Oh, you're assuming that CPI is computed in a rational scientific way. And that "housing" means housing. Neither is true.
Let's take a deep dive into how CPI housing is computed. You would think it could just be "track the price of every home sale in the US", or "track the price of every new rental". Ah, my naive friend, the US Bureau of Labor Statistics lives in the year 1900 as if data, electricity, compute, and statistics don't exist. Anyway, those numbers would be too scary and too objective.
First, the real kicker, CPI housing doesn't measure the cost of a house! At all. Their rationale is that the a house is an investment. So it has no relationship to the Case-Shiller index at all.
What do you measure if you don't measure the cost of housing? Well, you're left measuring rents, but that's not satisfying when a lot of people buy a home. So.. why don't we measure the imaginary rent that someone thinks that their house that they bought 30 years ago might be able to fetch.
3/4ths of the housing index is computed from a single quantity called the "Owners’ equivalent rent of residences". They ask people in a survey: “If someone were to rent your home today, how much do you think it would rent for monthly, unfurnished and without utilities?”
Let that sink in for a moment. People who aren't renting their home. Who aren't real estate agents. Who have no idea what the cost to rent their home is. Who are homeowners and haven't rented a home in decades probably. "Bob, what do you think this would rent for?" "Ugh, $1400/month" Great! Let's base our entire understanding of the economy on the random answers of clueless people.
But, remember, the Bureau of Labor Statistics needs to do everything like we're in the pre-electricity era cruising around on the latest model of horse. So they can't ask too many people. Only a few thousand people, and they keep asking the same people over and over again. Seriously, they ask the same few thousand people every 6 months what their rent is and what their imaginary rent is. They even admit this doesn't change much. And they admit this means all the data they get is garbage, it's too skewed by the particular properties of those few thousand people they randomly selected once.
They could just ask 100,000 people or at least pick a different group of a few thousand every time? And then everything averages out eventually.
Doing that with your horse and buggy is way too hard. Instead they adjust the junk responses they get for the type of housing, quality, etc. With a whole bunch of magic numbers that come from yet another survey of clueless people, the Department of Energy’s Residential Energy Consumption Survey, which was last run in 2020 (although I think housing CPI is computed with the 2015 run) and asked about 5000 households a bunch of questions + got some billing info from their utilities.
They compute monthly CPI with an amazing formula that I think is the only use of a 6th degree square root I've ever seen in statistics (and I have CS/math degrees and do ML research; it's probably legit, but I'm not working backwards to figure out how they derived this).
Oh, we're done yet! Housing also gets older all the time, we need to correct for this. That's a whole pdf. That I've never read. And I'm not sacrificing my sanity for HN. https://www.bls.gov/cpi/quality-adjustment/updating-housing-... But it has a model from January 1997 with 110 degrees of freedom. That's not confidence inspiring to say the least.
Ok, so we ask clueless people to give us meaningless numbers that we then adjust based on complex largely meaningless formulas to compute an imaginary rent that doesn't exist instead of just looking at the actual cost of housing. How could this get worse? With the magic of "imputation"! A beautiful word from statistics that means "Oh, we didn't collect enough data and we have no idea what the actual number should be, so we're just going to make a number up and pretend like it's a data point we gathered based on some idea of what we think it ought to have been". Because the best data is the data you make up according to your gut.
No one with two neurons to rub together would ever adopt this methodology in the modern world.
Of course none of these numbers match any of the reality that anyone experiences. Their documentation on how anything is computed reads like standup comedy.
It's surprising how much you discount something you don't take the time to understand. Clearly your single set of experiences is superior to the century of work by thousands of professionals, probably all with far more knowledge in this area.
You should start your One True CPI and license it; you'd be rich.
I took a few seconds to understand why the 6th riot was used; it's clearly correct and if you took a moment to understand instead of discounting it, youd learn. Hint: it's 6 things multiplied together - ever hear of geometric means? Similarly I looked at the paper you discounted because the degrees of freedom somehow offended you. It too is a good model. And, as credentials go, I have a PhD and decades of modeling experience. Even so, I'd expect someone with your claimed expertise to be able to understand these.
If there were some vastly better method, traders could make money on the difference. They look at it, and cannot. There's lots of papers and research on this.
Multiple other actors,like the MIT Billion Prices project started, and others followed, measure it, and get equivalent numbers.
So, if you're so sure the BLS is stupid and wrong, there should be some group getting it right. Care to find them?
Another simple experiment you can do (and I've done) is to take a reasonable set of monthly purchases, weight costs like that, go back 40 - 50 years, and pull up ads for those goods. Pop it all in excel, and compute compound inflation over that time. Do it honestly. Then check BLS over that time. TADA! Guess what? It's pretty good. I've even gotten several friends to do this when they claim all this nonsense. They've all been surprised how well it works.
> Clearly your single set of experiences is superior to the century of work by thousands of professionals, probably all with far more knowledge in this area.
You're behaving like I am alone in criticizing the BLS for this insane methodology. There are countless papers out there about how crazy the CPI is, and specifically CPI Housing.
> You should start your One True CPI and license it; you'd be rich.
Predicting CPI and corrections to it is an established enterprise. As is predicting better inflation numbers.
> Multiple other actors,like the MIT Billion Prices project started, and others followed, measure it, and get equivalent numbers.
We're talking about housing yet you're quoting a project that specifically excludes all housing? How is this relevant? Why do I feel like you haven't read their paper?
They very explicitly show that their benchmark doesn't work outside of goods. Like they look at medical costs and show that they aren't predicted well. Sadly they don't look at housing. The project also ran at a time of very low inflation in the US so there's not much to see. It's easy to have agreement with CPI when the numbers are all fairly low, not much room to disagree at all. We can also argue how much agreement their numbers actually have with CPI.
> I took a few seconds to understand why the 6th riot was used; it's clearly correct and if you took a moment to understand instead of discounting it, youd learn. Hint: it's 6 things multiplied together - ever hear of geometric means?
Ok, I was having fun at that point. But, let's be serious. It's still stupid. And I stand by my statement, never in my life would I take a geometric mean here.
You're telling me that you would do this today? You wouldn't say set up an actual Bayesian model and compute its posterior to estimate all of these quantities? You would compute all of these indices and corrections separately? All of these little errors add up, you have no idea what's going on with them. You wouldn't want to get some confidence interval out? You wouldn't want to be able to see how variations in your assumptions can change the results? You would impute the missing numbers with whatever felt good? Come on. This isn't acceptable methodology in any field that I've ever been a part of.
> Similarly I looked at the paper you discounted because the degrees of freedom somehow offended you. It too is a good model. And, as credentials go, I have a PhD and decades of modeling experience. Even so, I'd expect someone with your claimed expertise to be able to understand these.
You're telling me that the best you can think of model wise is a linear country-wide model with an R^2 of 0.5? After decades of modeling experience that's all you've got? And that one R^2 is enough to make you believe this is a good model? That's the standard of evidence you use before you commit trillions of dollars to a model? Give me a break! This is the initial analysis an undergrad would make in a mediocre course project they didn't put much effort into. Even our undergrad courses have far higher standards than that paper.
> Well, you're left measuring rents, but that's not satisfying when a lot of people buy a home.
It's perfectly satisfactory when the accepted rationale for many, many home purchases is as an investment vehicle, rather than as just as home (which is what rent covers).
> No one with two neurons to rub together would ever adopt this methodology in the modern world.
You're seriously making the claim that the BLS has nobody with your intelligence in a powerful position?
> You're seriously making the claim that the BLS has nobody with your intelligence in a powerful position?
It has nothing to do with intelligence. And of course statisticians at BLS know that this is trash.
Any economics or statistics student could do be a better job. And certainly, no statistician worth their salt would compute anything with the methodology that BLS uses. I cannot imagine a paper being published based on this methodology. It's not that reviewers would reject it, it's that no one in a position to submit such a paper would ever do so because of the shame associated with doing such shoddy work. I'm not sure I'm doing a good job of explaining how incredibly subpar this is.
BLS has other priorities. Changing the methodology isn't easy at all. All of your old numbers don't make any sense anymore. You can't go back and recompute them. And people would fight over what methodology to use.
Everything they do follows this insane totally outdated approach, it would be really confusing to have parts that are modern and parts that are ancient; perhaps even more confusing than the current mess.
There are no incentives to change anything. Who will make them change something? Congress? They have a worksheet, they run it, they get numbers, garbage in/garbage out, people use the numbers regardless.
Hardly anyone knows how these numbers are computed. Hardly anyone cares that they are unscientific trash. Pretty much everyone stops at the title associated with the number.
> Any economics or statistics student could do be a better job. And certainly, no statistician worth their salt would compute anything with the methodology that BLS uses.
So every bond trader, Fed observer, academic economist, investment bank economist, pension fund, union, etc, are all aware of this bad methodology in the BLS CPI, and are saying nothing? All these people with heaps of money on the line in one of the most looked-at statistics published are… ignoring the problem?
> Hardly anyone knows how these numbers are computed. Hardly anyone cares that they are unscientific trash. Pretty much everyone stops at the title associated with the number.
Really? Because MIT's Billion Prices Project (BPP; an index of internet prices, scraped daily) found the CPI and the BPP are very tightly intertwined:
> So every bond trader, Fed observer, academic economist, investment bank economist, pension fund, union, etc, are all aware of this bad methodology in the BLS CPI, and are saying nothing? All these people with heaps of money on the line in one of the most looked-at statistics published are… ignoring the problem?
No, they make money off the problem by computing better numbers internally. And there are countless papers out there criticizing the BLS for their shoddy methodology.
> Really? Because MIT's Billion Prices Project (BPP; an index of internet prices, scraped daily) found the CPI and the BPP are very tightly intertwined:
Funny someone else brought that up. That's "true" as long as you don't look at any of the details and don't read their publications.
They don't measure housing for one thing, and that's what we were talking about here. Only goods. They even show that services aren't well estimated at all. And even some categories of goods aren't. And they unfortunately ran at a time of historically low inflation. It's much easier to predict a number that is very low and has very small range.
So.. the MIT project is about the wrong category of goods. Computed in a totally different way by the BLS. Nothing in my post applies to it. And they explicitly show that they don't agree with CPI on categories like it (simply because their project isn't relevant to them).
There is no internationally agreed upon method for housing/shelter, with pros and cons for each:
> International statistical agencies have unanimously adopted the net acquisition approach for durables, but there is no consensus about the best approach to the treatment of OA in the CPI16 (Table 1). Rental equivalence is the most popular approach among countries belonging to the Organisation for Economic Co- operation and Development.17 Johnson’s (2015) recent review of the U.K. CPI proposes using CPIH, which includes the costs of OA and is based on a rental- equivalence approach, as the U.K.’s main measure of inflation. Several countries in the European Union have refrained from incorporating OA into their CPI, although Eurostat is currently conducting a pilot study for the euro area based on the net acquisition approach. Australia and New Zealand use a net acquisition approach, while Sweden and Finland—like Canada—are using a partial user-cost approach. No country has adopted a full-fledged user-cost approach.
People also often confuse the price of the asset (e.g., land) with the price of the home/structure.
> The [US] BLS views housing as a mostly “investment” item as opposed to a consumption item. So, for instance, when you consume a hot dog and have to replace it then the cost of replacement is a direct reflection on your well-being. A $1 hot dog that costs $2 one year later is a material change in living standards, all else equal, since the hot dog is an asset that you literally consume. A house is much more complex. [...]
> Of course, anyone who owns a house knows that it’s not that simple. You do basically consume your house over time. For instance, my home has appreciated substantially since I purchased it just 5 years ago and underwent a hellish remodel. At that time the cost of replacement was roughly $300 per square foot. But in the ensuing years the cost of replacement has increased to $400 per square foot. As my physical home falls apart over the years I will need to replace it. But the key point is that, as I replace these components the housing market is likely to revalue the total home value to account for this investment. So even though I am consuming my house over time I am very likely to recoup those costs.
The house (home) is not being sold for one million dollars. The land is being sold for $950K, and the buyer is getting $50K worth of structure.
Similarly if you are a home owner your home/structure depreciates over time so you have to do maintenance on it over time. And because of various shortages (e.g., lumber, plumbing fixtures) the cost of upkeep has gone up. Upkeep is a part of the CPI: the "C" in CPI stands for consumer.
And why the component is called Shelter and not Housing, at least in Canada:
Canada has always had far better statistics than the US. Both for CPI and for unemployment. This often leads to Canada looking worse, for example, by having a higher unemployment rate, when instead, Canada just has a much better way of counting who is actually unemployed.
> There is no internationally agreed upon method for housing/shelter
Agreed. But, I think that we can also agree that basing most of the number on asking people what the imaginary rent of their home might be is totally worthless.
> In the Canadian CPI paper there is some explanation towards the complexities of shelter / owner accommodation:
And you'll note that Canada doesn't use Owners' Equivalent Rent, never mind making it the majority of the index. Instead it actually computes what it would likely cost to rent that dwelling: mortgage, replacement cost, property taxes, maintenance, and other costs. That's a far more rational way of doing it.
And plenty of folks in Canada think like you do: that the official way Is Wrong and Some Other way should be used—often pointing to the US as the way it should be done. And yet here is someone in the US saying the Canadian way should be the way. ¯\_(ツ)_/¯
At the end of the day, you have some things that are a huge component of the CoL for some people (housing, car) that are fairly modest for others who have a paid-off mortgage or live in a LCoL area and either don't own a car or drive a paid-off old one.
Assuming the inflation of those items differ significantly from a broader basket of goods, there's no way to have a single measure that reasonably represents price increases for the two groups.
Owners equivalent is an insane, distortionary metric here when you can just look at the market data for similar sized properties. No one who was going to rent their house would do a survey of what their neighbors with similar sized homes think they should charge for it. They’d look at what the market rate is for their neighborhood. This is, of course, what property management companies actually do on behalf of owners.
Further, there is no internationally agreed upon method for housing/shelter, with pros and cons for each:
> International statistical agencies have unanimously adopted the net acquisition approach for durables, but there is no consensus about the best approach to the treatment of OA in the CPI16 (Table 1). Rental equivalence is the most popular approach among countries belonging to the Organisation for Economic Co- operation and Development.17 Johnson’s (2015) recent review of the U.K. CPI proposes using CPIH, which includes the costs of OA and is based on a rental- equivalence approach, as the U.K.’s main measure of inflation. Several countries in the European Union have refrained from incorporating OA into their CPI, although Eurostat is currently conducting a pilot study for the euro area based on the net acquisition approach. Australia and New Zealand use a net acquisition approach, while Sweden and Finland—like Canada—are using a partial user-cost approach. No country has adopted a full-fledged user-cost approach.
Seems like any particular system has pros and cons. Someplace else in this thread someone mentioned that Canada's system for Shelter is better, but even StatCan recognizes limitations of what they use:
And plenty of folks in Canada think like you do: that the official way Is Wrong and Some Other way should be used—often pointing to the US as the way it should be done. And yet elsewhere in the thread someone in the US saying the Canadian way should be the way. ¯\_(ツ)_/¯
You’re completely missing my point: we already have firms who do what the BLS claims to be doing that have greater predictive power and determinative impact on what people play for housing. Ignoring that reality to shrug and go “lots of countries measure things in ways that don’t make any sense so what can you do” is not a refutation of anything.
> You’re completely missing my point: we already have firms who do what the BLS claims to be doing that have greater predictive power and determinative impact on what people play for housing.
I agree that what the BLS is doing is not predictive. Because that's not the point of the CPI: it's measuring what happened in the past.
That's why the numbers that were released on October 12, 2023 were for what happened in September 2023, not November 2023.
See also determinative, which is what the BLS is trying to measure, i.e. what would be actually paid were the housing made available as a rental. No firm trying to calculate rents for a new rental would do what the BLS does, which is why their number are bullshit. If you can’t acknowledge this, it seems like you’re just being intentionally obtuse.
> Sometimes I feel like people read some econ 101 and the admittedly intuitive idea that inflation is “always and everywhere a monetary phenomenon”, but didn’t stop to think that it might be a lot more complex than that.
This is the case with almost every armchair economics comment. The same thing happens with the complexity of affordable housing: “It’s all just supply and demand! Econ 101!”
Let's take a deep dive into how CPI housing is computed. You would think it could just be "track the price of every home sale in the US", or "track the price of every new rental". Ah, my naive friend, the US Bureau of Labor Statistics lives in the year 1900 as if data, electricity, compute, and statistics don't exist. Anyway, those numbers would be too scary and too objective.
First, the real kicker, CPI housing doesn't measure the cost of a house! At all. Their rationale is that the a house is an investment. So it has no relationship to the Case-Shiller index at all.
What do you measure if you don't measure the cost of housing? Well, you're left measuring rents, but that's not satisfying when a lot of people buy a home. So.. why don't we measure the imaginary rent that someone thinks that their house that they bought 30 years ago might be able to fetch.
3/4ths of the housing index is computed from a single quantity called the "Owners’ equivalent rent of residences". They ask people in a survey: “If someone were to rent your home today, how much do you think it would rent for monthly, unfurnished and without utilities?”
Let that sink in for a moment. People who aren't renting their home. Who aren't real estate agents. Who have no idea what the cost to rent their home is. Who are homeowners and haven't rented a home in decades probably. "Bob, what do you think this would rent for?" "Ugh, $1400/month" Great! Let's base our entire understanding of the economy on the random answers of clueless people.
But, remember, the Bureau of Labor Statistics needs to do everything like we're in the pre-electricity era cruising around on the latest model of horse. So they can't ask too many people. Only a few thousand people, and they keep asking the same people over and over again. Seriously, they ask the same few thousand people every 6 months what their rent is and what their imaginary rent is. They even admit this doesn't change much. And they admit this means all the data they get is garbage, it's too skewed by the particular properties of those few thousand people they randomly selected once.
They could just ask 100,000 people or at least pick a different group of a few thousand every time? And then everything averages out eventually.
Doing that with your horse and buggy is way too hard. Instead they adjust the junk responses they get for the type of housing, quality, etc. With a whole bunch of magic numbers that come from yet another survey of clueless people, the Department of Energy’s Residential Energy Consumption Survey, which was last run in 2020 (although I think housing CPI is computed with the 2015 run) and asked about 5000 households a bunch of questions + got some billing info from their utilities.
They compute monthly CPI with an amazing formula that I think is the only use of a 6th degree square root I've ever seen in statistics (and I have CS/math degrees and do ML research; it's probably legit, but I'm not working backwards to figure out how they derived this).
Oh, we're done yet! Housing also gets older all the time, we need to correct for this. That's a whole pdf. That I've never read. And I'm not sacrificing my sanity for HN. https://www.bls.gov/cpi/quality-adjustment/updating-housing-... But it has a model from January 1997 with 110 degrees of freedom. That's not confidence inspiring to say the least.
Ok, so we ask clueless people to give us meaningless numbers that we then adjust based on complex largely meaningless formulas to compute an imaginary rent that doesn't exist instead of just looking at the actual cost of housing. How could this get worse? With the magic of "imputation"! A beautiful word from statistics that means "Oh, we didn't collect enough data and we have no idea what the actual number should be, so we're just going to make a number up and pretend like it's a data point we gathered based on some idea of what we think it ought to have been". Because the best data is the data you make up according to your gut.
Feel free to read the horror show for yourself https://www.bls.gov/cpi/factsheets/owners-equivalent-rent-an...
No one with two neurons to rub together would ever adopt this methodology in the modern world.
Of course none of these numbers match any of the reality that anyone experiences. Their documentation on how anything is computed reads like standup comedy.