The September Effect

The September Effect gets its name from the observation that average historical performance in Septembers has been significantly below the performance of all other months during the year. Since 1945, the S&P 500's performance in September has been -0.6% (that isn't a typo; it's been a loss of 0.6%!!!).

This is truly amazing performance because from January 1945 through September 2021, the S&P 500 has been up over 2,200%. With the overall markets up so much, it's extremely unusual to find any 30-day period which has consistently performed so poorly. If we string together the historical September performances, the S&P would be down a stunning 99.1% (remember that markets can only go down a maximum of 100%).

Exhibit 1

One dollar was all September could take. The other months added twenty-two.

What $1 in the S&P 500 became, January 1945 to September 2021 · linear scale

Both figures are stated in the text of this article: the S&P 500 up over 2,200% from January 1945 to September 2021, and the same index down 99.1% if the historical September returns are strung together. The dollar outcomes are those percentages compounded from $1; no external data is used. The scale is linear and in dollars, so the two outcomes are drawn at the same size: one dollar of loss is the most the September path could ever reach — a position cannot fall below zero — against twenty-two dollars of gain for the other. Only the two endpoint values are data; the dashed lines joining them are schematic, and the path the index actually took between those dates is not shown.

And the average September performance becomes even more negative as we go back further in history because we pick up the substantial negative market performance in September of 1929 at the start of the Great Depression. And the performance numbers for the Dow index are worse than the S&P's—the Dow has been down 0.9% in Septembers since 1945. Furthermore, the September Effect appears to be a global phenomenon, with an average of 60% of global markets experiencing negative returns historically in September.

This year's September performance was driven by two events: 1) a sudden rise in long-term interest rates as the Federal Reserve (the US Central Bank) appears to be starting to unwind its quantitative easing policy, and 2) a liquidity event—the impending Evergrande bankruptcy. From late 2008 until late 2014, in response to the Great Recession, the Federal Reserve began a policy called "quantitative easing." The Fed instituted that policy again in March of 2020 in response to the Covid pandemic. Typically, the Fed controls short-term interest rates and money supply through purchases of short-term Treasury instruments.

Quantitative easing is a policy that the Fed instituted to control longer-term interest rates and the overall economic risk appetite by buying longer-term Treasury instruments as well as riskier securities such as mortgage-backed securities. Quantitative easing essentially lowers interest rates and increases the desire to take financial risk in the economy. This desire to take increased risk translated into increased purchases of equities, which partially resulted in the run-up of equities over the last few years.

The Fed announced in late September that it would start slowing its quantitative easing purchases soon and end quantitative easing altogether by early next year. This caused a substantial rise in long-term interest rates (a sell-off of bonds), which resulted in a sell-off of equities as investors expected that higher interest rates would slow down the economy and result in lower future corporate profits. The other major event was the impending bankruptcy of China Evergrande, a large real estate developer with hundreds of billions of dollars of debt.

The markets worried that this bankruptcy would be so large that it would cause a global liquidity event. Essentially, there are multiple global banks and other financial institutions that have lent to Evergrande and will likely take losses when Evergrande defaults. The problem is that there are extensive linkages across financial institutions.

So even if, for example, only J.P. Morgan has directly lent to Evergrande, Citibank and Bank of America may be exposed because they've lent to J.P. Morgan.

These types of linkages connect all major financial institutions worldwide including banks, insurance companies, and investment funds. So, whenever a major event like this occurs, all financial institutions eventually become exposed to the invent, and these events are sometimes called financial contagion. As a result, investors rush to safety, or "liquidity," and they sell their holdings of risky securities and move into cash instruments.

This is why these events are also commonly known as "liquidity events," and we see decreases (often substantial decreases) in equity markets and markets for other risky securities worldwide whenever liquidity events occur. There is no systematic explanation for the September Effect. Behavioral events such as investors rebalancing their portfolios after they come back from their summer vacations are drawn upon to explain the September Effect, but data have not supported any of these explanations so far.

From a practical standpoint when one encounters poor portfolio performance in September, it is important not to act hastily. Just note that the September Effect seems to be a persistent phenomenon and that the average historical performance in all other months of the year have been significantly better than September.