3 ms·

I'm not sure if the response to difficulties in measurement is to build a system out of faith and assumptions. We can analyze different policy decisions, eithe

by jljljl 11y ago

I'm not sure if the response to difficulties in measurement is to build a system out of faith and assumptions. We can analyze different policy decisions, either by modeling human behavior (based on experiments which can be controlled and analyzed), or through natural experiments (e.g., What happens when a bunch of states and municipalities alter their minimum wage policies?)

duncan_bayne 11y ago
The problem is that the models don't relate well to reality, and it's too difficult to isolate confounding effects from natural experiments.
cturner 11y ago
> What happens when a bunch of states and > municipalities alter their minimum wage policies What kind of sizes to you need to be statistically significant? How can you be confident that there are not network effects in play that you don't know about. Immigration patterns, organised crime networks, political corruption, the subtleties of quality in supply chain companies, measurement variation, the culture of particular trade unions. What if someone else is running other economic experiments in parallel to you? What if there's network effects created by the contrast between the areas you're changing? In municipality A you raise wages and not in muni B. Labour will feel pull to flow from one to the other while you're running the experiment. You'll never come close to modeling all this. Worse yet - these are fundamentally political issues that relate to practice of power and allocation of resources. Everyone has an agenda in how they play out, and this further muddies the water of the experiment and the analysis. > I'm not sure if the response to difficulties in > measurement is to build a system out of faith and > assumptions. I have these concern also. As I look at it - everyone is operating on faith and assumptions. The Austrian position is strange in that it's honest about it.
- jljljl 11y ago
  I disagree with this argument. You are posing two extremes: One where everything can be known with certainty, and one where everything is purely based on faith, and dismissing the potential of there being something in the middle. You are right, economics can't model every potential variable that influences outcomes. Large scale studies of minimum wage admit that policy and economic differences between local municipalities can cause the effect of a minimum wage change to vary. But you can model for most major differences, such as "Did adjacent municipalities adjust their wages at the same time? Did this have an impact on employment". Ultimately you can get much closer to an understanding of how minimum wage impacts an economy than you would from a purely faith based approach, and you can land at a recommendation like "In general small increases in the minimum wage do not raise prices or hurt employment, given the following conditions". That's much more valuable than the Austrian position that purely bases itself on faith.
  - cturner 11y ago
    "You are posing two extremes: One where everything can be known with certainty, and one where everything is purely based on faith, and dismissing the potential of there being something in the middle." That's not what I'm posing. My claim is that both positions involve faith, but the Austrians are conservative in their assumptions and systematic in how they build from it. To me that is more trustworthy than non-systematic and statistically insignificant approaches. Your minimum wage paragraph doesn't bring me closer. I think we've reached agree-to-disagree. Fortunately, this is only the internet. What we say doesn't matter, my opinions are of no consequence.

by jljljl 11y ago

duncan_bayne 11y ago

The problem is that the models don't relate well to reality, and it's too difficult to isolate confounding effects from natural experiments.

cturner 11y ago

> What happens when a bunch of states and > municipalities alter their minimum wage policies What kind of sizes to you need to be statistically significant? How can you be confident that there are not network effects in play that you don't know about. Immigration patterns, organised crime networks, political corruption, the subtleties of quality in supply chain companies, measurement variation, the culture of particular trade unions. What if someone else is running other economic experiments in parallel to you? What if there's network effects created by the contrast between the areas you're changing? In municipality A you raise wages and not in muni B. Labour will feel pull to flow from one to the other while you're running the experiment. You'll never come close to modeling all this. Worse yet - these are fundamentally political issues that relate to practice of power and allocation of resources. Everyone has an agenda in how they play out, and this further muddies the water of the experiment and the analysis. > I'm not sure if the response to difficulties in > measurement is to build a system out of faith and > assumptions. I have these concern also. As I look at it - everyone is operating on faith and assumptions. The Austrian position is strange in that it's honest about it.

jljljl 11y ago
I disagree with this argument. You are posing two extremes: One where everything can be known with certainty, and one where everything is purely based on faith, and dismissing the potential of there being something in the middle. You are right, economics can't model every potential variable that influences outcomes. Large scale studies of minimum wage admit that policy and economic differences between local municipalities can cause the effect of a minimum wage change to vary. But you can model for most major differences, such as "Did adjacent municipalities adjust their wages at the same time? Did this have an impact on employment". Ultimately you can get much closer to an understanding of how minimum wage impacts an economy than you would from a purely faith based approach, and you can land at a recommendation like "In general small increases in the minimum wage do not raise prices or hurt employment, given the following conditions". That's much more valuable than the Austrian position that purely bases itself on faith.
- cturner 11y ago
  "You are posing two extremes: One where everything can be known with certainty, and one where everything is purely based on faith, and dismissing the potential of there being something in the middle." That's not what I'm posing. My claim is that both positions involve faith, but the Austrians are conservative in their assumptions and systematic in how they build from it. To me that is more trustworthy than non-systematic and statistically insignificant approaches. Your minimum wage paragraph doesn't bring me closer. I think we've reached agree-to-disagree. Fortunately, this is only the internet. What we say doesn't matter, my opinions are of no consequence.