More Research of Price Expectations Needed Now: Essential in enabling Japan to end deflation

Consulting Fellow, RIETI

In November 2009, the government announced that the Japanese economy was in moderate deflation. This is the second time this century that Japan has dipped into deflation, coming three and a half years after the last deflation period that continued from March 2001 through June 2006. Taking a broader view, however, it would be fair to say that the Japanese economy has been in a continuous spell of deflation since the mid-1990s. The annual growth rate of Japan's consumer price index (CPI), measured by the general index excluding fresh foods, remained negative or around zero throughout those years except for 1997 when the consumption tax rate was raised and 2008 when crude oil prices soared. According to my estimates, the CPI growth rate contains an upward bias of slightly less than 1 percentage point on average because quality improvements in products, i.e. effective declines in prices, are not fully reflected in CPI figures. With this taken into account, the downward trend of prices has been continuing for about 15 years.

Unlike in the Great Depression, a huge price fall of tens of percent has not been observed in the ongoing deflation period, which is just as mild as the previous one. However, current deflation is also different from the previous one in that this time around we have not seen a string of bank failures resulting from a serious aggravation of the problem of bad loans. Thus, very few point to the possibility of Japan being in "debt-deflation," a process in which deflation causes more deflation, as defined by Irving Fisher. But there has been an active debate on whether the introduction of inflation targeting and the use of non-traditional monetary policy tools are effective as a means to fight deflation.

In the previous deflation period, heated debates occurred from time to time over the cause of the deflation and policy responses. However, as a prerequisite to get the right prescription for deflation, it is necessary to fully understand, either theoretically or empirically, a mechanism by which price expectations are formed, and the impact of such expectations on the real economy. The expected rate of price increase (expected inflation), which reflects people's expectations about changes in the price level over a period of time in the future, varies significantly across economic entities and depends on the period covered. In any case, however, a change in the expected inflation leads to a change in the real interest rate even if the nominal interest rate remains unchanged, thereby having a significant impact on the real economy including household consumption and capital investments.

Deflation expectations - i.e. when the expected inflation is negative - lead to an increase in the real interest rate and effective debt burden, which would negatively impact the economy by inducing the postponement of durable goods purchases and cause an effective transfer of income from entities with net liabilities to those with net assets. The expected inflation increase also affects the actual rate of price increase through wage and price setting decisions.

Furthermore, it affects a central bank's function of stabilizing prices. A situation in which the expected rate of price increase is relatively stable over a medium- to long-term time horizon is described as being "well-anchored" and fosters strong public confidence in the central bank. On the other hand, when the expected inflation is unstable, the central bank's price stabilization function does not work properly.

* * *

In an attempt to explain the mechanism of how price expectations are formed, various theories, including traditional rational expectations models, have been developed. And many empirical studies have been put forward overseas to verify whether such a mechanism is actually functional and whether there exist conditions for the mechanism to function in reality. In Japan, however, quite often accumulated data and empirical evidence is insufficient to support economic policies in general. Price expectations are no exception and at the time of the previous deflation period, policy discussions were often predicated on the assumption that the expected inflation is equal to the current inflation.

Needless to say, "expectations" for future prices are not visible and the inflation expectation must be measured in some way. While some methods have been developed to fulfill this need (see Shimizutani, S. Kitai to Fukakujitsusei no Keizaigaku [Economics of Expectations and Uncertainty] for details), many attempts have been made overseas to determine the expected inflation based on data collected through surveys of households and corporations. In the United States, a group led by the University of Michigan has been collecting and measuring data on households since the 1960s, which has been used to identify the expectations formation mechanism and served as a reference for policy implementation. In Japan, efforts have been made since 2000 to quantitatively measure the expected inflation.

One such attempt, in which the expected inflation over a short-term time horizon was measured in the midst of the previous deflation period, is a monitor survey on citizens' lifestyles ( Kokumin Seikatsu Monita Chosa ) conducted by the Cabinet Office. The figure shows quarterly changes in the expected inflation from June 2001 through March 2004. In this survey, respondents were asked to provide specific figures as to how much they thought the prices of daily commodities had changed from a year before and will likely change in the next year.

The figure shows that the current rate of price change as perceived by survey respondents (perceived inflation) moved closely in tandem with the CPI growth rate, indicating that households accurately grasp the actual movement of prices. Meanwhile, the expected inflation was slightly negative in 2001 and 2002 as if confirming that the Japanese economy was in mild deflation, but suddenly rose to 1% in the January-March quarter of 2003. Subsequently, the expected inflation dipped but rose to slightly below 1% in the January-March quarter of 2004. Those findings from the survey are subject to bias because respondents were not selected at random. However, its innovative approach of directly asking households, in the sample quantitative questions, about their expected price changes has since been adopted in other surveys including the Monthly Consumer Confidence Survey by the Cabinet Office and the Opinion Survey on the General Public's Views and Behavior by the Bank of Japan (BOJ).

Expected Inflation in Japan Expected Inflation in Japan

Note: Created by the author based on the research paper co-authored with Masahiro Hori published in International Economics and Economic Policy , vol. 2, 2005

* * *

Now, what factors determine such changes in people's price expectations? Determinants for the expected inflation includes the current rate of price increase as perceived at present (adaptive expectations), the expected inflation as perceived in the past (expectations inertia), income, and exogenous factors such as the monetary policy and the outbreak of a global incident. Joint research conducted by Masahiro Hori of the Cabinet Office and myself found that changes in the expected inflation can be partially explained by adaptive expectations and expectations inertia.

Meanwhile, when the central bank changed its monetary policy, only a small percentage of households changed their expected inflation even if they were aware of the policy change. However, those who did revise price expectations raised their expected rate of price increase by nearly 1 percentage point each when the BOJ introduced quantitative easing in March 2001 and when the central bank began considering purchasing asset-backed securities in 2003. Their expected inflation increase also rose by 1.2 percentage points in response to the September 11, 2001 terrorist attacks in the U.S., and 1.4 percentage points in response to the subsequent outbreak of the Iraq War in 2003. It is inferred that these incidents caused a huge shock to households' price expectations probably due to their memory of the oil shocks in the past. The jump in the expected inflation in early 2003 can be almost entirely explained by the outbreak of the Iraq War.

A drastic change in the foreign exchange rate can be cited as an example of exogenous shocks that affected households' price expectations. According to joint research conducted by Tatsuhiro Yogi of the Bank of the Ryukyus and myself, the Nixon Shock in 1971, which resulted in a depreciation of the U.S. dollar by 17% against the yen, pushed up the expected inflation by an estimated 4% to 5% in pre-reversion Okinawa, where the U.S. dollar was the currency in use at the time.

Thus, the expected inflation, while being subject to impacts through diverse channels, is strongly affected by an exogenous shock. A change in the monetary policy is no exception. Even though simply being aware of a policy change did not result in a meaningful change in households' price expectations, its impact on those households that revised their expectations was just as large as the impact caused by the 9/11 terrorist attacks or the Iraq War. Therefore, if the government and monetary authorities intend to reverse the ongoing deflationary expectations by means of monetary policy, it is important to take a step that is bold and easy to understand so that as many households as possible will get and understand the message and change their price expectations.

In the joint research with Hori, which also examined the effects of deflationary expectations on households' consumption behavior, it was confirmed that only those with housing loans reduced consumption, postponed the purchase of durable goods, and linked deflationary expectations to anxiety over losing their jobs which resulted in a negative impact on their consumption behavior. Likewise, in the U.S., Frederic S. Mishkin, a professor at Columbia University, showed that the impact of the Great Depression on household purchases of durable goods and houses varied depending on the amount of financial assets or liabilities held.

Empirical findings discussed in this article concern just the mechanism in which households form their short-term price expectations. Efforts have been made to collect data on companies' expected rate of price increase, for instance, through the Cabinet Office's Annual Survey of Corporate Behavior. But companies' expectations are even more difficult to measure quantitatively than households' expectations. Fortunately, however, Japan's efforts to collect and accumulate data on the expected inflation have progressed since the previous deflation period. By actively utilizing these data, we should accumulate knowledge and understanding of the mechanism for the formation of price expectations and then proceed to make policy debates based on such knowledge and understanding. Now is the time to make strenuous efforts so as to enable Japan to have meaningful policy debates, develop an effective "exit" policy, and derive valuable lessons from the ongoing deflation.

>> Original text in Japanese

* Translated by RIETI.

March 24, 2010 Nihon Keizai Shimbun

May 18, 2010