Worried About Inflation? Get Money Right First

October 21, 2009 By Qing Wang Hong Kong & Steven Zhang Shanghai Fear of Inflation As economic recovery is underway, there appears to have been increasing concerns among some market participants about potentially high inflation in China in 2010, especially in view of the rapid monetary expansion. The growth rate of broad money, M2, reached 29% in September, the fastest pace since 1997. Some market commentators predict that such an extraordinarily rapid monetary expansion is bound to result in high inflation sooner or later as, they argue, inflation is after all a monetary phenomenon. Out of this concern about serious inflation down the road, these market commentators are calling for early and aggressive policy tightening by the authorities. This fear of inflation appears to have weighed on market sentiment of late. A Bit of Theory The ‘quantify theory of money' is the theoretical foundation for gauging the inflation outlook by examining the expansion of money. The macroeconomic relationship between money, economic activity and prices is usually expressed in its most basic form by the well-known equation of exchange: MV = PY, where P is the price level of goods and services, Y represents national output or income, M is the supply of money, and V the velocity of money. With a stable velocity, a stable relationship between nominal national income (PY) and M exists. The ‘quantity theory of money' argues that money is directly related to nominal income or output, and hence prices (i.e., inflation). It should, however, be pointed out that only purchasing power that is actually used for transactions can influence nominal income. Therefore, a quantity relationship between prices or GDP and money should more precisely refer to that part of the money supply that becomes effective purchasing power. Get the Measurement of Money Right in China There are two popular indicators of money supply in China: M1 and M2. However, neither provides a precise measurement of the money as defined by the ‘quantity theory of money', in our view. Compared to many other countries, the definition of M1 in China is too narrow and rather unique, in that it only includes cash and demand deposits by enterprise and does not include demand deposits by households. While the definition of M2 in China is de jure broadly in line with the international standards, it is de facto too broad to be considered as representing ‘purchasing power used for transactions'. Here is why: Reflecting the underdevelopment of capital markets as a result of ‘financial repression' for decades, a large part of households' deposits at banks - that comprise the M2 - are actually long-term savings (or a form of financial investment) instead of for transaction purposes, and therefore do not represent actual and perhaps even potential purchasing power. This is the key reason why China's M2-GDP ratio is 193%, one of the highest in the world. And Chinese households' deposits account for about 85% of GDP, while their financial exposure to the stock market only accounts for about 25% of GDP. When the long-term saving component of M2 is stable, the change in M2 primarily reflects the change in money supply/demand for transaction purposes. However, when the long-term saving component becomes unstable, the change in M2 will not necessarily reflect the change in money supply for transaction purposes, and drawing implications of the change in the headline M2 growth to the inflation outlook could become rather tricky and even misleading, in our view. This has certainly been the case in China in the last 4-5 years. The Chinese stock market has undergone a rapid and profound development in recent years and become a meaningful asset class into which Chinese households are actively seeking to diversify their financial portfolio that has long been dominated by bank deposits. This has led to an unstable relationship between the money supply and CPI inflation in recent years, as the change in the headline M2 growth is influenced by the vagaries of households' bank deposits as a result of a massive rebalancing of households' financial portfolio between cash and stocks. Specifically, when the stock market rises, households' deposits (and thus M2) tend to decline and vice versa. To get the measurement of money right in China for gauging the inflation outlook entails estimating the underlying household deposits that are free from the influence of the stock market and for transaction purposes only. To this end, we run a regression with households' deposits as dependent variables and nominal GDP as an independent variable and use the fitted value from the regression equation - which can explain about 97% of the variation of the actual household deposits - as a proxy for the household deposits that are used for transaction purposes. The difference between the headline households' deposits and the fitted value can therefore be viewed as a proxy of households' deposits for investment purposes. (In this context, strictly speaking, the households' bank deposits for investment purposes should be interpreted as a deviation from the historical average amount of households' bank deposits for investment purposes.) The portion of households' bank deposits for investment purposes is heavily influenced by the stock market performance: when the stock market rises, households' bank deposits tend to decline and vice versa. After pinning down the portion of household bank deposits for investment purposes, we adjust the headline M2 to estimate the ‘true M2' that reflects transaction purposes only. And it is the change in this part of M2 that should have direct implications on inflation. Indeed, the true M2 growth demonstrated a much closer correlation with CPI inflation than the headline. Of particular note, despite the record-high headline M2 growth of about 29%Y in 3Q09, we estimate that the true M2 growth was only about 20%Y, which is substantially below the recent peak of 26%Y reached in 3Q07. Note that, back in 3Q07, the headline M2 growth was only 18%Y. What explains these large discrepancies between the headline M2 and the true M2 growth rates? First, despite the seemingly well-behaved headline M2 growth in 3Q07, the overall demand for money has actually declined sharply, because households chose to buy shares over holding cash in the form of bank deposits. As such, the underlying money supply significantly outpaced money demand, as is reflected in the rapid increase of the underlying true M2 growth, generating strong inflationary pressures (we first discussed this topic in China Economics: Tighter Policy on Structural Shift in Money Demand, July 3, 2007). The situation so far this year is just the opposite: despite the rather strong headline M2 growth since 4Q08, the underlying overall demand for money has actually increased sharply, because extreme caution and risk-aversion amid ‘the most severe financial turmoil since the Great Depression' has caused households to hold cash over buying risk assets. As such, the underlying money demand may have outpaced the supply of money, as is reflected in the not-so-rapid increase of the true M2 growth in 3Q09. In conclusion, the strong headline M2 growth overstates the true underlying monetary expansion, as it fails to account for the change in M2 caused by the shift in asset allocation by households between cash and stocks. The true M2 growth is estimated to be about 20% in 3Q09. While this is much lower than the headline M2 growth and thus less alarming, it still looks quite high judging by the historical trends of the true M2. Looking at the relationship between true M2 and CPI would make one wonder whether a repeat of the episodes of high inflation during 2003-04 and 2007-08 cannot be avoided. Alternatively, could we expect a repeat of the situation in 2000-01 where inflationary pressures were quite moderate despite relatively high true M2 growth? To this discussion, we turn next. Friedman versus Keynes Followers of Milton Friedman's monetarist school believe that inflation is a monetary phenomenon, and rapid monetary expansion will sooner or later result in high inflation. However, the students of the Keynesian school would argue that when aggregate demand is weak, the output gap large and unemployment high, strong inflationary pressures are unlikely to emerge even if there were to be strong money supply growth. This is because the supply of money may be stuck in a ‘liquidity trap' without being able to effectively stimulate aggregate demand. It is difficult to estimate the output gap for such a rapidly growing economy like China, where structural changes are constant. However, the change in exports can be treated as a proxy for the output gap in China, in our view. Much weaker exports represent a powerful negative demand shock that is deflationary. In particular, China's past experiences suggest that a significant decline in export growth should have a meaningful disinflationary/deflationary impact on the economy. China has suffered three episodes of deflation in the last decade or so: one during the Asian Financial Crisis, the other in the aftermath of the NASDAQ stock bubble burst, and the current one. The deflation either coincided with or occurred in the immediate aftermath of a collapse in export growth. While the latest data suggest that the sharp decline in exports has stabilized and the negative growth rate has started to narrow, it still makes the decline in exports the largest in China's history. Looking forward, we expect China's export growth to turn positive towards year-end; however, the recovery in 2010 is unlikely to be very robust, as Morgan Stanley's global economics team envisages a tepid recovery in the G3 economies in 2010 (see Global Forecast Snapshots, September 10, 2009). We forecast China's exports to expand by 10% in 2010, which is much lower than the pre-crisis average of 23%. The lackluster export growth will continue to constitute a strong headwind containing inflation pressures in 2010, as was the case in 2009, in our view. In other words, the backdrop for external demand suggests a repeat of the 2000-01 situation in 2010, where inflationary pressures were rather muted despite relatively high M2 growth. In summary, as a monetarist, one should be worried about potentially high inflation, albeit less so than warranted by the surge in headline M2 growth. However, as a Keynesian, one also has strong reasons not to be concerned about inflation, given the persistence of weak external demand and attendant excess production capacity that limits firms' pricing power. Between the two offsetting forces, which plays a dominant role and what's the net impact on inflation? We turn next to econometric modeling, with a view to quantifying the impact. The Model: A Hybrid of Two Schools We construct an ordinary least square (OLS) regression equation with CPI inflation as the dependent variable and the true M2 growth (with a two-quarter lag) and export growth (with a one-quarter lag) as the independent variables. By specifying such an equation, we hope to capture both the inflationary effect of monetary expansion and the disinflationary effect of weak exports. A ‘good' model should ideally have the following three features: the model can explain the bulk of the variation in CPI inflation (i.e., a high R-squared); a positive and statistically significant coefficient for true M2 growth; and a positive and statistically significant coefficient for export growth. The regression results turn out to be quite good, especially in view of the rather simple structure of the equation. Both the estimated coefficients have the right positive signs with about 99% significance level, and the R-squared is 0.82. The Inflation Forecasts To make a forecast based on the inflation model, we plug into the equation our forecasts of M2 and export growth forecasts for 2010. We forecast the true M2 and export growth in 2010 to be 18% and 10%, respectively, which are in line with our forecasts of real GDP growth of about 10%. The model generates a path of CPI inflation forecasts through end-2010. The model-based CPI inflation forecasts suggest that fear of serious inflation in 2010 is unwarranted. Specifically, the average CPI inflation in 2010 will be about 2.5%. The inflation rate will turn positive in 4Q09 and starts to rise rapidly to 2.4%Y in 1Q10 and 2.6%Y in 2Q10, as the lag effects of strong true M2 growth in 3Q09 and 4Q09 are only partly offset by continued weak exports. CPI inflation will likely peak in 3Q10, at about 2.8%Y, as the pick-up in export growth since 2Q10 will add to inflationary pressures despite some moderation in M2 growth. Policy Calls Our policy calls hinge on our outlook for inflation in 2010. Specifically, we expect that the current policy stance should remain broadly unchanged toward year-end and turn neutral at the start of 2010, as the pace of new bank lending creation normalizes from about Rmb10 trillion in 2009 to Rmb7-8 trillion in 2010. Policy tightening in the form of a RRR hike, base interest rate hike, or renminbi appreciation is unlikely until after mid-2010, in our view. We cannot rule out the possibility of 1-2 base interest rate hikes accompanied by RRR hikes in 2H10, as CPI inflation reaches 3%Y and the export growth recovery proves to be more durable, likely by early 3Q10. Given the de facto USD peg new renminbi regime that has been in place for over a year now, the exact timing of China's rate hike would also be influenced by the timing of the US Fed's first rate hike, in our view (see China Economics: An Exit Strategy for the Renminbi? June 9, 2009). In general, we do not expect China to hike rates before the US Fed does. Morgan Stanley's US economics team expects the US Fed to stay on hold until mid-2010 (see US Economic and Interest Rate Forecast: Recovery Arrives - but Not a ‘V', Richard Berner and David Greenlaw, September 8, 2009). Where We Can Be Wrong If the global economic recovery in 2010 were to be much stronger than expected, both China's export growth and global commodity prices could surprise to the upside, likely resulting in stronger inflationary pressures and earlier policy tightening. If, however, we turn out to be wrong, it would suggest that both global and Chinese economies would be in a much better shape than is currently envisaged under our baseline scenario. We will devote a separate follow-up note focusing on the implications to inflation stemming from potential supply shocks (e.g., high international commodity and domestic food prices). Stay tuned.