What the General Theory actually argued
Keynes published The General Theory of Employment, Interest and Money in 1936 [1]. Its chapter 3 states the central claim: output and employment are set by effective demand, and Say's law, under which demand always matches whatever is supplied, is equivalent to assuming there is no obstacle to full employment. Keynes rejected the two classical links that were supposed to remove such obstacles. Lower money wages would not reliably restore full employment, he argued in chapter 19, because a wage cut also cuts incomes and spending; whatever good it does comes mainly through raising the quantity of money in terms of wage-units and so lowering interest rates, which a flexible monetary policy could do directly. And the interest rate cannot be counted on to balance saving and investment, because the demand for money as a liquid asset competes with lending.
The book's second idea is that the economy can settle well below full employment, with no automatic force pulling it back, when private confidence collapses. Investment rests on fragile long-term expectations and on 'animal spirits', a spontaneous urge to act, as much as on calculation (chapter 12). Keynes also discussed, as a possibility rather than his central case, a floor below which interest rates would not fall because almost everyone would rather hold cash, and added that he knew of no example of it hitherto (chapter 15). That case is what is now called the liquidity trap.
Hicks formalized Keynes in 1937 [2] with what became the IS-LM framework, translating the argument into a simultaneous equilibrium in the goods market and the money market. The framework clarified when each instrument is strong and when neither can reach the target alone, and the mechanism is worth stating rather than memorising. Fiscal expansion is powerful when the LM curve is FLAT and the IS curve is STEEP: a flat LM means money demand is highly interest-elastic, so the extra spending lifts the interest rate only slightly, and a steep IS means investment barely responds to whatever rate rise does occur -- together, little crowding out. Monetary expansion is powerful in the mirror case, a STEEP LM and a FLAT IS: a steep LM means a given change in the money supply moves the interest rate a long way, and a flat IS means interest-sensitive spending responds strongly to that move. The diagram leaves out the fundamental uncertainty behind investment that chapter 12 puts at the center, which is why Post-Keynesians treat it as a simplification of Keynes rather than a summary.
Aggregate demand as a macroeconomic stabilizer
The Keynesian multiplier is the mechanism by which an initial change in spending propagates through the economy as the recipients of that spending become spenders themselves. If the marginal propensity to consume is 0.8, then a one-unit increase in government spending produces an eventual total increase in output of five units, assuming no capacity constraints or crowding out. In practice, multipliers are lower because leakages into imports, taxes, and saving reduce the propagation. The empirical debate about the size of fiscal multipliers has not been resolved, but estimates are consistently higher for slack economies than for full-employment ones [3].
The policy implication is straightforward: when private spending falls and the economy is below potential, there is a case for public spending to fill the gap. The argument does not require that government spending is always efficient or that public debt is costless. It requires only that unused capacity exists, that spending can mobilize it faster than relative-price adjustment, and that the borrowing cost of fiscal expansion is lower than the output cost of prolonged slack.
Critics have argued that Ricardian equivalence undermines the multiplier: forward-looking households will save more in anticipation of future taxes, offsetting the stimulus. The evidence on consumption responses to transitory payments does not support full Ricardian equivalence. Households spent a substantial part of the 2001 US tax rebates within months, and those with low liquid wealth or low income spent more [4], consistent with the Keynesian mechanism rather than the full-information benchmark.