2 ms·

Why don't they just measure it month-to-month. Are they trying to make the whole thing more dramatic by reading the yearly change? They make it seem like the Fe

by seydor 4y ago

Why don't they just measure it month-to-month. Are they trying to make the whole thing more dramatic by reading the yearly change? They make it seem like the Fed can only do division by 12 months to make up its mind.

MattGaiser 4y ago
Probably to reduce volatility in the measurement and because their main policy tool of interest rate hikes takes a while to take effect.
- seydor 4y ago
  i would expect them to have a model better than moving average
  - imtringued 4y ago
    They have the CPI. They can calculate whatever they need from that.
  - QuarterReptile 4y ago
    What are the assumptions of the model? Who would make the model? Who would trust the model? How long until the unchanging underlying assumptions blow up under the added pressure of fiscal policy for the most important currency in the world leaning on them? Inflation is one of the most important metrics for healthy governance, right up there with 'what wars are we fighting?' You really don't want it to be any more game-able than it has to be.
somenameforme 4y ago
You can answer this with reductio ad absurdum. Imagine they could do it by the second. Should they? That actually may not be self evident to many. The reason you would not want to is that as you start minimizing time frames, you start measuring more noise (variance) than signal. Inflation is measured by looking at how the price of a standardized "basket of goods" changes. And this can go up and down quite quickly over short periods of time, for reasons that have nothing to do with generalized inflation. A yearly rate is long enough to make sure you're getting mostly signal, not so long that it loses meaning, and you also get to automatically normalize for various seasonal trends without relying on magic numbers.
- saurik 4y ago
  The speed of my car can also go up and down quite quickly and yet it would be crazy to try to drive it using a moving average of how fast I was going over the prior 60 seconds. "I dunno. I realize I slammed on the brakes ten seconds ago, and my speed over the past minute has certainly gone down a bit? But I'm still going really fast... I should keep holding this brake down." Meanwhile, of course, you are stopped on the exit ramp of the highway. I'm sure you'll figure it out, though, in another 50 seconds.
hartator 4y ago
Month-to-month might have some noise that makes it hard to read. We are still at 5.0% year-to-year inflation and 0.1% month-to-month inflation. This is still bad.
wcoenen 4y ago
> Why don't they just measure it month-to-month Because there are seasonal variations, e.g. gasoline tends to be a bit more expensive in summer in the USA https://www.eia.gov/energyexplained/gasoline/price-fluctuations.php https://www.eia.gov/energyexplained/gasoline/price-fluctuati...

by seydor 4y ago

MattGaiser 4y ago

Probably to reduce volatility in the measurement and because their main policy tool of interest rate hikes takes a while to take effect.

seydor 4y ago
i would expect them to have a model better than moving average
- imtringued 4y ago
  They have the CPI. They can calculate whatever they need from that.
- QuarterReptile 4y ago
  What are the assumptions of the model? Who would make the model? Who would trust the model? How long until the unchanging underlying assumptions blow up under the added pressure of fiscal policy for the most important currency in the world leaning on them? Inflation is one of the most important metrics for healthy governance, right up there with 'what wars are we fighting?' You really don't want it to be any more game-able than it has to be.

somenameforme 4y ago

You can answer this with reductio ad absurdum. Imagine they could do it by the second. Should they? That actually may not be self evident to many. The reason you would not want to is that as you start minimizing time frames, you start measuring more noise (variance) than signal. Inflation is measured by looking at how the price of a standardized "basket of goods" changes. And this can go up and down quite quickly over short periods of time, for reasons that have nothing to do with generalized inflation. A yearly rate is long enough to make sure you're getting mostly signal, not so long that it loses meaning, and you also get to automatically normalize for various seasonal trends without relying on magic numbers.

saurik 4y ago
The speed of my car can also go up and down quite quickly and yet it would be crazy to try to drive it using a moving average of how fast I was going over the prior 60 seconds. "I dunno. I realize I slammed on the brakes ten seconds ago, and my speed over the past minute has certainly gone down a bit? But I'm still going really fast... I should keep holding this brake down." Meanwhile, of course, you are stopped on the exit ramp of the highway. I'm sure you'll figure it out, though, in another 50 seconds.

hartator 4y ago

Month-to-month might have some noise that makes it hard to read. We are still at 5.0% year-to-year inflation and 0.1% month-to-month inflation. This is still bad.

wcoenen 4y ago

> Why don't they just measure it month-to-month Because there are seasonal variations, e.g. gasoline tends to be a bit more expensive in summer in the USA https://www.eia.gov/energyexplained/gasoline/price-fluctuations.php https://www.eia.gov/energyexplained/gasoline/price-fluctuati...