Chapter 10 provides insight about whether the periods most likely associated with narrative intensity based on corporate novel events align with statistical breakpoints identified by structural change tests in the relationships driving SP500 and firm-level returns, the VIX volatility index, trading volume, and equity ETF flows. Popular breakpoint tests of structural change are applied to each of the stock market relationships based on common fundamental/risk relationships explored in the literature. The Chow test allows for the narrative intensity periods to be imposed ex ante in testing for breakpoints. The Bai and Perron unknown multiple breakpoint test identifies the most likely points of temporal instability in the time-series relations for comparison to the narrative intensity periods without imposing them ex ante. The analysis finds that structural breaks, in particular those found in aggregate- and firm-level returns, volatility, and fund flow regressions are at least somewhat aligned with the periods of highest, and moderate, KU narrative intensity from Chapter 6.

This chapter offers concluding thoughts and avenues for future research that might incorporate novel events and narrative dynamics. The key takeaways from the book are provided. Suggestions for creative ways to conduct narratological analysis of corporate disclosure statements and internal and external communications are made. Two different econometric approaches that allow for novelty and narrative dynamics when structural change is unanticipated are discussed. The chapter closes with some insights about where the economics and finance professions may turn to deal with the evidence presented throughout the book.

This chapter applies a manual textual analysis of scapegoat effects in the stock market under the Novelty–Narrative Hypothesis. Information contained in Bloomberg News daily market wrap reports proxy for the story-weights investors place on observable and unobservable micro and macro fundamentals. Findings of dramatic variation in the narrative-based attention investors place across particular fundamentals suggests instability, in both frequency and magnitude, characterizing their forecasting strategies of future stock market outcomes. Time-varying properties of the narratives are found to be connected to shifts in the underlying variables particularly level- and gap-effects. Evidence suggests that adding narrative dynamics as proxied by scapegoat effects to benchmark macro-finance models improves fundamentals' ability to explain stock price fluctuations under uncertainty. Measures of model fit and forecasting accuracy provide the most empirical support for varying-coefficient scapegoat models based on micro and macro novel events and narrative effects, or attention-weights, attached to fundamentals.

Why do researchers turn to unstructured financial text to better understand stock market behavior? How can stock market news reports help to reveal novel events and associated narratives while allowing for unanticipated change and true uncertainty? Chapter 4 discusses the particular features of textual analysis that are attractive to researchers investigating these questions. The benefits of soft information, broader and richer information sets, textual tone, unstructured data, and novel event identification are all discussed within a narratological framework. Focus will be paid to the textual data sources of stock market news reports released by Dow Jones, The Wall Street Journal, Barron's, MarketWatch, and Bloomberg News. Word clouds from Bloomberg News stock market wrap reports based on a lexicon dictionary of unique entities and nonrepetitive events for the last twenty-seven years are presented with accompanying histograms of event frequency. The chapter motivates the benefits from employing the RavenPack news analytics platform featured predominantly throughout the empirical analysis of Chapters 5 through 10 and 12.

This chapter explores higher order relations between overarching narratives and subnarratives coursing through the stock market, that is, between macro unscheduled and micro unscheduled events. The macro KU events data from Dow Jones financial news is introduced. First, the absolute count of unscheduled macro events reported for the USA is dwarfed by their micro KU counterparts. Second, the macro KU indices share many of the same internal relationships as the corporate KU indices. Both KU macro and micro count data display similar signs of correlations between their associated ESS, relevance, and ENS scores suggesting that both narratives and subnarratives may have a common internal framework. Third, sentiment, relevance, novelty, and inertia indices for macro and micro events are interacted through principal components analysis to create KU micro and macro “narrative” series. The comovement of the two new variables over time is striking that suggests that macro narratives and micro subnarratives are positively related. The fourth key finding is that there is strong statistical evidence of “causality” only in the direction from macro KU events to corporate KU events.

Chapter 1 sets the stage for the Novelty-Narrative Hypothesis as applied to stock market outcomes throughout the book. The importance of nonrepetitive events, narrative dynamics, and investor emotion for understanding market instability is introduced through insights from great early thinkers of modern-day financial markets – think Knight and Keynes – but also through extant evidence from other disciplines such as cognitive psychology and sociology. Benefits of textual news analytics in assessing the role of novelty and narratives under uncertainty are introduced with brief descriptions of the Bloomberg and RavenPack data employed throughout. The main analytical features of Chapters 5 through 12 are previewed emphasizing the methodological benefits of big data textual analysis applied to millions of financial news reports. The chapter highlights key findings from empirical investigations connecting Knightian uncertainty indices derived from novel corporate events and narrative proxies to periods of temporal instability in stock returns, volatility, trading volume, and fund flow relationships. The chapter closes with foreshadowing a Kuhnian paradigm shift in macro-finance.

Chapter 9 extends the analysis by presenting statistical tests shedding light on the comovements between corporate KU event indices and actual stock market outcomes at the aggregate and firm level. The first stage of the analysis is based on correlation statistics between the micro KU indices and the SP500, the CAPE price-to-earnings ratio, the VIX options implied volatility index, SP500 trading volume, and equity fund ETF flows. Findings suggest a strong underlying relation between novel corporate events and stock market outcomes with many hypothesized signs consistent with the Novelty-Narrative Hypothesis. The second stage of the analysis focuses on unscheduled corporate events and the variance in analyst projections of long-term growth prospects for individual Dow Jones Industrial Average 30 firms. Results suggest a significant inverse relationship between the dispersion of firm-level analyst forecasts over time and both the count of corporate KU events and variation in KU event groups. Lastly, Granger causality tests find that the significance runs from lagged KU event and variation indices to future values of analyst growth estimates.

Chapter 6 introduces the Knightian uncertainty sentiment, novelty, and volume indices based on unscheduled corporate news events. The corporate KU Sentiment Index is presented and plotted against US stock market valuation levels over the last 20 years. KU event-months with the highest/lowest sentiment scores are identified. Similarly, the corporate KU Novelty and Relevance Indices are introduced with graphical and descriptive analysis. Taken together, the three filters for highest/lowest sentiment, highest novelty, and highest relevance are interacted with the baseline KU Index from Chapter 5 to identify periods characterized by the highest narrative intensity. Periods of moderately high narrative intensity are also identified. These points of interest will serve as benchmarks for identified breakpoints found in formal structural change tests for stock returns, volatility, volume, and equity index fund flow relationships in Chapter 10.

This chapter suggests that the field of macro-finance is primed for a scientific revolution, or paradigm shift, in Kuhnian terms away from determinate modeling structures toward partially open ones that allow for novel Knightian uncertainty-related events, narrative dynamics, and the unanticipated structural change they engender. The chapter presents a discussion on recent developments representing aggregate behavior and outcomes in financial markets based on model ambiguity and intervallic change that is allowed to occur at times and in ways that would be difficult for any researcher to anticipate ex ante. The main findings of the preceding analyses of Chapters 4 through 12 will be woven into the discussion throughout. The modeling frameworks of Knightian Uncertainty Economics and Imperfect Knowledge Economics are discussed as potential approaches to be adopted in the post-paradigm shift in macro-finance.