Midterm Elections

There isn't much new in terms of sentiment that we haven't already discussed lately, so I want to touch on something that has generated a lot of questions over the past couple of days.

Traders have come to the belief that the market simply does not go down in the days leading up to and immediately following midterm elections.  So whatever bearish setup might be out there (if there are any left) will likely be steamrolled by this positive seasonality.

Seasonality can be a powerful force, but it's rarely a reason to abandon everything else.  Anyway, let's go back to 1922, the furthest back that I can find reliable dates, and see how the market performed around these elections.

I used Dow Jones Industrial Average prices prior to 1950, and S&P 500 prices thereafter.  Also, election dates were exchange holidays prior to 1970, so the market was closed on election day leading up to then.

As most of you have probably heard by now, the seasonality surrounding these dates has been exceptionally positive.

The day of and day after have been OK, but it's the period leading up to and immediately after the elections that have been so impressive - partly because of the usual end-of-month / beginning-of-month bullish bias but also probably due to the elections themselves.

Since the elections ceased to become a holiday, the market has suffered a bit more on the day of the election, with generally muted returns, but it has still had an overall positive expectation.

As for the multi-day window surrounding the midterms, we saw a negative return only 2 out of 22 years and overall the median return was an impressive +2.5%.

The "Max Loss" column shows the worst dip (using closing prices) during those multi-day windows, and we see the unusual circumstance where the max loss is actually a positive number - in other words, it would be the equivalent of the S&P not closing below yesterday's close of 1185 anytime between today and next Thursday.  Only 4 of the years saw a loss of more than -1%, and 3 of those were just barely.

The maximum gain, though, was excellent with a median of +3.6%.  There were only two years where the market didn't rally at all, but otherwise 16 out of the 22 had rallies of at least +2% sometime during the next week or so.

I don't want to rely on this too much, as it's only a seasonal bias and it's fairly well-known, but the market has clearly been bucking any bearish influences over the past two weeks so this is one semi-concrete reason to expect it to continue to do so.