Economic Indicators

The Weekly Economic Index (WEI) is a composite of weekly year-over-year percentage changes in 10 economic indicators: Redbook same-store sales; Rasmussen Consumer Index; new claims for unemployment insurance; continued claims for unemployment insurance; adjusted income/employment tax withholdings; railroad traffic originated; the American Staffing Association Staffing Index; steel production; wholesale sales of gasoline, diesel and jet fuel; and, weekly average US electricity load. Does WEI usefully predict U.S. stock market and government bond returns? To investigate, we relate WEI to performance data for SPDR S&P 500 (SPY) as a proxy for the stock market and iShares Barclays 20+ Year Treasury Bond (TLT) as a proxy for government bonds. Since WEI measurements are nominally Saturdays but released mid-day the next Thursdays, we align its values with SPY and TLT prices as of the close on the next Thursdays. Using weekly data as described during January 2008 (limited by WEI) through mid-February 2023, we find that: Keep Reading

Can investors exploit information about business credit tightening/loosening as reported since 1990 in the Federal Reserve’s quarterly Senior Loan Officer Survey to time the U.S. stock market? In the January 2023 draft of his paper entitled “Profitable Timing of the Stock Market with the Senior Loan Officer Survey”, Linus Wilson examines the power of “Net Percentage of Domestic Banks Tightening Standards for Commercial and Industrial Loans to Large and Middle-Market Firms” to predict S&P 500 Index next-quarter returns. A positive (negative) reading means that credit conditions are tightening (loosening) for large and medium-sized firms. Specifically, he relates January survey results to subsequent April-June stock market returns, May survey results to July-September returns, August survey results to October-December returns and November survey results to January-March returns. He considers the full sample of 32 years, two subperiods of 15 years and three subperiods of 10 years. For portfolio tests, he uses the first 15-year subperiod to model allocation decisions to the S&P 500 Index/3-month U.S. Treasury bills (either long-short the stock index or long-only the index) and applies the model to a July 2005 through March 2022 test period. Using quarterly survey results, monthly S&P 500 Index levels and monthly estimated S&P 500 dividends (from Shiller’s data) during April 1990 through March 2022, he finds that: Keep Reading

Do monthly retail sales data reliably predict U.S. stock market behavior? To investigate, we relate monthly change in retail sales to monthly S&P 500 Index return. We consider both seasonally adjusted (SA) and non-seasonally adjusted (NSA) retail sales series, as compiled by the the U.S. Census Bureau. Using monthly levels of retail sales and the S&P 500 Index during January 1992 (limited by retail sales data) through October 2022, we find that: Keep Reading

A subscriber asked whether the correlation between returns on stocks and bonds is elevated when inflation is above 5%, such that equities and fixed income offer little diversification protection. To investigate, we calculate the U.S. overall inflation rate from monthly values of the consumer price index over the prior year to find months with the inflation rate over 5%. We then compute monthly total return correlations for the following two pairs of funds when inflation is above or not above 5%:

1. Fidelity Fund (FFIDX) and Fidelity Investment Grade Bond Fund (FBNDX).
2. SPDR S&P 500 ETF Trust (SPY) and iShares iBoxx $ Investment Grade Corporate Bond ETF (LQD).

Using monthly dividend-adjusted prices for FFIDX and FBNDX since January 1980 and for SPY and LQD since July 2002, all through September 2022, we find that:

Keep Reading

Is the level, or changes in the level, of Federal Reserve (Fed) holdings of U.S. Treasuries (bills, notes, bonds and TIPS, measured weekly as of Wednesday) an indicator of future stock market and/or Treasuries returns? To investigate, we take dividend-adjusted SPDR S&P 500 (SPY) and iShares Barclays 20+ Year Treasury Bond (TLT) as tradable proxies for the U.S. stock and Treasuries markets, respectively. Using weekly Fed holdings of Treasuries, and SPY and TLT total returns during mid-December 2002 through late October 2022, we find that: Keep Reading

Are many accepted equity factors/return anomalies artifacts of the secular decline in interest rates during their discovery sample periods? In their September 2022 paper entitled “The Factor Multiverse: The Role of Interest Rates in Factor Discovery”, Jules van Binsbergen, Liang Ma and Michael Schwert study the role of the secular decline in interest rates since the early 1980s in the discovery of equity factors/return anomalies. They use value-weighted long-short portfolios and monthly reformation for all factors/anomalies. They apply duration-matched fixed income portfolio return adjustments to returns for each anomaly portfolio to model returns for the latter if there had been no interest rate decline. They then classify each anomaly as false positive (present for unadjusted returns, but not adjusted returns), false negative (present for adjusted returns, but not unadjusted returns) or robust to the effect of interest rates (present for both unadjusted and adjusted returns). Using monthly returns for 153 accepted factors/anomalies over respective original test periods and for 1,395 potential undiscovered factors/anomalies based on firm accounting variables during July 1962 through December 2020, along with contemporaneous yield data for zero coupon U.S. Treasury bonds and notes, they find that:

Keep Reading

Does the U.S. stock market respond predictably to Federal Open Market Committee (FOMC) announcements, typically released between 14:00 and 14:20 EST? In the August 2022 version of their paper entitled “The FOMC Announcement Reversal”, Tommaso Baglioni and Ruy Ribeiro examine the relationship between pre-FOMC announcement returns and post-FOMC announcement returns. Specifically, they test a reversal strategy that buys (sells) E-mini S&P 500 just before announcement at 13:50 EST when the return during the 24 hours before the announcement is negative (positive) and closes the position at the end of the trading day. They buy at the ask and sell at the bid to account for trading frictions. They compute average cumulative return per round trip transaction and Sharpe ratio as average return divided by standard deviation (standardized to reflect one trading day and the number of hours the position is open). They consider two subperiods (October 1997 through March 2011 and April 2011 through January 2020). They also look at interactions of strategy performance with four measures of economic conditions: market uncertainty (VIX), economic policy uncertainty, monetary policy uncertainty and consumer sentiment. Using intraday E-mini S&P 500 prices, exact FOMC announcement release data and measures of economic conditions on FOMC announcement dates during mid-October 1997 through January 2020 (a total of 180 scheduled FOMC announcements), they find that:

Keep Reading

Is the gap between the yield on the 10-year constant maturity U.S. Treasury note (T-note) and the 10-Year breakeven inflation rate (a measure of expected inflation over the next 10 years derived from T-note yield and 10-Year Treasury inflation-indexed constant maturity securities yield) indicative of future stock market or U.S. Treasury bond yields? To investigate, we relate monthly values of this gap (the expected real T-note gap) and changes in the gap to future monthly returns for SPDR S&P 500 ETF Trust (SPY) and iShares 20+ Year Treasury Bond ETF (TLT). Using monthly values for the four series during January 2003, limited by the breakeven inflation rate series, through July 2022, we find that: Keep Reading

Which variables deserve greatest focus when predicting stock market returns? In their July 2022 paper entitled “Searching for the Best Conditional Equity Premium Model”, Hui Guo, Saidat Sanni and Yan Yu exhaustively explore combinations of 18 previously identified potential stock market return predictors to isolate the most powerful subset. They focus on a best subset selection method with a penalty on complexity, thereby suppressing data snooping bias and selecting a manageable subset. For robustness, they consider four alternative variable selection methods. Using quarterly U.S. data for these 18 variables and S&P 500 Index levels/returns during 1947 through 2020, they find that: Keep Reading

Why have inflation, economic uncertainty and geopolitical uncertainty not driven up the price of gold? In their brief May 2022 commentary entitled “Why Isn’t the Gold Price Higher?”, Paul Gambles and James Fraser review the times when gold is, and is not, a good investment. Based on historical gold prices, with focus on recent decades, they conclude that: Keep Reading