Three thoughts on your variant of the proposal:
I think that 2 above is the biggest reason why targeting investment could never work on a practical basis. The data is just too noisy, lagged, arbitrary, inaccurate.
- Remember the classical dichotomy, between nominal (measured in $) and real (measured in physical, or inflation-adjusted units), and the (long-run) neutrality of money. Monetary policy cannot determine any real variable in the long run. The attempt to do so would cause accelerating inflation or deflation. (The attempt to peg employment with a vertical LR Phillips Curve was one example of this). So REAL investment won't work as a target, but NOMINAL investment might.
- It is VERY hard to measure investment (even gross investment, let alone net investment). Where do you draw the line between consumption and investment; current expenses and business investment? Human capital? Home improvements? Cars? It's also noisy, on a monthly basis. Somebody buys a big order of airplanes, and up it goes. (Depends on size of country, but a guy at the bank of Canada once told me that US is a big enough country that this noise gets smoothed out, but Canada isn't, so never believe Canadian monthly GDP figures). And data comes in with a long lag.
- Some fluctuations in (real) investment are good, and you don't want to try to smooth them out, or to tighten monetary policy whenever investment increases. In Canada, I think of the oil sands investment, big enough to show up in national figures. Maybe a housing boom as well, if baby-boomers suddenly all reach the age of wanting houses.
- True. This would imply that this scheme should only be used to correct money-related reasons for a shortage or excess of investment, and not to try to change the underlying dynamic of the economy. For example, if underinvestment is due to an imbalance between the availability of long-term debt and the demand for long-term investment; or due to uncertainty that short-term debt can be rolled over in the future. Arguably the central bank has a role in reassuring markets that debt of various maturities will always be accessible.
- I had not appreciated this factor to the extent that Nick makes clear. Although I would specifically exclude residential investment from the target, this only addresses a small part of the point. As someone who runs a business myself, I am well aware of the inadequacies of standard accounting measures of capital investment (for example software is usually considered a revenue expense even though nobody would buy software on that basis). So we might need to invent some kind of proxy to understand the actual levels of productive investment. I would think that in the current economy it's clear that investment is below optimal levels, but without further evidence I would be hard pressed to defend that point.
- Again true, though this is also true of inflation targeting which tends to be targeted on a 2-year horizon. As Nick says, point 2 is probably the main issue.