State of Inflationary Confusion

No One Is Average

In the US we have two different inflation measures, produced by agencies of two different cabinet departments. The Federal Reserve prefers to look at the Commerce Department’s Personal Consumption Expenditures (PCE) Index, because they believe it is more comprehensive and nuanced than the Labor Department’s Consumer Price Index (CPI).

Are they right? In a moment I will talk about the differences, which are important; but I think the Fed is exactly backwards here. Neither measure is particularly foolproof, but the flexibility and adjustments that make the Fed prefer PCE also take the index further from reflecting the average citizen’s economic condition. This bias shows up in Fed policy, and not in a good way.

That doesn’t mean the CPI is wonderful, though. Unlike some, I don’t believe it is intentionally manipulated. I think the wonks (and I say that in a complimentary way, as a fellow wonk) who compile price data do a nearly impossible job as well as anyone can. Browse through the methodological explanations on the CPI home page and you’ll quickly see how much effort goes into that work. They have a whole “data available” shopping list:

• Price indexes are available for the US, the four Census regions, nine Census divisions, two size of city classes, eight cross-classifications of regions and size-classes, and for 23 local areas. Indexes are available for major groups of consumer expenditures (food and beverages, housing, apparel, transportation, medical care, recreation, education and communications, and other goods and services), for items within each group, and for special categories, such as services.

• Monthly indexes are available for the US, the four Census regions, and some local areas. [You can see those indexes here.]

• More detailed item indexes are available for the US than for regions and local areas.

• Indexes are available for two population groups: a CPI for All Urban Consumers (CPI-U) which covers approximately 94 percent of the total population and a CPI for Urban Wage Earners and Clerical Workers (CPI-W) which covers 28 percent of the population.

• Some series, such as the US City Average All items index, begin as early as 1913.

All that data gets worked into “baskets” that try to match the spending habits of typical consumers. That’s where the effort starts going wrong. The problem is quite simple and beyond anyone’s control: None of us are average.

We all spend our money differently, for an endless variety of reasons that change all the time. When you say inflation is higher than CPI shows while your neighbor says inflation is no big deal, you can both be right. Worse, even someone with spending patterns identical to yours can experience an entirely different inflation rate simply because they live in a different city or state. Or they choose to send their kids to a more expensive school. Or they spend a larger amount on health care and less on goods but more on services. It can get quite nuanced.

Reducing this complexity to one number and then using that number to guide monetary policy is asking for trouble. And trouble is what we get.

Hedonic Guesswork

CPI isn’t entirely useless. It can show us broad price trends over long periods. Those trends can reveal some things, as shown in this 20-year American Enterprise Institute chart that’s making the rounds this month.

What jumps out to me is that the highest inflation is in the goods and services over which people have the least discretion. This is particularly burdensome to lower-income Americans. The disinflation that so vexes the Fed impacts more optional purchases. Here’s how my friend Barry Ritholtz describes the pattern:

It is notable that the two big outliers to the upside are health care (hospital, medical care, prescription drugs) and college (tuition, textbooks, etc.).

Clothes, cars, TVs, cell phones, software – technology in general – showed disinflation or outright deflation in prices. (Housing and food & beverage have been right at the middle of inflation levels.)

Wages have barely ticked over the median inflation measure, but that did not stop some people from blaming the correction on rising wages.

Reading the pundits, I cannot tell which fate awaits us: the robot-driven apocalypse where we are all out of work, or the inevitable spike in wages that sends rates much higher and kills the market. Perhaps both – higher wages sends employers into the waiting arms of our automated future.

You can quibble with this data. Have TV prices really fallen 99%? No, unless you hedonically adjust, because today we can buy TVs of a quality that didn’t exist in 1997. If you use hedonic prices, adjusting for quality and technological sophistication, then you can argue that the price of TVs is down 99%. But we all know that we are paying less for TVs. Same for other technology goods. But you simply cannot argue that we are paying the same now for new vehicles as we did 20 years ago, even though the cars we buy today are technologically vastly superior to what we could buy back then. These hedonic price adjustments are guesswork.

Still, the broader point seems right. Inflation is a real problem for some people and no big deal for others, yet the Fed uses inflation measures to impose a single monetary policy on everyone. Is it any wonder that policy doesn’t work for many of us?

PCE Versus CPI

The Federal Reserve prefers to use core PCE rather than core CPI. The Cleveland Fed has a very good basic explanation of the differences between the two indexes. A glance at their charts, below, will show that core PCE (Personal Consumption Expenditures) is significantly lower than core CPI (Consumer Price Index). After the charts, I will quote a few paragraphs from the Cleveland Fed.

What accounts for the difference between the two measures? Both indexes calculate the price level by pricing a basket of goods. If the price of the basket goes up, the price index goes up. But the baskets aren’t the same, and it turns out that the biggest differences between the CPI and PCE arise from the differences in their baskets.

The first difference is sometimes called the weight effect. In calculating an index number, which is a sort of average, some prices get a heavier weight than others. People spend more on some items than others, so they are a larger part of the basket and thus get more weight in the index. For example, spending is affected more if the price of gasoline rises than if the price of limes goes up. The two indexes have different estimates of the appropriate basket. The CPI is based on a survey of what households are buying; the PCE is based on surveys of what businesses are selling.

Another aspect of the baskets that leads to differences is referred to as coverage or scope. The CPI only covers out-of-pocket expenditures on goods and services purchased. It excludes other expenditures that are not paid for directly, for example, medical care paid for by employer-provided insurance, Medicare, and Medicaid. These are, however, included in the PCE.

Finally, the indexes differ in how they account for changes in the basket. This is referred to as the formula effect, because the indexes themselves are calculated using different formulae. The details can get quite complicated, but the gist of the matter is that the PCE tries to account for substitution between goods when one good gets more expensive. Thus, if the price of bread goes up, people buy less bread; and the PCE uses a new basket of goods that accounts for people buying less bread. The CPI uses the same basket as before (again, roughly – the details get complicated).

Now, in conversations with my friend and fellow wonk Peter Boockvar, he has pointed out other, more nuanced differences. The inclusion of government-priced medical care, such as Medicare, where the government manipulates what they will pay, significantly reduces the healthcare inflation of the PCE. And as noted above, the PCE assumes that if the price of something – beef, for instance – goes up, consumers will buy less beef and more chicken, which is cheaper.

You can make the argument that the PCE is biased toward returning lower inflation numbers, but that tendency is almost beside the point. The technical differences between the two indexes make for extraordinarily economically dense discussions, and I’m sure the issues are debated heatedly at certain conferences that focus on such things, but both measures are honest attempts to understand what inflation is and how it affects us. Neither index necessarily reflects the inflation that you are personally experiencing, or the inflation of your particular area or region. And similar differences pertain in every country of the world.

But in most countries, inflation affects the bottom 50% more than it does the top 50% by income. Because there are certain necessities of life that must be purchased, and because many of those goods and services (such as housing, and health care) have higher than average inflation, the bottom half suffers a much higher inflation rate than the overall national average.

And yet, a national inflation policy geared to the lower 50% would aggressively skew monetary policy in a negative fashion.

Sidebar: PCE and CPI use different measures and percentages for housing costs. CPI uses something called Owner’s Equivalent Rent, which is a hypothetical number based on certain assumptions. Here’s a thought project. At one time the US used actual house prices to measure inflation, as Europe does today. If we had been using actual house prices during the period 2000 to 2008, the Fed would have been raising rates aggressively, trying to lean into the inflation caused by the increase in house prices, thereby likely avoiding the housing bubble but creating a recession earlier than 2008, for entirely different reasons.

Tell me again why 12 people sitting around a table should set interest rates based on data they don’t understand, in a market that is way too complex? More loans are based on LIBOR than anything else. And we trust a rather complicated market process, which can somewhat be manipulated, to set the price of LIBOR. Manipulating interest rates in the broader market would be far more difficult and would lead to interest rates that are more reflective of what is going on in the marketplace. Just saying…

Disinflation Fixation

In theory, we want “price stability,” which would mean the absence of either inflation or deflation. When Greenspan was asked, when he was chairman of the Fed, what is meant by price stability, he answered “Zero.” None of this 2% inflation target. “Price stability” is the obsession of central bankers everywhere, and in some places is their legal mandate. They currently like having just a little inflation but not too much. The Fed wants 2% and can’t even deliver that, if you define inflation by CPI or PCE. People think that 2% coming. Maybe so.

However, maybe we should all think about this issue differently. Last week I ran across a December 2017 Project Syndicate article by good friend William White, formerly Bank of International Settlements chief economist and now with the OECD. Bill is my favorite central banker in the world. (The full article is well worth reading.)

White says central banks rightly responded to 1970s inflation by clamping down hard, but then failed to adjust when conditions changed. That oversight led directly to some of today’s problems.

From the late 1980s onward, low inflation was largely due to positive supply-side shocks – such as the Baby Boomer-fueled expansion of the labor force and the integration of many emerging countries into the global trading system. These forces boosted growth while lowering inflation. And monetary policy, far from restricting demand, was generally focused on preventing below-target inflation.

As we now know, that led to a period of easy monetary conditions, which, together with financial deregulation and technological developments, sowed the seeds of the 2007 financial crisis and the ensuing recession. The fundamental analytical error then – as it still is today – was a failure to distinguish between alternative sources of disinflation.

The end of the Great Moderation should have disabused policymakers of their belief that low inflation guarantees future economic stability. If anything, the opposite has been true. Having doubled down on their inflation targets, central banks have had to rely on an unprecedented array of untested policy instruments to achieve their goals.

The fixation on keeping inflation low, according to White, has driven up global debt ratios, squeezed lender margins, and forced lending activity into an opaque shadow banking sector. All these effects raise systemic risks that will probably bite us eventually.

Here’s Bill again, with a thought-provoking point I’ve formatted in bold.

These developments constitute a threat not just to financial stability, but also to the workings of the real economy. Moreover, one could argue that easy money itself has contributed to the unexpectedly strong disinflationary forces seen in recent years. Owing to easy financing and regulatory forbearance, aggregate supply has risen as “zombie” companies have proliferated. Meanwhile, aggregate demand has been restrained by the debt headwinds – yet another result of easy monetary conditions.

This insight isn’t intuitive to many economists. Easy credit – as the Fed gave us for the last decade – should raise inflation, not reduce it. Bill says no; it allows zombie companies to survive and overproduce, while putting consumers in so much debt that their spending gets constrained. So it pushes inflation down instead of up.

Wrap your head around that thought. It answers some riddles that otherwise make little sense. But it also highlights the difficulty of formulating sane policy. Yes, it’s important to let zombie companies die. Creative destruction and all that. But real people work for the zombies, earning real money that lets them buy goods and services and keep the economy moving. So what do you do? None of the choices are painless.

Too often, we simply redistribute the pain to those least able to bear it, who are understandably unhappy – hence the present social and political tensions. They all trace back to economics.

Is data boring? Yes. It’s often wrong and misleading, too. But ignoring it to fly by the seat of our pants isn’t the right response, either. We need much better understanding and application of all these numbers, and I see very few economists trying to deliver either. That’s the core problem.

I am getting close to going on too long here, so the prescription for what we should be doing will come in future letters. Suffice it to say that using data that is fundamentally flawed as a “guide” to monetary policy creates the rather strange outcomes that we see. I get extraordinarily angry when central banks and big government proponents argue that it is capitalism and free markets, rather than manipulation and inappropriate regulation, that are the problem. The monetary policymakers never see themselves as the problem. This blind spot is just a corollary to one of my favorite Paul Simon quotes: “A man hears what he wants to hear and disregards the rest.”

Dallas, SIC, and San Diego

This last week was rather exhausting, with late planes and long flights. But I am now back in Dallas and think I will take a little time to relax before plunging into the week on Monday morning. With the exception of speeches and business all day Tuesday, my focus will be on the final plans and my topics for the Strategic Investment Conference in San Diego March 6–9. The conference will be well attended, and as usual almost half of the attendees have been to more than a few SIC conferences, so I will be among old friends. I will spend much of the week talking with the speakers and getting a sense of what they will say, so that I can make sure that I have everything and every topic in its proper order. I must say, what I have heard so far has shown me a surprise here and there, as it seems many are expecting changes and adapting their own businesses and outlooks. If nothing else, it promises to be a week of revelation.

I know my editors are working to pare back my letter, as we are trying to keep it shorter, so I’m going to help them this week by limiting my personal comments to wishing you a great week and hoping that you will find time to be with friends and family.

Your ready for a little rest analyst,

John Mauldin

[email protected]

© Mauldin Economics

www.mauldineconomics.com

© Mauldin Economics

Read more commentaries by Mauldin Economics