Imagine a couple in their late fifties. They own a home and rental property, hold RRSPs and report a healthy household net worth as measured by Statistics Canada’s Survey of Financial Security (SFS).
What the data does not tell you? The properties are in his name, the savings are mostly his and she has spent two decades working part time to raise their children. On paper, the household looks fine. In reality, she is poor and he is rich. Canada’s statistics were never built to see the difference.
Canada doesn’t measure wealth at the individual level. It measures it at the household level using what Statistics Canada calls the economic family unit. The result is that the gender wealth gap — the disparity between men’s and women’s accumulated assets net of debt — cannot be quantified.
While the gender pay gap has rightly received decades of attention, the wealth gap remains absent from policy debates because we don’t properly measure it. That must change. Statistics Canada must reform the SFS and begin collecting individual wealth data.
Wealth is not a luxury indicator
Income tells you what someone earns. But wealth tells you what they could survive on if the paycheque stopped. It is also the foundation of retirement security, the collateral behind a business loan, the cushion that lets someone leave a bad job or an unsafe relationship. Gender-equity policy should, therefore, target the wealth gap as directly as it has targeted the wage gap, including by reforming retirement subsidies and family law to protect those — often women — who lack asset ownership.
None of this is possible when the gender wealth gap cannot even be assessed. In Canada’s wealth surveys, including the SFS, couples report their combined non-pension wealth as a single figure. Who owns what is never asked. This masks inequality within unions and, since more than half of Canadians are partnered, it masks a great deal.
Some may argue the fix is simple: divide household wealth equally between partners. But this assumption would be deeply misleading. Couples do not necessarily share wealth. Research I conducted with Hélène Belleau found that in Quebec fewer than three in 10 couples with unequal incomes equalized their retirement savings in 2015. Even the residence was registered to only one partner (usually a man) among more than 20 per cent of couples living in an owner-occupied home.
Aren’t these disparities irrelevant, given that matrimonial property laws mandate the equal split of assets at divorce? They are not. An increasing share of couples are not protected by this measure because they live in common-law unions. And even among married couples, divorce settlements are negotiated, not automatic. The partner with fewer individual assets and less economic leverage often leaves with less.
Matrimonial regimes also usually only apply to non-inherited assets accumulated during the marriage; the growing age at first union and the higher prevalence of intergenerational transfers mean a rising share of household wealth is not covered at all.
The upshot is that property ownership inside modern unions is highly individualized and unequal. Our measures of the gender wealth gap should reflect that reality.
What machine learning reveals
Together with colleagues Diana Peña and Mamadou Diallo, I recently published the first Canadian estimates of the gender wealth gap capturing this hidden inequality. This was achieved not through better data, but by compensating for its limitations with computation.
Using the results of five waves of the SFS conducted between 2005 and 2023, we trained a machine learning model on solo-dwellers and used it to predict the individual wealth of each partner in coupled households. We then combined these estimates with the one domain where the SFS does record individual ownership: employer-sponsored pensions.
The results were striking. In 2023, women’s median non-pension wealth was roughly $28,000 lower than men’s, a gap of about 20 per cent. In wealth percentile terms, women ranked nearly four points below men. That was an improvement from seven points in 2005, but still far from parity.
The gap varied sharply by province. In Alberta, women trailed men by about seven percentile points in 2023, a figure virtually unchanged over two decades. In British Columbia, Quebec and Ontario, the gap was roughly half that.
Pension wealth has largely converged across most provinces, although even that finding is not exactly cause for celebration because it reflects losses among men as much as gains among women.
Have Quebec’s policies helped?
Quebec offers an important test case for policy impact. Since the early 2000s, the province has invested heavily in subsidized child care, parental leave and child benefits, policies shown to improve women’s employment and earnings.
Our analysis suggests these initiatives may also have contributed to a narrowing in the gender wealth gap, particularly among younger cohorts that spent most of their working lives under this framework. However, the evidence is not definitive, since Ontario experienced comparable progress without the same policy set.
One explanation for these mixed results is that it may be too early to fully observe the impact of Quebec’s policies on wealth, which accumulates over decades and even generations. More fundamentally, however, we had to estimate individual wealth because the data do not exist. The inconclusive findings could reflect the uncertainty inherent in estimating individual wealth.
To meaningfully assess the effect of Quebec’s policies, and the broader Canadian trends in gender equity, we need to stop guessing at the size of the gender wealth gap. We urgently need better data.
The fix is straightforward and overdue
Canada does not need to invent a new methodology to solve its data problem. Models already exist internationally.
Germany’s socio-economic panel, for instance, asks about each household member’s individually owned assets and their share of jointly held ones. The UN’s EDGE project goes further, developing guidelines to distinguish legal ownership from the right to use or sell an asset, a crucial nuance when a home is registered in the name of one partner but lived in by both.
Statistics Canada should reform the Survey of Financial Security to collect individual-level wealth data. This would not require rebuilding the survey from scratch, only adding questions about who owns what within the household. The cost of doing so is modest. The cost of not doing so is borne by women whose economic vulnerability our current statistics cannot see.
Closing a gap requires first being able to measure it. Right now, Canada is navigating the gender wealth gap by starlight. It is time to turn on the lights.
