Oh, Nattie, you beast! You are the “widow maker.“. You have hurt us so many times.
Natural Gas
US natural gas futures capped a fourth monthly loss in February after prices plummeted to lows not seen since June 2020. An unusually mild North American winter has dampened demand for the home-heating fuel as production soared, prompting the longest streak of monthly declines for gas contracts since the beginning of the pandemic. Analysts expect prices to remain depressed for awhile, given that forecasts show unseasonably warm weather for much of the next month. Meanwhile, US inventories continue to hover well above typical seasonal levels, with stockpiles in the latest week sitting about 27% above the five-year average — adding further pressure to prices. – Bloomberg
If the stock market is defined as the S&P500, the answer is: “On average, no.”
The following table illustrates that of the 13 leap trading days (February 29) since 1950, when the markets were open, the average leap day return is -0.06 percent versus 0.0356 percent for the total population of 18,686 trading days.
Only 5 of the 13 leap days (38 percent) have generated positive returns versus 53 percent for all trading days since 1950. Four of the past five leap days (2020 fell on a Saturday) have seen negative returns, with the last three consecutive leap days in the red.
Statistically Meaningful?
Is the return differential meaningful? Statisticians would say no.
Based on the results of our two-tailed t-test, there is insufficient evidence to conclude that a significant difference exists in the mean value between leap year returns.
The t-statistic of -0.302 indicates the direction and magnitude of the difference between the average returns, relative to the variation observed within the groups. However, the p-value of 0.768 falls short of the critical value at the 0.05 alpha level (95% confidence level) of 2.179, suggesting the hypothesis that the two means are not statistically different cannot be rejected. This implies that the observed difference in means between leap returns and all trading day returns is not statistically significant.
The significant disparity in sample sizes statistically swamps any of the signals one might gain from the differences in the data.
Traders Are Not Academics
Nevertheless, traders and algorithms are not academic statisticians. They look for patterns and run with them. In fact, it’s about time we get used to the fact that the new AI models, especially the LLMs, derive their logic from empirical observations, i.e., more data and patterns in the data, and less on traditional logic.
We have noticed an interesting anomaly with respect to the U.S. Presidential election in the political betting markets over at PredictIt.org. We’ve been trading on this site since the 2004 election.
The site is pricing the Democratic Party to win the presidential election at $.54 (think of it as a probability), with the Republican Party priced at $.48. However, PreditIt has Donald Trump priced at $.46 to win the election with President Biden priced at $.42.
Not So “Risk-Free”
Doesn’t that present an arbitrage opportunity? Sell the Democratic Party at $.53 and buy Biden at $.43, banking $.10 no matter the final outcome? But risk-free? Not so fast.
It seems the market is pricing a significant chance of Biden dropping out of the race and being replaced by a stronger candidate with a better chance of beating Trump. That is so far off the conventional wisdom’s radar it will shock many.
Nevertheless, the market is pricing Biden with only a 74 probability of being his party’s nominee versus Trump’s 91 percent.
If you have strong conviction it will be Biden and Trump on the ballot on election day, do the arbitrage. Even better, sell all the other candidates, which could bank you more but at a larger potential loss.
We heard from a friend this weekend, ‘There is no way in hell it won’t be Biden and Trump on November 5th.” We asked if he would like to bet on it. The response was a resounding no. That would have been an easy $.30 plus if he bet us straight up, hedging with a buy of both Biden and Trump to win their party nominations.
Risk-free arbitrage opportunities are rare and include counter-party risk, which many overlook or don’t consider.
Global Macro Monitor 