Inflation, explained

Inflation refers to a general rise in the level of prices. Its opposite is deflation, a general fall in the price level.

It's important to distinguish these ideas conceptually from the rise or fall of the price of some particular class of goods. If there's a huge fad for kale so people start eating more kale and less spinach, leading to soaring kale prices and falling spinach prices, that's not "kale inflation" and "spinach deflation." It's shifting tastes. Inflation is when everything gets more expensive.

The difference was in some ways easier to understand in the days when the value of currency was pegged to the price of gold. In most of the 19th Century, one dollar was defined in terms of a certain quantity of gold. When global production of gold surged, the value of gold relative to other commodities would fall. That meant dollar-denominated prices of just about everything would rise. That was inflation. Conversely, when the world went a while without any major new gold discoveries the price of gold relative to other commodities would rise. That meant dollar-denominated prices of just about everything would fall. That was deflation.

Today, currencies are not linked to the price of gold so the causal mechanisms of inflation are more complicated.

But the same principle applies. Inflation in the United States is a systematic fall in the price of dollars relative to goods and services, not just an increase in the price of something that happens to be important to you.

Inflation is also not the same as a fall in living standards. Inflation was generally higher in the late-1960s than its been in the 21st century, but in the late-sixties wages rose considerably faster than prices leading to rising living standards. In the 21st century, both wages and prices have risen slowly leading to very small increases in living standards.

In the United States, the main measurements of inflation are various kinds of consumer price indexes. These are ultimately based on the work of a small army of individuals who are employed by the Bureau of Labor Statistics to go around to different stores and look at what different items cost. They look in a great deal of geographical detail at both prices and locations. So they can tell you what canned corn costs versus frozen corn, and what corn costs in Chicago versus in Miami.

When the data-gathering is done, the number crunching begins.

To go from a giant list of prices to a useful price index you need to decide how much weight to give to all your different data points. Americans eat a lot more chicken than lamb, so a change in the price of chicken is more important than a change in the price of lamb. All the different prices tracked are weighted according to their size in the overall consumption of the American population. BLS decisions about how important the price of different things are may vary considerably from your personal experience. Gasoline and milk are both a big deal to the average American consumer, but if you're lactose intolerant and walk to work they may not be very important to you.

The other bit of number-crunching is the effort to adjust for differences in quality.

During the mid-aughts when most Americans were shifting from low-definition televisions to higher-definition televisions, many people found themselves buying more expensive HDTVs. A very superficial analysis might have concluded that the price of televisions was rising. In reality, the price of both HD and non-HD televisions was falling. BLS quality adjustment methods are intended to take into account this kind of shift and record it properly as a decline in prices. In the case of goods that have well-defined feature sets (like televisions) this can be done in a reasonably unambiguous way but it can be more challenging for other categories.

Milton Friedman famously wrote that "inflation is always and everywhere a monetary phenomenon." In other words, inflation is always a consequence of monetary policy. Prices are rising because the people in charge of managing the currency are choosing to allow prices to rise.

This is an important theory of inflation because it certainly doesn't seem to be true.

It sure looks like lots of things cause inflation. China gets richer and its citizens develop a taste for beef, so beef prices rise. Instability in Iraq curtails oil production so gasoline prices rise. And these commodity price hikes can propagate through the economy. Since doing almost everything involves some transportation, higher gasoline prices can put pressure for higher prices on all kinds of goods and services. Next thing you know, inflation is everywhere.

Friedman's argument — especially relevant in the 1970s when oil crises were quite severe — is that this is an illusion. Those kind of events can cause the price of specific commodities to rise, but with appropriate monetary policy there's no reason that average prices should rise. Adequate control of the money supply would simply mean that higher gasoline prices are offset by falling prices for other things. Oil scarcity should make oil more expensive relative to things that aren't oil, but there's no reason it should lead to a systematic decline in the purchasing power of dollars.

An important 1997 paper by Ben Bernanke, Mark Watson, and Mark Gertler argued that trying to stabilize prices under these circumstances would be a mistake. They argued that the impulse to respond to commodity price shocks (and especially oil shocks) by reducing the money supply to curtail inflation tends to exacerbate recessions unnecessarily. Their argument, in essence, is that while you certainly could address a crisis of gasoline affordability by raising interest rates to create more unemployment to push the price of other things down as unemployed people cut back on spending it's not clear why you would want to. Basically you're taking a bad situation and "offsetting" it by creating a second bad situation.

However, even though Friedman and Bernanke are giving different policy advice they are not really disagreeing about the inflation. In both cases, it is the Middle Eastern crisis that creates the policy dilemma — do you let prices rise or do you let unemployment rise? And in both cases it is the decision to resolve the dilemma in favor of price increases that leads to the inflation.

Different government agencies produce a lot of different price indexes. Broadly speaking, these indexes all show the same large-scale trends over time. Inflation was high in the 1970s and has been low recently. Electronics have gotten cheaper and medical care has gotten more expensive. But the indexes measure somewhat different things, and do diverge over time and are especially likely to diverge over longer spans of time.

Indexes differ in the following main ways

Here are some of the most important price indexes that you'll see in the news

The Consumer Price Index is the most frequently-reported and widely discussed inflation index in the United States. It covers the population of Americans who live in metropolitan areas — cities and suburbs. It covers prices that consumers pay directly. The price of business equipment is not considered. Nor is the price the government pays for things. Nor are indirect prices considered. If your health insurance company raises your premiums or the government raises your bus fare, that's in the CPI. If your doctor starts charging your insurer more or if the buses the transit agency buys get more expensive, it isn't. Of course in the long-term, more expensive buses and higher doctor prices should pass through to consumers but in the short-term they might not.

CPI handles owner-occupied housing with what's called owners' equivalent rent. Basically, they look at what it would cost you to rent a house that's similar to yours and then count that as the price you are paying to yourself in your role as your own landlord. The actual sale price of houses is ignored. Since most people don't rent, this means weird things can happen during housing bubbles or busts — the sale price of houses can plummet while the CPI says houses are getting moderately more expensive.

CPI is used in the United States to set benefit levels for Social Security and some other social welfare programs, and is also used to set tax bracket levels. Importantly, even though it is widely reported in the press it is not used by the Federal Reserve to make monetary policy.

A common suggestion in policy circles is that the government should stop using the CPI and switch instead to a chained version of the index. A chained index more aggressively accounts for consumer substitutions between similar kinds of goods. If bad weather destroys the pear crop, pushing pear prices up an unchained price index would record that as a healthy dose of inflation. A chained index observes that many consumers will simply eschew these expensive pears and eat apples instead, indicating a more modest dose of inflation.

Virtually all economists agree that chained price indexes do a better job of capturing the concept of monetary inflation. But the consequences of shifting from CPI to Chained CPI — lower Social Security benefits and higher taxes — are very controversial, so actual policy proposals around chained versus unchained are contentious.

The PCE Deflator is used by the Fed when setting monetary policy, and also by the Commerce Department in reporting elements of GDP.

It differs from the CPI in two main ways. One is that it is a chained index. The other is that rather than measuring prices paid by consumers it measures prices of things that consumers consume. In other words, rather than measuring your copayment for a tooth cleaning it measures the price the dentist charges for a tooth cleaning. The practical upshot of this is that medical care is a bigger element of the PCE basket. Since the health care industry is, in fact, a large segment of the American economy the Federal Reserve feels this broader index is a better gauge of monetary conditions.

If you ever read a reference to inflation in a European country, you are almost certainly reading about the Harmonized Index of Consumer Prices (HICP). HICP is used by the European Central Bank and the Bank of England to set monetary policy, and is also widely used to index various forms of government benefits.

HICP differs from US consumer price indexes in two ways. One is that it seeks to cover rural as well as metropolitan consumers, overcoming the logistical difficulties inherent in covering rural prices with a statistical sampling method. The more important difference is that rather than imputing rental prices to owner-occupied housing, HICP simply leaves owner-occupied housing out. That means housing is considerably less-weighted in Europe than in the United States.

A separate issue is that the consumption basket of the average European is quite different from the average American. Europeans spend much less on health insurance premiums and out-of-pocket medical care, and also have lower after-tax incomes, so basic commodities like food loom much larger in European consumer price trends.

The broadest measure of inflation is the GDP Deflator. This is a chained index used by the Commerce Department and other agencies to report on trends in national output.

It differs from more commonly discussed price indexes in that it considers the price of everything, not just the price of stuff consumers buy. The price of business equipment, in particular, is part of GDP. In many cases, an increase in business equipment prices would be passed on to consumers. But many companies operate internationally, so this is not necessarily the case. Since businesses buy a lot of computer hardware, and since the price of computer hardware has tended to fall, the GDP deflator is generally lower than consumer price indexes.

Most major price indexes are sometimes published in terms of a "core" sub-index that leaves out food and energy prices. There are a lot of myths around this, including the myth that when the Federal Reserve thinks about inflation it's ignoring food and energy prices.

What's actually happening is this. Central banks try to achieve a low level of long-term inflation, including all goods. But to do this, they need indicators of likely future inflation. Most central banks think that food and energy prices swing too wildly to be useful future predictors of inflation. That's primarily because these prices have a lot to do with the weather. An unusually hot summer leads to more air conditioner use and higher electricity prices. An unusually cold winter leads to high heating oil prices. An early freeze ruins the orange crop and leads to expensive fruit.

All this stuff matters to people's lives. But the weather is unpredictable. An unusually cold winter doesn't tell us anything about likely conditions in the fall. So in the short-term, it often makes sense to ignore these short-term swings.

Methodologically speaking, throwing out food and energy prices is pretty crude. The Cleveland Federal Reserve Bank publishes more mathematically sophisticated price indexes like the trimmed-mean CPI and the median CPI that attempt to achieve the same goal. These work by essentially throwing out whichever prices happen to be fluctuating most wildly in any given month, rather than judging in advance that food and energy are excessively unstable. The idea is that outlier price swings in any set of goods is probably due to some kind of strange one-off that doesn't tell us much about the future.

The short answer is that they don't. An inflation index can tell you that the price of an iPhone 5 has fallen considerably since the introduction of the iPhone 5S. But it can't tell you much about the fact that 10 years ago there were no touchscreen smartphones with LTE service, whereas today such smartphones are broadly accessible to the middle class.

This is not, per se, a methodological failing. The job of a price index is to track prices not to track innovation.

But it's important to keep this in mind, especially when considering price trends over very long time horizons. You can use the CPI to translate a 1964 income into a 2014 income, but that doesn't change the fact that many classes of goods were unavailable 50 years ago. This is most obvious in the realm of digital gadgets. But it's also true of important aspects of medicine (cochlear implants), transportation (airbags), foods (pluots, broccolini), and other realms of everyday life.

This chart shows William's view of the true CPI (in blue; he seems to just be adding a basically arbitrary fudge factor) versus the official one:

The implication of Williams' view — spelled out in this chart — is that the total output of the American economy has been steadily shrinking since about 2000. This notion of a continuous generational recession is hard to square with other data and aspects of human behavior. There are more people with jobs in America today than there were 15 years ago. There are also more houses and more cars on the road.

If America's larger population and workforce are producing less total stuff, what is everyone doing at work all day? And how are American companies earning profits? Why has the stock market gone up? Why do immigrants keep coming here? Governments really can manipulate inflation statistics, but they can't brainwash people in this way.

MIT's Billion Prices Project scrapes online retail websites to do a fast, automated inflation index with somewhat different coverage than the official government CPI. BPP really does show an inflation rate that differs from the government inflation rate. But it is much closer to the official rate than to Williams' rate. And the tendency is for CPI and BPP to converge in the medium-term, suggesting that despite the methodological differences they are both capturing the same underlying trends.

A person with good credit can obtain a 30-year mortgage at an interest rate that is considerably below Williams' estimate of the true rate of inflation. That means that if Williams is right, the inflation-adjusted interest rate on a 30-year mortgage is negative and has been consistently negative for years. Many auto loan rates would also qualify as negative in inflation-adjusted terms if Williams is correct.

If true, this would represent a major blunder on the part of every single bank in America. Williams thinks that JP Morgan, Wells Fargo, Bank of America, etc. are all in the business of handing out free money to their customers. Of course, bankers make mistakes. But this would be a weird one. Major banks have lots of money at stake in getting this right, and employ many economic researchers to help them understand issues that are relevant to their business.

A somewhat confusing bit of economics jargon is that inflation-adjusted prices are often referred to as real prices as opposed to nominal prices that are not adjusted.

Use of inflation adjustments is often important to get historical comparisons right. The 1982 film ET earned just under $800 million at the box office, considerably less than the $1.2 billion earned by 2013's Iron Man 3. But adjusted for inflation, ET's earnings were more like $2.2 billion in today's dollars. The use of nominal figures to assemble lists of all-time top earners will tend to bias any list toward recent projects. Inflation adjustment can improve that.

On the other hand, one complexity with so-called "real" prices is that in the real world almost all prices are nominal. It is certainly possible to write contracts that include automatic CPI adjustments. But it is typical for mortgages, salaries, 2-year cell phone contracts, and a vast array of other important economic transactions to be conducted in nominal terms. Obsessive application of inflation statistics can obscure that.

It can also obscure the importance of relative price shifts. The fact that movie tickets have gotten a lot more expensive since ET's day is itself an interesting fact about the movie industry. Other forms of entertainment — televisions, for example — have gotten cheaper.

On an individual level, people dislike inflation for the obvious reason that it's nice to buy things cheaply.

But this logic doesn't really apply on a social level. What's nice for you is for your (nominal) income to rise faster than the (nominal) prices of the things you buy. But if the price of everything falls, then everyone's incomes are going to have to fall too. Your expenses are an income stream for the people you buy things from. And your income is someone else's expense.

So economists' concerns about inflation are different from the ordinary person's search for a bargain. The basic problem with inflation, from a social viewpoint, is that when inflation is high everyone needs to spend a lot of time thinking about inflation. Any salary negotiation, any price decision, and any medium-term contract needs to build in assumptions about inflation. Success in business ends up having a lot to do with skill or luck in forecasting inflation, rather than in management of real enterprises. And not since everything will be pegged to inflation (or pegged to the right index…) lots of distortions will enter the economy.

By contrast, low and stable inflation lets us do what do now — go about our daily lives not worrying much about inflation.

Note that while this is a very real problem, the economist's worry about inflation is also a fairly modest one. The ordinary person's worry about the "rising cost of living" (prices rising faster than incomes) is a much more serious problem than the economist's worry about calculation costs. But even though both issues are related to prices, only the calculation costs problem is truly a question of inflation per se. In the 1990s, the inflation rate was considerably higher in the United States than in Japan. But overall US economic growth was robust, while Japan was mired in a depression. Under those circumstances, Japanese people had much more cause to worry about the rising cost of living than did Americans.