IS-LM Model: Equilibrium and Shifts
Table of Contents
The IS-LM model explains how spending decisions and the demand to hold money interact to determine output and an interest rate in the short run. Its practical question is why a fiscal expansion can raise production while also raising borrowing costs, and why a monetary expansion can have different effects depending on economic conditions.
This guide derives both curves, solves a hypothetical economy, and checks how its equilibrium changes. It also separates the textbook assumption of a fixed money supply from a central bank that implements an interest-rate target. The numerical results illustrate the model; they are not estimates of the U.S. economy.
What Is the IS-LM Model?
IS represents goods-market equilibrium: planned purchases equal production. LM represents money-market equilibrium: desired real money holdings equal the available real money stock. Their intersection identifies an output level and interest rate that satisfy both conditions at once. MIT’s introductory notes develop these two markets for a closed economy with predetermined prices and wages. [1]
On a standard IS-LM diagram, real output Y is on the horizontal axis and the interest rate is on the vertical axis. A point on IS clears the goods market; a point on LM clears the money market. Only an intersection clears both.
IS refers to investment and saving; LM refers to liquidity preference and money supply. These names describe equilibrium conditions, not two independent predictions of the interest rate. The model connects the spending adjustment taught by the fiscal multiplier effect with the money-market equilibrium.
Set the Assumptions and Units First
Use a closed economy, excluding exports, imports and international capital flows. Hold the price level fixed while solving each short-run equilibrium, and assume firms can respond to demand by changing production. Government purchases and lump-sum taxes are given. Consumption rises with disposable income; planned investment falls when its financing rate rises.
Investment responds to a real interest rate, whereas the opportunity cost of holding non-interest-bearing money is a nominal interest rate. To use one rate in the calculations below, make the additional assumption that expected inflation is zero. Then the nominal rate i and real rate r coincide. Sticky current prices alone do not require zero expected future inflation. With nonzero expected inflation, keep the two rates separate: r is approximately i minus expected inflation. [2]
| Variable | Meaning and units |
|---|---|
| Y, C, I, G, T | Real output, consumption, planned investment, government purchases and taxes; USD billions per year. |
| M | Nominal money stock, in USD billions; a stock, rather than annual spending. |
| P | Dimensionless price level; P = 1 is the base-price level. |
| m = M/P | Real money balances, in billions of base-price dollars. |
| i | Annualized interest rate in percent: enter 5 for 5%, rather than 0.05. |
The model’s money is a simplified non-interest-bearing liquid asset held against an interest-bearing alternative. It is not an assertion that bank reserves, the monetary base and M2 are interchangeable or move one-for-one. Modern U.S. implementation requires the separate discussion below.
Derive the IS Curve From Planned Spending
Let C0 be autonomous consumption, c the marginal propensity to consume, I0 autonomous investment, and b the sensitivity of investment to the rate. Assume 0 < c < 1 and b > 0. The spending relationships are:
I = I0 − b × i
Y = C + I + G
Substitute consumption and investment into the equilibrium condition and collect the income terms:
Y = [C0 − cT + I0 + G − bi] / (1 − c)
That produces a downward-sloping IS curve. A rate decline raises investment; the resulting income and consumption response raises equilibrium output. Higher government purchases or autonomous private spending instead increase output at every given rate, shifting IS right. A cut in lump-sum taxes also shifts IS right in this specification. The separation between behavioral equations and equilibrium is central to the university teaching treatment. [3]
Do not confuse this condition with the GDP accounting identity. Recorded expenditure equals recorded output even away from planned equilibrium because unsold production can count as unintended inventory investment. IS describes combinations where planned purchases match production. In the closed economy, that is also the condition for planned national saving, Y − C − G, to equal planned investment.
Derive the LM Curve From Money Demand
Suppose real money demand is L(Y, i) = kY − hi, where k and h are positive. More output means more transactions and greater desired money holdings. A higher nominal rate raises the opportunity cost of holding money, reducing desired balances. Set demand equal to the fixed real supply:
i = (k/h) × Y − (M/P)/h
The LM curve slopes upward in the output-rate diagram. As income increases, people want more money for transactions. With the stock fixed, the rate must rise enough to discourage other desired money holdings. This is a relationship among money-market equilibria, rather than a claim that the central bank raises its target whenever output rises. [1]
Holding P fixed, a larger M shifts LM down or right: a lower rate is consistent with equilibrium at any given Y. A smaller M shifts it up or left. A change in income shifts money demand in a separate money-quantity diagram, but it produces a movement along the existing LM curve when M, P and the money-demand function remain unchanged.
Solve a Numerical IS-LM Equilibrium
Consider a hypothetical economy with the following annual spending functions and money stock. Keep P = 1, expected inflation zero, and all behavioral coefficients fixed:
C = 100 + 0.75 × (Y − 200), so taxes are 200 and the marginal propensity to consume is 0.75.
I = 300 − 20i, G = 250, and M/P = 600.
L(Y, i) = 0.5Y − 40i.
The coefficient 20 means a one-percentage-point rate increase reduces annual planned investment by $20 billion. The coefficient 40 means it reduces desired real money balances by $40 billion. Because Y is an annual flow and money is a stock, the coefficient 0.5 has units of years. The coefficients deliberately preserve that distinction.
1. Calculate the goods-market schedule
Substitute the inputs into Y = C + I + G:
0.25Y = 500 − 20i
Y = 2,000 − 80i
2. Calculate the money-market schedule
Insert the fixed real money supply into money demand:
Y = 1,200 + 80i
3. Find the intersection and check both markets
Both schedules must give the same Y. Thus 2,000 − 80i = 1,200 + 80i, which gives 160i = 800 and i = 5%. Substitution gives Y = 2,000 − 80 × 5 = 1,600 billion dollars per year.
| Item | Calculation | Result |
|---|---|---|
| Consumption | 100 + 0.75 × (1,600 − 200) | 1,150 billion/year |
| Planned investment | 300 − 20 × 5 | 200 billion/year |
| Planned purchases | 1,150 + 200 + 250 | 1,600 billion/year |
| Real money demand | 0.5 × 1,600 − 40 × 5 | 600 billion |
Planned purchases equal output, and real money demand equals the available stock. Solving just IS would leave many possible output-rate pairs; solving just LM would do the same. The intersection removes that indeterminacy.
Fiscal Expansion and Interest-Rate Crowding Out
Start again at Y = 1,600 and i = 5%. Raise annual government purchases from 250 to 300, leaving taxes, the real money stock and all behavioral functions unchanged. IS shifts right; LM stays fixed.
The new IS equation is Y = 2,200 − 80i. Equating it to LM gives 2,200 − 80i = 1,200 + 80i, so i = 1,000 / 160 = 6.25% and Y = 1,700. Output rises by $100 billion per year, while the rate rises by 1.25 percentage points.
Investment falls from 200 to 300 − 20 × 6.25 = 175. This $25 billion annual decline is the interest-rate crowding out effect in this example. Higher output increases transaction demand for the fixed money stock; the higher rate required along LM restrains private investment. The model makes the size of this response depend on its behavioral slopes. [4]
Compare the Fixed-Rate and Joint-Equilibrium Multipliers
If the rate were held at 5%, the new IS curve would instead give Y = 2,200 − 80 × 5 = 1,800. That $200 billion increase equals the simple fixed-rate spending multiplier: [1 / (1 − 0.75)] × 50 = 4 × 50 = 200. This is a counterfactual comparison; it is not a joint equilibrium with the original real money stock of 600.
With LM unchanged, the actual model multiplier is 100 / 50 = 2. Do not equate the $25 billion investment decline with the $100 billion output shortfall relative to the fixed-rate comparison. The investment loss also reduces induced consumption: 4 × 25 = 100. In the actual new equilibrium, consumption is 1,225 and 1,225 + 175 + 300 = 1,700.
The results depend on the assumptions, rather than establishing that fiscal policy always has a multiplier of two. A steeper LM generally produces more rate pressure for a given fiscal shift; a flat LM can eliminate this particular rate-induced offset.
A Textbook Monetary Expansion and the Policy Mix
Reset to the original economy, including G = 250. Now increase nominal M from 600 to 700 while P remains 1, raising real balances to 700. Keep the spending functions unchanged. This shifts LM down or right, while IS stays fixed.
The new LM equation is Y = 1,400 + 80i. Equating it to the original IS gives 2,000 − 80i = 1,400 + 80i, so i = 600 / 160 = 3.75% and Y = 1,700. Investment rises to 300 − 20 × 3.75 = 225; money demand is 0.5 × 1,700 − 40 × 3.75 = 700.
The economy moves along unchanged IS toward a lower rate and higher output. This is a comparative-static result for an exogenous money supply, not a forecast of the effect of a Federal Reserve reserve purchase.
Finally, combine the two hypothetical changes: G = 300 and real balances = 700. Compare that joint policy mix with each change separately, always using the same original baseline:
| Scenario | Output (USD billions/year) | Rate (%) | Investment (USD billions/year) |
|---|---|---|---|
| Original: G = 250, m = 600 | 1,600 | 5.00 | 200 |
| Fiscal only: G = 300, m = 600 | 1,700 | 6.25 | 175 |
| Money only: G = 250, m = 700 | 1,700 | 3.75 | 225 |
| Both: G = 300, m = 700 | 1,800 | 5.00 | 200 |
For the combined case, 2,200 − 80i = 1,400 + 80i restores the 5% rate and gives output of 1,800. The larger money stock accommodates higher money demand at that unchanged rate. Investment remains at 200, while consumption becomes 1,300. The policy mix determines the composition of spending as well as total output.
IS and LM Shifts vs Movements Along a Curve
Ask which relationship changed at a given output-rate pair. A variable already plotted on an axis generally moves the economy along a schedule; a change in a held-fixed determinant changes the schedule itself.
| Change, other assumptions fixed | Curve shift | Joint-equilibrium result |
|---|---|---|
| Higher G or lower lump-sum T | IS right | Higher Y and i; movement along LM. |
| Lower autonomous private spending | IS left | Lower Y and i; movement along LM. |
| Higher M at fixed P | LM down/right | Higher Y, lower i; movement along IS. |
| Higher P at fixed M | LM up/left | Lower Y, higher i; movement along IS. |
These signs assume the stated positive response coefficients and an unconstrained intersection. Changes in consumption behavior or money-demand sensitivities can also change slopes, not just produce parallel shifts. A simultaneous fiscal and monetary change need not have the same interest-rate result as either policy alone.
What Changes in a Liquidity Trap?
The ordinary linear LM equation above has finite rate sensitivity and slopes upward. A liquidity trap introduces a different region: near a lower bound, money and short-term safe bonds can become close substitutes, so additional money is willingly held without a further rate decline. Textbook LM then has a horizontal segment. [5]
If IS intersects that flat segment, increasing money supply can leave the output-rate intersection unchanged. A fiscal expansion can raise output without raising the rate while the new intersection remains on the flat segment. In the simple spending model, the fixed-rate multiplier then applies. A sufficiently large expansion that reaches the upward segment changes that conclusion.
Zero is the elementary model’s lower-bound approximation, not a universal limit on all nominal rates. For a concrete historical counterexample, the ECB announced a deposit-facility rate of −0.10%, effective June 11, 2014. That was a specific policy rate in the euro area, not a U.S. rate or a claim that rates can fall without limit. [6]
The flat-LM result concerns the elementary money-for-short-term-bonds channel. Expectations, longer-term yields and credit conditions can still matter. After its target reached 0%–0.25% in December 2008, the Federal Reserve used forward guidance and asset purchases aimed at broader financial conditions. Simon Potter’s 2017 New York Fed speech describes those tools; it does not imply a guaranteed output response. [7]
Expected inflation also matters: a low nominal rate can coexist with an elevated real borrowing rate if prices are expected to fall. The zero-expected-inflation assumption used in the numerical example should not be carried into every lower-bound episode.
IS-LM vs an Interest-Rate-Targeting Model
The Federal Reserve’s ample-reserves framework primarily controls short-term rates through administered rates, especially interest on reserve balances, while the FOMC specifies a federal funds target range. It does not implement each policy-rate choice simply by fixing a scarce reserve quantity and waiting for a textbook LM intersection. [8]
A dated operational example is the Fed’s July 2026 Monetary Policy Report: it described reserve-management purchases to maintain ample reserves and administered rates keeping the effective federal funds rate within its target range. That illustrates why supplying reserves to support rate control is distinct from the numerical fixed-M experiment above. [9]
For a short-run exercise that holds the policy rate at a chosen target, draw a horizontal policy-rate curve instead of imposing the fixed-money-supply LM. Under the same single-rate and zero-expected-inflation simplifications, IS determines output at that chosen rate. Cutting the target shifts the horizontal curve down. This is a modeling abstraction, not a claim that a target range is literally one rate.
| Framework | What pins the rate relationship? | Meaning of the curve |
|---|---|---|
| Ordinary textbook LM | Fixed real money stock and money demand | Upward slope: more output raises the clearing rate. |
| Liquidity-trap segment | Lower-bound constraint and asset substitutability | Flat region: more money need not lower the rate. |
| Fixed policy-rate target | Assumed central-bank rate choice | Horizontal at the chosen target, which may be above the lower bound. |
A horizontal target curve therefore does not diagnose a liquidity trap. Nor does it describe the central bank’s entire reaction function: an MP curve can instead incorporate systematic responses to output and inflation, as in Romer’s teaching model. [2] Actual business financing rates also depend on expected future rates, maturity and credit conditions; fixing the overnight target does not hold every borrowing cost constant. [10]
Connect IS-LM to the Aggregate Demand Curve
Return to the original fixed-money-supply model: G = 250, nominal M = 600, and the original spending and money-demand functions. Now compare alternative fixed-price equilibria by varying P alone, keeping expected inflation at zero. These are separate hypothetical price-level settings, not a timeline of inflation.
A higher P reduces M/P, shifts LM up or left and lowers equilibrium output along IS. Repeating that exercise traces a downward-sloping aggregate demand curve in output-price-level space. This derivation depends on holding nominal money fixed; Chinn’s university notes state that policy assumption explicitly. [11]
In this example, substitution of i = 0.0125Y − (600/P)/40 into Y = 2,000 − 80i gives Y = 2,000 − Y + 1,200/P. Therefore:
| Price level P (base = 1) | Real money (USD billions) | Output (USD billions/year) | Rate (%) |
|---|---|---|---|
| 0.8 | 600 / 0.8 = 750 | 1,750 | 3.125 |
| 1.0 | 600 / 1 = 600 | 1,600 | 5.000 |
| 1.2 | 600 / 1.2 = 500 | 1,500 | 6.250 |
Moving between these price-level settings is a movement along AD, even though LM changes in the supporting IS-LM diagram. A change in G or nominal M instead changes output at a given P and shifts AD. IS-LM alone does not determine the actual price level; add aggregate demand and aggregate supply to analyze that additional question.
How to Analyze an IS-LM Policy Scenario
- Choose the framework. State whether money supply or the policy-rate target is held fixed, and whether prices can change.
- Separate interest rates and units. Identify the nominal money-demand rate, real investment rate, expected inflation and stock-versus-flow measurements.
- Identify the initial shock. Fiscal or autonomous spending changes affect IS; money supply or money-demand changes affect LM in the textbook case.
- Solve both conditions. Compute the new output and rate together, then substitute back into planned spending and money demand.
- Check the boundary conditions. A lower bound, a central-bank response, capacity pressure or an external sector can change the result.
Use the loanable funds market to examine saving and investment schedules directly. Its horizontal axis measures a saving/investment flow, whereas IS-LM’s horizontal axis measures output. For an open-economy extension involving exchange rates and capital flows, see the Mundell-Fleming framework in forex fundamental analysis. Those external channels are deliberately excluded from the solved example.
Common Mistakes
- Entering 0.05 into a percent-based equation. The example’s coefficients require i = 5 for a 5% annualized rate.
- Solving one market only. An output-rate pair must satisfy both schedules to be a joint equilibrium.
- Shifting LM merely because income rises. At unchanged money supply, price level and money-demand behavior, that is movement along LM.
- Applying the simple multiplier with an unchanged upward LM. The resulting rate response changes investment and the total output effect.
- Equating investment crowding out with the full output shortfall. Reduced investment also affects induced consumption in this model.
- Calling every horizontal rate curve a liquidity trap. A chosen policy target can be horizontal well above a lower bound.
- Treating a larger reserve balance as an automatic M2 or GDP increase. The textbook money-supply experiment omits the institutions and transmission channels needed for that inference.
Limitations of the IS-LM Model
The example deliberately fixes prices, expectations and behavioral coefficients and uses one financing rate. It omits default risk, changing credit spreads, banking constraints, supply shocks, foreign trade and policy lags. Its precise numbers follow from chosen assumptions, not measured policy effects.
IS-LM is useful for checking the internal logic of a demand-side argument. It does not establish how much spare capacity exists, how quickly production responds, or how future taxes and policy commitments alter today’s decisions. A demand expansion near capacity can create price pressure that the fixed-P calculation does not capture.
Likewise, the model’s short-run output change does not by itself describe long-run productive capacity or welfare. Government investment may affect future supply, while the elementary G term only records current purchases. Keep conclusions proportional to the model and use evidence about the actual policy regime and economy before applying them to an investment decision.
Frequently Asked Questions
Sources
- MIT OpenCourseWare: Simple Notes on the ISLM Model, Spring 2011. Closed-economy goods and money markets, pp. 1–5.
- David Romer, UC Berkeley: Short-Run Fluctuations, January 2012 revision. MP rules, nominal versus real rates, and expected inflation at a lower bound.
- Leigh Tesfatsion, Iowa State University: The IS-LM Model, October 12, 1996. Behavioral assumptions, planned equilibrium and linear schedules, sections D–E.
- Roberto Rigobon, MIT: Managerial International Macroeconomics, March 6, 2016 preliminary notes. Policy effects, liquidity trap and slope-based limiting cases, chapter 3.
- Paul Krugman: Thinking About the Liquidity Trap, December 1999. Elementary flat-LM and asset-substitution mechanism, section 1.
- European Central Bank: Monetary Policy Decisions, June 5, 2014. Historical negative deposit-facility rate, effective June 11, 2014.
- Simon Potter, New York Fed: Implementing Monetary Policy with the Balance Sheet, November 6, 2017. Historical lower-bound response; speech expresses the author’s views.
- Federal Reserve Board: Interest on Reserve Balances Frequently Asked Questions. Relevant operational-framework answers posted September 23, 2025.
- Federal Reserve Board: Monetary Policy Report, July 10, 2026, Part 2. Ample reserves, reserve-management purchases and administered-rate implementation.
- Federal Reserve Board: Monetary Policy—What Are Its Goals? How Does It Work?. Policy transmission and the role of expectations in longer-term rates.
- Menzie D. Chinn, University of Wisconsin–Madison: Aggregate Demand–Aggregate Supply, Fall 2020. Deriving price-level AD with nominal money held fixed, p. 1.
Sources accessed November 14, 2026. All numerical model calculations are original hypothetical examples, rather than empirical policy estimates.
Disclaimer
This article is for educational and informational purposes only and does not constitute investment or policy advice. Economic models simplify actual institutions and behavior; the hypothetical multipliers and equilibrium values are not forecasts. Consult primary evidence and qualified professionals when evaluating a real policy or investment decision.
