Sentiment of wall street journal and SPY returns. The WSJ is getting gloomier over time, but seems to be perking up as we close out the year
## [1] "I was curioius about the relationship between the WSJs tone, and the market. So I wrote a script which scrapes the wall street journals archive of headlines for everyday from 2007-01-03 to 2020-11-12. Example here: https://www.wsj.com/news/archive/2020/11/12 . It then does a sentiment analysis using a dictionary thats specialized for finacial language (the loughran dictionary). Various comparisons are made to the SPY. Could be all jank and meaningless, but its fun"
## [1] "click and drag to zoom in on time ranges. All plots are linked, so comparisons can be made across them"
The wall street journals increasingly gloomy tone seems to be driven by more negative headlines, not less positive ones. Theyre also increasingly focusing on legal disputes, and no less certain than ever, lol
Sentiment peaked back in 2007. Sad! The good ol’ days, before the fin crisis lol
Cumulative returns SPY (cash) aka daily close prices
Sentiment wsj sentiment again, for comparison to spy prices. Notice dramatic drops in sentiment during the fin crisis, and during the covid market panic earlier this year. Picking up as we approach year end
Correlations among sentiment and particular word categories, and SPY returns
## [1] "sentiment analysis using a dictionary that's specialized for finance. 'sentCount' is the sentiment, 'LaggedSent' is sentiment lagged once, so we can see if yesterday's sentiment has a relationship with stuff today, 'returns' is spy returns"
## [1] "We see that the WSJs sentiment has a small, positive, statistically significant correlation with SPY returns. I guess it lends a little credence to the idea that the market responds to news, or it may just be that the WSJ is doing a good job of covering the market. Maybe feedback, who knows? We dont see the lagged sentiment having a significant relationship, so that suggests maybe the market doesnt have much memory? Or just that this is janky. The lack of significant relationship between the market (spy returns) and our measure of uncertainty is interesting. We always hear that the market hates uncertainty. But that isnt supported here. Thats sorta what id suspected. When youre uncertain, you dont move in any predictable direction, id think. We also see predictable relationships between positivity, negativity, and the market. Market goes up with positive news, and/or the tone of the news goes up with the market. Inverse for negativity. Probably over-interpreting noise, but whatever"
there’s a ton left that could be said and analyzed, but that’s enough for now - @joe_r_odonnell