To follow up on IS-LM, I think some clarification and more generalized criticism are needed. Also my title for that post went a bit too far (as I later realized) which I rectified.
The first major flaw of the model is that it does not differentiate between interest rates demanded by savers from banks and the interest rates demanded by banks from entrepreneurs undertaking investment projects. If the model did differentiate between those two kinds of interest rates there could be no IS curve.
Also, even if we abstract from the problems with curves, what I meant in the previous post but did not express exactly correctly is that the volume of investment loans is not determined solely by the savers’ income and liquidity preference. Savers may have high liquidity preference and demand high interest rates from banks for foregoing the hoarded money but if entrepreneurs can offer investment projects which are productive enough for banks to finance banks can offer savers the right interest rates.*
If entrepreneurs cannot offer such projects it means that hoarding is entirely normal from the economic standpoint. It means that hoarders do not want to buy more than the economy currently produces at the prices at which the additional consumer goods would be offered. Thus, prices should fall. Attempts to get someone else to spend instead of hoarders are essentially redistribution plain and simple.
Thus, the IS-LM model by its very design completely ignores one of the two decentralized mechanisms that can take the economy out of the recession without redistribution – entrepreneurial innovation.
On my second point from the previous post, I was somewhat sloppy when I talked about a straightforward contradiction. According to the model, if the LM curve is taken into account, income determines not saving but actually how much consumption the receivers of monetary income will forgo. The rates they will demand from banks are determined by their liquidity preference.
This already presents a problem for IS-LM because it renders the IS-LM graph meaningless in terms of how things actually happen in time. In other words, IS-LM attempts to make synchronous three phenomena happening asynchronously under the very logic of the model. Under the logic of the model people first forgo consumption, then face the choice of whether to hoard the money or invest it according to their liquidity preference and only as a result of their choice the exact volume of investment is achieved. The representations of those three phenomena on the same graph creates a dynamic nightmare. It can then be asked whether it even makes sense to talk about people first forgoing consumption and then making a decision whether to hoard or invest it but this is too much to demand from such a comparative-statics model.
But the deeper issue is that the decisions to forgo consumption are modeled as mechanical and exogenously determined. People just do not consume some part of their income according to the exogenously set marginal propensity to consume. Of course the marginal propensity to consume postulated by the model might change and it may be said to account for changes in saving but that marginal propensity to consume may itself be not an exogenous variable but may itself partly depend on income. In such a case (which is quite plausible) the whole edifice of IS-LM collapses because a scenario can be easily imagined when as a result of hoarding by some people the total income falls in the first period but then the reduced income may lead to more consumption in the second period out of the received income.
*I neglected this aspect for simplicity but this logic applies to the decisions of entrepreneurs whether to hoard or reinvest their profit. Their decision whether to hoard or reinvest it will depend not just on their liquidity preference but also on whether they can think of sufficiently productive investment projects that they may undertake.