OK

'Inflation is a tax on the poor' is a meme often used to justify policies which hurt the poor. Is inflation really the cruelest tax?

Most often it is conservatives and libertarians who invoke the possibility of hyperinflation to justify austerity measures and cutting government spending on programs that benefit the poor and the middle class. Since conservatives in particular seem less concerned with the possibility of hyperinflation when trillions of dollars are pumped into backdoor bailouts of the banks or wasted on foreign military adventures or subsidies for Big Agriculture and Big Oil, I'm a bit suspicious when they claim to be protecting the poor from future inflation by cutting government services and transfer payments.

So what's the evidence? Does inflation hurt the poor the most? Find out underneath the mating snails.

Historical data on inflation is available from many sources. I grabbed data from the Federal Reserve on annual average CPI from 1913 through 2010 (1982-1984=100).

Data from the Census Bureau on the Share of Aggregate Income Received by Each Fifth and Top 5 Percent of Households can be found here.

The correlation coefficients between CPI and share of aggregate income of households, from the beginning of the series in 1967 to 2009, are as follows:

Poorest 20%       0.707

Lower middle       0.582

Middle                0.601

Upper middle       0.660

Richest 20%     (-0.635)

For the poorest quintile the correlation of income share with inflation over the period 1967 to 2009 was strongly positive. That is, periods of high inflation, such as the 1970's, coincided with periods of a stable or slightly growing share of national income for the poorest 20%. Conversely, during the long period of low inflation starting in the early 80's, the share of total income of the poorest 20% persistently declined. The same is true for the middle three quintiles, which also had strongly positive correlations. Only for the top quintile, the top 20% of households, was the correlation between inflation and income share strongly negative (-0.635). For the richest households periods of high inflation corresponded to periods of declining income share, and periods of low inflation corresponded to periods of growing income share.

All of these results are statistically significant (p<.001). There is less than one chance in a thousand that comparison of sets of unrelated variables would, by chance, show as much of a relationship.

The Census Bureau also has data on the share of income by quintile of families, instead of households, and that data goes back to 1947. Over that longer period a similar pattern, though less pronounced, still holds. Comparison of income share of families with the CPI data for the period 1947 to 2009 yields the following correlations:

Poorest 20%       0.486

Lower middle       0.162

Middle                0.199

Upper middle       0.294

Richest 20%     (-0.272)

Again it was the poorest 20% whose share of aggregate income was most closely correlated with inflation, and again the income share of the richest 20% was negatively correlated with inflation.

The correlations for families for the lower middle and middle quintiles are too small to meet the test for statistical significance, and the significance of the results for the upper middle and top quintiles is modest (p<0.05, or less than one chance in twenty that the correlation is the result of chance). But, as with the household data, the positive correlation between inflation and the income share of the poorest 20% is highly statistically significant (p<0.001).

Based on these data sets from the Fed and the Census Bureau I'm inclined to conclude that inflation hurts the rich and helps the rest.

This conclusion is supported by analysis of historical data on the income share of the top decile in other countries. I found a link to such data here. For the years (as noted) where I had continuous data for both, I calculated correlations between CPI data (sourced as noted) and income share of the top 10% for a few countries as follows:

Australia       -0.414          (p<0.01)           1949-2003     CPI data here    

Canada        -0.331          (p<0.01)            1941-2000     CPI data here

France         -0.294          (p<0.1)             1956-1998     CPI data here

Germany       -0.170          (n/a)                1961-1998     CPI data here

Japan           -0.051          (n/a)                1956-2002     CPI data here

Spain           -0.414          (p<0.01)            1957-2002     CPI data here

UK               -0.506          (p<0.001)          1961-1998     CPI data here

US               -0.362          (p<0.001)          1917-2004     CPI data here

The results for Japan and Germany are not statistically significant, and the significance of the other results varies as shown. I note that all of the correlations of the income share of the top 10% in the countries tested are negative. Since the correlation for income share of the bottom 90% is necessarily of equal magnitude and opposite sign as the correlation for the top decile, these results seem to indicate that inflation tends to increase the income share of the bottom 90% at the expense of the richest 10%.

I have no training in economics or statistics, so all of the above analysis may be wrong. I may have made bad choices about what data sets to compare, failed to control for different base years in inflation data, or misinterpreted the implications of the correlations. I'm on particularly shaky ground when making bold assertions about statistical significance. If I have gone astray in my choice or interpretation of data, or otherwise miscalculated, I hope someone will set me straight in the comments.

But if I'm not wrong then I contend that those who have propagated the dubious meme that inflation hurts the poor the most (I'm talking to you, Ron Paul) should be challenged to explain why historical data seem to indicate that the meme is just flat wrong.

EMAIL TO A FRIEND X
Your Email has been sent.