Predicting and stabilising Exchange Rate movements

The Government should relentlessly improve the national productivity through measures such as developing infrastructure, supporting innovation, pursuing foreign direct investment, combating any existing state sector inefficiencies and corruption and improving the policy making capacity of Government institutions.  Such measures will enhance real national income leading to lower rate of inflation

In a central banking monetary regime exchange rate is an ‘endogenous’ variable, i.e. it is determined by several other short run and long run economic variables. Thus, it cannot be directly controlled by the central bank or any Government entity.  It cannot be directly controlled by any private entity either.  However, there can be an indirect control by the central bank and other relevant institutions by managing variables such as foreign exchange inflows and outflows, money supply, real output and inflation.

Short run as well as long run

Exchange rate movements are short run as well as long run. The short run is roughly one year or less, while the long run is over one year.  Short run movements are driven by foreign exchange inflows and outflows arising from international trade, services such as remittances, travel and tourism, and inflows and outflows into the financial and capital accounts.  Since Sri Lanka permits some speculative capital inflows such as investments by foreigners in domestic equity and bond markets, the differential between domestic interest rate and foreign interest rate also plays a role in the determination of the short run exchange rate.  Because those inflows and outflows fluctuate without any predictable path, in theory, the short run variations in the nominal exchange rate cannot be predicted, and it follows a ‘random walk’ around the long run trend. A random walk is a path that no one can predict. 

In addition to those economic factors, adverse expectations by market players also affect the short run exchange rate. In some extreme cases, when market players feel that the Central Bank is acting irrationally to defend a specific fixed nominal exchange rate, such market players could mount a speculative attack against theCentral Bank  In a recent Daily FT article Dr. Wijewardena has explained one such case that happened in Thailand in 1997 and the grave economic fallout that followed the attack.

Long run fluctuations arise from economic variables such as money supply, real output and inflation. The long run nominal exchange rate between the US Dollar (USD) and Sri Lanka Rupee (LKR) has been gradually depreciating since the setting up of the Central Bank of Sri Lanka in 1950.  This is primarily due to relatively higher money supply in Sri Lanka than its trading partners facilitated by the central bank as it has been enjoying the ‘monetary independence’.  This was exacerbated by the lower real national income in Sri Lanka.

The long run trend of the exchange rate, especially exchange rate between the USD and LKR can be easily found by anyone who type USD LKR into Google search App which provides a graph of fairly long run behavior of the nominal exchange rate, as well as the daily rate. Same information is found on the website of the Central Bank of Sri Lanka, and from its daily announcements. The graph shows an upward trend indicating the gradually depreciating nominal exchange rate.

As was explained earlier, the two main variables determining the long run USD LKR nominal exchange rate are the inflation rates in the USA and Sri Lanka.  Historically, Sri Lanka inflation has been higher than the US inflation rate leading to a gradually depreciating exchange rate.  The underlying causes of long run inflation rates are relatively higher money supply coupled with relatively lower growth of real national income. This has generated a historical upward trend in USD LKR exchange rate.  

Forecasting model

Given Sri Lanka’s current situation, in a forecasting model, the trend or ‘time’ variable can replace all other economic variables such as inflation, money supply and national income. Thus, a simple ‘regression equation’ with ‘time’ as the only predictor provides a decent path for the long run nominal USD LKR exchange rate. Thus, exchange rate can be predicted by the equation exchange rate (e) = a + b. f(T); where ‘a’ is a constant and ‘b’ is the trend coefficient.  ‘T’ is the time variable where T =0 in 1950, and T=1 in subsequent days, weeks, months or years depending on the frequency of prediction required, and f(T) is a linear or non-linear function of time T.

The equation for exchange rate combining both short run and long run behavior for any given year ‘T’ can be written as e = a + b. f(T) + c, where ‘c’ is short run fluctuations of the exchange rate. Since short run fluctuations around the long run path of the nominal exchange rate fluctuate randomly, ‘c’ could be a negative number or a positive number or zero in any given day, month or year.  In the long run it averages out to zero.  

In addition to the path predicted by the above equation, there has been violent and sharp depreciations of the nominal exchange rate from time to time.  A good example is what happened in 2021 and 2022 when the exchange rate jumped from 200 Rupees for one US Dollar to 350.  This was because the Monetary Board of the Central Bank attempted to directly control the exchange rate using official foreign exchange reserves disregarding ‘endogeneity’ of the exchange rate and eventually bankrupting the country.  To account for those incidents, the regression equation can further be modified to e = a + b. f(T) + c +d.D, where ‘d’ is the coefficient attached to a ‘dummy’ variable D. The dummy variable takes the value zero if the central bank and the governing board are competent in monetary and international economics or it takes the value one, if the central bank and the governing board are not so competent. The coefficient ‘d’ will have a large positive value implying large sporadic depreciation of the exchange rate if the central bank and the governing board are not competent, and no impact on the exchange rate if they are competent.

Serious economic consequences 

There are serious economic consequences of the depreciating exchange rate. The depreciation in turn increases the price level and inflation thus creating a vicious cycle of such depreciation leading to inflation and then inflation leading to further depreciation.  This depreciation increases the cost of essential consumer goods such as fuel, food, medicine and other essential services.  Once prices of those goods and services rise, they remain high due to the phenomenon called ‘downward rigidity of prices.’ The increase in consumer goods prices further impoverishes poor segments of the population and fixed income earners such as Government employees.  The Government becomes helpless to provide any assistance to those segments of the population due to fiscal constraints, and to avoid the vicious cycle of higher wages and higher Government handouts feeding the inflation and leading to further depreciation of the currency. However, the Government should find some mechanism to support those vulnerable groups by reducing expenditures somewhere else and should have long run strategies to combat the vicious cycle of high inflation and depreciation of currency.

The path for Central Bank

The Central Bank may charter a path for long run nominal exchange rate based on long run inflation rate consistent with ‘price stability’ defined in monetary economics underpinned by growth in money supply and real output growth thus ensuring a stable long run real exchange rate. The short run fluctuations around the long run nominal path arising from strengths or weaknesses of external sector variables may be managed either through permissible action or strategic inaction. The Government should relentlessly improve the national productivity through measures such as developing infrastructure, supporting innovation, pursuing foreign direct investment, combating any existing state sector inefficiencies and corruption and improving the policy making capacity of Government institutions.  Such measures will enhance real national income leading to lower rate of inflation.  The ultimate result will be having a stable real exchange rate in the long run while facing moderate fluctuations in the short run around the long run path of the nominal exchange rate. When this happens the regression equation predicting nominal exchange rate explained above will have to be replaced by a structural model with economic variables explaining the exchange rate movements, instead of ‘time’ variable.

(The author worked at the Central Bank of Sri Lanka as the Director of Economic Research, and an Assistant Governor. At present he teaches Economic Theory and Policy at the University of Iowa, USA. He may be contacted through [email protected].)