Exchange Rate Exposure

Dominique and Tesar 2006 – Journal of International Economics 68

Summary by Paolo De Angelis

As any introduction, the first part of the article describes and conjects about the main thesis: a possible relationship within exchange rate exposure and firm value.
The Authors talk about their work and how they builded such strong hypothesis to demonstrate which is the connection and doing that they explain two objectives: understand how much these firms are exposed to exchange rate fluctuation and investigate why some firms are exposed and some not.
Then there is an explanation about tools used to do the research and statistic technique used to test different ideas about correlation among main features characterizing these frims.
A very important inference is about the firms involved and not; talking about Ford Motor Company the Authors explain how not only international operating firms are influenced by exchange rate but also local ones, infact exchange rate does influence sales competition in the country where local firms operate.
As a resume it explain the third hypothesis about firms exposed due to the level of trade that they do; but, instead of the two previous hypothesis, using proxies because of data lacking.
All of this using two diffused indexes: a mark-up index and the Herfindhal Hirschman Index, that is very suitable because of his power to evaluate firms not only relating to market share but squaring it to avoid both arithmetical calculation problem and becoming results more representative of real situation.
Thus make in evidence that despite of too industrialsized index, a cross-country regression describes that country, where industries is concentrated, are more exposed, regarding to (cross-country) datas.
At the end the Authors provide an interesting assumption about the continous changing of firms exposed to exchange rate risk. Considering that, big firms have well-managed hedging systems so we can deduct that every firm is continously exposed but according to D. and T. their adaptive strategies permit to observe an ongoing phenomenon of exposition.
Afterthat there is a very short but important analitical explanation about the exchange rate exposure existence and influence. Using CAPM and its hypothesis Domnique and Tesar demonstrate why the B2, i (delta)St is different from zero and this is a part of a even more bunch of ideas that in the last 20 years are discouraging the Capital Asset Pricing Model.
Doing that they find a good reason, same as the real world, that there is a truly relation, between excess returns and exchange rates,analytical speaking. Even do this Beta will be linked with: “… a set of factors that could proxy for plausible channels for exposure.”
The importance of the third part of the paper is based on a wide description about how datas had been gathered. There are not a lot of words to spend on it but is important to understand how the two Economists constructed their theory. Personally this short part of the paper is the key to realize the same results, specially about different source and methods they used to organize the researches.

The core of the article is well designed to compare datas and assumption like a swinging path throrough the essay. The first question is about finding the relevant exchange rate to compare the results. The most diffused is the trade weighted exchange rate which is unlikely to be affect by possible mistakes about semi-exposure of some firms to the currency, but as suggested by Williamson it could be solved creating specifing exchange rates.
The Authors analyze two graph showing the relations among country’s firm or industry and different exchange rate; thus making a comparison about different exposure related on different country and indicating with “any exch rate” a sort of correlation among dual exchange rates.
It is easy to observe how many difficulties are in this comparison because of two reason: the time exposure to exchange rate and the hedging process done by international firms. Infact in a third graph using an industry specific trade based exchange rate the Authors improved the low possibility to find a relation in the sequent analysis because of using both import\export rates there are not relevant results (as shown in Fig. 2 on page 11).
The definition of market index to use to compare results takes a relevant discussion. The first is about value weight index that eliminate macroeconomics and negative exchange rate’s effect on cash flows.
But there is another best index called “equally weight index” that is like to compare different sized firms because of weighted evaluation of their portfolio or fund composition (example found on Investopedia: Rydex S&P Equal Weight ETF).
Then the Authors talk abuout World Index of Datastream which unfortunately reveals poor in explenating return effect.
Secondly there is a possible correlation among market index as a whole and exchange rate, infact it seems that marginal firm-level exposure is small even though aggregate market-level exposure is high. Anyway, looking also at possible multicollinearity, is biased because the market index absorb the impact of movement in exchange rate. This will probably suggest a Local index instead of an International one.
The Fig. 4 showed in section 4.2 is one of the clear explanation aboout how D & T are doing their regression. Here, using rolling selection of variables, the return horizon influence the exposure at 5% ( interval of confidence at 95%) on dollar trend with differences in every country but (except japan) the overall result is a defined increasing from 1 to 52 weeks. This mean that weekly exposure well evidences the relation searched.
“Averaging across significant dollar exposure betas across countries, the data suggest that a 1% change in the exchange rate is correlated with a one-half percent change in stock returns.” This expression indicate the result on observation of magnitude: easily, skimming that firm with low exposure the Authors undestand how in some countries like Thailand and China there are few organisation exposed to exchange rate but these are well exposed. Furthermore it involves in previous decision about Equally WLI.
To conclude the first part of regression, two ways are used to analyze datas: first considering only last five years it measures the extent of exposure; then dividing datas into separate subperiods and evaluating betas consistency in these sub-samples. As a result there is not one specific period for all the firm but it vary on subperiod and subsamples of firm. Infact this is a solid confirmation about continuos hedging and reacting of firms at anytime and different level of exposure.

HP1: dollar exposure is probably not exposed to firm’s charateristics and size, but quite exposed to industry’s affiliation , in particular for electric gas and water industry. Infact for others the figures show different exposure signs but here is definitely one way exposure, nevertheless unuseful.

HP2: finally a clear proof of international activity firm’s benefit gained by hedging and dollar exchange rate fluctuation. As shows in Tab 5 international exposed firm are likely to be favoured by their exposure. Confirming initial conjecture given by the Authors.

HP3: relates closely to international trading and links with dollar exchange rate exposure. There are not strong evidences about this exposure but fragmented one like for importing country ( Italy, Germany and Japan) and, excluding paricular positive HHI’s information about France and UK, for Japan again because of markup index.

At the least these tables show a relevant information about three indicators considered: 1- international activity 2- industry level exports 3- industry level imports and the positive HHI value confirm the country by country regression’s results.

Reading conclusions the idea which come is basicly the lack of a particulare sector or industry , rates or size, that affect and characterize exposure level at exchange rate. Anyway it seems that small and international oriented firms are likely to be influenced but the concusion is definitely that it depend on feature of the firms.

Personally the results does not surprise me: since the first conjucture given by Dominiquez and Tesar it is clear that variety of firm and industry’s composition in the world, different culture and econmic structure, would probably have a weak link and a non-regular behaviuor.
Infact globalization is a recent phenomenon and only few country are strongly based on international events; moreover after Bretton Woods, regarding to the dollare exchange rate.
But results given by this article are based on proxies and adjusted datas because of the uncomplete availability of them.
I conclude saying that this work gives a strong confirmation of exchange rate exposure and open the path to undesrtand why after 40 years, the dollar’s fluctuation affect worlwide economic and puzzle me about hedging vantages.