Introduction

 

The OECD countries nowadays are going through a major financial and social crisis. Even if most of them have the political and economic means to overcome a situation of emergency, some stay stuck in this quagmire, unable to raise their growth. From the public opinion point of view, growth is a crucial thing: it needs to be increasing in order to have a feeling of safety. However, in France for instance, growth is somehow stagnating. How come growth does not increase in these wealthy countries? In connection with this, another symbol of a country social and economic health is the Gross Domestic Product (GDP). Indeed, the GDP is the economic identity card for a country business and represents its economic situation. What truly composes the GDP? Where do its influences come from? Can it help growth increasing?

The dataset (cross sectional data) was collected on the OECD website http://www.oecd.org/statistics and gathers information (17 variables) on 31 OECD countries.

 

 

1.   Problem Statement

 

Following a macro economic point of view, the Gross Domestic Product (Y) may be considered as the sum of:

• The consumption (C)

• The investment (I)

• The public debt (G)

• The net export, or goods trade balance: X – M (X: exports, M: imports)

 

Y = C + I + G + (X-M)

 

• Private consumption stands for the final consumption expenditure of households (as well as non-profit institutions). Households’ final consumption expenditure thus represents the expenditure, including imputed expenditure, incurred by resident households on individual consumption of goods and services.

 

• Investment stands for the gross fixed capital formation which is measured by the total value of a producer’s acquisitions (minus disposals) of fixed assets during the accounting period (plus certain additions to the value of non-produced assets) realized by the productive activity of institutional units.

 

• Public Debt will be developed in the following section

 

• Net exports are computed such as the difference between exports and imports. The level of exports and imports for OECD area includes flows within the area shared by the member countries. The net trade however is not affected (since the non-consolidated flows in exports and imports cancel each other out in the balance). We shall notice that among our 31 OECD countries, 10 countries have a deficit in their trade balance: imports exceed exports (capture9).

 

The GDP, within a country, is usually computed by the national statistical agency, which has an unlimited access to a large number of sources. However, most countries follow established international standards while making the computations. The international standard for measuring the GDP comes from the 1993 System of National Accounts. It was established by the International Monetary Fund, the European Commission, the Organization for Economic Cooperation and Development, the United Nations and the World Bank.

 

Then how can GDPs of two countries be compared?

The GDP is measured in the currency of the country. When trying to compare the value of output in two countries using different currencies, it requires certain adjustments. The usual method is a conversion of the GDP value of each country into U.S. dollars in order to compare them afterwards. Conversion into dollars may be done either using market exchange rates or purchasing power parity (PPP) exchange rates. The PPP exchange rate is the rate at which the currency of one country would have to be converted into that of another to purchase the same amount of goods and services in each country. There is a large gap between market and PPP-based exchange rates in emerging market and developing countries. For most emerging market and developing countries, the ratio of the market and PPP U.S. dollar exchange rates is between 2 and 4. In that sense, the data we have contain GDPs in US dollars, calibrated ex-post. This is because non-traded goods and services tend to be cheaper in low-income than in high-income countries. For advanced economies, market and PPP exchange rates tend to be much closer. These differences mean that emerging market and developing countries have a higher estimated dollar GDP when the PPP exchange rate is used.

 

A certain methodology exists in terms of computation for international comparisons. Indeed, when we want to make an international comparison of the level, the structure, and the rhythm of development based on macro economic indicators of results, we need to solve two main problems:

 

• We need to make sure that the content of the indicators may be compared with each other, by using the National Accounting System (implemented in most of the OECD countries. We are already sure because the data we have has been calibrated with US$ and we also have the PPP for each country.

 

• We need to make sure the currency units may be compared with each other by whether using official exchange rate (even if it influences negatively the real values, since exchange rates do not represent the currency purchasing power), whether by using « parities » with which the GDP per head is converted in common currency of comparison. This is why we have PPP and external exchange rates.

 

 

2.   Public debt

 

A public debt is the total of financial commitments, such as loans, taken by the State, its public services, or public firms. When the fiscal profits are lower than the public administration expenses, then public deficit appears. Public debt increases every time the public deficit is financed by a loan. A public debt thus represents the accumulation of financial needs of a State. Whether a State is able to pay its debt back or not is evaluated by financial grading agencies. Within the public debt, we may distinguish two different concepts: the domestic public debt, which is held by the economic agents living in the responsible state, as well as the external public debt, held by foreign lenders. Knowing these concepts, what can be said about the OECD countries in term of public debt?

 

The twentieth century has witnessed very large fluctuations of the public debt. Right after the Second World War, the public debt reaches the highest level ever recorded. However, the prosperity period until the seventies gives the public debt a break. This is only after this period that the debt for the OECD countries increases a lot, around 1980. Within 25 years, the public debt came from representing 35% of the GDP to 70% in the euro-zone. In the eighties, the growth went on decreasing while the interest rates increased a lot, so that the gap between the interest rates and growth rate became positive. States of the OECD decided, because of the growth decrease, to use public spending into reflation policies instead of catching up their deficit (like Reagan in the USA, who lowered the taxes while public spending were increasing). In the nineties, the debt rate went on increasing, like in France for instance: between 1992 and 1997, it raised of 25 points of the GDP. Indeed, in Europe, interest rates were very high because of the German reunification, even if growth decreases a lot (or is negative). However, in the end of the 1990s, there is a phenomenon of a huge reduction of the gap between the interest rate and the growth rate (which allows the USA to stabilize their debt rates).

 

The public debt of the OECD countries is still a current problem nowadays, all the more so since the 2008 economic crisis. The main problematic currently studied is about the consequences public debts may have on the growth in general. OECD-wide total financial liabilities now exceed 1 000% of GDP. Countries are not equal in debt levels and leverage that reflect more national characteristics. The problem is that high indebtedness can create vulnerabilities to shocks, exposing households, firms and governments to maturity, currency and liquidity mismatches as well as potential solvency problems. Furthermore, high indebtedness can expose the economy to asset price movements, which can amplify shocks and macroeconomic instability. Indeed, the economic crisis caused by the US subprime mortgage market resulted in deep economic recession in many countries of the world. Governments and central banks reacted to this new Great Recession by fiscal and monetary policy expansion, coordinated across countries, while banking sector bailouts prevented the collapse of the financial system. Even if these actions helped smooth the cycle, fiscal loosening and banking sector bailouts contributed to a sharp increase in many countries’ public debt-to-GDP ratio.

 

Relying on this background, Reinhart and Rogoff (2010) pointed out the existence of strong negative effects of high public debt on economic growth. Using simple descriptive statistics, they demonstrated forcefully that economic growth slows down considerably if the public debt-to-GDP ratio exceeds 90%. Many empirical papers published afterwards validated the debt 90% threshold. Cecchetti et al. (2011) found a threshold of 86% of GDP for a panel of 18 OECD countries and for the period from 1980 to 2010. Padoan et al. (2012) decided to study a longer period (1960 to 2010) and found similar effects for a similar group of countries. Covering a mix of advanced and emerging market economies, Kumar and Woo (2010) found a turning point at 90% of GDP. Checherita and Rother (2010) reported similar results for countries in the euro zone. There are also articles against this assumption, proving that the threshold could vary under 70% and have consequences on the GDP, which is why we could wonder if the threshold is the only determinant for the growth slow down due to the public debt in the OECD countries. Indeed, we choose to study the link between GDP and key variables among which debt may play a main role. Those key variables with respect to OECD countries are embodied by every variable a government can meet when it faces a growth slow down, such as variables concerning the population, the inflation, or the business exchanges. We want to know which one of these variables are the determinants of the GDP. We will not focus on GDP fluctuation since our data are not time series but we will seek the determinants of the factual situation.

 

 

3.   Hypotheses

 

General economic theory provides that GDP is closely related to the key variables we chose to analyze and also gives an insight on how the interactions are effective and to what extent. Thus, we tend to verify the following hypotheses:

• When consumption raises, the GDP increase (H1).

• When investment raises, the GDP increase (H2).

• A positive trade balance has a positive impact on the GDP. Indeed, the more you export, the more your income increases and so does the GDP (H3).

• An increase in the debt is in favor of a bad economic health (based on the literature review) and does not help economical growth (H4). 

 

 

4.   Descriptive statistics

 

The data collected on OECD database provides information on economic variables of 31 countries. In order to identity the relevant variables to fit in the ones contained in the Y = C + I + G + (X-M) formula, we investigated every definitions and it turns out that:

• Total domestic demand (in percentage of change since previous year) will be used for consumption. The demand’s evolution reflects the potential change in the consumption from one year to another. These data can be used in our model since it is not time series data but just an evolution rate.

• Foreign Direct Investment (FDI) will be used for investment. FDI does not reflect domestic investment but provides a good insight to figure the investment power of each country.

• Public debt in percentage of GDP will be used for debt.

• Goods Trade Balance will be used for the net exports (Exports – Imports).

The United States have the highest GDP of OECD countries, i.e. 14419.4 billion of US dollars and Japan has the highest debt rate in percentage of GDP with 192.7%, while France is at 95.8% and has already transgressed Maastricht criteria regarding the debt non-exceed of 70% of the GDP. Domestic demand has highly fluctuated in OECD countries between 2007 and 2008. For instance, in Korea domestic demand has increased by 7.2% when Greece saw its demand decrease by 6%. In France, domestic demand has stayed pretty stable with a fluctuation of +1.4%.

Eventually, we have 8 missing data regarding the goods trade balance; 12 countries have a positive net exports meaning that exports exceed imports and 10 countries have a negative net exports.

Details on descriptive statistics are available in Appendix A (Figure A1).

 

We also investigated the link between the GDP and the public debt. Without searching for explanatory relationship between these two variables, here is how they are connected: 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

     

 

                                                                          Figure 1. Scatter plot: Public debt in percentage of GDP 

 

 

The use of the log(GDP) instead of the row GDP variable will be explained in the following section when determining our model to be used.

Every country on the right side of the vertical line (x=100) has a public debt exceeding 100% of its GDP. It can be noticed that these data reflects the situation in 2010 and some European countries were already in trouble regarding their debt. In particular, it is the case for Italy, Iceland, Portugal and Greece. Moreover, Iceland did even go bankrupt.

 

 

5.   Determination of the model:

 

Our dependent variable to be analyzed is GDP. When looking at its distribution we figured out that it was more relevant to use the logarithm instead of the row variable. Indeed, the log provides a “bell-shaped” curve associated to a normal distribution, stabilizing the variance. We conducted the same process for the variable Foreign Direct Investment (FDI), using log(FDI).  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2. Distribution (Kernel density) of GDP (left) and log(GDP) (right)

 

 

Analyzing our model, some variables seemed to be more useful than others. Since their p-values were not significant in the first regression we ran (Appendix B: Figure B1.), we tested whether employment rate, purchasing power parities, real net national income, effective exchange rates and goods trade balance could be removed at the same time. The results were ambiguous (Appendix B: Figure B2.). Indeed, the hypothesis of coefficient equal to zero excluding the 5 variables together was supposed not to be rejected. But because of the p-value of 0.0535, which is rather ambiguous, we investigated some correlations that could be at the origin of the problem.

 

As a matter of fact, the goods trade balance and the real net national income were strongly correlated (Appendix C: Figure C1.). So excluding the gtrbal from the test, results prove that we can be sure to remove from the regression our 4 variables (Appendix B: Figure B3.).

 

Moreover, our dataset gathers information that can be considered as redundant involving a non-necessary use of every variable. For instance, we witness an inner relationship between imports and gtrbal since Goods Trade Balance = Exports – Imports. Thus, according to the basic economic model of GDP: Y = C + I + G + (X-M), it seems relevant to use gtrbal in our final model.

 

After the series of tests, we obtained significant results with a first model (Appendix B: Figure B4.). The major problem in our analysis we encountered was that: every coefficient is significant, but R² exceeds 98%, which is very suspect. Indeed, we probably have a problem of endogeneity in our analysis since we should not be higher than 40% (approximately) of the model that explains the variance. Therefore, let’s look for the weak link in our regression. By weak link we mean the independent variable that contains to much information about the GDP so that it explains it by essence.

 

Imports / FDI / Trades balance:

 

As a matter of fact, imports are strongly correlated with logfdi and gtrbal: -0.66 (Appendix C: Figure C2.). We performed some tests to seek what variable can be removed at the same time. Imports, logfdi and gtrbal cannot be removed at the same time (Appendix B: Figure B5.).

As explained in the previous section, imports and good trade balance are redundant information. On the overall, we have 8 missing data on 31 countries. On the 23 remaining, 10 countries have a deficit in their trade balance: imports exceed exports (Appendix A: Figure A2.).

 

Debt / Population growth:

 

There is a negative correlation of -41% between the debt and population growth (Appendix C: Figure C3.). Then, is there an inter-generational transfer regarding countries’ debt? This question has long been discussed: is the debt a burden for the next generation?

From a generation to another, debts and lends are transferred. For a given generation, the sum of the lends is equal to the sum of the debts. Transfers are thus across the generations. They happened between those who held the public debt titles and those who do not have any. In order to create a debt, a state relies on obligations (treasury bonds) on the stock markets. These obligations are bought by other agents (banks, or insurances), who place the households savings of those who have enough money to save. The state’s process of debt creation leads to a backward-redistribution coming from every taxpayer, including the poorest ones (the VAT) and going to the public debt holders, meaning the richest households.

 

Thanks to these tests and endogeneity control, we were able to determine the best model to use, considering our variables and aim to seek GDPs’ determinants.

 

 

6.   Econometric analysis

 

Our goal is to base our econometric analysis on the economic equation of the GDP:

Y = C + I + G + (X-M). In that sense, here is the most appropriate use of our independent variables:

• Use the purchasing power parities (ppp) and demand for consumption (C). Indeed, the variable demand gathers the total domestic demand by percentage of change from previous year. This variable reflects the evolution of the level of potential consumption within each OECD country, associated to the purchasing power comparison between countries.

• Use the log of FDI (logfdi) and the government financial balance deficit or surplus (finbal) to explain the potential investment (I) of each country.

• Use the debt (G)

• Use the Goods trade balance (gtrbal) for the net exports (X-M).

When regressing according to the whole economic model, only two variables (debt and logfdi) are significant (Appendix B: Figure B6.). After having performed some tests, ppp and finbal were removed (Appendix B: Figure B7.). Moreover, results with loggdp regressed on demand logfdi debt and gtrbal were still non-significant for 2 variables (Appendix B: Figure B8) then we removed gtrbal from the model and obtained the final econometric model to use:

 

loggdp = b₀ + b₁.demand + b₂.logfdi + b₃.debt + u_i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Interpretations:

 

The p-values of our three independent variables are significant.

• When the domestic demand increases by one unit from one year to another, the GDP potentially raises by 16 billion of US$.

• When the FDI increases by 1%, the GDP increases by 52%.

• When the debt increases by one unit, the GDP increases by 1,2 billion of US$.

The F-statistic provides good results and R2 = 69% meaning that 69% of the GDP variance is explained by our model.

 

Tests:

 

We had good results from several tests we performed:

• The Breusch-Pagan test provided that we could not reject the hypothesis of constant variance: prob > chi2 = 0.3674 (Appendix B: Figure B9.).

• The White’s test to check for homoscedasticity provided that we could not reject the hypothesis of homoscedasticity: prob > chi2 = 0.8109 (Appendix B: Figure B10.).

• The Ramsey RESET test provided that we had no omitted variable in our model:

Prob > F = 0.3281 (Appendix B: Figure B11).

 

Conclusions & discussion

 

This project investigated the determinants of OECD countries’ GDP. It results that we were not able to base our econometrics analysis on the whole basic economic formula for the GDP since the goods trade balance could not been taken into account. For sure, the net exports embody a wide part of the GDP since our economies are based, among other parameters, on economic trade. Hypotheses H1 and H2 were thus verified but we were not able to verify H3 even if we can be pretty convinced that a positive goods trade balance will make the GDP rise.

 

On the other hand, our final model provided very good results since our model explains 69% of the GDP variance.

One result was not expected: the sign of the debt coefficient in our regression is positive: H4 was not verified. Based on the wide literature treating this negative relationship (Reinhart & Rogoff, 2010, Cecchetti et al., 2011 and Padoan et al., 2012), we expected that the more the debt increases, the less growth there will be in the considered country. Indeed, mostly other countries finance the debt; and these liabilities should provide help for economical growth. But other determinants should be taken into account since the debt process is complex and has many parameters interconnected. In addition, we can think that in the short term, an increase in the debt will be in favor of more economic growth (GDP increases) since in enables more investment, but in the long term, it can be the opposite, becoming a signal of bad economic health.

 

We found it original to base the consumption on the demand evolution. We thought that demand truly reflects the evolution of consumption in these countries.

 

 

 

References

 

[1] Boulhol, H., de Serres, A & Molnar, M. (2008). The Contribution of Economic Geography to GDP per capita, OECD Journal: Economic Studies.

 

[2] Callen, T. (2012). Gross Domestic Product: An Economy’s All, Finance & Development, Department of the International Monetary Fund.

 

[3] Cecchetti, S. G. et al. (2011). The real effect of debts, Bank for International Settlements Working Papers No 352.

 

[4] Checherita, C., Rother, P. (2010). The impact of high and growing government debt on economic growth: an empirical investigation for the euro area, European Economic Review, 56(2012), 1392–1405.

 

[5] Dewan, E., Hussein, S. (2001). Determinants of Economic Growth (Panel Data Approach), Working Paper Economics Department Reserve Bank of Fiji.

 

[6] Égert, B. (2012). Public Debt, Economic Growth and Nonlinear Effects: Myth or Reality?, OECD Economics Department Working Papers, No. 993, OECD Publishing.

 

[7] Égert, B. (2013). The 90% Public Debt Threshold: The Rise and Fall of a Stylised Fact, OECD Economics Department Working Papers, No. 1055, OECD Publishing.

 

[8] Folcut, O, Despa, R. & Mustea, R. (1992). Handbook of the International Comparison Programme, Department of economic and social development statistical division UN, New York.

 

[9] OECD (1979), Quarterly National Accounts:  A report on the sources and methods used by OECD Member Countries, Paris.

 

[10] OECD/Eurostat (2012), “Component expenditures of GDP”, in Eurostat-OECD Methodological Manual on Purchasing Power Parities, OECD Publishing.

 

[11] Padoan, P. C., Sila and P. Van der Noord (2012). Avoiding debt traps. Financial backstops and structural reforms, OECD Economics Department Working Papers, No. 976, OECD Publishing.

 

[12] Reinhart, C. M. & Rogoff, K. S. (2010). Growth in a Time of Debt, American Economic Review, American Economic Association, vol. 100(2), pages 573-78

 

[13] Sutherland, D. et al. (2012). Debt and Macroeconomic Stability, OECD Economics Department Working Papers, No. 1003, OECD Publishing.

 

[14] Turner, D. and F. Spinelli (2013). The Effect of Government Debt, External Debt and their Interaction on OECD Interest Rates, OECD Economics Department Working Papers, No. 1103, OECD Publishing.

 

[15] Turner, D. (2013). Methodological notes referring to the OECD press release: Contributions to GDP growth, OECD Quarterly National Accounts