IAS 107 Intermediate Macroeconomics March 17, 2011 Inflation Lecture Cleaned-Up Transcript
J. Bradford DeLong
The main meat of this lecture is Inflation Economics.
Here we see the consumer price index for all urban consumers in the United States of America since 1950. The price level today is 10 times as high as it was in 1950. If you had a dollar in 1950, you could have bought about ten times as much stuff as you can with a dollar today.
You can argue over whether this is right. There are lots of things that we consume and that we like to consume and that we spend our time doing today that are to a first approximation absolutely free today that that were simply unavailable in 1950. In 1950 the number of people who could watch television was in the hundreds. Things that people used to pay substantial amounts of money for in the 19th century or even in 1950 are available to us virtually free through either broadcast media or the internet.
Back in 1890, New England Utopian Edward Bellamy wrote a book--Looking Backward--that was one of the five bestselling novel of the late 19th century. That Looking Backward was a best-selling novel is worth noting because it is execrable. It is close to the worst novel you could imagine. Complete cardboard characters. No interesting plot. Completely unconvincing romantic subplot. Implausible contrivances. Author stepping out of his role to preach at you over and over again. No possible reason for anyone to read it.
Except that millions and millions of people in America did.
They read Looking Backward because it painted a picture, gave them a wonderful dream, told them what life would be like in the rich world of abundance that Bellamy foresaw for 2100.
And of all the marvels of wealth and abundance that Bellamy's protagonist sees in 2100 after he is transferred forward in time by a totally implausible plot device is this:
The house he winds up staying in in 2100 has its own telephone.
Not only does it have a telephone, but it has a speaker phone.
And not only does it have a speaker phone, but you can call up anyone on it.
And not only can you call up anyone on it, you can call up an orchestra that is currently playing in the city.
And you can put that orchestra on the speaker phone so that you can listen to music in your own home.
And you can not just listen to music--you have a choice: you can choose among four different orchestras.
And when confronted with this marvel of technology, Edward Bellamy's protagonist says:
It appears to me, Miss Leete, that if we could have devised an arrangement for providing everybody with music in their homes, perfect in quality, unlimited in quantity, suited to every mood, and beginning and ceasing at will, we should have considered the limit of human felicity already attained, and ceased to strive for further improvements...
Transformations like this are not in the consumer price index because the Bureau of Labor Statistics cannot count them. It can count by how much the prices of those things that people bought last year changed between last year and this year. It can weight those relative price changes by their share of spending last year. That is what the consumer price index does: it tells you how much more expensive than it was last year is it this year to buy our representative basket of goods and services. As new goods and services are invented, the BLS adds them to the basket gradually over the course of time. But the value of adding a new good to the basket--if making it possible to do something that previously had an infinite price--the Bureau of Labor Statistics cannot calculate that. It had an infinite price last year, so the decline in price from last year to this year is infinite. But it had a zero quantity last year, so you multiply infinity by zero to get the contribution of the invention of a new good to the consumer price index. And you cannot multiply infinity by zero to get a sensible answer. So the BLS does not try.
They do do the best job that they can.
Economists believe that the consumer price index by and large, understates how fast we're growing rich that it overstates the actual cost of living. We squabble over what the overstatement of inflation in the consumer price index is. Is it something like a percentage point per year? Maybe. Maybe it is more. Maybe it is less. Keep a CPI inflation-overstatement number of 1% per year in the back of your head. That is what I do.
Our inflation rate has not been even. See this 10% inflation rate rate at the start of the 1950s? This is the Korean War. In 1950 dictator Kim II-Sung of North Korea decided that it was time to reunify the country. He sent his Soviet Union-armed army south. First the American army and then the United Nations force intervened in response. The United States rearmed massively to fight first this hot war and then the Cold War. We had an enormous surge of government demand for goods that were already in short supply. Prices jumped by almost 10% in a year.
After the Korean War ebbed, the rate of inflation fell back to its normal level for the 1950s--1.5% or 2% per year or so.
And then, in the 1960s, inflation started to creep up. It crept up late in Lyndon Johnson's and into Richard Nixon's administration. At the end of the 1960s the Federal Reserve decided that inflation was too high. It decided to raise interest rates, raise unemployment, create a little excess supply on the labor market, create a few more workers looking for jobs who'll say, "I know you are paying this guy $8 an hour, but I'll work for $7 an hour," create a little more excess capacity in the factories so businesses will be a little bit hungrier for business and say, "We know that we were going to charge you $20 for this product but because we have excess capacity and want to sell more, we'll only charge you $19."
And so inflation falls.
The Federal Reserve thinks that it's gotten things under control.
It breathes a sigh of relief.
Then the 1973 Arab-Israeli War comes along. With it comes an oil embargo from the Middle East. The world price of oil triples. U.S. inflation rises to a peak of 12%.
The Federal Reserve hits the economy on the head with a brick again. It raises interest rates again. It so diminishes investment spending, GDP, and employment. With excess supply in the economy, the inflation rate falls back to 5%. The Federal Reserve concludes again that the situation is now under control and will return to normal. But the problem is that the late 1970s sees a new normal--with an inflation rate now of 6% per year rather of the 2% per year in the 1950s and early 1960s.
The economists in the Carter administration squabbled about what to do. When inflation that was at 1.5% or 2% per year, that didn't strike anybody as a problem. At 6% per year inflation, everyone has to change their prices all the time. Memories of what things ought to cost are no longer reliable. But reducing the inflation rate below 6% would require pushing the unemployment rate up. What was the best thing to do?
The Carter administration was still squabbling about what to do when the Iranian revolution arrived. As the Ayatollah Khomeini consolidates control the world price of oil again triples. Inflation in the United States hits a peak of 14%.
At this point newly appointed Federal Reserve Chairman Paul Volcker decides: enough is enough. Perhaps a 6% inflation rate was not something to worry about that much. But a 14% inflation rate certainly is.
Moreover, Volcker believed, the history of the previous 20 years had taught people that the inflation rate always went up.
It was necessary, Volcker decided, to disrupt this pattern of ever-rising inflationary expectations.
The next three years saw what we economists call the Volcker disinflation. The Federal Reserve did indeed raise interest rates and so hit the economy on the head with a brick to create excess supply, reduce inflation, and disrupt inflationary expectations. It keep hitting and hitting. The unemployment rate rose to its post World War II peak of 10.8%. And, when Paul Volcker was finally convinced that he had done enough to break the cycle of inflationary expectations, he eased off.
And he succeeded. He created our post-1984 relatively low inflation regime in which everyone expects there'll be a little bit of inflation. Inflation has been between 2% and 3% in average years--maybe a bit more in years when something unusual is happening like Saddam Hussein's invasion of Kuwait that caused another spike in oil prices, a bit less in depressed times like today.
That is the history of inflation in the United States over the past 60 years.
The next question to ask is: why is inflation a problem? In inflation, prices go up but incomes go up too. What's the big problem? If you have price stability, then your prices stay the same and your wages rise at the rate of real economic growth. If you have inflation, prices rise at the rate of inflation and wages rise at the rate of inflation plus the rate of real economic growth. How does that the fact that a dollar now is really the equivalent of a dime in 1950 or a nickel in 1913 hurt us? A dollar is the equivalent of a dime on both sides: both when we're earning money and when we're buying things.
It would seem to be a big zero.
Now it's not quite a big zero. Even though prices go up and incomes go up too, not all prices and incomes go up evenly. If you're holding dollars in your pocket or in a no-interest bank account, inflation is a clear loss to you. Inflation is a sneaky and underhanded way for the government to impose a tax on all of the cash that you have in your wallet. Inflation is also a big redistribution from those who did not expect inflation back when they made their contracts to those who did expect inflation when they made their contracts, and who as a result incorporated inflation allowances or inflation indexation into the bargains that they made.
More important, I think, is a loss of information carried by prices. When the price level is stable, you can go shopping for a refrigerator after a decade and you remember what a refrigerator should cost, so you can figure out whether this is a good deal or not without much effort. If prices are rising by an uncertain amount that varies from year to year between 5% and 15%, you go into a store after 10 years to buy a refrigerator and you have no idea what it should cost because all your memories of previous refrigerator shopping are inapplicable.
This loss of information carried by prices puts sand in the gears of the market mechanism. It is certainly not a good thing. How bad a thing--economists argue about that.
Most important of all, perhaps, is a redirection of effort and energy to gaming the system. People running businesses and making investments should be focusing on how to make their company's production operations more efficient. They should be focused on which lines of business are likely to produce the highest social returns for more investment. They shouldn't be focusing on what the rate of inflation is going to be, how they can forecast it, and how can they profit from or at least avoid loss from the inflation process. This redirection of effort and energy away from actually making things better and cheaper and more efficiently, to figuring out how to deal with the complexities of the monetary system is probably the biggest loss that comes from an ongoing process of inflation.
High inflation is a very big negative for politicians in office because inflation is taken correctly to be a sign of an incompetent government. A government that presides over high or even moderate inflation is a government that's likely to lose the next election. Thus politicians tend to be quite firmly focused on inflation control--especially after the defeats of Jimmy Carter and James Callaghan in the US and Britain at the end of the 1970s to opponents Ronald Reagan and Margaret Thatcher, who promised to control inflation and then did so.
How should we think about the process by which you get inflation? The easiest way to think about it is simply to go to the quantity theory of money. We all hold money in our pockets, in our bank accounts, in the unspent balances on our credit cards. We all hold money for one reason, to spend it.
If you don’t want to spend it, there are other better things to hold your wealth in than cash money. It is usually much better to hold it in a five-year treasury bond or a five-year treasury index security than to hold it in cash. Money is dominated as an asset if you're holding it for any purpose other than spending it.
And so as economist Irving Fisher a century or more ago said, money has a velocity for which we use the letter V. And the velocity V of money is the pace at which people are used to spending the money in their pockets. V tends to go up when interest rates rise. V tends to go down when interest rates fall. V changes as the banking technology of the economy changes. V is larger if more spending is done by the government, etc, etc. But you take the total amount of money in the economy. You multiply it by the velocity of money (V). You divide it by the level of real spending in the economy (Y). You get the price level (P).
Want to keep the price level from going up? Keep the money stock (M) from going up. And that was Milton Friedman's central career insight: that inflation is ultimately a monetary phenomenon.
How exactly does adding to the stock of money boost the price level? What happens when the central bank expands the money supply? It engages an open market operations: it buys bonds for cash. That pushes down the interest rate on bonds. That tends to make investment and exports grow.That pushes to push up the level of demand in the economy. Workers demand higher wages. Firms find that they can insist on higher markups. And the price level rises.
So there is a link between the income-expenditure model of how the central bank affects production and the inflation economics quantity-theory model of how the central bank's increases in the money stock influence the price level. To look at that link we would like a nice measure of the supply and demand balance, or the gap between the economy's current level of production (Y) and its potential level of production (Y*) so that we could see whether income and expenditure are putting upward or downward pressure on inflation. And we're lucky in that we do have such a measure: the unemployment rate.
Recall Okun's Law: the graph that plots the relation between changes in real GDP along the horizontal axis and changes in the unemployment rate along the vertical axis. Take the change in output relative to the current level of output, subtract off the population growth rate, subtract off the rate of growth of the efficiency of labor, and then multiply the result by -0.5. That is what the change in unemployment should be over a year.
This relationship has broken down somewhat over the past two years. But, by and large, when the level of production in the economy is growing faster than potential output, unemployment will fall. When the level of production is falling or growing more slowly than potential output, unemployment will rise. So we are going to focus on the relationship between the unemployment rate and the change in the price level--the rate of inflation.
We start from our income expenditure model and figure out how fiscal policy, monetary policy and the economic environment have together produced the level of real GDP in the economy. We can compare that to the level of potential output and use Okun's Law to calculate the unemployment rate. But how then do we go from the unemployment rate to the level of inflation?
We write down another equation:
π = E(π) + b(u*-u)
We write this little π for the inflation rate--the proportional rate of change in the price level. Why π? Because we wrote P for the price level. π is the same letter in Greek as in P as in Latin. We're going to write E(π) for the expectation of inflation rate, for what people in the market thought inflation would be when they made their plans and bargains six months or a year or two years ago, or whenever We write u for the unemployment rate. We write u* for the natural rate of unemployment--the non-accelerating inflation rate of unemployment--the rate of unemployment at which production is equal to the economy's potential output and markets are in balance, with producers and purchasers neither being surprised by excess supply or surprised by excess demand.
We write these variables down in this equation:
This equation tells us that the inflation rate this year is going to be what people expected the inflation rate to be plus this parameter b times the difference between the natural rate of unemployment and the actual rate of unemployment.
π = E(π) + b(u*-u)
We think about this equation like so: We think that people last year made their plans about what they were going to ask for in terms of wages and charge in terms of prices. Then this year rolls around. Either the unemployment rate is higher than the natural rate of unemployment or it is lower. Either capacity utilization is lower than its normal value or higher.
If the unemployment rate is higher than its natural rate and if capacity utilization is lower then workers say: "Oh, wow there are lots of unemployed workers out there who might compete with with me for my job. I shouldn't ask for wage increases as high as I had planned on. I should go to the bosses and volunteer to accept a 2% lower wage increase this year--because we understand that we could be replaced and we also understand that business is slow. So whenever capacity utilization is low and unemployment is high, actual inflation is going to undershoot expected inflation. Conversely, whenever the unemployment rate is low and capacity utilization is high, actual inflation is going to overshoot expected inflation.
That is the Phillips Curve.
There are other things that you could throw into this equation. Most important are supply shocks that push up some prices without lowering others: the triplings of world oil prices in 1973 and 1979 had direct effects boosting inflation even though they did not act directly on either expectations or on the rate of unemployment.
How well does this Phillips Curve do? Most of the time expectations of inflation are what economists call adaptive. People just take a look at what the current rate of inflation is and they say inflation next year is probably going to be about the same as inflation this year. Thus people adapt over time if inflation goes up. If inflation this year is higher than what inflation was last year, then people looking forward two years from now are not going to expect that inflation will return to its old value. Adaptive expectations are a pretty good rule of thumb for the United States recently, but not a good rule of thumb for places where inflation is doing something weird or getting out of control--like say Zimbabwe in the late 2000s.
If expectations of inflation are adaptive, then the change in the inflation rate should be strongly correlated with the unemployment rate. When unemployment is high inflation should be falling--if the Phillips Curve is worth much. When unemployment is high inflation should be rising--if the Phillips Curve is worth much.
How does it do?
This is the graph we get for the post-World War II U.S. Not the tightest fit in the world, but by and large and on average unemployment rate is bigger than 6.5%, you can bet that the inflation rate is going to fall. Whenever the unemployment rate is less than 5% you can bet that the inflation rate is going to rise. In between it is not clear whether inflation is more likely to rise and fall.
That is our uncertainty about what the natural rate of unemployment is--and the natural rate of unemployment shifts over time.
This is a relationship that you can take to the bank--not in six months, not in one year, but over a five-year period.
This is the Phillips Curve.
If we are going to graph the Phillips Curve, you need to know three things: the current expected rate of inflation, the natural rate of unemployment, and the slope parameter β. The Phillips Curve goes through the point on the graph where inflation is equal to expected inflation and unemployment is equal to the natural rate--in fact, you might as well take those as the definitions of the natural rate and of expected inflation. And the parameter? We simply look and see. There are theoretical reasons to believe that the more volatile inflation is the steeper the Phillips Curve will be, but we have no good derivation from microeconomic foundations or from behavioral foundations of what the slope of the Phillips Curve is. We just have the data. And the data says: about 1/2, for the U.S. since World War II.
If expected inflation rises, if something happens that makes people think that inflation will be higher than they'd thought before--like the election of François Mitterrand to the presidency of France at the start of the 1980s--the Phillips Curve will jump up. If expected inflation falls--as happened after Paul Volcker had hit the economy on the head with a brick to create excess supply to make firms stop raising their prices, and after people realized he was serious--the Phillips Curve will shift down.
Let me just make one more note: Japan in the 1990s. The actual inflation rate hit 0 and then went a little bit further and stuck. It didn't fall much below 0 even though expected inflation was 0 and unemployment was far above Japan's natural rate. This Phillips Curve relationship does not seem to apply to today to situations of inflation, when consumer prices are actually falling. But in the United States during the whole period since World War II the inflation rate has been positive with the exception of one quarter in 2009.
Another note about what determines the location of the Phillips Curve. When inflation is above expected inflation we expect the Phillips Curve to jump up over time. When inflation is below expected inflation we expect the Phillips Curve to shift down over time. But there are times when inflation expectations are what the Federal Reserve calls "well anchored": just because inflation is high or low this year doesn't mean people are going to change their expectations of inflation. They have confidence the Federal Reserve has a stable inflation target and is going to push the economy to that target.
The natural rate of unemployment is usually stable: 5 or 6% in the US today. Sometimes it won't be. Back in the 1970s when baby boomers like me were entering the labor force, and people in their 20s, they're really lousy at job search. They don't know how to work the system. They tend to mouth off in front of their bosses and get fired. A labor force that has a lot of twenty-somethings will have people who won't keep their jobs very long on average and will have a hard time finding their jobs after they lose them. So we should expect the natural rate of unemployment in an economy to be higher when a baby boom generation is in its 20s than when a baby boom generation is like now in its 50s. And that seem to be true. That the natural rate of unemployment fluctuates with the demography.
And there are worries about structural unemployment. If the structure of the economy is changing rapidly so that a bunch of industries are shrinking rapidly and a bunch more are growing rapidly, those workers who are losing their jobs in the declining industry is going to have a hard time picking up the skills needed in the expanding industries. Won't that push up the natural rate of unemployment? Yes.
And if your unemployment rate is high for a long time--as it is going to be in the United States, people who would normally be busy finding and working at jobs won't really be looking for work or will have lost so many of their job related skills or have lost so much of the ability to actually show up at 9 o'clock five days a week that people will be loathe to hire them. Doesn't a prolonged period of high unemployment itself raise the natural rate of unemployment? Yes.
If you look at the actual history of unemployment and inflation in the United States since 1950 or so, you get what at first glance looks like a complete and total mess. But you can use the idea of the Phillips Curve to make sense of the dynamics of unemployment and inflation in the United States over the past 60 years.
Starting in 1950 we see the Korean War start. By 1951 the unemployment rate is down to less than 3.5% because of expansionary fiscal policy, and our inflation rate is up to 8%. The inflation rate comes down after this big boom of wartime related inflation is over. And then from 1955 to 1960, you can perform--– you see the economy performs a loop. As the unemployment rate falls, the inflation rate tends to rise, as the unemployment rate rises, the inflation rate tends to fall. And by the start of the 1960s, you're more or less back to normal.
So you look at the economy in the 1950s and, yes, there is a relationship between unemployment and inflation, there's a Phillips Curve. It's not a terribly tight relationship, but those times when unemployment was higher were times when inflation was lower, and those times when unemployment was lower where times when inflation was higher.
In the 1960s the Phillips Curve fit is much tighter. Unemployment falls, and inflation rises. It looks like there was a quite stable Phillips Curve in the 1960s.
And then in 1970 something happened. Things went haywire. Unemployment rose and inflation rose too.
What happened in 1970 was the breakdown of static inflationary expectations. Recall that the position of the Phillips Curve is going to change if either the natural rate of unemployment changes or if expected inflation changes. The natural rate of unemployment wasn't changing much over this time period. Neither was the economy's expectations of inflation--until 1970.
In 1970 people wised up. They noticed that inflation was rising, and so they began the process of raising their expectations of inflation. Year by year they raised their expectations of inflation and year by year the Phillips Curve headed up.
And let me stop there. Have a good vacation. I will see you all in a week and a half.