Post date: 2024-06-24 11:05:30 GMT
Predict macro "surprises" and maximize profits
Government "surprises" often cause abrupt changes in T-bill rates, significant market fluctuations, and other unfavorable macroeconomic impacts for investors. This note examines whether the value of upcoming CPI announcements can be predicted by analyzing institutional activity and changes in market liquidity in the U.S.
Using AbleMarkets' estimates of institutional activity and liquidity shifts, we can accurately forecast the direction of T-bill rates as far as two days ahead. Our findings indicate that institutions tend to sell off their holdings up to five days before a T-bill rate increase linked to a CPI announcement. Market makers also decrease liquidity levels ahead of such rate increases. For T-bill rate decreases driven by CPI announcements, institutions typically purchase U.S. market assets for at least two days leading up to the announcement, while market makers reduce liquidity just one day prior.
AbleMarkets builds Artificial Intelligence solutions to help investors better navigate the markets. Among AbleMarkets solutions, NewsAI, Institutional Flow AI and Changes in Liquidity deliver fundamental, institutional and market makers’ perspectives to AbleMarkets’ clients. As this note illustrates, Institutional Flow and Changes in Liquidity estimates are highly predictive of CPI announcements as reflected in the T-Bill rates. The results are consistent with prior studies: see Aldridge (2016) and Aldridge and Krawciw (2017).
The NewsAI index does not reflect the CPI changes until the day of the announcement itself. NewsAI, therefore, is not helpful in predicting CPI changes. However, there is clear evidence that institutional investors are ahead of the curve on CPI announcements at least 5 days ahead of the announcements and trade accordingly.
In fact, 5-day trend in Net Institutional Activity explains a considerable portion of the variation in the T-Bill rates on the CPI announcement dates as our regression Model 1 shows (results in Table 1).
In addition to the institutional activity, AbleMarkets estimates changes in market liquidity created by the market makers. We observe that one day ahead of the CPI announcements, the directions of the changes in market liquidity are also highly predictive of the direction of the T-bill rates on the day of the CPI announcements, but only when no news about the clear market direction is available, as documented in Model 2 column of Table 1.
In fact, the predictive power of the liquidity changes on the day preceding the CPI announcement is even higher than the institutional activity trend. This signifies that both institutions and market makers are able to accurately predict the upcoming CPI announcements at least one day ahead. (Changes in market liquidity were not predictive when measured two or more days ahead of CPI announcements.)
Finally, taken together, market liquidity and institutional flow trends delivered the most powerful performance, explaining 22% of variation in the upcoming changes in the T-Bill rates on the dates of the CPI announcements (Model 3, Table 1).