US Debt, the gift that keeps on giving

October 31, 2006 5:34 AM

Borrow and spend, the delayed pain Ever thought of giving the future generations a little something to remember you by? How about the gift of debt? Whatever lens you want to look at the graph, you can blame the rise on: Out of control spending on the military or social programs; The end of the US being the swing producer of oil; Expensive wars; The money no longer being tied to gold or silver and instead being a fiat currency

No matter what the reason for the spending, the GAO warns: If the United States government conducts business as usual over the next few decades, a national debt that is already $8.5 trillion could reach $46 trillion or more, adjusted for inflation. Or If you don't like the graph due to the static nature, how about a clock so you can watch the numbers move upward?
posted by rough ashlar (80 comments total)

From the 'business as usual' link:

with the first baby boomers becoming eligible for Social Security in 2008 and for Medicare in 2011, the expenses of those two programs are about to increase dramatically due to demographic pressures. People are also living longer, which makes any program that provides benefits to retirees more expensive.

Medicare already costs four times as much as it did in 1970, measured as a percentage of the nation's gross domestic product. It currently comprises 13 percent of federal spending; by 2030, the
Congressional Budget Office projects it will consume nearly a quarter of the budget.

Economists Jagadeesh Gokhale of the American Enterprise Institute and Kent Smetters of the University of Pennsylvania have an even scarier way of looking at Medicare. Their method calculates the program's long-term fiscal shortfall — the annual difference between its dedicated revenues and costs — over time.

By 2030 they calculate Medicare will be about $5 trillion in the hole, measured in 2004 dollars. By 2080, the fiscal imbalance will have risen to $25 trillion.

Will Medicare survive? Or will 'the powers that be' simply default on sick Americans?
posted by Fuzzy Monster at 5:40 AM on October 31, 2006

It'll be interesting to see if Bush can break the 14-digit barrier before the end of his term is up. That may, in fact, be what he'll be able to point to as the only unambiguous accomplishment of his presidency.
posted by psmealey at 5:43 AM on October 31, 2006

He's come this far without taking any damn advise from anyone, reguardless of credentials or expertise. Why should he start now?
87% of all statistics are made up on the spot and meaningless anyway.
posted by shnoz-gobblin at 5:47 AM on October 31, 2006

Bush can only spend so long as:

A. Incoming tax revenue is capable of paying the interest on current debt
B. We can find a country or bank willing to lend to an out-of-control debtor

Current spending is unsustainable insofar as the above conditions are economically unsustainable.
posted by Blazecock Pileon at 5:51 AM on October 31, 2006

Or will 'the powers that be' simply default on sick Americans?
posted by Fuzzy Monster at 5:40 AM PST

Plenty of stories about the lack of care for Vets.
So there is 'defaulting on 'obligations' already.

I'd bet the angle for the medicare cuts will be "the majority of cost is for end of life care, when the human is no longer a wage earner....cut the expense that results in little to no gain to the treasury." That and "preventitive care".
posted by rough ashlar at 5:52 AM on October 31, 2006

Plenty of stories about the lack of care for Vets.
So there is 'defaulting on 'obligations' already.

Good point, rough ashlar.
I agree that politicians will cut Medicare further. But I don't know how Medicare could be cut without disgruntled Boomers rising up in a massive Grey Wave. It's a huge problem: no wonder politicians on both sides of the spectrum have been ducking this issue. Kudos to GAO Chief David M. Walker for trying to bring this to America's attention.
posted by Fuzzy Monster at 6:02 AM on October 31, 2006

Will Medicare survive? Or will 'the powers that be' simply default on sick Americans?

No -- Medicare will collapse along with the rest of the health system. Medical costs are one of the things that are killing America -- land of the best medical system in the world, so long as you can afford it.

Not only is this cost hurting things badly in the budget via Medicare and Medicaid, it's hurting the entire economy. But we don't *dare* do anything about it, that might hurt the markets.

Medicare & Medicaid can't be fixed, becaue the problem is intractable -- costs are going to rise faster than inflation. We can't fix the problem here, we need to fix the entire health care system.

This won't happen anytime soon, and I honestly think it will take something like the collapse of the entire country to change it.
posted by eriko at 6:06 AM on October 31, 2006

Silly people. The conservavites are simply expecting the liberals to, in their time, legalize euthanasia, resulting in lower spending. Then, after the die off, they will revive their 'right to life' slogans and demand a return to 'morals'.

This works just like the old 'tax and spend' nonsense where the so-called 'conservatives' spend so they can force the liberals, in their turn, to tax, then win on an anti-tax platform (while accusing the liberals of having done the spending).
posted by Goofyy at 6:10 AM on October 31, 2006

Before delcarling doomsday, compare situation when Clinton was in his final months in office. In fact, money comes in when there are sufficient taxes to raise money and when spending is not way out of bounds...you can cut back on services, or you can bring back or restore or add taxes, and additionally, you can trim military spending. After all, how many bases do we have in how many countries and how many troops spread wordwide and for What purpose? In Europe to protect agains the Russian menace?
posted by Postroad at 6:17 AM on October 31, 2006

This works just like the old 'tax and spend' nonsense where the so-called 'conservatives' spend so they can force the liberals, in their turn, to tax, then win on an anti-tax platform (while accusing the liberals of having done the spending).

That bears repeating. Here in Maine our economy's in trouble, and it's blamed on the "tax and spend" local democrats who barely hold a majority. The reality is that the state has to make up all of the shortfall caused by the federal spending cuts over the last 10 years or so. It's just shifting the burden to and fro and blaming the other guy.
posted by SteveInMaine at 6:20 AM on October 31, 2006

.
posted by xammerboy at 6:21 AM on October 31, 2006

How about the gift of debt?

This is actually ironic, since debt can be a good thing for countries.
posted by smackfu at 6:30 AM on October 31, 2006

A related prior post with many good comments.

We're in truly deep shit.
posted by Malor at 6:30 AM on October 31, 2006

Will Medicare survive? Or will 'the powers that be' simply default on sick Americans?

Or maybe we'll just default on our loans and our new overlords will repossess us.

If that sounds depressing, just think of the jobs it could create--we can have the call centers located back here again!
posted by leftcoastbob at 6:31 AM on October 31, 2006

This is actually ironic, since debt can be a good thing for countries.
posted by smackfu at 6:30 AM PST

Please feel free to explain how the present level of debt is 'a good thing'.
posted by rough ashlar at 6:40 AM on October 31, 2006

This is actually ironic, since debt can be a good thing for countries.

Debatable. I've heard from the right as well as from the left that gifts of debt to improvished nations (particularly those in Africa) do nothing more than prolong the cycle of immiseration rather than provide seed money to stimulate self-sufficient enterprise. I grant you, in the face of famine or civil war, it may be the only decent thing to do, but this far from a certainty in most economic circles.

Also there's the "Confessions of an Economic Hitman" argument that makes the debtors even worse off than they were before. His example was this: American government goes into Peru telling them that we'll "give" them foreign aid (loans), on the condition that they use the proceeds to pay Bechtel and Brown & Root to build - largely unneeded - power plants. This way, the American companies get the both the proceeds of the loans and Citibank gets the interest income on the debt. In this case, debt is good. Good for us. For Peru, not so much.
posted by psmealey at 6:41 AM on October 31, 2006

Maybe Medicare, Medicaid, and Social Security can be cut & restarted at the state level (where they should have been al along).

But I also kinda suspect only some larger collapse will change our governments spending habits. Spending is just way to easy if your the worlds reserve currency.
posted by jeffburdges at 6:46 AM on October 31, 2006

But we don't *dare* do anything about it, that might hurt the markets.

The same "we" is protecting paying corporate interest, for instance in the form of duties/tarifs, whatever can make stuff that is made outside more expensive, so that it becomes less competitive with internal interests.

So I guess the market is already very much influenced, more then we may imagine : I ask myself what is the point of obsessing with maintaning the free market theory operative (or a second-best of free market) when this implies betting our welfare for the future , while we are NOT getting what we really need from the same system ?

We can't fix the problem here, we need to fix the entire health care system.

it's the whole goddamn approch that is broken. Why are all the benefits of scale, low costs of mass production and advancement in technlogy being privatized ?

For instance if producing a goddamn DVD costs more or less $1 a copy and $15-20 are asked per piece, how comes the sale price isn't approaching marginal cost of production per unit ? Why should I sustain ALL the risk of their production NOT being entirely sold by paying for 20 copies of the SAME good ?

Similarly in medicine, why should I sustain a monopoly privilege (aka patent) over a drug , whose cost has been paid over and over and over again possibily an hundred thousand times, WITHOUT getting some sensible benefit ?

Similarly in transportation, in Italy public money times and again paid the construction and maintenance of highways ; yet our govt decided it was a good idea to give "concessions" to privates, allowing them to get the tolls cash flow (which already existed btw, but the money went back to State) while tolerating an inferior maintenance and few new works ; why didn't these privates PAY for the privilege of accessing such an enormous cash flow ? One can't exact tolls without an highway system, which costs trillions indeed...why didn't privates pay for that ? Where is the value going, as it is certainly not going back into public benefit or coffers ?
posted by elpapacito at 6:47 AM on October 31, 2006

Kinda off on a tangent, but I hear the Supreme Court will hear a case regarding the awarding of very large punitive damages against corporations for egregious actions. (a tobacco company and the state of Oregon in this case)

The argument goes that punitive damages should not be in excess of about 5 times the compensatory damages awarded.

I say let's apply that thinking to the tobacco company executive and legal representative salaries as well, huh? Five times the lowest paid worker, outsourced to India or not.
posted by nofundy at 7:01 AM on October 31, 2006

Needs to calculate total unfunded liability, not just cash balance.
posted by delmoi at 7:28 AM on October 31, 2006

Or maybe they'd decide that America should be building machine tools. I remember hearing about that, a long time ago. Once, it was considered a vital part of both the economy and national security. Nowadays, I think they practice faith-based economics with regards to such old fashioned concerns as machine tools. Besides, the Chinese will always be ready to sell us that thingamagig we need to repair the all important widget, right?
posted by Goofyy at 8:30 AM on October 31, 2006

I wouldn't pay much attention to anything on the "clock" page (last link). For a start, they claim that there haven't been any budget surpluses since 1993 — but that's because they're looking at nominal debt instead of real debt, which is a pretty basic mistake.
posted by matthewr at 8:40 AM on October 31, 2006

should be building machine tools

no machining, no guns.

I'm no macro-economist but I see are two kinds of government debt; international capital investment with the expectation to be repaid, and simple under-taxation of domestic wealth.

With the obscene L-curve we've got in this country

A lot of the Bush tax cuts just went into the real estate market, driving up valuations for everyone. Fat lot of good it did there.

Every T-bill sold is a dollar that could and should have been taxed -- either that or we've got to cut spending so much on stuff.

But hey, I've got a funded treasurydirect.goc account now so I've become The Man.
posted by Heywood Mogroot at 8:46 AM on October 31, 2006

I would care a lot more about the whole "National Debt" thing if I understood what a debt actually means at the nation-state level.

At a very simplistic level, a nation controls the currency supply. Why should it ever go into debt at all? And before you say "printing money causes inflation", allow me to point out that the money supply has been increasing at 6% per year for quite a while now, and inflation is nothing near that.

Concepts such as money supply, the (lack of a) gold standard, and unsecured debt make the whole topic extremely complex.

--------------------

In my experience, bringing up the National Debt is usually an appeal to emotion. No one likes to be in Debt, and the thought that our nation could be in Debt as a whole is an emotionally potent idea. And of course, as the OP said, "Won't someone think of the children!?!"

As a whole, the issue leaves me cold. It's clear that not one person in a hundred (including myself) have any idea what the implications of a National Debt *might* be, and that the 1% who have a clue are in violent disagreement about what the implications actually *are*.
posted by tkolar at 8:48 AM on October 31, 2006

For a start, they claim that there haven't been any budget surpluses since 1993 — but that's because they're looking at nominal debt instead of real debt, which is a pretty basic mistake.

? annual deficits are positive or negative regardless of inflation.

The actual numbers:

09/30/1993 $4,411,488,883,139.38
09/30/1994 $4,692,749,910,013.32
09/29/1995 $4,973,982,900,709.39
09/30/1996 $5,224,810,939,135.73
09/30/1997 $5,413,146,011,397.34
09/30/1998 $5,526,193,008,897.62
09/30/1999 $5,656,270,901,615.43
09/29/2000 $5,674,178,209,886.86
09/28/2001 $5,807,463,412,200.06
09/30/2002 $6,228,235,965,597.16
09/30/2003 $6,783,231,062,743.62
09/30/2004 $7,379,052,696,330.32
09/30/2005 $7,932,709,661,723.50
09/29/2006 $8,506,973,899,215.23
posted by Heywood Mogroot at 8:52 AM on October 31, 2006

"Government is not the solution to our problem, government is the problem."
--Ronald Reagan.

In the past 8 years:

*Largest increase in discretionary spending of any administration since LBJ

LBJ %25.2
Nixon %-16.5
Reagan %11.9
Clinton %-8.2
Bush 43 %35.8

*Largest increase in defense spending of any adminstration since LBJ

LBJ %26.7 (Vietnam build-up)
Nixon %-32.2
Reagan %34.8 ("Star Wars" / Cold War)
Clinton %-16.8
Bush 43 %44.5 (post 9/11)

*Largest entitlement program since the "Great Society" Medicare was established (presecription drug program)

*War on Iraq premised on the idea of bringing Democracy to every nation on the globe.
posted by stbalbach at 8:54 AM on October 31, 2006 [1 favorite]

6% per year for quite a while now, and inflation is nothing near that

Certainly wages aren't rising anywhere near that, but prices (eg. rents) sure are.

As a whole, the issue leaves me cold

The issue is completely abstract, other than if & when the wheels come off, and your personal share of the annual interest payments on the debt. Regardless if we ever pay it back/down or not, the Interest. Must. Flow.
posted by Heywood Mogroot at 8:58 AM on October 31, 2006

LBJ was required by his Congress to institute a 10% surtax (mostly socking the rich) to cover HIS military buildup. I wonder how much support for the war(s) there would be if we actually had to pay for it.
posted by Heywood Mogroot at 9:00 AM on October 31, 2006

Heywood Mogroot: "? annual deficits are positive or negative regardless of inflation."

You're missing the point. The clock link guy is using nominal national debt figures to determine whether there was a budget surplus or deficit. This is wrong — real figures should be used.

The official nominal national debt increase (ΔB) is overstated by π×ΔB, where π is inflation.

For example, say the official (nominal) debt is unchanged. According to the "clock" link, this means the budget deficit is zero. Say inflation is 10%. At the end of the year, the real value of the debt has decreased by approx· 10%. In reality, the government has therefore run a budget surplus of 10% of B, where B is the initial level of debt.
posted by matthewr at 9:25 AM on October 31, 2006

From this link:

The National Debt will cause the United States to go bankrupt

We have to pay back all of the $5.6 ($9) trillion in debt

The interest payments on the national debt are a burden to future generations.

Foreign ownership of the debt causes money to flow out of the U.S.

The national debt is out of control.

All these five statements are myths. The 'realities' are on the linked page.

posted by matthewr at 9:31 AM on October 31, 2006

In reality, the government has therefore run a budget surplus of 10% of B, where B is the initial level of debt

Only if wages / actual productivity (not the fake GDP they are using now) are rising that 10% too. Rising prices -- by themselves -- do not ameliorate the situation.
posted by Heywood Mogroot at 9:40 AM on October 31, 2006

(this is IMO . . . what I need to learn about macro-economics is more than I know)
posted by Heywood Mogroot at 9:41 AM on October 31, 2006

Only if wages / actual productivity (not the fake GDP they are using now) are rising that 10% too.

Why? I have no idea what this sentence means. What alternative measure of economic growth (instead of the "fake GDP") are you proposing?

Rising prices -- by themselves -- do not ameliorate the situation.

The nominal national debt was, in 1791, $75.6m. You're saying that if it was $75.6m in 2006, the debt "situation" would be unchanged. Hmmmm.
posted by matthewr at 9:52 AM on October 31, 2006

what I need to learn about macro-economics is more than I know

Indeed. There are plenty of online sources of macroeconomic education — I think MIT's OpenCourseWare scheme has some first-year macro, and Berkeley puts some first-year lectures on iTunes for free.
posted by matthewr at 9:55 AM on October 31, 2006

The 'realities' sorta miss the mark, too.

1) is a dodge; surely the borrowing capacity, without harmful effects -- of the USG is not infinite

2) Sure, we don't have to "pay it back", but we DO have to pay the interest. Interest never rests, and when you are BORROWING to pay it you are truly f'ed.

3) "Thus we are just paying interest to ourselves."

In this case we are paying interest to people who should have been taxed instead. What a great racket.

4) Foreign holders of the debt are at $2.1T now. At 5% this is $105B/year in interest payments. At 8% this will be $170B/yr. I wish we could actually sell them something, other than capital or land, to recoup this outflow.

5) The debt-to-GDP ratio for the U.S. is relatively modest compared (1) to other nations and (2) historically

GDP is a bogus number these days; factor out the GDP rise due to one-time deals like mortgage equity extraction (2-3% of GDP for the past few years) and the picture looks a lot worse.

And comparing ourselves to Italy and Poland is disengenuous. We, as a country, have fallen far since 1980 are we are accelerating.
posted by Heywood Mogroot at 9:57 AM on October 31, 2006

At a very simplistic level, a nation controls the currency supply. Why should it ever go into debt at all? And before you say "printing money causes inflation", allow me to point out that the money supply has been increasing at 6% per year for quite a while now, and inflation is nothing near that.

Actually, a lot of us (me included) have been jumping up and down and screaming about this for years. We HAVE been seeing inflation, very intense inflation. It's just been in the stock market, up until it crashed in Y2K, and then in real estate. The problem with that kind of inflation is that it's seductive... people LIKE it when their house is worth 20% more every year.

The dollar has dropped like a rock since Y2K, and the only reason it's not at least 30% lower is because Japan and China have been mopping up excess dollars to keep their own currencies low. Each bank has hundreds of billions of dollars and dollar-equivalents. By doing this, they in essence import our inflation... they print more of their currency to buy ours, making their currency cheaper and easier to get. This causes a powerful economic stimulus over the short term, but leads to very nasty inflation eventually. It preserves the status quo, and allows our Fed to abuse the system longer. However, this means that their economies are ultimately taking the hit, too. It's short-term thinking.

They're going to get tired of mopping up our dollars; if they buy a carload of them now, a freightcar shows up next month. Buying the freight car means a whole train shows up..... eventually it'll be metaphoric ocean liners full.

Ultimately, this means the world sends us their valuable goods and services, and we give them green paper (or electronic blips) in exchange. This is fantastic for us over the short term, but when people realize just how badly they've been robbed, they're gonna be PISSED. And then we'll have to deal with the shock of goods and energy that are suddenly twice as expensive... when the government is already in debt to such a proportion that it can't possibly be paid back.

It appears that we're just about out of options, and I think the next five or ten years are going to be very, very painful.

This has been building for many years, ever since Greenspan took over. His one and only solution to all problems was 'more liquidity!'. This was building all through the Clinton years, with the Clinton government's enthusiastic assistance. The Bushies didn't do anything to fix it, and certainly share in the culpability, but it's not all their fault. When the shit really hits the fan, the party in power is gonna take the blame, and folks aren't going to realize that, while both parties got us into this mess, it was most especially Clinton's administration to blame.

For a start, they claim that there haven't been any budget surpluses since 1993 — but that's because they're looking at nominal debt instead of real debt, which is a pretty basic mistake.

The United States government, one of the most complex fiscal entities ever created, uses cash-basis accounting. You cannot trust anything they say; their numbers are outright fabrications. Have been since LBJ and the Great Society. The one and only entity that's actually got any kind of grasp on the true financial situation of the US Government is the GAO... and they're saying 46 trillion.
posted by Malor at 9:59 AM on October 31, 2006 [1 favorite]

Couldn't they just shift the military spending towards medicare spending? I mean, what's the difference between propping up a military industrial complex versus propping up a pharmaceutical industrial complex? Oh yeah, one aims to save lives.
posted by furtive at 10:05 AM on October 31, 2006

In reply to Heywood:
1. You're misrepresenting what the author said. No-one claimed the government could incur "infinite" debt as you say: rather, "large amounts of debt".
2. Why are we "truly f'ed"? See (1).
3. No idea what this means. The interest payments are not going outside of the US economy. Some negative income distribution effects may occur, as the author mentions in problem 1.
4. "this proportion has remained virtually unchanged since 1980. They typically reinvest their interest payments in the U.S."
5. Show me one academic economist (as opposed to 'random dude on the internet') who believes that GDP figures are wrong enough to be "bogus". There may well be biases; whether these biases add up to a significant overstatement of correct GDP is far from a foregone conclusion.
posted by matthewr at 10:20 AM on October 31, 2006

matthew, if you want to see just how bogus the GDP is, just look at the computer numbers. Most of the 'legendary' productivity gains during the 90s came from this source, and it was simple numbers games.

Back in the late 80s and early 90s, computer were moving very quickly, and the government at the time tried to capture this effect by adjusting the dollar figures. They figured that if we shipped an 8Mhz 286 last year, and a 16Mhz 286 this year, well, computers are twice as good. But they don't cost any more than last year's model, so productivity (and the GDP component from computers) doubled!

That argument isn't completely ridiculous, because at the time, computer speed was a terrible limitation, and you really COULD get twice as much work done with a 16Mhz 286. But once computers got fast enough to run all the basic office applications well, the productivity gains from CPU speed boosts pretty much stopped. But it hasn't stopped in the government figures. Each year, the computer GDP goes up a GREAT deal more than the actual dollars being spent. The last time I looked at this issue, four or five years ago I think, the actual cash being spent was like 50 billion/year, but according to the government, the 'real' figures were north of 450 billion.

Beware anytime the government uses the word 'real' with regard to money, because 'real' always means 'fake'. I'm serious.

They do the same thing with cars, although that's more to limit the inflation figures. Cars get more expensive every year, but according to the government, they don't get more expensive because they have more features.... even though we don't have any choice about buying them, and can't opt out and get a cheaper car. According to the government, there has been very little auto inflation since the 1980s, when you could buy a cheap new car for about $7500.

They do a similar trick with housing... to try to hide housing inflation, they came up with this bizarre fiction called 'owner's equivalent rent', which basically was a way of adjusting the price down so there wasn't any inflation. I don't remember the details now, but look it up.

The worst thing they're doing is in tracking the CPI. For a long time, they tracked the cost of a 'basket' of goods and services. Last year sometime, they decided to change that. They said, "consumers will buy what's cheap", and now they change the basket of goods and services every month. I'm serious. If beef goes up, but chicken stays flat, there's no inflation. If vegetables go up, but okra is down, everyone will buy okra. No inflation.

In other words, they have completely abandoned the original reason to track inflation... to measure declining standards of living from increasing prices.

Why? Because if inflation is high, it costs the government a huge amount of money. They stand to gain many many many billions from lying about it. So, they lie.
posted by Malor at 10:39 AM on October 31, 2006 [1 favorite]

Malor: "The worst thing they're doing is in tracking the CPI."

I assume what you are describing in this paragraph is the change to "chain-weighted" measures of inflation, replacing the Laspeyres-index CPI measure. You say it is the "worst thing". In fact, it is an excellent thing.

It makes no sense to use a completely fixed basket of goods and services — clearly, we don't buy the same things now that we did in 1900. To get around this, the basket was changed every five or six years. This still isn't fast enough to take account of consumption changes.

If the relative prices of chicken and beef changes in favour of chicken, of course consumers will buy more chicken. Why would they not? The old CPI Laspeyres index would not take this substitution bias into account, hence introducing an upward bias to inflation measures. Chain-weighted measures use a constantly-varying 'basket', so they do not suffer from this problem, and are therefore a more accurate measure of cost-of-living.
posted by matthewr at 11:14 AM on October 31, 2006

The clock is actually going down. :P
posted by fusinski at 11:17 AM on October 31, 2006

Ha! I hadn't noticed that, fusinski.
posted by matthewr at 11:20 AM on October 31, 2006

The substitution bias is another way to say declining standard of living. It's enshrining a declining standard as the normal way of doing things.

Consumption patterns are not inflation. Yes, inflation causes consumption changes. But arguing, as the government does, that the consumption changes caused by inflation means that no inflation is taking place is an outright lie. It's a deliberate attempt to hide declining standards of living.

In other words, they have completely abrogated their original responsibility. The CPI is now a fictional number that's just whatever they want it to be this month. It no longer measures anything meaningful.

How you can argue that this is a good thing -- how you can confuse consumption and inflation -- is entirely beyond me.
posted by Malor at 11:25 AM on October 31, 2006 [1 favorite]

The CPI, chain-weighted or not, doesn't attempt to measure standard of living. Instead, it measures cost of living.

Thanks to chain-weighting, it now measures more accurately the prices of what people are actually buying. How would using a fixed basket be an improvement on chain-weighted measures? A rise in the price of Spam may have increased the cost of living of the average consumer in 1940, but nowadays it could double and the average consumer wouldn't care.

As well as the substitution problem, there's the problem of new goods. Given the fixed basket was only updated every five years, if a good was introduced just after the basket was updated, the rapid falls in its price over the five years would be completely ignored. For example, say the price of DVDs and DVD players fell sharply during the five years after its introduction and then stabilised. The CPI would completely ignore the fall and only include the stable-prices period, thus overstating inflation.

Incidentally, how does a basket that changes every five-six years avoid your criticism about substitution bias?
posted by matthewr at 12:16 PM on October 31, 2006

of course consumers will buy more chicken

Soylent Green is the endgame, man, Soylent Green.
posted by Heywood Mogroot at 12:16 PM on October 31, 2006

Yes, inflation causes consumption changes.

Ceteris paribus, this is wrong. Relative price changes cause consumption changes. If all prices (including the price of labour) rise at the same rate (via the ceteris paribus assumption), the consumer's budget constraint and choices are unchanged.
posted by matthewr at 12:25 PM on October 31, 2006

Matthewr:All these five statements are myths. The 'realities' are on the linked page.

Your link seems to be out of date from around the year 2000. Things have gotten drastically worse in the last six years. Interest on the public debt represents about 24% of personal income tax revenue. In other words, you could have a 24% tax cut if it weren't for the interest on debt. This is tax money that is going from workers' labor to the wealthy that hold treasury bonds. The portion of debt held by foreigners has doubled in the last six years to about 45%. That means that your taxes are going not just to wealthy Americans, but increasingly to foreigners.

The debt as a percentage of GDP is also increasing rapidly. Historically debt has ranged from about 100% at the end of WWII to about 24% in the mid-70s. Today it is back up to 38% and is projected by the GAO to reach 75% in the next 15 years. These are not good numbers.
posted by JackFlash at 12:36 PM on October 31, 2006

1. You're misrepresenting what the author said. No-one claimed the government could incur "infinite" debt as you say: rather, "large amounts of debt".

from the link:

"All of these factors mean the federal government can incur large amounts of debt."

But how large is too large, and are we there yet? More importantly, we have some massive demographic obligations on the ten-twenty year horizon, are we getting our ship in order?

2. Why are we "truly f'ed"? See (1).

When you are charging card A to pay card B, the endgame is near. We're not there yet, but outside of fiscal meltdown scenario the number I am interested in is how much of my taxes are going toward interest on the debt.

This year it was 20% of the budget, so that I pay $x000 toward these interest payments, x being a pretty large number to me. This is no way to run a railroad.

3. No idea what this means. The interest payments are not going outside of the US economy. Some negative income distribution effects may occur, as the author mentions in problem 1.

This is a pretty important issue, especially given how present tax law has moved the tax burden from capital to labor.

4. "this proportion has remained virtually unchanged since 1980. They typically reinvest their interest payments in the U.S."

The question is how long this game can go on, and their tolerance for exposure, and when the intend to monetize this debt back into their own economies. Japan Inc is on the hook for the most, $630B or so.

5. Show me one academic economist (as opposed to 'random dude on the internet') who believes that GDP figures are wrong enough to be "bogus". There may well be biases; whether these biases add up to a significant overstatement of correct GDP is far from a foregone conclusion.

From my reading, OER and hedonics make the GDP figure bogus. You are free to disagree, of course.

Now, I don't pretend to be an expert -- I need a lot more education here.

But for my money something is rotten in the state of Denmark, part of the endgame being the breaking of the Greenspan SSTF plan of 1986, a theft of $2.1T+ from wage earners to those who are and will be benefitting from the Bush tax structure modifications of 2001-2003.
posted by Heywood Mogroot at 12:43 PM on October 31, 2006

The CPI, chain-weighted or not, doesn't attempt to measure standard of living. Instead, it measures cost of living.

The reason they measure it is to determine how much the cost of living is being impacted by inflation.... to determine what people are losing from the government's abuse of the money supply. Its PURPOSE is to measure the loss of living standards. It does that through measuring the cost of a representative basket of goods. They aren't measuring "the minimum things needed to survive". That is not the CPI's purpose. They are measuring "the cost of this basket of goods", in an attempt to discover what monetary policy is doing to the standard of living.

If the yardstick changes at the measurer's convenience, they can use the number to say anything they want to say. It loses all meaning. Your approach means we are measuring absolutely nothing and attaching a great deal of importance to it. The economy could collapse, and people could be starving in the street.... but there's no inflation because sawdust is very cheap this month.

Given the fixed basket was only updated every five years, if a good was introduced just after the basket was updated, the rapid falls in its price over the five years would be completely ignored. For example, say the price of DVDs and DVD players fell sharply during the five years after its introduction and then stabilised. The CPI would completely ignore the fall and only include the stable-prices period, thus overstating inflation.

Oh my GOD you are confused about what inflation is. Inflation is prices going up from too much money chasing too few goods. It's really pretty simple. You DON'T want new goods included, because the steep price falls of new goods happen because of economies of scale and manufacturing efficiencies. Their price drops have nothing whatsoever to do with monetary policy.

Doing things your way automatically understates inflation, by a huge degree.

Ceteris paribus, this is wrong. Relative price changes cause consumption changes. If all prices (including the price of labour) rise at the same rate (via the ceteris paribus assumption), the consumer's budget constraint and choices are unchanged.

Faulty assumptions lead to faulty conclusions. Inflation doesn't work that way. It doesn't start all in the same place, and it doesn't affect the entire economy at once. It's not like you can take a single reading and determine inflation.

If you COULD do that, they could measure just one thing, and that would be inflation. Say, the price of soybeans. The reason they HAVE a basket is because it affects different segments of the economy at different times.

In other words, the simple existence of the CPI's basket refutes this assertion completely.

As far as the 5-year schedule goes... I think that's much too frequent. The basic staples of life haven't changed that much in the last fifty years. The basket does not need frequent updates. Adjusting it every five years isn't great, but it's still 60 times better than monthly. A changing yardstick results in bad measurements.

Doing it once a decade at census time might make sense.

JackFlash: remember that that's 38% of a wildly overstated GDP. The figures back in WW2 were based against GNP. GNP probably wasn't that great a number in absolute terms, but it was enormously better than the modern GDP, which has very little to do with reality at all.
posted by Malor at 1:44 PM on October 31, 2006 [1 favorite]

Its PURPOSE is to measure the loss of living standards.

No, this is not the purpose of consumer price index. The purpose is, as I said, to measure the cost of living. Measuring living standards involves looking at incomes and prices and, potentially, all sorts of other things. One common measure of living standards is the HDI, which has nothing in common with the CPI. The idea that merely because the CPI has risen by 3% this year, living standards have necessarily fallen, is just nonsense. Even with the narrowest definition of living standards as merely the goods and services purchased, you would need to look at wage inflation to be able to say anything about living standards.

If the yardstick changes at the measurer's convenience

It doesn't. If you want to construct an index of consumer prices, looking at what actual consumers are buying seems like a good idea to me.

people could be starving in the street

Once again, cost of living, not living standards.

Oh my GOD you are confused about what inflation is

Well, this paragraph is just wrong. You're confusing the cause of inflation with what inflation itself is. You may argue, as Friedman famously and controversially did, that inflation is caused by increases in the money supply. That's entirely different from what inflation is, which is the rise in the general level of prices. The general level of prices surely includes goods that couldn't be purchased a few years ago. Why would it not?

Their price drops have nothing whatsoever to do with monetary policy.

And?

Inflation doesn't work that way. It doesn't start all in the same place, and it doesn't affect the entire economy at once.

First, that's not the point I was addressing. What you said was "inflation causes consumption changes". This is wrong. If real income fell, perhaps consumers would switch to inferior goods. But that's separate from inflation itself. Relative price changes cause consumption changes.

The basic staples of life haven't changed that much in the last fifty years.

As you say, the CPI doesn't measure "the minimum things needed to survive". It measures what the rise in nominal prices of what people are actually buying. People certainly aren't buying the same things as they were 50 years ago.

Doing it once a decade at census time might make sense.

Why does changing it every 10 years make sense when apparently every 5 years does not?

The basket does not need frequent updates.

In order to measure the prices of what consumers are actually buying, it obviously does.
posted by matthewr at 2:34 PM on October 31, 2006

Neat discussion. Some quick fact checking:

This year it was 20% of the budget

That page says interest on the debt in 2006 was $405 billion. Wikipedia says that the federal budget for 2006 was $2.6 trillion, which means interest was 15% of the budget. But this number is actually still incorrect, because it doesn't count the interest earned by the various trust funds (mainly SS). I'm a little confused by this, since I was under the impression that the trust funds are spent, not invested, and thus any interest they earn is just money pulled out of tax revenue, but that's how the CBO calculates it (see page four), and also how Daniel Gross calculated it in a recent NYT op-ed on this topic, so I'll trust the CBO definition of interest payments, which they say was $220 billion in 2006, or 8% of the federal budget.

Doing things your way automatically understates inflation, by a huge degree.

matthewr had it right the first time. He's referring to "new goods bias", an understood phenomenon. I'm not an expert in it, but there's evidently disagreement as to how big of an effect it has on CPI, as Dean Baker describes. I find Baker's argument that it only distorts the cost of living for rich people convincing, but there's no question about the direction of the distortion: it overstates CPI.
posted by gsteff at 2:43 PM on October 31, 2006

Wikipedia says that the federal budget for 2006 was $2.6 trillion

Apologies, I meant 20% of the tax receipts, not expenditures.
posted by Heywood Mogroot at 6:32 PM on October 31, 2006

which they say was $220 billion in 2006, or 8% of the federal budget.

yes, the trillion+ in the SSTF is earning interest along with all the other treasuries we have out.

This SSTF is a general-fund obligation no different than other bonds at the moment, and being a living, breathing taxpayer for the next 30-odd years I'm on the hook for it. . . . 10+ years from now future admins are going to need to find space in the budget to repay these bonds, not bank them. The way things are going now, we'll be f-ed then, but I'm more than willing to, say, come back in 2009 and review the situation with you. Not like we're going to do anything about the problem before then anyway.
posted by Heywood Mogroot at 7:36 PM on October 31, 2006

a theft of $2.1T+ from wage earners

Its not theft....they just don't know where the 2.3trillion + is
posted by rough ashlar at 8:17 PM on October 31, 2006

They do a similar trick with housing... to try to hide housing inflation, they came up with this bizarre fiction called 'owner's equivalent rent', which basically was a way of adjusting the price down so there wasn't any inflation. I don't remember the details now, but look it up.

What it basically means is this: the government assumes in its figures that a certain percentage of all homeowners are renting to themselves. There are several areas where this kind of chicanery happen—they're called imputations, and are explained far better, and in more detail here. Basically:

Imputations are a part of GDP that the government decides to estimate value, where no cash actually changed hands.

With rental imputations (explained here in the government's FAQ).

The BEA treats homeowners as businesses, which pay rent to themselves. Therefore, homeowners contribute to the real estate industry's GSP even if not employed by the industry. In addition, like businesses, homeowners' property taxes paid to state and local governments are included as part of real estate TOPI.

It's all pretty fucked up.
posted by Civil_Disobedient at 3:40 AM on November 1, 2006

I don't know enough about the rental imputations to have a strong opinion either way, but in general there's nothing wrong with trying to include non-cash transactions in GDP.

All of the following things ought, in an ideal world, to be included in GDP. They aren't included because they're very hard to measure or quantify.

The black market ("shadow economy")

Subsistence production
If a farmer grows some crops and consumes them himself, it's very hard for the government to count it, but it clearly is the production of a 'final good'. Obviously this applies mainly to the developing world.

Domestic work

posted by matthewr at 4:18 AM on November 1, 2006

Once again, cost of living, not living standards.

That's a meaningless phrase by itself. If you're not measuring against a fixed reference point, you're not actually measuring anything. If food and energy become so expensive that the poorest people are starving in the streets, there's still no inflation. All they can buy is sawdust, but it's cheap this month!

Inflation moves through an economy slowly, and hurts the people at the bottom the most. The people closest to the money, the rich people near the banks, make out like bandits, because they can use the new cash being made available to extract goods and services at the old rate. They can, in essence, use their privileged position to compete for resources unfairly. This makes those goods and services get more expensive. People farther away from the banks stop buying as many of those goods and services; their quality of life is impaired by the inflation.

But, if we use your method, there's no inflation, because they're buying chicken instead of steak. Their quality of life has declined, but your preferred CPI doesn't cover that.

With a useful CPI, one that measures a fixed basket of goods, consumers can look at how much things went up this year, compare it to their wages, and instantly understand roughly how they did. With your flavor of CPI, they can make no useful projections whatsoever. The data means nothing.

Inflation is not just rising prices. It did not exist in any significant way when we were on solid, unabused commodity money; rather, the overall long term tendency was toward a very slow deflation. (as manufacturers get more efficient, their goods gradually get less and less expensive.) Inflation is purely an artifact of, originally, monetary debasement, and then later outright fiat currency. It is rising prices due to monetary policy. Steady price rises over the long term are not normal economic behavior.

The reason prices go up is not because the goods are harder to make, but because the dollars used to measure them are weaker.

We're now getting into a new era where we may actually start to truly run short of things, primarily oil. This will likely start to cause price increases that aren't caused by monetary disorder. But up until this point, there has never been long-term scarcity of any common good, yet somehow their prices always go up every year. When the money was solid, that didn't happen.
posted by Malor at 4:28 AM on November 1, 2006

Oh, by the way, this is my understanding of the SSTF.

Say you and your spouse are saving for college. Each month, you put $200 in a jar over the fridge. Quietly, your spouse takes the $200 out and spends it, and replaces it with an IOU for $200 + interest. Calling that an asset, at that point, would be pure fantasy. You'll still have to pay the tuition when it comes due, and if you think you have the money to pay for it, you're probably going to make very bad budget decisions.

That is exactly what the government is doing, and that's why we were running a 'surplus' during the Clinton years. We weren't, we were just robbing Social Security.
posted by Malor at 4:30 AM on November 1, 2006

Let's have some definitions:

"Inflation is a sustained rise in the general level of prices in the economy — the rate at which the price level increases."

"The consumer price index measures the average price of consumption, or equivalently the cost of living." The basket is supposed to "represent the consumption basket of a typical urban consumer".

Do you disagree with any of these definitions? The definition of cost of living and the purpose of the basket seem to conflict with most of your most recent comments.

If the typical urban consumer lives on sawdust alone, the price of sawdust is what the CPI will measure.

Their quality of life has declined, but your preferred CPI doesn't cover that.

So what? Inflation is not quality of life. "Cost of living" is an awkward phrase, and it can give the misleading impression that it means the cost of maintaining a fixed standard of living (for some definition of standard of living).

It's important to measure inflation for several reasons. For example, to be able to measure economic growth. If nominal GDP growth is 3% and inflation is also 3%, the economy has in reality not grown at all. All that's happened is the numerical value of prices has changed. Based on this information alone, we can't say anything at all about living standards.

But up until this point, there has never been long-term scarcity of any common good

Eh? Scarcity is the fundamental problem of economics. All economic goods are scarce.
posted by matthewr at 6:02 AM on November 1, 2006

Wikipedia's article on inflation is very good.

What I'm trying to say here is a mild variation on monetarist inflation. I can't sync your argument up with any other I know of. It's simplistic to the point of being ludicrous, and very supportive of a measurement that means nothing.
posted by Malor at 6:37 AM on November 1, 2006

As far as I'm can tell, my "argument" is aligned with the prevailing economic orthodoxy. Whether you're monetarist, Keynesian, Marxist or anything else has little to do with the definitions of inflation (as opposed to the causes) and the consumer price index.

The basket is supposed to "represent the consumption basket of a typical urban consumer".

Do you disagree with this? If not, how do you reconcile this with your view that "the basket does not need frequent updates"? If the basket is supposed to represent a typical consumer's consumption, how could updating it to reflect the typical consumer's consumption do any harm?
posted by matthewr at 6:54 AM on November 1, 2006

It causes harm because it hides what it's supposed to measure -- inflation, and the resulting decline in standards of living.

If beef goes up and consumers switch to chicken because they can't afford beef anymore, your method of measurement will say there's no inflation. If ALL foods go up and consumers, out of desperation, switch primarily to cheaper foods just to get enough calories to stay alive, your measurement method says there's no inflation. If prices continue to go up and they buy shoe leather to chew on to try to assuage the hunger pangs a bit, there's STILL no inflation.

Consider this in bold, 72-point type: A measurement system with a changing yardstick does not measure anything.

The reason they change the basket monthly is NOT to 'better track consumer buying habits', but rather to hide the very strong inflation that is taking root in the system. They do things like excluding gas in months it goes up a lot, because it's an 'unusual event', but then they INCLUDE it in subsequent months when it goes back down.

If you believe this change is to improve accuracy, you are sadly mistaken. The basket needs occasional recalibration, just as all measurement devices do, but the monthly change isn't about that.

Is is, rather, so that the government can more easily lie about inflation, and save billions on cost-of-living adjustments.

Soon, we will have no inflation, because the dog food that senior citizens are forced to buy is cheaper this month.
posted by Malor at 7:27 AM on November 1, 2006

how could updating it to reflect the typical consumer's consumption do any harm?
It causes harm ...

The purpose of the basket is to "represent the consumption basket of a typical urban consumer". Do you disagree with this? How could not updating it improve its representation of the typical consumer's consumption? If the typical consumer's consumption patterns don't change over time, updating it makes no difference to anything. If the patterns do change, updating it obviously means it is more representative of the typical consumer's consumption.

what it's supposed to measure -- inflation, and the resulting decline in standards of living.

Argh! Since we're resorting to bold here: the CPI is not intended to measure standards of living. Never has been, never will be. Consumer price inflation does not even, of itself, imply real wage decreases, let alone falls in living standards.

They do things like excluding gas in months it goes up a lot, because it's an 'unusual event', but then they INCLUDE it in subsequent months when it goes back down.

If they actually excluded goods for purely arbitrary reasons, of course this would be a bad thing. I haven't seen any evidence for this. More importantly, it's a different from constantly updating the basket, which is what we're talking about.

A measurement system with a changing yardstick does not measure anything.

First, it's not measuring what you think it is (see above). Second, the yardstick is not changing. The CPI measures the average nominal price of consumption. Consumption patterns may change, this doesn't mean a "yardstick is changing". See the definitions of CPI and inflation in my previous comment.
posted by matthewr at 7:51 AM on November 1, 2006

The CPI's basket of goods can't change from month to month, or else the CPI is entirely meaningless. It measures the change in buying the same things over time. Econ 101.
posted by oaf at 8:12 AM on November 1, 2006

Just to make sure we're all talking about the same thing here, the chained CPI measure I'm talking about it is known as C-CPI-U by the BLS.
posted by matthewr at 8:20 AM on November 1, 2006

To provide a bit more background, the C-CPI-U is a type of 'superlative' index. These superlative indices are intended to correct the substitution bias inherent in the old-style Laspeyres index, which causes inflation to be overstated. They also avoid the understatement of inflation inherent in Paasche indices (the opposite of Laspeyres). These are all Econ-101 terms, and I'm sure they have Wikipedia articles with formulae.

A simple way of balancing these two biases is to use a Fisher index, which is the geometric mean of the Paasche and Laspeyres indices.

Rather than using Fisher's method, the C-CPI-U is calculated using a Tornqvist formula. This is the geometric mean of the prices in the comparison and reference period, weighted by the geometric means of average expenditure shares in the two periods.

Wikipedia has a good article giving a simple explanation of the difference between Paasche and Laspeyres indices.

Why would inflation be more biased in either direction using a superlative index?
posted by matthewr at 8:41 AM on November 1, 2006

A measurement system with a changing yardstick does not measure anything.
*ahem relativity cough* ;)

Define [the CPI] Away by Lew Rockwell.

If my CPI measures steak one month and chicken the next, what exactly can I conclude about the value of my cash dollars?

If it measured steak both months, I could see whether my steak/dollar rate is changing, but if I instead have TWO equations:

cpi(chicken, ...) = x
cpi(steak, ...) = y

my set of equations is underdetermined and I cannot detect a change in my dollars' buying power relative to either steak OR chicken.

"standard of living" is not measured by CPI, but I'm not interested in some average standard of living. I am interested in the value of my dollar relative to my actual spending habits and it seems like the C-CPI-U does not let me determine this. Even if the old CPI did not measure goods I actually ever purchase, I could at least figure the prices of my goods compared with the CPIs measured goods and get the same data.

How can I do this if different sets of data are measured every time I want to calulate the actual inflation of my cash money?
posted by sonofsamiam at 9:13 AM on November 1, 2006

In my first comment I said that the C-CPI-U ought to be a better measure of inflation than the old CPI measure was.

Let's compare three types of price index:

p = price, q = quantities purchased (hence, weights), 0 = base year, t = current year.

Laspeyres (Old CPI): Σp_tq₀ / Σp₀q₀

Paasche: Σp_tq_t / Σp₀q_t

As you can see, the Laspeyres index uses base-year quantities (i.e. the base-year basket), while the Paasche index uses current-year quantities. The Laspeyres index overstates increases in the cost of living because it assumes consumers don't substitute less expensive goods for more expensive ones, when of course they do.

On the other hand, the Paasche index understates cost-of-living increases because although it accounts for substitution, it doesn't reflect any loss of welfare due to substitution (my previous comments rather bypassed this point).

Incidentally, while the old CPI was a Laspeyres index, the GDP deflator is a Paasche index.

Clearly, a measure that avoids these biases is a good thing.

C-CPI-U (*): √(P_L + P_P)

Where P_L is the value obtained using the Laspeyres index and P_P the value from Paasche.

The C-CPI-U gets rid of the substitution bias, thus providing a more accurate measure of inflation. A proof of the way it does this can, I believe, be found in Diewart (1976).

As per my original point, the change to the C-CPI-U is an improvement over the old Laspeyres CPI.

Malor, if you disagree, please indicate exactly which bit I've got wrong.

*In fact, what I've written is a Fisher index, when the C-CPI-U uses a Tornqvist index. But typing the Tornqvist formula in HTML entities is not a lot of fun and there's no theoretical difference.
posted by matthewr at 10:02 AM on November 1, 2006

it doesn't reflect any loss of welfare due to substitution (my previous comments rather bypassed this point).

In other words, it's a completely useless number. Which is what I have been saying at length.

If you aren't considering 'loss of welfare' (ie, loss of living standards), then there simply is no reason to compute the number at all.

Remember, this number is used to determine cost of living adjustments for Social Security. Many private retirement systems are based off it as well. By changing how it is computed, they have doomed old people to a perpetually-declining standard of living, because the market basket they choose each month is whatever went up the least.
posted by Malor at 11:53 AM on November 1, 2006

Did you only read that sentence, or the rest of the comment as well?

The reason for computing the Paasche index (in this case) is in order to then calculate the C-CPI-U. Which gives a better measure of inflation than the old CPI did because substitution biases are effectively removed. Which was my original point — the new C-CPI-U is more accurate than the old number.
posted by matthewr at 12:31 PM on November 1, 2006

If you use a number that includes substitution bias, even if you're averaging it with a number that doesn't include substitution bias, you still end up with the bias. You're just square rooting the sums of the two numbers, meaning that the profound understatement caused by goods substitution is lessened somewhat, but not eliminated.

The Laspeyres index does not overstate inflation. It tracks the strength of the dollar against a fixed base quantity. That measure is useful. Particularly with food and energy, which are the critical components of inflation (because their consumption, past a certain point, cannot be delayed), there are almost never 'new' products to substitute.

You have a fixation with new goods, but it's just not warranted. Almost all new goods these days are electronic toys, few of which are necessary to life. Tracking the essentials that people MUST buy to live is what matters. The whole idea of inflation 'ex food and energy' is laughable, a game to distract people from the inflation that really hurts them. If cars inflate, you can just delay buying a new one, or buy one with fewer features, or buy a used one. If bread or gas inflate, you suck it up and buy them anyway.

And you keep dodging the fact that when computing the market basket, the politicians will always choose the goods that go up the least. Always. They're not tracking what people really buy, their numbers are just tracking what's cheapest. It's understating inflation by design.

It's trying to use manufacturing efficiencies of new goods to hide their bad monetary policy.
posted by Malor at 12:55 PM on November 1, 2006

Let me add this, just for clarity:

The proper measure of inflation is: how much more does it cost me to live THE SAME WAY this year than it did last year?
posted by Malor at 12:58 PM on November 1, 2006

"the profound understatement caused by goods substitution is lessened somewhat, but not eliminated"

Wrong.

Diewart (1976) proved that Fisher and Tornqvist indices are 'superlative.' They better approximate the true Cost of Living than either Laspeyres or Paasche indices.

"The Laspeyres index does not overstate inflation."

Wrong.

From a Subcommittee of the House of Representatives [link]: "[there was] virtual unanimity [in the subcommittee] that a price index that tracks the cost of purchasing a fixed market basket of goods and services, such as the CPI now does [Laspeyres], represents an upper bound on changes in the true cost of living. I doubt there exists a professor teaching microeconomics who doesn't routinely demonstrate this characteristic of fixed-weight price indexes to his or her classes."

Also, from Gregory Mankiw: "a Laspeyres index tends to overstate the increase in the cost of living"

Later, in an example where the price of oranges rises:
"Because the CPI [not C-CPI-U] is a Laspeyres index, is overstates the impact of orange prices on consumers: by using a fixed basket of goods it ignores consumers' ability to substitute apples for oranges."

Is Mankiw wrong? Does the fact that he's a Professor of Economics at Harvard change your answer?

"the politicians will always choose the goods that go up the least"

Politicians are just as able to decide to exclude some goods with the old CPI as they are with the new one. What does political interference have to do with the Laspeyres vs. C-CPI-U debate? Note the distinction between excluding goods and changing weights.
posted by matthewr at 2:09 PM on November 1, 2006

"the profound understatement caused by goods substitution is lessened somewhat, but not eliminated"

Note that when I'm saying this wrong, what I mean is that although substitution bias is not necessarily entirely eliminated, we cannot say it tends toward understatement as opposed to overstatement. The C-CPI-U is a better estimator of true cost of living changes.
posted by matthewr at 2:24 PM on November 1, 2006

No, it doesn't change my argument or my position in the least.

That argument is, simply, wrong. A bunch of politicians likes a different method of measuring inflation that results in a lower inflation number. Wow, that's a shock. By 1998, the monetary dysfunction in this country was in full swing, and I'm sure they realized that a lower inflation number was to their profound benefit.

Drinking yak milk and eating dog food may be substitutes taken by consumers because they can't afford anything else, but that doesn't mean their standard of living stayed the same, and it doesn't mean that there was no inflation.

You are supporting, by your passionate arguments, starving seniors to death.

At least, with a basket that changes every five years, they have to pick something and stick with it for awhile. It makes it much harder to hide inflation. Being able to change month to month just means the number is an outright lie.

I will not change my position on this. A measurement with a changing yardstick is useless. It's the CONSUMER price index. It's supposed to be something THEY can use to make meaningful decisions. It isn't, it's a lie, and I'm done talking about it.
posted by Malor at 4:13 PM on November 1, 2006

Yes, C-CPI-U will produce a lower figure for inflation. Yes, this will make politicians happier. Does this mean it is a flawed measure? No. The fact that politicians are in favour of it doesn't make it a less accurate estimator.

Economists have shown that C-CPI-U is less affected by substitutional bias than the old CPI — therefore, it is a better measure of inflation. Your response has been rhetoric about seniors and dog food.
posted by matthewr at 4:26 PM on November 1, 2006

Ok, I will say one more thing, a thought experiment.

Say there are 100 goods being tracked. Each month, 10 of the goods go up 10%. 80 of the goods are steady-state, and 10 of the goods decline by 5%.

Each month, AFTER knowing the results for that month, the "economists" choose the basket that consumers "would prefer the most"... say they're choosing 50 as a 'representative sample'. That means they will choose 40 goods that stayed steady and 10 goods that declined. Negative inflation!

Over time, the overall price will increase by 1% a month, but they can claim by using this system that prices are DECREASING by 0.5% a month.

It is, in other words, a lie. They ALREADY do this by excluding gas when it goes up sharply, and then including it again when it drops.
posted by Malor at 4:32 PM on November 1, 2006

And also note.... by choosing their basket carefully, they can pick almost any number in a very broad range. If they're not limited to a basket of 50, they can choose anything from +10 to -5.
posted by Malor at 4:34 PM on November 1, 2006

You're confusing changing the basket with changing weights. This confusion may well have been reinforced by my first few comments in the thread; if that's the case, I'm sorry. The C-CPI-U only involves a changing basket in the sense that weights are constantly updated in the Paasche component of it.

In an ideal world, we wouldn't have to deal with baskets at all — if measurement costs are zero, why not just measure the price of every good, and weight the price by the quantity good. This seems pretty uncontroversial. Since we're in thought-experiment territory, let's ignore baskets and pretend we can measure everything.

Let's also say that there are two goods in this fictional economy, oranges and apples. Due to some exogenous reason, the price of oranges suddenly rises relative to that of apples.

Laspeyres index
The Laspeyres index assumes that consumers buy the same quantities of oranges and apples, despite the fact that the relative price has changed. In reality, rational consumers obviously buy more apples. The Laspeyres index still gives the same weight to the pricier oranges despite the fact that fewer oranges are now bought. It therefore overstates inflation¹.

Paasche index
The Paasche index, by contrast, looks at the current quantities of oranges and apples. It doesn't take into account the fact that consumers' welfare may have been reduced by substitutions. It therefore understates inflation².

The Old CPI
The old CPI was a Laspeyres index. Laspeyres indices overstate inflation. Therefore, the old CPI overstated inflation.

The C-CPI-U
By taking a geometric mean of the Paasche and Laspeyres indices, the overstatement and understatement effects are reduced or eliminated and the approximation of true inflation increases.³

Baskets

Note that changing the basket has nothing to do with any of this. Let me present the ideal-world zero-measurement-cost scenario, and then a more realistic one.

Ideal-world
The basket includes all goods and services consumed. No new types of good or service are introduced between base-year and current-year, so we don't have to worry about the whole new-goods thing. The Laspeyres component of C-CPI-U uses the weights from the base-year, causing overstatement. The Paasche component uses the weights from the current year, causing understatement. The two effects cancel one another out in C-CPI-U.

Real-world
Economists can't afford to measure every price and quantity of every good/service consumed. Therefore they look at the top 100 goods by consumption of the average consumer. Again, the Laspeyres component of C-CPI-U uses the original 100 goods and the weights from the base-year, causing overstatement. Economists now work out what the top 100 goods by consumption are in the current year. They use this basket and the current-year weights to compute the Paasche index, which inherently understates inflation. Again, the two effects cancel one another out in C-CPI-U.

Obviously, economists must impartially work out the new basket. But with the Laspeyres index, they had to impartially work out the old basket. Of course an economist could theoretically choose a basket to prove whatever he wanted. But this problem applies just as much to the old CPI as the new one. The process of choosing a basket ought simply to involve working out what the top 100 goods are at the given point in time for the average consumer. If politicians interfere with this, it makes no difference whether you're using CPI or C-CPI-U. If you don't trust the measurer, changing the details of the measure hardly makes anything worse. The reality is that most real economists do think the BLS does a pretty good job of estimating inflation.

Conclusion
In the real-world as well as an ideal world, the C-CPI-U is more accurate, so is a better measure of inflation. Meanwhile, the world keeps spinning, dogs eat dog food and seniors live long and fulfilling lives.

^1,2 These statements appear almost word-for-word in every single undergraduate microeconomics textbook. I don't think you want to argue with them.
³ Shown in Diewart 1976.
posted by matthewr at 5:38 PM on November 1, 2006

That argument is, simply, wrong.

I've just realised that you are saying that Laspeyres indices do not overstate inflation. Please find a single source that backs up this assertion.

On my side, I have:

a standard Macroeconic teaching text by a Harvard Professor (Mankiw, Macroeconomics). Note that this isn't the work of some renegade, it's very orthodox.

Wikipedia: "The Laspeyres index systematically overstates inflation"

"As economists have long known, a Laspeyres index .. tends to overstate the rise in the cost of living by not allowing substitution to occur." Diewart (1998)

I could spend all day merrily finding more and more examples to support this. I can guarantee that you will not be able to find a single source supporting your assertion that Laspeyres indices do not overstate rises in the cost of living.

posted by matthewr at 5:50 PM on November 1, 2006