Defining Ideas

The Politics Of The New Deal

Tuesday, July 19, 2016
Image credit: 

Editor’s note: This is a summary of the paper, “The New Deal and Executive Control of the Distribution of Federal Funds Across States,” presented at the Hoover Institution Executive Power and the Rule of Law Conferences in March and June 2016. Price Fishback, the Thomas R. Brown Professor of Economics at the University of Arizona, is the corresponding author at pfishback [at] Valentina Kachanovskaya just completed here Ph.D. in Economics at the University of Arizona and may be reached at kachanovskaya [at] We thank Charles Calomiris for his encouragement and suggestions in starting into this project. We would also like to thank Alex Field, John Wallis, and the participants at the conferences for their insights and comments.

There has been an extensive discussion in the study of administrative law about the transition from direct Congressional control of regulation of economic activity and spending to more control of such activity through the executive and administrative bodies (DeMuth Sr. 2013, Epstein 2013; Price 2015). One of the key areas where the Constitution gives Congress control of federal government economic activity is in the “power of the purse.” A key question that has arisen in many recent debates is the extent to which increased executive discretion might be abused for political purposes that favor the executive. For example, if the President were given greater authority to allocate spending across the states, would he or she favor swing states with more spending to curry favor with their voters? To address that question, we examine changes in Congressional legislation that controlled the distribution of federal funds across the states during the 1920s and 1930s and the impact of increased executive discretion during the New Deal on the allocation of funds.

In the 1920s the lion’s share of federal spending distributed across states was for highways, rivers and harbors, veterans, and defense. Congress maintained relatively tight control over the distribution of spending across states. The highway program established a specific formula for the maximum federal spending in each state and required matching state expenditures. Congress directly approved all river and harbor projects of the Chief Engineer of the Army and the irrigation projects of the Bureau of Reclamation. Rules were established for the relatively small amounts of up to $1 to $2 million distributed for the Shepard-Towner child health program, agricultural experiment stations, agricultural extensions, and agricultural and mechanical colleges.

As the economy sank into the Great Contraction, President Hoover and the Republican-dominated Congress between 1929 and 1932 ramped up annual real outlays of federal spending by 88 percent, largely through existing government programs. In January 1932, they created the Reconstruction Finance Corporation (RFC) to provide emergency financing and gave it authority to make loans to aid in financing agriculture, commerce, and industry. They expanded the lending authority to include making loans for relief to state and local governments in July of that year. The claims of emergency appear to be strongly correlated with the loosening of Congressional control over the distribution of funds.

In the election of 1932, the economy’s continued decline contributed to a landslide in favor of Franklin D. Roosevelt and large majorities of Democrats in both houses of Congress. In the first hundred days of the Roosevelt administration, the federal government consistently used the emergency language to establish the “New Deal,” a large-scale experiment in which dozens of new programs and agencies were established in an attempt to resolve a wide range of specific problems in the economy. Congress declared emergencies in many areas and gave the President extensive discretion over the allocation of funds across states. This was particularly true for monies spent on conservation work by young adults in Emergency Conservation Work (later the Civilian Conservation Corp (CCC)), the Federal Emergency Relief Administration (FERA), and the Civil Works Administration (CWA). The FERA and CWA were strongly intertwined and distributed 45 percent of the federal money that went to the states in fiscal years 1934 and 1935.

By 1935 FERA and CWA administrator Harry Hopkins had become dissatisfied with the extent of control that he had over the distribution of funds within the states by the FERA. As a result, Roosevelt and Congress negotiated over the future distribution of relief funds. The outcome was a compromise in which the Roosevelt administration gained greater control of the emergency relief funds distributed by the WPA but Congress established tighter rules over the distribution of Aid to Dependent Children (ADC), Old-Age Assistance (OAA), and Aid to the Blind (AB) under the Social Security Act of 1935. The new rules allowed the states to choose the benefit levels in each program, while the federal government provided matching funds of up to $15 per month per relief case. As a result, the states’ benefit choices largely determined the distribution of federal funds in these programs (Wallis, Fishback, and Kantor 2006).

The questions we address here are are: How did the distribution of funds across states change when Congress relaxed control over the allocation of some spending programs during the New Deal? Did they follow the rhetoric in President Roosevelt’s speeches and increase the amounts of federal money spent to promote relief, recovery, and reform? Or did they use the funds to promote the reelection of President Roosevelt, as the New Deal’s critics claimed? Can one establish an empirical connection between increased executive discretion over the allocation of funds during the 1930s and the more politicized uses of funds?

A large literature has developed to address these questions by examining the variation in the per capita distribution of New Deal funds across states and counties during the New Deal. The studies find that the distribution of per capita New Deal funds was influenced by efforts to promote relief and recovery, to improve federal land, to reward areas with clout in Congress, to respond to the demands of the states, and to promote Roosevelt’s reelection by spending more where there were more swing voters.

The existing literature on the federal distribution across states has one large gap. It has not been able to address the central causal question of whether greater executive discretion resulted in more politicized spending. Did the New Deal’s cross-state distribution of funds follow similar patterns to those that had been pursued under prior administrations, or did it involve a major policy shift that reflected the ceding of power from Congress to the Executive? We address this question by assembling a new dataset that contains two types of information. First, we examined Congress’s enabling legislation for spending on major programs before and during the New Deal to determine which programs gave the President more discretion. Second, we dug into various government publications and collected annual information on federal spending by state from fiscal years 1924 through 1940, using regression analysis on the data to compare and contrast the factors that influenced spending before and during the two periods.

Our statutory analysis shows that Congress consistently described the New Deal legislation as emergency legislation designed to combat the Great Contraction. The use of the term calls to mind the necessity for speedy action and helped to justify Congress’s granting increased authority to the executive branch in the distribution of funds.

Our regression analyses of cross-sectional data and an annual panel data set included various controls, such as the extent of federal land, changes in income, the level of income, and several other factors. The results show that the major change between the Hoover and Roosevelt administrations was a move to a strong positive relationship between the amount of per capita funds and swing voting across the states.

This new increased responsiveness to the needs of swing states did not occur equally across all programs. The increased allocations to swing state voters largely appears through the distribution of funds from the relief programs, which were being administered by the federal government for the first time in American history. In both the Hoover and Roosevelt eras, Congress gave the President extensive discretion over the distribution of most of the relief funds. The RFC’s distribution of relief loans in fiscal year 1933 under Hoover, however, was not positively related with swing voting. In contrast, under the Roosevelt administration, the distribution of total relief was strongly positively related to swing voting in both the cross-sectional and annual panel analysis.

The story about the link between executive discretion and politicized spending becomes somewhat more complicated when we look at the individual relief programs. The strongest positive relationship with swing voting was found for the FERA and CWA relief programs under the First New Deal. During the second New Deal there was a positive but weaker relationship for the WPA. One might be tempted to infer that it was the new discretion awarded to the executive that allowed the Roosevelt administration to use the relief funds to attract more swing voters. That may be true, but our results also show that the SSA public assistance grants—which were not under the control of executive discretion—were also strongly correlated with swing voting. The executive had very little control over these grants because Congress set them up as matching grants and the states were the units deciding how much would be spent. This finding suggests that there may be other factors than executive discretion that caused funds to shift to swing states in the 1930s. Although our regression analysis controlled for many observable characteristics (such as income and employment), it is conceivable that omitted variables reflecting differences related to states’ circumstances may be driving some of the shift toward swing states in the FERA and CWA programs, too.

The statutory analysis shows that the decision rules for programs for veterans, reclamation dams, and rivers and harbors stayed the same in the 1920s and 1930s. In the cross-sectional regression analysis, none of the three were very responsive to swing voting in either period after controlling for federal lands and state grant demands, but the annual variation in the Bureau of Reclamation funds shows a stronger positive relationship with swing voting under the New Deal despite the absence of an explicit rule change. The relationships for highways are also mixed but in this case the Roosevelt administration was given more discretion because the National Industrial Recovery Act of 1933 largely kept the distribution formula from previous highway acts but eliminated the state matching feature. Without the state matching, the annual highway distributions under the New Deal were highly responsive to swing voting, but the cross-sectional analysis (which does not take account of the effect of the distribution formula) does not find the same strong positive relationship.

Overall, our results provide some evidence that the Roosevelt administration used its expanded discretion in some areas of relief, and possibly in spending on highways, to allocate more money where there were more swing voters. These patterns added fuel to the fiery criticisms of the New Deal as a means of getting Democrats reelected by its opponents. Yet, it remains a bit unclear whether executive discretion, or some other factor, is driving our findings about the shifts in funding toward the swing states. The state/federal public assistance programs under the Social Security Act had strong matching rules in which each state’s decisions largely determined how much was spent there. Executive discretion was not a factor in those allocations. Yet, the distributions across states in those programs also showed a strong increasing tilt toward swing voting states. Furthermore, in some areas where Congress gave Roosevelt more discretion, we do not see a strong relationship between spending and swing voting across states.

It remains to be seen whether those areas of spending might have been less susceptible to political pressure. Although we regard our conclusions as mixed with respect to the central question of whether executive discretion drove more politicized spending, our results point future research toward explaining the differences we have identified across programs in the cross-state changes in spending.


Anderson, G. M., and Tollison, R. D. 1991. ’Congressional Influence and Patterns of New Deal Spending, 1933-1939’, Journal of Law and Economics, 34 (April): 161-75.

Arrington, L. J. 1970. ’Western Agriculture and the New Deal’, Agricultural History 49 (October): 337-316.

Couch, James and William Shughart II. 1998. The Political Economy of the New Deal. New York: Edward Elgar.

Crafts, Nicholas and Peter Fearon (eds.) 2013. The Great Depression of the 1930s: Lessons for Today. Oxford: Oxford University Press.

DeMuth Sr., Christopher. 2015. “Can the Administrative State be Tamed.” Unpublished Working Paper, 69 pages.

Epstein, Richard. 2013. “The Perilous Position of the Rule of Law and the Administrative State.” 36 Harvard Journal of Law & Public Policy 36: 5-19.

Fishback, Price. 2015. “New Deal Funding: Estimates of Federal Grants and Loans Across States by Year, 1930-1940.” Research in Economic History 33 (2015): 41-109.

Fishback, Price.and Valentina Kachanovskaya, V. 2015. The Multiplier for Federal Spending in the States During the Great Depression. Journal of Economic History, Vol. 75: 125-162.

Fishback, Price, Shawn Kantor, and John Wallis. 2003. “Can the New Deal’s Three R’s Be Rehabilitated? A Program-by-Program, County-by-County Analysis.” Explorations in Economic History (October): 278-307.

Fleck, Robert. 1999a. “Electoral Incentives, Public Policy, and the New Deal Realignment.” Southern Economic Journal 63 (January): 377-404.

_____. 1999b. “The Value of the Vote: A Model and Test of the Effects of Turnout on Distributive Policy,” Economic Inquiry 37 (October): 609-23.

_____. 2001a. “Inter-party competition, intra-party competition, and distributive Policy: A Model and Test using New Deal Data.” Public Choice 108 (July): 77-100.

_____. 2001b. “Population, Land, Economic Conditions, and the Allocation of New Deal Spending.” Explorations in Economic History 38 (April): 296-304.

______. 2008. “Voter Influence and Big Policy Change: The Positive Political Economy of the New Deal.” Journal of Political Economy : 1-37.

Price, Zachary. 2015. “Waiving Rules of Law.” Unpublished paper for Conference on the Role for Executive Authority & Discretion. Hoover Institution Program on Regulation and the Rule of Law. Stanford University, March.

Public Works Administration. 1939. America Builds: The Record of PWA. Washington, D.C.: Government Printing Office.

Reading, Donald. 1972. “A statistical analysis of New Deal economic programs in the forty-eight states, 1933-39,” unpublished dissertation, Utah State University.

Stromberg, David. 2004. “Radio’s Impact on Public Spending.” Quarterly Journal of Economics 119 (February): 189-221.

Wallis, John. J. 1981. Work Relief and Unemployment in the 1930s. Unpublished PhD. dissertation, University of Washington.

_____. 1987. “Employment, Politics, and Economic Recovery During the Great Depression,” Review of Economics and Statistics, 69 (August): 516-20.

_____. 1998. “The Political Economy of New Deal Spending Revisited, Again: With and Without Nevada.” Explorations in Economic History 35 (April), 140-70.

Wallis, John .J., Fishback, Price V. and Kantor, Shawn E., 2006. “Politics, Relief, and Reform: Roosevelt’s Efforts to Control Corruption and Manipulation During the New Deal” in Corruption and Reform edited by Edward Glaeser and Claudia Goldin. NBER Volume. Chicago: University of Chicago Press.

Wright, Gavin. 1974. “The political economy of New Deal spending: an econometric analysis,” The Review of Economics and Statistics, Vol. 56, No. 1, pp. 30-38.