As for models; their output is exactly as accurate as the hard data and variables plugged into them. Variables are the easiest to massage since most cannot dispute their accuracy; they are educated guesses at best. Hard data can be vetted by some but taken as accurate buy most. Figures never lie, but liars can figure.
There is definitely a bit of that possible. To clarify the issue a bit. All models are just a variables and equations. The issue is how to estimate the value for a variable from data. To take an example, suppose one of your variables is 'income'. How can we get the hard data for income? Should we just look at wages or should we include benefits like healthcare etc. Both assumptions could be valid and sometimes which assumption is selected can impact the outcomes of a model.
A good model states their assumptions clearly and also how sensitive are the results to certain estimates. Moreover, a good model also allows for multiple results - that is the outcome is not pre-assumed (sometimes one can see the outcome of a model without even having to look at the hard data because there's only one mechanism in the model.)
Once upon a time in my dim past I was schooled on a Monte Carlo model system and don't know anything really about current versions, but what I do know something about is human behavior where agendas and beliefs are concerned. Doesn't seem to matter the agenda but more often than not, folks ultimately seek out confirmational bias and invent "facts" to support them or their personal versions of truth. A tweak here and a tweak there and eureka; there's the exact answer we have been looking for! How many financial theories have there been, each the soup d'jour? How many tweaks have there been to the CPI? Sorry, models are fine, but modelers---not so much and too often so by my reckoning anyway.
I don't fully understand how the model predicts the impact of tariffs of different goods on inflation. I understand the assumptions in the model, but not how the model predicts the impact. Can you explain how this is done? (I am imagining this the econometrics part, which might be too abstract, even so if you could describe at top level.) Thanks!
Good question. The authors - CQRS - first write down the model with all assumptions and functional assumptions. For example, there's a function that takes the amount of labor and the amount of intermediate goods, and gives out total output of a new good. This mathematical function have parameters that tell us how much of the new good will be produced from a given amount of labor and a given amount of intermediate goods.
To figure out the value of these parameters, CQRS 'calibrate' the model. Some of the parameter values come from other research (for example, how much do we discount the future can be taken from other studies). The remaining parameters are then usually set in such a way as to match real-world data. Basically, economists have certain real world-data (output, inflation, prices etc). The CQRS had model-equivalent values that match real world-data. So CQRS test different parameters value until their model gives numbers that are very close to real-world data.
Now CQRS have a model with parameter values. This model can now generate data that looks like the US economy. To test the impact of tariffs, CQRS can change the tariff parameter and see what the model outputs are. In this way, CQRS can compare what will happen in the economy at different tariff rates.
Let me know if this was clear/any follow up questions!
Tariffs can, do change relative prices. They cannot affect the average price level except by affecting the way the central bank makes monetary policy. Now one of the things that a reasonable central bank will do is to try to adjust the price level so as to permit changes in relative prices (from tariffs or anything else) not to cause markets in whihc there are downwardly sticky prices to fail to clear -- for unemployed resources to appear. While that adjustment is going on, while the price level is adjusting is inflation.
Tariffs + wise central bank monetary policy = (temporary) inflation + (hopefully) no unemployment
Tariffs + unwise central bank monetary policy = no inflation + unemployment
I write mainly about US monetary policy, US fiscal policy, trade/industrial policy, and climate change policy.
I have my opinions about which US political party is by far the least bad and they are not hard to figure out, but I try to keep my analysis of the issues non-partisan.
Keynes said, “Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back.”
Simple answer; do everything possible to avoid purchasing tariffed goods. When a tariffed good retailer attempts to pass on an obvious tariff initiated price increase refusal to purchase is the very best answer. By doing so the consumer is doing their bit to send the message to the exporting country America means business.
As for models; their output is exactly as accurate as the hard data and variables plugged into them. Variables are the easiest to massage since most cannot dispute their accuracy; they are educated guesses at best. Hard data can be vetted by some but taken as accurate buy most. Figures never lie, but liars can figure.
There is definitely a bit of that possible. To clarify the issue a bit. All models are just a variables and equations. The issue is how to estimate the value for a variable from data. To take an example, suppose one of your variables is 'income'. How can we get the hard data for income? Should we just look at wages or should we include benefits like healthcare etc. Both assumptions could be valid and sometimes which assumption is selected can impact the outcomes of a model.
A good model states their assumptions clearly and also how sensitive are the results to certain estimates. Moreover, a good model also allows for multiple results - that is the outcome is not pre-assumed (sometimes one can see the outcome of a model without even having to look at the hard data because there's only one mechanism in the model.)
Once upon a time in my dim past I was schooled on a Monte Carlo model system and don't know anything really about current versions, but what I do know something about is human behavior where agendas and beliefs are concerned. Doesn't seem to matter the agenda but more often than not, folks ultimately seek out confirmational bias and invent "facts" to support them or their personal versions of truth. A tweak here and a tweak there and eureka; there's the exact answer we have been looking for! How many financial theories have there been, each the soup d'jour? How many tweaks have there been to the CPI? Sorry, models are fine, but modelers---not so much and too often so by my reckoning anyway.
Hey, thanks for this, very interesting!
I don't fully understand how the model predicts the impact of tariffs of different goods on inflation. I understand the assumptions in the model, but not how the model predicts the impact. Can you explain how this is done? (I am imagining this the econometrics part, which might be too abstract, even so if you could describe at top level.) Thanks!
Good question. The authors - CQRS - first write down the model with all assumptions and functional assumptions. For example, there's a function that takes the amount of labor and the amount of intermediate goods, and gives out total output of a new good. This mathematical function have parameters that tell us how much of the new good will be produced from a given amount of labor and a given amount of intermediate goods.
To figure out the value of these parameters, CQRS 'calibrate' the model. Some of the parameter values come from other research (for example, how much do we discount the future can be taken from other studies). The remaining parameters are then usually set in such a way as to match real-world data. Basically, economists have certain real world-data (output, inflation, prices etc). The CQRS had model-equivalent values that match real world-data. So CQRS test different parameters value until their model gives numbers that are very close to real-world data.
Now CQRS have a model with parameter values. This model can now generate data that looks like the US economy. To test the impact of tariffs, CQRS can change the tariff parameter and see what the model outputs are. In this way, CQRS can compare what will happen in the economy at different tariff rates.
Let me know if this was clear/any follow up questions!
It does, thanks.
Love these primers, please keep doing them!
Tariffs can, do change relative prices. They cannot affect the average price level except by affecting the way the central bank makes monetary policy. Now one of the things that a reasonable central bank will do is to try to adjust the price level so as to permit changes in relative prices (from tariffs or anything else) not to cause markets in whihc there are downwardly sticky prices to fail to clear -- for unemployed resources to appear. While that adjustment is going on, while the price level is adjusting is inflation.
Tariffs + wise central bank monetary policy = (temporary) inflation + (hopefully) no unemployment
Tariffs + unwise central bank monetary policy = no inflation + unemployment
Tariffs are just hidden taxes passed down the supply chain. Consumers always end up footing the bill—especially in import-heavy economies
Hi
As you are not already a subscriber, may I invite you to subscribe (for free) to my substack, "Radical Centrist?"
https://thomaslhutcheson.substack.com/
I write mainly about US monetary policy, US fiscal policy, trade/industrial policy, and climate change policy.
I have my opinions about which US political party is by far the least bad and they are not hard to figure out, but I try to keep my analysis of the issues non-partisan.
Keynes said, “Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back.”
I want to be that scribbler.
Thanks,
Simple answer; do everything possible to avoid purchasing tariffed goods. When a tariffed good retailer attempts to pass on an obvious tariff initiated price increase refusal to purchase is the very best answer. By doing so the consumer is doing their bit to send the message to the exporting country America means business.