What Was the Reinhart-Rogoff Excel Error?

In 2010, two Harvard economists published a paper that influenced austerity policies worldwide. Years later, a graduate student discovered the conclusions rested partly on a basic Excel mistake.

Contents:

1. Quick Answer

2. What the Paper Actually Claimed

3. What the Excel Error Was

4. Why It Mattered

Quick Answer

In their 2010 paper "Growth in a Time of Debt," Carmen Reinhart and Kenneth Rogoff argued that countries with public debt above 90% of GDP experienced sharply negative economic growth. The conclusion was partly driven by an Excel formula that excluded five countries from the calculation by accident, because the authors had not extended their AVERAGE range to cover all the rows in the dataset. A 2013 replication attempt by University of Massachusetts graduate student Thomas Herndon uncovered the error.

What the Paper Actually Claimed

Reinhart and Rogoff's central finding was that high debt loads were toxic to growth. Specifically, they reported that countries with debt-to-GDP ratios over 90% had a mean real GDP growth rate of -0.1%. This was a striking number, and it gave political cover to policymakers in the U.S., UK, and Europe who were pushing for spending cuts in the aftermath of the 2008 financial crisis.

The paper was enormously influential. Paul Ryan cited it in budget proposals. The IMF referenced it. UK Chancellor George Osborne used it to justify austerity measures.

What the Excel Error Was

When Herndon requested the original spreadsheet to replicate the results, Reinhart and Rogoff shared it. He found three separate problems, but the most embarrassing was the spreadsheet error.

The authors had organized their data with countries in columns and years in rows. They then used an AVERAGE formula to calculate mean GDP growth for the high-debt category. The formula looked something like this:

= AVERAGE(L30:L44)

The problem was that the data for that column ran through row 49. The range stopped at row 44, so five countries (Australia, Austria, Belgium, Canada, and Denmark) were simply left out of the average entirely. It was a drag-to-fill mistake: when copying the AVERAGE formula across country columns, the range wasn't extended to capture all the rows.

When Herndon included the missing countries and corrected the other issues (a selective exclusion of some post-war data, and a non-standard weighting method), the average growth rate for high-debt countries came out to +2.2%, not -0.1%.

That is the difference between "high debt causes economic contraction" and "high debt is associated with slightly slower but still positive growth." Completely different policy implications.

Why It Mattered

The Excel error alone did not invalidate every finding in the paper, and Reinhart and Rogoff have argued that their broader research across more datasets still supports a link between high debt and slow growth. The other two methodological issues Herndon identified were arguably more consequential than the spreadsheet mistake itself.

But the AVERAGE range typo became the story, because it was so tangible. It is the kind of mistake any Excel user can picture making, and it is a vivid reminder that a single misplaced cell reference in a spreadsheet can have consequences far beyond the screen.

The episode is now a standard cautionary tale in data science and economics courses. It shows why audit trails matter, why formulas should be checked against the full extent of the data, and why peer reviewers should request raw files alongside written results.

A quick way to protect against this type of error is to reference entire columns (like L:L) or use structured table references that expand automatically when rows are added, rather than hardcoded ranges that silently fall short.