"Propensity to consume" and "multiplier effect"
An exchange between Minister of Finance Naoto Kan and Yoshimasa Hayashi (a member of the Liberal Democratic Party who formerly served as the state minister of economic and fiscal policy) at a House of Councillors Budget Committee meeting in January drew significant attention. Mobilizing their "expertise" in economics, the two lawmakers engaged in a heated debate over the effect of an economic stimulus policy of the ruling Democratic Party of Japan. I would like to frankly welcome the fact that a debate based on economics took place in the political arena. Seizing it as a rare opportunity, I would like to discuss the effects of the DPJ's "subsidies-to-households policy" in light of the trend of research in economics. The subsidies-to-households policy is a policy in which the government provides cash to individual households under the names of "teigaku kyufukin (Supplementary Income Payment)" and "kodomo teate (child allowance)." Based on textbook Keynesian economics, these kinds of subsidy programs help increase household consumption and thereby stimulate the economy.
Of the amount of subsidiaries received, the ratio of the portion spent on consumption is referred to as "propensity to consume." Meanwhile, the increase in the gross domestic product (GDP) resulting from the provision of subsidies as a ratio to the amount of the subsidies provided is called a "multiplier." It is generally perceived that the propensity to consume is around 0.7, which implies approximately 70% of the amount of subsidies is spent on consumption. According to the Keynesian principle of effective demand, an increase in consumption results in an increase in income, which leads to a further increase in consumption. As a result of this multiplier effect, the provision of subsidies pushes up GDP by an amount greater than the amount of subsidies. When the propensity to consume is 0.7, the multiplier is assumed to be about 2.3, which means the provision of subsidies would increase GDP by an amount equal to 2.3 times the total amount of the subsidies. For instance, when the government distributes 2 trillion yen just like the Supplementary Income Payment (SIP) program, 1.4 trillion yen would be spent on consumption in the first round, which finally results in a 4.6 trillion yen increase in GDP.
A new macroeconomics and economic stimulus effects
In contrast to Keynesian economics, the new macroeconomic theory, which takes dynamic optimization behavior into account, tells us the DPJ's subsidies-to-households policy has no economic stimulus effect. Because subsidies are financed by issuing government bonds, the government should eventually raise taxes to redeem the government bonds. Thus, rational households regard the subsidy just as a temporary income change, since it is inextricably combined with tax hikes in the future. This would induce households to save the money they received for tax burdens in the future, and their propensity to consume would be zero. Needless to say, where there is no effect to increase consumption, there is no effect to stimulate the economy.
These differences in the policy evaluation have been subject to debates since the second half of the 1970s and have come to be known as the "consumption function controversy" between Keynesian and neoclassical economists. However, these debates seem to have run their course, and neoclassical economics has become mainstream. Many economists are now skeptical about the economic stimulus effects of subsidies and believe that evaluating government policies based on textbook Keynesian economics is an outdated approach.
Policy debates falling two laps behind
However, things are not as simple as that. Since the second half of the 1990s, researchers have become more sophisticated in their approach, and their examination of whether tax rebates and/or subsidies can stimulate consumption has produced positive effects. In the United States, it has been shown that consumers' propensity to spend the 2001 and 2008 tax rebates, respectively following the burst of the information technology bubble and the Lehman debacle, was 0.2 to 0.4 (Johnson, Parker, and Souleles, 2006; Shapiro and Slemrod, 2009). This would be consistent with the evidence from the SIP program in Japan, in which the propensity to consume is estimated to be 0.33 by the Cabinet Office (Cabinet Office, 2010).
Theories have also been developed to explain these empirical observations. A mechanism for consumption increase under the dynamic optimization framework is gradually being elucidated. One candidate for the mechanism is the precautionary savings theory, which is an extension of the traditional economics approach, and another is a behavioral economics approach that takes irrationality into consideration.
This means that on the frontier of economics, it is acknowledged that a policy of providing subsidies to households is effective to some degree as a tool to stimulate consumption and the world of economics is now departing from the new macroeconomic paradigm and moving into the next stage. Thus, Keynesian economics, which remains the standard in Japan, is not just one but two laps behind the latest trend of economics.
Fortunately or unfortunately, the policy implications of the latest research findings are closer to those of textbook Keynesian economics, than to those of the new macroeconomics. In this sense, policy evaluation in Japan is pointed in the right direction. However, when premised on a behavioral economics theory, for instance, one-time subsidies such as the SIP program and regularly paid-out subsidies such as child allowance would differ in their impact on the economy, even if the two types of subsidies are equal in total amount. In order to properly analyze such differences, it is essential to incorporate theoretical and empirical developments.
Naturally, anywhere in the world there is a certain time lag before the findings from cutting-edge academic research can have an impact on government policies in practice, but policy debates need to evolve before they fall two laps or more behind.
