The Palestine Economic Policy Research Institute (MAS) Mission Strategic Objectives Board of Trustees

Transcription

1 Palestine Economic Policy Research Institute Money Supply Estimation, Foreign Exchange Risk, and Other Challenges Facing the Palestinian Economy in the Absence of a National Currency Osama Hamed 2015

2 The Palestine Economic Policy Research Institute (MAS) Founded in Jerusalem in 1994 as an independent, non-profit institution to contribute to the policy-making process by conducting economic and social policy research. MAS is governed by a Board of Trustees consisting of prominent academics, businessmen and distinguished personalities from Palestine and the Arab Countries. Mission MAS is dedicated to producing sound and innovative policy research, relevant to economic and social development in Palestine, with the aim of assisting policy-makers and fostering public participation in the formulation of economic and social policies. Strategic Objectives Promoting knowledge-based policy formulation by conducting economic and social policy research in accordance with the expressed priorities and needs of decision-makers. Evaluating economic and social policies and their impact at different levels for correction and review of existing policies. Providing a forum for free, open and democratic public debate among all stakeholders on the socio-economic policy-making process. Disseminating up-to-date socio-economic information and research results. Providing technical support and expert advice to PNA bodies, the private sector, and NGOs to enhance their engagement and participation in policy formulation. Strengthening economic and social policy research capabilities and resources in Palestine. Board of Trustees Samir Hulileh (Chairman), Ghassan Khatib (Deputy Chairman), Magda Salem (Treasurer), Luay Shabaneh (Secretary), Ismail El-Zabri, Jawad Naji, Jihad Al Wazir, Lana Abu Hijleh, Sabri Saidam, Nafez Husseini, Khaled Osaily, Mohammad Nasr, Bassim S. Khoury, Nabeel Kassis (Director General of the Institute - ex officio member) Copyright 2015 Palestine Economic Policy Research Institute (MAS) P.O. Box 19111, Jerusalem and P.O. Box 2426, Ramallah Tel: /4, Fax: , info@mas.ps Web Site:

3 Money Supply Estimation, Foreign Exchange Risk, and Other Challenges Facing the Palestinian Economy in the Absence of a National Currency Lead Researcher: Research Assistants: Osama Hamed Diana Abu Aloyoun Arwa Abu Hashhash "This study has been commissioned and reviewed by the Palestine Monitory Authority (PMA). Partial Funding from PMA is gratefully acknowledged" Palestine Economic Policy Research Institute (MAS) Jerusalem and Ramallah

4 Abstract The absence of a national currency and the circulation of more than one foreign currency subject the Palestinian economy to serious challenges. The lack of a national currency makes it very difficult to estimate money supply, and hence the Seigniorage lost to the issuers of the circulating foreign currencies. It also increases the vulnerability of the Palestinian economy to economic shocks while limiting the ability of Palestinian policy makers to deal with these shocks. As a result of these factors, the circulation of more than one currency makes the Palestinian economy highly vulnerable to foreign exchange risk. In this paper we attempt to estimate Palestinian currency in circulation, money supply (M1), and the currency composition of both, as well as the Seigniorage generated by the Palestinian economy. We estimate Palestinian currency in circulation by using Jordanian currency-gdp ratio and the medium currency-gdp ratio for Arab countries. We evaluate the accuracy of our estimates by assessing the strength of Palestinian currency demand using wage currency composition and wage frequency data for Palestinian workers. We estimate Palestinian money supply by combining our estimates for Palestinian currency in circulation with current account deposits at the Palestinian banking system. We estimate the seigniorage generated by the Palestinian in 2012 by calculating the change in currency holdings in 2012 from its 2011 level, and adding the result to the change in total cash holdings by commercial banks in the same year. We develop a methodology to separate currency demand into its medium of exchange and store of value components. Using this methodology, along with our estimate for Palestinian currency holdings, we calculate the currency composition of Palestinian currency holdings and money supply. We also investigate the vulnerability of the Palestinian economy to external shocks and foreign exchange risk, as well as possible mechanisms to decrease such vulnerability, including direct targeting of external shocks. Additionally, we analyze the impact of foreign exchange risk on households and firms and the effectiveness of commercial banks as financial intermediaries as well as possible regulatory mechanisms to reduce such impact.

5 Table of Contents 1. Introduction 2 2. Estimating Palestinian Money Supply Previous Estimates of Palestinian Money Supply Estimating Money Supply Based on Jordanian Data Estimating Palestinian Money Supply Based on Data from Arab Countries 5 3. Medium of Exchange and Store of Value Components of Money Supply Inventory Approach to Currency Demand Palestinian Wage Frequency and Currency Holding Patterns Estimating the Components of Palestinian Currency Holdings 9 4. Currency Composition of Palestinian Money Supply Seigniorage Generated by the Palestinian Economy Currency Denomination of Product Prices and Wages Currency Competition Currency Denomination of Product Prices Currency Denomination of Wages Vulnerability to Foreign Exchange Risk Vulnerability of Households and Firms Impact on Financial Intermediation Credit and Liquidity Foreign Exchange Induced Risks Monetary Management Vulnerability to External Shocks Direct Targeting of External Shocks The Need for a Lender of Last Resort 25 Annex A: Weighted Average of Currency Holdings a (in NIS) 26 Annex B: Wage Currency Denomination for Selected Employers 28 References 31

6 1. Introduction In the absence of a national currency, three main currencies circulate in the West Bank and Gaza Strip (WBGS): the new Israeli shekel (NIS), the U.S. dollar (dollar), and the Jordanian dinar (JD). The absence of a national currency makes estimating Palestinian money supply very difficult and increases the vulnerability of the domestic economy to external economic shocks. It also severely limits the effectiveness of monetary policy as a macroeconomic stabilizing instrument. The circulation of three currencies simultaneously makes the Palestinian economy highly vulnerable to foreign exchange risk. In section 2 of the paper we survey previous estimates of Palestinian money supply. We then make estimates using the Jordanian currency-gdp ratio and the median currency-gdp ratio for Arab countries. In section 3 we develop a methodology for separating money supply into its medium of exchange and store of value components. We then use this methodology to calculate the two components of Palestinian money supply and to evaluate the accuracy of money supply estimates made in section 2. In section 4 we calculate the currency composition of Palestinian money supply based on the methodology developed in section 3. In section 5 we estimate the Seigniorage generated by the Palestinian economy. In section 6 discuss the currency denomination of prices and wage in the West Bank and Gaza Strip. In section 7 we investigate different types of foreign exchange risk experienced by the Palestinian economy and policy options to address these risks. In section 8 we explore issues related to monetary management in the absence of a Palestinian currency, including the need for a lender of last resort. Some of the data used in the paper were collected by MAS research assistants. Other data were produced specifically for this paper by the Palestinian Central Bureau of Statistics (PCBS) and the Palestinian Monetary Authority (PMA). 2

7 2. Estimating Palestinian Money Supply 2.1 Previous Estimates of Palestinian Money Supply There are numerous measurements of money supply. In general, money supply consists of currency in circulation and bank deposits. The narrowest measurement (M1) includes mainly current account (demand) deposits and currency in circulation. A broader measurement (M2) also includes savings accounts, small certificates of deposits (time deposits), in addition to the assets that are part of M1. M3 incorporates additional types of bank deposits. In the absence of a national currency, it is challenging to estimate Palestinian currency in circulation, and hence domestic money supply, because of the lack of currency in circulation data. Between 1967 and 1994, Palestinian money supply consisted largely of currency in circulation because bank deposits were negligible in this period. All banks that were in operation in the West Bank and Gaza Strip (WBGS) on the eve of Israeli occupation in 1967 were closed by the Israeli military authorities. Between 1967 and the early 1980s, banking services in the WBGS were limited to a few Israeli bank branches that catered mainly to Israeli settlements in the WBGS. Between 1981 and 1993, the Bank of Palestine was allowed to re-open one of its branches in the Gaza Strip, and a Jordanian bank (Cairo-Amman Bank) was given permission to open a few branches in the West Bank. By the end of 1993, total WBGS bank deposits was only $ 219 million, which represented 9.4 % of GDP 1. The Bank of Israel estimated currency in circulation in the WBGS in 1969 at $117 million, or 60.6 % of GDP (Arnon, 1996). Another estimate of WBGS currency in circulation in the period was made by Hamed (1993). This estimate was limited to the period and covered only the Israeli currency circulating in the WBGS. The estimate assumed that the Israeli currency served as the main medium of exchange in this period while the Jordanian currency served as a store of value. The Palestinian currency in circulation was estimated under two alternative assumptions. The first alternative assumed that the currency-gnp ratio for the WBGS was the same as Israel s. Based on this assumption, Palestinian currency in circulation in 1987 was estimated at $55.9 million, accounting for 3.2 % of Palestinian GNP. The second alternative assumed that the Palestinian currency-gnp ratio was equal to the sum of Israeli currency in circulation and Israeli demand deposits divided by Israeli GNP. Based on this assumption, currency in circulation in the WBGS in 1987 was estimated at $133.4 million, which represented 7.5% of Palestinian GNP. Hamed (1993) argued that the second estimate was more realistic because checks, credit cards, and other alternatives to currency as means of payment were much more widely used in Israel than the WBGS. The Palestinian banking system has expanded substantially since By the end of 2013 the number of bank of branches and offices reached 237, and total bank deposits were $8,666.2 million, accounting for 86.6% of Palestinian GDP. The expansion of the Palestinian banking system has undoubtedly decreased Palestinian demand for currency as a store of value and may have decreased demand for currency as a medium of exchange slightly. Nevertheless, the unstable political and economic environment, and the fact that the use of alternatives to cash as a medium of exchange (such as credit cards and checks) are very limited, means that demand for currency has continued to be strong in the post-1994 period 2. There have been a number of estimates for currency in circulation and money supply in the post period. Erickson von Allman and Fisher used the Middle East and North Africa median currency-gdp ratio of 5.5 % to estimate the 1999 currency in circulation in the WBGS at $ For more information about the Palestinian banking system in the period , see Hamed (1996). It has been long established that demand for currency tends to decline as a country s payment system becomes more advanced and to decrease during periods of economic and political uncertainty. Demand for currency is also known to vary inversely with the frequency of wage payment, real interest rates, and inflation rate. See Baumol (1952) and Tobin (1958). 3

8 million. Arnon (1996) put Palestinian currency in circulation in 1995 at $395 million, accounting for 12.3 % of GDP. This figure was obtained by estimating a money demand function using data from Jordan, Syria, and Egypt. Wazir (2011) estimated Palestinian currency in circulation for the period using the Jordanian currency-demand deposits ratio. Using this method, they estimated Palestinian currency in circulation in 2008 at $1,766 million, which is equal to 28.3 % of Palestinian GDP in the same year. 2.2 Estimating Money Supply Based on Jordanian Data One way to estimate Palestinian currency in circulation, and hence money supply, is to assume that the Palestinian currency-gdp ratio is equal to the corresponding Jordanian ratio. This assumption can be justified by the cultural similarities between the two countries. To estimate Palestinian money supply using Jordanian data, we need to account for recent fluctuation in Jordanian currency in circulation due to the influx of Iraqi and Syrian refugees. We also need to take into consideration the fact that the Jordanian currency (JD) circulates in both Jordan and the WBGS. To address the issue of refugees, we decided to use the 2011 Jordanian currency in circulation as a basis for calculating Palestinian money supply. We chose 2011 because the number of Syrian refugees was still limited, while the number of Iraqi refugees was at a low level due to relative political stability in Iraq on the eve of the US military withdrawal. This puts the Palestinian currency-gdp ratio at 9.98%. To account for the circulation of JD in the WBGS, we use the following formula to calculate Palestinian money supply (M1) : M pi = (CU jd11 / GDP j11 + GDP p11 )* GDP pi + DD pi, where M pi = 2012 Palestinian money supply (M1) in year i CU jd11 = total JD in circulation in 2011 GDP j11 = Jordanian GDP in 2011 GDP p11 = Palestinian GDP in 2011 GDP pi = Palestinian GDP in year i DD pi = total demand deposits at the Palestinian banking system at the end of year i Table 1: Palestinian Money Supply (M1) Based on Jordanian Data Jordanian 2011 GDP in million JD (GDP j ) 20,477 Jordanian currency in circulation in million JD (CU jd ) 3,019 Jordanian currency-gdp ratio in % (CU jd /(GDP j +GDP p ) 9.98 Palestinian GDP (mn $) 9, ,279 Palestinian currency in circulation (mn $) ,125.6 Palestinian demand deposits (mn $) 2,773 2,984 Palestinian money supply (mn $) 3, ,109.6 Sources: Jordanian GDP data is from Central Bank of Jordan A, Jordanian currency in circulation from Central Bank of Jordan B, Palestinian demand deposits data is from the PMA, and Palestinian currency in circulation is from Annex A. Based on the above formula, we estimated Palestinian money supply to be $3,748.6 million in 2011 and $4,109.6 million in 2012 (see Table 1). Currency in circulations accounts for $975.6 million of money supply in 2011 and $1,125.6 in

9 2.3 Estimating Palestinian Money Supply Based on Data from Arab Countries Table 2: Currency GDP Ratios for Arab Countries Country Currency-GDP ratio Country Currency-GDP ratio % % % Qatar 1.47 Djibouti 9.24 Kuwait 3.02 Jordan 9.98 UAE 3.26 Tunisia Bahrain 3.68 Yemen Oman 3.88 Egypt Lebanon 4.79 Iraq Saudi Arabia 4.78 Algeria Sudan 7.58 Morocco Mauritania 8.69 Sources: calculated using data obtained from the IMF (2014). Palestinian currency in circulation and money supply can also be estimated using currency circulation data from other Arab countries. The currency in circulation-gdp ratio varies considerably between Arab countries. It ranges between 1.47% for Qatar and 20.71% for Morocco (see Table 2). Using the 8.69 % median for the 17 Arab countries for which we obtained data, we calculated Palestinian money supply using the following formula: M pi = (CU/GDP ) a * GDP p + DDP p, where M pi = Palestinian money supply in year i. (CU/GDP ) a = the median of currency-gdp ratio for Arab countries GDP pi = Palestinian GDP in year i DDP pi = Palestinian demand deposits in year i Table 3: Palestinian Money Supply (M1) Based on Arab Countries Data Median Currency-GDP Ratio for Arab Countries ( %) Palestinian GDP (mn $) 9, ,279 Palestinian Currency in Circulation in(mn $) Palestinian Demand Deposits in (mn $) 2,773 2,984 Palestinian Money Supply in (mn $) 3,623 3,964 Sources: Palestinian GDP data is from the PCBS, Palestinian demand deposits data is from the PMA, and the median currency-gdp ratio is from Table 2. Our Palestinian money supply estimate based on the median currency in circulation-gdp ratio for Arab countries is $3,623 million for 2011 and 3,964 million for 2012 (see Table 3). 5

10 3. Medium of Exchange and Store of Value Components of Money Supply Economic agents demand for a currency depends on its value as a medium of exchange and as a store of value. A country s medium of exchange currency demand is a function of the country s GDP, interest rates, and the availability of alternatives to currency as a means of payment. Its demand for currency as a store of value depends on the depth of its financial system and the degree of economic literacy of its population. In this section, we use wage frequency and banking data for Palestinian workers to assess the relative strength of Palestinian currency demand in order to evaluate the accuracy of the Palestinian currency holdings estimates made in section 2. We also develop a methodology for separating currency demand into its medium of exchange and store of value components, and use this methodology to estimate the shares of each of the two components in Palestinian currency holdings. This separation allows us to calculate the currency composition of Palestinian money in section Inventory Approach to Currency Demand Advances in payment systems, financial deepening, and increased financial literacy have resulted in a sharp decrease in the currency-gdp ratio in developed countries in the last few decades. As an example, the UK currency-gdp ratio at the end of 2013 was 2.7%. In contrast, most developing countries still have relatively high currency GDP ratios (see Table 4). To better understand the persistence of high currency-gdp ratio in developing counties let us separate currency in circulation into its medium of exchange and store of value components. We can do so by using the inventory approach to currency demand to estimate the medium of exchange component. The inventory approach, which has its theoretical foundations in Baumol (1952) and Tobin (1956), starts with the assumption that income streams received by households and expenditures made by them are not synchronized. This makes it necessary for households to hold liquid assets for transaction purposes. The transaction balance does not have to be in cash. A portion of it can be held in income-generating financial assets that can easily be converted to cash. At the beginning of each period, the household is assumed to set aside a certain percentage of its income as cash and use the rest to acquire financial assets that are converted to cash uniformly throughout the period. The conversion of financial assets to cash is assumed to have pecuniary (such as a broker s fee) as well as non-pecuniary costs (time and effort). The number of conversions per pay period, and hence the average household cash holdings, is assumed to be a function of the frequency of wage payment, market interest rates, and the average cost per withdrawal. A rise (fall) in the cost of conversion, other things being equal, decreases (increases) the number of conversions made each period and hence increases (decreases) the average cash holdings. A rise (fall) in market interest rates, other things being the same, increases (decreases) the number of withdrawals, hence decreasing the average cash holdings. The inventory approach has usually been used in the economic literature to estimate transaction demand for money using data on income, market interest rates, wages, and the cost of cash withdrawal from income-generating assets. In this paper, we utilize the inventory approach to separate currency-gdp ratio into its medium of exchange and store of value components. In this section, we use the inventory approach to estimate the upper bound of medium of exchange component and then use the result to calculate the lower bound of the store of value component for a number of developing countries. In section 3.3, we use the wage payment frequency patterns and banking habits of Palestinian workers to estimate the medium of exchange component of the currency-gdp ratio. 6

11 Table 4: A Lower Bound for the Store of Value Component of the Currency-GDP Ratio for Some Developing Countries Country Currency- GDP Ratio % Store of Value Component % Bolivia Thailand India Peru Jordan Tunisia Iraq Morocco Sources: calculated using data from the IMF (2014). As an illustration of the way household currency holdings are determined under the inventory approach, let us suppose that a household receives $1000 in wages per month and these wages are the only sources of income for the household and are paid on a monthly basis. Let us also suppose that the household keeps $250 of its income in cash (to cover one week expenditures) and deposits the rest in an interest-bearing bank account, and then withdraws equal amounts from the account on a weekly basis until the next pay period. Assuming that the household s expenditures are spread out uniformly throughout the month, the household is expected to have an average cash balance of $125 ($250/2). If the same household is paid twice a month, its average household balance is expected to be half that amount ($62.50). If all households in a country receive their income streams in cash on a monthly basis, we should expect the medium of exchange component of the country s currency-gdp ratio to be around 4.17% 3, which is the upper limit for this ratio 4. If some households receive their income streams more frequently (bi-weekly, weekly, or daily), and/or invest some of them in income generating assets, the medium of exchange component currency-gdp should be lower. In Table 4, we subtract 4.17 (the upper bound of the medium of exchange component of the currency-gdp ratio) from the currency-gdp ratios of a number of developing countries to get the lower bound of the store of value component for these ratios. Based on Table 3, a large percentage of currency in circulation in most of these countries is used as a store of value. In Jordan, currency holdings that serve as a store of value account for 5.81% of GDP. In Morocco, it is 16.54%. 3.2 Palestinian Wage Frequency and Currency Holding Patterns A country s wage payment frequency patterns and banking habits provide valuable information about the magnitude of its households currency holdings, and hence its currency-gdp ratio. If a country s households receive their wage earnings infrequently (frequently) and/or do not (do) invest a significant portion of their earnings in income-generating financial assets, we should expect the medium of exchange component of the country s currency-gdp ratio to be relatively high (low). Information about the willingness of a country s households to invest wage earnings in income-generating assets also provides us with an important indication about the size of the store of value component of its currency-gdp ratio. In the rest of this section, we analyze wage frequency and banking data for Palestinian workers obtained from the PCBS 5. Based on this Monthly GDP divided by 2. Technically, the resulting ratio is the currency-income ratio and not the currency-gdp ratio, but the differences between the two ratios are usually small and are not relevant to the issues since they result in an overestimate of the medium of exchange component of the currency-gdp ratio. The April 2014 round of the PCBS Labor Survey included a number of questions proposed by MAS about wage payment frequency, currency denomination of wagers, and banking habits. 7

12 analysis, we make some conclusions about the relative strength of the medium of exchange and store of value components of Palestinian demand for currency. These conclusions are in turn used to evaluate the accuracy of the Palestinian money supply estimates we made in section 2. The wage payment frequencies and banking data analyzed in this section will be used in section 3.3 to estimate Palestinian households currency holdings. Table 5: Payment Methods by Type of Employer Employer Type Cash Transfer Wire Check % % % Public Sector Private Sector NGOs Israel and Settlements Source: calculated based on PCBS data. Most Palestinian workers in the WBGS are paid either in cash (52.8%) or wire transfers (45.1%). Only 2.1% are paid by check. Paying workers in cash is much more common in the private sector (75.0%) and for Palestinian workers in Israel and Israeli settlements (74.4%) than I the public sector (10.1%) (see Table 5). Very few of the workers who are paid in cash deposit their pay in bank accounts, which are the main income-generating financial assets in the WBGS. The workers who are paid in cash and keep their pay in cash account for 52.0% of all workers. Additionally, a vast majority of workers who are paid by wire transfer or check convert their wages to cash immediately. Such workers account for 41.5% of all workers. This means that 93.5% of workers keep no balances at commercial banks, limiting the role of the banks operating in the WBGS as financial intermediaries. Table 6: Payment Methods by Economic Sector Cash Transfer Wire Check Sector in % in % in % Agriculture Manufacturing Construction Commerce-Hotels Transport-Storage Others Source: calculated using PCBS data. The percent of workers paid in cash varies significantly by economic sector [see Table 6]. In construction and agriculture, the percentages of workers paid in cash are 83.4% and 82.2%. In transport and storage, on the other hand, the share of workers paid in cash is 57.1%. Table 7: Frequency of Wage Payment by Type of Employer Employer Type Monthly Bi-Weekly Weekly Daily % % % % Public Sector Private Sector NGOs Workers in Israel and Settlements Sources: calculated based on PCBS data. 8

13 Almost all public sector workers (98.4%) are paid monthly (see Table 7). The percentage of workers paid monthly is also very high for NGOs (91.1%). In contrast, the percentages of monthly wage payment at the private sector and among Palestinian workers in Israel and Israeli settlements are 61.0% and 67.8%. For all workers, the shares of monthly, bi-weekly, weekly, and daily wage payments are 73.4%, 2.8%, 15.6%, and 8.2%, respectively. Table 8: Frequency of Wage Payment by Economic Sector Sector Monthly Bi-Weekly Weekly Daily % % % % Agriculture Manufacturing Construction Commerce-Hotels Transport-Storage Others Sources: Calculated using PCBS data. The frequency of wage payment varies widely by economic sector. At commercial banks, all workers are paid on a monthly basis. In the construction sector, only 49.8% of workers are paid monthly while 9.6%, 30.7%, 9.9% are paid bi-weekly, weekly, and daily basis, respectively (see Table 8). We have shown in this section that 94% of Palestinian workers hold their earnings in cash. This indicates a strong demand for currency as a store of value in the WBGS. We have also shown that 73% of all workers are paid monthly, indicating strong demand for currency as a medium of exchange. Given the strength of demand for currency as a store of value as well as a medium of exchange, we should expect the Palestinian currency-gdp ratio to be relatively high, justifying the use of the relatively high Jordanian currency-gdp ratio in estimating Palestinian currency holdings and money supply. 3.3 Estimating the Components of Palestinian Currency Holdings The wage frequency patterns and banking habits of Palestinian workers presented in section 3.2 indicate the Palestinian medium of exchange currency-gdp ratio to be relatively high. To estimate this ratio, we first calculate the weighted average of cash holding of Palestinian workers using the following formula: CU = Σ (K i W i /2) where, CU = average daily currency holdings in NIS. K i = the share of group i as a percentage of all workers in the sample, where i takes the following values: workers paid daily, workers paid weekly, workers paid bi-weekly, workers paid monthly. W i = the average wage of group i. Based on our calculations (see Annex A), the average household currency holdings derived from estimate wage earnings of NIS 2, Assuming similar currency holding patterns for other income streams received by households, this implies that the medium exchange component of the Palestinian currency-gdp ratio is around 3.14 % (after dividing 37.68% by 12 to find the ratio as a percentage of annual GDP). This is relatively high, given that the theoretical upper limit for the medium exchange component of the currency-gdp ratio is 4.17% (see section 2.1), supporting our conclusion in section 3.2 that the medium exchange component of Palestinian currency demand is relatively high. This puts the value of the medium of exchange currency holdings of Palestinian households in 2011 at $306.9 million. 9

14 When we subtract our estimate of the medium of exchange component of Palestinian currency- GDP from the estimate made in section 1 of Palestinian currency-gdp ratio based on Jordanian data (9.98%), we get an estimate of 6.84% for the store of value component of the Palestinian currency-gdp ratio. This put the store of value currency holdings of Palestinian households in 2011 at $668.6 million. 10

15 4. Currency Composition of Palestinian Money Supply Table 9: The shares of the three circulating currencies in Palestinian Currency Holdings and Money Supply, Currencies NIS Dollar JD All Medium of Exchange Currency Holdings (mn $) Store of Value Currency Holdings (mn $) Total Currency Holdings (mn $) Share in Total Currency Holdings* ( %) Demand Deposits (mn $) 1, Money Supply (M1 mn $) , ,748.6 Share of Money Supply *(%) Sources: calculated using data obtained from the PCBS and the PMA. *Percentages do not add up because of holdings in other currencies Using the methodology developed in section 3, we estimate the shares of the NIS, dollar, and JD in the medium of exchange component of Palestinian currency holdings at 91.4%, 5.6%, and 3.0%, respectively [See Annex A]. Using our $306.9 million estimate of the medium of exchange component of Palestinian currency holdings [see section 3.3], we conclude that $280.8 million of these holdings are in NIS, $17.9 million are in dollar, and $9.2 million are in JD. To estimate store of value currency holdings denominated in each of the three circulating currencies, let us assume that the shares of the three currencies in the store of value component of Palestinian currency holdings is comparable to their shares in the interest bearing bank deposits at the Palestinian banking system. This is not an unreasonable assumption in view of the fact that interest bearing bank deposits serve as the main store of value for Palestinian households. At the end of 2011, the NIS accounted for 21.2% of interest bearing deposits at the Palestinian banking system while the dollar and the JD accounted for 40.6%, and 34.2%, respectively. Based on these shares and our estimate of the total store of value currency holdings of Palestinian households using Jordanian data obtained above ($668.6 million), we estimate the NIS, dollar, and JD store of value currency holdings in 2011 at $141.6 million, $271.5 million, and $228.7 million, respectively. To estimate the currency composition of Palestinian currency holdings, we add the medium of exchange and store of value holdings denominated in each of the three currencies to find total currency holdings denominated in the same currency. We then use these results to calculate the shares of the three currencies in total Palestinian currency holdings (see Table 9). Based on our calculations, we estimate the shares of the NIS, dollar, and JD in Palestinian total currency holdings in 2011 to be 43.3%, 29.7%, and 24.4%, respectively. To estimate the currency composition of Palestinian money supply (M1), we add the medium of exchange holdings denominated in each of the three currencies with the current account deposits denominated in the same currency, and then divide the result by our estimate of the Palestinian money supply made in section 1.2. Based on our calculations, we estimate the shares of the NIS, dollar, and JD in the Palestinian money supply (M1) in 2011 at 44.8%, 32.5%, and 17.2%, respectively (See Table 9). 11

16 5. Seigniorage Generated by the Palestinian Economy Seigniorage can be defined as revenues associated with issuing a currency. A country s annual seigniorage is usually calculated by subtracting the value of the country s monetary base at the end of the year in question from its level at the end of the previous year. Seigniorage consists of two components. The first is derived from the increase in total currency outstanding, and the second is generated by income earned by investing required bank reserves. Seigniorage increases with economic growth. Moderate inflation may also result in higher seigniorage revenue 6. A country s currency is usually issued by its monetary authorities, who use the seigniorage generated by the currency to cover operating cost and turn over the remainder to the country s Treasury. Most of the seigniorage generated by the Palestinian economy in the period was retained by the Bank of Israel and the Central Bank of Jordan, the issuers of the main currencies in circulation in the WBGS during this period. Palestinian seigniorage received by the Bank of Israel was derived mainly from the use of the Israeli currency as the main medium of exchange, while that retained by the Central Bank of Jordan was obtained mostly from the use of the JD as a store of value and the investment of required reserves on Palestinian deposits in the Jordanian banking system. The Palestinian seigniorage revenues retained by the Bank of Israel in this period were estimated by Hamed (1993) as part of a study about the unbalanced economic relations between the Palestinian and Israeli economies in the period The study assumed that the WBGS monetary base consisted of currency in circulation only, which is not an unreasonable assumption since WBGS bank deposits were negligible in this period. Hence, it calculated the annual seigniorage by finding the change in currency in circulation in two consecutive years. The study estimated the average seigniorage for the period at 1.4 % to 4.2% of Palestinian GDP 7. Measured in 1987 US dollars, the study estimated total Palestinian seigniorage retained by the Bank of Israel in this period at $.7 billion to $1.8 billion dollars. The Palestinian economy has been able to retain some of its seigniorage in the post-1994 period because Starting in 1995 deposits at the Palestinian banking system have been subject to PMA reserve requirement and the seigniorage generated by investing bank reserves have been retained by the PMA. Additionally, there has been a substantial increase in deposits at the domestic banking system since Some of the increase in bank deposits was due to decreased Palestinian dependence on currency notes as a store of value and to attracting Palestinian deposits from the Jordanian banking system to the Palestinian banking system. There have been some attempts to estimate Palestinian seigniorage in the post 1994 period. Arnon (1996) estimated Palestinian seigniorage for 1995 as part of a study about potential seigniorage revenues under an independent Palestinian state. Based on a regression analysis of Jordanian, Egyptian, and Syrian time series data, Arnon concluded that the long run average of seigniorage that can be generated by a currency issued by a future Palestinian state to be 1.2% of Palestinian GDP 8. Von Allman (2001) estimates the seigniorage that could be generated by a Palestinian currency issued under a currency board arrangement, assuming that the Palestinian currency is fully backed by foreign reserves and the reserves are invested on the international market at a 5% annual rate of return. Von Allman s estimate, which assumes a Palestinian currency in circulation- GDP ratio of 5.5%, put the 1999 Palestinian seigniorage at $12 million, or.3% of GDP. While the PMA currently retains some of the revenues generated by the Palestinian economy by subjecting the Palestinian banking system to interest free reserve requirements and investing these A high inflation rate may decrease seigniorage because of its negative impact on real money demand. The low bound estimate is based on the assumption that the WBGS currency-gdp ratio was equal to Israel s. The high bound assumed it was equal to the sum of Israeli currency in circulation and demand deposits divided by Israeli GDP. According to Arnon (1996), a Palestinian currency would generate substantial seigniorage in the fist five years, averaging 22.9% of GDP. Seigniorage would decrease quickly after that to a long run average of 1.2% of GDP. These estimates assume implicitly that the Palestinian currency would not be introduced under a currency board arrangement. 12

17 reserves, the Palestinian economy still loses substantial seigniorage to the issuers of the currencies that circulate in the WBGS. Issuing a Palestinian currency would allow the Palestinian economy to capture some but not necessarily all of the lost seigniorage. The capacity of the new currency to reclaim Palestinian seigniorage would depend on the foreign exchange regime under which the currency is issued 9. Table 10: Lost Palestinian Seigniorage Currency Holdings of Palestinian Households Based on Jordanian Data (mn $) ,125.6 Based on Data from Other Arab Countries (mn $) Currency Holdings of Commercial Banks (mn $) Total Currency Holdings in the WBGS Based on Jordanian Data (mn $) 1, ,856.6 Based on Data from Other Arab Countries (mn $) 1,347 1,711 Seigniorage Retained by Issuers of the Three Circulating Currencies Based on Jordanian Data (mn $) 384 Based on Data from Other Arab Countries (mn $) 364 Sources: household currency holdings are from Table 9; currency holdings of commercial banks are from the PMA We estimated the 2012 Seigniorage retained by the issuers of the currencies that circulate in the WBGS. We did so by calculating the change in currency in circulation in 2012 and adding the result to the change in total cash holdings by commercial banks in the same year (see Table 10). Based on our calculations, Palestinian Seigniorage retained in 2012 by the issuers of the three circulating currencies ranged between $364 million to $384 million, or 3.2 % to 3.4% of Palestinian GDP. 9 The issue of choosing a foreign exchange regime, and hence the capacity of a Palestinian currency to capture the seigniorage generated by the Palestinian currency, is not within the scope of this paper. 13

18 6. Currency Denomination of Product Prices and Wages 6.1 Currency Competition The dollarization literature, which deals with situations where foreign currencies circulate simultaneously alongside national currencies, has addressed since the 1970s some of the issues related to currency selection in a multi-currency environment. However, the micro foundations of currency selection by economic agents were not modeled until the 1990s, in an area of research currently known as currency competition. The currency competition literature makes a distinction between network and non-network benefits of a currency. Network benefits are assumed to be a function of the number of its users and are attributed to economic agents increased understanding of prices quoted by others. Non-network benefits are derived from the intrinsic properties of the currency, such as the redemption value under the gold standard. Some currency competition models (Dawd,1993) assume that non-network benefits are the same for all individuals, while others (Stenkula, 2003) assume that they vary across individuals due to differences in emotional attachments (such as nationalistic feelings) and preferences regarding economic issues (such as inflation rates). The circulation of more than one currency within a country s borders is likely to occur if the nonnetwork benefits of circulation currencies are relatively high and vary significantly across economic agents. To minimize currency conversion cost and exposure to foreign exchange risk, a household (firm) operating in a multi-currency environment is expected to select the currency that minimizes the mismatch between its income (revenues) and its expenditures (costs) 10. When an economic agent switches from one currency to another, he/she incurs switching costs associated with becoming accustomed to reckoning, keeping accounts, and quoting prices in the new currency. Hence, a switch is likely only if the sum of the network and non-network benefits resulting from the switch exceed the switching costs (Dawd, 1993). In section 6.2 and 6.3, we will present Palestinian product prices and wage payment patters and will try to explain them using currency competition models. The patterns presented in these sections will be used in section 7 in our evaluation of the vulnerability of the Palestinian economy to foreign exchange risk. 6.2 Currency Denomination of Product Prices Except for a limited number of items, goods and services in the WBGS have been priced in Israeli currency since Soon after occupying the WBGS, the Israeli government implemented a number of measures that established the lira (the Israeli currency at the time) as the main medium of exchange in the WBGS. These measures included making the lira legal tender, closing all commercial banks that were in operation on the eve of the Israeli occupation, and curtailing WBGS trade with the rest of the world to make the Palestinian economy dependent on Israel 11. The persistence of Israeli currency as the medium of exchange in the post-1994 period (the NIS, by then) should not be surprising, given the WBGS s continued overdependence on Israel for imports, and the fact that most Palestinian imports are denominated in the NIS because they originate in Israel or are imported from other countries through Israeli firms 12. Consequently, the vendors of most imported goods have continued to price their products in the NIS to minimize currency conversion cost and foreign exchange risk exposure. This in turn has created network benefits large enough to make most other vendors in the WBGS use the NIS to price their products See Ekman, Israel kept two border crossings between the West Bank and Jordan and Palestinians were allowed to export agricultural products through these crossing to Jordan and Gulf countries, but no imports were allowed. In 2013, Israel accounted for 71.4% of Palestinian imports. 14

19 The items that were not priced in the Israeli currency in the period tended to be high ticket items that were sold on installment plans (such as electric appliances ) or under long term contracts (such as rental agreements). These items were usually priced in JD because of the high depreciation rate of the Israeli currency for most of the period. This practice has continued in the post-1994 period despite the fact that the foreign exchange rate of the Israeli currency has been quite stable in this period. Anecdotal evidence indicates that most rental agreements are currently denominated in the JD or the dollar. Electric appliances and cars are priced in the dollar or the JD, according to conversations with some vendors in Ramallah and Al- Bireh. The dominant role of the JD and dollar in rental agreements may be explained by the importance of bank credit to financing the construction of new commercial and housing rental units, and the fact that most credit extended by the Palestinian banking system is in dollar. At the end of the second quarter of 2014, the dollar accounted for 85% of bank credit to housing and real estate activities. The JD accounted for another 8.5%. The share of the NIS was only 6.4%. The tendency for appliance and auto vendors to price their products in the dollar or the euro may be explained by their need to match the currencies in which their products are invoiced, since these products are usually imported directly. 6.3 Currency Denomination of Wages Table 11: Currency Denomination of Wage by Type of Employer Employer Type NIS Dollar JD % % % Public sector Private sector NGOs Israel and Settlements Source: calculated using PCBS data. Most Palestinian workers (93.0%) are now paid in the NIS. The shares of workers paid in the dollar and the JD 4.5% and 2.5%, respectively. The share of workers paid NIS is particularly high for the public sector (98.7%) because the PNA accounts for most public sector employees, and PNA wages are denominated in the NIS. In contrast, the shares of workers paid in the NIS for the private sector and NGOs are 92.2% and 38.1%, respectively (see Table 11). The decision of the PNA to pay its employees in NIS should not come as a surprise in view of the currency composition of its revenues. In 2013, Palestinian tax revenues collected by Israel on PNA s behalf, which are denominated in NIS, accounted for 45.9% of total PNA revenues. In the same year, revenues generated by taxes and fees collected directly by the PNA, which are denominated mainly in NIS, account for another 34% of total PNA revenues. Currency denomination of private sector workers varies widely. In the manufacturing and agricultural sectors, where most prices (and hence revenues) are denominated in NIS, the shares of workers paid in NIS are 97.9% and 96.3%, respectively. In contrast, commercial banks pay their employees in either the dollar or the JD, whose exchange rate is fixed in relation to the dollar (see Annex B). This can be justified by the need to minimize exposure to foreign exchange risk since interest income accounts for most of their revenues, and the fact that the bulk of their interest earning assets are denominated in either the dollar or the JD. Wages of workers employed by the companies listed on the Palestinian Stock Exchange are also denominated in either JD or dollar, even though their product prices (and hence their revenues) are denominated in the NIS. Of the 23 listed companies, 15 pay their workers in the dollar, 5 in the JD, and 3 in the NIS. To understand the currency denomination of wage payments by listed companies, we need to keep in mind that the dollar and dinar are the main stores of value in the WBGS and are likely 15

20 to be so for most shareholders. To minimize foreign risk for shareholders, most listed companies use the dollar or JD to price their shares and pay their employees. Paying employees in dollar or JD is particularly common among NGOs because of their dependence on donors as a source of revenues. Only 38.2% of those employed by NGOs are paid in the NIS, while 50.2% are paid in the dollar and 11.6% are paid in the JD. The dollar and JD are also widely used in paying wages by universities and private schools. All Palestinian universities pay their employees in the JD except one, which pays its workers in dollar (see Annex B). Of the 8 private Ramallah schools MAS contacted, only 2 pay their employees in the NIS while 4 make their wage payments in dollars and 2 in JDs. Private schools use of the dollar as a currency of wage payment may be explained by their dependence on foreign donors. The use of the JD as wage payment currency in most universities is driven by the fact that student tuition fees, which have in recent years become the main source of their revenues, are denominated in JD. The universities use of the JD in setting their tuition fees is a legacy of the pre-1994 period, when university budgets were denominated in the JD because JD denominated funds received from donors in Arab countries accounted for most revenues. 16

21 7. Vulnerability to Foreign Exchange Risk 7.1 Vulnerability of Households and Firms The circulation of three major currencies in the WBGS subjects households as well as firms to substantial foreign exchange risk. The foreign exchange rate of the JD is fixed in relation to the dollar and the JD-dollar parity has remained effectively the same since Hence the foreign exchange risk experienced by Palestinian households and firms is determined mainly by the degree of fluctuation in the dollar-nis exchange rate, which since 2005 has been determined by market forces. Figure 1: Quarterly Inflation Rates and Changes in Dollar Purchasing Power: Sources: calculated using data obtained from the PCBS and the PMA. The exposure of households to foreign exchange risk is manifested through increased variability of household purchasing power and consumption expenditures. The payment of wages in three different currencies decreases the accuracy of the Palestinian inflation rate as a measurement of the purchasing power of wages. This in turn distorts household consumption choices and decisions made by firms regarding their input mix and output levels. To account for the impact of changes in the NIS exchange rates on the purchasing power of WBGS households, we adjusted PCBS inflation rates for changes in the NIS-dollar and NIS-JD by subtracting the percentage change in the foreign exchange rates from the inflation rate (see Figure 1). Based on our calculations, it is clear that wage payment in more than one currency increases the variability of household purchasing power, which in turn increases the variability of consumption expenditures and real GDP. For the period , the standard deviation for the unadjusted inflation rate is 1.6%. When we adjusted for changes in the NIS-dollar and NIS-JD exchange rates, the standard deviation increased to 3.70% and 3.64%, respectively. Paying wages in more than one currency also increases the variability of relative wages. A depreciation (appreciation) of the NIS against the dollar (JD) decreases (increases) the wages of workers paid in NIS relative to those paid in dollar (JD). This in turn distorts the effectiveness of relative wages in allocating labor among different sectors of the economy. The impact of NIS depreciation (appreciation) is particularly strong on households that live in rental housing units because rents are usually denominated in the dinar or dollar while most wages are denominated in the NIS. 13 In 1988 Jordan had a serious balance of payment crisis that resulted in the devaluation JD against the U.S. dollar by 40%. 17