Before prescribing anything, get the diagnosis straight, because two different problems hide under one word. Poverty is about how little some people have. Inequality is about how spread out incomes are. They can move together, but they don't have to. If every income doubled overnight, poverty as hardship would shrink dramatically while the gap between rich and poor, in dollars, would double.

Economists measure poverty two ways. Absolute poverty, like the official U.S. poverty line or the World Bank's line for extreme poverty, asks whether people can meet a fixed standard of need. Relative poverty, asks how far people fall below others around them. The free-market view tends to focus on the first: whether the floor is rising, whether the poor live better than they did, and whether they can climb.

Inequality is measured by lining up every household from poorest to richest and asking what share of total income each slice receives.

The Gini coefficient turns the picture into one number: the area between the curve and the diagonal (A) divided by the whole area under the diagonal (A + B). A Gini of 0 means perfect equality; a Gini of 1 means one household gets everything. For U.S. household money income in 2023, the Census Bureau put it at 0.485. Measures that count taxes and government benefits, or that compare what households consume rather than earn, show less inequality, so it matters which number you are reading.

Why insist on the distinction? Because the policies that shrink the gap fastest are not always the ones that raise the floor fastest. Chapter 1 made the point with a pie: a fight over slices can shrink the pie, while a growing pie can give the smallest slice more than it had.Ch 1 The rest of this chapter asks which policies actually raise the floor.

Start with how taxes are structured. A progressive tax takes a larger share of income as income rises; the federal income tax is the main example. A proportional tax takes the same share from everyone. A regressive tax takes a larger share from low earners than from high earners. A sales tax is regressive even though everyone pays the same rate at the register, because poorer households spend nearly all of their income while richer ones save some of theirs.

Those are the official taxes. The ones that matter most at the bottom are not called taxes at all. Almost every means-tested program shares one design feature: benefits shrink as earned income rises. That is reasonable, since help should flow to those who need it most. But shrinking a benefit as earnings rise is, mathematically, a tax on earning more. Economists measure it with the effective marginal tax rate: the fraction of one additional dollar of earnings that is lost to taxes and to benefits that phase out. If earning $1 more costs 40 cents in taxes and lost benefits, the effective marginal tax rate is 40%, whatever the official bracket says.

Now stack several programs together: food assistance, housing, child care, health coverage. Their phase-outs pile on top of one another, and the effective marginal tax rate on a low-income worker can climb past 60%, 80%, even 100%. When it crosses 100%, earning more leaves a family with less. That is a welfare cliff: a threshold where one extra dollar of earnings ends a benefit worth far more than a dollar, so total resources drop.

A worker facing a cliff who turns down a raise, declines extra hours, or refuses a promotion is not being lazy. She is responding rationally to the incentive the policy built, exactly as labor-supply logic predicts: people supply more labor when the net return to working rises, and less when it falls. A cliff makes the net return to crossing it negative. The tragedy is that a program meant to raise well-being can trap a family below the cliff, discourage advancement, and lower the output and mobility that well-being rests on.

In 1975 the economist Arthur Okun offered an image that every debate about redistribution still uses. Moving money from the rich to the poor, he said, is like carrying water in a leaky bucket. Some of it never arrives. Taxes on the rich discourage some of the work, saving, and investment that create income. Benefits that phase out discourage work at the bottom, as Concept 2 showed. And running the programs costs money. Okun himself favored some redistribution anyway; his point was that the leaks are real, and that how much a policy leaks should decide whether it is worth it.

The free-market case goes further, and its strongest argument is arithmetic. Growth compounds, and small differences in growth become enormous over a working life.

That is the core of the case: the policies that matter most for the poor over a lifetime are the ones that keep the economy growing and let the poor share in the growth by working, which is why the rest of this chapter is about work and ownership rather than transfers. Two more arguments add to it. The first is a prediction about politics, and it has a name: Director's Law. In 1970 George Stigler named it for his Chicago colleague Aaron Director: public programs are made mainly for the benefit of the middle class and paid for with taxes that fall heavily on both the poor and the rich. The logic is the median voter's. A program needs a majority to pass, and the votes are in the middle, so programs sold as help for the poor drift toward whoever is numerous and organized enough to claim them. Chapter 19 calls this public choice.Ch 19 The second argument is about people: help that replaces work can quietly take away something money can't buy: the skills, routines, and self-respect that come from earning.

If the best path out of poverty is a job, the first market solution costs almost nothing: stop blocking the path.

Licensing. In the 1950s, about 5% of American workers needed a government license to do their jobs. Today the figure is more than 20%. Some licenses protect the public from real harm. Many mainly protect the people who already hold them, which is the rent-seeking from Chapter 11: a chain across the river, stretched across the entrance to a career.Ch 11 Licensed workers are about a quarter less likely to switch occupations, and the burden falls hardest on people without the money or time for long training. Jestina Clayton learned to braid hair as a child in Sierra Leone and started a braiding business in Utah. The state told her she needed a cosmetology license: 2,000 hours of training that could cost as much as $18,000, none of it about braiding. In 2012 a federal judge struck the rule down as "wholly irrational."

Housing and place. Where you live shapes what you can earn, and zoning that blocks new homes keeps the poor out of the cities with the best jobs (Chapter 12).Ch 12 The Moving to Opportunity experiment gave some families in high-poverty public housing vouchers to move to lower-poverty neighborhoods. Children who moved before age 13 earned about 31% more in their mid-twenties than similar children who didn't move, and were more likely to attend college.

The first rung. The minimum wage, from Chapter 13, raises pay for workers who keep their jobs, but where it bites hard it can price the least experienced out of the entry-level jobs where people learn to work.Ch 13 It is also poorly targeted: many minimum-wage earners are not poor, and many poor households have no one working. Keeping the bottom rungs of the ladder open is itself anti-poverty policy.

If transfers are needed, design them so that working always pays. The Earned Income Tax Credit (EITC) is the clearest example of a transfer that works with incentives instead of against them, and it is one of the largest anti-poverty programs in the United States. Its defining feature is that you must earn to receive it, and it is built in three stages. In the phase-in range, the credit rises with every dollar earned: for a parent with one child, each dollar of wages brings 34 cents of credit on top. That is a negative marginal tax rate on work; the government pays you to earn. Across a middle plateau the credit holds at its maximum. Then a gradual phase-out shrinks it slowly to zero, a modest tax on additional earnings and never a cliff.

Milton Friedman proposed a bolder design in 1962: a negative income tax. Scrap the patchwork of separate programs, each with its own rules and cliffs, and replace it with one cash payment run through the tax system. Every negative income tax has two settings: a guarantee (what a family with no earnings receives) and a phase-out rate (how much the payment shrinks for each dollar earned). The payment reaches zero at the break-even income, the guarantee divided by the phase-out rate.

Here is the hard trade-off every designer faces. You want a generous guarantee, a low phase-out rate so work pays, and a low total cost. You can have any two. A high guarantee with a low phase-out rate means the payment reaches far up the income scale, which is expensive. Keep the cost down with a low phase-out rate and the guarantee must be small. Keep both the guarantee and the cost reasonable and the phase-out rate must be steep, which taxes work.

What does the evidence say? Early studies credited the EITC with drawing many single mothers into work during its expansion in the 1990s. A careful recent reappraisal by Henrik Kleven finds that most EITC expansions show no clear effect on employment, and attributes the 1990s gains largely to welfare reform and a strong economy instead. Either way, the EITC does not punish the first hours of work, which a cliff does. For cash with no work requirement, the evidence is clearer. In the negative income tax experiments of the 1970s, men heading households worked slightly less, and their wives cut back more. In a large recent study, people given $1,000 a month for three years were about 4 percentage points less likely to be in the labor force, worked 1 to 2 fewer hours a week, and earned about $1,900 less a year; the biggest change in how they spent their time was more leisure. Cash buys freedom, and some people use part of it to work less. Whether that trade is worth it is a value question; that it happens is not seriously in doubt.

Work is one path out of poverty. Ownership is the other, and it is the path the poor are most often shut out of. Chapter 12 called capital the input you have to wait for, because its payoff comes over time.Ch 12 That waiting is exactly what makes it powerful for the people who can afford to start.

Now stretch the clock past one lifetime. Wealth is the one kind of income a family can pass down, so its compounding doesn't have to stop at 65. A parent's down payment, a grandparent's savings, a college paid for without loans: each lets the next generation start its own clock earlier and with more. Wealth, unlike a paycheck, can be built by more than one generation at a time.

That arithmetic is the heart of an argument by the economists Trevon Logan of Ohio State and Darrick Hamilton (then at Ohio State, now at The New School): unequal wealth today comes largely from unequal access to wealth-building policies whose benefits compound across generations. Their history is a list of the policies that built American wealth and of who was kept out of them. Federal land policy in the 1800s went overwhelmingly to white families; the Southern Homestead Act of 1866, meant in part for freed people, was undermined by discrimination and poor administration. After World War II, the GI Bill paid for college and helped veterans buy homes, but most colleges were closed to Black veterans and the benefits went overwhelmingly to white ones. Federal mortgage insurance, which helped build the American suburb, was kept out of Black neighborhoods by redlining and restrictive covenants. And some wealth that was built was destroyed outright, as in Tulsa in 1921.

The result shows up in the data. In 2022 the median white family had $285,000 in wealth and the median Black family had $44,900: about 16 cents on the dollar. The income gap is far narrower, about 60 cents on the dollar in 2024. In a 2015 study, Hamilton and his coauthors found that Black families headed by a college graduate had less wealth than white families headed by a high-school dropout. Their conclusion is that education alone can't close a gap like that, because it is wealth that pays for education more than the reverse. As a remedy, Logan and Hamilton argue that the most just approach would be a comprehensive program of reparations.

Start with who owns the asset that compounded fastest. In 2022, 58% of American families had some money in the stock market, a record, mostly through retirement accounts. But the share falls steeply with income: in 2019, only 15% of families in the bottom fifth held any stock, compared with 92% in the top tenth. The bottom half of American households, ranked by wealth, owns about 0.6% of the country's stocks and mutual funds.

For some families the barrier starts earlier still. In 2023, 4.2% of U.S. households, about 5.6 million, had no bank account at all, and another 14.2% had one but still relied on check cashers, payday lenders, or money orders. The most common reason given for having no account was not distrust. It was not having enough money to meet the bank's minimum balance: a price barrier. Worldwide, about 1.3 billion adults are unbanked, and half of them come from the poorest 40% of households.

Markets have been tearing these barriers down, and the evidence is striking.

Mobile money. In 2007 a Kenyan phone company launched M-Pesa, which let people store and send money by text message through small agents in shops. Economists Tavneet Suri and William Jack found that access to it lifted about 194,000 Kenyan households, roughly 2%, out of extreme poverty. The gains were largest for households headed by women, about 185,000 of whom moved from farming into business or retail. No grant, no program: a cheaper, safer way to save and send money.

Stablecoins. A stablecoin is a digital dollar that can be sent across the world in minutes, where sending $200 abroad through traditional channels cost 6.5% on average in early 2025. In countries with high inflation, it is also a way to save in a currency that holds its value. A 2025 U.S. law, the GENIUS Act, requires dollar stablecoins to be fully backed by cash and short-term Treasury securities. The risks are real: they are not federally insured, and TerraUSD, a coin with no real reserves, collapsed in 2022.

Cheap investing. Account minimums, trading commissions, and high fund fees used to price small savers out of the stock market. Competition has driven many of them toward zero. Today a teenager with $20 can own a slice of the whole market through a low-cost index fund or fractional shares.

Government can help too, mostly by getting out of the way. Many programs still impose an asset test: save too much, and you lose your benefits. Supplemental Security Income, for low-income people who are elderly or disabled, cuts off anyone with more than $2,000 in savings. That limit was last raised in 1989; if it had kept up with inflation since the program began, it would be close to $10,000. A rule like that is a cliff for saving: it tells the poor that building a cushion is against the rules.

Meanwhile the tax code subsidizes saving for those who already have something to save. Deductions for retirement contributions and mortgage interest, and lower tax rates on capital gains, are worth the most to people in high tax brackets with large accounts and expensive homes, and nothing to a worker who owes no income tax. The free-market fix is not a new program but a level one. A step in that direction starts in 2027: the federal Saver's Match will deposit up to $1,000 a year straight into the retirement accounts of low- and moderate-income savers, paid even if they owe no tax.

The boldest idea is to give every child a stake, and it comes from Hamilton himself: he and the economist William Darity proposed baby bonds in 2010. Connecticut was the first state to adopt them. Its program invests $3,200 for each child born into the state's Medicaid program since July 2023. At 18 to 30, the child can claim it, projected at $11,000 to $24,000, to buy a home, start a business, pay for training, or save for retirement. In 2025 Congress created federal child accounts, often called Trump Accounts: $1,000 from the Treasury for each citizen child born from 2025 through 2028, invested in low-cost index funds, to which families can add up to $5,000 a year and employers up to $2,500. Philanthropists Michael and Susan Dell pledged $6.25 billion to add $250 for millions of children in lower-income areas. These programs are new, and whether they change what young adults do with their lives is not yet known. But they do what no transfer at 40 can: they put the poorest families' clocks on the same starting line as everyone else's.

Access is the first step, not the last. Even people with open accounts and good options often don't save as much as they mean to. Why that happens, and what can be done about it without taking away anyone's freedom to choose, is a question for later in the book.