Box from the forthcoming Bank of Israel Annual Report for 2021: What Explains the Increase in Tax Receipts in 2021?
- Tax receipts in Israel in 2021 were high not just compared to 2020—a year in which they declined—but also relative to what was expected according to the long-term growth trend over the five years that preceded the COVID-19 crisis.
- The NIS 67 billion increase in tax collection compared with 2019 was partly due to the “natural” increase of about NIS 34 billion. The deviation of macroeconomic variables from the trend of the previous five years contributed a further NIS 29 billion.
- Due to the significant weight of the exceptional components—consumer goods imports, increases in financial asset prices, and the strong activity in the real estate market—tremendous caution is necessary in relying on this revenue in planning toward the next budget.
Tax receipts in Israel in 2021 were high not just compared to 2020—a year in which they declined—but also relative to what was expected according to the long-term growth trend in recent years. There are a number of complementary explanations for this increase in tax revenue: structural factors, exceptional factors, and one-off factors. In order to answer the question of how much of the increase in receipts was due to each of these components, we analyze the development of the relevant macroeconomic variables in the past two years compared with the preceding five years.
Structural factors: The structural growth of tax revenues is due, first and foremost, to the “natural” growth factors in the tax base: GDP growth, which is due to the increase in the number of workers and in GDP per worker. Long-term structural processes, such as an increase in the motorization rate and fuel consumption in Israel (tax-intensive activities) and changes in income distribution (because the marginal tax on those with higher salaries is greater) have in the past contributed to an expansion of the tax base. The decline of the volume of the shadow economy and the permanent streamlining of Tax Authority activities may also lead to a structural increase in revenues. Another structural change in tax receipts may be due to legislative changes that raise the statutory tax rates or reduce tax benefits.
Exceptional factors: Some tax revenues change with the business cycle, and are therefore referred to as cyclical factors. The tax response to changes in the business cycle tends to be nonlinear. Tax receipts fall sharply during an economic slowdown and increase with elasticity greater than one (relative to GDP) when the level of activity begins to recover from the trough. The part of tax revenue that is beyond the linear effect (which is included in the structural factors discussed above) is defined as a cyclical effect. This revenue incudes taxes on domestic consumption, corporate taxes (derived from their profits), payroll taxes derived from the volume of the wage bill, and so forth. The group of exceptional factors also includes taxes due to anomalous fluctuations in the asset markets (real estate and financial assets), where the cycles do not necessarily match the general business cycles in the economy.
Two phenomena that characterized the Israeli economy in particular over the past two years were an increase in the export of high-tech services and the expansion of activity and increase in prices in the housing market. These two phenomena are not necessarily commensurate with the general business cycle in the economy, but they do have a cyclical nature, such that an increase in certain years may moderate, or even reverse, in the medium term. Wage increases and capital gains of employees in high-tech services began even before the crisis, and became even more prominent during the crisis, even compared with other advanced economies. Friedman (2017) found that wages in the high-tech industries tend to be more elastic than wages in other industries in response to changes in demand and in firms’ profitability. This is because of employment stability in the industry, which is due to the high level of human capital that is typical of its employees. For this reason, the growth in tax receipts from the high-tech industries observed in recent years may moderate in the coming years, and may even go through a downward correction. The pace of building starts and land marketing also accelerated in recent years, and this process may continue for a number of years. However, it is difficult to know what the horizon of this process is—when it will end, and whether it will be halted sharply or gradually subside. We therefore relate to these factors as exceptional factors.
One-off factors: These reflect one-off events that are exceptional in size, such as large transactions that carry a significant tax payment, the response of tax payers to legislative or regulatory measures, and tax assessments to large companies related to events that took place in previous years. Some of the revenue this year was apparently due to legislative changes that caused economic activity to be brought forward. For instance, the increase in taxes on investors in the real estate market in November encouraged them to bring real estate transactions forward to the months preceding the tax increase, meaning that amounts that would have been collected in respect of those transactions at later times were collected in 2021 instead. This component is unexpected, short-lived, and may even have a negative impact on revenue in the following years.
When calculating the increase between 2019 and the following years, we must also take into account the fact that in 2019, net income taxes were relatively low due to increased tax refunds of about NIS 2.5 billion. These were due to the decline in the corporate tax rate in previous years, and to temporary benefits issued in 2017 for withdrawing dividends (including to self-incorporated individuals), which encouraged people to bring tax payments for 2018 and 2019 forward to 2017.
The components’ contributions according to the Bank of Israel tax model
The tax model, which helps formulate the government revenue forecast as part of the Bank of Israel Research Department’s staff forecast, includes a number of activity variables and a number of financial variables that are intended to explain past tax receipts and forecast tax revenue for the medium term. The model includes quarterly GDP with a lag, GDP with a lag of five quarters, GDP growth with a lag of two quarters, the portion of the average wage that is not explained by GDP growth, the portion of consumer goods imports that is not explained by GDP growth, seasonality variables (a dummy variable for certain quarters), the Bank of Israel interest rate, the change in the general shares index on the Tel Aviv Stock Exchange, the same change with a lag, and the number of housing transactions with a lag of three quarters. An examination of the parameters estimated in the model up to 2019 compared with a re-estimation using actual data up to 2021 showed that there were no significant changes in the parameters, so that most of the changes in tax revenue in recent years can be attributed to the sharp changes in the variables themselves.
In order to understand the contribution of the macroeconomic variables to changes in tax revenue, we used two tax model projections. The first projection assumed that the development of the explanatory variables in the model in 2020 and 2021 was similar to their behavior between 2015 and 2019 (a routine period in which the economy operated in a full employment environment). The second projection was made using the actual development of the variables. The difference between the forecast contribution of each variable based on past activity and based on its actual contribution enables us to quantify the contribution of the macroeconomic changes to overall tax revenue.
The exercise, the results of which are presented in Table 1, emphasizes the high weight of the contraction of GDP in the reduction of tax revenue in 2020 (Column 2). It also emphasizes the effect of the strong recovery in 2021, which offset most of the negative contribution from the previous year (Column 3). Two additional variables that contributed much to the increase in revenue in 2021 are the sharp increase in consumer goods imports, beyond what is explained by the long-term correlation with GDP, and equity prices on the Tel Aviv Stock Exchange. In contrast with the large negative contribution of GDP in 2020, imports and equity prices had a negligible negative effect on revenue, while in 2021 they increased exceptionally. In addition to these two variables, the increase in the number of housing transactions explains the increasing tax revenue as far back as 2019. However, real estate tax (betterment and purchase) revenue in 2021 was particularly anomalous (about NIS 7 billion more than the increase in total taxes that is explained by the anomalous number of transactions). The comparison of the contribution of these variables to increased revenue compared to their contribution during the routine years, hints at anomalous profit-taking in the capital market, and the anomalous levels of consumer goods imports and housing transactions, which may decline in the coming years.
In conclusion, the main contributing factor to the NIS 67 billion increase in tax revenues in 2021 compared with 2019 was the “natural” growth of tax revenues by about NIS 34 billion over the period. During those two years, there were no significant legislative changes, so that most of the structural growth in tax revenues was due to that natural growth, which would have taken place anyway had the macroeconomic variables changed in accordance with past trends. The exceptional factors, represented by the deviation of activity variables from the pre-crisis trend, explain an additional NIS 29 billion, divided as follows: NIS 13 billion are due to the increase in consumer goods imports that are not explained by GDP; NIS 8 billion are due to the price increases in the capital market; and NIS 10 come from the housing market. GDP and the deviation of the average wage from its long-term correlation with GDP were offset between 2020 and 2021, and over all they made a negative contribution of about NIS 3 billion. One-off factors reduced revenue by about NIS 2 billion, and the remaining unclassified increase in revenue is estimated at about NIS 6 billion—mainly due to unexplained increase in 2020. Due to the significant weight of the exceptional components in the increase in revenue in 2021, it is necessary to be very cautious in relying on this revenue when planning toward the next budget. Based on the past, it is reasonable to assume that the tax burden as a percentage of GDP will return to a scale similar to the precrisis level.
 P. Dudine & J. T. Jalles. (2018). "How Buoyant is the Tax System? New Evidence from a Large Heterogeneous Panel". Journal of International Development, 30(6): 961–991.
 In other advanced economies as well, there are empirical attempts to distinguish between structural and transitory changes in macroeconomic activity, in order to determine the level of the structural deficit, in particular by assessing the extent of revenue growth derived from the cyclicality of asset prices (physical and financial). For instance, see Morris, Richard, and Ledger Schuknecht (2007), “Structural Balances and Revenue Windfalls—The Role of Asset Prices Revisited,” ECB Working Paper 737 (Frankfurt, European Central Bank).
 For more details, please see Chapter 1 in Bank of Israel Annual Report for 2021 (forthcoming).
 For more information, please see Chapter 9 in Bank of Israel Annual Report for 2021 (forthcoming).
 For more information in the relatively high elasticity of wages in the high-tech industries and the cyclicality that has been typical of these industries since the 1990s, see Yoav Friedman (2017), “Information Technology Industries: Employees, Wages, and Dealing with Shocks”, Economic Quarterly, 61(1): 145–178 (June 2017) (in Hebrew).
 A significant example of one-off tax revenue took place in 2017, when the government decided to temporarily lower the tax on dividends to owners of self-incorporated individuals. This policy encouraged the distribution of exceptional volumes of dividends, and tax revenues that totaled about NIS 11 billion. Due to these payments being brought forward, tax receipts from dividends were weaker, and tax rebates increased, in the following years . For more information, see Chapter 6: “The Public Sector and its Financing” in the Bank of Israel Annual Report for 2018.
 For details on the tax model, see A. Brender and G. Navon (2010), “Predicting Government Tax revenues and Analyzing Forecast Uncertainty”, Israel Economic Review, 7(2): 81–111. The parameters of the model were estimated using a sample from the years 1995–2019.
 Two possible factors for the increase in tax revenue from wages, which are not included in the model, are the concentration of the increase in the average wage in 2021 mostly on those with higher wages, on which the marginal tax is higher than the average in the economy; massive amounts of capital raised by high-tech companies, which were partly translated into the exercise of options by employees and tax payments by some company owners in respect of the issuances. Precise data on these two components are not available at this stage, but simulations show that they may amount to a few billion shekels.