The Quiet Depression

Take a household in the lowest-income tenth of the country. Put them next to a household in the highest-income tenth. Run both of them forward from 1997 to today. The high-income household is roughly 46 percent better off in real terms. The low-income household is almost exactly where it started. After housing costs, it is worse. That is not a recession. That is not a bad year. That is what twenty five years of a particular policy regime produces, and it is the story this piece is about.

One important clarification before we start. Everything in this piece is about income, not wealth. Income is what you earn. Wealth is what you own. They are not the same, and in Britain they increasingly do not even move together. Plenty of low-income households sit on valuable housing wealth. Plenty of high-earners have almost no net worth outside their salary. This article is about the flow of money into and out of your bank account each month. A separate piece on household wealth is coming, and it will tell a different, and in some ways worse, story.

And one more clarification, because it matters. This is not a redistribution piece. The top decile did not pull ahead because anyone else was robbed. It pulled ahead for the same reason real wages stagnated everywhere below it: policy. And it is worth saying plainly what the highest earners are already contributing. According to HMRC's own published liabilities, the top 1 percent of income taxpayers pay roughly 29 percent of all income tax. The top 10 percent pay around 60 percent. The bottom 50 percent, combined, pay about 10 percent. The top decile is not the problem. It is the single largest source of funding for everything this piece argues the state is failing to deliver. The argument here is not to tax them harder. It is that the bottom and middle deciles have not moved in real terms for a quarter of a century, and the fix for that has to be growth, not redistribution.

The argument is simple. British take home pay has not failed evenly. It has failed in a specific shape: flat at the bottom, flat in the middle, pulling away at the top, and well behind every serious G20 peer except Italy and Japan. And the shape does not belong to one government. It runs straight through Blair, Brown, Cameron, May, Johnson, Truss, Sunak and Starmer. The architecture was set in the late 1990s. Every administration since has preserved it.

We are going to look at two things. First, what happened to real income by decile inside the UK since 1997, before and after housing costs. Then, how British real wages compare to the rest of the G20 over the same period. The UK data comes from the ONS Households Below Average Income series and the IFS Living Standards work. The international data comes from the OECD and ILO wage series. The numbers are not in dispute. What they mean, is.

1. Twenty five years, by decile

The chart below is the single most important chart in this piece. It splits British households into ten equal income buckets, from lowest to highest, and asks how much real net income each bucket has today compared to the same bucket in 1997. Toggle the switch to see the difference between income before housing costs and income after them. BHC is what arrives in your bank account. AHC is what is actually left once the landlord or the mortgage has taken their cut. For most people in the bottom half of the country, AHC is the one that matters.

Three patterns jump out, and they line up neatly with three political eras.

The first was a real, if expensive, boom. Between 1997 and 2007, every decile moved up, and the bottom two moved up fastest. This was the tax credit decade: a system that raised bottom decile take home pay without raising gross wages, by topping up the difference out of general taxation. It worked, on its own terms. It is also the only window in the full quarter century where the distribution actually moved in favour of people on lower incomes. The bill for it quietly doubled the working age welfare budget, and the trick only works for as long as the taxpayer is willing to keep writing the cheque.

Then it stopped. Between 2008 and 2016, real household income for deciles one to eight is virtually indistinguishable year on year. A line on a chart that should be climbing just sits there. Only the top two deciles kept moving, partly because their pay held up better and partly because their wealth sat in assets the Bank of England was busy reflating with QE. Historians will eventually describe this stretch the way they now describe the 1930s: the longest wage stagnation in peacetime British history.

Then it stopped again. From 2017 to 2023, there was a modest recovery to 2019, then covid, then the Liz Truss mini-budget, then the 2022 inflation wave. Each of those took a slice out of real income. By 2023 the lowest income decile had, after housing costs, less real income than it had in 2004. Not less than last year. Less than nineteen years ago.

2. The growth table: who gained, who held, who lost

If the line chart is the picture, the table below is the invoice. It takes the same three eras and asks, bluntly, how much better off each decile is at the end of each one. All figures are cumulative real growth in net household income, before housing costs. After housing costs, add a few more percentage points of pain to the bottom half.

Figures are real net household income, equivalised, before housing costs, CPIH deflated. Values are indicative, rounded from HBAI, ETB and IFS living standards series. AHC figures reduce D1 to D5 by an additional three to six percentage points across the full period because of housing cost growth.

Two things worth pausing on.

First, look at the gap between D10 and D1. On paper, it barely moved in percentage terms. The top decile had 5.8 times the income of the bottom decile in 1997; it still has something close to that today in BHC terms. But that is before you account for housing. Once you do, the gap widens sharply. The top decile mostly owns. The bottom decile almost entirely rents. When housing costs grow faster than wages for twenty five years straight, the arithmetic stops being subtle. Housing is where the distribution actually broke, and it broke for reasons that have nothing to do with what the top decile was or was not doing.

Second, look at the middle of the table. D3 to D7 grew within a three percentage point band of each other across a quarter century. That is 25 years in which a primary school teacher, a shift manager at a supermarket, a mid-level engineer, an NHS band 6 and a small business owner all ended up in essentially the same place. Not asset-wealthy enough to benefit from soaring house and share prices. Not low-income enough for the transfer system to reach them. Just a broad middle, carrying a growing share of the income tax base and told every few years that the economy is doing well. Above them, D10 is doing the genuine heavy lifting on income tax, paying roughly 60 percent of the total. Below them, the transfer system does what transfers do. The middle is the gap the system does not see, and this is the silent casualty column.

3. The G20 league table

One defence you hear a lot is that wage stagnation is a problem everywhere. That the whole developed world saw real wages stagnate together, and the UK is just part of that story. It is not true. If you pull the OECD and ILO real wage series for every G20 economy and index them all to 1997 = 100, you get the chart below. Every bar shows where a country's average real wage sits today compared to where it sat at the start of the Blair era. The vertical line is the 1997 baseline. Anything to the right of it is real wage growth. Anything to the left of it is real wage decline.

Start with the G7, because that is the comparison Britain usually makes with itself. Over the full 25 years, British real wages grew roughly ten percent. The US managed about eighteen. Germany, thirteen. France, sixteen. Canada, nineteen. Of the seven largest advanced economies on earth, only Italy and Japan did worse. Japan has been a cautionary tale about deflationary wage stagnation in every economics textbook written since 1995. We are now in the same paragraph as Japan.

Then extend the frame. Australia grew by a quarter. South Korea grew by roughly half. Even Mexico and Turkey, for all the volatility their currencies and inflation have thrown up, delivered more real wage growth to their workers than the UK did to its own. China is literally off the chart. India has more than doubled its real wage level. Essentially every G20 economy that did not start the period with a fully mature Western European labour market is now growing real wages faster than Britain.

Look at where the UK bar sits. Fifteenth out of the nineteen countries in the chart where the data is actually comparable. The countries Britain beats are Italy, Japan, Argentina and, on a technicality, Saudi Arabia. That is not a peer group. That is a warning list.

4. Why: the architecture that produced this shape

The official explanation is that wages stagnated because productivity stagnated, and no one is quite sure why productivity stagnated. That is the kind of answer you give when you do not want to name the people who made the decisions. The truth is harder and more useful. Productivity stagnated because the policy regime of the last 25 years was designed, decision by decision, to produce the shape of the chart at the top of this piece. Four of those decisions carry most of the weight.

The mass expansion of higher education

In 1997, the Blair government announced the 50 percent target. Go to university. Take out a loan. The country will be wealthier for it. University intake nearly tripled over the next two decades. Fees appeared in 1998, tripled in 2006, and were raised again in 2012. And then the thing the policy was supposed to deliver, a wage premium for graduates, quietly collapsed. Because when you flood the market with graduates faster than the economy creates graduate jobs, a degree stops paying for itself. The median graduate in 2023 earns roughly what a median A-level educated worker earned in 2003. The attached debt is around £50,000. This was not a market failure. This was a policy that sold young people a wage premium while dismantling the conditions that created it.

The tax credit architecture

Between 1999 and 2010, Gordon Brown built the most sophisticated transfer system in British peacetime history. It did exactly what it was advertised to do: it lifted bottom decile take home pay without the political cost of raising gross wages. That is why the lowest two income deciles grew faster than the middle from 1997 to 2007. The part nobody advertised is that it also let employers pay lower gross wages, safe in the knowledge that the Treasury would top up the difference. When austerity eventually pared back the top-ups, the gross wages underneath them had already re-anchored lower. The floor did not come back with the subsidy.

The housing supply settlement

Net housing additions ran below household formation in every year of the last 25 apart from a handful. Mortgage costs moved from roughly 3x income to 4.5x for the median first time buyer. Private rents grew faster than wages in every decade since 1997. The effect on AHC incomes for the bottom five deciles was direct: everything they gained in gross, they paid back in housing. This is why the AHC line for D1 barely moves at all.

The size of that transfer is the entire story of why the AHC chart looks the way it does. Our earlier piece, They Imported Millions. You Paid the Price., set out the supply side numbers in detail. The UK added 11.1 million people between 1997 and 2024 and built roughly 4.8 million net new homes. Home ownership fell from a peak of 70.9 percent in 2003 to 63.0 percent today even as prices doubled. The outstanding mortgage stock grew from around £500 billion to £1.73 trillion. The average first time buyer is now 33, up from 29 in 1997. And 45.8 percent of all new mortgage lending is now at more than 4.5 times income, up from a small single digit share when this policy regime began.

Every one of those numbers is a direct subtraction from AHC real income, concentrated in the bottom six deciles. A bigger mortgage relative to income at a higher multiple over a longer term, entered later in life, is how wage growth gets silently transferred from workers to rentiers and to the banks that finance the loans. The decile chart above is, in large part, just the wage side of the housing shortfall ledger. If you want the full accounting, read them together.

The post 2010 preservation of the regime

And here is where most accounts lose their nerve. The Conservative governments of 2010 to 2024 campaigned, loudly, against the previous regime. In practice, they preserved every load-bearing part of it. University fees went up, not down. Tax credits were renamed Universal Credit and kept. Housing supply stayed below household formation. Income tax thresholds were frozen from 2021, extending the reach of fiscal drag into every middle decile. Employer NI was pushed up in 2025. Four different Prime Ministers wore different rosettes and delivered, almost line for line, the same policy settlement. The chart above is not a partisan verdict. It is a cross-party one. That is the editorial point this think tank keeps making, and it is the only framing that honestly matches the data.

5. Who this matters to

It matters to Gen Z. If you are 22 and you feel like the system is stacked against you, that is not imposter syndrome, that is the data. Graduate starting wages are lower in real terms than they were for your parents, and the debt attached to them is higher than it has ever been. The people who stacked the deck against you are not your grandparents. They did not sit in cabinet meetings and vote to triple university intake, freeze the personal allowance, or let housing supply fall behind population growth for three decades. The Treasury did. The Department for Education did. Successive ministers of both colours did. Direct your anger at them.

It matters to boomers. You are routinely told, particularly online, that your prosperity came at your children's expense. It did not. You bought houses at three times your income because the housing supply settlement made it possible. You got free university because the country could afford to educate fewer graduates into better jobs. You had functioning labour markets because successive governments were not yet running the economy for rentiers. Every one of those conditions was dismantled after your generation benefited from it. Your children are worse off because the system changed, not because you cheated.

And it matters most to Gen X and older millennials. Look at that middle band of the decile chart again, the one where five deciles are stacked on top of each other going almost nowhere for fifteen years. That is you. Too young for the free university. Too late for the cheap house. Too busy working, paying tax and raising kids to notice that your real wage had not moved in a decade and a half. You are the generation everyone else's argument forgets. When the commentariat talks about boomers vs Gen Z, they are describing a fight at one end of a room that you are holding up with your tax contributions. This think tank has said it before and will keep saying it: Gen X and older millennials are the single biggest policy failure of the last quarter century, and the only group the political class seems unable to name.

6. What would fix it

Before we list the fixes, two points worth saying clearly. The first is that the state did not stand back while real wages stagnated. It was an active participant. It expanded higher education and crushed the graduate premium. It subsidised low wages through tax credits and let gross pay re-anchor downward. It froze tax thresholds and let fiscal drag do the work of a stealth tax rise. It designed a planning system that protects a builder oligopoly and keeps housing supply scarce. Every one of those is a state intervention that made working-age people poorer. The fix is not more state. It is removing the distortions the state already installed.

The second is that none of what follows is about taking anything from anyone. The top decile already pays close to 60 percent of all income tax, and the top 1 percent pays 29 percent of it on their own. They are not the reason the middle and the bottom have not moved in real terms for twenty five years. Policy is. So the answer is not to squeeze the people already funding most of the system. It is to grow the real income of everyone below them, by removing the structural drags that have held that income down. Growth, not redistribution. Bigger real wages across D1 to D9, not smaller real wages in D10.

Four things, in order of how much real money they would put back in ordinary people's pockets.

• Break the planning cartel. The British housing shortage is not a market failure. It is a regulation. The Town and Country Planning Act replaced an as-of-right system with case-by-case discretion in 1947, and every decade since has added more discretion, more consultation, more appeal rights, more reasons for a "no". That is how you get a country that added 11.1 million people and built 4.8 million homes. The fix is zone-based approval: if your plot is inside the zone, and the build meets the code, you get your permission. Period. Strip the Green Belt designation from the roughly 30 percent of it that is not green, not open, not in public use, and sits inside existing commuter belts. Cut the consenting period to a statutory 12 weeks. The builder oligopoly survives today because only firms of a certain size can navigate the process. Simplify the process and the oligopoly breaks on its own, without a single new subsidy or quango.
• Strip the regulatory moat that protects the big five housebuilders. Section 106, Biodiversity Net Gain, affordable housing quotas set per-site, phased build-out clauses, nutrient neutrality. Each of them, individually, sounds reasonable. Together, they add a fixed compliance cost that a Persimmon can absorb and a small builder cannot. The result is a top-five market share of 55 to 60 percent and land banks measured in years rather than months. Remove the moat, and you multiply the number of firms that can credibly build a house by an order of magnitude. Volume then comes from competition, not from a minister's target.
• Unfreeze the tax thresholds and stop using fiscal drag as a stealth tax. The frozen personal allowance, higher rate threshold and child benefit taper are, right now, the single largest transfer of real income from middle earners to the Treasury. Indexing them to CPI would return roughly £640 a year to a worker on median wages, and meaningfully more to anyone fiscal drag is pulling into the 40 percent band. This is not a new tax cut. It is stopping an old stealth rise from quietly expanding every year.
• Reprice further education and let the market signal the rest. Stop guaranteeing loans for any course at any institution regardless of graduate outcomes. If a degree's median earnings cannot service the loan, the state should not underwrite the loan. That alone forces universities to price courses where the market actually values them, and it restores the wage premium the 1997 policy quietly eliminated. No central planner picks winners. The market does, through the prices students and lenders are willing to pay.

None of the four requires a bigger state. Three of them are actively smaller. The planning fix is deregulation. The housebuilder fix is deregulation. The threshold fix is the Treasury giving back what fiscal drag has been quietly taking. The education fix is ending a universal state guarantee and letting the market do the allocation. The reason the four feel politically impossible is not that they are impossible. It is that the beneficiaries of the current regime, the listed housebuilders, the mortgage lenders, the Treasury itself, are concentrated, organised and loud. The losers are dispersed across every decile below the top, do not know they are a class, and have no one in Westminster naming them as one. This think tank exists, in part, to name them.

Methodology and sources

UK decile data is drawn from the ONS Households Below Average Income series (HBAI), the ONS Effects of Taxes and Benefits on Household Income (ETB), and the IFS Living Standards, Poverty and Inequality dataset. Equivalisation uses the modified OECD scale. BHC deflates by CPIH; AHC uses a rent inclusive deflator. Values in chart 1 are indexed for clarity and should be treated as rounded to the nearest full percentage point. G20 real wage growth uses the OECD Average Annual Wages dataset where available (G7, Australia, South Korea, Mexico, Turkey) and the ILO Global Wage Report for remaining economies (China, India, Brazil, Indonesia, Russia, South Africa, Argentina, Saudi Arabia). Emerging market series are subject to structural breaks and should be read as direction of travel rather than precise point estimates.