欧洲货币联盟单一货币的影响

加拿大Assignment代写范文:“欧洲货币联盟单一货币的影响”，这篇论文主要描述的是在欧洲货币货币联盟的保障下，欧元的使用比较的稳定，但是在欧洲货币联盟如此优秀的联盟，也饱受着限制价格收敛的困扰，使得这种完美的欧洲货币联盟也无法持续的保持。

Let x be the degree of price dispersion among a group of countries where j∈ (EMU, Non-EMU) and r∈ (≥1999, 1998). EMU identifies the treatment group (here the 11 original EMU countries), Non-EMU refers to the control group of countries outside the eurozone; ≥1999 represents the treatment period (i.e. 1999 and after), 1998 refers to the pre-treatment period before the introduction of the euro

We want to estimate the effect that the introduction of the single currency has had on price dispersion in the eurozone ? denote this ?treatment effect? Δ. Ideally the treatment effect would be estimated by comparing the outcome for the treatment group after receiving the treatment,, with the outcome the same group would have experienced had the treatment not occurred:However, the ideal estimate S* suffers from a missing data problem because we cannot simultaneously observe both outcomes for the same group of countries. With non-experimental data, the missing counterfactual,, needs to be replaced by an observable variable that serves as proxy.

The Difference-in-Differences (DD) approach makes use of data for a control group that does not receive the treatment but experiences some or all of the other influences that affect the treatment group. The DD strategy assumes that over time, the no-treatment average outcome for the treated follows the same path as the no-treatment average outcome for the non-treated: E() = E(). Algebraically the DD estimate can be written.By comparing the before-after change in the outcome of the treated with the before-after change in the outcome of the non-treated, the DD approach i) nets out fundamental differences in the two groups; and ii) eliminates common trends in the outcome variable affecting both groups. However, as Meyer (1995) notes, the DD methodology rests on the assumption that there are no other factors during the treatment period which affect the treatment and control groups differently. This is one of the main threats to the validity of inferences from this approach since changes in other laws/policies or macroeconomic conditions are not likely to always influence both groups in the same way. Meyer (1995) also notes that the inclusion of multiple pre-intervention periods or additional control groups can reduce the severity of this threat, but data limitations prevent such measures from being used in this study.

In its simplest form, the DD approach will be implemented in a linear regression framework for each good/car model as shown in Tables 4A/5A.In the regression above, the dependent variable is the estimated coefficient of variation of common-currency prices for a given group, and i) D≥1999, ii) DEMU and iii) are dummy variables equal to one when i) r = ≥1999, ii) j = EMU and iii) r = ≥1999 and j = EMU simultaneously, zero otherwise. The coefficients in eq. (3) are linked directly to those in eq. (2) as highlighted by Lutz (2003).The dummy variables capture influences that are not directly measured but are specific to the treatment and control groups and/or to periods before and during the treatment. The coefficient β captures time-variant influences that are likely to affect both groups such as improvements in transport technologies.If β is large in absolute value, it suggests that period-to-period changes in the dependent variable are not unusual and further evidence on its variance over time might be warranted, however, Tables 4 and 5 indicate that β is typically close to zero and insignificant for the majority of goods under study. The coefficient γ captures time-invariant influences that vary between groups such as the distance between markets or whether countries share a common language. Tables 4 and 5 show that γ is also typically close to zero and insignificant, suggesting the treatment and control groups are relatively similar. However, most important is the coefficient δ, as this corresponds directly to the DD effect in eq.

In addition to the simple OLS regression in eq. (3), it is possible to extend the DD methodology by combining observations across goods/models into a grouped panel. DD results for categories of goods can then be produced by running a pooled-OLS regression using the framework below.The appeal of the DD approach comes from its simplicity as well as its potential to circumvent many of the endogeneity problems that typically arise when making comparisons between heterogeneous groups. However, the DD approach also has its limitations.Bertrand et al. (2003) suggest that the standard DD approach may suffer from a serial correlation problem which produces standard errors that may understate the standard deviation of the DD estimator. This leads to over-rejection of the null-hypothesis that the treatment has no effect. One method of overcoming the serial correlation problem, which is particularly applicable to this study, involves including dummy variables for each post-treatment period as shown in equation.

In addition to solving the serial correlation problem, this methodology allows the researcher to compare one period with another and thus determine when the treatment effect was largest ? for this study, it also makes it possible to see whether the introduction of euro-notes and coins decreased price dispersion. The drawback of this methodology is the number of variables included in the regression. Due to the low number of observations available for each good/car model, this procedure is not applicable to the individual regression that stem from eq. (3). However, the grouped panel datasets used in eq. (4) provide sufficiently high degrees of freedom to make the analysis useful.

Other criticisms of DD estimation revolve around whether the treatment and control groups accurately represent the true population. In this study, Figure 3 provides evidence to support such concerns, however, these issues can only be resolved at the data collection stage. One other prominent concern has been whether DD estimates ever isolate a specific intervention or causal relationship. Advocates have attempted to improve the rigidity of their DD conclusions by including additional control variables and thus isolate a specific intervention. However, due to the large number of potential variables involved, and the negligible results of Lutz (2003), I leave this as an avenue for future study.

The results of the DD estimates of the single currency effect are presented in Tables 4-6. First consider Table 4A, which contains estimates for the 25 goods included in the DrKW dataset. Each individual regression consist of just 12 observations and suffers from low degrees of freedom and potential serial correlation.Nevertheless, a significant majority (18 of the 25 goods) report falling price dispersion after the introduction of the single currency. In addition, the median DD estimate for all 25 goods is notably negative thus providing further evidence to support the European Commission?s claim that the single currency has reduced price dispersion within the eurozone. However, if we only consider the two statistically significant results, the evidence for price convergence is far from overwhelming - one (shampoo) is consistent with the EC?s claim, while the other (camera film) reports price divergence.

At a more aggregate level, Table 4B provides evidence for different categories of goods within the DrKW dataset. Of the three categories, non-tradable goods have experienced the largest fall in price dispersion followed by non-durable goods. In contrast, durable goods experienced a slight rise in price dispersion following the introduction of the euro. These results can be explained by past fluctuations in the levels of price dispersion. As Figure 2 and a number of empirical studies [Rogers (2001) and Sosvilla et al. (2002)] suggest, non-tradable goods tend to have much higher levels of price dispersion than tradable goods. Similarly, durable goods tend to have the lowest levels of price dispersion because their higher prices encourage early arbitrage. Thus, there is greater potential for price convergence among non-tradable goods since prices have converged less in the past. Interestingly, this result differs from both the preliminary analysis and the findings of the other studies mentioned above. Important here is the introduction of euro-notes and coins.

Table 6A indicates that price convergence increased after 2001 for both non-tradable goods and non-durable goods while it fell for durable goods. This may be because durable goods are more expensive and thus tend to be purchased using non-cash payment methods ? therefore the introduction of notes and coins would have little effect on durable good purchases. In contrast, the lower-prices of non-tradable and non-durable goods are more readily associated with cash purchases. Thus, when notes and coins were introduced in 2002, more consumers may have become aware of price differences due simply to the physical act of handing over money. This increase in price transparency raises competition between national markets and could causes prices to converge. However, if purchasing methods can explain the disparity in the data, this implies that consumers are suffering from a form of money illusion.

An alternative explanation that preserves the rational consumer hypothesis, relates to price rounding of low value items. Before the introduction of euro notes and coins, goods are sold in national currencies and are subject to rounding conventions. For example, you might have had the 15 franc or 2 DM cup of coffee before the introduction of euro notes and coins. However, once you change from national currencies to a single currency, everybody rounds, say to two euros. This doesn?t mean consumers are aware of what, say, a coffee costs in other countries, but nevertheless brings about price convergence and is perhaps a more plausible argument than the money illusion hypothesis. An extension of this line of reasoning relates to the pricing strategies of multinational companies. For low price items in particular, it may be more cost effective to adopt the same pricing strategy in different markets e.g. charge the same for a Big Mac or 1.5 litres of Coke whether you are in Paris, Berlin or Athens.

Table 4B also provides a DD estimate that is pooled across all 25 DrKW goods. Although statistically insignificant, the DD estimate implies that price dispersion has fallen 13.2% in the five years following the introduction of the single currency10. According to results of the dynamic regressions presented in Table 6A, the rate of convergence has remained relatively stable between 1999 and 2003. Overall then, the DrKW results appear to provide evidence in favour of the price-equalising effects of the euro, although not at statistically significant levels.

Moving to the car price data, it is immediately obvious from Figure 1 and a comparison of Tables 4 and 5 that car prices have significantly lower levels of price dispersion than the goods in the DrKW dataset. There are two potential explanations for this difference: i) cars have already experienced price convergence because they represent a profitable form of arbitrage due to their high prices; ii) the inclusion of taxes in the DrKW dataset may increase price dispersion. Despite the differences in the datasets, Table 5A indicates that the majority of car models (11 out of 18) experienced price convergence during the treatment period. However, unlike the DrKW data, a comparison of the statistically significant results shows 3 models (Seat Ibiza, Volkswagen Golf and Honda Civic) that indicate price divergence, compared with just 1 (Ford Fiesta) corresponding to the price convergence hypothesis.

In terms of car size, Table 5B indicates that large car models have experienced the greatest fall in price dispersion over the treatment period. In contrast the more popular, medium sized models encountered diverging prices while dispersion among the smaller models remained relatively stable. Since prices increase with car size, these findings are not consistent with the result that arbitrage costs fall as the price of the good increases. However, when assessing the profitability of an arbitrage transaction it is also important to consider the quantity demanded. For cars, it is the medium sized models that are most highly demanded, and thus represent the most profitable arbitrage transactions. Presumably then, it is these medium sized models that drove the earlier price convergence found in Gaulier and Haller?s (2000) study of car prices between 1993 and 1999. However, as prices converged within the medium sector, arbitrage agents should increasingly focus on the smaller and larger segments where price dispersion remains relatively high. The results from this study suggest that such a move occurred over the period 1999-2003.

Now consider the broader DD estimate measured across all car models - shown in Table 5B. This result indicates that price dispersion increased by 3.9% for the five years since the introduction of the single currency. While this contradicts the EC?s claim of price convergence under the single currency, it carries little weight due to the very low statistical significance of the result-indicated by a t-value that is very close to zero. However, the dynamic analysis in Table 6B also casts doubt over the price convergence hypothesis. The results show that prices tended to converge between 1999 and 2001 but then diverged following the introduction of euro-notes and coins in 2002. Since cars are rarely purchased using cash, the introduction of tangible notes and coins should have produced little discernable effect on the rate of price convergence. Yet, the price divergence observed in 2002 and 2003 is surprising. Once more, the reliability of data for the control group can be used to explain such results. As has already been indicated in Figure 3, much of the rapid price convergence within the non-eurozone group can be accounted for by exchange rate movements. Nevertheless, after 2001, price convergence in the eurozone group is not nearly as fast as the control group and so relative price dispersion increased.

In summary, the car price data provides mixed evidence of price convergence within the eurozone. If we consider both datasets together, the majority of goods (29 out of 43) produced negative DD estimates ? indicative of price convergence.In contrast if we consider only the statistically significant results the inference swings towards price divergence with 4 out of 6 reporting positive DD estimates.

Conclusion

The results in this papers suggest that the single currency has led to a widespread narrowing of price differences across a range of goods, albeit at statistically insignificant levels. This suggests that the use of different currencies was a barrier to further economic integration in the European Union, thus supporting the European Commission?s claim. In addition, the results indicate that for some lower priced goods, the introduction of euro-notes and coins did act as a further spur to price convergence, suggesting consumers may not be perfectly rational but instead suffer from a form of money illusion.

The results of this paper are consistent with the conclusions of the majority of other studies regarding price dispersion in common currency areas, yet they differ to those of Lutz (2003) whose work this study most closely resembles. Differences in data can be used to explain at least some of this contrast. One could argue that the inclusion of a wider range of goods make the results of this study more indicative of general macroeconomic price trends. In addition, the use of five years of post-treatment data as opposed to just three, make this study more likely to pickup on price convergence patterns than Lutz?s work. However, it is possible that the data and empirical strategy used in this paper are unrepresentative of true conditions. Therefore it is important to assess the validity of results.

The first criticism of this study relates to the basis on which its conclusions have been drawn. Unlike Lutz (2003), this study bases its conclusions on the sign (±) of all DD estimates, rather than focusing on just the statistically significant results. This was justified on the grounds that each regression had very low numbers of observations and thus were unlikely to produce statistically significant results. However, the disadvantage of this method is that it produces less reliable conclusions. Indeed, if this study followed the lead of Lutz and focused on just significant results, then its conclusion would be one of price divergence rather than convergence.

Clearly, the robustness of the study's conclusions are sensitive to the data used, and hence to its reliability. Unfortunately, as section (3) of this study indicates, the data used is far from ideal. Therefore, a simple method of improving the strength of results would be to use a dataset that contains more of the ideal properties described in (3). Here, the Economist Intelligence Unit?s dataset represents a considerable improvement. By including a wider range of goods, the EIU?s data is likely to be more representative, whilst the inclusion of a greater number of countries make comparisons between multiple control groups possible ? thus adding considerable weight to conclusions.

The methodology used in this paper can also be criticised as it fails to address one of the key issues relating to price convergence ? inflation. While other studies have shown that inflation among low price economies was responsible for price convergence, the DD methodology can only conclude that price convergence has occurred and provides no evidence to substantiate the Balassa-Samuelson hypothesis. This is a notable downfall of the DD methodology since policymakers remain concerned about the inflation effects of joining the single currency.

Nevertheless, this paper has shown that over the five years since the introduction of the euro, prices have converged quite steadily. However, this pattern is unlikely to continue indefinitely, since even in a perfect monetary union, there will be a limit to price convergence. In the UK for example, large price differences still exist - supermarkets in Newcastle and London charge very different prices. This is partly because consumers do not have perfect information, but even if they did, price dispersion would still exist due to transport costs. One useful source of further study might therefore be to compare price convergence in border areas with those inland. until data becomes available to make such comparisons possible, policymakers will continue to discuss the long term impacts of the single currency on price convergence.

51due留学教育原创版权郑重声明：原创留学生作业代写范文源自编辑创作，未经官方许可，网站谢绝转载。对于侵权行为，未经同意的情况下，51Due有权追究法律责任。

51due为留学生提供最好的加拿大Assignment代写服务，亲们可以进入主页了解和获取更多assignment代写范文 提供美国作业代写服务，详情可以咨询我们的客服QQ:800020041哟。-xz