Inflation experiences and financial decisions--论文代写范文精选

在每个调查中,我们构建人均债务和债券资产的数量,以及收入和资产净值。然后运行评估结果，提供进一步的细节数据。下面的essay代写范文进行详述。

Abstract 
So far our results show that differences in inflation experiences generate differences in beliefs about future inflation. To what extent do these differences in beliefs affect the economic decisions of households? Since differences in inflation expectations generate disagreement about real rates of return on assets and liabilities with nominally fixed rates, households with higher experience-based inflation expectations should be more inclined to borrow and less inclined to invest at nominally fixed rates than households with lower experience-based 25 inflation expectations.

To estimate the effect of learning-from-experience forecasts on financial decisions, we turn to a different data source, the Survey of Consumer Finances (SCF), which provides detailed information on households’ financial situation. We rely on the data set constructed by Malmendier and Nagel (2011), which comprises both the modern triennial SCF from 1983- 2007 and older versions of the SCF from 1960-1977. For comparability with our baseline estimation in Table 1, we aggregate the microdata again at the cohort-level.8 

In each survey wave, we construct per-capita numbers of debt and bond holdings (in September 2007 dollars), as well as income and net worth, for all birth-year cohorts. We then run the estimation on the resulting cohort panel. Appendix E provides further detail about the data set and the construction of our variables. Table 2 provides summary statistics for the key variables in our analysis. Households’ primary fixed-rate liability is mortgage debt, shown in column (i). Prior to 1983, the SCF often provides mortgage information only for households’ primary residence, not for other real estate owned by the household. To construct a measure that is consistent over time, we define fixed-rate liabilities as the sum of fixed-rate mortgage balances secured by the primary residence. 

On the asset side, we measure households’ holdings of long-term bonds, shown in We also tabulate separately the summary statistics for mortgages that were newly taken out or refinanced in the same year during which the survey took place. (The survey is carried out from June to September.) We use these alternative outcome variables when focusing on the flow rather than level of liabilities. We split these (re-)financing volumes into fixedrate and variable-rate (re-)financings, as shown in columns (iii) and (iv) of Table 2. 

The information whether mortgages have variable rates or fixed rates is only available starting in 1983, but variable-rate mortgages were largely non-existent in the U.S. prior to the 1980s (see Green and Wachter (2005)). Finally, columns (v) and (vii) show family income and net worth, which we use as control variables. In the years before 1983, the coverage of household assets in the SCF is not as comprehensive as from 1983 onwards. For the sake of comparability over time, our measure of net worth uses only categories of assets and liabilities that are available in all survey waves: financial assets, defined as stocks, bonds, and cash, including mutual funds and DC accounts, plus equity in the households’ primary residence. 

Table 3 presents the results of the estimations. In each column we regress the log of the respective cohort-level per-capita nominal position on the learning-from-experience inflation forecast, constructed using our point estimate of θ = 3.044 from Table 1. We control for the logs of income and net worth. All regressions include dummies for the survey year and for age. Column (i) shows the estimation results for the regression using the total size of households’ fixed-rate mortgage positions as the outcome variable. We find that it is positively related to the learning-from-experience inflation forecast, and the point estimate of the coef- ficient is more than four standard errors above zero. As predicted, households whose experiences lead them to expect higher inflation and, hence, lower real interest rates take on more fixed-rate liabilities. 

The magnitude of the effect is large: a one percentage point difference in the learning-from-experience forecast corresponds to a 0.35 change in the log of the fixed-rate mortgage balance, which is between a third and a quarter of a standard deviation of the dependent variable (see Table 2). This magnitude is comparable to the variation associated with a one-standard-deviation change in log income. In column (ii), we estimate the effect of inflation experiences on fixed-rate investment, i.e., the size of households’ nominal bond positions. 

Here, the sign of the coefficient is negative, indicating that households with higher learning-from-experience forecasts of inflation take smaller positions in long-term bonds. The size of the coefficient is on the same order of magnitude as the coefficient in column (i), but the estimate is not statistically different from zero. Taken together, the results in column (i) and (ii) show that households with higher learning-from-experience inflation forecasts tilt their exposure to liabilities rather than assets with nominally fixed rates. As shown in columns (iii) and (iv), we also obtain similar results when restricting the sample period to 1983-2007, when the SCF data is of higher quality. In columns (v) and (vi), we refine the analysis in two ways. （essay代写)

First, we focus on the extent to which households have recently taken out new mortgages or refinanced them, rather than The SCF sample includes 19 surveys during the period from 1960 to 2007, and 18 of those have information on holdings of long-term bonds. The data on borrowing and bond holdings is aggregated to per-capita numbers at the cohort level. Each cohort is assumed to recursively estimate an AR(1) model of inflation, with θ = 3.044, as in Table 1, column (i). We use the resulting learning-fromexperience forecast of inflation to explain log fixed-rate mortgage borrowing and log long-term bond holdings in OLS regressions. Log mortgage borrowing in columns (v) and (vi) comprises only loans taken out or refinanced in the year in which the survey was carried out. Standard errors reported in parentheses are clustered by time period.

The results in column (v) and (vi) are consistent with this prediction. We find that households with high learning-from-experience forecasts of inflation are significantly more likely to take out new fixed-rate mortgages and to re-finance at fixed rates. A one percentage point difference in the learning-from-experience forecast corresponds to a roughly 1.33 change in the log of the fixed-rate mortgage balance, which is more than a third of a standard deviation of the dependent variable according to Table 2. We also find that experience-based inflation expectations are negatively related to the volume of new variable-rate mortgages, but the estimated coefficient in column (vi) is not statistically significant, and the point estimate is small—a one percentage point difference in the learning-from-experience forecast corresponds only to about a ninth of the standard deviation of the dependent variable. 

Overall, the findings in this section confirm that learning from inflation experiences not only affects the expectations of individuals, but also helps understand patterns in the financial decision-making of households. These results also speak to the relevance of expectations data from household surveys (as opposed to surveys of professional forecasters) for real economic variables. For the decisions that we study here—asset allocation to long-term bonds and mortgage financing—we estimate a significant effect of individuals’ perceptions on aggregate outcomes. On the asset side, this might seem surprising because of the widespread delegation of portfolio management to professional fund managers. However, households control much of the asset allocation decisions in the economy when they allocate their wealth to mutual funds that are restricted to invest in certain asset classes, including the choice stocks versus bonds. On the liabilities side, the results might be less surprising as households are the main players as buyers and sellers in the residential real estate market and as borrowers in residential mortgage financing. In summary, the findings in Table 3 show that the learning- 30 from-experience model helps understand household behavior in these important markets.（essay代写)

51Due网站原创范文除特殊说明外一切图文著作权归51Due所有；未经51Due官方授权谢绝任何用途转载或刊发于媒体。如发生侵犯著作权现象，51Due保留一切法律追诉权。
更多essay代写范文欢迎访问我们主页 www.51due.com 当然有essay代写需求可以和我们24小时在线客服 QQ:800020041 联系交流。-X（essay代写)