Author: chogan@directresearch.com

 

The biggest eye-opener that I ginned up in the last year was a graph of new and used car prices.  As above.

Functionally, all it told me is that official U.S. price statistics are worthless for tracking trends in how much you have to pay to buy a car.

But, as it turns out, if you look a little deeper, the goofy official U.S. car price data have a lot of company.  That is, many items in the U.S. Consumer Price Index now have “quality” adjustments of the type that generate those odd car price trends.  U.S. Bureau of Labor Statistics (BLS) price indices, for such items, net out a BLS estimate of the change in the “quality” of those items over time.  The upshot is that much of the spending in the U.S. Consumer Price Index is now somewhat divorced from the actual prices that you, the consumer, must typically pay.

I am hardly the first one to have noted this.  I think the BLS numbers for home computers, in particular, have drawn a lot of attention.  There, “quality” includes attributes like processor speed, disk size and speed, installed memory, and so on.  As computers have gotten better, but prices have risen in line with overall inflation, the BLS has recorded that as a massive, ongoing decline in the dollar price of home computers.  Per quality-adjusted unit.  And, because computers are part of the CPI, this means a lower overall CPI increase.

My only real point is that the U.S. Consumer Price Index (CPI) is increasingly less relevant as a measure of “the cost of living”.  In any real-world sense.  It doesn’t track what it costs to get by in America.  Not unless you know where I can buy a brand-new 1993 Toyota Corolla, at the 1993 price, instead of what’s currently being offered on new-car lots.

The CPI measures something, and I’m sure that whatever it measures makes good sense to the folks who measure it.  But if you’re merely a typical U.S. consumer, and, say, need to own a car and a phone to get by in life, you can’t take the increase in the CPI as any accurate measure of what is happening to the cost of living.  For you, the typical U.S. consumer.

Sketchy details follow.

---

Did you know that the price of a typical new car in 2019 was the same as it was in 1993?

Lines:  Price index data from the U.S. Bureau of Labor Statistics, via the Federal Reserve of St. Louis FRED system.  Toyota Corolla lowest MSRP from Cars.com, history of the Toyota Corolla.

That, according to the Bureau of Labor Statistics, the official source for U.S. price data, and the keeper of the super-important U.S. Consumer Price Index.

I ginned up the graph above trying to make sense of new and used car prices, Post 1836.

I think we can agree that something happened about half-way up, on the graph above.  Or halfway across, depending on your perspective.  The lines diverge.

What happened, exactly, we might reasonably disagree about.  But the title of the graph gives a broad hint.  Just prior to that, BLS began embedding a quality adjustment in its car price data.

Whatever it was that happened, it’s clear that past that point, what the BLS tracked as the price of a car (orange and gray lines) had almost no relationship to the price of a car, meaning, what you actually have to pay to buy a car (yellow bars).

---

Did you know that the price of the typical cell phone fell 50% in just the past four years?

Source:  https://data.bls.gov/timeseries/CUUR0000SSEE041

From the same folks who produced the car data above.  Again, quality-adjusted data.

Such a bargain now.   Maybe I’ll finally bite the bullet and buy an Apple phone, now that they’re half-price.

Or are they?

Might as well toss this one in, too:  The price for an internet connection is easily 10% less now than it was in 1997.  Again, this is supposed to be an index of the price, in dollars.  I have no clue what the big dip is, mid-graph.  But this is what the BLS says.

Base source:  BLS data query, for finding data series, https://beta.bls.gov/dataQuery/find

---

Did you know that the dollar price for major kitchen appliances in 2024 is essentially unchanged from where it was in 1998?

Source:  https://beta.bls.gov/dataQuery/find?fq=survey:[cu]&s=popularity:D&q=appliances

You get the drift.  BLS data, quality adjusted.  Per the BLS, you have to shell out fewer dollars, today, to buy a fridge, than you did 25 years ago.

Quality-adjusted.

---

Inarticulate conclusion: I am not a Luddite, I think.

I drive an Eee-Vee, for gosh sakes.  I’m not ready to toss my wooden clogs into the industrial machinery of Progress.  Nor am I here to kvetch about the accuracy of BLS’s methods for making these quality adjustments.  (That’s a completely separate issue).

Instead, I want to ask a question.  Can I buy a brand-new 1993 Toyota Corolla?  At the 1993 price?

No?  Then maybe the CPI is no longer a good cost-of-living index.  Or, alternatively, maybe it’s a bit harder for the younger generations to get by than you might think, based solely on the official numbers.

I’m not one to say that the sky is falling because of these quality adjustments embedded in the U.S. CPI.  Practically speaking, I would say that the CPI understates the actual change in “the cost of living”.  Where living is defined as living like the average American.  I.e., has a place to stay, drives a car, uses a cell phone, wears clothes, and so on.

You can see the full list of what’s adjusted in this fashion on the BLS website:   https://www.bls.gov/cpi/quality-adjustment/.   Broadly speaking, BLS embeds some sort of quality adjustment into most of the items in these categories:

• Cars/trucks
• Clothing
• Appliances
• Electronics
• Information services (internet, phone, cable service).

By and large, BLS has no such adjustments for:

• Car/truck repairs, parts, or rentals.
• Food
• Energy
• Health care
• Misc household stuff

And there’s a separate, seemingly quite different adjustment for rent and the rental equivalent of home ownership.

By inference, then, payments linked to the CPI as a “cost of living adjustment” won’t rise fast enough to keep up with the actual cost of living. To some degree.  This includes most notably Social Security payments, but also most Medicare payments to health-care providers, and in general a whole lot of salary and contractual payment items in the private sector.

In my darker moments, I’m sure this is considered more a feature than a bug, by the Federal government.  At some level, it doesn’t much matter if these quality adjustments are right or wrong on their own merits.  They are saving Uncle Sam some money (via, e.g., reduced Social Security outlays), and they are being tolerated.  In a rational world, legislation that would weaken those adjustments would have to be scored as costing the Federal government money.  Hence, a tough sell, and the net effect is that the legislative branch turns a blind eye to it. 

Meanwhile, getting back to cars. I believe that most of the quality-adjusted items in the CPI have, in fact, gotten a lot better in recent years.  I’ve argued that for cars, specifically, many times, on this blog.  It’s self-evident for phones.  I’m pretty sure my current internet service is a lot faster than it used to be.  And so on.

Back to cars, well, in fact, cars are … more now.  Passenger vehicles are bigger, faster, and get better mileage than they did 25 years ago.

Source:  https://www.epa.gov/automotive-trends/highlights-automotive-trends-report

But at the same time, if you need a car and a phone to hold a job, you don’t have the option of buying a new 1993 Toyota Corolla (-equivalent), at an appropriate discount to the current model year.  Your option is to buy the better quality modern car (phone, internet, clothing) currently offered for sale.  At the current higher price.

So if your metric is “I gotta have a car”, then the cost of “a car” has indeed risen a lot faster than the BLS says.  Per the original graph, it (yellow bars) seems to have risen right in line with inflation/the overall price level (blue line).

What you can buy now is a better car than you could buy 25 years ago.  Safer.  Better gas mileage.  More bells and whistles.  (E.g., I don’t know the last time I even saw a car that didn’t have power windows and AC, both of which used to be luxury add-ons.)  But it’s still, at root, just a car.

A weird side-effect of this is that, per the BLS, recent generations of Americans are, in essence, victims of forced hyper-consumption.  When I was a youth, I drove cars that were absolute pieces of crap.  But they were cheap.  A kid today has no option but to buy a much better, far more sophisticated vehicle.  And those simply are not going to be as cheap as beater used cars were when I was in their shoes.

You have no choice but to buy a much better fill-in-the-blank than people were buying two decades ago.  And you have more-or-less no choice but to pay accordingly.  The net result of which is that the — income shares — of these items stayed about the same.  A phone was a relatively expensive object four years ago, and it’s just about as expensive now.

But, when it comes to cars, phones, internet service, and so on, as far as the Federal government is concerned, dollar prices have been flat-to-falling for past couple of decades.  And the reason we now pay many more dollars for “a car” is that we have, collectively, decided to buy bigger, better cars.

And while it is true that cars (phones, internet service, …) are better (bigger, more capable, faster, more efficient) than they used to be, you, as the consumer, or Americans, as consumers, don’t really have the option to keep your level of consumption constant.  Every time you replace an item of that sort, you have to replace it with what’s offered in the marketplace.  And if the market offers you ever-more-capable, ever-more-expensive items, well, if you want to replace your old one, that’s what you have to buy.

The bottom line is that if you accept the BLS quality adjustments as fundamentally correct, then you have to believe that we are, in effect, trapped on a treadmill of ever-increasing levels of consumption.  For some items, at least.

And yet, it’s an odd sort of treadmill.  It’s not as if you’re now required to own two phones, if you want to make phone calls.  It’s that you are required to own one, but your only option is to buy one that’s twice the phone it was, four years ago.  At about the same price you’d have paid four years ago.  Which the BLS then handily marks down as a simultaneous doubling of your phone consumption, coupled with a 50 percent cut in the price of a phone.

That’s just bizarrely at odds with perception, I think.  If the price of chicken goes from $2 to $1 a pound, then I correctly perceive that the cost of my chicken dinner has been cut in half.  By contrast, I sure wouldn’t notice much difference in functionality between a four-year-old phone and a brand-new one.  But per the BLS, that price has also been cut in half.  Even though I’m paying more for a new phone today than I would have paid four years ago.

Conclusion.

I guess that’s as far as I can take this train of thought.

If you’re of the opinion that a phone is a phone, then the BLS numbers don’t “fit” your experience.  You’ve seen no huge fall in cell phone prices these past four years.  By contrast, if you are really into phones, then maybe the BLS got everything right and there has been a halving of quality-adjusted phone prices over the last four years.

At root, my issue with these quality-adjusted items is that the declining price per unit is coupled with an offsetting mandatory increase in units consumed.  Per the way the BLS reckons it. With the result being a bunch of price cuts, on paper, that in no way, shape, or form reduce the amount of money I have to shell out, to own a (car, cell phone, and so on).

It’s kind of a good-news, bad-news joke.  The good news is that the price of X has fallen in half, as BLS measures it.  The bad news is that you can’t buy X.  All you can buy now is 2X.  Again, per the BLS method.

Sure, the BLS-measured price of X is way down.  But you can’t pocket that money.  You can’t use it elsewhere.  This makes a BLS-estimated price decline in (say) cell phones fundamentally different from a decline in (say) the price of chicken.  Unlike the chicken, the only way to take advantage of half-price cell phones is to buy twice as much cell phone.

So the joke’s on you.  Your out-of-pocket is the same, but BLS tells you your life has gotten easier.  Because cell phones now cost half what they used to (per quality-adjusted unit).  With the catch being, you now have to buy twice as much cell phone.

The effect of these ongoing quality improvements is that material goods are getting better.  But as the BLS measures it, that makes it appear as if material goods are getting cheaper.

But they aren’t.  Not if the relevant unit is “a working car” or “a functioning cell phone”.  And as a result, changes in the CPI understate the actual change in the cost of maintaining a typical American lifestyle.

My gut reaction is that the BLS numbers help to paint too rosy a picture of what it takes to get by in the modern world, versus the world of a few decades ago.  The march of progress has made these objects and services better.  But it’s almost a matter of opinion as to whether that has made them cheaper.