A Primer on Standard of Living and Cost of Living

by Russ Roberts on September 8, 2006

in Standard of Living

In the comments to my post on CPI bias, Don Lloyd accused me of needlessly confusing standard of living and cost of living. Another commenter, Cyberike, had a different complaint:

How much does it cost for the necessary expenses of life and how do the costs compare over 30 years?

Certainly I live in a better house than I did in 1975, but my costs
have gone up by a factor of 5 (500%). My electric bill is outrageous.

But I want to make a new argument:  you have to include those things that are now necessary that did not exist 30 years ago.

I had no cable bill, no internet bill, no cell phone’s. Those items
right there add about $200 to my monthly expenses. I see no one taking
these types of items into account, yet they are certainly factors in
the cost of living. Have they added to my quality of life? Yes. Do we
consider them necessities? Also yes.

I think Cyberike’s concern is a common one. Sure, I have more stuff, but I still feel like I’m struggling. So what’s the meaning of a higher standard of living?

I want to use these comments as a jumping off point for some general observations on the standard of living and the cost of living.

Standard of living is inevitably linked to the cost of living. When we try and see whether we’re better off than we were 30 or 100 years ago, we immediately have a problem. Our incomes are higher which suggest a higher standard of living. But we know that the general level of prices has risen over time, what is called inflation.

So you can’t look at the growth of income alone to measure the growth in material well-being. You have to take account of the increase in prices to accurately measure how much more stuff we can have now compared to before.

(And an obvious challenge is that there is no "we." Some of the things that have gotten a lot more expensive than others
may not be items you consume. And there may be some items that have
actually fallen in price, despite the general rise in prices, that you
consume a great deal of. So each person’s well-being is very difficult
to calculate. The presumption is that if you measure the prices of a broad enough range of goods and services, it will capture the average rise even if it will not be accurate for each individual. And since we’re usually looking at some measure of average income, we’re going to be in the ballpark.)

Conceptually, measuring the change in prices over time is pretty straightforward and has been studied for a long time. You go out and see what people buy. You look at the prices of those things. Then you do it again down the road. The average change in prices, weighted by the shares of the goods and services in total expenditure, is the measure of inflation. You divide the new higher income level by the new higher price level. If incomes have gone up by a bigger percentage than the average price, the standard of living is higher.

The average level of prices is the "cost of living" as economists use the phrase. You have to bring in the cost of living if you want to talk about the standard of living, because of inflation.

In practice, measuring the change in prices is quite difficult. Some of the difficulty is because the mix of goods and services—the "basket" of goods and services as it is sometimes called—isn’t really the same as it was a year ago or ten years ago. One reason is that people will tend to substitute toward goods that have become relatively cheaper and move away from goods that have gotten relatively expensive. So if you’re weighting the prices by expenditure, do you use the old weights of a year ago or the new weights? Economists have been aware of this problem for as long as they have talked about price indices and there are various ways of correcting for this problem.

But the harder problem is that because our standard of living is improving, because we’re better off and more real income, we don’t just  want more than we had before or a wider range of stuff than we had before. We want better versions of what we had before. We want safer cars and higher-quality personal stereos and bigger televisions and houses with more bathrooms and more living space, and less-invasive surgery and shirts that don’t wrinkle. Virtually everything is constantly getting improved, from the deeply important to the trivial, from prostate surgery to dental floss (flavored! smoother! on a stick!).

So just looking at the change in prices, even for the "same" good or service overstates the change in prices because it’s not the same good. Suppose you buy an iPod for $300 that holds 10,000 songs replacing a portable CD player that carried 16 songs for $100. The price of a portable stereo certainly didn’t triple. It probably went down.

Conceptually, this problem is also easy to solve. You use more refined categories of the good than "portable stereo." But then they improve each of the sub-categories. The iPod gets better. It holds more songs than it did before. It gets a color screen. It plays videos. This, too is conceptually fixible via what economists and econometricians call hedonics. But in practice, it’s not just challenging to fix. It’s impossible. The Bureau of Labor Statistics simply doesn’t have enough people to spend the millions (trillions?) of hours it would take to accurately correct for all quality changes. So this inevitably biases the CPI upward, overstating inflation and understating cost of living growth.

The problem gets worse when totally new products come along such as the iPod. How do you correct for the change in quality between a washboard and a washing machine or a horse and a car?

Over small enough periods of time, these distortions are relatively small. Over longer periods of time—a decade or three, a century—it becomes meaningless. How do you compare a dentist’s fee for coping with a tooth infection in 1900 that can kill you to the fee for a modern root canal? How do you compare an iPod to a strolling band of musicians, the only portable stereo available in 1900?

The tools of economics and statistics are strained beyond credibility. So any estimate of our change in our standard of living over 30 or 50 or 100 or 1000 years are crude at best. But there are good guesses and bad guesses. Surely, the average American is many times more comfortable materially than in 1900 and even perhaps compared to 1970.

Ironically, the faster our standard of living grows, the faster the quality and variety of choices that are available to us and the more our measure of the cost of living is overstated and our actual increase in material well-being is understated.

(And yes, I know some people think the CPI is understated. Let’s see them quantify that bias and show it’s anything close to the magnitude of failing to correctly deal with the improvement in quality.)

Which brings us back to Cyberike’s complaint about life today compared to the 1970′s:

I had no cable bill, no internet bill, no cell phone’s. Those items
right there add about $200 to my monthly expenses. I see no one taking
these types of items into account, yet they are certainly factors in
the cost of living. Have they added to my quality of life? Yes. Do we
consider them necessities? Also yes.

These are not the cost of living, as economists define it. But they are the cost of living in the everyday sense of the phrase. "I can’t live without my cell phone." Or e-mail. Or my laptop or my GPS system. "I have to have my digital camera for my trip." Can you imagine having to wait for your film to be developed? Who would put up with that? Well, we could obviously, but few of us choose to. Or can you imagine putting three kids in one bedroom? Impossible. Actually, it’s not. Some people do it today. Americans did it all the time 100 years ago. But the average American family today doesn’t do it. Not because we have to give each kid his or her own bedroom. Because we’re wealthy enough to choose to do so.

That’s called a higher standard of living. You’d feel it more if you went back to 1970 and with your grainy 20" TV and four channels to choose from, your smaller house, your more invasive surgery and your wooden tennis racket and your shorter life span. Some of those things are trivial and don’t really make us any happier. Some of them just become part of our expectations and the thrill of that iPod or that cell phone with a camera and voice-dialing has faded long ago like most material thrills. But most of us, as Don pointed out, would find the choice between 1970 and today an easy one to make. Because there are parts of the comparison that aren’t trivial, such as living longer. And because we’re human. We prefer more to less even though we get used to the "more" and forget how much we longed for it when we had "less."

We can talk all day about the significance of a higher standard of living and whether it makes us happier. But what we’ve been talking about in the recent posts here at Cafe Hayek is that if you want to measure our increased command over goods and services over the last five or 30 or 100 years, it’s harder to measure than it looks and there are good reasons to believe that our overestimation of the cost of living means that our measures of the improvement in our standard of living are understated.

And that brings me to one last point about these issues. No matter how much we have. No matter how great our purchasing power, we will always find it challenging to get buy on what we have. New stuff comes along that seems like a necessity. Of course it’s not a necessity. A house bigger than 2000 square feet is not a necessity. A car with a CD player and drink holders isn’t a necessity. But we will always want to use our expanding standard of living to buy more of what we have and new stuff we hadn’t dreamed of. When the peasant gets his first thatched hut and no longer has to live under a threatening sky, he doesn’t say, that’s it, I’m satisfied, nothing else is necessary to make me happier.

Thomas Sowell said it best. You can read about it here.

We are never satisfied. Something inside us wants more and better. And more and better beyond that. A spiritual fully-realized human being is aware of that drive and recognizes it and searches for something more meaningful and deeply satisfying than an iPod that plays video. But it is that drive for more and better that produces a longer life expectancy along with better dental floss. They both spring from the same source. And that source, that drive, those desires were in that medieval peasant in his thatched hut, in us today in our McMansions and will be in our great-great-great grandchildren tomorrow in whatever way they choose to live their lives.

Be Sociable, Share!



39 comments    Share Share    Print    Email


Ghost September 8, 2006 at 11:57 am

Let me restate Russ's point more forcefully: Economists reject the idea of a moving, subjective standard of "necessity" because it IS a moving target. The goal is to think about and measure cost of living in an absolute sense without regard to changing standards of necessity.

That doesn't mean that some people don't feel poor if they can't buy a new car every 6 years or have a cellphone. But it becomes a question of moral and social philosophy as to what people "should" want or have. Russ wants to separate out the objective question: Are people much richer — in purely material terms — than they were 30 years ago? [YES!!] from the subjective question as to whether they feel richer than their forefathers felt.

You can then say: Ok, people are richer today, but expectations of minimal decency have so changed since 1970 that people feel poor relative to their new needs without needlessly complicating the basic calculation of cost of living.

Bret September 8, 2006 at 12:36 pm

I really liked this post. To me, it has an excellent balance of the measurement issues ("But in practice, it's not just challenging to fix. It's impossible"), leading to difficulty in comparing statistics over long time frames ("Over longer periods of time—a decade or three, a century—it becomes meaningless"), and the implications of those when talking about cost/standard of living.

dj superflat September 8, 2006 at 1:23 pm

awesome post, i like the turn to the spiritual at the end. and i just don't see how anyone could argue with the analysis, but i'm sure someone will.

Don Lloyd September 8, 2006 at 1:43 pm


"…This, too is conceptually fixible via what economists and econometricians call hedonics. But in practice, it's not just challenging to fix. It's impossible. The Bureau of Labor Statistics simply doesn't have enough people to spend the millions (trillions?) of hours it would take to accurately correct for all quality changes. So this inevitably biases the CPI upward, overstating inflation and understating cost of living growth…."

It is not for a lack of time and resources that hedonic quality adjustments cannot accurately be used to offset price increases due to monetary supply inflation, but for the fact that it is fundamentally impossible.

Hedonic quality CAN adjust if the number of strawberries in a carton increases, but NOT if the taste of the strawberries improves.

Quality changes can only affect relative prices, not overall price levels if the supply of goods, the supply of money, and the demand to hold money are all held constant.

Assume that a closed economy has only 10 goods, and that the ONLY change over the next year is an improvement in quality for each of the 10 goods. For simplicity assume that all 10 goods are packaged in such a way so that they all have the same price.

Assume that the improvements in quality are determined in the following way:

Start by dialing up the quality of all 10 goods by 10%, or by what seems like 10%, it doesn't matter.

Go back and adjust the individual good quality levels recursively until all 10 goods have the same market price.

With no change in the number of goods, or the supply and demand for money, the 10 new prices will be identical to the prices that existed before the quality of the goods was increased.

We have a substantial increase in the standard of living due to the increased quality of goods, but the prices and the cost of living are unchanged.

This is an inherent flaw in the use of hedonic quality adjustments, and, in fact, if hedonic adjustments didn't result in an understatement of the price effects of monetary supply inflation, you can bet that the government wouldn't make use of them as it attempts to reduce future COLA adjusted outlays.

That said, this in no way invalidates the arguments against the significance of the supposed lag in real wage rates. It is entirely possible for the aggregate statistical real wage rates to lag or even fall even while every individual, both actual and hypothetical time-displaced individuals, to be left better off, possibly substantially.

Real world statistics reduces a complex system of an infinite number of variables to a few numbers with a substantial loss of information. It is not at all surprising that the superficial statistical results may not align with the experience of the people and things that are being aggregated.

Regards, Don

kebko September 8, 2006 at 2:38 pm

I have a question.

Let's say that in 1996, I & 4 other who jointly manage a trust or some real estate holdings,etc. get together for our annual meeting. We buy 5 airplane tickets for $400 each & spend another $200 each on hotels, etc. to meet for a day. Total cost of the meeting is $3,000.

In 2006, we have the same meeting via teleconference. The 5 airplane tickets & travel expenses would now cost $3,600. But instead, the total cost of the teleconference is $300.

It seems to me that the final statistical result of this change will be that economic activity decreased by $2,700, and inflation over the 10 years would appear to be cumulatively 20%. Where this kind of substitution leads to more productivity with producers, it would lead to productivity increases & lower inflation. But, where it wasn't attached to cost of goods, wouldn't the end statistical result show the opposite of reality, and no productivity increase? Would the 90% increase in productivity instead show up as a $2,700 slowdown in the economy?

Another example would be individual investors who can instantly access all of a company's SEC filings, compared to years ago when that task would have been time consuming, slow & expensive. There is probably less spent now on dispensing SEC filings to investors, but the amount of information that is accessed dwarfs the earlier amounts. Since these transactions are largely not cash based now, it seems like a change like this would not be statistically tracked at all. In dollars, less activity is happening now, but in terms of quality & amount of service, what investors have now is immeasurably more.

Can the changes from all of these internet & telecommunications technologies be incorporated into the statistics?

Cyberike September 8, 2006 at 2:51 pm

Excellent, excellent post which does exactly what I had wished: give some kind of analysis which includes those issues. I wanted to stimulate some debate while expressing a feeling that, somehow, we are falling behind in a way that the numbers do not capture. Meanwhile, I don't think I can argue with the many well reasoned counterpoints and rebuttals.

I think my point has been well received: what are considered necessities change over time, and the additional cost has not been factored into metrics like the CPI. Remember that at one time, things like electricity were not considered neccesities.

One question: is insurance factored into CPI? My homeowners insurance has nearly doubled over the last 6 years. It it also something we can say has doubled in quality over the same time?

ghost September 8, 2006 at 3:05 pm


I feel strongly that these perceived necessities should NEVER be factored in. If they go up so much in price yet you feel they are essential and therefore feel worse off, you should consult a priest, rabbi, or psychiatrist. No policy should ever be based on these perceptions. If you are materially better off, you ARE economically better off. Beyond that we start asking about the meaning of life. NO economists should be involved in that debate.

But then, that's just one man's opinion…

spencer September 8, 2006 at 4:27 pm

Before I accept your argument that the CPI understates inflation and so we are understating the improvement in standards of livings I would like to see a reply to some different inputs.

In the CPI durables — essentially all the things you have been talking about– have a weight of about 12% and new cars are 5.1 percentage points of that. since 1980 the average increase in durables in the cpi has been 1.7% Ok, assume they have overstating inflation and the actual price increase for durables was actually zero. This would reduced reported inflation rate by about 0.2 percentage points annually.

On the other home owners equivalent rent accounts for about 25% of the cpi. Over the past decade it has grown at about a 3% rate while the best measure of home prices — the OFEHEO measure of repeat sales — has grown at a 7% annual rate. So housing has been understated by four percentage points a year and this would increase the overall cpi by one percentage point a year — or about four times the impact of the durables you keep talking about.

Moreover, in the CPI medical care — some 6% of the CPI — only measures out of pocket spending on medical care. It does not include the bulk of hospital care, etc that medicare and insurance pays for. So we know the cpi significantly understates the impact of medical care on the costs of living. On the other hand the CPI is not for medical care is not quality adjusted, so maybe these two factors roughly offset each other.
Do you have evidence to the contary?

All the communication and computers you discuss has a weight in the CPI of 3%,
about the same weight as education where no one denies that prices have been soaring.
Why does falling prices in one count, but rising prices in the other not count?

Ok, the CPI does not pick up new technologies when the early adapters start using them, but by the time they become common they are generally included in the CPI and there are strong arguments that the hedonic pricing overstates the fall in prices of these items.

Services excluding shelter has about the same weight in the cpi as durables. In this a see a mixed bag. On the one hand things like ATMs and other electronics mens we can access banking 24 hours a day and this is a big improvement in living standard not reflected in the cpi. But on the other service in many areas of the economy — like retail — has clearly deteriorated.
I have no idea if these conflicting trends roughly offset each other, do you?

But what I see is you selectively citing
evidence to support your conclusion and ignoring evidence to the contrary. You may be right, but I see plenty of evidence to the contrary.

It is like your post on housing the other day. You were citing data on the size of homes to demonstrate how our standard of living had increased But you were using the size of new homes and ignoring the stock of homes. Housing is a long lived asset and even when we are building 2 million new homes the stock of total housing only inceases about a quarter to a half percentage point annually. In 2005 the median house in the US was build in 1973
and the median square footage of owner occupied homes grew from 1,788 square feet in 1975 to 1,858 square feet in 2005 — a compound growth rate of about a quarter of a percentage point a year — a massively different rate of growth then you were showing.

spencer September 8, 2006 at 5:49 pm


erdos September 8, 2006 at 7:04 pm

*** searches for something more meaningful and deeply satisfying than an iPod that plays video. ***

How bout an iPod that plays video AND has 20 hours of battery life?? You can't really get more meaungful than that.

Ilkka Kokkarinen September 9, 2006 at 1:01 pm

You might be interested in the post "Are You Better Off Today than you would have been 25 or 50 years ago?" at GMR Musings, in which GMR crunches the numbers and answers the question "If you wanted to offer your family today the average standard of living 50 years ago, how big a salary would you need for this?"


JJo September 9, 2006 at 10:23 pm

I think the standard of living is higher because the upper echelon has risen as the result of mass media and corporate power. Think about having a million people purchase your product. What do you do with the profits? There's always 'more' and 'better' for those people. And there's always a higher step on the ladder for the rest of us.

Aaron Krowne September 9, 2006 at 11:22 pm

I don't see how this post was a valid primer at all. You jump right in again and conflate standard of living and cost of living, assuming ab initio that they must be the same metric. Why? What is the pressing need for looking at the data this way?

A pure cost-of-living CPI (which we do not have now) would be quite useful for portraying the median financial situation. But hedonics and weighting tricks (like geometric weighting) have been thrown in because of an a priori, bureaucratic decision that the CPI was "too high".

Despite all of this alleged confusion and ambiguity, the economy is sending us some real clear signals about the well-being of the average person. For example, the negative savings rate stubbornly won't go away. I would love to know why so much debt is being taken out if the typical person is actually /financially/ better off. I'm not talking about some vague notion of "material" welfare or "joy", but /financial well-being/.

I can distinguish between the two–but anyone who believes hedonics belong in a CPI metric cannot.

Don LLoyd:

You're talking to a brick wall, I'm afraid. Your excellent conceptual proof that quality is in fact extra-financial doesn't matter. Russ and those who take his position in this debate expressly want to lower the CPI, and as they've directly declared, they're willing to go as far as mixing totally different kinds of effects as long as it produces the desired downward adjustment.

Frankly, the true CPI would be too embarassing for the political and economic establishment, so the message is "stop complaining about your well-being, shut up and keep consuming. Here's more credit cards, and don't you have some home equity you could be withdrawing?"

JohnDewey September 10, 2006 at 7:59 pm

Aaron Krowne,

I do not understand what you mean by "pure cost of living CPI". What would it include and what would it exclude? Are you suggesting that it should be based on a basket of goods and services comprised of exactly the same items in 1980 that are included in 2006?

JohnDewey September 10, 2006 at 8:17 pm

Aaron Krowne: "I would love to know why so much debt is being taken out if the typical person is actually /financially/ better off."

I don't know that we can get into the heads of all the debtors, but here's a suggestion. We've taken on more debt because interest rates are so low. Tax deductibility of mortgage interest has made borrowing even cheaper.

Consider an upper middle income family that's socking away $25,000 in retirement savings. Why didn't they instead apply a few years retirement savings to increase the down payment on their McMansion? It's likely their investment returns far exceed the after tax cost of that mortgage. To me, that's a legitimate reason to take on debt. I know a number of my Boomer friends have made exactly that decision in the past decade.

Michael Sullivan September 11, 2006 at 1:31 pm


In fact it's a very hard question, which I think both Don L. and Russ recognize.

The biggest flaw with doing a complete set of hedonic adjustments is that it doesn't reflect the real "cost of living".

Goods at the same quality as that found in 1970 do not necessarily exist in 2006, though if they did, they would be cheap.

Sometimes the market overcorrects. If I had a much smaller income, I might well like to buy a house that is only 600 square feet, as were not so uncommon in the first half of the 20th century. They are pretty uncommon now. If I can find one today outside of a very rural or depressed city area, the land alone will be valued based on it's ability to support a 2000+ sf house (after tearing down the shack), probably pricing it out of my range. If I really want to live in that house, I can't today live in a reasonable suburb or thriving city neighborhood. My only choices are to live in an apartment, a slum or the wilderness, all of which involve big tradeoffs from my original goal (which I could have achieved in 1915) of living in a 600 sf house in a suburb/village.

Same thing with quality improvements in consumer goods. There's a baseline below which prices don't go for computers. A 1995 computer today, while you can buy one on the scrap heap for much less than it cost in 1995, doesn't actually provide the same experience as it did in 1995, since it won't work with software that you can buy now, or on websites that exist now. What it costs to get "a computer that will do most of the stuff that people expect computers to do" is about the same today as it was in 1995. What people expect is different of course, but you don't really have the option of getting the same quality as in 1995, you get to have either a much better experience or a much worse one.

But quantifying this difference is not feasible to do exactly (or even all that accurately) because of the difficulty of interpersonal utility comparisons (I refuse to say "impossibility" — raspberries offered to Don L. and any other catallarchists on his side of that debate).

I actually think the current CPI is a reasonable estimate of the "cost of access to normal society". But we do need to be aware that it is measuring that, and not some arbitrary universal standard of good.

If real wages for the median worker fall according to CPI, a median worker will feel strapped and like their access is eroding, and this will be based on a reality. But that's not the same as saying that his standard of living is objectively worse. It's possible that it would be for a given person with a very odd utility map, but those people will be rare (if they weren't, the market would likely have provided them better tradeoffs). Most people wouldn't take Don B.'s trade.

The problem is that different measurements measure different things, and our sound bit politics tend to conflate all of it to produce a lot of meaningless comparisons.

If you want to study poverty, and who is truly being deprived, you need a different kind of metric than CPI. If you want to measure whether the bulk of citizens feel like they have access to middle class american society and to what extent, I think median wages relative to CPI is a pretty good measure.

It's only when we claim that that measure represents a declining absolute standard of living that we go wrong.

A CPI with a lot of hedonic adjustments would be very useful, OTOH, for determining how many people are truly left behind and in poverty — the attempt being to establish an absolute standard of living. But turning around and using that figure to measure how the median household is doing in terms of access wouldn't be reasonable either.


JohnDewey September 11, 2006 at 2:18 pm

Michael Sullivan: " actually think the current CPI is a reasonable estimate of the "cost of access to normal society"."

Is CPI-U – or CPI-W – a reasonable estimate for determining the many billions in income to transfer across generations? Your comment seems to be seeking a middle ground in the inflation index debate. For those on either side of the huge inter-generational transfers, can there be a middle ground?

Michael Sullivan September 15, 2006 at 4:40 pm


It sounds like you're asking "What should we be using to index social security or other pensions?" I think CPI is reasonable for that, since it basically measures access to what american society considers a standard consumption level. OTOH, I'm not sure the social security system is entirely reasonable. It's not at all clear to me that intergenerational transfer of wealth in that way is particularly efficient. I would like to see social security either disappear (gradually, keeping promises already made or at least 70-80%), or become a forced pension plan with private accounts, combined with a just-above-poverty means-tested safety net for those who mess up or who don't work enough to put much away. I don't think eithee of these is a realistic hope.

I do think that a lot of hand wringing about median wages failing to outstrip CPI is silly, because the standard consumption level will naturally rise if the bulk of the people are making more money.

For me, the interesting factor about median wages falling behind CPI is that it's so radically different than the *mean*, and so radically different from what we've typically experienced in the past (during non recessionary times). Something unusual is going on, and it would be helpful to figure out exactly what.

Prema Rachel February 19, 2008 at 1:21 pm

Wonderful posts I came across whilst looking for comparisons in cost of living between now and ten years ago. (obviously, up, up, up.. oil products) I do wonder what such well educated and informed subscribers would offer if they were asked to factor in the declining overall health of our environment. Is there an apex to our standard/cost of living to where the sharp curve upward, (in factoring in those new commodities designed to make life easier and more efficient,) starts to plummet, as factors such as increasing wealth for fewer people and declining middle classes reduce the majorities of populations into two class systems, rich and poor. My question is how can one define ones standard/cost of living, even though, there is much more available tools for well being (advances in health, science and technology) if fewer and fewer people have access to them? It is all well and good to say today I can expect to survive a heart attack, unless of course I'm poor, or today I can drive to my friends house, the car and gasoline will only consume a fraction of my income, unless I'm poor and those things now consume over 23% of my income.

Would live to hear your comments. Thank You/

Prema Rachel

{ 1 trackback }

Previous post:

Next post: