The US dollar is widely used in transactions all over the world. Does this widespread use make inflation behave differently in the US than in other countries?
To investigate this issue, I will look at the data in the light of a very widely used theory of inflation.
The most widely used theory of inflation in the long run says that inflation occurs when too much money chases too few goods. The reasoning is simple: The productive capacity of a country depends exclusively on non-monetary factors – such as its endowment of productive resources, its technology, its policies and its institutions. When more money is printed and injected into circulation by a country’s monetary authorities, it makes absolutely no difference to the country’s output of goods and services. So, there’s more money to spend, but on an unchanged output of goods and services. Consequently, people bid up the prices of the unchanged output of goods and services.
The theory can be taken to the data in the form of the following equation: predicted inflation = growth rate of the quantity of money - growth rate of inflation-adjusted output. (For sticklers, this is the quantity theory of money with the assumption that the velocity of money is constant.)
It is easy to get annual data from the World Bank – for various countries and for several decades for each country – on:
(For most – though not all – countries, the data are from 1960 to 2020.)
Then, it is straightforward to calculate the annual growth rates of our three variables for each country. It is then simple to calculate predicted inflation = growth rate of the quantity of money - growth rate of inflation-adjusted output for each country and compare it to actual inflation = growth rate of the overall level of prices.
library(tidyverse)
library(WDI)
library(plotly)
mydata <- WDI(indicator = c("GDP" = "NY.GDP.MKTP.KN", "CPI" = "FP.CPI.TOTL", "money" = "FM.LBL.BMNY.CN"))
# The data download is blissfully automatic, but it takes a while.# The downloaded data includes various country groups. I need to delete them.
country_groups <- c("Middle East & North Africa", "IDA & IBRD total", "East Asia & Pacific", "Europe & Central Asia", "Sub-Saharan Africa (excluding high income)", "Sub-Saharan Africa", "Africa Eastern and Southern", "Africa Western and Central", "Latin America & Caribbean", "Euro area", "High income", "Heavily indebted poor countries (HIPC)", "IBRD only", "IDA total", "IDA blend", "IDA only", "Latin America & Caribbean (excluding high income)", "Least developed countries: UN classification", "Low income", "Lower middle income", "Low & middle income", "Middle income", "Middle East & North Africa (excluding high income)", "Upper middle income", "North America", "Pre-demographic dividend", "Early-demographic dividend", "Late-demographic dividend", "Post-demographic dividend", "Latin America & the Caribbean (IDA & IBRD countries)", "Middle East & North Africa (IDA & IBRD countries)", "East Asia & Pacific (IDA & IBRD countries)", "South Asia (IDA & IBRD)", "Sub-Saharan Africa (IDA & IBRD countries)", "Europe & Central Asia (IDA & IBRD countries)", "Small states", "Other small states", "OECD members", "European Union", "Fragile and conflict affected situations", "Central Europe and the Baltics", "Arab World", "World", "East Asia & Pacific (excluding high income)", "Europe & Central Asia (excluding high income)", "South Asia")
mydata2 <- mydata |>
filter(!(country %in% country_groups)) |> # The country groups are gone
group_by(country) |>
na.omit() |>
summarise(year.beg = min(year),
year.end = max(year),
GDP.beg = GDP[year == year.beg],
GDP.end = GDP[year == year.end],
GDP.growth = (((GDP.end/GDP.beg)^(1/(year.end - year.beg)) - 1) * 100),
CPI.beg = CPI[year == year.beg],
CPI.end = CPI[year == year.end],
inflation = (((CPI.end/CPI.beg)^(1/(year.end - year.beg)) - 1) * 100),
money.beg = money[year == year.beg],
money.end = money[year == year.end],
money.growth = (((money.end/money.beg)^(1/(year.end - year.beg)) - 1) * 100),
inflation.predicted = money.growth - GDP.growth
) |>
mutate(USA = (country == "United States"))Each dot represents a country. The horizontal axis shows the inflation rate predicted by our theory: predicted inflation = growth rate of the quantity of money - growth rate of inflation-adjusted output. The vertical axis shows the actual inflation.
Were our theory perfect, every dot would have been on the line of equality. Sadly, that is not the case. But, for an economics theory, this is pretty good. Most dots are close to the line of best fit. And for most countries the actual inflation is not that different from the inflation predicted by the theory.
fig1 <- ggplot(mydata2, mapping = aes(x = inflation.predicted,
y = inflation, color = USA, label = country)) +
geom_point(alpha=0.3) +
geom_smooth(method = "lm", se = FALSE) +
labs(
y = "Inflation (%)",
x = "Money Growth (%) - GDP Growth (%)"
) +
annotate("text", x = mydata2[mydata2$country == "United States", ]$inflation.predicted, y = mydata2[mydata2$country == "United States", ]$inflation, label = "United States") +
scale_y_continuous(trans='log2') +
scale_x_continuous(trans='log2') +
theme_minimal() +
theme(legend.position="none")
ggplotly(fig1)Finally, is the United States different? Is the link between the quantity of money and inflation especially loose for the United States because of the widespread use of the US dollar in cross-country trade?
No, the blue dot representing the United States is very close to the line of best fit. It’s a very typical country as far as inflation is concerned.
A look at the data shows that the theory of long-run inflation that most economists rely is quite good as an explanatory device. The excess of the rate at which money increases in an economy over the rate at which the economy’s inflation-adjusted output increases is a good predictor of inflation.
The United States is subject to the same inflationary forces as other countries despite the wide use of the US dollar.