New study highlights the real cost of political interference in banking

GAINESVILLE, Fla. – Banks across the world have long faced a common threat – political interference. Previous studies have shown that the world average of government ownership of banks is almost 42 percent, putting banks at risk of political capture.

Following the 2008 financial crisis, when many governments stepped in to save their banks from crisis, deep political divides formed and caused political interference to become the number one industry concern among bankers, non-bankers and regulators, according to a study by PwC and the Centre for the Study of Financial Innovation.

Aside from the ethical implications of political interference, Assistant Professor of Finance Nitish Kumar questioned another troublesome aspect of this relationship between politicians and banks – the cost.

“A growing literature examines how political favors arise through the banking sector in the form of preferential lending to particular groups of borrowers,” Kumar said. “However, very little is known about the real cost of such interference.”

In a new study, “Political interference and crowding out in bank lending,” Kumar found that politicians in India influence banks to increase lending to farmers before state elections at the cost of lending to manufacturing firms. To abstract away from macroeconomic factors that might influence bank lending, Kumar compared states with an upcoming election to neighboring states with no upcoming elections in a given year.

Specifically, Kumar found that the average agricultural sector lending during election years was 9.4 percent higher than during non-election years, while lending to manufacturing firms during the same period was 2.7 percent lower compared to non-election years. Such lending distortions were greater in electoral districts where farmers had more political weight, where election contest was expected to be tight and where incumbents had more influence over banks.

“These cross-sectional results lend creditability to the preferred explanation for observed behavior: politicians influence banks to lend more to farmers before elections in order to win their votes,” he said.

While the initial lending was a boost to politicians’ agricultural constituents, Kumar found that those in the areas of higher lending before a state election were more likely to be bailed out during the financial crisis of 2008.

“This suggests that even the favorably targeted segment of voters incur a long-term cost of political interference,” he said.

With higher lending rates to the agricultural sector, another important cost Kumar found was the effect of the credit squeeze on manufacturers. In finding a funding loss of 2 percent during election years, Kumar details that firms are forced to cut investments by 3.6 percent, thus pushing production in election years down 1.9 percent compared to non-election years.

He also notes that lower rates of production have an adverse effect on the utilization of fixed assets, causing capital utilization to fall by 1.9 percent in election years. This decline in capital utilization causes cascading effects on the industry and country as a whole.

“A fall in plant utilization rates during elections renders some productive capacity idle,” he said. “I estimate that had firms maintained their utilization level during the election year and put all their productive resources to use just like in the off-election years, the sector would have added another 0.15-0.20 percent to India’s GDP.”

Had India’s GDP been higher, it could have made a difference in a secondary finding from Kumar’s study – how productivity differences affect economic growth. While much research has attributed differences in output between workers to production technology available to firms in developed versus developing counties, Kumar provides support for another factor researchers have been investigating – capital misallocation.  

“Huge differences in output per worker…has been largely attributed to differences in firm-level production technology,” he said. “More recently, the focus has shifted to how misallocation of resources across firms, including differential access to bank loans, can affect aggregate total factor productivity. The evidence in this paper provides an example of such misallocation.

“By fixing such distortions in the banking sector, we might be able to put resources into more productive use.”

“Political interference and crowding out in bank lending,” is forthcoming in the Journal of Financial Intermediation.