3 ms·

I don't understand the details on how inflation is measured here, but I expected the US Dollar curve to be a constant 1, after the normalization. However, it se

I don't understand the details on how inflation is measured here, but I expected the US Dollar curve to be a constant 1, after the normalization. However, it seems to go down. What does that mean? That the dollar loses value faster than "money is printed"?

kqr 4y ago
You're thinking of the inflation adjustment the wrong way around. A dollar earned in 1980, if, saved until today, buys less than it would have then.
- leethargo 4y ago
  Yes, your second sentence makes sense to me. I guess I imagined the inflation-adjusted value of X to be defined at time t to be something like: value(X, t) / value(USD, t). So, when I substitute USD for X, I get constant 1.
  - aidenn0 4y ago
    That was almost right; the correct denominator is value(USD, SOME_FIXED_TIME). Nominal (i.e. non-inflation-adjusted) use value(USD,t) as the denominator implicitly when priced in dollars (hence the term "dollar denominated")
compumike 4y ago
Here's what the site does behind the scenes: real_price($ASSET, t) = nominal_price($ASSET, t) * (price_level($NOW) / price_level(t)) Where price_level(t) is the CPI-U series (with interpolation). If t = $NOW, at the right side of the chart, then the fraction goes to 1, so that real_price($ASSET, $NOW) = nominal_price($ASSET, $NOW) (Though many comments are requesting alternative normalization schemes!) For cash (USDOLLAR), nominal_price(USDOLLAR, t) = 1 for all t -- the nominal price of a dollar bill is always 1. So https://totalrealreturns.com/s/USDOLLAR https://totalrealreturns.com/s/USDOLLAR is plotting a curve that looks like 1/price_level(t). (Actually it's 1*price_level(now)/price_level(t), because of the normalization above.) And you can download price_level(t) from https://download.bls.gov/pub/time.series/cu/ https://download.bls.gov/pub/time.series/cu/.

3 ms·

I don't understand the details on how inflation is measured here, but I expected the US Dollar curve to be a constant 1, after the normalization. However, it se

by leethargo 4y ago

kqr 4y ago
You're thinking of the inflation adjustment the wrong way around. A dollar earned in 1980, if, saved until today, buys less than it would have then.
- leethargo 4y ago
  Yes, your second sentence makes sense to me. I guess I imagined the inflation-adjusted value of X to be defined at time t to be something like: value(X, t) / value(USD, t). So, when I substitute USD for X, I get constant 1.
  - aidenn0 4y ago
    That was almost right; the correct denominator is value(USD, SOME_FIXED_TIME). Nominal (i.e. non-inflation-adjusted) use value(USD,t) as the denominator implicitly when priced in dollars (hence the term "dollar denominated")
compumike 4y ago
Here's what the site does behind the scenes: real_price($ASSET, t) = nominal_price($ASSET, t) * (price_level($NOW) / price_level(t)) Where price_level(t) is the CPI-U series (with interpolation). If t = $NOW, at the right side of the chart, then the fraction goes to 1, so that real_price($ASSET, $NOW) = nominal_price($ASSET, $NOW) (Though many comments are requesting alternative normalization schemes!) For cash (USDOLLAR), nominal_price(USDOLLAR, t) = 1 for all t -- the nominal price of a dollar bill is always 1. So https://totalrealreturns.com/s/USDOLLAR https://totalrealreturns.com/s/USDOLLAR is plotting a curve that looks like 1/price_level(t). (Actually it's 1*price_level(now)/price_level(t), because of the normalization above.) And you can download price_level(t) from https://download.bls.gov/pub/time.series/cu/ https://download.bls.gov/pub/time.series/cu/.