Why Crowding Out Varies
An increase in government spending can support demand, but its effects on private spending depend on the conditions in which it occurs. If the additional spending is financed by borrowing, the government competes for funds with private borrowers. A higher real interest rate can encourage private saving while making some private investment projects less attractive. The balance between those responses determines how much private investment is displaced.
The interest-rate channel is not equally strong in every setting. It depends on how readily private saving responds to interest rates, how sensitive investment is to borrowing costs, whether the economy has unused productive resources, and how monetary conditions affect market rates. The simple model below isolates the first two forces; the discussion then explains the limits imposed by capacity and monetary conditions.
A Local Loanable-Funds Model
Suppose government borrowing increases by \(D>0\), measured in billions of dollars. Let \(\Delta r\) be the resulting change in the relevant real interest rate, measured in percentage points. Around the initial equilibrium, assume the change in desired private saving is \(\alpha\Delta r\), and the change in desired private investment is \(-\beta\Delta r\). Here \(\alpha\) and \(\beta\) are nonnegative response coefficients, measured in billions of dollars per percentage point.
This is a local, linear model. It holds other influences on saving and investment fixed, assumes the rate change affects both schedules as specified, and treats the additional government borrowing as a shift in the demand for funds. It does not by itself describe changes in output, prices, or the central bank's policy response.
Deriving the Interest-Rate and Investment Effects
Substituting the two response equations into the market-clearing condition gives \(\alpha\Delta r=D-\beta\Delta r\). The investment response appears with a negative sign because a rate increase reduces desired investment: if private investment falls, private saving need not rise by the full amount of the new government borrowing.
When \(\alpha+\beta>0\), division gives the rate change. Substituting that change into the investment equation gives the amount of investment displaced. The ratio of displaced investment to additional government borrowing is the model's crowding-out share.
The denominator condition matters. If \(\alpha=\beta=0\) and \(D>0\), neither private saving nor private investment changes when the rate changes. The market-clearing equation would require \(0=D\), which is impossible. In that case this model has no market-clearing rate change; the displayed formulas do not apply.
For \(\alpha+\beta>0\), a larger saving response \(\alpha\) reduces the rate increase and reduces the share of borrowing that displaces investment. A larger investment sensitivity \(\beta\), holding \(\alpha\) fixed, increases the displaced share. In the limiting case \(\beta=0\) and \(\alpha>0\), investment does not respond to the rate, so this interest-rate channel produces no investment crowding out. In the limiting case \(\alpha=0\) and \(\beta>0\), the model predicts that private investment falls by the full amount of new borrowing.
Measure \(D\) as the increase in government borrowing, not simply as the increase in spending if taxes or other financing also change.
Use \(\alpha\) for the private-saving response and \(\beta\) for the private-investment response, with units consistent with the rate change.
For positive \(D\), the formula requires \(\alpha+\beta>0\). If both coefficients are zero, the model cannot clear the market through an interest-rate change.
Separate the rate change, the amount of investment displaced, and the fraction of government borrowing it represents. Do not treat this local financial model as a complete forecast of output.
Worked Example: A Strong Saving Response
A hypothetical economy experiences an increase of \(\$12\) billion in government borrowing. Near the initial equilibrium, private saving rises by \(\$3\) billion per percentage-point increase in the real interest rate, while private investment falls by \(\$1\) billion per percentage point. Using the local loanable-funds model, calculate the rate increase, the change in private saving, the change in private investment, and the crowding-out share.
Method. Set \(D=12\), \(\alpha=3\), and \(\beta=1\), with dollar amounts in billions. The units of both response coefficients are billions of dollars per percentage point. Since \(\alpha+\beta=4>0\), the model's formulas apply.
Rate change. The denominator is \(3+1=4\). Therefore, \(\Delta r=12/4=3\) percentage points.
Saving change. The increase in private saving is \(\Delta S_p=3\times3=\$9\) billion.
Investment change. The change in private investment is \(\Delta I=-1\times3=-\$3\) billion.
Crowding-out share. The displaced investment is \(\$3\) billion out of \(\$12\) billion of additional government borrowing, so the share is \(3/12=0.25\), or 25%.
Check. The market-clearing condition is \(\Delta S_p=D+\Delta I\). Its right-hand side is \(12+(-3)=9\), equal to the \(\$9\) billion increase in private saving. The common mistake avoided here is treating all \(\$12\) billion of government borrowing as an equal reduction in private investment. In this example, private saving supplies most of the additional funds.
Comparing Saving and Investment Sensitivity
The rate increase and the amount of investment displaced are related but distinct outcomes. A more responsive saving schedule can limit the rate rise because households and firms supply more funds as the return increases. A more responsive investment schedule makes a given rate increase reduce investment by more. The formulas show that the displaced fraction depends on the relative sizes of the two responses, not on either coefficient alone.
| Change in conditions, holding other factors fixed | Implication in the model |
|---|---|
| Higher \(\alpha\): saving responds more to the rate | Smaller rate increase and smaller fraction of borrowing displaced from investment |
| Higher \(\beta\): investment responds more to the rate | Larger fraction of borrowing displaced from investment |
| \(\beta=0\), with \(\alpha>0\) | No investment crowding out through this rate channel |
| \(\alpha=0\), with \(\beta>0\) | The model assigns the full increase in borrowing to a fall in private investment |
Worked Example: Same Borrowing Increase, Different Investment Sensitivity
Two hypothetical economies each face an increase of \(\$8\) billion in government borrowing. In Economy A, \(\alpha=1\) and \(\beta=3\); in Economy B, \(\alpha=3\) and \(\beta=1\). All coefficients are in billions of dollars per percentage point. Compare each economy's interest-rate increase, private-investment change, and crowding-out share.
Method. Apply the local model separately, first checking that the sum of the response coefficients is positive in each economy.
Economy A. Here \(\alpha+\beta=1+3=4>0\). The rate increase is \(8/4=2\) percentage points. Private investment changes by \(-3\times2=-\$6\) billion, so the crowding-out share is \(6/8=0.75\), or 75%.
Economy B. Here \(\alpha+\beta=3+1=4>0\). The rate increase is also \(8/4=2\) percentage points. Private investment changes by \(-1\times2=-\$2\) billion, so the crowding-out share is \(2/8=0.25\), or 25%.
Check. In Economy A, saving rises by \(1\times2=\$2\) billion and \(8+(-6)=\$2\) billion. In Economy B, saving rises by \(3\times2=\$6\) billion and \(8+(-2)=\$6\) billion. Both satisfy market clearing. The rate increase is the same because the sums \(\alpha+\beta\) match, but the investment responses differ because the investment sensitivities do not. The common mistake avoided is inferring the amount of crowding out from the rate change alone.
Available Resources and the Meaning of Crowding Out
The loanable-funds calculation describes a financial channel. It does not establish that every fiscal expansion must reduce current output or private spending by the amount shown. The economy's available labor, capital, and productive capacity also matter. When substantial resources are idle, an increase in demand may lead firms to produce more without immediately bidding those resources away from private uses. Near capacity, by contrast, meeting additional public demand is more likely to require higher prices, reallocation of resources, or less private production.
A separate, simple expenditure-multiplier calculation can illustrate the distinction between a demand impulse and the real output response. In the basic model with a fixed marginal propensity to save \(m\), no interest-rate response, and no other leakage or supply constraint, the government-spending multiplier is \(1/m\). Its use requires those assumptions; it is not a formula for the crowding-out share.
Worked Example: Fiscal Demand With Idle Capacity
A hypothetical economy has a marginal propensity to save of \(0.25\). A temporary increase in government purchases is \(\$10\) billion, and the estimated output gap—the amount by which output could rise before reaching the stated capacity benchmark—is \(\$45\) billion. Under the simple multiplier model with interest rates, taxes, prices, and other leakages held fixed, calculate the unconstrained output response and assess whether it fits within the stated output gap.
Method. The assumptions permit use of the simple multiplier \(1/m\). First calculate the multiplier, then multiply it by the spending increase, and only then compare the result with the capacity margin.
Multiplier. With \(m=0.25\), the multiplier is \(1/0.25=4\).
Unconstrained output response. The model implies an output increase of \(4\times\$10=\$40\) billion.
Capacity comparison. The stated output gap is \(\$45\) billion, which exceeds the model's \(\$40\) billion response by \(\$5\) billion. Under the assumptions and the stated capacity benchmark, the calculated response fits within the available margin.
Check. The calculation compares like quantities: a model-implied output change of \(\$40\) billion and a stated output margin of \(\$45\) billion. It does not claim that the \(\$10\) billion spending impulse itself becomes \(\$10\) billion of output. The common mistake avoided is subtracting the spending impulse from the output gap without first applying the specified multiplier.
Monetary Conditions and the Interest-Rate Channel
The relevant rate for investment decisions is the real borrowing cost, meaning the nominal rate adjusted for expected inflation. When expected inflation is stable, a rise in nominal market rates generally corresponds to a rise in real rates; when expected inflation changes, the two need not move together. A nominal rate movement by itself therefore does not always reveal the change in the real rate that matters for this channel.
Monetary conditions affect whether fiscal borrowing translates into a higher real market rate. With monetary policy held fixed in the sense of no offsetting change to the relevant rate conditions, additional borrowing may put upward pressure on market rates. If monetary conditions prevent or limit that increase, the interest-rate channel of crowding out can be weaker. If they reinforce upward pressure, it can be stronger. These are conditional statements: the loanable-funds formulas do not specify a central-bank reaction or determine the outcome under a particular monetary policy.
There is also a distinction between a rate being held steady and the absence of all crowding out. A steady relevant rate means this particular interest-rate channel predicts no investment decline from a rate increase. It does not rule out other changes in private spending, resource allocation, or prices. Similarly, even when the model predicts some investment displacement, it does not determine the net effect of fiscal policy on total output without a broader model.
Common Exam Traps
- Using the formula when both response coefficients are zero. For \(D>0\), the model has no market-clearing rate change if \(\alpha=\beta=0\). Check \(\alpha+\beta>0\) before dividing.
- Confusing the rate change with the crowding-out amount. Investment changes by \(-\beta\Delta r\); the rate change alone is not a dollar amount of displaced investment.
- Ignoring units. If \(\alpha\) and \(\beta\) are in billions per percentage point, \(\Delta r\) must be expressed in percentage points for the product to be in billions.
- Assuming all public borrowing displaces private investment. That is the model's result only in the special case \(\alpha=0\), \(\beta>0\). With positive \(\alpha\), some borrowing is supported by additional private saving.
- Equating demand with output. A spending impulse is not automatically the same amount of output. Apply the appropriate multiplier under its assumptions before comparing an output response with available capacity.
- Inferring a real-rate change from a nominal rate alone. Expected inflation can change, so the real borrowing-cost movement may differ from the nominal movement.
Check Your Understanding
Use the model and assumptions stated in each question. Monetary amounts are in billions unless stated otherwise.
- Government borrowing rises by \(15\). Private saving responds by \(2\) per percentage point and investment falls by \(3\) per percentage point. Find the rate increase, the changes in saving and investment, and the crowding-out share. (Answer: \(\Delta r=3\) percentage points; \(\Delta S_p=+6\); \(\Delta I=-9\); crowding-out share \(=60\%\).)
- Borrowing rises by \(9\), with \(\alpha=0\) and \(\beta=3\). Calculate the rate increase and investment change, and state why this is the full-crowding-out case in the model. (Answer: \(\Delta r=3\) percentage points; \(\Delta I=-9\); the crowding-out share is \(100\%\) because private saving does not respond and investment supplies the entire adjustment.)
- Two economies each face \(D=10\). Economy A has \(\alpha=4,\beta=1\); Economy B has \(\alpha=1,\beta=4\). Calculate the rate increase and crowding-out share for each, and identify which has greater investment displacement. (Answer: A: \(\Delta r=2\) percentage points and \(20\%\) crowding out; B: \(\Delta r=2\) percentage points and \(80\%\) crowding out; B has greater investment displacement.)
- In the simple multiplier model, the marginal propensity to save is \(0.20\), government purchases rise by \(\$6\) billion, and the output gap is \(\$40\) billion. Find the unconstrained output response and determine whether it exceeds the stated gap. (Answer: Multiplier \(=5\); output response \(=\$30\) billion; it is \(\$10\) billion below the stated gap.)
- In a local model, \(\beta=2.5\) billion per percentage point. If monetary conditions keep the relevant real interest rate unchanged, what investment change does this interest-rate channel predict? If instead the real rate rises by \(1.2\) percentage points, what is the predicted investment change? (Answer: With no rate change, \(\Delta I=0\); with a \(1.2\)-percentage-point rise, \(\Delta I=-2.5\times1.2=-\$3.0\) billion.)