The Effect Of Durable Goods On The Growth Globally

The present System of National Accounts (SNA93) treats durable consumption goods as

consumption goods rather than investment although rentals for owner occupied households is imputed into GDP. We argue that households de facto treat the purchase of durable goods as investments and thus, the treatment of durables as capital assets conceptually does not differ from the present treatment of owner occupied dwellings. This is not captured by the economic analysis based on current statistical conventions.

The purpose of this paper is to estimate the effect of durable goods and ICT on euro area economic growth and productivity change; when expenditure on consumer durables is recorded as capital investment. The capitalization of consumer durables impacts both the levels and growth rates of the capital stock, productivity and GDP. Our growth accounting computations demonstrated that the capital services of durables contributed one-tenth of economic growth and one-eight of labour productivity growth in 1995-2004. ICT's impacts were larger, i.e., one-fifth of GVA growth and one-sixth of labour productivity growth.

The purpose of this paper is to estimate the effect of durable goods and ICT on euro area (EA) GDP growth and productivity change from 1995 to 2004. In this exercise expenditure on consumer durables is recorded as capital investment. This impacts both the levels and growth rates of the capital stock, productivity and GDP. The advantage of this treatment is that it makes the treatment of consumer durables symmetric to that used in the Systems of National Accounts to account for owner occupied dwellings. As we also account for the effect of ICT the true proximate sources of growth are highlighted.

We argue that households de facto treat the purchase of durable goods as investments. However, this is not captured by the economic analysis based on current statistical conventions. The present System of National Accounts (SNA93) treats consumer durables as a part of private consumption, whereas Dale Jorgenson consistently treats consumer durables as capital inputs both on the output and the input sides. Charles Hulten recently defined investments as such expenditures that are made at the expense of current consumption in order to increase or maintain future consumption.

The results of this paper also show that the new treatment of consumer durables increases annual GVA growth by 0.08 percentage points and labour productivity growth by 0.07 percentage points as the new growth of gross value added (GVA) is two and labour productivity growth is 1.2 per cent. Furthermore, our growth accounting computations demonstrated that the capital services of durables contributed one-tenth of economic growth and one-eight of labour productivity growth. It was no surprise that ICT's impacts were larger, i.e., one-fifth of GVA growth and one-sixth of labour productivity growth.

The combined contribution of ICT and durable capital deepening is the most important component of EA labour productivity growth. The role of other capital deepening is nearly as big. Previously we thought that the deepening of other capital carried by far the largest contribution.

As the outcome of this paper is that the alternative treatment of durable goods as well ICT has a considerable effect on economic growth and productivity, it is not difficult to find a policy recommendation or justification for this paper. This paper emphasises that in fiscal as well as ECB.

The reclassification of durable goods has an effect on GDP. The reallocation of consumer durables to gross fixed capital formation (instead of private consumption) increases output (and possibly intermediate consumption), since investment from the output approach point of view provides a service flow to production. From the income approach point of view this treatment affects two components: operating surplus and consumption of fixed capital. Together these effects are by definition exactly the same size as the service flow effect of the output approach. From the expenditure point of view, durable goods should first be re-classified from private consumption to investment. Second, the value of the service flow has

to be classified as private consumption.

The capitalisation of durable goods has also been suggested to be considered during the currently ongoing SNA update. The proposal was rejected because “consumer durables are not regarded as assets in the system because the services they provide are not within the production boundary”. However, the Inter-Secretariat Working Group on National Accounts proposed to record capitalised consumer durable goods in satellite accounts. Moreover the group recommended showing consumer durable goods as a memorandum item in the balance sheet but not in the totals of non-financial assets.

There are five sources of labour productivity growth. The first one is durable goods' capital deepening, i.e., the share weighted increase of durable good capital services per hour worked. The second source is the share weighted deepening of ICT capital. The third source is the share weighted deepening of other capital. The fourth component is the improvement in labour quality, which is defined as the difference between the growth rates of labour services and hours worked, multiplied by labour’s income share. The fifth component is a general advance in multi-factor productivity, which increases labour productivity point for point.

The EA aggregate is a simple aggregation of available EA Member States (EA-MS).

However, the EU KLEMS database does not include data for all the EA-MS, i.e. Greece, Ireland, Cyprus, Luxembourg, Malta, Portugal and Slovenia do not have any data available in the database. Since Slovenia is a member of the EA only since 2007, and Cyprus and Malta since 2008, it is not necessary to include them in the analysis. Thus, Greece, Ireland and Luxembourg are the only countries for which no data is available and they represent approximately five percent of the EA-GDP in 2006. Therefore, levels in this paper are underestimated by approximately five percent.

The private consumption data is the so-called “Table 5 data” of the ESA 95 transmission programme. This data is available for almost all of the MS. The detail of the data is the two-digit level of the COICOP classification. As discussed later this data is broken down in more detail to estimate the share of durable goods. Unfortunately, more detailed data than 2-digit level data is not available from the international databases. The data is not PPP corrected and therefore, we had to perform the PPP correction ourselves. This has been explained in more

details in the following sub-section.

Stocks of consumer durables

Private consumption can be divided into services and goods that can be classified durable, semi-durable or non-durable. Owing to the lack of detailed expenditure data on durables, we used the same annual shares of consumer durables in each two-digit COICOP consumption group as in our previous work (see table 1) and multiplied these shares with the national two-digit current price consumption expenditure figures of the EA countries.

Estimation of output and value added

In this paper, consumer durables are treated in the same way as imputed rents in the national accounts. In principle, the logic of capitalising durable goods follows exactly the same logic as imputed rents. The SNA postulates that heads of households who own the dwellings that the households occupy are formally treated as owners of unincorporated enterprises that produce housing services consumed by those same households. As well-organised markets for rented housing exist in most countries, the output of own-account housing services can be valued using the prices of the same kinds of services sold on the market, in line with the

general valuation rules adopted for goods or services produced on one’s own account. In other words, the output of housing services produced by owner-occupiers is valued at the estimated rental that a tenant would pay for the same accommodation, taking into consideration factors such as location, neighbourhood amenities, and so forth, as well as the size and quality of the dwelling itself.

The weights of alternative rates of return for durable goods have been calculated from the annual Monetary Union Financial Accounts (MUFA). Three different categories of assets have been used in the calculation: currencies and deposits, shares, and debt securities (including mutual funds). The returns of the currencies and deposits were calculated by using one-month Euribor (Euro Interbank Offered Rate). The returns of shares were calculated by using the Dow Jones Euro STOXX price index, and finally, the returns of debt securities were calculated by using the three-year EA Government benchmark bond yield.

Treating consumer durables as investments has a surprisingly large impact on the level of EA gross value added. The ratio of the output of consumer durables to unrevised GVA (both at current prices) varies between 6.45 and 9.64 per cent annually. On average it is 8.03 per cent in the years 1995-2004 (table 3). The GVA impact is lessening towards the end of the period since the output of consumer durables only stayed level although nominal GVA increased by a quarter in the decade we are observing. The impact of consumer durable assets on the EA

capital stock cannot yet be estimated since the capital stocks underlying the capital service calculations have not been released in the EU KLEMS database.According to the growth accounting results published by the EU KLEMS consortium in November 2007 the EA gross value added (GVA) grew in volume terms on average by 1.92 per cent annually in the years 1995-2004 (table 4). This growth stemmed nine-tenths from the combined effect of the inputs and the rest was attributed to multi-factor productivity (MFP) . One third of economic growth came from labour services and almost sixty per cent from capital services (of which twenty percentage points was related to ICT capital services). Capitalising durables does not radically alter our general perception of the proximate sources of EA economic growth. The relative contributions of the inputs and the

residual remain similar. There are, however, important differences. Economic growth was actually faster than previously perceived (2.00 per cent annually and not. Furthermore, the capital services of durable goods were one-tenth of economic growth. This naturally implies that the contributions of the other inputs were lower.

Another way of looking at economic growth is to decompose it into the impacts of labour input and labour productivity (table 5). Hours worked increased in the observation period at the brisk rate of 0.79 per cent per annum. The new treatment of consumer durables boosted economic growth by 0.08 percentage points annually and labour productivity growth by 0.07 percentage points. Using equation 4 we found that of the new labour productivity growth estimate of 1.20 per cent annually as much as 0.15 percentage points, or one-eighth, was attributed to the share weighted increase of durable good capital services per hour worked by our calculations. One-sixth of labour productivity growth stemmed from ICT capital deepening. Again, the contributions of the other inputs turned out to be lower than earlier

thought. The residual remained unchanged.

Conclusions

The purpose of this paper was to estimate the effects of ICT and durable goods, when they are treated as investments, on EA GDP and productivity growth. The increasing use of technology and the breakthrough of home/entertainment technology in the past few decades emphasises the importance of this kind of analysis. Capitalising consumer durables has a surprisingly large impact on the level of EA economic growth. In relation to unrevised GVA the share is around 8 per cent on average in the years 1995-2004.

Various Factors Influencing the Growth of Durable Goods

The Role Of Superstars

Our calibration of the persistence of the superstar shock is guided by the Gini index for wealth we want to match. However, we cannot estimate this persistence from standard data sets. This is not the case for the process governing regular households’ shocks, which is very close to the idiosyncratic component of the earnings process estimated by Storesletten, Telmer, and Yaron 2004) using data from the PSID. What is the contribution of superstars in shaping the overall level

of wealth inequality? To answer this question, we eliminate the superstar shock and recalibrate all economies so that they still produce the same aggregates. The results are shown in Table 5. The Gini index for wealth falls to 0.645, whereas in the benchmark economy with superstars the Gini index for regular households is 0.75 (its counterpart in the data is 0.77, see Table 1). Importantly, wealth inequality is slightly lower in the benchmark economy than in the one-asset economy and the difference in the Gini indices is about the same magnitude as in the specification with superstars. Thus, adding the superstars not only helps us bring the overall Gini coefficient for wealth closer to that of the data, but it also helps us obtain the implied level of wealth inequality for households outside the top 1 percent of the earnings distribution. Eliminating superstars, however, does not change our main result: the existence of illiquid assets and credit frictions mitigate somewhat the effect of uninsurable idiosyncratic labor risk on wealth inequality resulting in slightly lower inequality.

The Role Of the Persistence Of the Earnings Process

Our benchmark model with illiquid assets and credit frictions delivers just slightly lower wealth inequality than the standard one-asset economy used in the literature. This result is obtained using an earnings process with very high persistence (consistent with the empirical evidence). To determine if our result is robust to this specification, we simulate our model economy using a different transition matrix for the earnings process holding the earnings shocks and the stationary distribution constant. In particular, we keep the probability of becoming a superstar unchanged but assume the probability of being one of the six regular types is the same for all types. The transition matrix is shown in Table 6. We recalibrate the relevant parameters so that aggregate statistics remain the same (the parameter values are in the notes to Table 7). We call this economy

the volatile benchmark economy. With lower persistence in the earnings process, wealth inequality is much lower than before (the Gini index is 0.635 in the volatile bechmark case). Houses are more equally distributed (the Gini coefficient is 0.256) and financial assets are less concentrated (the Gini coefficient is 0.863). Housing wealth as a fraction of total wealth decreases in all quintiles – since earnings shocks are not persistent, households accumulate proportionally more liquid assets. As with the persistent earnings

process, housing wealth as a fraction of total wealth decreases with wealth but the differences across quintiles are less extreme. In the one-asset economy, wealth is also less concentrated (the Gini index is 0.676). Importantly, inequality is still higher in the volatile one-asset economy than in the volatile benchmark economy. Moreover, the difference is now more pronounced. In order to understand why this is the case, it is useful to construct a measure of permanent earnings in our model. Since we abstract from aggregate uncertainty, for any household whose earnings shock in period t is e, we can write permanent earnings, e, as the sum of current and future earnings: The (normalized) permanent earnings shocks with the volatile process and the original process are: {1.00, 1.01, 1.01, 1.02, 1.04, 1.09, 7.38} and {1.00, 1.22, 1.49, 1.84, 2.32, 3.04, 13.07}, respectively. Regular households are much more similar with the more volatile earnings process, they save less, which leads to lower inequality. For the same reason, houses are more similar across households, which results in less differences in wealth composition across quintiles. Furthermore, with less savings, more households are likely to be affected by the frictions of our model, which explains the larger difference in inequality between the benchmark economy and the one-asset economy in this case. However, the difference in wealth inequality across models is still modest in magnitude.

This calibration allows us to illustrate further the predictions of our model regarding the distribution of houses. Table 8 presents key distributional statistics for homeowners in the data (first panel) and in the model with both persistent and volatile earnings (second panel and third panels respectively). With persistent earnings, the Gini index for earnings is lower in the model than in

the data (0.408 vs. 0.479 in the data). This implies lower levels of inequality for any dimension of wealth but, nevertheless, the model with persistent earnings captures the remarkable similarity of the distributions of earnings and houses observed in the data. This feature of the data is not specific to 1998 as shown in Table 9 (although houses are becoming slightly more concentrated than earnings in the recent years). In our model, houses cannot be more unequally distributed than earnings for homeowners because the return to owner occupied housing falls with the size of the house. Note that with volatile earnings, the Gini coefficient for houses is less than half the coef-ficient for earnings (0.179 vs. 0.449).14 What causes this difference? The distributions of earnings and houses are quite close with persistent earnings in our model because permanent earnings and current earnings are highly correlated and households acquire houses according to their permanent income.15 While permanent income still guides house purchases with volatile earnings, current earnings in this case are not highly correlated with permanent income and the distributions are not

alike. In summary, high persistence is necessary to obtain a Gini index for wealth close to the one observed in the data. Nevertheless, earnings persistence cannot be estimated directly using the Survey of Consumer Finances.16 Our analysis in this section suggests that the distribution of houses might be used to discriminate among earnings processes that differ in their persistence. That is, the distribution of houses gives us indirect evidence of the persistence of the earnings process. Changes in the down payment and the adjustment cost Over the last few decades, there has been a significant reduction in the down payment required by

financial institutions as well as a proliferation of home equity loans. In our model, a decrease in the parameter θ captures these financial changes (although we cannot disentangle one from the other). We analyze the effects of financial liberalization on aggregate ratios and on the wealth distribution by simulating our model economy for different values of the down payment requirement (keeping all other parameters constant). A decrease in the down payment requirement relaxes the borrowing constraint. Thus, fewer households are constrained and their purchases of houses increase. Therefore, inequality in houses should decrease (see Table 10). However, because a decrease in θ implies higher borrowing in the economy, financial assets become more concentrated and overall, wealth inequality worsens. In general, the observed effects tend to be small. This is because changes in the down payment affect mainly liquidity constrained households, who are concentrated at the bottom of the wealth distribution.

Since their asset holdings amount to a very small fraction of aggregate wealth, the effect of changing the down payment on total wealth is not large. For instance, the Gini of wealth with no down payment is 0.81, while with a 20 percent down payment it is 0.801. When the down payment is 100 percent, the Gini coefficient for wealth is substantially lower, 0.736. Table 10 demonstrates that the distribution of financial assets across quintiles is significantly more concentrated for lower down payments. The counterpart to this result is in Table 11, where we show that the portfolio of poor households becomes substantially more illiquid as down payments fall. Table 11 also indicates that with lower down payments, the housing stock increases, the capital stock decreases and the interest rate rises (from 3.473 for a 100 percent down payment to 3.996 for no down payment).

With a rental market, as down payments decrease, homeownership increases (results not tabulated for brevity). For example, with a down payment of 50 percent, homeownership is only 49 percent, while with a 5 percent down payment the rate is 83 percent. As a result, inequality in housing decreases considerably more than without a rental market (the corresponding Gini coefficients for housing for a 50 percent and a 5 percent down payment are 0.68 and 0.55 respectively). As before, because there is more borrowing, financial assets become more concentrated with lower down payments. The effect on overall inequality is even smaller in this case and can be non-monotonic in the down payment. For example, going from a 50 percent to a 20 percent down payment leads to less inequality (the Gini index for wealth goes from 0.8116 to 0.809), while going from a 20 percent to a 5 percent down payment increases inequality slightly (the Gini index increases from 0.809 to 0.8121). However, with or without a rental market, when down payments decrease the housing stock increases, the capital stock decreases and the interest rate rises (the equilibrium interest rate in the choice economy with a 50 percent down payment is 3.87 percent while the interest rate is 3.95 percent with a 5 percent down payment).

We also investigate the effect of changes in the degree of illiquidity of houses. In Table 10, we report aggregates with a higher adjustment cost, with no adjustment costs and for an economy with liquid houses and no down payments (ρ = 0, θ = 0). As before, all remaining parameters are kept at their benchmark values. For a given down payment, lowering ρ makes the durable more attractive for households of all wealth levels, which leads to an important increase in the housing stock. Thus, the change in aggregate wealth composition seems more dramatic than the effect of lowering down payments. Since the housing stock increases, the capital stock decreases and the interest rate sharply rises (from 3.71 to 4.25 when going from 10 percent to 0 percent in transaction costs). However, the effect on the wealth distribution is negligible. In terms of wealth composition (see Table 11), decreasing the degree of illiquidity increases the housing wealth to total wealth ratio substantially for the lower quintiles.

Final Comments

In this paper, we explicitly model the existence of illiquid houses that serve as collateral for loans in the context of a heterogenous agents model with uninsurable earnings risk. Our goal is to asses whether the availability of houses mitigates the effect of uninsurable idiosyncratic labor risk on wealth inequality. In our model, households can save in the form of a liquid asset and in the form

of illiquid houses that can be purchased on credit (minus a down payment). Our economy delivers slightly less wealth inequality than the one-asset economy analyzed in the literature. We show that the standard model is indeed equivalent to an economy with liquid houses, a zero down payment and a perfect rental market. The difference in terms of wealth inequality between both economies is small because the frictions of our model (required down payments and adjustment costs) mainly affect the poor who only account for a small fraction of aggregate wealth. Importantly, our two-asset economy has one main advantage over the standard one-asset framework:it allows us to study wealth composition issues. Our model is able to reproduce all main patterns of the U.S. distribution: (1) wealth is more concentrated than earnings, (2) financial assets are more concentrated than wealth, (3) households’ portfolios become more liquid as wealth increases, and (4) the distribution of houses and earnings for homeowners are very similar. A fairly persistent earnings process is necessary to obtain the latter result. We also show that the easing of collateral credit has an important effect on the portfolios of the relatively poor.In this study, we abstract from some important issues that remain topics for future research. The most obvious and potentially important one, is the omission of life-cycle effects. In the data,a household’s portfolio composition varies with age and it would be interesting to analyze whether or not the model can account for the life-cycle patterns of wealth holding and wealth composition. Also, some of the aggregate effects discussed in this paper may be amplified when including life cycle considerations. An further extension could deal with the interaction between collateral credit and earnings ability. For example, access to collateral credit could increase the probability of becoming a superstar.