Figure 5 does not provide a very clear picture of how net capital flows have developed over time and what cumulative implications are at the macro level. Figure 6 therefore shows the cumulative net capital flows for FDI, portfolio and loans, and errors and omissions as well as the grand total of private sector capital flows.
Between 1995 and the first quarter of 2018, 700 billion dollars left Russia. This is twice as much as all of the fixed capital investments in 2017 and could obviously have boosted growth significantly if it had been invested in Russia instead. That does not mean that zero flows would have been optimal for the investors making these decisions, but it shows clearly that these flows are extremely important for the macroeconomic development of Russia. Most of the capital left in the form of portfolio flows and loans, but at the end of the sample, all three categories contribute to the outflows. FDI was for a long time the only component that recorded a cumulative net inflow over the period, but after the global financial crisis, there has been a steady outflow in this category as well and these outflows accelerated in 2014. More generally, the global financial crisis represents a very clear shift in capital flows, and outflows then accelerated when sanctions were introduced in 2014 before there was some levelling off in 2018.
Figure 6:Capital outflows from private sector.
Source: Central Bank of Russia and author’s calculations.
The question is then what factors may be behind these capital flows. In principle, we should expect flows to be correlated with the returns and risk on investment in Russia versus the rest of the world. There are different ways of trying to gauge expected returns and risk, but some relatively straightforward measures can be derived from the stock markets in Russia and abroad. Here, we use daily data on the Russian dollar index RTS and the S&P500 index from the US stock market. We also add daily data on oil prices since this is an important determinant of growth in Russia and also a source of foreign capital that can either be invested at home or abroad. From these data, we compute the daily returns and rolling 20- and 60-days standard deviations of our series and take quarterly averages of these measures to generate series with the same frequency that we have for capital flows. This then allows us to run a regression with net capital outflows being explained by the returns and volatility of the Russian and US financial investments as well as oil that are shown in Table 3.
Table 3:Correlates of private capital flows.
Note: Volatility diff is RTS volatility minus S&P volatility and Return diff is S&P return minus RTS return, so both coefficients are expected to be positive.
The regression results are quite interesting. The most statistically significant variable is the volatility on the Russian market, which has the expected positive sign that indicates that increased volatility increases net capital outflows. The other statistically significant variable is returns in the US market, but there is no offsetting effect from returns in the Russian market. The oil price variables are also not significant, which is perhaps a bit surprising given their importance for growth and investments. However, it could be the case that high oil prices both generate foreign exchange earnings in Russia that could leave the country as capital flows and encourage inflows into the Russian economy, and this estimate reflects that these two forces cancel each other out.
In principle, the relative volatility and return between the domestic and foreign market should matter for flows, and if the regression is run on these variables instead, the importance of volatility is further enhanced while the return variable becomes statistically insignificant. However, the overall explanatory power of such a regression is greatly reduced and is the reason the more detailed specification discussed earlier is preferred. The exact causal links and mechanisms cannot be investigated fully in this setting since there may be an effect going from capital flows from Russia to volatility in the Russian stock market. In the end, however, it is clear that volatility is an important correlate of capital flows that warrant a closer look.
Determinants of Returns and Volatility
The next item to investigate is how returns and uncertainty in the Russian stock market have developed and to what extent this can be understood by external and domestic factors. Again, the stock market here is viewed as a way to measure returns and uncertainty more broadly that would be correlated with capital flows, investments and likely also consumer confidence (which is not analysed further here but is an important demand side factor for growth). There are several factors that we can expect will affect returns and volatility on the Russian stock market. First, stock markets today are linked globally, and the developments on global markets are captured by the US market’s S&P500 index. We also know that many of the companies on the Russian stock market are linked to the energy sector, and therefore, international oil prices should matter for the valuation of the RTS. The S&P500 and Brent oil price are exogenous factors, so we can run a regression explaining variation in the return and volatility of the RTS with these variables as explanatory variables.
Table 4:Correlates of stock market returns and volatility.
Source: Author’s estimate based on market data.
Table 4 confirms that US stock market returns and changes in oil prices have a significant impact on returns in the Russian market. The estimation shows that coefficients are quite robust to estimating the relationship since the start of the RTS index in 1995 or focusing on the years after the global financial crisis.10 In the case of returns, the lags of US returns and oil price changes are significant, which is somewhat contrary to regular arguments about efficient markets that would immediately include all new information. The reasons for this apparent anomaly could include rather mechanical explanations such that the markets are located in different time zones, to market frictions that would lead to a somewhat delayed response.11 The coefficients on the lags are slightly smaller in the more recent years, which could be a result of reduced frictions, but the coefficients are still highly significant in both samples.
For volatility in the Russian market, the volatility in the US market and the volatility in oil prices are also highly significant and together explain about a third of the Russian volatility. The coefficients are again stable across the two samples and do not indicate a structural break in the relationship between the earlier and later time period. Note that the full set of explanatory variables that are included in the table were allowed to enter the first set of regressions, but insignificant variables were omitted from the final estimation to generate robust models from which we can compute residuals in the next stage.
The residuals computed from the estimated model mentioned earlier show the returns and volatility in the RTS that are unexplained by the external factors that are included in the model. This would thus include both domestic and foreign policy events that are not captured by changes in the US market or oil prices. Of course, the residuals will also include company-specific factors that influence the expected performance of the Russian stock market that we would not think of as Russian domestic or foreign policy events. For this reason, the residuals are noisy signals of these factors, but we can still use the residuals to look at what happens in the market at times when we know there are important policy events taking place and we have at least filtered out two important external sources of variation in the Russian market.
The