Are Those Inflation Numbers Right?

I originally wrote this as a reply to a comment on Facebook. The person had two major claims.

  1. Our government redefined the definition of inflation twice in the 90’s to make sure inflation was always true
  2. If we used the original definition of inflation then we would have had 5% inflation or higher for the past 10 years

I’d be surprised if US inflation has actually been running at 5% or higher on average for the last 10 years if that’s what you are saying. The problem gets worse if you go extend that all the way back to the 1990s or use a higher rate than 5%.


1. If I look at nominal GDP from Q4 2009 to Q4 2019, I’m seeing US nominal GDP of 14.628T in Q4 2009 and 21.729T in Q4 2019. If I assume 5% geometric average of inflation, that implies an ~ avg 1% decline in real GDP annually and a 9% fall in real GDP over that 10 years in total. I strongly suspect that’s not true.


2. The Billion Prices Project typically followed CPI pretty closely when I followed it. I couldn’t find a recent update to it, but at least when I used to follow this project it tracked official CPI well. I’m not sure how closely they followed the BLS methodology and just had more timely and larger datasets. I don’t know of anyone that has built a bottom-up alternative CPI that has held up to any scrutiny. If somebody has done this, I’d be very interested to know what they did.


3. According to the official numbers, I’m seeing a cumulative increase in prices of 13.6% in Eurozone, and 19% in the last 10 years (A 5% inflation rate implies a 62.8% change in prices over 10 years). EUR/USD average exchange rate in 2009 was 1.39, 1.33 in 2010…1.12 in 2019. While there are many events and short term speculation that affect exchange rates, a higher US inflation rate would tend to lead towards deprecation of USD versus EUR which we’re not seeing. The Yen has appreciated against the USD in the last 10 years (in 2009 Dollar bought 93.68 Yen,in 2019 the exchange rate was 109 but the reported inflation rate in Japan is something like 0.5% annually for the last 10 years) I suppose other countries could have similarly flawed methodologies as well, but I know even less about how inflation is calculated in Japan or Eurozone.

Bridging the Gap Between Old and New

Sometimes no good deed goes unpunished, but sometimes doing the favor is its own reward. A friend of mine has a very old computer with data that he wanted to retrieve. This computer is old enough where USB did not yet exist and it has have no Ethernet connectivity. We need a bridge from the distance past (at least in computer terms) to the present.

He tried to move files with floppies to a somewhat less ancient computer with a floppy drive and USB. This process worked for a while, but then the 20+ year old floppy drive or controller failed. Efforts to make combinations of different hard drives and controllers work failed. With that, the files seemed to be stuck in a late 1980s Windows 3.1 machine. The two hard drives in the machine were pretty weird by today’s standards. One requires an ISA based controller card and the other was a hard drive integrated on to a giant ISA card.

In a few cardboard boxes lie their ticket out of the 1987 model PC. My first glimmer of hope was a 25 pin to 9 pin null modem cable. I connected the 25 pin end to the Windows 3.1 computer and the 9 pin end to a Windows 98 machine. Then, I fired up HyperTerminal on 98 and Terminal on Windows 3.1. No data seems to be moving at all. Seemed like a hardware problem since I could not even transmit text.

I figured there must be a way. Maybe I needed a 9 pin to 9 pin cable. I could easily find this on Amazon for a one time use. However, maybe I don’t need to buy anything. I have a pile of junk in cardboard boxes after all. I look around and I think I have it. I found a 25 pin to 25 pin adaptor. I put that on the end. Then, I find a 9 pin to 25 pin. So now I have a mess of cables — 9 pin to 25 pin to 25 pin to 9 pin.

Old Computers connected via a weird null modem cable

So, I connect these cables together to see if I can even a bit of text across. Now, we’re in business. I am able to transmit binaries with XModem from Windows 3.1 Terminal to Hyperterminal on Windows 98.

19200 bps serial terminal connection

Later on I was able to do even better than terminal. While Terminal worked, it was kind of annoying.

  • Can only transmit one file and then you have to manually select the next file. I zipped each directory at a time to make it go faster.
  • Does not transmit file name so you need to type the name of the file on the receiving computer
  • Maximum transfer rate of 19200 bps from the Windows 3.1 version of terminal

I remembered a very old piece of software called Laplink. This software was designed to do exactly what I needed to do. This program doesn’t even run in Windows. I ran it in DOS on the Windows 3.1 and in a Windows 98 DOS box. This resulted in a vast improvement in speed (about 8x) and the ability to copy entire directory trees.