An assessment of the Indian rupee crisis

2018 has not exactly been kind to emerging market (EM) currencies. At the current junction, some EM currencies, such as the Argentine peso, the Turkish lira and the South African rand are even performing far worse than during the taper tantrum back in 2013 (see figure 1). The Indian rupee (INR) has been the worst performing currency in Asia, losing more than 12% against the US dollar compared to the beginning of the year and registering a historic low of 72.7 against the USD in September (see Figure 2). In this Special, we will make an in-depth analysis of the INR currency crisis and what to expect for the INR going forward. Should we brace for more rupee weakness or will authorities be able to stop the free fall?

A combination of factors are at the root of the global sell-off of emerging market currencies. First, in April US yields and the USD were pushed higher on the back of ongoing Fed monetary tightening, which has resulted in a sudden stop of portfolio inflows toward emerging markets. Through 2018, it became apparent that the impact of FOMC US rates decisions on investor sentiment towards emerging markets has become more and more important, and we believe this will continue to be case in 2019.

Second, EM currencies have been bearing the brunt of the protectionist measures launched by the Trump administration, as these measures have fuelled a global risk-off sentiment with investors. On 24 September, the Trump administration added fuel to the fire by implementing yet another round of tariffs[1] on USD 200bn, raising total US trade measures on China to more than 253bn worth of Chinese imports. Beijing has called off talks to de-escalate the trade tensions, and imposed reciprocal tariffs targeting USD 60bn of US exports to China. In total, China has retaliated 113bn of US export to Chinese shores, but most likely is also targeting the US by raising the amount of non-tariff barriers. Going forward, we expect more EM currency volatility on the back on ongoing trade tensions between the two largest global economies (Giesbergen and Erken, 2018).

A third determinant of EM currency weakness are stubbornly high oil prices, which rose from just below USD 65 per barrel in February to more than USD 80 per barrel at the moment of writing (figure 2). Sagging supply from Venezuela and Iran are starting to feed into higher oil prices. High crude prices are especially troublesome for large net oil importers, such as South Africa, Indonesia, and especially India. In India, crude and refined petroleum alone constitute a whopping 17 percent of total import (figure 3). Higher oil prices result in a deteriorating current account metrics, which may lead to further capital outflows and stress on the currency.

The final factor causing EM currency weakness has been related to country-specific turmoil in e.g. Turkey (concerns about central bank independency and inflation), Brazil (upcoming elections), South Africa (elections) and Argentina (low reserves, high inflation). These country-specific events have exacerbated the adverse sentiment of investors towards emerging markets in general, which resulted in a large sell-off of not only the assets of the countries experiencing the domestic turmoil, but others as well. Countries, such as Indonesia and India, were dragged into global EM turmoil, as investors are obviously still placing the so-called ‘fragile five’ countries[2] from taper tantrum in the same basket. The question is: is the major sell-off of the INR justified if we take a closer look at the economic fundamentals.

[1] The tariffs of 10% in the 200bn package will take effect on 24 September and from 1 January the tariffs will rise to 25%.

[2] The fragile five are Brazil, Indonesia, India, South Africa and Turkey.

The EM Vulnerbility Heatmap approach

Rabobank’s first EM Vulnerability Heatmap provided a framework that helped us successfully navigate the relative performance of EM currencies during bouts of broad-based weakness. Our ‘new and improved’ EM Vulnerability Heatmap should help carry that baton going forward. Of course, domestic issues and events will play a huge role that can, and will, lead to substantial outperformance or underperformance relative to that implied by our heatmap but this framework should provide a good starting point to assess likely baseline performance.

The basic structure is the same as our previous editions but we have reviewed our input variables and split the framework into three main groups: economic variables, market variables and portfolio flow variables. Without going into details on the results or methodology of the EM Vulnerbility Heatmap, it becomes apparent that India is not among the best performing emerging market in terms currency fundamentals, but certainly not among the worst performers either. This underlines that relatively healthy emerging markets, such as India and Indonesia, have been dragged into the current EM currency turmoil by countries performing much worse.

The underperformance of the INR also becomes clear when we use the outcome of the EM Vulnerbility Heatmap and link this to the actual currency losses this year (figure 4). When combining both dimensions, we get an indication whether a EM’s currency is performing as expected based on its fundamentals or is over/underperforming relative to other emerging markets. We see that for example India is underperforming: it is ranked 10th in terms of vulnerability based on the Heatmap approach, but looking at the losses this year it comes at place 7.

An iterative portfolio-INR model

We can also use models to provide us with an idea what level of the INR is justified based on the fundamentals. We have experimented with the variables that we have used in the EM Vulnerbility Heatmap, but not every variable proves to be statistically relevant for India. The annex gives a detailed description (data, functional form, assumptions) of a two-equation iterative model, which integrates a portfolio investment equation with an error-correction equation for the INR. The model produces the following results.

We expect portfolio investment to recover to levels in 2019 of INR 23bn on a monthly basis (Figure 5). The trajectory remains negative, with average investment sliding to approximately 14bn in subsequent years. Admittedly, our model is not able to pick up the high volatility on a monthly basis, so this means that the INR can experience more volatility going forward than we have projected in our baseline. Nevertheless, if we strip out the underlying volatility, our model picks up the trend reasonably well. This is in line with earlier results, where we estimated a model for Indian portfolio investment on a quarterly frequency (see Erken, Hayat and Heijmerikx, 2018).

In the medium to longer term, India should brace for more rupee weakness (Figure 6) due to a further pickup in inflation[3], structurally lower portfolio inflows and a further increase in oil prices. Ultimately, the steady state INR rate in our models would lie somewhere around 74. However, our models expect a gradual process toward this level which will take a stretch of many years, which fits the pattern of fast-growing emerging markets. This is not something to worry about. In the short term, however, our models point at INR levels of 67. This means that the large gap between the current value of the rupee (hoovering between 72 and 73) and its predicted value based on the underlying fundamentals is solely driven by bearish investor sentiment. Based on these outcomes, we argue that the INR has been oversold and we expect strengthening in the upcoming period.

[3] Domestic dynamics at the moment have been so strong that the output gap has closed far more quickly than expected. The result is that ongoing high domestic dynamics will also continue to put upward pressure on price levels.

So what would have to happen to reach levels currently seen for the INR based on the underlying determinants? We have used our model to run four scenarios for portfolio flows and the INR (Figure 7). If India would relive the impact of demonetisation in combination with a severe oil price shock, this would push the INR to levels between 72 and 73. Hence, we need to make quite bold assumptions to arrive at INR levels seen in September, which underlines our main conclusion that the current levels of the INR are out of sync with its fundamentals and the current weakness is mainly driven by market sentiment.

Of course, things could be significant worse. In a cumulative stress scenario - which encompasses an economic shock similar to demonetisation, an oil price shock, an acceleration of the Fed tightening cycle and political turmoil - capital flight would accumulate to more than INR 1000bn until early 2020 and this would push the rupee to levels close to 77, according to our model. Below we elaborate on each of these scenarios in more detail.

Scenario 1: demonetisation

In our first scenario, India would relive the economic growth damage of demonetisation and lose 1.1ppts in the current fiscal (FY 2018/2019) compared to our base case. As a consequence, the INR would face a 3% depreciation against the USD, which would shift the baseline (dark blue line) to August levels (orange line in figure 5). Ultimately, the effect peters out and both the baseline forecast and scenario 1 will come together at a level somewhat below 74.

Scenario 2: oil price shock

In our second scenario, we add an oil price shock to our demonetisation scenario. We assume that oil prices will spike at 120 USD in January 2019, before levelling off to approximately 100 in the subsequent months (see Figure A.2 in the annex). The decreasing trajectory is based on insights by Middeldorp, Groenewegen and De Vreede (2018), who developed a VAR model to estimate the production response in case of an oil price shock. Based on their insights, roughly 50% of the initial price shock is reversed after two years due to a counterbalancing impact of additional shale production in the US.

We find that the oil price pushes down the INR by 2% to 3% structurally (light blue line). However, the impact of higher oil prices affects the INR with a certain delay in our models. Therefore, the INR depreciation only begins to materialise in 2019 and maxes out in 2021.

Scenario 3: faster Fed tightening

In our baseline, we assume a gradual pickup in US inflation over the next few years. Currently, we have 4 hikes penciled in for this year (see Marey, 2018). There is also an alternative scenario, where inflationary pressure would take off much more profoundly than anticipated. This could be initiated by low labour market slack resulting in an acceleration of wage pressure, Trump’s tax cuts - which overheats the US economy - and price pressure from import tariffs imposed on 250bn worth of Chinese imports. As inflation overshoots, the Fed is forced to tighten monetary policy much more quickly and significantly in order to break a potential wage-price spiral (light-blue line in Figure A.3). In this scenario, we assume that the RBI won’t cut any rates in 2019 and 2020, but will not follow the Fed in their hiking spree as well. Ultimately, a faster than expected Fed tightening would result in additional structural weakness of the INR of almost 3% (see light green line in Figure 7).

Scenario 4: political turmoil

In the final scenario, we assume that the opposition succeeds in forging a grand coalition and beats Modi’s BJP in the general elections next year. Furthermore, we assume that losing the elections by Modi will result in government instability and the political reform agenda of the BJP will be put on hold by the newly-elected consortium. We incorporate this scenario in our models by doubling the value of our IHS Markit political risk measure from 2.4 now to 5 over a short horizon (see Figure A.4). After a newly-formed coalition has been installed, the political risk will level off to a value of 3, one standard deviation above current levels. In this scenario, the INR will lose almost 4%, but will make up for previous losses from 2021 onwards and appreciate somewhat more than 1% in the longer term (see grey line in Figure 7).

Limitations

It is important to stress that the INR trajectories in our scenario analysis are estimated ceteris paribus, which means that we leave all other parameters unchanged. For instance, we only assess the direct impact of a higher oil price on the INR, whereas we know that high oil prices also lead to a deterioration of the current account and fiscal metrics, which produces second order or even third order effects on the INR. To keep matters simple, however, we abstract from these feedback loops. Incorporating all these effects without one framework would take us a step further towards dynamic computable general equilibrium (CGE) modelling, something which we would like to pick up in the future.

The INR rate is not completely at the mercy of external factors. The Indian government has been pursuing different strategies to prop up the rate of the INR. On Friday 14 September, Finance Minister Arun Jaitley announced that the government will curb non-necessary imports, boost export and re-emphasize its commitment to the fiscal deficit target of -3.3% for fiscal year 2018/2019. Moreover, the government wants to stimulate capital inflows, by relaxing rules for investment in Masala bonds, corporate bonds and allow manufacturing to raise short-term external commercial borrowings (ECBs).[4] So far, the measures have hardly resulted in strengthening of the INR. Therefore, we certainly expect more government announcements going forward.

RBI policy

The big question is whether the RBI is also going to step in with its monetary policy. So far, the RBI has been intervening with INR 20bn on the foreign exchange market in the last couple of months to prop up the INR rate (Figure 9). Keep in mind, however, that interventions on the currency market are also a form of monetary tightening, as the central bank will buy local currencies in order to swap them for foreign currency (e.g. US dollars) on the forward market. Therefore, central banks will not be too keen to heavily use this type of intervention, as there is a downward risk of derailing the economic growth momentum. The RBI will also take into consideration what to do with its most important policy rates in the October Monetary Policy Committee (MPC). Admittedly, inflation has been on a downward trajectory and with flexible inflation targeting as its main policy objective, one might expect that the RBI wants to keep its rates on hold. We believe, however, that the RBI cannot run the risk to under-deliver. We do not expect the Fed to halt its tightening cycle before 2020, especially given the ongoing strong data that the US economy is showing. It does not take much imagination to see how under-delivering by the RBI vis-à-vis Fed rates could easily spur another round of capital outflows, result in further weakening of the INR and prop up inflationary risks. What’s more, we believe that the inflationary pressure in India has far from disappeared. The low inflation rates in July (4.2%) and Augustus (3.7%) were mainly due to very favourable food price developments and high core inflation of 6% should be taken as a sign on the wall.

[4] Investment in Masala bonds issued this financial year are exempt from withholding tax and restrictions on issuing Masala bonds by banks are relaxed. Furthermore, the government wants to remove the exposure limit of 20% for foreign portfolio investors in corporate bonds.

Unorthodox policies

If the current set of measures and monetary policy will fail to stop the INR from sliding further and further, the government will have to think of more unorthodox measures, such as capital controls or bilateral swap arrangement. Capital controls could encompass, e.g., a cap on outward direct investment for firms. However, authorities probably do not want to resort to such extreme measures, as it could take years to regain trust of investors once it has been damaged.

In the longer term, the government should think of ways to lower the dependency of the Indian economy on oil import, thereby making the economy and INR more resilient against fuel price swings. Social unrest is already increasing, as surging oil prices are eating away real disposable income of people in developing and emerging economies. Although India has been focusing on increasing investment in renewable energy, at the current levels this is still a drop in the ocean. Speeding up the energy transition in order to secure energy supply and make India less dependent on external volatile fluctuations might be the single best strategy for the Indian rupee in the long run.

Emerging market currencies have been taking quite a beating this year and the worst may not be over yet. Fears over an escalating China-US trade war, higher oil prices, a tightening monetary cycle by the Fed and country-specific events in emerging markets have fuelled a global risk-off sentiment, crushing many EM currencies. Indonesia and India were dragged into global EM turmoil, as investors are obviously still treating the so-called ‘fragile five’ countries[5] from taper tantrum as the same asset class.

In this Special we use two approaches to delve deeper into the fundamentals of the INR in order to examine if the current weak levels are justified. Our newly developed EM Vulnerbility Heatmap approach shows that India is not among the best performing emerging markets in terms currency fundamentals, but certainly not amongst the worst performers either.

Our integrated portfolio-INR model shows that in the medium to longer term, India should brace for more rupee weakness due to a further pickup of inflation[6], lower structural portfolio inflows and a pickup of oil prices. Ultimately, the steady state rate of INR would lie somewhere around 74. However, these models expect a gradual process which will take years and years and fits the pattern of a fast-growing emerging markets. This is not something to worry about. In the short term, however, our models arrive at INR levels of 67. This means that the large gap between the current value of the rupee (hoovering between 72 and 73) and its predicted value based on the underlying fundamentals is solely driven by bearish investor sentiment. Scenario analyses indeed shows that we need to make quite bold assumptions, such as a demonetisation shock in combination with a severe oil price shock, to justify INR levels seen in September.

Based on all these outcomes, we argue that the INR has been oversold and we expect the rupee to strengthen in the short term. Moreover, Indian authorities are vigilant in preventing the INR from sliding further against the USD. Finance Minister Arun Jaitley already announced the first steps in September, which encompass curbing unnecessary imports, boosting exports and relaxing rules to in order to attract more external investment. As these measures have not proved very fruitful, we certainly expect more government announcements going forward. We also expect the RBI to step in and hike their most import policy rates by 25bps during the October MPC. Although stable inflation is the RBI’s main objective (rather than a stable rupee itself) and inflation has been relatively subdued the last two monthly, it does not take much imagination to see how under-delivering by the RBI vis-à-vis Fed rates could easily spur another round of capital outflows, result in further weakening of the INR and prop up inflationary risks.

[5] The fragile five are Brazil, Indonesia, India, South Africa and Turkey.

[6] Domestic dynamics at the moment have been so strong that the output gap has closed far more quickly than expected. The result is that ongoing high domestic dynamics will also continue to put upward pressure on price levels.

Ahmed, S. & Zlate, A. (2014). Capital Flows to Emerging Market Economies: a Brave New World?, Journal of International Money and Finance, 48, pp. 221-248.

Baek, I. M. (2006). Portfolio investment flows to Asia and Latin America: Pull, push or market sentiment?, Journal of Asian Economics, 17(2), pp. 363-373.

Cheung, Y. W., Chinn, M. D., & Pascual, A.G. (2005). Empirical exchange rate models of the nineties: Are any fit to survive?. Journal of international money and finance, 24(7), 1150-1175.

Cheung, Y. W., Chinn, M. D., Pascual, A. G., & Zhang, Y. (2018). Exchange rate prediction redux: new models, new data, new currencies. Journal of International Money and Finance.

Engel, C., Mark, N. C., West, K.D., Rogoff, K., & B. Rossi (2007). Exchange rate models are not as bad as you think. NBER Macroeconomics annual, 22, 381-473.

Erken, H.P.G., R. Hayat and M. Heijmerikx (2018). India: trade wars and capital flight. Rabobank.

Dahlhaus, T, & Vasishtha, G. (2014), The Impact of US Monetary Policy Normalization on Capital Flows to Emerging-Market Economies, Bank of Canada Working Paper no. 14-53.

De Vita, G. and Kyaw, K.S. (2008). Determinants of capital flows to developing countries: a structural VAR analysis, Journal of Economic Studies, 35(4), pp. 304-322.

Froot, K.A., P.G. O’connell and M.S. Seasholes (2001). The portfolio flows of international investors, Journal of financial Economics, 59(2), pp. 151-193.

Grinold, R., & Kroner, K. (2002), The equity risk premium: Analyzing the long-run prospects for the stock market, Investment Insights, 5(3), pp. 7-33.

Grinold, R., Kroner, K, & Siegel, L. (2011). A Supply Model of the Equity Premium, in Hammond, B., Leibowitz M. & Siegel L., eds., Rethinking the Equity Risk Premium, Charlottesville, VA: Research Foundation of CFA Institute.

Koepke, R. (2015). What drives capital flows to emerging markets? A survey of the empirical literature. International Institute of Finance (IIF) Working paper.

Koepke, R. (2018). Fed policy expectations and portfolio flows to emerging markets, Journal of International Financial Markets and Money, forthcoming.

Le, Q.V & Zak, P.J. (2006). Political risk and capital flight, Journal of International Money and Finance, 25, pp. 308-329.

Montiel, P. & Reinhart, C.M. (1999). Do capital controls and macroeconomic policies influence the volume and composition of capital flows? Evidence from the 1990s. Journal of International Money and Finance, 18 (4), pp. 619-635.

To develop a model which can serve as an anchors for our INR forecast, we use an integrated portfolio investment-INR model. The model consists of two equations. The first equation is a capital flows model based on research earlier this year (Erken, Hayat and Heijmerikx, 2018). The second equation is an error-correction model (ECM) for the INR. Both equations contain two unknown variables, but by solving them within one system, we can ensure that we obviate endogeneity issues to a certain extent. These issues lie at the heart of many projections of financial market variables.

Equation I: portfolio investment

Portfolio investment is driven by both pull and push factors (Koepke, 2015). Pull factors are variables such as economic growth, stock market performance and country/political risk, while push factors are external factors, such as global risk aversion, global economic growth and global interest rates.[7]

There is wide consensus in the literature that US interest rates affect capital flows to emerging markets (e.g. Montiel and Reinhart, 1999 and Ahmed and Zlate, 2014).[8] If the USD strengthens, USD denominated private and public debt and commodities will become more expensive by local currency standards. Moreover, a rise of the US-EM interest rate differential makes emerging markets less attractive for investors. To proxy expected changes in the US Fed policy rate, we follow Koepke (2018) and Dahlhaus and Vasishtha (2014). Koepke, for example, uses Eurodollar futures as a proxy for the expected Fed policy rates. Koepke finds that a change in the expected future Fed policy rate has a large and negative effect on portfolio flows in emerging markets. Similar to Koepke, we use the 1 year Eurodollar futures price to gauge the market’s expectation of the Fed policy rate 12 months. To proxy the interest rate differential between India and the US, we use the difference in (3 month) interbank rates between India (Mibor) and the US (Libor).

Economic growth is also likely to have a positive effect on capital inflows (e.g. De Vita and Kyaw, 2008). Higher economic growth implies higher earnings for firms and more revenues for governments. In addition, higher stock prices are likely to have a positive influence on capital flows because the stock market tends to be a leading indicator of economic growth and reflects investor sentiment. Evidence of stronger capital flows due to higher asset prices has for example been found by Froot et al. (2001) and Koepke (2018). Exchange rate weakness is likely to be a negative for capital flows, as it eats away the returns of foreign investors. The effect of a country’s fiscal deficit is not clear cut. On the one hand, a higher deficit indicates a need for external capital, which implies that a higher deficit leads to higher capital inflows. On the other hand, a higher fiscal deficit implies higher public debt in the future, which implies higher sovereign risk. The literature indicates that the first effect dominates (e.g. Baek, 2006). Political risk is likely to be negative for capital flows since it implies more uncertainty and increases the country-specific discount rate on an asset and thus reduces its value. Le and Zak (2006), for example, find evidence that an increase in political risk leads to capital flight in the 45 developing countries..

Model and data

Ultimately, our portfolio model takes the following form:

[7] Koepke explains that the push and pull framework is broader than the simple internal and external dichotomy. However, to keep things simple, we broadly refer to push and pull factors as internal and external factors.

[8] This relates to the interest rate differential between emerging markets and the US as well as tightening of US monetary policy. The difference between the two is important, because the interest rate differential between the US and emerging markets could go down without a change in US rates.

where Pt represents portfolio investments in equities and securities in India (six month moving average), is the expected Fed policy rate based on euro-dollar one year future contracts,is the US 3 month interbank rate (LIBOR), is the Indian 3 month interbank rate (MIBOR), yt is the volume mutation of gross domestic product (GDP), s is the value of India’s stock market index (measured by the BSE Sensex), Et is the exchange rate (INR per USD), FDt indicates India’s fiscal deficit as a % of GDP and Dtd is a demonetisation dummy (picking up the impact of capital flight due to Modi’s announcement in November 2016 to replace 83% of all bank notes in circulation), PRt is an indicator from IHS Markit that gauges political risk in India and finally, c is the constant term, and finally, et is a Gaussian error term.[9] We have tested each variable on stationarity using augmented Dickey-Fuller tests. In case of non-stationarity, we express the variable in first differences (indicated in the formula by ). Table A.1 shows all the data sources per variable. We estimate equation (1) using simple OLS and monthly data covering the period 1999m2-2018m7 (indicated by t).

[9] More specifically, PR indicates IHS Markit’s political risk measure (ranging from 0.1 to 10), which is the average of three equally weighted sub components of political risk. IHS Market defines these subcomponents as: Government instability (the risk that the government will change), Policy instability (the risk that the government’s broad policy framework will shift) and State failure (the risk that the state is unable to function effectively). A higher number here indicates more political risk.

Results

Table A.2 shows our estimation results. Column (1) shows the results of our base model, which contains both interest spread variables and two ‘pull’ factors: the change in economic growth and the exchange rate. All variables, except the expected interest rate differential show a statistically significant effect. The US interest differential (α₁) has a coefficient of -5.5, which means that if financial market expectations about a Fed rate hike in one year changes by 1ppt, portfolio investments in India decrease by INR 5.5bn. The actual interest rate differential (α₂) has a coefficient of 10, which means that if the spread decreases by 1 ppt (for example due to a US rate hike), portfolio flows decrease by INR 10bn. Economic growth (α₃) has a coefficient of 21.89, which means that an increase of growth by 1ppt has a positive impact on portfolio investment of INR 22bn. This variable only picks up shocks to economic growth. Finally, the exchange rate has a significant effect of -2.66. This means that a 1% depreciation results in portfolio outflows of INR 2.66bn. Of course, there is also reverse causality: when investors shift their portfolios away from India, they will sell rupees and buy other currencies for their alternative investment, which puts downward pressure on the INR. This is the key reason for us to solve equation (1) together with the INR equation, which includes portfolio investment as an independent variable. Due to the two-way causality, it is important to include this self-reinforcing mechanism in our model.

In column (2) we include the stock market as a variable. The coefficient of 0.72 for the stock market variable (α₅) implies that if returns on the stock market increase by 1ppt, portfolio investments to India increase by INR 720 million. Column (3) shows the impact of the change of the fiscal deficit on portfolio flows (α₆). A deterioration of the fiscal deficit by 1ppt compared results in a capital outflow of INR 15bn. A change of the political risk index (α₇) by 1ppt results in capital outflow of INR 550mln. In column (4) we wrap up our model by including a demonetisation dummy (α₈), which captures the 15.5bn drop of portfolio flows after PM Modi’s announcement to ban 83% of all currency notes late 2016.

Equation II: INR model

Academic scholars have devoted much time and energy in developing solid exchange rate models, but only with limited success. According to Cheung, Chinn and Pascual (2005) and Cheung et al. (2018) there is no model that consistently outperforms a random walk, although Engle et al. (2007) find this a too strong criterion to completely dismiss exchange rate models as being useful. In the literature, there have been various approaches (see Cheung et al., 2018): the covered interest rate parity, the purchasing parity, the sticky price model, the real interest rate differential, the yield curve and the sticky price monetary model augmented by risk and liquidity factors. We use different elements from these approaches as building blocks for our own model.

Model and data

Our INR model takes has the following error-correction specification:

where E is the INR/USD exchange rate, y represents growth of GDP volumes for India (IN) and the United States (US), π is inflation measured by the consumer price index, P portfolio investments in equities and securities in India. As portfolio investments on a monthly frequency are very volatile, we use a six monthy moving average. O represents the Brent crude oil price and M is the money supply (M1), which is defined as currency with the public and deposit money of the public. The indicator picks up a time trend. The parameter λ is our error correction parameter. If the ECM parameter is statistically significant and negative, this means that out-of-equilibrium inferences of the independent variables in the short term are pushed towards the long-term steady state. Table A.3 shows the data sources of each variable used.

Results

Table A.4 gives an overview of the estimation results of equation (2). The interpretation of the effects on the INR is complicated by the trend variables and the error-correction mechanism. Therefore, we will elaborate on the effects after discussing the estimation results. Our base model in column (1) only contains the growth and inflation differential between India and the US and a simple time trend. All variables show the expected sign and are statistically significant. The negative coefficient (α₁) for the growth rate differential means that if Indian economic growth rises vis-à-vis the US, this pushed down the INR/USD rate, which means that the INR appreciates. Reversely, coefficient α₂ indicates that if inflation in India outpaces US inflation, this props up the rupee rate (i.e. the INR depreciates). In the long run, the growth rate differential (α₇) has a significant impact on the INR rate, but the inflation differential does not. The error correction parameter (λ) is statistically significant and negative, which means that short-term deviations are pushed to the long-run equilibrium.

In column (2), we include portfolio investment as a variable (α₃). As expected, the effect is negative: higher portfolio investment result in an appreciation of the INR vis-à-vis the USD (i.e. the INR/USD rate goes down). The effects of the other variables remain stable and statistically significant. In column (2) the inflation differential also has a significant impact on the long-term INR rate, whereas this was not the case in column (1).

In column (3), we add the oil price in both the short (α₄) and long-run specification (α₉). Although the effect of the oil price is positive and statistically significant, we need some serious lagging before the impact on the INR materializes. In the last column, we add money supply (M1). We pick up a negative impact (α₅), which means ns that higher liquidity would result in an appreciation of the INR. This is rather counter-intuitive, as one would expect that higher liquidity would lead to lower interest rates and a weakening currency. However, higher liquidity demand could also be the result of higher domestic demand and economic growth. In this case, more liquidity could go hand-in-hand with an appreciation of the INR.

Besides the variables shown in Table A.4, we have experimented with different functional forms and other control variables, such as the current account deficit or the VIX, neither of which produced fruitful results. Ultimately, the fit of the INR model is depicted in Figure A.1.

Our base case

In order to use both the portfolio and INR equations in an iterative process to produce forecasts, we need to extrapolate the underlying parameters of both models. The forecasts of the interest rate, inflation, economic growth and the oil price are based on Rabobank official forecasts.

For this year, we expect the Fed to raise interest rates in September by 25 bps and 25bps in December and we have one more hike by the RBI in the cards in October. In the subsequent years, we expect the Fed to raise rates in September 2019 by 25bps and in March 2020. We expect economic growth in India to arrive at 7.9% for this year, 8.2% for 2019 and around 8% in the years thereafter. In the US we expect economic growth of 2.8% for this year, 2.3% for 2019 and rates of approximately 2% in the following years. Inflation in India is expected to remain high due to unfavourable oil price dynamics, a closing output gap and a weak INR. We expect inflation to arrive at 4.9% this year and 5% next year, before picking up towards 6% in 2021. In the US we expect inflation to average 2.7% and weaken to 2% next year remain around that rate. We expect the oil price to average 80 USD this month, before levelling off to USD 72 per bbl in October next year.

Forecasts of India’s stock market are based on the Grinold and Kroner model (Grinold and Kroner, 2002 and Grinold, Kroner and Siegel, 2011). In this model, future returns for the stock market of a country are determined by the sum of that country’s expected (i) real economic growth, (ii) inflation rate, (iii) dividend yield and share buybacks, and (iv) the re-pricing of valuation levels (measured by price earnings ratios). The idea behind this model is that in the long run, corporate profits cannot grow faster or slower than the economy and that valuation levels return to their long term values. For India, we expect (for the foreseeable future) a real economic growth of 7.9%, an inflation rate of 5% and an average dividend yield of 1.3% (same as the 10 year average for the BSE Sensex) and a repricing of approximately 6% (based on a price earnings ratio for the BSE Sensex of 23.3 at the time of writing, which we expect will revert to its long-term average of 18 over the forecast period).[10]

With respect to the Indian fiscal metrics and money supply, we use ARIMA modelling to extrapolate the series. We expect the fiscal deficit to stabilize around -3.8% and money supply growth to continue a stable rate at 13.8%.

Scenarios assumptions

[10] Taken together, we expect the Indian stock market to return approximately 8% (7.9% + 5% + 1.3% -6% = 8%). We disregard the effect of share buybacks because data on this variable more difficult to attain.