Bruce Sacerdote has a very interesting new NBER working paper looking at growth in American consumption, income and wages. His study was motivated by the apparent disconnect between reports of stagnant real income and what seems to be rapid improvements in living standards:
In the first part of this descriptive paper I consider growth in household consumption for households with below median income. The pessimistic narrative on real wages is somewhat at odds with casual empiricism about material goods consumption. If you spend time working with high school students, you notice that even in low income areas, many of the students have cell phones and have access to cable TV and internet service at home. Access to these powerful and modern tools suggests that low income families have seen important gains in at least some areas of consumption. The quality and variety of home appliances and electronics (TVs) in the average home is surely vastly superior to what people owned in the 1970s. American homes have become more spacious and cars are both higher quality and there are more cars per family.
Last week I purchased a one-year-old Nissan Maxima SL (with 6600 miles) for $26,350, and was stunned by how much cars had changed since I bought my last car (a one-year-old Altima, in 2008.) I recall reading that (in the 1990s) Bill Gates used to drive a Lexus LS400. I would much prefer this 2016 Maxima to the car driven by the richest guy in the world, just a couple decades ago. The monthly payments are about $480, which means it’s within reach of tens of millions of Americans. There are an amazing number of safety innovations since I last bought a car.
I recently passed by a store selling TVs, and was floored by the improvement in picture quality since I last bought a set (in 2008). For the first time in my life, good TVs don’t seem inferior to what you get at a movie theatre. I don’t think I even need to talk about calculators, cameras, flashlights, calendars, movie cameras, iPods, GPSs, encyclopedias, weather reports, timers, clocks, TVs, taxi summoners, and a dozen other appliances, all of which are incorporated into your telephone. My scanner recently broke, but I quickly realized that I don’t need one–the iPhone can scan as well.
Some argue that all these quality improvements don’t matter, “you can’t eat quality”. But even there it seems living standards are rising fast:
The Hamilton (1998) technique of calculating CPI bias is ingenious in that it only requires a modest number of data inputs. Hamilton’s insight is that food’s share of the household budget over time should depend only upon income and the price of food relative to all other goods. If incomes are not rising, changes in food’s budget share should be attributable to changes in the relative cost of food. In reality, PSID (Panel Study of Income Dynamics) data show food’s budget share falling over time even holding the relative price of food and CPI adjusted income constant. This is a strong indication that the CPI is over adjusting and that true real incomes are indeed rising. It is easy to calculate the CPI bias as that adjustment to CPI needed to explain the difference between true food budget share and predicted food budget share. Hamilton finds that CPI is overstated by 3 percent per year from 1974-1981 and 1 percent per year from 1981-1991.
Sacerdote estimates that consumption for families with below median incomes has risen by an astounding 164% since 1960, which is about 1.7% per year. That seems far more plausible to me than reports that real incomes are stagnant. Admittedly the fastest growth occurred between 1960-72, but even in the post-1972 period, consumption is up by 81.5% in the “below median income” category.
READER COMMENTS
Andrew_FL
Apr 19 2017 at 1:52pm
Quality improvements do matter, the problem is that the magnitude by which something is “better” is subjective and cannot, therefore, be objectively quantified.
You will disagree with this correct point, because denying the subjectivity of utility is necessary for you to protect your ideology.
Jon Murphy
Apr 19 2017 at 2:07pm
@Andrew_Fl
Well, we can quantify quality to some extent (in fact, that’s one of the things that the BLS does). For example, we can measure median square footage of homes now vs the 1960’s. Or what the standard options on cars now vs 1960’s. Quality of food can be quantified by the instances of food-borne illnesses, etc.
It is certainly true that utility is subjective, but even that we can quantify to some extent (as an aside, just because something is subjective doesn’t mean it can’t be quantified). We can reasonably assume that the price one pays for a good or service is at least equal to (or below) the utility one gets for the good/service vs other possible uses. For example, if I spend $12k on a car, we can reasonably assume that car provides me with at least the same or more utility than the utility provided by other uses of that $12k. Another example: if I spend $40 on Red Sox tickets, we can reasonably assume that ball game provides me with at least $40 worth of utility (in this case, entertainment).
Alex S.
Apr 19 2017 at 2:35pm
Given the paper’s result, how does your inner Bayesian feel about the Great Stagnation Hypothesis?
People aren’t angry with the level of their real wealth, but at it’s current growth rate.
Perhaps today’s social angst is because b****ing and moaning is a luxury good. The marginal cost of complaining was high 200 years ago and before that, because so much time had to be spent just surviving. Now, thanks to such high living standards, we can just hop on the internet, complain, and feed off others’ complaints. The same rationale may apply for why even poor countries have comparable levels of happiness with wealthy nations.
Christian List
Apr 19 2017 at 4:34pm
[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment and your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.–Econlib Ed.]
Andrew_FL
Apr 19 2017 at 4:49pm
“Well, we can quantify quality to some extent (in fact, that’s one of the things that the BLS does).”
The BLS are doing pseudoscience.
And there is no such thing as “dollars of utility”.
Thaomas
Apr 19 2017 at 5:10pm
How do these insights apply to the optimal target for NGDP growth?
Jon Murphy
Apr 19 2017 at 6:35pm
@Andrew_FL:
And there is no such thing as “dollars of utility”.
Sure there is. Price, like I discussed above. Prices give us a good idea of how much utility a person gets from an item. That’s why prices are good rationing tools. That’s why us free-marketers like prices. It is subjective person to person, but it can be roughly quantified.
Andrew_FL
Apr 19 2017 at 8:28pm
“Prices give us a good idea of how much utility a person gets from an item.”
No they don’t.
“That’s why prices are good rationing tools.”
No, it isn’t
“That’s why us free-marketers like prices.”
Fortunately the case for a free price system doesn’t depend only on “free-marketers” whose understanding of value theory is so fundamentally flawed as yours.
Jon Murphy
Apr 19 2017 at 8:41pm
@Andrew_FL
No they don’t.
Sure they do. It’s simple cost-benefit analysis. if a person is willing to pay a certain price for a good/service, then one can reasonably assume that they receive a greater than or equal to benefit than the cost they pay. It’s rough, to be sure, but it does give us some kind of quantification. In their great price theory textbook, Exchange and Production, Armen Alchian and William Allen make this point explicitly (other texts make similar points).
Let me try by asking you a question: what is the typical free market argument for allowing prices to rise following a natural disaster?
Shane L
Apr 20 2017 at 7:33am
I broadly agree with Scott’s point. I have wondered lately, however, if there is something strange about housing in this respect. Here in Ireland we have steeply rising accommodation prices in response to the recovering economy (and, presumably, population growth as net migration turns positive again). Unemployment has fallen a lot and median incomes are starting to rise, but many people are concerned with an apparent increase in homelessness, especially in Dublin city. (A homelessness charity has published data suggesting this apparent increase is real.)
It can be odd to see homeless people with mobile phones and good quality sleeping bags. In some respects they also seem to have benefited from the rise in quality of various goods, yet in lacking a home (and a safe place to accumulate goods) they still seem seriously poor.
My argument with left-leaning friends that things have improved in various ways does not go down well with them since they point to this rising homelessness. What matter that sleeping bags are lighter and more insulating now, if one has to use them on urine-stinking concrete in the freezing city streets? I’m not sure what is happening with homelessness, if it is going to decline in the long run or will persist. Many goods are cheaper and higher-quality now than in the past. I gather that modern homes have typically better insulation, for example, but I’m not sure that they are relatively cheaper than in the past. All a bit puzzled by this.
Scott Sumner
Apr 20 2017 at 7:35am
Andrew, I’ve never argued that utility is objectively measured.
Perhaps you are confusing me with someone else.
Alex, In my view, the Great Stagnation hypothesis applies to measured gdp. So no effect.
Thaomas, No effect on optimal NGDP.
Travis Allison
Apr 20 2017 at 10:13am
Scott, I don’t understand your response to Alex. You said that the Great Stagnation hypothesis applies to measured GDP. But if inflation is understated by 1% per year, then measured real GDP is much greater than previously believed.
Ricardo
Apr 20 2017 at 10:31am
The link appears to be broken.
Andrew — everyone accepts your point about the nonexistence of a utilometer. But a simple regression allows us to characterize quality improvements in an objective way. For example, taking cotton as the baseline for a shirt, substituting into silk might imply a price factor of 1.2. These factors come from observed market prices.
Andrew_FL
Apr 20 2017 at 11:47am
@Jon Murphy-
“Let me try by asking you a question: what is the typical free market argument for allowing prices to rise following a natural disaster?”
The correct argument is that in order for resources to be allocated efficiently prices must reflect relative scarcities. If you’ve been using a different one based on higher prices maximizing utility you are wrong, and should maybe consider becoming a socialist.
“In their great price theory textbook, Exchange and Production, Armen Alchian and William Allen make this point explicitly (other texts make similar points).”
If Alchian and Allen did indeed make your incorrect point, then I’ve lost a lot of respect for Armen Alchian. But I suspect you have misunderstood them.
@Ricardo-
“everyone accepts your point about the nonexistence of a utilometer.”
No, they don’t, because there are utilitarians present who apparently don’t understand that the existence of such is at the core of their ideology.
“But a simple regression allows us to characterize quality improvements in an objective way. For example, taking cotton as the baseline for a shirt, substituting into silk might imply a price factor of 1.2. These factors come from observed market prices.”
I deny that this works and say that to say that what one is measuring here is quality is pseudoscience.
@Scott Sumner-
“Andrew, I’ve never argued that utility is objectively measured.
Perhaps you are confusing me with someone else.”
You have argued for utilitarianism, of which objective utility is a logic antecedent.
Jon Murphy
Apr 20 2017 at 12:01pm
@Andrew_FL
The correct argument is that in order for resources to be allocated efficiently prices must reflect relative scarcities.
That is correct. And what do we mean by “allocated efficiently”?
Jon Murphy
Apr 20 2017 at 12:14pm
@Andrew_FL
In reading your comment, I think you may misunderstand what I am saying.
All I am saying is the pretty standard economics argument, that individuals seek to maximize their utility. Judging by your “socialist” comment, I took it to mean you think I’d advocating maximizing some kid of social utility. That is, most emphatically, not the case. There is no such thing as “social utility.”
So, given that individuals seek to maximize their individual utility, we can, though the observation of their human actions (to use Mises’ term), get a rough idea of how the value something (that is, that kind of utility someone gets from something).
The tl;dr version: so long as we assume people are generally rational, we can reasonably conclude that their actions seek to improve (ie generate utility) their lives, and thus their behaviors at various price points can give us rough ideas on how much value (or utility) they gain from such a transaction.
Andrew_FL
Apr 20 2017 at 3:40pm
“and thus their behaviors at various price points can give us rough ideas on how much value (or utility) they gain from such a transaction.”
This does not follow.
Jon Murphy
Apr 20 2017 at 4:19pm
@Andrew_FL:
Why?
Fred_PA_2000
Apr 20 2017 at 11:07pm
I thought there were elements to this argument that don’t seem to be getting any commentary here. (This may merely confirm my amateur status in these matters. But if so, I would appreciate others setting me straight.)
Isn’t there an argument that stagnation in disposable income is being seriously affected by — even caused by — an ever increasing tax bite?
And for the sake of fair argument, shouldn’t we acknowledge the effects of government tax & transfer programs? That incomes at the top aren’t growing as fast as they might appear if they’ve been subjected to more & more aggressive haircuts. And life at the bottom may be better than earned income might imply once we add in ever more generous transfer payments.
Finally, that while take home pay may appear stagnant, benefits (esp. healthcare) have expanded substantially over recent decades.
It seems to me that there’s more here than just a debate over the true rate of inflation.
Brandon Berg
Apr 21 2017 at 4:59am
Fred: Is there, in fact, an ever-increasing tax bite? Other than the expiration of the temporary payroll tax holiday a couple years ago, the last middle-class payroll tax increase was in 1986, and middle-class income tax rates have decreased as well.
As you can see here, effective federal tax rates for the middle and lower classes are the lowest they’ve been for decades.
Maybe middle-class taxes are increasing at the state level, but they’d have to increase quite a bit to cancel out the federal tax cuts.
Rick Hull
Apr 21 2017 at 5:40am
Jon,
You seem to be missing the supply side of prices. Let’s reason from a price change. If the price of a silk shirt goes up relative to cotton, has the shirt’s relative quality increased, or was there a silk shortage, or?
Jon Murphy
Apr 21 2017 at 8:51am
@Rick-
With respect, I do not see how that has anything to do with the discussion Andrew_FL and I are having.
I mean, I am not missing the supply side; I’m not talking about the supply side. It is irrelevant. I’m talking about demand, consumer demand. If the price of silk falls, then those who value silk less (that is, get less utility from a silk shirt compared to other items) may now be temped into buying it, since the cost is closer to that consumer’s utility received from his point of view.
Jon Murphy
Apr 21 2017 at 8:52am
Again, to be clear: I am not saying that the price perfectly quantifies utility. What I am saying is that price can give us a rough idea of who values what.
Rick Hull
Apr 21 2017 at 2:38pm
Jon,
When prices move and people’s buying behavior changes as a result, this does give us some insight via revealed preference, but only in terms of ordinal preference or preference ranking.
But you have a LONG way to go in order to reach cardinal utility (measurement of utility, rather than ranking of preference) from price observations.
Jon Murphy
Apr 21 2017 at 5:23pm
Thus why I said “rough.” It gives us an idea, but it is incorrect to say, as Andrew_FL said, that we cannot know it.
Rick Hull
Apr 21 2017 at 7:55pm
How do you plan to graduate from preference ranking to measures of utility? I think you’re 100% stuck at preference ranking. That’s the rough spot, and it’s all you’ve got to work with.
As such, Andrew_FL’s point remains essentially unchallenged:
Floccina
Apr 24 2017 at 2:49pm
I think the measurement problems are due to the fact that, as difficult as it is to value new inventions, it is even harder to value improvements in quality. I have bought vehicles with over 150K miles on them and they have gone for another 100K without much repair. In the 1970’s a car with 100k miles on it would coast a lot in repairs. A CPI adjuster could easily miss the latter but he know he must adjust for the new inventions.
Also hard to adjust for is cleaner air and water ways.
BTW IMHO a great thing that could be done for lower income people is to Wipe the Mosquito varieties that bite humans.
Floccina
Apr 24 2017 at 2:55pm
BTW if inequality means that many of the top 20% earners buy a new car every few years, just for the novelty, and that fills the market with great used cars that somewhat mitigates the growth in inequality.
Comments are closed.