Aggregate demand and aggregate supply

• Define aggregate demand, name its components and explain what determines each of them.
• Distinguish a movement along AD or SRAS from a shift, and explain demand and cost shocks with labelled AD–AS diagrams.
• Draw and explain the monetarist/new classical curves (SRAS and vertical LRAS) and the three-section Keynesian AS curve.
• Explain short-run equilibrium, deflationary (recessionary) and inflationary gaps, and long-run equilibrium in each model.
• Explain what shifts LRAS and the Keynesian AS, and show the result for full-employment output.
• Compare the assumptions and the policy implications of the two models and judge which fits a given situation.

The aggregate demand and aggregate supply (AD–AS) model explains two things together: an economy's real GDP and its average price level. It looks like a demand and supply diagram, but its curves describe the whole economy, so the reasons for their shapes and their shifts are different.

• Aggregate demand (AD): Planned expenditure on domestically produced final output at each average price level: C + I + G + (X − M).
• Consumption (C): Household spending on goods and services. Newly built housing is counted as investment, not as consumption.
• Investment (I): Spending on newly produced capital goods and changes in stocks (inventories), including new housing. Buying an existing share or a second-hand building is not investment in this sense, because nothing new is produced.
• Government spending (G): Public purchases of goods and services. Transfer payments such as pensions are not part of G: they enter aggregate demand only when the people who receive them spend the money.
• Net exports (X − M): Exports minus imports. Imports are subtracted because C, I and G include spending on goods made abroad, which is not part of domestic output.
• Short-run aggregate supply (SRAS): Planned production at each average price level while money wages and other input prices have not yet adjusted.
• Potential output: The real GDP an economy produces when its resources are fully employed, also called full-employment output and labelled Yf. Full employment does not mean zero unemployment: frictional, structural and seasonal unemployment remain.
• Long-run aggregate supply (LRAS): In the monetarist/new classical model, a vertical line at potential output, whose position is fixed by the quantity and quality of resources, technology and institutions.
• Keynesian AS: One aggregate supply curve with three sections: horizontal where there is spare capacity, upward-sloping as bottlenecks raise costs, and vertical at full-employment output.
• Deflationary (recessionary) gap: Actual real output below potential output. The name does not mean that prices are falling.
• Inflationary gap: Actual real output above potential output, which puts upward pressure on costs and prices and cannot be sustained.
• Equilibrium: The price level and real output at which planned spending equals planned production. An equilibrium need not be desirable, equitable or at full employment.

Aggregate demand (AD) is the total planned spending on an economy's output at each average price level. Suppose the price level falls while money incomes, the money supply and every other condition stay the same. Three economy-wide effects raise the real output demanded. First, a given amount of money and other financial wealth buys more, so households are wealthier in real terms and consume more (a wealth effect). Second, households and firms need less money for their transactions, so more is saved and lent; interest rates fall, which encourages investment and consumption financed by borrowing (the interest-rate effect). Third, if foreign prices and the exchange rate are unchanged, domestic goods become cheaper relative to foreign goods, so exports rise and imports fall (the international trade effect). Together these explain a movement down the AD curve.

AD is not the demand curve for one good multiplied by a large number. A fall in the price of one good leads buyers to substitute towards it. A fall in the whole price level leaves relative prices unchanged, so the slope of AD rests on the three economy-wide effects above. The effects have a limit that is worth knowing: a falling price level (deflation) raises the real burden of existing debts and can lead buyers to postpone purchases, so in practice lower prices do not always raise spending.

AD shifts when a determinant of one of its components changes while the price level is unchanged. A rise in consumption, investment, government spending or net exports shifts AD to the right, because more real output is demanded at every price level. A fall in any of them shifts AD to the left.

Consumption. Greater consumer confidence about future income and job security raises C, because households are more willing to spend and to borrow. Lower interest rates raise C by making credit cheaper and saving less rewarding. A rise in household wealth, such as higher house or share prices, raises C through the wealth effect, although asset gains are unevenly distributed. A cut in income taxes raises disposable income and therefore C. A high level of household indebtedness lowers C: more income goes on repayments, and indebted households may use a tax cut to repay debt and not to spend. If households expect prices to rise, they bring some purchases forward, which raises C now.

Investment. Firms compare the expected return on new capital with the cost of financing it. Lower interest rates reduce that cost and raise I. Stronger business confidence, meaning higher expected sales and profits, raises I. New technology raises I because firms must buy new equipment to use it. Lower business taxes leave more profit after tax to pay for and reward investment. Heavily indebted firms invest less, because lenders are less willing to extend credit and profits go on servicing debt. Lower interest rates alone need not raise investment in a depressed market: if firms expect weak sales, cheap credit does not make new capacity profitable.

Government spending. G depends on the government's political and economic priorities: its decisions about health, education, defence and infrastructure, and about whether to support or restrain the economy. A decision to buy more locally provided maintenance services raises government purchases directly. A transfer payment to a household is different: it raises disposable income, and only the part that is then spent enters C.

Net exports. A rise in the incomes of trading partners raises their demand for imports, which are this country's exports, so X − M rises. A recession abroad lowers exports even if domestic firms have not raised their prices. A depreciation of the exchange rate makes exports cheaper abroad and imports dearer at home, which raises X − M; an appreciation does the opposite. Trade policy matters as well: a new tariff abroad on this country's goods lowers exports, while a domestic tariff on imports lowers M.

The accounting has to stay consistent. An imported household appliance raises C and M by the same amount, so it adds nothing to aggregate demand for domestic output. Not every extra unit of spending buys domestic output.

Short-run aggregate supply (SRAS) is the total real output firms plan to produce at each average price level in the short run, the period in which money wages and other resource prices are fixed. If the price level rises while wage contracts and other input prices have not changed, producing extra output becomes more profitable, so firms expand output along SRAS. This is a movement along the curve: its cause is a change in the average price level, with the conditions of supply held fixed.

SRAS shifts when costs of production change at every price level. The main cause is a change in the costs of factors of production: higher money wages that are not matched by productivity, or dearer energy, raw materials and imported components, raise unit costs. At each price level production is less profitable, so firms supply less and SRAS shifts left. An indirect tax works in the same way: a higher tax on firms' output or inputs adds to their costs and shifts SRAS left, while a subsidy lowers costs and shifts it right. Lower wages or resource prices, or higher productivity, also shift SRAS right. A policy can act on more than one curve: government spending on training raises G, and so AD, at once, and raises productive capacity only later.

After an adverse cost shock, firms supply less at the old price level than buyers plan to purchase. The excess demand pushes the price level up, planned spending contracts along AD, and the economy settles at a higher price level and a lower real output. A cost shock therefore creates a difficult trade-off. Expanding AD to restore output adds to the pressure on prices; reducing AD to restrain prices deepens the fall in output. Neither response repairs the original shortage of inputs. When costs go on rising, the result is cost-push inflation.

In the short run, macroeconomic equilibrium is where AD crosses SRAS. At that price level the real output buyers plan to purchase equals the real output firms plan to produce. If the price level were higher, firms would produce more than buyers wanted: unsold stocks would build up and firms would cut prices and output until the two plans matched. If it were lower, buyers would want more than firms produced, stocks would run down, and prices and output would rise.

The equilibrium changes whenever one of the curves shifts. With an upward-sloping SRAS there are four cases, and each moves the price level and real GDP in a known direction.

Nothing guarantees that the short-run equilibrium is at potential output, Yf. When equilibrium real GDP is below Yf there is a deflationary (recessionary) gap: aggregate demand is too low to employ all resources, so unemployment is above its natural rate. The extra unemployment is cyclical unemployment. When equilibrium real GDP is above Yf there is an inflationary gap: firms are using overtime and working their resources harder than can be sustained, unemployment is below its natural rate, and wages and prices are under upward pressure. When equilibrium real GDP equals Yf the economy is at full employment.

The lab below sets one of these cases in motion. A cut in interest rates raises consumption and investment, AD shifts right from AD1 to AD2, and the short-run equilibrium moves up along SRAS. Change the starting point to see the same shift close a deflationary gap or open an inflationary one.

The lab draws only the upward-sloping SRAS. It does not draw the Keynesian AS curve, and it does not shift SRAS or LRAS. Its flatter SRAS setting is still not the Keynesian AS curve: on the horizontal section of that curve, an increase in AD raises real GDP with no rise in the price level at all.

The monetarist/new classical model separates the short run from the long run. In the short run wages are fixed and SRAS slopes upward. In the long run all wages and resource prices are free to change, and the model assumes that they are flexible in both directions. If the price level doubles and wages and other costs double as well, firms' real costs and real profits are unchanged, so they produce the same output as before. Long-run aggregate supply (LRAS) is therefore vertical at potential output, Yf: in the long run output is determined by the quantity and quality of factors of production and by technology, not by the price level.

Long-run equilibrium in this model is where AD crosses LRAS, with SRAS passing through the same point. Real GDP equals Yf, so the economy is at full employment. Full employment does not mean that everyone has a job. At Yf unemployment equals the natural rate of unemployment: frictional, structural and seasonal unemployment remain, and only cyclical unemployment is absent.

Begin below potential output, with a deflationary gap and idle resources. Unemployment above the natural rate weakens workers' bargaining position, and because wages and other input prices are assumed to be flexible, they fall over time. Production becomes cheaper, SRAS shifts right, and output returns to Yf at a lower price level. Begin instead with an inflationary gap. Firms compete for scarce labour and materials, so wages and other costs rise, SRAS shifts left, and output falls back to Yf at a higher price level. LRAS does not move in either case, because the quantity and quality of productive resources are unchanged.

Two implications follow. Any gap is temporary, because the economy corrects itself without government intervention. And a rise in AD cannot raise real GDP in the long run: once wages have caught up with prices, output is back at Yf and only the price level is higher.

The adjustment is a result of the model's assumptions; it is not a promise that a recession will be short or painless. Wage contracts, resistance to pay cuts, debts fixed in money terms and weak spending can delay it for years, and the cost of that delay is borne by the people who are unemployed while wages fall.

Keynesian economics rejects the assumption that wages and prices fall quickly when demand is weak. Money wages are sticky downward: contracts, minimum wage laws and workers' resistance to pay cuts stop them falling, and firms cut output and jobs before they cut prices. The Keynesian model therefore draws a single aggregate supply curve, the Keynesian AS, with no separate short-run and long-run curves.

At low output, unemployed workers and idle machinery (spare capacity) allow firms to meet extra demand with no pressure on costs. Across this horizontal section an increase in AD raises real output and leaves the price level unchanged. As output rises, bottlenecks emerge: some skills, materials and types of equipment become scarce while others are still unemployed, so additional output becomes more costly and the curve slopes upward. An increase in AD now raises both real output and the price level. At full-employment output, Yf, every resource is in use and the curve is vertical: additional demand raises the price level without increasing real output.

Because wages and prices do not fall, a low-demand equilibrium can persist below full employment. Nothing shifts the supply curve to the right. Investment can stay low because firms expect weak sales. A wage cut, if it came, would also reduce workers' incomes, and so might lower consumption and make debts harder to repay. A deflationary (recessionary) gap, with its cyclical unemployment, can therefore last until aggregate demand rises. This is the Keynesian case for demand management: expansionary fiscal or monetary policy to close the gap.

Equilibrium at full employment occurs when AD meets the Keynesian AS exactly at Yf. If AD rises beyond that level an inflationary gap opens: planned spending exceeds what the fully employed economy can produce, and the result is a higher price level with no extra output.

An increase in AD can bring unemployed workers and idle equipment into production. That raises actual output without any increase in what the economy is able to produce. Productive capacity grows only when the determinants of long-run aggregate supply improve. There are four, and spending more money does not by itself deliver any of them: skills have to be usable and equipment has to work.

• Quantity and quality of factors of production. A larger or more active workforce, net investment in machinery and infrastructure, newly usable natural resources and more enterprise raise the quantity of resources. Education, training and better health raise the quality of labour (human capital), and newer capital goods raise the quality of capital. In each case more can be produced when resources are fully employed.
• Improvements in technology. Better production methods and equipment allow the same workers and capital to produce more output, which is a rise in productivity.
• Increases in efficiency. Resources are used with less waste, or move to the uses in which they produce most, for example when competition forces firms to cut costs or workers move from declining to expanding industries. The same resources then yield more output.
• Institutional changes. Secure property rights, a legal system that enforces contracts, a sound banking system, less bureaucracy and less corruption make it safer and cheaper to invest, trade and start firms, so more of the economy's potential is realised.

When productive capacity grows, LRAS shifts right, from LRAS1 to LRAS2, and full-employment output rises from Yf1 to Yf2. The same improvements usually let firms produce more at each price level, so SRAS shifts right as well. With AD unchanged, the new long-run equilibrium has a higher real GDP and a lower price level. If AD is also growing, real GDP rises with little change in the price level, so the economy reaches a higher real GDP at a lower price level than it otherwise would.

One caution: LRAS marks potential output, not current output. A shift of LRAS alone does not move the short-run intersection of AD and SRAS; the economy would simply have more spare capacity. Actual output moves towards Yf2 only if costs fall (SRAS shifts right), AD rises, or the model's automatic adjustment takes place.

In the Keynesian diagram the same growth in capacity moves the vertical section of the AS curve to the right, from Yf1 to Yf2, and lengthens the horizontal section. Equilibrium output rises only if aggregate demand is high enough to use the new capacity.

The two models differ in one assumption: whether wages and prices fall quickly enough to restore full employment. The table sets out what follows from each answer.

The policy implications follow directly. For Keynesians a deflationary gap calls for expansionary demand-side policy, and the cost of doing nothing is persistent cyclical unemployment. For monetarist/new classical economists demand-side policy is at best unnecessary, because output returns to Yf by itself, and at worst inflationary; lasting increases in output require supply-side policies that shift LRAS.

The choice of model also decides who bears the cost of a mistake. If the automatic adjustment is slow, the cost falls on the workers who stay unemployed while wages fall. If Keynesian stimulus is too large or arrives too late, the cost falls on households whose wages and savings lose value as prices rise.

Let AD be P = 140 − 0.5Y and SRAS be P = 20 + 0.5Y. Setting them equal gives 120 = Y and P = 80. Let potential output be 140, so current output is 20 below potential.

• AD becomes P = 150 − 0.5Y: 150 − 0.5Y = 20 + 0.5Y gives Y = 130, P = 85. Output and the price level both rise.
• From the initial state, SRAS becomes P = 30 + 0.5Y: Y = 110, P = 85. The same final price level comes with lower output.
• LRAS moves from Y = 140 to Y = 155. The long-run equilibrium on the unchanged AD curve is now at Y = 155, where P = 140 − 0.5 × 155 = 62.5, down from P = 70 at Y = 140. The short-run equilibrium stays at Y = 120, P = 80 until SRAS shifts, so the gap between actual and potential output widens from 20 to 35.
• Automatic adjustment at unchanged AD towards Y = 140 needs P = 70 and an SRAS that passes through that point, for example P = 0.5Y. The equations say nothing about how long the adjustment takes.

Dearer imported energy raises unit costs across the economy, so SRAS shifts left. With AD unchanged the model predicts a higher price level and lower real output, which is the combination the Bank of England described in November 2022: inflation in double figures together with a projected fall in GDP. The case illustrates the mechanism. It does not show that every price rise in 2022 had one cause, and the Bank also pointed to a tight labour market at home.

The case also shows why a cost shock is hard for demand-side policy. Raising Bank Rate reduces AD, which lowers inflation over time but adds to the fall in output. Supporting AD would have protected output and added to inflation. Neither produces more energy.

• AD = C + I + G + (X − M). Transfer payments are not part of G; imported consumption raises C and M together.
• A change in the price level is a movement along a curve; a change in a non-price determinant shifts the curve.
• A rise in AD raises real GDP and the price level on an upward-sloping SRAS. An adverse cost shock shifts SRAS left: real GDP falls and the price level rises.
• A rightward LRAS shift raises full-employment output (Yf1 to Yf2). In the long run real GDP is higher and the price level lower, provided demand is sufficient. Show SRAS moving with it, or AD rising, to explain how actual output gets there.
• A deflationary (recessionary) gap is output below Yf; it does not prove that prices are falling. Disinflation means slower inflation; deflation means a sustained fall in the price level.
• Monetarist/new classical: flexible wages return output to Yf, so in the long run a change in AD alters only the price level. Keynesian: wages are sticky downward, so a deflationary gap can persist and a rise in AD can close it without inflation while there is spare capacity.