Lessons from the 2014 Consumption Tax Rate Increase: The decision-making process is more important than the timing of implementation

Faculty Fellow, RIETI

The Japanese government is discussing whether to again postpone a consumption tax rate hike, which is presently scheduled for April 2017. One focal point of the debate is the risk that a tax rate hike could dampen consumption. This article considers what impact a tax rate hike would have on consumption, based on research conducted with David Cashin, an economist at the Federal Reserve Board of Governors.

♦  ♦  ♦

Very few items are exempt from Japan's consumption tax, which is levied at a single rate, and there is strong pressure to pass the tax onto retail prices. Therefore, when a consumption tax rate hike is announced, it is expected that the price of goods will rise by the same margin as the tax increase.

The impact that the expected price increase has on consumption can be examined using the Life Cycle/Permanent Income Hypothesis (LCPIH), a standard economic theory of how consumption is decided. The LCPIH assumes that 1) consumers' total lifetime consumption will equal lifetime income and 2) consumers budget to try to avoid changes in consumption levels. For households, it is optimal to allocate their lifetime income over time and maintain the same consumption level.

The same hypothesis also suggests that tax hikes cause optimal consumption to decline permanently, through two means.

The first is the so-called income effect, which suggests that falling lifetime income changes consumption. If a consumption tax rate hike causes prices to rise, then all else being equal, real lifetime income falls. Consumption will then fall by the same ratio to stay within the consumer's lifetime budget.

The second means by which optimal consumption declines is called the intertemporal substitution effect, where changes in relative prices cause changes in the timing of consumption. Even if we assume consumption changes when a price increase is expected, consumers will try to consume more while prices are low. Therefore, consumption levels are relatively high until prices rise, and fall after the actual tax hike happens.

The intertemporal substitution effect needs to be distinguished from the so-called last-minute demand surge. A last-minute demand surge is the consumption of durable goods, such as cars and television sets, as well as storable non-durable goods, such as canned foods and washing detergent just before a tax rate hike. This is similar to the intertemporal substitution effect in the sense that purchasing rises while prices are low, but in a last-minute demand surge, because there is a change in purchase timing, expenditures change but do not cause consumption to do so. The increase in expenditures resulting from a last-minute demand surge is evened out by a rebound decline after the tax increase and has no effect on long-term consumption levels. This is a major factor in economic fluctuations in the short term, but is outside the scope of analysis here.

Both the income effect and intertemporal substitution effect are reactions to tax rate hikes, but they occur at different times. The figure shows how consumption changes at the time a tax rate hike is announced and when it is implemented. The income effect happens immediately when a tax rate hike is announced, as consumption levels fall. This is because consumers recognize the tax rate hike as a decrease in lifetime income, even before it actually happens.

Income Effect and Intertemporal Substitution Effect by the Consumption Tax Rate Hike
Income Effect and Intertemporal Substitution Effect by the Consumption Tax Rate Hike

The intertemporal substitution effect, on the other hand, causes consumption levels to rise at the time of announcement and fall after the tax rate hike is implemented. In other words, as the figure shows, consumption falls in two steps: when the tax rate hike is announced and when it is implemented.

The long-term impact of a consumption tax rate hike on consumption can be measured by combining changes in consumption levels upon the announcement of the hike and upon its implementation. Therefore, we measured the breadth of the consumption drop by using micro-data from the Family Income & Expenditure Survey. To eliminate any effect from last-minute demand surge and the subsequent rebound decline, we counted only expenditures of non-storable, non-durable goods as consumption.

♦  ♦  ♦

A critical point to actually measuring how far consumption declines is to identify the timing of the announcement. The decision to raise taxes generally goes through many policy-making processes, so we cannot say at what stage households recognize an "announcement" of a tax rate hike. First of all, if each household recognizes the tax rate hike at different times, we cannot uniformly identify the time of the announcement.

In the case of the 2014 tax rate hike, however, the October 1, 2013 press conference by Prime Minister Shinzo Abe can be considered as the time of the announcement. Right up until then, there were high expectations that Abe would decide to postpone or cancel the consumption tax rate hike, so the announcement had a surprise effect. This was a clear announcement to practically all households.

When we measured changes in consumption at the time of the announcement in October 2013 and upon implementation in April 2014, adjusting for seasonality, household attributes, etc., we found consumption fell by 4.1% in October 2013 and 1.0% in April 2014. In other words, the consumption tax rate hike resulted in a combined 5% downturn in consumption.

A 5% consumption drop seems like a major change, given that the margin of the tax hike was only 3%. However, in October 2013, the plan was to eventually raise the consumption tax rate to 10%. This means that consumers would have seen a tax rate hike of 5%. If that is the case, the tax rate hike margin and the consumption decline margin were about the same, which positively supports the LCPIH.

Although a consumption tax rate hike has a large impact on consumption, the decline in consumption is actually relatively small at the point when the tax is implemented. The April 1997 consumption tax rate hike from 3% to 5% reveals similar results. According to our estimates, the decline in consumption when it was implemented in 1997 was 0.7%.

It is very interesting to compare the more recent consumption drop at the time of announcement to the 1997 case. This is because no such drop in consumption could be found at the time of announcement in the 1997 case. If the LCPIH holds true, there should have been an income effect. But it is conceivable the reason there was no detected change in consumption was because there was no clear announcement and every consumer became aware of the tax rate hike at different times. Viewed at the macroeconomic level, no clear drop in consumption was observed.

In the 2014 case, conversely, there were high expectations of a tax rate hike postponement, and then there was a clear announcement just before the scheduled time of increase, so many consumers changed their consumption levels at the same time and there were large changes in consumption even at the macroeconomic level.

♦  ♦  ♦

We can derive the following policy implications from these results. First, actual consumption tax rate hikes do not of themselves have a major impact on consumption, because the timing of consumption does not change very much as a result of relative price changes. Conversely, if a tax rate hike has already been announced, consumption is not likely to recover even if it is postponed somewhat.

Second, to avoid sudden changes in consumption, major revisions of information on the tax rate hike should be avoided. Policies such as tax rate hikes, which have a great impact on people's lives, cause unnecessary fluctuations in macro consumption if they are changed just before implementation without a policymaking process. A more desirable strategy would be to announce a long-term schedule for tax system changes and then reliably follow through.

The third implication is the importance of consumption indicators. The reason we are able to observe a consumption decline at the time of announcement is because we looked at the consumption of non-storable, non-durable goods. Looking at consumption expenditures as a whole, there was an increase at the time of announcement, in part because of a last-minute demand surge. To correctly determine the state of consumption, it is necessary to observe appropriate indicators.

A further consumption tax rate hike is inevitable in Japan, where the population is aging rapidly. If we assume the LCPIH holds true, it will not be possible to avoid a drop in consumption as a result of a tax rate hike. However, it is possible to avoid unnecessary consumption fluctuations if we adequately consider not the timing of implementation, but the process by which the tax rate hike decision is made.

>> Original text in Japanese

* Translated by RIETI.

May 23, 2016 Nihon Keizai Shimbun

June 22, 2016