You can create Classical or Keynesian type of models for these countries. In these model you can make specific specifications for the exchange relationships among other relationships. An old book, "The Macroeconomics of Open Economies" by Richard Morley Elgar, 1988 can help you with simple exchange rate specifications. Another practical book: "Practical Application of Approximate Equations in Finance and Economics," by Manuel Tarrazo, Quorum Books, 2001 has examples of models for Klein, Sargent and others.

Some things to be careful about are 1. When you compare countries, you must have some idea as where they stand with regards to their Natural Rate of Output, or Non-accelerating inflation rate of unemployment, NAIRU. 2. When you deal with subsets of countries (16 in your case), and you do comparative changes (dynamic or static), you have to control for the rest of the world (ROW). This may mean that you have to do a general equilibrium rather than a partial equilibrium model.

ISSUE: Effect measurement of trade volume stimulated by changes in exchange rate

EXCHANGE RATE & TRADE VOLUME: It is a common conception that weak exchange rate stimulates exports and, thus, contributes to an increase in trade volume, i.e. export more = improve trade balance; it may also have been said that strong exchange rate (strong currency) may have the opposite effect. This "general statement" may not be true all the time. If the economy is predominantly engaging in assembly industry for export where components have to be imported before assembly and re-export to earn export money----strong local currency would be helpful in importing "relatively cheaper" components. Weak currency in this type of economy would not be helpful to raise export volume. With this potential scenario in mind, your 16 countries must be homogeneous if you are looking at them as one group or divided into comparable subgroups if you are doing comparison study. make sure that the 16 countries are not a mix bag of apples and oranges.

EFFECT MEASUREMENT BY TWO-COUNTS POISSON DISTRIBUTION: You are looking for the effect of change of exchange rate depreciation, i.e. what is the effect if the exchange rate depreciates, this is to compare the change in trade volume resulted from changes in exchange rate at t1 (prior period) to t2 (subsequent period). One possible tool is comparing two counts in Poisson distribution:

Z = (R1 - R2) / sqrt ((R1 / t1) + (R2 / t2))

... where R1 = N1 / t1; R2 = N2 / t2. Ni represents frequency counts and ti represents time period. The following decision rule applies: H0: R1 = R2 and HA: R1 not = R2. The null hypothesis assumes that there is no change of the event. the alternative hypothesis claims that there is a change in the event as the result of the introduction of the stimulus, i.e. currency depreciation. The design is a straight forward 2X2 table: (t1; t2) X (N1;N2).

REFERENCE: See Test 8 in Kanji 100 Statistical Tests. An article by Krishnamoorthy and Thomson may also be useful for introductory reading. See Attached files.

Hello dear Sir;

As you know sometimes; governments reduce the value of their money against common reference Money (Such as USD, Euro, GBP, JPY and so on) them-self in order change  the balance of Export versus Import share but keep in mine there are a lot of items such as price of precious metals (especially Gold), inflation and De-valuation of currencies leaded to fluctuation of exchange rates. Also you can always use forecasting models for exchange rate of currencies.

But in my case it was; cost estimation and feasibility  study of one Canadian mine; I have considered "one item" in modelling of Cash flow, DCF, NPV, PP, ROR and so on: fluctuation of exchange rate of CAD versus USD and I have applied Sensitivity Analysis and Spider Diagram in order to show the effect of exchange rate on  the mentioned responses of economical model.

I hope this short reply could be useful.

Regards;

Hamed

The best econometric model to address this question is the ARDL model. It can be used to assess both the short-run and the long-run effects of exchange rate depreciation on trade balance. But, such model applies especially for time series analyses. As you have a panel data, the panel cointegration model proposed by pedroni could be used instead. The FMOLS and DOLS techniques may produce good results.

You can estimate a panel VAR with fixed effect that will account for the effect of trade balance exclusively due to exchange rate variations. My concern is, you're bound to encounter the problem of negative values for trade balance that will prevent you from taking the logarithm of this variable. Logarithmic conversion of raw data has many advantages, as it prevents heteroscedastacity in panel VAR estimation and helps us avoiding the problem arising from different units of measurement used for different variables. I suggest that you use exports and imports as different variables and test the hypotheses relating to the dynamic behavior of exports and imports due to exchange rate appreciation/depreciation. Cheers.

I agree with Juehui Shi's comment, which is very straight. I also have written a paper on the effect of real exchange rate on export performance......., which is coming in journal soon. Depending on data quality, you should be aware how you want to calculate exchange rate may be the major issue. In terms of method, your data quality will determine which to follow?

Allow for heterogeneity in the elasticity of devaluation and incomes. My other thoughts are in