![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
This paper puts forth a Neo-Kaleckian open economy model of two countries in order to investigate adjustment of US–China external imbalances. First, a stylized fixed mark-up model is presented, and discussed based on graphical analysis. Second, we present estimates of bilateral income and price elasticities of imports. Third, we employ the model for simulation analysis. Specifically, we randomly distribute expenditure change across government, investment and imports and calculate the exchange rate change necessary to lead to an equal change in the bilateral external imbalance. Doing so repeatedly allows us to estimate probability distributions of endogenous variable changes.
Notes
1. We will not review the literature on causes of global imbalances. See, for example, Eichengreen (Citation2006), Ocampo (Citation2007) and Vasudevan (Citation2010).
2. The question can be seen through the lens of transfer theory. Assume the expenditure changes in the two countries implies a financial transfer from deficit (transferor) to surplus (transferee) country. The question then is whether the transfer will be effected; that is, whether the trade balance improves by exactly the amount of the expenditure change. If output is demand-determined and the exchange rate is fixed, it will always be under-effected unless it directly and sufficiently impacts autonomous expenditures on imports. A fully effected transfer implies that the transferor country finances the expenditure change with increased net export earnings. Since import propensities are much smaller than unity, the endogenous import changes following income changes cannot be sufficient to that end. If, however, the transfer to a significant part directly affects autonomous import demands – which add to the endogenous trade changes – the transfer can be effected, and can even be over-effected. See Johnson (Citation1956) for a relevant discussion.
3. Other papers have focused on cyclical as well as financial issues. Here, we will do neither– the model is presented purely in terms of (real) flows. Moreover, the model is not scaled by the capital stock, and concerns only the medium run. It follows that the rate of utilization is endogenous, but does not have to be in the long period. For an early contribution to that debate, see Auerbach and Skott (Citation1988).
4. An appendix details the equations of the full model, with the US, China and the rest of the world indexed by 1, 2 and 3, respectively. Table shows a symbolic Social Accounting Matrix (discussed further below), which might be helpful to quickly illustrate the accounting of the model.
5. The model is fairly straightforward, and to save space we leave a brief discussion of stability for the appendix. Here, it might suffice to say that the model is dynamically stable if both countries have positive multipliers in isolation; this result is unaffected when output prices are taken into account, since mark-up, nominal wage and labor productivity are constant.
6. US goods exports and imports to and from China are from US Census Bureau. Unit value index of exports and imports of US (Base: 2005=100) is used to deflate export and import series. Seasonally adjusted quarterly GDP (Base: 2005=100) is used to proxy US income (Source: Bureau of Economic Analysis). Similarly for China, GDP series is drawn from National Bureau of Statistics of China. Since quarterly GDP of China is available only from 1992 onwards, earlier data points are constructed using the GDP growth rate by Abeysinghe and Rajaguru (Citation2004). GDP is deflated by the Consumer Price Index (Base: 2005=100) and seasonally adjusted. The real exchange rate is measured with CPI of China and CPI of USA respectively; price indexes are from International Financial Statistics (IFS).
7. The results from ADF unit root test suggest that all variables are non-stationary, i.e. I(1), except the trade balance, which is found to be stationary.
8. A trend term is included in import and import equations, since we find a strong trend in these series.
9. This method applies the maximum likelihood procedure to determine the presence of cointegrating vectors in non-stationary time series. If cointegration is detected, then normalized cointegrating vectors provide estimates of long-run relationship. Johansen’s cointegrating analysis involves estimating a Vector Error Correction Model in reduced form. Johansen has proposed two likelihood ratio statistics, the trace static and the maximum eigenvalue statistic, both of which determine the number of cointegrating vectors based on significant eigenvalues of the matrix of parameters. The trace statistic tests the null of r cointegrating vectors against the alternative of more than r cointegrating vectors, while the maximum eigenstatistic tests the null of r against the alternative of exactly r + 1 cointegrating vectors. If the variables under consideration are cointegrated, i.e. the long-run relationship is established, then the cointegrating vector is normalized with respect to the relevant variables.
10. The SAM underlying the model is an aggregated version of the SAM used in von Arnim (Citation2010). The raw data stem from a number of sources, including UN SNA, BEA, IMF Direction of Trade Statistics (DOTS) and IMF International Financial Statistics, (IFS). Functional distribution of income for China is taken from Rada (2010).
11. The focus of the simulations is on the trade linkages and trade elasticities. As can be seen in Table 8, the distributive-demand calibration is chosen with equal parameters in both countries. That is not to argue that this must be the case, but in order to choose parameter values that are plausible, and broadly supported by empirical estimates. Importantly, while the savings differential is assumed to be the same, the aggregate private savings propensity differs drastically. The (gross) aggregate savings propensity is 14% in the US, but 50% in China; see the flows of funds row (FOF) in Table 6.
12. The probability distribution functions are estimated with a (normal) Kernel Density Estimation (KDE) procedure. The bandwidth is chosen ad hoc, such that tails of the distributions are not ‘too far’ from the lower and upper limits of the sample data.
13. The inflation statistics are probably of less concern. The key issue is that inflation in both countries is ‘imported,’ since all other items in the cost decomposition are held constant.
14. FIRE: Finance, Insurance, Real Estate.
15. The global supply of reserves could be affected through emergence of new currencies, including the Euro and a (fully convertible) Yuan; the global demand for reserves could be affected through reduced necessity for emerging market countries to ‘self-insure’ against balance of payments crises and resulting IMF conditionality. Discussion of these proposals goes beyond this paper. Ocampo (Citation2007) provides an overview.