Cars and chips
Rich countries have seen several months of high inflation readings. The Fed and the ECB are affirming their determination to stay the course with asset purchases and low policy rates. But central banks are increasingly under pressure to justify these efforts in the face of price rises.
To my mind, the debate suffers from the lack of a good theory of prices. Inflation hawks' worries about price rises becoming embedded in expectations seems not to have much bearing in the present decade. I am more sympathetic to the dovish view that price rises have to do with supply-chain hiccups, and so are likely to be transitory, but it is hard to offer an argument based on economic theory that this is the case.
As a step toward the theory that seems to be missing, this post looks at prices for used cars, a particular example that illustrates how we might think about prices more generally.
Vehicle prices and supply chains
This graph shows the inventory of new and used cars and trucks owned by US auto dealers, measured as a multiple of monthly sales. Cars depreciate as they are used, so there is an ongoing flow demand for vehicles. The graph says that dealers usually keep between two and three months' worth of sales on hand. That number was closer to three until mid-2017, then had been falling steadily to around two months by the end of 2020. There also seems to be a seasonal dip at the end of each year.
As the COVID-19 pandemic hit in early 2020, car sales stopped and inventories spiked. As the pandemic went on, the excess was worked off and stocks leveled off at just below two months' worth of sales.
Then, starting in 2021, inventories went through the floor. At the same time, the price of cars has jumped up, which is contributing to the high readings of the overall consumer price level of recent weeks. This is apparent in the component price indexes:
I show the series going back to 2016 to highlight how unusual the recent price rises have been. Note that the biggest jump is in the price of used vehicles (blue line). The price of new vehicles (red line) has also risen, but nowhere near as dramatically.
What is going on? This piece from the FT has the goods: A year ago, computer chip makers asked auto manufacturers to forecast demand. Chip-making is a complex production process, and planning requires long lead times. Adjusting those plans is like turning a battleship: it can not be done quickly. Auto makers provided forecasts that assumed low demand for autos. Chip makers looked at those forecasts and shifted production away from the chips for cars. Instead they responded to increased demand for electronic devices.
But auto makers' forecasts turned out to be underestimated. So now the production of new cars is way below demand, hamstrung by the shortfall in chip supply. One consequence is the very low inventory of autos in the first graph above. Another consequence is the very high price of used cars in the second graph.
These pieces can be put together in a way that gives a bit more insight into price-making. In the second graph above, I have also shown the price level for semiconductors. It is quite flat over the entire period, including during the pandemic. But we know that there are supply shortages. This suggests that the market for semiconductors is not cleared by price: if demand is higher than supply, then the demand simply goes unmet. But cars are different: there the price does adjust to clear the market. Ask anyone who has needed to replace their car this year.
The diagram below tries to capture the whole story. On the left is the market for chips; on the right, the market for cars. For both cars and chips, inventories are measured going across and price going up and down. For each market there are two price curves, the price at which dealers sell (the higher ask price) and the price at which dealers will buy (the lower bid price). This is perhaps more intuitive for cars: if you sell your car to a dealer and then buy it back, you will pay a higher price than you received. That is the source of the car dealer's profit. Chips, on the left, work differently: only quantities adjust, not prices. (The bid price for chips is a bit abstract—a manufacturer "buys" chips, we could say, by buying inputs, including labor, and consuming capital to produce them.)
These diagrams capture the current situation pretty well. Chips are not available for car manufacturers because chip makers are still switching back from making other kinds of chips. That's reflected in shortages, not in higher prices for chips. But as a result, there is also a shortage of new cars. Because used cars can substitute for new cars, inventories of both are low. And because prices of cars respond to inventories, car prices are high. I've shown a vertical line near the low point of car inventory to indicate this. In effect, the high price of cars is clearing the market for chips.
Toward clearer thinking about prices
This is not yet a comprehensive theory of prices. But it seems to me that we need more attention to supply chains and market-making to get to a better understanding of inflation. More on this soon, I hope.