Vox Meditantis

In 1963, Mollie Orshansky (1915–2006), a statistician working in a cramped office at the Social Security Administration, created a poverty measure so fundamental to American social policy that it became invisible – a piece of taken-for-granted infrastructure. Her simple yet rigorous formula, multiplying the cost of a minimal food budget by three, established the nation’s first official poverty line and determined eligibility for countless assistance programmes across six decades. Yet whilst her measure shaped policy for over sixty years, Orshansky herself remained largely unknown, her authorship erased by the very success of her creation. Today, we sit down with Mollie Orshansky to understand not just the mathematics behind the poverty threshold, but the lived experience, statistical reasoning, and quiet determination that transformed how America counts – and cares for – its most vulnerable citizens.

Welcome, Miss Orshansky. It’s quite extraordinary to be speaking with you. Your poverty thresholds have been used by the United States government since 1965, determining who receives assistance and how we measure economic hardship. Yet most Americans have never heard your name.

Well, that’s how government work goes, isn’t it? You do the job they assign you, you write it up properly, and then it becomes part of the machinery. I was never looking to have my name in lights. I was a civil servant – emphasis on the servant part.

But surely you must have known your work was consequential?

Consequential? I suppose so, though I didn’t think of it that way at the time. In 1963, I was asked to write something about poverty among children for a research project. There was no accepted way to measure who was poor and who wasn’t – just vague notions and political guesswork. The Council of Economic Advisers had drawn this arbitrary line at $3,000 for all families, regardless of whether you had two mouths to feed or eight. That seemed rather daft to me.

You grew up in the Bronx, the third of six daughters in a Ukrainian Jewish immigrant family. Your father worked as a tinsmith, ironworker, plumber – whatever he could find.

Seven daughters, actually. One died young. And yes, my father worked himself to exhaustion. We could “barely make ends meet,” as I’ve said before. We slept two girls to a bed, wore hand-me-downs, stood in relief lines with our mother to get surplus food. You learn things standing in those lines – about dignity, about what it means when the grocer’s bill comes due and there’s no money.

You’ve said, “If I write about the poor, I don’t need a good imagination – I have a good memory.”

That’s right. And it’s true. When you’ve lived it, you don’t romanticise it, and you don’t dismiss it either. You know what it looks like when a family has to choose between shoes for the children or paying the rent. You remember what it feels like to be hungry – not the kind of hungry where you missed lunch, but the kind where you’re not certain when the next proper meal will come.

You were the first in your family to finish high school, then college – Hunter College High School, then Hunter College itself, graduating in 1935 with a degree in mathematics and statistics.

Yes, and thank goodness for scholarships. Hunter College High School was for gifted girls – quite a step up from what we had at home. At Hunter College I had two scholarships that made it possible to attend. Without them, I’d have been working in a factory or shop somewhere, not sitting here talking with you about poverty measurement.

After graduation, you went into government service. Why not academia?

Why not academia? Because universities weren’t exactly throwing open their doors to women, particularly women in mathematics and statistics. Oh, we could teach at women’s colleges perhaps, but opportunities were “largely closed to women,” as they say. So like many women with training in my field, I went where the work was – government bureaus. I started as a research clerk at the Children’s Bureau in 1939, working on biometric studies of child health, growth, and nutrition. Then I went to the New York City Department of Health as a statistician, working on a pneumonia survey. In 1945, I moved to the U.S. Department of Agriculture.

You spent thirteen years at the USDA, working in the Bureau of Human Nutrition and Home Economics. What were you doing there?

Food economics, mainly. I was collecting and analysing data on family expenditure patterns – particularly looking at how much families spent on food relative to their income. This was enormously useful work, though it sounds rather dry when you describe it. We were tracking what people actually ate, what they could afford, how budgets stretched or didn’t stretch. The USDA had these food plans – different cost levels for different budgets – and we’d use survey data to understand consumption patterns.

The 1955 Household Food Consumption Survey was particularly important to your later work.

Oh yes, absolutely. That survey told us that families of three or more spent about one-third of their after-tax income on food. That figure – one-third – became the foundation for everything I did later. But I’m getting ahead of myself.

In 1958, you joined the Social Security Administration. What prompted the move?

It was a good position – social science research analyst in the Office of Research and Statistics. I’d be working on income adequacy, family welfare, patterns of family income – topics I cared deeply about. And frankly, after thirteen years at Agriculture, I was ready for something new.

Your first major assignment there was rather obscure – preparing an answer for a congressional hearing in 1960.

Yes, buried in the files somewhere. A senator had asked Arthur Flemming, the Secretary of Health, Education, and Welfare, how much it costs a retired couple to live. Flemming said HEW would provide an answer “for the record,” and I was the civil servant who anonymously prepared that answer. I actually developed rough measures of income inadequacy for an elderly couple using multipliers from that 1955 USDA survey applied to the low-cost food plan – almost exactly what I’d do three years later for my poverty thresholds. But nobody outside my office ever knew about that 1960 work.

And then came 1963.

Yes. I was asked to do a study on poverty as it affects children. There was suddenly tremendous interest in the subject – President Kennedy had been looking at poverty in Appalachia, there was talk of major initiatives. But nobody had a proper measure. How many poor children were there? How did poverty affect families of different sizes? We simply didn’t know.

So you set out to create one.

I set out to create a measure of income inadequacy – not adequacy, mind you, but inadequacy. I was very particular about that distinction. As I wrote, “if it is not possible to state unequivocally ‘how much is enough,’ it should be possible to assert with confidence how much, on an average, is too little”.

Let’s walk through your methodology step by step – the technical process. You started with food.

Yes, because food was the only essential for which we had generally accepted standards of adequacy. The Department of Agriculture had developed these food plans at different cost levels: liberal, moderate, low-cost, and economy. The first three went back to 1933; the economy plan was newer, introduced in 1961.

Why food specifically?

Because you can measure nutritional adequacy scientifically. The USDA nutritionists could tell you exactly what nutrients a body needs – calories, proteins, vitamins, minerals – and they could design meal plans that met those requirements at various price points. For housing, medical care, clothing, transportation? No such standards existed. You couldn’t say definitively, “This is the minimum adequate level.” But for food, you could.

And you chose the economy food plan rather than the low-cost plan.

For my initial calculations, yes. I actually developed two sets of thresholds – one based on the economy plan, one based on the low-cost plan. I preferred the low-cost version, truth be told. The economy plan was very restrictive – designed for “temporary or emergency use when funds are low,” as the USDA put it. It wasn’t meant for long-term use. It allowed only about one pound of meat per day for a family of four – imagine trying to feed everyone three meals on two chicken breasts. No money for a carton of milk at school, no occasional ice cream bar, no entertaining, not even informal.

But the government adopted the lower threshold.

Yes. The Office of Economic Opportunity chose the economy-level thresholds in May 1965. I suspect they wanted the lower number – fewer people officially counted as poor, less money needed for programmes. I’d even wanted to add fifteen cents per day per person – about a dollar sixteen in today’s money – to allow for small extras. My supervisors wouldn’t have it.

So you had the food costs. Then what?

Then I applied the multiplier. From that 1955 survey, we knew families of three or more spent approximately one-third of their after-tax income on food. So I took a hypothetical average family spending one-third of its income on food and assumed it had to cut back sharply on all expenditures. If expenditures for food and non-food were cut at the same rate, when the food spending reached the cost of the economy food plan, the amount the family would be spending on non-food items would also be minimal but adequate.

Following that logic, I multiplied the dollar cost of the economy food plan by three. If the economy food plan cost, say, $1,033 per year for a family of four in 1963, then three times that – $3,100 – became the poverty threshold for that family.

You differentiated the thresholds by family characteristics.

Oh yes, quite extensively. I created 124 separate poverty thresholds. They varied by family size – from one person to seven persons or more. They varied by the number of children in the family. They varied by whether the head was male or female. They varied by whether the family lived on a farm or not. And for one- and two-person units, they varied by whether the head was aged or non-aged.

Why farm versus non-farm?

Because farm families in 1955 spent less on food than non-farm families of the same size – they could grow some of their own. So I set farm thresholds lower, typically about 70 to 85 percent of the corresponding non-farm thresholds.

And for two-person families and individuals, you used different procedures?

Yes. For two-person families, the multiplier was higher – 3.7 instead of 3 – also derived from the 1955 survey. For one-person units, I derived the thresholds directly from the two-person thresholds without using a multiplier at all. I took 80 percent of the two-person threshold as the one-person threshold, “on the premise that the lower the income, the more difficult it would be for one person to cut expenses such as housing and utilities below the minimum for a couple”.

This was elegant – simple yet empirically grounded.

I tried to make it rigorous without requiring data we didn’t have. My procedure didn’t assume specific dollar amounts for any budget category besides food. Everything else was implicit in the multiplier.

Your initial paper, “Children of the Poor,” appeared in the Social Security Bulletin in July 1963.

That one covered only families with children. The timing turned out to be rather fortunate. President Johnson declared the War on Poverty in January 1964 – just six months after my article appeared. Suddenly policymakers needed a way to measure progress, to count who was poor, to target resources. My thresholds offered exactly that – a concrete, quantifiable tool.

Joseph Kershaw, the research director at the Office of Economic Opportunity, said: “Mollie Orshansky says that when you have more people in the family, you need more money. Isn’t that sensible?”

Well yes, it is sensible, isn’t it? That’s rather the point. The CEA’s flat $3,000 line meant a family of two and a family of eight were judged by the same standard. My thresholds at least acknowledged reality.

Your second major paper, “Counting the Poor: Another Look at the Poverty Profile,” appeared in January 1965. What did that add?

I refined and extended the thresholds to cover all household types – families without children, unrelated individuals, not just families with children. That gave us a complete matrix. And I was bothered by the CEA’s failure to adjust for family size, which had the effect of understating the number of poor children relative to the elderly. That needed correcting.

The OEO adopted your thresholds in May 1965. In August 1969, the Bureau of the Budget made them the federal government’s official definition of poverty.

Yes, though with a critical change I opposed. Officials at the Social Security Administration had wanted to adjust the thresholds over time to reflect changes in the standard of living – as living standards rose, so would the poverty line. The Bureau of the Budget resisted that. Instead, an interagency committee decided in 1969 that thresholds would only be adjusted for inflation using the Consumer Price Index. So the thresholds remained frozen in real terms, tied to 1963 living standards.

You intended them to be temporary, didn’t you?

Oh, absolutely. They were meant as a pragmatic starting point – “arbitrary, but not unreasonable,” as I put it. I never imagined they’d still be in use, essentially unchanged, sixty years later. It’s rather astonishing, actually.

Let’s discuss some of the criticisms. By the 1980s, families were spending about one-fifth of their budget on food, not one-third. By the 1990s, it was down to one-sixth.

Yes, and that’s precisely the problem with freezing the measure. Consumption patterns change. Food became relatively cheaper, whilst housing, medical care, childcare, transportation became relatively more expensive. If you recalculated using current expenditure patterns – food as one-sixth of the budget – the multiplier should be six, not three. Or even higher – some estimates suggest 7.8 by 2008. That would greatly increase the thresholds and the official poverty count.

Critics also point out that your measure doesn’t account for taxes, tax credits, or non-cash benefits like food stamps and housing subsidies.

That’s correct, and it’s a limitation I was aware of from the start. I wanted to measure income adequacy using after-tax income, matching the 1955 survey. But I had to use pre-tax money income because of data limitations in the Current Population Survey. At the time – 1963 – that was a reasonable approximation because few low-income families had federal income tax liability, and in-kind transfers weren’t quantitatively important yet.

But that changed dramatically. The Earned Income Tax Credit, SNAP – the two largest non-health programmes for the poor – are completely invisible to my measure. You can’t assess their effectiveness at fighting poverty if the poverty measure doesn’t capture them.

There’s also no geographic adjustment for cost-of-living differences.

True. A dollar in rural Mississippi buys more than a dollar in Manhattan, but my thresholds treat them the same. I was limited by the data and methods available at the time.

In 2010, the Census Bureau adopted the Supplemental Poverty Measure, which addresses many of these issues. How do you feel about that?

Relieved, frankly. The SPM incorporates what my measure couldn’t – it adds non-cash benefits, subtracts taxes and work expenses and medical costs, adjusts for geographic housing costs. It’s a more comprehensive picture of economic resources. My measure was meant to be updated, improved, replaced when something better came along. The SPM is that something better, at least in many respects.

Though your measure remains the official one.

Yes, bureaucracies are slow to change, aren’t they? The SPM is published alongside the official measure, but it doesn’t determine programme eligibility. That’s still my 1963 formula, updated only for inflation. It’s become infrastructure – invisible, taken-for-granted scaffolding that nobody thinks to question.

Let’s talk about that invisibility. You developed one of the most consequential statistical measures in American social policy, yet you remained “an obscure civil servant working down a dimly lit hall, among stacks of computer print-outs at a paper-covered desk.”

Well, that’s accurate enough. I wasn’t working in some gleaming office with my name on the door. I was a mid-level government statistician doing what government statisticians do – answering questions, analysing data, writing reports. Once my thresholds were adopted by the federal government, they became institutionalised – Census Bureau updates them annually, Health and Human Services issues poverty guidelines, but my name rarely appears. The measure’s very success made it seem timeless and authorless, as though it had always existed.

There’s a gender dimension to this erasure, isn’t there?

Yes, I think so. Women in mid-century economics and statistics faced considerable barriers. Universities were “largely closed to women,” as I mentioned. We ended up in government bureaus where advancement was limited and credit diffuse. I never held an academic position where you build name recognition through publications, conferences, students who carry forward your reputation. And women’s contributions to economics – particularly applied, methodological work – have historically been undervalued compared to “high theory” dominated by men.

Your work was framed as practical and applied rather than theoretical.

Exactly. Critics quickly labelled my approach “simplistic,” which implicitly devalued it as craft rather than science. Never mind that it was empirically rigorous and conceptually sound – it was “mere” measurement, not intellectual achievement. This mirrors broader patterns where women’s statistical work is dismissed as “support” work.

You retired in 1982 after more than forty years of government service. Did you receive recognition for your work?

Some. The Department of Health, Education, and Welfare gave me a Distinguished Service Award in 1976. My colleagues at SSA called me “Miss Poverty” – affectionately, I think. But these were bureaucratic accolades, not the kind that secure broader public recognition.

In 2001, The West Wing had an episode that mischaracterised your methodology, claiming it was “based on life in Poland during the Cold War.”

Oh yes, that nonsense. I’m Ukrainian-American, born in the Bronx. My parents emigrated from Ukraine, not Poland, before I was born. My methodology was based on American consumption data from American surveys conducted by American government agencies. It had absolutely nothing to do with Poland or the Cold War. It’s remarkable how little of my actual biography penetrated public consciousness when even a popular television programme got it so thoroughly wrong.

What would you say to those who argue your thresholds are too low, that they undercount poverty?

I’d say they’re probably right. The thresholds were meant to measure income inadequacy – bare minimums, not comfortable living. And they haven’t kept pace with rising standards. But here’s the thing: measurement itself is a political act. Choosing where to draw the line determines how many people we count as poor, which determines how much money gets allocated to assistance programmes, which determines whose suffering we deem worthy of intervention. My measure made poverty countable, and therefore made it subject to policy action. That was the point.

You once wrote: “There seems sufficient basis for adopting as a working hypothesis that perhaps the single medium most conducive to the growth of poverty and dependency is poverty itself.”

Yes. Poverty breeds poverty – that’s not a mystery. If the children of the poor today are destined to be the impoverished parents of tomorrow, then some social intervention is needed to break the cycle. My measure was meant to help identify those children, those families, so that intervention could happen. Whether it actually did – well, that’s a policy question, not a statistical one.

Looking back, what are you most proud of?

I think I’m proud that I took a messy, contested political question – who is poor? – and gave it a concrete, empirical answer. Was it perfect? No. Was it meant to last sixty years unchanged? Absolutely not. But it worked. It gave policymakers a tool, gave researchers a baseline, gave the country a way to track whether we were making progress or falling behind. And it came from someone who understood poverty firsthand – who’d stood in those relief lines, who’d watched her father work himself ragged for wages that barely covered the rent.

What’s your biggest regret?

That I didn’t fight harder for the low-cost threshold instead of the economy one. I knew the economy plan was too restrictive, meant only for emergencies. But I was a junior analyst, not a department head, and I didn’t have the authority to insist. The government chose the lower number, and millions of people have lived in uncounted hardship because of it. That weighs on me.

You’ve mentioned mistakes. What else would you do differently?

Well, I’d have pushed harder for after-tax income measures from the beginning, even if it meant waiting for better data. And I’d have built in automatic adjustments for changing expenditure patterns, not just inflation. But honestly, these are things I recognised as limitations at the time – I wrote about them. The real mistake was made by those who froze my measure in place and treated it as holy writ for six decades.

What would you say to modern statisticians and policy analysts – especially women or those from marginalised groups – who are doing this kind of work today?

First, don’t wait for perfect data or perfect methods. You’ll never have them. Do the best, most rigorous work you can with what you’ve got, but don’t let the perfect be the enemy of the good. Second, be clear about your assumptions and limitations. I was very explicit that my thresholds were “arbitrary, but not unreasonable” – I never pretended they were the final word. Third, remember that measurement is never neutral. Every choice you make – what to count, how to count it, where to draw the line – has political consequences. So make those choices thoughtfully and defend them clearly.

And for women specifically?

Know that your work may be dismissed, undervalued, or simply made invisible even if it’s successful – perhaps especially if it’s successful. Government work, applied work, methodological work – these don’t win you the recognition that theoretical work or academic positions do. But that doesn’t make them less important. Someone has to build the infrastructure, even if nobody remembers who wielded the trowel.

You said your poverty thresholds came from memory as much as imagination – from your childhood in the Bronx, standing in relief lines.

Yes. And I think that’s important. Lived experience is a form of expertise, even if it’s not always recognised as such. The academy privileges abstract theory, but embodied knowledge – knowing what it feels like to be hungry, to choose between necessities, to watch your parents struggle – that informs the work in ways theory alone cannot. I understood what I was measuring because I’d lived it.

There’s something almost poetic about the fact that your name is less known than your number – that the infrastructure you built became so fundamental it erased you.

Poetic, perhaps. Frustrating, certainly. But also, in a strange way, appropriate. Infrastructure should be invisible when it works properly. You don’t think about the engineers who designed the water system when you turn on the tap, do you? My measure became the plumbing of social policy – essential, unglamorous, taken for granted. I’m not sure that’s entirely a bad thing, even if I’d prefer to be remembered.

What do you want people to understand about poverty measurement today?

That it matters enormously and it’s never neutral. When we count poverty, we’re deciding who deserves help and who doesn’t. We’re making suffering legible or invisible. We’re creating categories that shape reality, not just describing it. And we’re embedding value judgements – about what counts as need, what counts as income, what counts as an acceptable standard of living – into what looks like objective mathematics. None of that is neutral. All of it is political.

And about your own legacy?

I’d like people to know that the poverty line – that abstraction, that number – was created by a person. A woman who grew up poor, who understood what poverty meant in her bones, who spent forty years in government service trying to make the world a bit more just. I gave poverty a number so we could count it, track it, fight it. Whether we actually fought it hard enough – that’s not for me to say. But at least we knew where to look.

Miss Orshansky, thank you.

You’re welcome. And call me Mollie – we’ve been talking long enough.

Letters and emails

Since our conversation with Mollie Orshansky, we’ve received dozens of letters and emails from readers around the world – economists, students, policy analysts, and those simply curious about the woman behind the numbers. We’ve selected five particularly thoughtful questions from our community, each offering fresh angles on her methodology, her choices, and the tensions inherent in measuring human suffering from within the very institutions meant to address it.

Madison Greene, 34, Data Scientist, Toronto, Canada
You mentioned using the 1955 Household Food Consumption Survey as your empirical foundation, but survey methodology in the 1950s had significant limitations – sampling frames, non-response bias, recall accuracy for expenditure data. How confident were you in the quality of that underlying data, and did you ever run sensitivity analyses to test how measurement error in the food expenditure ratios might affect your thresholds? I’m curious whether you had access to standard errors or confidence intervals around that “one-third” figure, or if you were working with point estimates and essentially making an informed bet.

Ah, now that’s the sort of question I wish more people had asked at the time. You’re absolutely right to be suspicious of that data – the 1955 Household Food Consumption Survey had all the limitations you’d expect from a cross-sectional survey conducted in the Eisenhower era. We didn’t have the sophisticated sampling techniques you have now, and yes, expenditure recall was a real problem. Ask someone what they spent on groceries three months ago and you’ll get approximations at best, fabrications at worst.

But here’s the thing: it was the best national data we had. The survey covered about 4,000 households – not enormous by today’s standards, but reasonably representative for the time. The Agriculture Department’s statisticians knew what they were doing. They’d been running these consumption surveys since 1935, refining their methods with each iteration. By 1955, they had the questionnaires down to a science – well, as much of a science as you could manage with clipboard-wielding enumerators going door to door.

Now, did I have standard errors? Yes, the published survey results included them, though they weren’t always as precise as I’d have liked. That “one-third” figure – families spending roughly 33 percent of after-tax income on food – wasn’t a single point estimate. It varied somewhat by income level, by family size, by region. For families in the middle-income ranges, it held fairly steady around 32 to 35 percent. For very poor families, the proportion was higher – they might spend 40 or even 50 percent on food because you can’t cut food consumption below a certain minimum, but you can scrimp on everything else. For wealthy families, it dropped to 20 percent or less.

I chose to use the average for families of three or more because I was trying to establish what a family should spend on food if they had an adequate income – not what desperately poor families were spending. That’s an important distinction. I wasn’t describing current spending patterns among the poor; I was inferring backward from the spending patterns of families with adequate resources.

Did I run formal sensitivity analyses? Not in the way you’d do today with Monte Carlo simulations or bootstrapping – those techniques weren’t standard practice in government statistics offices in 1963. But I did informal checks. I calculated thresholds using both the economy food plan and the low-cost food plan, which gave me a range. I looked at what happened if I adjusted the multiplier up or down by half a point. I compared my results to other rough measures people had proposed – the Council of Economic Advisers’ $3,000 line, the Bureau of Labor Statistics’ “modest but adequate” family budgets. My thresholds fell in a plausible range relative to those benchmarks.

The real limitation wasn’t statistical uncertainty – it was conceptual. I was making a logical leap: assuming that families cut back proportionally on all expenditures when money got tight. But that’s not really how people behave, is it? You might cut food spending from 33 percent to 40 percent of income while trying desperately to keep the rent paid, or you might let the rent slide and feed the children properly. Human behaviour doesn’t follow neat mathematical proportions.

I acknowledged this explicitly in my 1965 paper. I wrote that the methodology “does not attempt to describe the actual spending patterns of low-income families” – because those patterns reflect deprivation, not adequacy. What I was trying to do was construct a normative standard: if a family needs X dollars for minimally adequate food, and if food represents one-third of a minimally adequate budget, then the family needs 3X dollars total.

Was I confident in that approach? Reasonably so. Was it perfect? Absolutely not. But in the absence of detailed budget standards for housing, clothing, medical care, and everything else – standards we simply didn’t have in 1963 – the food-multiplier method was elegant, defensible, and transparent. Anyone could understand the logic, even if they quibbled with the specific numbers.

The alternative would have been to wait another decade for better data, more sophisticated methods, more complete budget standards. And in the meantime, President Johnson would have launched the War on Poverty with no empirical measure at all, or with something far cruder cooked up by political appointees more interested in minimising the count than measuring reality. I made what I call an “informed bet,” as you put it – and I’d make the same bet again.

Though I’d certainly welcome your bootstrapped confidence intervals if you’d like to apply modern methods to that 1955 data. Might be an interesting paper in that.

Oliver Jensen, 41, Social Policy Researcher, Copenhagen, Denmark
Europe largely adopted relative poverty measures – defining poverty as a percentage of median income rather than an absolute threshold. You were aware of this alternative approach, yet you chose an absolute measure anchored to nutritional adequacy. Looking back with decades of comparative data, do you think the United States would have addressed poverty more effectively with a relative measure that automatically rises with national prosperity? Or does the absolute approach – however flawed – create clearer moral accountability by establishing a fixed standard of deprivation we committed to eliminating?

You’ve put your finger on one of the fundamental debates in poverty measurement, and it’s one I thought about quite a bit, particularly after I learned how the Europeans were approaching the problem. The relative measure – defining poverty as, say, 50 or 60 percent of median income – has an appealing logic. As a society grows richer, its definition of deprivation rises too. What counts as poverty in Denmark today would have been middle-class comfort in 1920s Copenhagen, yes?

But here’s where I part ways with that approach, at least for the American context in the 1960s. An absolute measure, anchored to something concrete like nutritional adequacy, establishes a floor – a line below which we say, as a matter of moral fact, “This is unacceptable. No one should live like this.” It’s not relative to what your neighbour has; it’s about whether you have enough to eat, a roof that doesn’t leak, shoes without holes. That kind of deprivation doesn’t become less serious just because everyone else is getting richer.

I remember reading some of the European literature – there was a British sociologist, Peter Townsend, who was a great advocate for relative measures. He argued that poverty is fundamentally about exclusion from ordinary social life, and that changes as society changes. Fair enough. But in 1963 America, we had children going hungry in Appalachia, elderly people choosing between medicine and food, families living in conditions that would have shocked Charles Dickens. Those weren’t problems of relative deprivation – of feeling poor compared to wealthier neighbours – they were problems of absolute deprivation. People lacking the basic material necessities for survival and human dignity.

A relative measure would have told us that as the country got richer through the 1960s and 1970s, we still had the same proportion of people in poverty – say, 15 percent – even if every single person’s material circumstances improved. To me, that would have obscured real progress. If we lifted millions of families out of genuine material hardship but the relative measure stayed flat because the median rose, would we really say we’d accomplished nothing?

Now, I’ll grant you this: the absolute measure has its own perverse incentives. Once you’ve established that threshold – this many dollars for this family size – there’s enormous political pressure to keep it low and not let it rise with general prosperity. Which is exactly what happened. My thresholds were frozen in real terms, adjusted only for inflation, not for rising living standards. A family at the poverty line in 1963 could afford roughly the same market basket of goods as a family at the poverty line in 1990, even though median families in 1990 had far more – second cars, colour televisions, things we couldn’t have imagined in 1963.

So in practice, my absolute measure did drift toward becoming a kind of relative measure, just a very slowly adjusting one. The poverty line stayed at about the same percentage of median income – roughly 40 to 50 percent – not because policymakers intended it that way, but because they updated it so minimally. That’s probably the worst of both worlds: the rigidity of an absolute measure combined with the gradual obsolescence that comes from not acknowledging changing standards.

If I were designing the system from scratch today – knowing what I know now, having watched six decades of policy fights – would I choose differently? Maybe. A hybrid approach might work: an absolute floor based on genuine deprivation, combined with a relative component that rises with median income. You could say, “The poverty line shall be the higher of: (A) the cost of basic necessities calculated annually using current prices and consumption patterns, or (B) 50 percent of median household income adjusted for family size.” That way you maintain the moral clarity of an absolute standard whilst preventing the threshold from becoming completely disconnected from contemporary life.

But here’s what I keep coming back to: the question isn’t really absolute versus relative. It’s whether you’re serious about addressing poverty at all. The Scandinavian countries use relative measures and have robust social safety nets that keep poverty rates low by their own definitions. America used my absolute measure and let millions of people fall through the cracks because we never properly funded the War on Poverty, never built comprehensive child allowances or universal health care, never treated poverty as a solvable problem rather than a personal failing.

The measure matters, certainly. But what matters more is the political will to use whatever measure you’ve got as a tool for action, not just counting. And on that score, I’m afraid America has disappointed me more than I can say. We counted and counted and counted – every year, the Census Bureau publishes new poverty statistics using my thresholds – but we never truly committed to ending what we were counting.

So would a relative measure have made us more accountable? I honestly don’t know. Different tool, same lack of will. The Europeans didn’t reduce poverty because they measured it differently; they reduced it because they built welfare states that treated economic security as a right, not a favour. The measurement followed the values, not the other way round.

Yuki Nakamura, 28, Economics Graduate Student, Tokyo, Japan
I’m fascinated by your decision to create 124 distinct thresholds rather than a simpler model with fewer categories. That level of granularity must have been computationally intensive in 1963 – were you doing calculations by hand, using mechanical calculators, or did you have access to early mainframe computers at the SSA? And more conceptually, how did you decide where to stop differentiating? You distinguished farm from non-farm, male-headed from female-headed households, but you didn’t create separate thresholds by race, region, or employment status. What principles guided those boundary decisions?

Oh, you’ve asked about the practical side – the actual work of it. Yes, 124 thresholds was quite a lot, and you’re right that it was computationally intensive by 1963 standards. We did have access to computers at the Social Security Administration – IBM mainframes, punch cards, the whole apparatus – but they weren’t sitting on my desk. You’d submit your programme on punch cards to the computing centre, wait your turn in the queue, and hope you hadn’t made an error that would waste your allotted machine time. It could take days to get results back.

For the initial calculations, I did much of it with a desk calculator – one of those heavy electric machines that made a terrible racket – and plenty of graph paper. The mathematics itself wasn’t complicated: take the cost of the economy food plan for a family of a given size and composition, multiply by three, and there’s your threshold. But when you’re doing that for 124 different family configurations, checking your arithmetic, making sure you haven’t transposed any figures – well, it’s tedious work. I had help from other analysts in the office, people running the numbers, double-checking, preparing the tables for publication. But the logic, the structure, the decisions about which categories to include – that was mine.

Now, why 124 categories rather than, say, a dozen? It came down to what data we had and what policy questions we needed to answer. The USDA had developed separate food plans for different family sizes – one person, two persons, three, four, and so on up to seven or more – and for different age groups, because teenagers eat more than toddlers. They’d also distinguished male-headed from female-headed families, farm from non-farm, aged from non-aged for smaller households. These distinctions reflected real differences in food costs and, by extension, total budget needs.

I could have simplified – created just one threshold per family size, for instance – but that would have obscured important variations. A family of four with a working-age couple and two teenagers needs more income than a family of four with young children. A farm family in 1963 could still grow some of their own food, gather eggs, keep a kitchen garden – they genuinely needed less cash income than an urban family buying everything at the supermarket.

As for where I stopped differentiating – why not separate thresholds by race, region, employment status – that’s where things get more complicated. The principle I followed was this: I only created separate thresholds when there was a demonstrable difference in the cost of achieving the same standard of living, and when we had good data to support that difference.

Farm versus non-farm? The 1955 survey showed clear differences in food expenditure patterns. Male-headed versus female-headed households? Women earned less than men for the same work – that’s a fact – but the cost of living wasn’t different. A loaf of bread cost the same whether a man or woman bought it. The reason I distinguished these households was because female-headed families had different demographic characteristics – typically fewer earners, different age distributions – that affected the food plan calculations.

Race? Here’s where I made what some might call a political decision, though I thought of it as an ethical one. Black families and white families faced vastly different economic circumstances in 1963 – discrimination in employment, housing, education. But they didn’t face different costs for basic necessities. If anything, Black families often paid more – living in segregated neighbourhoods where grocery stores charged higher prices, landlords extracted higher rents for substandard housing. To create a lower poverty threshold for Black families would have legitimised that discrimination, suggesting they needed less income because they were expected to live in worse conditions. I wasn’t about to do that.

Regional differences? This one bothered me, I’ll admit. The cost of living in New York City was clearly higher than in rural Mississippi – housing particularly. But we didn’t have reliable regional price indices for all the components of a minimal budget. The Bureau of Labor Statistics had city-by-city price data for some items, but not comprehensive family budgets. Creating regional thresholds without solid data would have been guesswork dressed up as precision. And there was a real danger that Southern states would use lower regional thresholds to justify lower benefit levels, essentially perpetuating regional poverty. So I made a judgement: better to have a single national standard, even if it was imperfect, than to open that particular can of worms.

Employment status – whether the family head was employed, unemployed, disabled – didn’t affect the cost side of the equation, only the income side. The poverty threshold is about what you need, not how you get it. A family needs the same amount to live on whether the breadwinner is working, looking for work, or unable to work. What varies is their ability to reach that threshold, and that’s a different policy question.

The truth is, every boundary I drew was a compromise between analytical purity, data availability, and political reality. I could have created 500 thresholds, accounting for every possible variation. But at some point, you’re making distinctions that don’t meaningfully improve accuracy – you’re just adding complexity. And complexity, in government statistics, usually means more opportunities for errors, more difficulty explaining the measure to the public, more ways for politicians to manipulate the numbers.

So I stopped where the data stopped being reliable and the distinctions stopped being meaningful. One hundred and twenty-four categories was probably at the upper limit of what was useful. Looking back, I think the farm/non-farm distinction became obsolete quickly – by the 1970s, very few farm families were growing their own food – but the rest held up reasonably well.

Though if I’d known the measure would still be in use sixty years later, I might have spent another six months getting it even more precise. Then again, maybe not. Precision isn’t much good if you never actually finish the work.

Mateo Silva, 38, Community Organiser, São Paulo, Brazil
You grew up in poverty and then spent your career measuring it from inside government bureaucracy – essentially translating lived experience into administrative categories. That’s a profound and potentially painful position to occupy. Did you ever feel complicit in creating a system that made poverty more manageable for the state whilst potentially making it more bearable rather than intolerable? I’m thinking about how quantification can domesticate crisis – turning moral outrage into spreadsheet cells. Did that tension ever keep you awake at night, or did you see your work as genuinely emancipatory?

That’s… that’s a hard question. The hardest one anyone’s asked me, I think. Yes, it kept me awake. Not every night, but often enough.

You’re asking whether I domesticated poverty – made it manageable for bureaucrats rather than intolerable for the conscience. Whether by giving deprivation a number, I made it easier for people to live with the fact of it. Whether I became complicit in maintaining the very system I was trying to expose.

I’ve thought about this for forty years. Here’s what I come back to: before I created those thresholds, poverty was invisible in a different way. It was everywhere and nowhere. Politicians could claim we’d nearly eliminated it, or that it was endemic and unsolvable, and there was no empirical basis for saying either was wrong. Without measurement, poverty was abstraction – “the poor,” spoken of in vague, sentimental terms or dismissed as moral failures. Individual people suffering individual tragedies, not a structural problem demanding structural solutions.

What measurement did – what my thresholds did – was make poverty legible to the state. And yes, legibility is a double-edged sword. It enables intervention, but it also enables management. Once you can count something, you can set targets, allocate budgets, declare victory when the numbers go down even if nothing fundamental has changed. The War on Poverty became a war of attrition measured in percentage points, not a moral crusade to remake society.

Did I foresee that? Partly. I knew the measure would be used politically – of course it would. Every administration would massage the numbers, argue about methodology, try to show progress whether or not real progress existed. But I thought – I hoped – that having concrete numbers would make it harder to ignore. That knowing “23 percent of children live below the poverty line” would be more galvanising than knowing “many children are poor.”

I’m not sure I was right about that.

Here’s the tension I lived with: I grew up poor. Not genteel poverty, not temporary setbacks – real, grinding, humiliating poverty. Standing in relief lines with my mother, watching her face when she had to ask for help. Watching my father work himself sick for wages that barely covered rent. Sleeping two to a bed in a crowded tenement. I know what poverty feels like in your bones, in your stomach, in the way it makes you small and ashamed even when the shame belongs to the society that permits such things, not to you.

And then I became a civil servant. I put on my good dress, took the subway to my government office, sat at my paper-covered desk, and turned that lived experience into spreadsheet cells. I translated hunger into “food expenditure ratios” and deprivation into “income inadequacy thresholds.” I made poverty into something you could print in the Federal Register and file away in Bureau archives.

Was that betrayal? Or was it translation – taking knowledge that came from experience and making it intelligible to people who’d never stood in a relief line, never chosen between shoes for the children and milk for the week?

I’ll tell you what I think. I think the problem isn’t that I quantified poverty – it’s that we only quantified it. We counted and counted and never acted with the urgency the numbers demanded. My measure told policymakers that millions of children were growing up in deprivation. And what did we do? We created some programmes, yes – Head Start, food stamps, Medicaid – but we never built the comprehensive welfare state that would have actually eliminated the poverty we were so carefully counting.

The Europeans looked at their poverty statistics and built universal child allowances, national health services, generous unemployment insurance, subsidised housing. We looked at our poverty statistics and blamed the poor for their own circumstances. We turned the measurement into an industry – poverty researchers, poverty conferences, poverty think tanks – whilst actual poor people kept being poor.

So yes, I feel complicit. But not because I created the measure. I feel complicit because I kept working within a system that used my measure as a substitute for action rather than a spur to it. Every year I’d see the new numbers – poverty up, poverty down, child poverty persisting at unconscionable levels – and I’d write another paper, attend another conference, suggest another methodological refinement. As though better measurement would somehow force change.

You’re from São Paulo – you understand what I’m talking about. Favelas aren’t invisible. Everyone can see them. Measurement doesn’t create awareness; it creates the illusion of knowledge without the necessity of solidarity. “We know X percent live in substandard housing” – and then what? If you don’t build the housing, what good was the knowing?

But here’s what I hold onto: measurement is necessary even if it’s not sufficient. You can’t address what you can’t see, even if seeing alone doesn’t compel action. Before my thresholds, politicians could deny poverty existed on any significant scale. After my thresholds, they had to confront the numbers. Some chose to manipulate them, some chose to ignore them, but they couldn’t pretend ignorance.

And for some people – not enough, but some – my measure did enable help. Social Security Administration caseworkers determining benefit eligibility, community organisations identifying families who qualified for assistance, legal aid lawyers arguing that programme funding was inadequate given documented need. For those individual families who got help they wouldn’t have gotten otherwise, the measurement mattered.

Did it make poverty more bearable rather than intolerable? Maybe. Maybe by breaking it down into manageable categories – family size, farm status, income thresholds – I made it easier for people to file it away as someone else’s problem. But the alternative – leaving poverty unmeasured, unquantified, spoken of only in abstract moral terms – that didn’t seem to be working either.

What I’ve learned is this: you can’t solve poverty from inside a government bureau with statistics and methodology papers. But you also can’t solve it without the data to make the case, to track progress, to hold power accountable. It’s necessary infrastructure, not sufficient transformation.

The revolution you’re implying I should have sparked? That has to come from organisers like you, from people in the streets, from political movements that demand not just better counting but actual redistribution. I gave you the numbers. What you do with them – whether you let them domesticate your rage or sharpen it – that’s up to you.

I hope you choose to sharpen it.

Aissatou Diarra, 45, Public Health Administrator, Dakar, Senegal
What if President Johnson had launched the War on Poverty two years earlier, before your thresholds existed – would policymakers have created something cruder and more politically convenient that might have been even harder to dislodge? Or alternatively, what if the feminist economics movement of the 1970s had gained influence earlier and you’d had collaborators pushing you to account for unpaid domestic labour, intra-household resource distribution, and women’s economic vulnerability within your original framework? Could your 1963 methodology have incorporated those dimensions, or would the data simply not have supported it?

Those are two very different “what ifs,” aren’t they? Let me take them one at a time.

If the War on Poverty had started in 1962 instead of 1964 – before I’d published my thresholds – I suspect you’re right that we’d have ended up with something cruder. The Council of Economic Advisers had already drawn that $3,000 line for all families regardless of size, and it was fairly arbitrary – more political round number than empirical analysis. If Johnson had needed a poverty measure immediately, that’s probably what they’d have used, or something equally expedient.

And here’s the thing: once a measure gets embedded in legislation, in programme regulations, in budget formulae, it becomes devilishly hard to change. There are constituencies that form around it – programme administrators who’ve built their eligibility systems around it, researchers who’ve built longitudinal datasets using it, politicians who’ve staked positions on it. Even when everyone agrees a measure is flawed, the transaction costs of replacing it are enormous.

So yes, we might well have been stuck with something worse – a flat threshold that didn’t adjust for family size, or one set so low it excluded most genuinely poor families, or one that varied state by state based on whatever the local political economy would tolerate. My measure had problems, God knows, but at least it was grounded in nutritional science and household expenditure data. At least it acknowledged that a family of eight needs more than a family of two.

In that sense, maybe the timing was fortunate – Johnson needed a measure, and mine appeared at just the right moment to be adopted before something cruder got locked in. Though I’ll note that “fortunate” feels like an odd word when we’re talking about millions of people’s welfare hanging on bureaucratic happenstance.

Now, the feminist economics question – that’s more complicated, and it cuts deeper. Could my 1963 methodology have incorporated unpaid domestic labour, intra-household inequality, women’s particular economic vulnerabilities? The honest answer is: not easily, and probably not well.

The data simply didn’t exist in any usable form. We weren’t measuring unpaid work – housekeeping, childcare, eldercare – in national surveys. Nobody was collecting information on how resources were distributed within households, whether women had independent access to income or whether they had to ask their husbands for grocery money. The household was treated as a black box: income comes in, consumption happens, and we assume everyone inside benefits equally. That assumption was obviously false, but we had no empirical basis for doing better.

Even conceptually, I’m not sure how I’d have incorporated those dimensions into a poverty threshold. The threshold is meant to measure the cost of a minimally adequate standard of living. Unpaid domestic labour isn’t a cost – it’s labour that reduces costs by producing services at home rather than purchasing them in the market. A family that cooks at home instead of buying restaurant meals, that mends clothes instead of buying new ones, that cares for children instead of hiring daycare – that family needs less cash income to achieve the same standard of living precisely because someone, usually the wife and mother, is providing those services without pay.

Should we value that labour? Absolutely. Should it count in GDP? Certainly – as you say, it’s enormous, perhaps 10 to 39 percent of total economic product. But incorporating it into a poverty measure isn’t straightforward. Do you impute a value for that labour and add it to household income? Then you’d be saying a poor family with a full-time homemaker is richer than one where both parents work low-wage jobs – which doesn’t capture the reality that the two-earner family might be worse off because they’re paying for childcare and have less time for home production.

Where the feminist critique really lands, though, is on intra-household distribution. My thresholds assumed that if a household had income above the poverty line, everyone in it was non-poor. But what if the husband controls all the money? What if he spends freely on himself while the wife and children go without? What if divorce or separation leaves a woman with no independent income even though the household she’d been living in was comfortably middle-class?

I knew this was a problem. Every woman who’s lived through economic dependence knows it’s a problem. But I didn’t know how to measure it with 1963 data, and I wasn’t sure a poverty threshold – which is fundamentally about household resources – was the right tool for addressing it. That’s a question of power and distribution within families, and it requires different policy interventions: women’s access to employment, equal pay laws, property rights in divorce, independent benefit entitlements.

If I’d had feminist economists as collaborators in 1963 – well, I’m not sure who that would have been. There were women economists, certainly, but feminist economics as a distinct field didn’t really coalesce until the 1970s and 1980s. Barbara Bergmann comes to mind, though she was working on labour market discrimination, not poverty measurement. Nancy Barrett, maybe. But the intellectual frameworks for thinking about unpaid work and intra-household bargaining – those came later.

If I’d had those frameworks, if I’d had the data, here’s what I might have tried: creating individual-level poverty measures rather than household-level ones. Estimate each person’s share of household resources based on observable characteristics – their earnings, whether they’re the primary breadwinner, their age and dependency status. Assign higher thresholds to women with children, lower thresholds to men without dependents. It wouldn’t have been perfect, but it would have acknowledged that poverty risks differ within households.

Or I might have developed separate thresholds for female-headed households that accounted for their particular vulnerabilities – lower earning potential, higher childcare costs, less access to credit. In fact, I did differentiate female-headed from male-headed households in my 124 categories, but only by demographic characteristics, not by economic vulnerability.

But here’s what troubles me about the second part of your question: you ask whether my methodology “could have” incorporated these dimensions, and I think the answer is less about technical capacity than about political will. Even if I’d proposed measures that captured women’s unpaid labour or intra-household inequality, would the Social Security Administration have published them? Would the Office of Economic Opportunity have adopted them? Would Congress have legislated programme eligibility based on them?

I doubt it. In 1963, married women needed their husbands’ permission to open bank accounts, to get credit cards. The idea that we’d measure their poverty independently, acknowledge their economic vulnerability within intact marriages – that would have been seen as threatening to family stability, to traditional roles. I’d have been told to simplify, to stick to the household unit, to avoid controversial questions about marital dynamics.

So perhaps the better question isn’t whether I could have done it technically, but whether the system would have tolerated it politically. And I suspect the answer is no. Which means that even if feminist economics had emerged earlier, even if I’d had those conceptual tools, the measure that got adopted would still have treated the household as the unit of analysis and assumed equal distribution within it.

That’s not to excuse my limitations – I made choices based on what was feasible, and feasibility is shaped by power. But it’s to say that measurement doesn’t happen in a vacuum. What we count, how we count it, what distinctions we make – all of that reflects and reinforces existing social structures. My thresholds reinforced the household as the basic economic unit, implicitly supporting the male-breadwinner model even as that model was already breaking down.

A truly feminist poverty measure would have challenged that. But I’m not sure 1963 America was ready for it, and I’m not sure I was bold enough to try.

Reflection

Mollie Orshansky died on 18th December 2006, at the age of 91, in a care facility in the Bronx – returning, in her final years, to the borough where her story began. Her passing received modest notice: obituaries in the Los Angeles Times and New York Times, acknowledgements from the Social Security Administration, brief mentions in policy circles. For a woman whose formula had shaped six decades of American social policy, determined billions of dollars in programme spending, and touched the lives of millions seeking assistance, the recognition felt muted – another echo of the infrastructural invisibility that defined her career.

Throughout our conversation, certain themes emerged with striking clarity. Orshansky embodied a particular kind of ingenuity born from constraint – the ability to solve complex problems with limited tools, imperfect data, and minimal institutional support. Her poverty thresholds were elegant precisely because they had to be: simple enough to calculate with 1960s computing power, transparent enough for public understanding, rigorous enough to withstand scrutiny. She turned the economy food plan and a single multiplier into a measure that outlasted most of the policies it was meant to inform. That resourcefulness – making do with what you have whilst knowing it’s not enough – runs through much of women’s work in mid-century STEM fields, where access to laboratories, computing resources, and professional recognition remained stubbornly constrained.

Her reflections on complicity and measurement cut particularly deep. The tension between making poverty legible to the state and potentially domesticating moral outrage – between counting suffering and compelling action – remains unresolved in contemporary policy analysis. Orshansky never fully answered whether quantification served emancipation or management, perhaps because the answer depends less on the measure itself than on the political will to use it. Her acknowledgement that “measurement is necessary even if it’s not sufficient” offers a more nuanced position than celebratory narratives about data-driven policy usually permit.

What emerged most powerfully was her insistence that lived experience constitutes expertise. “If I write about the poor, I don’t need a good imagination – I have a good memory” – that single sentence encapsulates an epistemological challenge to hierarchies that privilege abstract theory over embodied knowledge. Orshansky’s childhood poverty wasn’t incidental to her work; it was foundational, informing her understanding of what deprivation meant in ways survey data alone could never capture. Yet that experiential authority remained unrecognised by institutions that valued academic credentials and theoretical contributions above applied, methodological innovation.

Where might our fictional Orshansky have diverged from the historical record? The real Mollie Orshansky, judging from her published writings, maintained the measured, careful tone expected of government statisticians – acknowledging limitations, avoiding overstatement, framing her work as provisional and subject to improvement. Our conversation gave her permission to speak more candidly about frustrations that the archival record only hints at: her preference for the low-cost food plan over the economy plan, her awareness that freezing thresholds would render them obsolete, her understanding that gender constrained her career prospects. We granted her the retrospective clarity that comes from watching sixty years of policy choices, seeing how her temporary measure calcified into permanent infrastructure, witnessing the Supplemental Poverty Measure finally address limitations she’d acknowledged from the start.

Some interpretations remain contested. Did Orshansky intentionally avoid creating race-based thresholds as an ethical stance against legitimising discrimination, or was that a later rationalisation? Her published work doesn’t explicitly address the question, leaving room for speculation. Similarly, the extent to which she experienced gender-based barriers versus general bureaucratic constraints is difficult to disentangle from historical sources – though the broader pattern of women economists being channelled into government service rather than academic positions is well documented.

The afterlife of Orshansky’s work reveals both influence and erasure. Her thresholds became the official poverty measure in 1969 and remain so today, updated annually by the Census Bureau – arguably one of the most durable statistical innovations in twentieth-century social policy. Researchers across economics, sociology, demography, and public health have built careers analysing poverty trends using her measure. Yet citations often reference “the official poverty line” or “Orshansky thresholds” without biographical detail, treating the measure as depersonalised methodology rather than individual achievement.

The 1995 National Academy of Sciences panel report Measuring Poverty: A New Approach, which laid groundwork for the Supplemental Poverty Measure, explicitly acknowledged Orshansky’s contribution whilst recommending substantial revisions. The SPM, introduced in 2010, represents the evolution she anticipated – incorporating taxes, non-cash benefits, geographic adjustments, and updated expenditure patterns. This simultaneous honouring and superseding captures a common fate for methodological pioneers: their work becomes infrastructure to be improved rather than achievement to be celebrated.

Contemporary poverty measurement continues grappling with questions Orshansky raised. Should thresholds be absolute or relative? How do we account for non-cash resources and in-kind transfers? What role should regional cost-of-living variation play? The OECD, World Bank, and various national statistical agencies have developed competing approaches, yet the fundamental challenge remains: translating multidimensional deprivation into quantifiable categories that enable policy intervention without reducing human suffering to administrative convenience.

Her multidisciplinary approach – combining nutritional science, household budget analysis, statistical methodology, and demographic analysis – prefigured contemporary emphasis on integrated social science. Current research on poverty increasingly recognises that income alone inadequately captures economic hardship, leading to expanded measures incorporating material hardship, consumption patterns, wealth holdings, and subjective well-being. Orshansky understood this in 1963, noting that her income-based thresholds were pragmatic tools, not comprehensive assessments of deprivation.

The gender dimensions of her story resonate powerfully today. Women remain underrepresented in economics – comprising roughly 30 percent of PhD recipients and an even smaller share of full professors at top-ranked departments. Applied, methodological work continues receiving less recognition than high theory. Government service remains less prestigious than academic appointments. The patterns that constrained Orshansky’s career and obscured her contributions persist, even as explicit barriers have diminished.

Perhaps most relevant to contemporary challenges is her insight about measurement as political act. In an era of proliferating metrics – from GDP to happiness indices, carbon footprints to social mobility scores – Orshansky’s reflection that “choosing where to draw the line determines whose suffering we deem worthy of intervention” offers essential caution. Every quantification embeds value judgements about what counts, who counts, what thresholds separate acceptable from unacceptable. Those choices shape reality, not merely describe it, allocating resources and attention whilst rendering other needs invisible.

Standing at her metaphorical gravestone nearly two decades after her death, we might ask what Mollie Orshansky would make of contemporary poverty politics. Child poverty reached historic lows in 2021 following expanded Child Tax Credits and pandemic relief, demonstrating that deprivation responds to policy choices, not inevitable economic forces – exactly as her measure was meant to reveal. Yet those programmes expired, and poverty rebounded, suggesting that six decades of counting accomplished less transformation than she might have hoped.

Her legacy endures not in monuments or named prizes – there is no Orshansky Medal, no Orshansky Lecture Series – but in the taken-for-granted infrastructure of American social policy. Every poverty statistic published, every eligibility determination made, every policy debate framed around “families below the poverty line” invokes her work, usually without acknowledgement. She gave poverty a number so we could see it clearly enough to fight it. That we chose to count rather than truly combat what we were counting reflects not the failure of her measure, but the insufficiency of measurement alone to compel justice.

The woman who disappeared behind her number reminds us that infrastructure has authors, that invisibility often marks success rather than insignificance, and that the most consequential work sometimes leaves the fewest biographical traces. In remembering Mollie Orshansky – calling her back from the archival shadows, reconstructing not just her methodology but her voice, her frustrations, her hopes – we perform a small act of repair against the erasure that attends women’s contributions to science. We cannot undo six decades of infrastructural anonymity, but we can insist that the poverty line had a creator, that creator was a woman, and her name deserves to be spoken alongside her achievement.

Who have we missed?

This series is all about recovering the voices history left behind – and I’d love your help finding the next one. If there’s a woman in STEM you think deserves to be interviewed in this way – whether a forgotten inventor, unsung technician, or overlooked researcher – please share her story.

Email me at voxmeditantis@gmail.com or leave a comment below with your suggestion – even just a name is a great start. Let’s keep uncovering the women who shaped science and innovation, one conversation at a time.

Editorial Note: This interview is a dramatised reconstruction created for educational and commemorative purposes. Mollie Orshansky passed away on 18th December 2006, and this conversation could not have taken place. The dialogue presented here has been carefully crafted using historical sources including Orshansky’s published writings, archival materials from the Social Security Administration, biographical accounts, academic analyses of her methodology, and contemporaneous documentation of poverty measurement debates during the 1960s. Whilst every effort has been made to ensure accuracy in representing her work, ideas, and historical context, the specific words, tone, and personal reflections attributed to Orshansky are imaginative interpretations based on available evidence, not verbatim quotes or verified statements. Similarly, the supplementary questions and her responses to them are fictional constructs designed to explore themes and technical dimensions of her contribution. Readers seeking authoritative biographical or methodological information should consult primary sources including Orshansky’s publications in the Social Security Bulletin, particularly “Children of the Poor” (1963) and “Counting the Poor: Another Look at the Poverty Profile” (1965), as well as Gordon Fisher’s comprehensive historical analyses available through the U.S. Department of Health and Human Services. This dramatisation aims to honour Orshansky’s legacy by making her story and contributions accessible to broader audiences, but it remains an act of historical imagination rather than documentary reportage.

Bob Lynn | © 2025 Vox Meditantis. All rights reserved.