Confronting the Cost of War: The Political Economy of War Finance

Rosella Cappella, Principal Investigator

Introduction and Puzzle

How do states mobilize and pay for war? Without wealth, soldiers cannot be paid, weapons cannot be procured, and food cannot be bought. For the millennia that wars have been fought, ruling groups have had the task of acquiring resources to support their war efforts. Leaders must first decide where to get money. Should the state levy taxes, incur debt, or print? If leaders look outside their borders for finance, should they float debt on international financial markets or secure a direct loan from another state? Each alternative – taxation, domestic debt, printing, external funding – has different political and economic costs and benefits. Borrowing compounds the cost of war through high interest rates; printing can result in disastrous inflation; taxation combats high inflation and minimizes cost yet can be politically damaging; garnering money from abroad invites outside influence and fosters dependency. How do leaders weigh these costs and benefits?

This project explores the dynamics and determinates involved in war finance. The prevailing models in international relations assume that regime type dictates war finance: states that can borrow to finance war will and governments characterized by democratic institutions are most amenable to borrowing. These models, however, wrongly assume that leaders’ preferences as well as state capacity are constant. I challenge this set of assumptions. Institutions do not make decisions, people do. When financing a war, leaders are aware that their decisions have both political and economic costs. They have preferences that are shaped by their beliefs about inflation and the war. However, their preferences are bound by the ability of the state bureaucracy to extract revenue and cope with low currency reserves in the presence of a balance of payments problem. I argue that this interaction between preferences and capacity explains the extent to which a state finances a war by taxation, domestic borrowing, external funding, and printing. In sum, I find that the likelihood of a war financed by taxation increases when policymakers fear inflation, there is high support for the war, or the state is characterized by a high bureaucratic capacity to extract tax revenue. In contrast, wars will be financed by more indirect means such as borrowing, at home or abroad, and printing when support for the war is low or the state has a low bureaucratic capacity. Finally, states will borrow from abroad when they face a balance of payments problem.

To address the puzzle, I created two data sets. The first encompasses all interstate wars from 1823 to 2003 over six months in duration. The second encompasses all interstate wars between 1945 and 2003, regardless of duration. The unit of analysis is war finance by participant, aggregated over the entirety of a war. I account for how much of the war’s cost was met by taxation, domestic borrowing, funding from outside the state’s borders, printing, and plunder. Moreover, I account for the presence of borrowing and the location of borrowing, specifically, if the state borrowed from either the United States or Soviet Union. Primary and secondary sources allowed me to code a majority of observations in the data set. During the interwar years, the
League of Nation’s statistical publications is an excellent source, particularly for the various Sino-Japanese wars. In the post-World War II era, primary sources include documents collected from the Foreign Relations of the United States (FRUS). During the Cold War, the United States’ intelligence estimates from various agencies provide a plethora of information for the various “proxy wars.” Specifically, intelligence estimates are helpful to understand funding for the various Southeast Asian wars and Arab-Israeli wars.
Preliminary Findings
Preliminary findings have yielded fascinating descriptive statistics as well as indicative correlations. There are four prominent war finance trends for wars over six months long. First, war finance rarely takes the form of “tax versus borrow.” The majority of states also engage in printing or external extraction to confront the cost of war. Second, most wars are not financed by taxation.

If taxation is a component of a state’s war finance strategy, it generally covers only a very small percentage of the war cost. Thus, borrowing is the dominant form of war finance. Third, the location of borrowing has changed over time. Since the end of World War II, there has been a shift from borrowing from within the state to borrowing from abroad. Last, the presence of printing and plunder have also decreased over time.

Preliminary results also suggest findings consistent with my theory. As a state’s administrative capacity to extract tax revenue increases, the more likely a higher percentage of the war will be paid for out of tax revenue. Bureaucratic capacity, however, does not appear to be statistically correlated with the percentage of the war financed via domestic or foreign debt. In addition, states that have a low domestic production capacity and that are forced to procure imports from outside their boarders are more likely to finance a larger percentage of the war from external sources. In regards to leaders’ preferences, when states experienced at least 10% inflation in the previous interstate war, they are more likely to finance the current war with higher levels of taxation. This finding suggests that there is a learning mechanism between wars, as previous war finance affects current war finance. Finally, contrary to conventional wisdom, I find no orrelation between regime type and the percentage of the war financed by taxation or domestic debt. Democracies are no more likely to finance their wars by taxation or domestic debt than non-democracies. However, there is a strong relationship between regime type and the percentage of the war paid for by printing. Non-democracies are more likely to print to finance a war than democracies. This finding, in conjunction with the finding above that non-democracies are more likely to finance their wars via external funding, suggests that non-democracies may be less likely to ask their societies to fund a war.

Until now, there has been no existing statistical study of war finance. To date, the closest things to a war finance database are data sets measuring resource potential, the supposed pool of money that policy makers could draw upon to confront the cost of war. While these data sets are important building blocks to understanding a state’s potential financial and, consequently, military power, they do not address how states actually pay for war. Thus, the predominant tests of war finance theories have been qualitative in nature. The data set presented in this proposal significantly advances the study of war finance by providing the first quantitative analysis.

As the study of war finance has implications for scholars and policymakers, the data set is constructed to the standards and format set by the preeminent database in international relations, the Correlates of War Project in order to promote maximum accessibility and usability. Specifically, the data set is constructed using the interstate wars and coding listed in the Correlates of War Project (COW).

By addressing the source and nature of war finance as well as how states cope with the economic effects of the policy chosen, the war finance data set has implications for political scientists, economists, finance and war scholars, and policymakers. For example, this data set will allow scholars of various disciplines to analyze, discuss, and recommend solutions to one of the most pressing policy challenge to date, how to finance America’s wars. The current trend of American war finance emphasizes borrowing from both at home and abroad. The financing of wars in Iraq and Afghanistan – the longest war in U.S. history – were the first major wars to be funded entirely by borrowing. The data set will allow scholars to understand how the financing of these costly wars affects the national and global financial system, the larger economy and society at large. The data set also allows scholars to address leadership decisions and risk management. How do leaders weigh the political and economic cost when deciding a war finance policy and how do they cope with the consequences once the policy is enacted? Moreover, given the expanding deficit, the data set will allow scholars to understand the causes effect of war finance on a state’s sovereign debt.