In the Abrahamic religions, there is a profound mystery in how to reconcile belief in free will with faith in divine providence. Similar mysteries lie at the heart of political science.

For instance, over the past 45 years, every time there has been a change of party in the White House, the opposing party won the governorship of Virginia a year later.

2021 was no exception. Democrat Joe Biden took the White House, Republican Glenn Youngkin became governor of Virginia. In media circles this outcome was widely described as a ‘shock.’ Given the electoral pattern over the past several decades, it shouldn’t have been.

The outcome of that race was chalked up to debates over “critical race theory” in K-12 schools, among other things. While those narratives may not be wrong exactly, the historical pattern over nearly the last half-century suggests that even if these specific issues had not been salient, some other controversy would have risen to the fore, and the outcome for Democrats would have been roughly the same.

Similar patterns hold at the national level. For instance, every time there is a change of party in the White House, the new incumbent party loses seats in the House of Representatives during the subsequent midterm elections. Here, we can go all the way back to the creation of the Democratic and Republican two-party system, and there are only two exceptions to the rule:

Following the outbreak of the Great Depression, which FDR had been elected to fix, his party actually gained seats in the House and Senate. Likewise, in the immediate aftermath of 9/11, with its unprecedented ‘rally around the flag’ effect in full bloom, George W Bush’s party gained seats in the House and held their ground in the Senate.

In every other instance, the newly-elected party lost seats. It didn’t matter which party was in the White House. It didn’t matter if the economy was up or down or if there was war or peace. It didn’t matter if the president was popular or unpopular or what the big topics of contention were. All of that was immaterial to the basic outcome: the incumbent party experienced a net loss in the House regardless.

The magnitude of those losses varied. At the low end, three seats changed hands. At the high end, 125.

The pathways to those losses varied too: different demographics shifted in different cycles, apparently for different reasons. Different seats changed hands, driven by a wide range of factors. Yet, the net losses occurred like clockwork. The mean loss was 41 seats. The median loss was 26. The modal loss was between 10 and 20 seats.

A majority in the House of Representatives requires 218 seats. Democrats currently have 220. Assuming all seats get filled by the end of this cycle, should Democrats lose more than 2 seats net in the upcoming elections, they will lose their majority in the House. Put another way, Democrats would have to outperform the Lincoln and Kennedy administrations in retaining their seats if they want to keep the chamber.

It seems highly unlikely that they will be able to pull off such a feat. Joe Biden’s net favorability rating is lower than any president on record. Even most Democrats want Biden to retire when his term is over rather than seeking reelection. But it’s not just the president. In the generic ballot, the Democratic Party is also underwater. Democrats saw a slight boost after the supreme court overturned Roe v Wade, but the gains seem to have levelled off, and it’s unclear whether the small bump will sustain over the next four months. But even if it did, again, Democrats would still be net-negative going into the midterms.

It is no wonder that both formal modeling and prediction markets put Republicans as the likely winners of the House by a ratio of nearly 9:1. The Senate is more of a toss-up. There’s a real chance for Democrats to hold that chamber, although it seems like that will be a steep climb too.

If and when Democrats see significant losses in 2022, critics will certainly chalk them up to rising inflation, growing concerns about crime, Biden’s failure to pass most of his Build Back Better proposal, his lack of action on gun control or abortion, his reneging on campaign promises to broadly erase student loan debt, missteps related to Covid or foreign policy, his administration (and the Democrats more broadly) leaning too heavily into “identity politics”, a failure to appropriately leverage the January 6 committee findings against the Republicans, and more.

While these narratives may not be exactly wrong, again, it is likely that the party would have seen losses even if Biden had passed Build Back Better, waived student debt, abstained from ‘woke’ identity politics, and so on. Something else would have worked against the party instead – including, perhaps, their own accomplishments.

Although perceptions of being ineffective can lead to disillusionment, it can also be a liability if an administration is perceived as too effective. If a party rapidly passes a bunch of major legislation or otherwise implements dramatic changes, this often leads to even bigger blowback at the ballot box as Americans try to pump the brakes.

There seems to be a sweet spot of tangible accomplishments that an administration can point to, promises fulfilled on issues that voters prioritize most – where the change is not perceived as too dramatic, or as too many changes happening too fast. Hitting that sweet spot while avoiding self-sabotage can help minimize losses. But it’s generally difficult to discern exactly where that sweet spot is. And again, even for administrations that really seem to get it right, they still see net losses in the House during their inaugural midterms, just smaller losses.

As far as how the likely losses will come about in 2022:

Independent and moderate voters aligned themselves with the Democrats to an unprecedented degree in 2020, playing a pivotal role in Biden’s victory. These voters have shifted dramatically towards Republicans since.

Democrats have been seeing consistent attrition of non-white voters over the past decade. Polling and surveys suggest that these trends are likely to continue among Black, Hispanic and Asian voters.

Democrats also saw significant declines in vote share among younger voters (18-29) in 2020 as compared to 2018. This alienation among young voters seems like it may persist or grow through the midterms as well – although recent supreme court rulings may blunt this a little.

Democratic losses with working-class voters seem likely to continue apace.

In the aftermath of the race, there will be all sorts of stories told about why these voters shifted to the right. Some of those stories may have a lot of truth to them. Many others will be nonsense. But even compelling narratives about the election should be taken with a grain of salt. Again, if these particular groups didn’t shift towards the GOP, other voters would likely have shifted instead. That is, even if they’d retained these voters, the outcome of the midterms may not have changed much.

Losses are pretty much a guarantee for a party’s inaugural midterms. It doesn’t matter who is in charge or what they do, they will still lose seats in the House. How big the losses are, among whom, and why – which specific seats change hands, in which districts – these all remain to be determined by the specifics of the cycle. The broad outcome seems fixed, how we get there is not. In this, political science and theology seem to converge.

• Musa al-Gharbi is a Paul F Lazarsfeld fellow in sociology at Columbia University. His book We Have Never Been Woke: Social Justice Discourse, Inequality and the Rise of a New Elite is forthcoming with Princeton University Press. He is a Guardian US columnist