3 ms·
> If I have a model that doesn't work in smaller AND larger (whatever that means) contexts, it's a bad model. That seems like a very strange remark. First, mac
by quonn 11y ago
> If I have a model that doesn't work in smaller AND larger (whatever that means) contexts, it's a bad model.
That seems like a very strange remark. First, macroeconomics va. microeconomics is not about large and small, but about the whole vs. parts of it. And that indeed makes a huge difference.
Second, we use different models all the time, pretty much everywhere. So in engineering small local systems you would use classical mechanics. In physics, not. There are similar examples in computer science.
- rbehrends 11y ago> That seems like a very strange remark. First, macroeconomics va. microeconomics is not about large and small, but about the whole vs. parts of it. And that indeed makes a huge difference. To expand on that: A sovereign nation state has the power to tax, to legislate, and to regulate trade with other nations; outside the Eurozone, to issue its own currency. Households and businesses can do none of the above, but are subject to the decisions that their governments make on their behalf.
- yummyfajitas 11y agoIn physics, we view it as a bad thing and a deep flaw when the macro model (GR) and the micro model (QM) don't agree. Large amounts of effort in physics are devoted to deriving macro models from micro models - this is the entire purpose of statistical mechanics, for example.
- parasubvert 11y agoSame goes in economics and a large degree to why Keynes was thrown out by freshwater schools in the 70s. Only to be resurrected lately because the new unified models were empirically shown to have little macro predictive power, ie. They were wrong.
- yummyfajitas 11y agoBut the Keynesian macro models similarly have little predictive power. They've failed most of the modern tests which actually distinguish them from monetarism. (There have been several "tests" of Keynesian thinking where Keynesian predictions agree with monetarist predictions, but such tests can't distinguish between the two theories.) Most of the modern shift in economic thinking is based more on fashion than on new information. See, e.g., Scott Sumner on the topic: http://econlog.econlib.org/archives/2014/05/when_ideologies.html http://econlog.econlib.org/archives/2014/05/when_ideologies.... http://econlog.econlib.org/archives/2015/03/ways_of_thinkin.html http://econlog.econlib.org/archives/2015/03/ways_of_thinkin....
- parasubvert 11y agoGood old IS-LM has done fairly well the past decade. I'm not sure what you're saying: that both Keynesianism and monetarism suck, or that there is a debate between two credible schools of thought and an unclear winner?
- dmichulke 11y agoMeasuring a big system might be useful if you are interested in this information. However, steering the big system by looking at the big statistics while ignoring its much more deterministic effects on parts is both, an oversimplification and common practice. Just think of a tax where they talk about future income streams for government but totally deny the costs (especially for complex taxations or ambiguous legislation). Or changes to the way GDP is calculated. Or "economical stimulation" via war.