Divorce Financial Decision Modeling

How this engine works

You enter your own figures; the engine models two scenarios side by side and shows the twelve-month difference between them. The outputs are estimates built from your inputs and documented assumptions — not predictions, and not advice. Mood and stress are self-reported context that adjust the wording of the summary, nothing else. Inputs are processed in your browser. The full methodology, including what the engine does not claim, is on the Decision Center.

Step 2: Decision Forge — compare assumptions

Decision scenarios with reflection prompts

Each scenario in the tool above presents two options drawn from this event and models them side by side from the figures you entered. Before the comparison, the page names a cognitive-bias concept as an educational reflection prompt. It is a general prompt attached to the scenario rather than a finding about you: the page does not test whether the concept applies to your situation, and it does not indicate which option you should choose.

Self-reported context at decision time

The page does not create a psychological profile. Mood and stress may tailor wording and general next-step suggestions. They do not change the entered financial values or scenario math. They do not establish decision readiness. The named bias concept is a general reflection prompt; the page does not detect bias, assess decision capacity, diagnose a condition, or predict outcomes.

The dates a settlement is measured against

A divorce settlement is written once and then measured against dates nobody negotiates. Two of these are counted from the day the marriage ends, one is fixed by the calendar, and one of them is the opposite of what most people are told.

The deadlines that outrank everything else

Most divorce decisions can be revisited. A few run on clocks set by a statute or a plan document, and those are the ones to handle before any modelling is worth much.

What a divorce actually costs, and why no honest page gives you a number

No federal or state statistical agency publishes a median or average cost of divorce. This is structural rather than an oversight: divorce is exclusively a state-court matter, so no federal agency collects it. Every national figure in circulation traces back to self-selected reader surveys run by legal-marketing companies or to divorce-service vendors.

The most-cited of those is a Martindale-Nolo reader survey, and it is worth knowing what it actually found, because the figure that circulates is not it. In that survey full-scope representation averaged $11,300 with a median of $7,000, and 42 percent of respondents paid $5,000 or less. Do-it-yourself divorces averaged $925 with a median of $300; online services ran $150 to $750; mediation ran $3,000 to $10,000. The mean is dragged upward by a small number of contested cases, which is why the median is the more useful number and why the ranges you see quoted — $15,000, $20,000 — describe the tail rather than the middle.

The one figure you can rely on is your own state court’s filing fee, published as a statute or schedule by that state’s judiciary. Look yours up rather than budgeting from a national average that does not exist.

Retirement accounts split by two different mechanisms

This is where the largest avoidable tax mistakes happen, because the two account types use entirely different instruments and the wrong one is expensive.

Publishing or believing that a QDRO lets you take penalty-free money out of an IRA is one of the most common errors in this area, and it combines both mistakes at once.

Alimony, and the rule almost everyone gets backwards

For any divorce or separation instrument executed after 31 December 2018, alimony is not deductible by the payer and not includible in the recipient’s income. Instruments executed on or before that date keep the old treatment, and they keep it indefinitely.

The part that is routinely reported wrongly: modifying a pre-2019 agreement does not by itself flip it to the new rules. The current treatment applies to a modified pre-2019 instrument only if the modification expressly states that the change in law applies. Sites that tell you any modification triggers the switch are wrong, and acting on that belief can cost a payer a deduction they were entitled to keep (IRS Topic 452).

Social Security on an ex-spouse’s record

A benefit many people do not know exists, and a few of its rules are counterintuitive. You may claim on a former spouse’s record if the marriage lasted at least 10 years, you are currently unmarried, you are 62 or older, and your own retirement benefit would be smaller than the spousal amount.

Two features are unique to divorced spouses. If your ex has not yet filed, you can still claim once you have been divorced at least two years and they are 62 or older and fully insured — an exception that does not exist for a current spouse. And your claim does not reduce your ex’s benefit, does not reduce their current spouse’s benefit, and is not subject to the family maximum. There is nothing to negotiate and nothing to disclose (20 CFR 404.331).

How often marriages actually end, and why the familiar number is wrong

The claim that half of marriages end in divorce comes from dividing one year’s divorces by that same year’s marriages — two different populations, which produces a ratio rather than a probability.

Following actual people over time gives a different picture. In the BLS National Longitudinal Survey of Youth cohort, by age 55 about 87 percent had married at least once and 40 percent had divorced at least once; among those who married, 46 percent had divorced at least once. The spread by education is the more useful finding: first marriages ending in divorce ranged from about 27 percent among men with a bachelor’s degree or higher to about 56 percent among women with less than a high school education (BLS, September 2024).

The most recent national rate is 2.4 divorces per 1,000 population for 2023, from 672,502 divorces and annulments. That figure excludes California, Hawaii, Indiana, Minnesota and New Mexico, so it is a 45-state and DC rate rather than a national one — California’s absence alone makes that distinction material (CDC/NCHS). No 2024 or later figure has been published, so any page quoting one is not reporting data.