BPP (MIT) U.S. price index holes and potential or probable issues, in no particular order ----------------------------------------------------------------------------------------- BPP only surveys goods, not services. Services, depending on where you look, are about 70% of the economy. I have very substantial trouble giving much credence to an index that only measures about 30% of prices, and tries to proxy the rest. BPP only surveys online prices, which is the most price competitive market place around since it's so easy to do price comparisons. That hints at possible undervaluing of price increases. BPP can't survey anything in the medical field, where the highest inflation frequently exists. Which housing index do they use? There's a huge difference between Case Shiller and Dataquick for example. Do they always see the whole transaction (in a swap for example)? How do they pick up health insurance paid by the company? That's not exactly always public knowledge available on the internet, let alone how accounting for coverages changes isn't simple too (yes, the BLS is far from perfect too) BPP does not and can not survey many high end items where a login or similar must occur. Inflation is always higher on high end and very high end items, aka a potential bias... and a more closed methodology. Is collecting 500k prices per day really all that different than 100k, in the sense that does anyone really believe that with all the variables in measuring inflation that it matters much to it's accuracy? Who owns BPP and who and what are their connections and possible vested interests? I haven't looked as deeply as I'd like and the BPP methodology is over 90% opaque, but what about bias for recording specials vs. everyday prices or software bugs that pick up data from the wrong column on a dynamic page (and yes, I'm sure they do their best and I'm not accusing them), or how about postage and handling "creativity" where higher prices are hidden in P&H. And then there's a biggie - how the big and medium sized sites can and do tailor make prices depending on who you are in their database? A truly almost wild variable. We know nothing about the weights. Is housing a 41% equivalent? How do they weight alcoholic vs. non alcoholic beverages? Where are the precise numbers published (they're not, just charts)? Where's the BPP history prior to 2009? CPI goes back to 1913, and extensions back into the 1700s. We have few historical comparisons for judgments of long term accuracy, especially during periods of high inflation like the late 1970s. Yes, the BPP can be a bit helpful since it comes out a month earlier but it doesn't make any difference to regular folk. It's primarily for weird inflation freaks like me, or corporate models. BLS related There's nothing wrong with using a survey method that's based in the 50s (I'm older than that anyhow ;-). The huge majority of BLS surveys since the 50s are eyeballs on, and they can see actual ingredients lists to see any "hidden" inflation or any other games that are hidden to a computer only based approach. Yes, BLS misses cell phones but not only is that temporary, I also believe that only misses about 10% overall, considering all the other ways surveys are done. How do we know that BPP survey methodology is even as good as the BLS? The BLS at least does tell us some thing about its methodology. One huge example is BPP's possible inability to take quality "adjustments" into account. Yes, size or weight is comparable and adjustable when it changes but what about switches to or from cane sugar, or to or from steel to pot metal, or the actual ingredient balance in concrete. They're not there in person like BLS reps and their eyeballs are. We know that there are issues with BLS stuff like hedonics or medical CPI being 7% index weight while medical is 17% of GDP, but we know even less about BPP methodology than we do about the BLS. And again, I'm not necessarily accusing nor do I want to impugn BPP. I'm simply pointing out a few things off the top of my head where the BPP index has large holes and potential issues. ----------------------------------------------- BPP is a terrible inflation gauge that should be expected to dramatically under-report inflation. it uses non weighted prices, so a drop in the price of gum from $1 to 99c offsets the rise in price of a car from $50,000 to $50,500. it looks only at online prices and does not include services or most food. it ignores insurance, education, healthcare, and shelter. it massively overweights tech, the most deflationary sector of the economy. BPP mistakes using a lot of inputs for using meaningful inputs. if you live in a hard drive and eat iphones, it may be great, but it bears so little resemblance to an actual consumer basket as to be nearly meaningless and there is every reason to suspect it would dramatically under report inflation just by virtue of tech weighting and leaving out healthcare. hitssquad | July 21, 2014 at 4:00 pm @morganovich > it uses non weighted prices [...] it ignores [...] healthcare “it calculates health-care costs by weighting changes in the price of wages, energy and drugs.” “the BPP weights its index based on the BLS basket” > it massively overweights tech, the most deflationary sector of the economy. …Then we should expect it to drop in half every couple of years. It doesn’t seem to be doing that. It might, in fact, be exaggerating real inflation. Reply ? morganovich | July 21, 2014 at 4:21 pm 1. so wages, energy, and a small amount of drug data that is mostly unavailable (it does not have access to prescription prices) seems like a good proxy for health care costs to you? it’s a completely made up number. further, it is assigned 7% weighting vs 18% in the actual us economy. 2. and where is shelter in this mix? 3. further, it ONLY looks at online prices, which are the most price competitive. it sees the price of bulk coffee, but not the price of the coffee you buy on the way to work. it sees a can of tuna, but not a tuna sandwich near your office. thus, it misses the part of the economy that has the most inflation because it has fewer substitutes and higher cost of switching. the same is true of the service sector. there is simply no way BPP could be exaggerating inflation. it likely runs low by 2-3 points. it has been tuned to look like cpi but is is neither capable of being nor designed to be a real consumer price measure. with housing up 10 and rent up 5, how does one average that to get to 2.9% you can’t. they use owner equivalent rent to mask this and pretend that because some people do not move, the price is lower. but that is not a market price nor a price level. would you say that because only 10% of people buy a new car a price hike in cars from $20k to 22k is only 2% and try to pass that off as a price level? price level is what you face if you go buy somehting. the fact that you might not buy it does not change that. the market price is the market price. CPI was shifted to be a COLA index, not a measure of price level. it’s not an inflation measure. if it were, shelter would read more like +7% which, by itself, would add a solid point to CPI. weight healthcare properly and you likely pick up another 50bp. include that fact that switching from rib eye to flank steak may reduce price, it also ought to have an offsetting quality adjustment, and now CPI starts to look like it might be only around 50% of the actual price level change. hitssquad | July 21, 2014 at 5:01 pm @morganovich > it also ought to have an offsetting quality adjustment IIRC, the fact that it doesn’t is the reason Thomas Sowell says it exaggerates real inflation. 2. Artifact of hyper-regulation/crony-capitalism in the real-estate industry. Liberalize real estate and homes might be sold on Amazon, where their prices could be captured by the BPP. morganovich | July 22, 2014 at 3:07 pm “IIRC, the fact that it doesn’t is the reason Thomas Sowell says it exaggerates real inflation.” i think you are misunderstanding my point. let me see if an example will help. if, for example, you like to eat ribeye steak and its price rises, the BLS will assume that you shift your consumption pattern toward, say, flank steak. this is called geometric weighting. there are several problems with doing this and all of them cause a systematic understatement of inflation. first off, the assumption is that price shifts are supply driven. that price rise is taken as exogenous. but there are multiple reasons price could rise. if the hike is because cattle herd are smaller, then yes, perhaps such weighting is realistic, but, if the price hike is caused by a change in demand, then it has the wrong sign. sometimes, prices rise because demand goes up and fall because demand goes down. we see this with meat/pasta depending on the current hipness of atkins/paleo. we saw it with pinot noir and merlot after the movie sideways. we see it every year during bbq season. the blanket geometric assumptions are getting a lot of their adjustments 180 degrees wrong. but that is not the only issue. what i was getting at before is this: substituting cheaper goods should have an offsetting quality adjustment. this makes the whole geometric weighting idea absurd. people value rib eye more than flank steak. how do we know? the market tells us. RE is twice the price of FS. so, if we swap out RE for FS, we are left with a basket of goods a consumer values less. this is the very definition of a quality decline. yet this is not taken into account. thus, the system often weights in the wrong direction (always toward cheap and toward lower inflation) and often gets that wrong. but even when it gets it right, it fails to account for the fact that the old basket was preferred to the new one. both tendencies will ALWAYS cause an understatement of inflation. one can try to argue that positive quality adjustments are not sufficient, but there is no empirical way to do so, it’s pure supposition, and, around things like houses year to year and food, there is little difference and among things like clothing and furniture, quality (unless you buy very high end) has plummeted. one could argue the same thing about food that is now truck ripened, lower nutrition, tastes worse, and is full of hormones, steroids, and antibiotics unless you pay a fair bit more. beef in 1940 is like organic beef today. so, it’s not at all clear to me that the case for quality adjustments being too low is at all sound. some things have gotten better to be sure (cars, computers, phones, etc) but this is far from universally true. in many cases cheap and nearly disposable has replaced high quality stuff that lasts. i sure see it in clothing. 2. i’m not sure how your point about home sales being regulated is germane regardless of whether or not it is true. it’s a big cost. it’s not in the BPP. their healthcare estimate and weighting is a pure fiction. BPP pretty much HAS to read low. it measures all the most price competitive products and cannot read the most inflationary. hitssquad | July 23, 2014 at 8:08 am @morganovich > people value rib eye more than flank steak. how do we know? the market tells us. RE is twice the price of FS. That works in the short run, but not the long run. Prices ultimately approach the cost of production, regardless of consumer valuation. > lower nutrition, tastes worse, and is full of hormones, steroids, and antibiotics [...] beef in 1940 is like organic beef today. Organic beef isn’t higher quality, and today’s commercial beef is higher quality than anything from 1940. ( https://www.youtube.com/user/meatnewsnetwork/videos ) Organic beef simply costs more to produce, which is why it costs more in the store. There’s some Fear, Uncertainty, and Doubt (especially when the buyer is buying for more than just herself, and therefore feels guilty about possibly trading quality for money), so organic captures some of the market, despite not providing anything more of real value to consumers. Is the CPI evil for not falling prey to FUD marketing? -- from morganovich