Inflation is usually treated as a problem of excessive demand. Too much money and credit chase too few goods, prices rise, and the Federal Reserve responds by increasing interest rates to make borrowing more expensive, reduce spending, and slow investment. That logic is persuasive when inflation is demand-driven. But it becomes less straightforward or even logical when prices rise because the economy has suffered an exogenous shock triggering a loss of real supply. In the case of a supply shock that pushes up prices, the central bank responding to inflation by increasing interest rates, is not restraining inflation but quite possibly exacerbating it.
Consider a deliberately simple economy with two goods, A and B. If a supply shock, assume an asteroid hits a wheat field or a widget plant, making good B more expensive. Inflation goes up and households must spend more to obtain it. Their real income falls, and they respond by reducing consumption of A. If prices are flexible, weaker demand for A should put downward pressure on its price or at least slow its normal rate of increase. The initial shock may still raise the overall price level because the decline in A’s price does not fully offset the increase in B’s price. But the economy is already performing part of the adjustment on its own: scarcity raises the price of B, and the resulting loss of purchasing power reduces demand elsewhere.
The central question for a central banker is whether monetary policy can address supply side driven inflation. In straight cost benefit terms, the central banker must estimate before deciding to raise rates which number is smaller: the loss of output for the estimated duration due to demand suppression from higher interest rates or the loss of demand due to the risk of accelerating inflation above the post-supply shock price.
To simplify, the central banker can immediately raise interest rates and lower output immediately after the asteroid strikes the plant. However, the plant will need capital to rebuild which was just made more expensive and now likely will take longer to rebuild extending the duration of the interest rate and output shock. Raising rates immediately lowers output and extends the duration of the output loss.
Conversely, due to a step up in inflation, consumers may understand the asteroid problem or panic interpreting this as a structural shift in inflation expectations taking inflation expectations not just to the post supply level, say going from 2% pre-asteroid to 3% post-asteroid but to 4% or 5% if inflation expectations change.
The 3% inflation rate cannot be avoided due to the immediate supply shock, what the central banker cannot know before deciding on interest rates is how consumers and business will react and adjust their habits. If the central banker leaves rates unchanged and consumers adjust their expectations to only the 3% level, the central banker made the correct decision.
However, if consumer expectations change to say 5%, the central banker must now raise rates and importantly not just to the level they would have raised them immediately after the asteroid strike but above that level because inflation is not just at the post asteroid strike level of say 3% but is now at say 5% and stickier.
So the primary question is how consumers respond. Whether the shock spreads beyond supply shock good depends on how consumers understand it. If households recognize that B is more expensive because of a specific, identifiable disruption, they may adjust their behavior without concluding that every price in the economy will continue rising. They buy less of B, substitute where possible, and reduce spending on A and other discretionary goods to absorb the loss of real income.
Flexible prices make that adjustment easier: weaker demand for A can slow its price growth, while higher prices in and around B signal where resources are scarce and where additional supply is most valuable. The result may still be a higher overall price level, but not necessarily a self-sustaining inflationary process. The danger arises when consumers interpret the price of B not as a relative-price signal but as evidence that the purchasing power of money is broadly deteriorating. They then bring purchases forward, demand generalized wage increases, and encourage firms to raise prices beyond the sectors directly affected by the original shock.
Let us make this example more tangible. The large majority of changes in producer prices came from one specific product diesel fuel. The spike in diesel fuel has nothing to do with any surge in trucking or organic shift in the economy but everything to do with a variety of events outside the United States from Ukraine bombing Russian petro-chemical refineries to events in the Middle East. United States monetary policy has no has no direct ability to repair the physical supply disruption in the middle of Siberia. Most importantly, the price increase is not linked to expansionary fiscal or monetary policy generating excess demand. Prices are rising solely due to exogenous shocks akin to an asteroid strike.
To see the policy problem clearly, hold the initial supply shock constant. Suppose the economy was experiencing 2% inflation before the disruption, and the loss of supply pushes inflation to 3%. For this example, assume that increase occurs whether or not the Federal Reserve changes interest rates. The first percentage point is the consequence of the damaged supply itself. It is not something higher rates can eliminate. The United States and the Federal Reserve how no influence financially or otherwise Russian diesel refining capabilities.
The policy question concerns what happens next. If consumers understand the shock and adjust their spending, inflation will likely remain close to the 3% supply-shock path. If they instead conclude that prices will continue rising everywhere, they may bring purchases forward, demand higher wages, and encourage firms to raise prices beyond the sectors directly affected. Inflation could then move from 3% to 4%. The cost of holding interest rates steady is not the entire increase from 2% to 4%; it is the additional inflation above the 3% path created by the supply shock.
That is the point at which the Federal Reserve’s usual response becomes more difficult to evaluate. Raising interest rates will not create more wheat, repair a damaged refinery, or bring another widget factory online. It can only reduce demand, including demand for Good A, but the original shortage of Good B remains whether interest rates remain low or are increased.
Higher rates also make it more expensive to finance the rebuilding of the damaged plant, they may delay the return of supply and increase the price that the replacement facility must charge to earn an adequate return.
In other words, given a negative supply shock that increases inflation, raising interest rates not only has no impact on the price level it may actually exacerbate a return to equilibrium by making it harder to correct the price shock and return the economy to a price equilibrium.
The economy may therefore experience the worst of both effects: less consumption of A in the present and no faster recovery in B. Higher rates can be justified only if the benefit of preventing inflation from moving above the supply-shock path is greater than the lost output, weaker demand, and delayed supply recovery that the policy itself creates.
The policy decision can be expressed through a simple numerical example. Suppose the supply shock raises inflation from 2% to 3%, and the Federal Reserve leaves interest rates unchanged. Assume there is a 40% chance that consumers and firms misread the shock, inflation rises from 3% to 4%, and the resulting uncertainty and eventual policy correction reduce real GDP by 2% over the next two years. The expected economic loss from allowing that inflationary spiral would be 0.8% of real GDP: a 40% chance of a 2% loss.
Now compare that with raising rates immediately. Suppose higher rates reduce consumption and investment enough to lower real GDP by 0.7%. Suppose they also delay the rebuilding of the capacity producing B, creating an additional 0.5% loss over the same period. The total cost of tightening would then be 1.2% of real GDP. Under these assumptions, holding rates steady produces the smaller expected loss: 0.8% rather than 1.2%.
The calculation could easily point in the other direction. If there were a 70% chance of an inflationary spiral, and the resulting loss of real GDP were still 2%, the expected loss from holding rates steady would rise to 1.4%. In that case, raising rates would impose a smaller loss than allowing the spiral to develop. The point is not that one answer is always correct. The point is that the decision depends on the probability and cost of inflation above the supply-shock path, compared with the certain output loss and delayed supply recovery caused by higher rates.
Time matters as well. A rate increase may reduce demand quickly, while the damage to investment and supply recovery appears gradually actually extending the duration of the reduction in output. Conversely, expectations may remain stable for several months before suddenly becoming embedded in wages and prices. The Federal Reserve therefore has to compare not only the size of the two losses, but also their probability, timing, and persistence. It should raise rates when the expected damage from generalized inflation above the supply shock jump is greater than the economic loss caused by suppressing demand in response to a shortage it cannot directly repair.
A few final points in closing. First, given the recent history since 2020 of moving inflation expectations a strong argument could be made that consumers and business are much more concerned about prices that might trigger the increase in inflation expectations the central banker fears. That concern is not unfounded. There is also, however, a strong argument that consumers are seeing broad based ongoing price moderation and understand that the price increases are tied to a very narrow subset of products that are exogenous and will be less prone to materially change their expectations. It is not clear which expectation might dominate the consumer psychology and that might in itself be a finding.
Second, the current supply side price shocks are materially different than the Covid induced supply side price shocks. The Covid supply shock was broad based whereas the current supply side shock is very narrow. Furthermore, the return to a good equilibrium in this case will be very different for as for instance, key products are literally in the middle of war zones.
This leads to multiple implications such as likely differences in consumer expectations and economic recovery especially when coupled with the large fiscal stimulus. Here there is effectively no large acceleration of fiscal pumping and the price increases stem fro very narrow goods. While the exact estimates will differ, research found that approximately 50% of inflation was supply side driven. Currently there is no fiscal acceleration and the supply side shock is very narrow.
This implies that where as the previous example with fiscal stimulus did call for some degree of demand restraint, it is less clear that is the case here.
It is taken as a given handed down from God that elevated inflation requires an increase in interest rates, however, there is good theoretical reason to believe that in supply driven inflation shocks raising interest rates can both exacerbate and prolong inflationary shocks rather than returning both the macro and micro economy to equilibrium. Due to the specific sectors driving the inflation, it may be worth considering whether rate increases is the right medicine.