Lecture 2 Chinas Macroeconomic Model by Vector

By Joe Williams,2014-06-11 07:18
7 views 0
Lecture 2 Chinas Macroeconomic Model by Vector

     Lecture 3 Macroeconomic Model of China by Vector Autoregression VAR

2.2 Determination of the price level P and inflation.

The model of section 2.1 does not explain the price level P. Chow (2002, chapter 7) has a

    model explaining the price level P by an error-correction mechanism as follows.

First, estimate an equilibrium cointegration equation by regressing lnP on ln(M/Y) and

    denote the residual of this equation by u. Second, estimate an equation to explain

    inflation defined as lnP(t)-lnP(t-1)=?lnP(t) by its lagged value, the current and lagged

    values of ?ln(M/P) and by u(t-1). The variable u(t-1) provides an “error-correction” effect to bring lnP back to equilibrium specified by the cointegration equation. Using data

    from 1952 to 2005, the results for M1 and M2 are respectively

lnP = -0.2798(.0540) + 0.5311(.0262)*ln(M1/Y) + u (3.1)

    R2 =.8873 s=.1894 DW=0.1165

    lnP = 0.0046(.0055) + 0.7076(.0925)* ? lnP(-1) + 0.2087(.0480)* ?ln(M1/Y) - 0.1477(.0533)* ?ln(M1/Y(-1)) - 0.0521(.0243)*u(-1) (4.1) 2R =.6272 s=.0309 DW=1.8304


lnP = -0.2508(.0385) + 0.4169(.0149)*ln(M2/Y) + u (3.2)

    R2 =.9375 s=.1411 DW=0.1107

    lnP = 0.0036(.0055) + 0.6519(.0967)*? lnP(-1) + 0.2123(.0464)* ?ln(M2/Y) - 0.1298(.0547)* ?ln(M2/Y(-1)) - 0.0808(.0314)*u(-1) (4.2)

    R2 =.6514 s=.0299 DW=1.8116

Equation (4.1) or (4.2) is an example of one equation in a Vector Error Correction VEC

    model. Vector means that the model has more than one variable. In this case the variables

    are ?ln(P) and ?ln(M/Y). The model has at least one equation for the levels of the

    variables lnP and ln(M/Y), called a cointegration equation, as given by (3.1) or (3.2). The

    residual or error u(t-1) of the cointegration equation in the period t-1 affects the

    dependent variable ?ln(P), besides the lagged values of lnP and ln(M/Y) through an error-correction effect. We will employ VEC models extensively later in this paper.

    Results will be presented using M1 only since results from using M2 are similar.

    Table 7.1

    Data on Inflation and its Determinants

     General Retail Price Index GDP Index Currency in

    Price Index Preceding 1978=100 Circulation (100

    Year=100 million) End of


    1952 0.8227 99.6 22.0 38.55 1953 0.8506 103.4 25.1 39.60 1954 0.8705 102.3 26.6 41.19 1955 0.8793 101.0 28.3 40.13 1956 0.8793 100.0 32.3 57.03 1957 0.8926 101.5 33.7 52.80 1958 0.8947 100.2 41.2 67.59 1959 0.9028 100.9 44.6 74.98 1960 0.9308 103.1 43.9 96.10 1961 1.0820 116.2 30.9 125.67 1962 1.1229 103.8 28.9 106.66 1963 1.0567 94.1 32.0 89.76 1964 1.0177 96.3 37.2 80.26 1965 0.9904 97.3 43.5 90.82 1966 0.9875 99.7 50.9 108.25 1967 0.9801 99.3 44.5 121.97 1968 0.9809 100.1 44.2 134.12 1969 0.9698 98.9 52.7 137.29 1970 0.9676 99.8 65.0 123.56 1971 0.9603 99.2 69.5 136.23 1972 0.9581 99.8 71.5 151.02 1973 0.9639 100.6 77.5 166.33 1974 0.9691 100.5 78.3 176.36 1975 0.9706 100.2 84.9 182.70 1976 0.9735 100.3 82.6 203.82 1977 0.9934 102.0 89.0 195.37 1978 1.000 100.7 100.0 212.27 1979 1.020 102.0 107.6 267.71 1980 1.081 106.0 116.0 346.20 1981 1.107 102.4 122.1 396.34 1982 1.128 101.9 133.1 439.12 1983 1.145 101.5 147.6 529.78 1984 1.177 102.8 170.0 792.11 1985 1.281 108.8 192.9 987.83 1986 1.358 106.0 210.0 1,218.36 1987 1.457 107.3 234.3 1,454.48 1988 1.727 118.5 260.7 2,134.03 1989 2.034 117.8 271.3 2,344.02 1990 2.077 102.1 281.7 2,644.37 1991 2.137 102.9 307.6 3,177.80 1992 2.252 105.4 351.4 4,336.00

    1993 2.549 113.2 398.8 5,864.70

    1994 3.102 121.7 449.3 7,288.60

    1995 3.561 114.8 496.5 7,885.30

    1996 3.778 106.1 544.1 8,802.00

    1997 3.808 100.8 592.0 10,177.60

    1998 3.709 97.4 638.2 11,204.20

     1999 3.598 97.0 684.1 13,455.50

     2000 3.544 98.5 738.8 14,652.70

     2001 3.516 99.2 794.2 15,688.80

     2002 3.470 98.7 857.4 17,278.00

     2003 3.467 99.9 940.1 19,745.99

    2004 3.564 102.8 1031.3 21,468.30

Since economic reform started there were several episodes of inflation, all associated with a rapid

    increase in money supply, as we can see from Table 7.1. Money supply can be measured by the

    amount of currency in circulation.[Delete:, since checking accounts were and still are uncommon

    in China.] The episodes include 1985, 1988, and 1993. In all three cases there was a rapidly

    increase in money supply at an annual rate of about 50 percent, as compared with some 20 to 25

    percent in other years. The increase in money supply can be said to be a government policy but the effect of the policy was not necessarily intended. In 1984 when enterprise autonomy was

    introduced for state enterprise reform, state enterprises went directly to the banks for credits.

    There was insufficient control of the extension of these credits. In fact the banks were by mistake

    given autonomy in much the same way as the state industrial enterprises. Credits were extended

    as a result of political pressure and willingness on the part of bank managers to extend loans to

    enterprises in their home province for the purpose of promoting local economic development or

    simply expansion of economic activities. When credits were so extended by the local banks

    which were branches of the People’s Bank, the enterprises receiving the credits can convert them

    to cash for wage payment and other expenses. The People's Bank had to issue currency to honor

    the credits of its own branch banks. The end result was an increase in currency in circulation of

    about 50 percent from January 1984 to January 1985. Inflation in 1985 was 8.8 percent, high

    compared with the norm of the low inflation rate achieved up to 1984.

The increase in money supply in 1988 was also unintentional and resulted from the reform

    process. Reform gave enterprises and consumers much freedom in the latter part of the 1980's.

    Aggregate demand and national output were expanding at a high rate. Unless the government

made a serious effort to control the supply of credits, inflation would occur. The government did

    fail to control the supply of credits. In 1988 money supply again expanded at the rate of 48

    percent. Furthermore the government announced a policy to fix the prices of some important

    consumer goods at the level to prevail at the end of 1998. In response to the announcement

    producers hastened to increase prices before the deadline. In the fall of 1988 the price index was

    increasing at an annual rate of over 30 percent although the annual rate of inflation was 18.5 percent. This rate was much higher than the 8.8 percent rate in 1985 because of the delayed

    effects of the substantial increases in money supply in the years before 1988 which had not

    occurred before 1985. Inflation, together with corruption, contributed to discontent and student

    demonstration in the Spring of 1989.

Witnessing the serious inflation in fall of 1988, the government tried to restrict the increase in

    money and credit creation by assigning credit quotas to banks of different regions and provinces.

    In addition, interest rates on bank deposits were raised to attract deposits and to reduce the

    quantity of money in circulation, thus making the price level lower than otherwise. Because of

    the absence of a well-functioning modern banking system, the government applied mainly

    administrative means to control money supply, namely the assignment of credit quotas to banks

    in different regions. A second instrument of monetary policy was to increase the interest rates on

    saving deposits in order to induce people to put their money in the banks rather than spending it.

    Deposit rates were increased to over 11 percent per year to make the real rate of interest positive

    and attractive enough for depositors to increase their deposits. The policy did work and inflation

    stopped in 1990.

After Deng announced the policy of speeding up reform, rapid development and further opening

    of China' door during his Southern Expedition to Shenzhen in February1992, people began to

    invest more and the economy started booming. Banks received the green light to expand credit for investment projects. {delete: The increase in aggregate demand led to serious inflation in late

    spring 1993.} From the end of 1991 to the end of 1996 money stock increased from 3177.8 to

    8802.0 or at an average annual rate of 22.5 percent per year. The inflation rates in 1993, 1994 and

    1995 were 13.2, 21.7 and 14.8 percent respectively. [Delete: Money supply in 1992 increased by 36 percent. Inflation in terms of the official retail price index reached 13 percent in 1993 and 22

    percent in 1994.]

    We now turn briefly to the effects on output associated with the large increases in money supply in the above episodes. Since the average rate of annual growth of real output was about 9.5 percent from 1979 to 1998, any increase much above that average can be considered large. Output increased by 170/147.6 or 15.2 percent in 1984 and by 192.9/170 or 13.5 percent in 1985 when money supply increased very rapidly in 1983-4. It increased by 260.7/234.3 or 11.3 percent in 1988, and by 449.3/398.8 or 12.7 percent in 1994. This is a part of the history of China’s

    economic fluctuations as measured by changes in total output.

To slow down the inflation in the mid-1990s the administrative and economic means of issuing

    credit quotas to regional banks and raising interest rates on deposits were again applied. A strong

    and well respected administrator in the person of Zhu Rongji was appointed to head the People's Bank to solve the inflation problem. Zhu was determined and feared by the heads of the provincial branches of the People's Bank. He told them that they would lose their jobs if the credit quotas were exceeded. The expansion of credits was put under control. Inflation was slowed down to 15 percent in 1995, 6 percent in 1996, 1 percent in 1997, and -2.5 percent in 1998. The rates in 1998 and 1999 were affected by the reduction in aggregate demand due to the Asian financial crisis which started in July 1997. Money supply in these years was increasing slowly also. Zhu Rongji became Prime Minister in March 1998 partly on account of his successful performance in control inflation in the middle 1990s.

    When he took office in 1998, Premier Zhu announced two macro-economic policy objectives, fully recognizing the “formidable challenges due to the financial crisis in Southeast Asia.” The objectives were to maintain a real growth rate at 8 per cent or more in 1998 and an inflation rate of not more than 3 percent. As President of the People’s Bank and later Vice Premier, Zhu Rongji deserved much credit in reducing the rate of inflation in China from 22 percent in 1994 to being negative in 1998 by reducing the rate of growth in money supply. Realizing the need to

    stimulate the economy he was prepared to raise the inflation rate target to 3 percent by adopting a more expansionary monetary policy. By September 1998 currency in circulation in China reached 10 trillion RMB, an increase of only 16 percent from a year ago. Total credit also increased by 17 percent in the same period. Since inflation had slowed down, nominal interest rates also came down. In March 1998 one-year deposit rate went down to 5.2 percent and one-year official lending rate to 7.9 percent. Commercial banks were allowed to set their lending rates within a fairly narrow range, between 10 percent below and 20 percent above the official lending rate.

    In the years 1998 and 1999, China was affected by the Asian financial crisis, although only to a moderate extent. Economic growth was slower. Perhaps at the time it would have been better for the People's Bank to exercise a more expansionary monetary policy by increasing credits and money supply more rapidly. Some economists think that it is possible to restrict the expansion of credits but not possible to increase money and credits when there is no demand for them. The impossibility to extend credit during times of slow growth because of lack of demand is known as a liquidity trap. In the Chinese case the government was trying to raise aggregate demand following the Asian financial crisis by increasing expenditures on the building of infrastructure. Instead of financing these expenditures by issuing bonds, the government could and should have financed some of the projects by issuing more money, or non-interest bearing government debt. It would solve the problem of the liquidity trap if such a phenomenon did indeed exist.

    To summarize the effects of monetary policy, the increase in money supply from 8802 at the end of 1996 to 17278 in 2002 amounted only to an average annual increase of 11.8 percent as compared with 22.5 in the five years before. There was price stability, and in fact a slight deflation, during this period as the data in Table 7.1 show.

    In the 1980s and 1990's monetary policies in China were exercised mainly by the administrative means of assigning credit quotas and the economic means of setting interest rates. Credit creation and contraction through the commercial banks by changing reserve requirements and open market operations were not possible because commercial banks were not operating effectively. The use of more modern means of macro-economic policy making will require some time to develop as there are problems in reforming the banking system as will be discussed in chapter 13. One problem in monetary policy yet to be resolved is the practice of simultaneous control of interest rates on loans and on deposits, often leading to insufficient profit margins for the commercial banks. These rates should be market determined to a larger extent than as practiced in China. This problem is recognized by the People’s Bank and the deregulation of interest rates is taking place gradually. As new commercial banks appeared to compete with the four large state commercial banks in the turn of the century, there is a need for the People's bank to monitor all commercial banks in order to prevent them from being financially irresponsible. In 2003, a

    Commission was established to take over the function of the People’s Bank in supervising

    the commercial banks.

    Macroeconomic developments up to 2004 are also consistent with the Friedman proposition that increase in money supply first leads to increase in output and then to increase in prices. Money supply increased rapidly in 2002-3 mainly as a result of the large inflow of foreign exchanges acquired through large trade surplus and large inflow of foreign investment. The acquired foreign exchange was converted into Chinese currency or reserves in the banks. This led to rapid expansion of money and credit and to increases in investment and output in 2003. The annual rate of increase in M2 was about 13 percent in January 2002, rose to over 18 percent at the end of 2002 and to over 21 percent in mid-2003, and was reduced to about 18 percent in the beginning of 2004. China’s real GDP increased at a 9.1 percent rate in 2003 and a 9.8 percent rate in the first quarter of 2004 as compared with 8.0 percent in 2002 (see Table 7.1). Signs of inflation began to show in the last quarter of 2003, with the rate of inflation rising to over 5 percent annually in July 2004 (the highest rate in seven years). From January to October 2004 overall level of domestic consumer prices increased by 4.1 percent while that of capital goods in circulation grew by 14 percent, a year-on-year increase of 3.3 percentage points and 6.5 percentage points respectively.

    During the period of 2002-2004 the central government again relied on administrative means to restrict the credits extended by provincial banks, and to limit the amounts of provincial investment projects. The People’s Bank also raised the interest rate. These policies were only partially successful as economic forces are often difficult to control by administrative means. If the Chinese government had revalued the RMB, raising its value as compared with the US dollar, the trade surplus of China would have been reduced and there would have been less inflow of foreign exchange. Money supply would not have increased as much. If the banks had had less money to lend there would have been no need for the government to restrict their credits which were used to finance construction projects. The alleged “overheating” of the economy during this period would not have taken place.

    We have discussed monetary policy mainly in terms of controlling the quantity of money. Historically economists have found a relation between change in the quantity of money and the

    rate of inflation. Inflation can be measured by the ratio p / pof a general price index for period tt-1

    t to the index for period t-1, minus one, as we have done using data in column 2 of Table 1.

    /p) =lnp-lnp. For tt-1tt-1Alternatively it can be measured by the natural logarithm of this ratio ln(psuch a large proportional change in the price index as from 0.9308 in 1960 to 1.0820 in 1961, the

    first measure of inflation gives 0.162, and the second measure gives 0.151, not very different

    from the first. For smaller proportional changes in the price index, the two measures are closer.

    To examine the relation between money supply and inflation we use the data in Table 1 on (1) the

    retail price index p, (2) China's gross domestic product in real terms Y, (3) the amount of currency in circulation as a measure of money supply M, (4) the ratio of money supply and to real GDP

    M/Y, (5) the inflation rate as measured by lnp-lnpThe rate of increase in money supply, which tt-1.

    can be measured by lnM-lnM, forms the basis of much of our discussion of monetary policy tt-1

    and its effect on the inflation rate. Note the three inflation episodes of 1985, 1988 and 1993-1995.

    It is a useful exercise to estimate a regression of the inflation rate lnp-lnp on the rate of increase tt-1in money supply lnM-lnM of the same year and/or of the preceding year. tt-1

As we have pointed out, a better explanation of inflation is to use the rate of change of the ratio

    M/Y instead of the rate of change of money supply itself. This formulation is based on the

    quantity theory of money. The quantity theory states Mv=pY, or p=v(M/Y), where v is the income velocity of circulation of money, measuring the ratio of money income pY to the stock of money M. Secondly, we need to specify the possible delayed effects on the inflation rate of the

    change in the ratio M/Y of the same year and of the preceding year, and of the inflation rate of the

    preceding year. Thus (M/Y),, (M/Y) and ln (p/p) are three variables that can affect the tt-1t-1t-2

    current inflation rate. Change in the ratio M/Y in the preceding year enters as an explanatory

    variable because of its delayed effect. Inflation rate of the preceding years enters because it can

    capture the effects of the changes in M/Y further back in time, as the dependent variable of the

    preceding year already captured the delayed effects of changes in M/Y of the previous years. Thus we have two more variables to explain the inflation rate in addition to the ratio M/Y. Both are used to capture the delayed effects of M/Y. These delayed effects are imbedded in the formulation

    of a dynamic equation to be discussed in the next section. Chow (1987) had estimated such an

    equation using data up to 1993. This equation is useful not only for our understanding of the

    effect of monetary policy on inflation, but also for the formulation of monetary policy. We can

    determine the right credit quotas to impose, once its effect on inflation is quantitatively


7.3 Econometric Analysis of Inflation and of Monetary Policy in China

    Consider a general method for estimating the delayed effects of a variable x on a variable y. (Note the distinction between this low case y and real GDP, denoted by Y.) In the example of the explanation of the price level, y is the natural log of the price level p and x is the natural log of the ratio M/Y of money supply to real output. The relation between these two log variables can be

    justified by taking the logarithm of both sides of the equation p=v(M/Y) without considering the delayed effects of x=M/Y. Natural logarithm is used also because a linear relation between log p and log (M/Y) is a better approximation to the data than a linear relation between p and M/Y. In

    is affected not only by x but possibly by past values of both variables. the dynamic equation, yt t

    In the long-run, let us assume that there is an equilibrium relation between the two variables

    given by y-?-?x=0. The dynamic equation has an error correction mechanism built into it. The 01

    correction mechanism specifies that a positive deviation u=y-?-?x from equilibrium in the t-1t-101t-1last period will assert a negative effect on the change ?y=y-y of the dependent variable in the ttt-1

    current period. Thus the coefficient ? of this deviation uin the regression of ?y is negative. t-1 t The dynamic equation proposed attempts to explain ?yby ?x, past changes of both x and y and t t

    by this error-correction term, i.e.,

    ?y=?+??x+??x+??y+?u+? (1) t01t2t-13t-1t-1t

    where u=y-?-?x and ? is a random residual.. t-1t-101t-1t

    Let lnp be the dependent variable y and ln(M/Y) be the explanatory variable x. Using M2 for M Chow and Shen (2005) plotted lnP against ln(M2/Y) for the period 1952 to 2002 as shown

    in Figure 7.1.

    Figure 7.1 Plotting log(P) against log(M2/Y)













The points in Figure 7.1 fall along a straight line approximately. The exceptions are the years

    after 1998 when prices increased less than the fitted line would predict probably because of low

    aggregate demand after the Asian financial crisis of 1997-9 in such a period increase in money

    supply may have a smaller effect on prices.

If we fit a linear regression using the data of Figure 7.1 we obtain

    2R?0.9639log()0.71270.3738log(/),PMY??? Adjusted 2(0.031)(0.0102)log()0.4170.4637log()0.5768log()PMY???? Adjusted 2(0.109)(0.033)(0.0729)2R?0.9646

The test of the hypothesis that the coefficients of log(M2) and log(Y) are equal in magnitude but

    opposite in sign gives a F statistic of 7.88, rejecting the null hypothesis at 1% significance level. By using log(M) and log(Y) separately, we get 2

Report this document

For any questions or suggestions please email