now outperformed the market for five consecutive quarters.

Appendix

Below is a list of stocks that are rated favorably by the ProntoNLP-based sentiment signal.

[1] Alternative Data: NLP-Based Sentiment

Introduction

Market Dialed Back the Reward for Beating

Fig. 1 – The Reward for Beating Has Fallen Substantially

NLP-Based Sentiment Signal Continues to Generate Outperformance

Fig. 2 – NLP-Based Sentiment Has Produced Outperformance in Five Consecutive Quarters

Fig. 3 – Sector Weights in Favored Basket of Stocks from NLP-Based Sentiment Signal

Conclusion

Appendix