Opening
HOST
Welcome back. Today we are on financial crises and the short-run model: what happens to output and inflation when banks panic. Let's start with a weekend in late September 2008.
ECONOMIST
Ben Bernanke, head of the Fed, the central bank of the United States, met the leaders of Congress. He asked them to approve a huge rescue of the banks. Without it, he warned, there might be no economy by Monday.
HOST
And the Fed cut interest rates, I suppose.
ECONOMIST
It did. By the end of 2008, its rate was almost zero. Yet the slump was the worst since the Great Depression of the 1930s.
HOST
So here is today's question. Why did the 2008 crisis cause such a deep slump, even though the Fed cut its rate to zero?
ECONOMIST
A slump, or recession, is a time when output and jobs fall. By the end of this episode, you'll be able to do three things. First, explain why extra interest for risk can hit spending like a fall in demand, and why a zero rate could not undo it. Second, explain how falling prices, with the interest rate stuck at zero, can turn a recession into a spiral. Third, describe the other tools the Fed and the government used, and what the data show about the slow recovery.
HOST
And we'll follow the chapter's mind map part by part?
ECONOMIST
Yes. Part one is the crisis in the data. Part two builds the model: friction, deflation and the zero lower bound. Part three is policy and the aftermath. Each part has three to five leaves, small topics we take in order.
The crisis in the data
HOST
Part one of the map: The crisis in the data. The first leaf is the BAA spread. What's that?
ECONOMIST
It's a credit rating, a grade for how safe a borrower is. BAA bonds are loans to solid companies that are not completely safe. The safe loan we compare them with is the U.S. government's 10-year bond, called a Treasury. The spread is the gap between their interest rates.
HOST
So the spread is the extra interest a solid company pays, in case it doesn't repay.
ECONOMIST
Exactly. Economists call that extra interest a risk premium. From 1996 to 2007, the spread averaged 2.16 percentage points. A percentage point is the plain difference between two rates.
HOST
And in the crisis?
ECONOMIST
In December 2008, the spread reached 6.01 points, almost three times its usual size.
HOST
But the Fed was cutting its rate in 2008. So I'd guess companies paid less to borrow, too.
ECONOMIST
That's the natural guess, but it's wrong. From 2006 to 2008, the government's 10-year rate fell by more than a point, while the rate on BAA bonds went up by about a point. So the Fed's cuts reached the government, not the companies that build factories.
HOST
Next leaf: the Fed rate. Remind me, which rate does the Fed set?
ECONOMIST
The Fed is the Federal Reserve. It sets the fed funds rate, the rate banks charge one another for overnight loans. On Islandia, the imaginary island country of our examples, that's the policy rate of the Bank of Islandia, the island's central bank.
HOST
And after the crash?
ECONOMIST
The fed funds rate was almost zero by the end of 2008, and it stayed there. From 2009 to 2015, it was never higher than 0.19 percent.
HOST
So in the panic, money was very cheap for everyone.
ECONOMIST
A lot of people think that, but not for firms. Firms pay the Fed's rate plus the risk premium. In late 2008, the premium had about tripled, so the rate on company bonds went up, not down.
HOST
Next leaf: the price-earnings ratio. What does it measure?
ECONOMIST
A share is a small piece of ownership in a company. Take an index of share prices, one number that follows many shares. Divide it by the average company profits of the past ten years, after inflation. Over the long run, that ratio has averaged about 18.
HOST
And before the great crashes?
ECONOMIST
In September 1929, just before the crash, the ratio reached 32.6, almost twice its long-run average. In December 1999, at the peak of the boom in internet shares, it reached 44.2, even higher. Both times, prices then fell a long way.
HOST
So those were bubbles?
ECONOMIST
Many economists call them that: prices far above what profits can justify. Home prices seem to have done the same in the 2000s, and their fall is where the 2008 crisis began. When those prices crash, people feel poorer and spend less.
HOST
So in part one we saw three facts. Lenders asked much more for risk, the Fed's rate was stuck at zero for years, and share prices climbed far above profits before both great crashes.
ECONOMIST
Right. Next branch: a model that puts these facts together.
Friction, deflation and the zero lower bound
HOST
Part two of the map: Friction, deflation and the zero lower bound. The first leaf is an equation: the rate firms pay is the Fed's rate plus a risk premium. A jump in that premium works like a fall in demand.
ECONOMIST
A financial friction makes borrowing more expensive. Here it's the risk premium. In calm times it is small and steady, so in practice the Fed sets the rate firms pay. In a panic, it jumps.
HOST
And why is that like a fall in demand?
ECONOMIST
Remember the IS curve. Firms compare the rate on a loan with the return on new capital, what a new machine or shed earns each year as a share of its cost. Take Harbor Boatworks, the island's boatyard. When the premium makes its loans cost more, it builds fewer boats, even if the central bank's rate hasn't changed. So spending and output fall, as if islanders had decided to spend less. To keep the rate firms pay the same, the central bank must cut one extra point for each point of premium.
HOST
Next leaf: 2008. The Fed cuts to zero, but the premium jumps to 4 points, and output ends at minus 4 percent. How do we get there?
ECONOMIST
Let's build that number step by step, on Islandia. Here every rate is a real rate, after taking out inflation. We measure output as a percent gap from potential output, its normal level. It starts at zero. The return on new capital is 2 percent, and so is the Bank of Islandia's real rate. The premium is zero. Each point that firms pay above that return costs 1 percent of output.
HOST
Then the housing bubble bursts.
ECONOMIST
Islanders feel poorer and spend less, so demand falls by 2 percent of potential output. With no response, output would be minus 2 percent.
HOST
But the Bank cuts its real rate from 2 percent to zero, 2 points below the return on capital. That adds back 2 points, and output is back at potential.
ECONOMIST
Right, the cut cancels the housing crash. Now the panic adds a premium of 4 points, about the jump we saw in the data.
HOST
The Bank has already cut to zero, so I'd guess output stays at potential.
ECONOMIST
That's the tempting answer, but let's check. What do firms pay now?
HOST
Zero plus 4, so 4 percent.
ECONOMIST
That's 2 points above the return on capital. It costs 2 points of output, on top of the crash's 2. So output is minus 4 percent, twice as deep as the crash alone. Why doesn't the Bank cut further? Two leaves on, we meet the floor.
HOST
Next leaf: the new AD curve, where inflation drifts to minus 4 percent. What's an AD curve again?
ECONOMIST
AD means aggregate demand, total spending. The AD curve links inflation to output through the rule of Chapter 13. When inflation falls below the 2 percent target, the central bank cuts its real rate, one point for each point. This time the Bank follows that rule instead of cutting at once. So the 2-point crash and the 4-point premium make one sudden hit to demand, a shock, of 6 points.
HOST
Then output falls by 6 percent?
ECONOMIST
Only by 4 percent in the first year. As output falls, the Phillips curve pulls inflation down, half a point for each point of the gap, and the rule cuts the rate. Things settle where both agree: output minus 4, inflation zero.
HOST
So inflation is 2 points below target, and the rule cuts the real rate by 2 points, to zero.
ECONOMIST
Exactly. That takes back 2 of the 6 points, so the rule softens the hit. After that, inflation keeps falling and the rule keeps cutting, and each year about one third of the remaining gap closes.
HOST
Where does it end?
ECONOMIST
Output is back at potential only when the rule has cut the rate by the full 6 points. By the rule, inflation is then 6 points below the 2 percent target: minus 4 percent. Prices falling year after year is called deflation.
HOST
Is that what happened after 2008?
ECONOMIST
Not quite. U.S. inflation stayed low but stopped falling, perhaps because people trusted the 2 percent target. And the model needs cuts below zero in money terms, which the next leaf shows is impossible.
HOST
Next leaf: the zero lower bound. The real rate is the nominal rate minus inflation, and the nominal rate can't go below zero. What's a nominal rate?
ECONOMIST
It's the rate in the news, in money terms, like the 0.19 percent of part one. It has a floor, a lowest level, near zero, because islanders can always hold cash, which earns zero. So at a zero nominal rate, the real rate is minus inflation.
HOST
So in the 2008 leaf, did the floor stop our Bank at a real rate of zero?
ECONOMIST
It depends on inflation. With inflation at zero, as in a deep slump, a zero nominal rate is a zero real rate. Then the Bank was already at the floor, and minus 4 was the best it could do.
HOST
And if inflation were still at its 2 percent target?
ECONOMIST
Then a zero nominal rate is a real rate of minus 2. Firms pay that plus the premium of 4, so 2 percent, just the return on capital. The rate does no harm, and only the crash's minus 2 is left.
HOST
But to close the gap, firms must pay zero, so the Bank needed a real rate of minus 4.
ECONOMIST
Yes, so a nominal rate of minus 2, below the floor. That's the answer to today's question in one line: the premium called for a cut below zero.
HOST
And if prices start to fall?
ECONOMIST
Then it gets worse. With prices falling 2 percent a year, a zero nominal rate is a real rate of plus 2, so firms pay 6 percent.
HOST
That's 4 points above the return on capital.
ECONOMIST
So it costs 4 points, on top of the crash's 2: output is minus 6.
HOST
Last leaf of this part: the deflationary spiral of the Great Depression. Minus 3, minus 6, minus 9 percent. What are those numbers?
ECONOMIST
Output in three steps of our model. Before the crash, output is at potential and inflation is about zero. Step one: worried about a bubble in shares, the Fed raises its rate by 3 points. Output falls to minus 3 percent.
HOST
And step two is the crash of October 1929.
ECONOMIST
Yes. The crash cuts spending by another 3 points, so output is minus 6 percent. Step three is the dangerous one. The slump pulls inflation down by 3 points, from about zero to minus 3.
HOST
So prices are falling, and the Fed keeps its nominal rate where it was.
ECONOMIST
Exactly. So the real rate rises by 3 points, and output falls to minus 9 percent.
HOST
And then? I'd guess output recovers, as it did under the rule.
ECONOMIST
Here's the surprise: it's the opposite. Under the rule, falling inflation brought rate cuts. Now the nominal rate is frozen, so falling inflation raises the real rate. From there, inflation falls even faster, the real rate rises again, and output falls again. Each year's gap is one and a half times the last: minus 6 became minus 9, and next year is worse. That's the spiral. In the real Depression, prices fell by about a quarter from 1929 to 1933.
HOST
So in part two we saw that the premium acts like a fall in demand. The zero floor blocks the cuts that would offset it, and falling prices can make the slump deeper and deeper.
ECONOMIST
Next branch: what policy did when its usual tool was stuck.
Policy and the aftermath
HOST
Part three of the map: Policy and the aftermath. The first leaf: money kept growing, as M1 growth went from 7 to 17 percent in late 2008. Why does that matter?
ECONOMIST
Because of the 1930s. M1 is cash, meaning notes and coins, plus the bank accounts people pay from. The economists Milton Friedman and Anna Schwartz blamed the Great Depression on a Fed that let the money supply fall from 1929 to 1933.
HOST
I'd guess money shrank again. Banks were scared and stopped lending.
ECONOMIST
Look at the data first: not this time. Between October and December 2008, the growth of M1 over the past year jumped from about 7 to about 17 percent. Growing money helps keep prices from falling, and part two showed why that matters.
HOST
Next leaf: quantitative easing. By 2015, the Fed's assets, everything it owned, were worth five times as much as before. What did it buy?
ECONOMIST
Quantitative easing, or QE, means buying assets other than short-term government bonds, on a huge scale. The Fed bought long-term government bonds and bundles of home loans, to push down the long-term rates that firms and families pay.
HOST
So the Fed printed trillions of dollars. High inflation must have followed.
ECONOMIST
That's a common belief, but it isn't what happened. The Fed paid mostly by adding to the deposits that banks keep at the Fed, called reserves. New cash paid for only about 15 percent. Cash gets spent in shops, but the Fed pays interest on reserves, so banks mostly held on to them. Inflation stayed low for more than a decade after QE began.
HOST
Next leaf: forward guidance. The expected rate was cut to 0.57 percent in June 2012. Expected by whom?
ECONOMIST
By financial markets. Forward guidance means telling markets how long the rate will stay low. A long-term rate is roughly the average of the short-term rates expected over the life of the loan. So a promise of low rates for longer lowers long-term rates today.
HOST
Can we put numbers on it?
ECONOMIST
Take a four-year loan. If markets expect short rates of zero, 1, 2 and 3 percent, the four-year rate is about their average, 1.5 percent.
HOST
And if the Fed promises zero for three years?
ECONOMIST
Then markets expect zero for three years, and 2 percent in the fourth. The average falls to one half of a percent.
HOST
And in 2012?
ECONOMIST
In January 2012, the Fed said its rate would likely stay very low until at least late 2014. Markets believed it: by June, they expected only 0.57 percent for the end of 2014.
HOST
Next leaf: potential GDP revised down 10 percent, and productivity growing 0.72 percent a year since 2004. Let's start with potential GDP.
ECONOMIST
Potential GDP is the total output the economy would produce with its workers and machines in normal use. The Congressional Budget Office, or CBO, estimates it.
HOST
About ten years after the crisis, GDP was back at potential. So I'd guess the economy recovered everything it had lost.
ECONOMIST
That's what the gap says, but look at how it closed. In 2017, GDP was less than 1 percent below the CBO's latest estimate of potential.
HOST
And against the old estimate?
ECONOMIST
Almost 11 percent below the potential the CBO expected back in 2007. So the estimate itself came down by about 10 percent. The economy never got back to the output expected before the crisis.
HOST
Why was it lowered?
ECONOMIST
Low investment after the crisis, and slower productivity. Total factor productivity, or TFP, measures how well an economy uses its machines and workers. In private business, it grew more than 1.5 percent a year from 1995 to 2004.
HOST
And since 2004?
ECONOMIST
Only 0.72 percent a year, less than half as fast. But the slowdown began before the crisis, so the crisis can't be its only cause.
HOST
Last leaf: policy, three topics: QE, fiscal stimulus and bubbles. What about fiscal stimulus?
ECONOMIST
Fiscal stimulus means the government buys more or cuts taxes. Normally, the central bank raises its rate when demand rises, which takes back part of the boost. At the floor, the rate stays at zero, so in our model each extra dollar of purchases adds a dollar of output. How much it adds in real life is still debated.
HOST
And bubbles. Should the central bank raise its rate to pop them?
ECONOMIST
That's debated too. On one side, low Fed rates from 2002 to 2006 may have helped the housing boom grow. On the other, bubbles are hard to spot at the time, and a rate rise hits the whole economy. Many economists prefer a capital requirement instead. Capital here means the owners' own money in a bank, a cushion against losses. Islandia's banks keep a thin cushion, and the requirement makes it thicker.
HOST
How does that help with bubbles?
ECONOMIST
When a bubble bursts, the losses hit the owners' cushion first. So lenders worry less, and the premium jumps less.
HOST
So in part three we saw the Fed avoid the mistake of the 1930s and try new tools, while governments added spending. But output stayed below the old expectations, and productivity had already slowed.
ECONOMIST
Exactly. Let's sum up.
Wrap-up
HOST
Let's go back to our three goals. First: why the premium hits like a fall in demand, and why a zero rate couldn't undo it.
ECONOMIST
Firms pay the central bank's rate plus the premium, so a higher premium cuts investment like a drop in demand. Offsetting it needed a nominal rate below zero. That's impossible, because people can always hold cash, which earns zero.
HOST
Second, the spiral.
ECONOMIST
With the nominal rate stuck, falling prices raise the real rate. That deepens the slump, which makes prices fall faster, again and again.
HOST
And third, the other tools, and what came after.
ECONOMIST
Money kept growing, the Fed bought long-term assets and promised low rates, and the government added spending. Still, output stayed below what was expected before the crisis. So here is the big message. A jump in the risk premium works like a collapse in demand, and the zero floor stops the central bank from offsetting it. Falling prices can then turn a recession into a spiral. So central banks and governments need tools beyond the interest rate.
HOST
Here's a question to think about. Suppose Islandia's central bank aimed for zero inflation instead of 2 percent. In the next crisis, how far below zero could it push the real rate, and why does that matter?
ECONOMIST
Pause the episode and think about it for a moment.
HOST
Okay, we're back. What's the answer?
ECONOMIST
It couldn't push the real rate below zero at all. At a zero nominal rate, the real rate is minus inflation: minus 2 with 2 percent inflation, but only zero with zero inflation. On Islandia, the best output at the floor falls from minus 2 to minus 4 percent. That's one reason central banks aim for a little inflation, not zero.
HOST
Where can students try all this themselves?
ECONOMIST
Open the chapter's Colab notebook to run every example yourself. The Model Apps let you move the premium without code, and the Data Apps show the charts.
HOST
And next time?
ECONOMIST
In Chapter 15, we meet DSGE models, the modern models of business cycles, the ups and downs of the economy. We ask: do productivity shocks cause recessions, and how do sticky wages and prices change the answer? Sticky means slow to change.
HOST
That's it for today. Thanks for listening, and see you next time.