Divorce Rate

Common statistics

When you hear about divorce rates, it is often presented in a certain way. For instance, you might hear that the divorce rate is around 2.9 out of 1000 people getting divorced. These numbers require a bit of comprehension to understand, afterwhich the person reading such a statement may begin questioning, how that statistic is applicable to the average individual looking to get married.

The "2.9 out of 1000 people getting divorced" indicates the number of divorces per 1000 individuals in a given population within a year. It offers a snapshot of divorces relative to the population size.

While a high divorce rate number appears jarring at first, understand that such a number does nothing distinguish the cohort of marriage that it stems from and for what reason that might have arisen. This is especially important considering the duration of said marriages. It also does nothing to illustrate the reasons and the potential changes in said reasons for initiating a divorce.

The most commonly reported annual statistics we can find are typically, duration of marriages that divorced, total number of marriages commenced, age group of brides and grooms (separately), age group of divorcees and perhaps age differential of marriages. None of this is presented in a multifactorial manner, with time being the only common column/variable shared between datasets, making establishing any direct links between divorce and other factors indirect, and therefore, relatively unreliable.

2 original datasets

Combination of divorce duration data and marriage count data from UK's office for national statistics.

Click the image on the right for the edited dataset that is a reformat of the source datasets for the plot.

Population of...

People or Marriages?

The average person, especially men, does not wish to get divorced (and rightfully so). It comes with immense consequences.

For such a person, it is typical for them to want to quantify the probability of divorce and this would reasonably be estimated from the ratio/proportion of marriages that end in divorce. Already, we see a chasm between the information that we regularly receive, versus what a prospective husband might want to know.

Almost expectantly, this data is not readily available to us, and can only be obtained by working backwards, using the duration of marriages prior to divorce reported every year and cross referencing them with the yearly reported number of commenced marriages.

Note that the resolution of the resultant ratio/proportion of marriages ending in divorce is dependent on the size of the duration buckets in the data we work with. In England's case, it is a resolution of 1 year, which is excellent compared to other countries, that typically keep it to a 5 year resolution (except for 10 in America until 2000).

Divorce per marriage cohort

The only useful statistic...

The proportion of a given fixed set of marriages that have ended in divorce is the most, not necessarily the only, useful statistic when it comes to depicting the health of marriages. Divorce rate per population does nothing to untangle the inevitable confluence of marriages from different cohorts.

For example, say about 415 487 marriages that commenced in 1970 - in fact, you can see exactly that on the interactive plot on the left, where the scrubber on the bottom adjusts the cohort of marriage's lifetime over time.

If we were to look at as of 1980, 63 417 of those marriages ended in divorce. In other words, 15.263% of those marriages ended in divorce. By 1990, 26.136% of those marriages had terminated in divorce. By 2000, 33.245% of them had dissolved. Inevitably, and intuitively, an increasing portion of marriages end in divorce with time. This increase inevitably plateaus, and that equivalently, meets common sense.

If you scrub through the graph on the left, or alternatively, the area plot on the right, you will notice that there is also clearly a difference in the manner in which the rate at which marriages divorce. For instance, 25.948% of marriages started in 1963 ended in 30 years (by 1993). This is in contrast to 42.313% of marriages started in 1980 ended in 30 years (by 2010). Compare that again, with the 47.44% of marriages started in 1990 ending in 30 years (by 2020). In other words, the year in which marriages commence has a clear correlation with the success rate of that marriage.

Main takeaway

∼45% and above

With more time, the proportion of marriages that end in divorce increases. This increase generally decreases over time, reaching a state of "equilibrium" so to speak. In other words, the older a marriage, the less likely it is to end in a divorce.

About 45% of marriages in England & Wales end in divorce given the current visible horizon of divorce rates. It has significantly increased, as opposed to the facetious claim by ourworldindata that "since the 1990s, these curves appear to be falling once again, mirroring the findings from the US" whereinwhich the "likelihood of divorce has been falling, and the length of marriage has been increasing". Sadly, it appears that the somewhat obscured nature of these findings allows for unverified claims like this space to fester. It is clear that marriages started in the 1990 have a larger proportion annuled compared to those in 1980s now, with the little time. They even lazily add the unassuming detail that "it will take several decades before we have the full picture on more recent marriages and their eventual outcomes".

This does not mean that there has not been any form of improvement in recent years. Strictly looking at 10 year periods for instance, 20.405% of marriages commenced in 2010, ended by 2020. This is smaller than the 24.526% of year 2000 marriages that terminated by 2010. This is again, smaller than 26.448% of year 1990 marriages that terminated by 2000. The same can be said about the 20 year periods. 39.925% of year 1990 marriages terminated in 20 years, as opposed to 37.078% of year 2000 marriages in 20 years. However, in the same vein, you could also make a remark about the fact that 39.031% of year 2001 marriages ended in 20 years. Does that now mean there has been a statistically significant increase again from the year 2000 to 2001 of cohort of marriages ending in divorce? Probably not.

The reality is that it is more belieavable that the divorce rate for marriages has stabilized to be about 45%, just like New Zealand appears to have. This is unlike America, which very likely has a much higher divorce rate, by virtue of their labour bureau refusing to disclose said figures.

Divorce rates are calculable from 1960 onwards. So the dataset is far more extensive than what we are used to from other countries. Furthermore, the England & Wales dataset is the only dataset so far we have observed to have the most impressive resolution of data available pertaining to divorce, at a 1 year resolution. It allows us to pinpoint the stability of marriages between cohorts by 1 year, instead of the typical 5 year band that other datasets relegate themselves to.

Divorce Rate

Sources

Population of...

Divorce per marriage cohort

Main takeaway

Stat Skim

Why People Like us!

Contact