A recent working paper from the Bank of Japan paints a concerning picture of how households react to an economy in stagflation and offers important insights relevant to other economies struggling with similar economic conditions. Their modelling suggests that persistently low interest rates damage the wealth of savers and compresses the wealth distribution. In a low interest rate/low growth environment, households tend to invest less in stocks and shares, or business ventures, and tend to hold more of their savings in cash deposits (even at low or zero rates). In other words, economic growth in suppressed further, in a feedback loop, thus sustaining stagflation for longer (and in our view tending to create conditions which become impossible to escape from – and cutting rates into negative territory will not offer an escape path).
Japan experienced a stock market boom (bust) in the 1980s (1990s) followed by a financial crises. The Bank of Japan’s response to the collapse of the bubble and subsequent deflation is an early version of the kinds of quantitative easing programs now pursued by other major central banks. Despite this policy stimulus, Japan has gone through two decades of low growth, low interest rates and low inflation.
In this paper we set out to understand the way these macroeconomic events affected Japanese household financial decisions. This is interesting because the Japanese experience can provide some indication of how the developed world may be affcted by the onset of such a ‘secular stagnation’ episode characterised by persistently weak rates of inflation and economic growth.
Using data from the Japanese Survey of Household Finance (SHF) from 1981 to 2014, we start by documenting several key household portfolio facts that are unique to Japan. First, stock market participation is considerably lower than in the US: in 2014, 15.5% of all households participate in the stock market. Second, conditional on participation, stockholders hold a relatively small share of wealth in stocks as a percentage of total financial wealth, and a relatively large share of financial wealth in bonds and money. Third, whether one focusses on stockholders or non-stockholders, the share of wealth allocated to cash-like financial instruments is very high. For instance, even for stockholders the share of liquid bank accounts in total financial assets is between 20% and 40% (depending on the age group). Fourth, the gap between the
average wealth of stockholders relative to that of non-stockholders is much smaller in Japan than in the US.What can account for these facts? To answer this question we rely on counterfactual analysis based on a structural, quantitative, life-cycle portfolio choice model with an explicit role for inflation and money demand. Understanding money demand is essential to match Japanese household portfolios given the strong prevalence of money-like financial instruments in Japanese portfolios. Portfolio choice models that incorporate monetary assets are not readily available. Instead, we make explicit the choice between money (that earns a zero nominal return) and other assets like bonds and stocks that earn the historically observed rates of return.
Given that our purpose is to develop a tractable, quantitative, model that can be confronted with the data, we introduce money demand through the shopping time approach. Specifically, we assume that money provides liquidity services: a higher amount of money lowers the cost from having to undertake a given transaction for consumption purposes. Everything else we assume is similar to recent life-cycle models that feature intermediate consumption and stochastic uninsurable labor income.
We calibrate the structural parameters of the model by matching key data moments from the Japanese Survey of Household Finances. Using fixed costs of stock market entry and preference heterogeneity, the calibrated model matches quantitatively limited stock market participation, and the share of wealth in money, bonds and stocks over the later parts of the life cycle. Understanding the portfolio choices associated with that part of the life cycle becomes extremely important in counterfactual analysis as most wealth accumulation takes place at that stage of the life cycle. Armed with this model we can now run counterfactual experiments to better understand the key drivers of Japanese household portfolios. Our counterfactual analysis can informatively address our key questions: why do Japanese households hold so few stocks and such high money balances?
Perhaps the single most widely discussed aspect of Japan’s economic performance since 1990 has been its persistently low level of inflation which has averaged close to zero for almost 15 years. It has been argued that a low level of inflation encourages investors to hold nominal assets (such as money) rather than real assets (such as equities). Our counterfactual experiments confirm this intuition. Had inflation in Japan averaged 2% (as in the US) stock market participation would have risen from 15.3% in the baseline simulation to around 20%. Moreover, the share of stocks in young and middle aged stockholders’ portfolios would have been significantly higher mainly at the expense of lower money holdings, while the share of stocks in elderly households’ portfolios would remain unaffcted . Therefore, low inflation plays an important role in keeping the share of money in Japanese household financial assets very high and contributes to crowding out stocks and bonds from household portfolios.
The second legacy of Japan’s long stagnation since 1990 has been its poor history of realized (and plausibly expected) stock returns as compared to other countries. In the baseline calibration, the mean equity premium for Japan is set to 1.8%. Increasing this to 4% (a typical choice in many life-cycle models calibrated to US data) raises the mean financial wealth to income ratio substantially and increases the stock market participation rate to 50%, a rate that is very close to the recent US experience.
Another important feature of the calibrated structural model is the relatively high cost of stock market entry. Reducing this fixed cost from our estimate (9%) to 5% (as estimated for the US in the literature) leads to an increase in the stock market participation rate from 15.3% to 43% indicating that frictions in equity market participation can be a very important factor in limiting Japanese households’ investment in equities.
The final interesting aspect of the Japanese household portfolio data is the puzzling low mean wealth of stockholders relative to non-stockholders at least in comparison with the US Survey of Consumer Finances (SCF). As before, low realized (and expected) stock returns play an important role in explaining this fact. Returns to capital (the stock market) have important long term implications for the wealth distribution. In the US, realized equity returns have been high, benefitting stock owners. In Japan, by comparison, realized stock returns have been low, and the wealth of those who own stocks relative to those who do not, has not risen to the same extent, generating lower wealth differences. This is especially interesting given the finite nature of the life cycle: one need not rely on an infinite horizon model to generate substantial differences in the wealth distribution. Moreover, these substantial wealth differences can arise even from a relatively low mean differential in expected stock returns (2%).
Note that papers in the Bank of Japan Working Paper Series are circulated in order to stimulate discussion and comments. Views expressed are those of authors and do not necessarily reflect those of the Bank.