In a recent discussion paper, I argue that contrary to conventional economic thinking, the purchase of foreign exchange (FX) by the Brazilian Central Bank through sterilized interventions  have been having an expansionary impact on Brazil’s aggregate demand and inflation levels. It is important to reconsider the current massive sterilized interventions not only because they constitute an expensive and inefficient way of tackling currency appreciation, but also because they make it more difficult to reach Brazil’s inflation target, thereby requiring an even more restrictive monetary policy. 

In Brazil, the purchase of foreign exchange (FX) reserves through sterilized interventions—that is FX purchases that are followed by sales of Brazilian treasuries to mop up liquidity and raise the basic interest rate (Selic) to the target level set by the monetary policy committee (COPOM)—has been the object of several criticisms, including my own. The fiscal cost of maintaining reserves amounting to US$ 330 billion is more than R$ 50 billion, exceeding total budget savings promised by the government for this year. Moreover, sterilized interventions have a temporary effect. In order to have a permanent effect, the Brazilian Central Bank (BCB) must continue its purchases, further increasing the costs of holding large reserves.

Economists usually assume that sterilized interventions do not alter aggregate demand. In the first step of a sterilized foreign exchange purchase, the BCB purchases USDs with BRLs, increasing the monetary base. The second step involves withdrawing the BRLs it has issued from circulation by selling government bonds. Thus, the sterilized purchase of foreign exchange induces an exchange of dollar assets for domestic government bonds in the private sector’s portfolio, without issuing any money. Therefore, there is no reason for aggregate demand to expand. 

In contrast to this standard view, I argue in a recent article (Can Sterilized FX Purchases under Inflation Targeting be Expansionary?, Discussion Paper # 589, Department of Economics, PUC-Rio¹) that the BCB’s sterilized foreign exchange purchases have been having an expansionary effect on aggregate demand and consequently on inflation. 

In the previous description, the sterilized purchase of foreign exchange does not produce any net monetary issuance, and merely involves exchanging dollar assets for domestic government bonds. However, this is not how sterilization is performed under inflation targeting. Under inflation targeting, the BCB exchanges BRLs for government bonds up to the point in which the Selic rate established by the COPOM is restored. This does not necessarily assure the repurchase of all money issued. If aggregate demand (and thus money demand) has increased, the interest rate will be re-established before all the money issued has been repurchased. In other words, the way sterilization is usually conducted by central banks, in which the interest rate remains unchanged, may end up expanding the money stock. 

But why would aggregate demand increase? According to my analysis, aggregate demand increases because the recent flows of capital to the Brazilian economy have been destined to financing the private sector’s consumption and investment expenditures. One way of analyzing the effect is that capital flows constitute an “arbitrage” between high domestic financing rates and the low rates prevailing abroad. The larger amount of credit at lower rates expands aggregate demand. Simply restoring the rate of interest that existed before the capital flows is not neutral in terms of stimulating the economy. Even if the Selic rate returns to its initial level, sterilized foreign exchange purchases will have led to net monetary expansion. The greater supply of money will equal the new and higher demand for money generated by the combination of the same rate of interest with a higher nominal output. 

Another way of interpreting this effect, using a portfolio balance framework was suggested to me by two colleagues.² Sterilized foreign exchange purchases increase the share of government bonds in the private sector’s portfolio. To induce this portfolio change it is necessary to alter relative returns on assets, raising the return on government bonds (Selic rate) vis-à-vis other assets. As the BCB acts to keep the Selic rate unchanged, it is the returns on other assets that must fall, which is equivalent to increasing their price. That is, faced with sterilized interventions, private agents try to recompose their original portfolios, increasing the demand for other assets which raises their price, thus expanding aggregate demand. 

According to the above interpretation, the sterilization of the foreign capital flows that seek to take advantage of the high rates of interest paid on Brazilian government bonds does not have an expansionary effect because, in this case, there is an increase in external demand for government bonds. When foreign demand for government bonds rises, the increase in the supply of government bonds due to sterilization does not have the effect of altering the relative returns on the economy’s various assets. In contrast, the inflows that finance private sector expenditures—including not only external credit, but also foreign direct investment, primary and secondary stock purchases and other types of financing—do generate an expansionary effect. 

The figures for the Brazilian economy in 2010 seem to corroborate the importance of the expansionary effect of sterilized interventions. The monetary base increased by 25 percent (R$ 40 billion), while sterilization purchases amounted to R$ 80 billion.³ Credit became more plentiful and cheaper, leading the government to resort to macro-prudential measures—such as increase in reserve requirements and ceilings for loan-to-value and number of installments—to restrict credit supply growth. 

In sum, sterilized interventions not only constitute an expensive and inefficient way of curbing currency appreciation, but also expand aggregate demand, making it even more difficult for the BCB to fulfill the already threatened inflation target. It is time to rethink this policy and once again consider implementing a much more ambitious fiscal consolidation program that will not only help to depreciate the real exchange rate, but also allow the BCB to lower the Selic rate, thereby decreasing speculative capital flows. 

¹ http://www.econ.puc-rio.br/PDF/td589.pdf.
² These are Eduardo Loyo and Samuel Pessoa, who I thank without any implications.

³ The exchange rate at 2010 year-end was 1.6662 BRL per USD, or USD 0.60 per BRL.