Does Switching from NASDAQ to the NYSE Affect Investment-Cash Flow Sensitivity

(1)

Does switching from NASDAQ to the NYSE affect investment

–cash ﬂow sensitivity?

Chau-Chen Yang

a

_{, H. Kent Baker}

b,

⁎

,1

_{, Li-Chuan Chou}

c

_{, Bo-Wei Lu}

a

National Taiwan University, College of Management, Department of Finance, Roosevelt Road, Taipei 106, Taiwan, ROC

b

American University, Kogod School of Business, Department of Finance and Real Estate, 4400 Massachusetts Avenue, NW, Washington, DC 20016, United States

c_{National Taipei College of Business, Department of International Business, Chi-Nan Road, Taipei 100, Taiwan, ROC}

A B S T R A C T A R T I C L E I N F O Article history: Accepted 23 May 2008 Available online xxxx JEL classiﬁcation: G10 G14 Keywords: Cashﬂow sensitivity Exchange switching

We examine whether afirm's sensitivity of investments to cash flow changes when it switches from the NASDAQ to the NYSE over the period 1992–2002. Contrary views exist on the effect of listing switches on investment sensitivity to cashflow. Investment–cash flow sensitivity is a proxy for the degree of uncertainty of using internal funds tofinance a firm's investments. We use a least square dummy variable model to analyze panel data before and after switching to determine the impact of switching. Based on pooled data, our evidence is consistent with the view that NASDAQ-to-NYSE switchers have significantly lower investment–cash flow sensitivity, which means that firms rely less heavily on internal financing after switching andfind accessing external financing easier. Thus, firms may benefit from switching in terms of a lower cost of external capital due to such factors as increasing visibility, liquidity, and reputation.

1. Introduction

Each year some managers decide to change the marketplace for trading their_{firms' securities. Although their motives for switching} differ, a common reason is management's desire to improve afirm's visibility.Baker and Edelman (1992a,b)andBaker et al. (1999b)find a positive relation between listing switches and visibility. Improved visibility suggests greaterflow and accessibility of information about a firm, which in turn may reduce information asymmetry and the uncertainty about afirm's prospects. Barry and Brown (1986) note that reducing information asymmetry and risk may enablefirms to lower their cost of external capital and to reduce their financing constraints. Accordingly, whenfirms can easily acquire less expensive external capital by switching to a larger exchange, their investment sensitivity to cashflow should decrease.

An alternative view is that the investment–cash flow sensitivity of switchingfirms should increase.Dharan and Ikenberry (1995)find that firms moving the trading location of their stock to the American Stock Exchange (AMEX) or New York Stock Exchange (NYSE) generally experience poor stock returns after switching. The negative impact of switching may indicate a higher cost of external equity and greater investment–cash flow sensitivity. In addition, this poor post-listing performance may relate to managers timing the application for listing. Due to the stringent requirements for listing on the AMEX or NYSE,

managers of smallﬁrms may choose to apply for listing before a decline in performance.

In this study, we use a cashflow model to examine whether a signi_{ficant relation exists between listing switches and investment–cash} flow sensitivity for switching firms. Given the differing views on how switching may affect investment–cash flow sensitivity, we do not predict the direction of the impact. We focus our examination on a sample offirms moving from NASDAQ to the NYSE during the period 1992–2002.

We use a least square dummy variable model to analyze panel data before and after switching to determine the impact of switching. Based on the pooled data, investment sensitivity to investment opportunities does not appear to differ between before and after switching. Our results show, however, that NASDAQ-to-NYSE switch-ers experience significantly lower investment–cash flow sensitivity. An implication of thisfinding is that switching firms benefit from a lower cost of external capital, which may be due to such factors as improved visibility, liquidity, and reputation.

Our study contributes to the literature by examining the theoretical and empirical aspects of switching trading locations. There is a large body of evidence on the issue of investment–cash flow sensitivity. Similarly, several studies explore the microstructure issues associated with switching from one equity market to another. To our knowledge, no previous study addresses how these two areas may be related. Our exploratory study seeks tofill this gap. Our evidence should be of interest to researchers studying market microstructure and to managers contemplating switching the trading location of theirfirm's securities.

The remainder of our paper has the following organization. In Section 2, we provide an overview of the literature on listing changes and investment–cash ﬂow sensitivity. We describe our cash ﬂow

Journal of Business Research xxx (2008) xxx-xxx

JBR-06617; No of Pages 6

⁎ Corresponding author.

E-mail addresses:yang@management.ntu.edu.tw(C.-C. Yang),

kbaker@american.edu(H.K. Baker),lcchou@mail.ntcb.edu.tw(L.-C. Chou).

1_{Tel.: +1 202 885 1949; fax: +1 202 885 1946.}

Contents lists available atScienceDirect

Journal of Business Research

ARTICLE IN PRESS

(2)

model in Section 3. In Section 4, we discuss our data and provide descriptive statistics about the marketplaces and switching compa-nies. We present our empirical results in Section 5. In Section 6, we summarize our ﬁndings and conclude the paper by suggesting directions for further work.

2. Related literature 2.1. Listing changes

Managers face the decision of whether to switch marketplaces for trading theirfirms' securities. For the most part, managers of NASDAQ or AMEX firms may consider switching their listing to the NYSE. Researchers have studied the reasons for listing switches and how switches affect visibility, liquidity, and stock returns. The decision to switch often involves a complex trade-off between costs and bene_fits. Somefirms move from a smaller market to a larger one to achieve visibility gains. For example, news on a listing may attract attention to the company in the investment community.Merton (1987)contends that the publicity associated with listing reaches some investors who were previously unaware of the security. In addition,Barry and Brown

(1986) note that listing may temporarily increase information

availability due to enhanced interest, which, in turn, may reduce the uncertainty about aﬁrm's prospects and raise the stock's price.

Baker et al. (1999a)use several proxies to test whether changes in

listing from the OTC to the NYSE affect afirm's visibility. They find that increased visibility in afirm is primarily due to changes in market capitalization not due to the listing itself.Baker et al. (1999b)use similar methodology to examine a sample offirms switching from the AMEX to the NYSE. They find a positive and significant relation between switching and visibility gains.

Baker et al. (2002)show that_{ﬁrms listing their stocks on the NYSE or}

the London Stock Exchange (LSE) have signiﬁcant increases in visibility in the form of increased analyst coverage and media attention. Analyst coverage and media attention are their proxies for visibility. Analyst coverage after listing is related to a decrease in the cost of equity capital in a way similar to Merton's (1987)investor recognition hypothesis. This effect is stronger for NYSE than LSEﬁrms, which may compensate for the higher costs associated with NYSE listing.

Several studies examine the relation between listing switches and liquidity. For example,Baker and Edelman (1992a)ﬁnd evidence that moving from NASDAQ to NASDAQ NMS (National Market System) tightens spreads and increases volume, which leads to enhancing a ﬁrm's liquidity by lowering transaction costs. Baker and Edelman

(1992b)indicate that low versus high transaction volume AMEX stocks

have more to gain in terms of visibility and market interest by switching to the NYSE. The temporary increase in attention by institutional investors may increase informationﬂows and may reduce uncertainties about stock performance and risk. Liquidity may increase due to investor interest in the stock.

Other studies examine the market reaction to listing. For example,

Grammatikos and Papaioannou (1986)andEdelman and Baker (1990)

report thatﬁrms with low prelisting liquidity experience the most favorable market response to initial listing. Baker and Edelman

(1992b) ﬁnd that cumulative abnormal returns (CARs) rise during

the prelisting period for the AMEXfirms switching to the NYSE but begin to decline shortly after the approval date for the full sample. These price effects may be associated with the bene_{fits of switching.} Another explanation is thatfirms may be taking other actions that generate good news before listing and bad news after listing. After splitting the full sample into low-volume and high-volume sub-samples,Baker and Edelman (1992b)find that the market reacts more favorably for the low-volume group. This favorable reaction may reflect not only greater visibility and market interest but also temporary increases in both informationflows and expectation of liquidity gains as well as reductions in systematic risk.

2.2. Investment–cash ﬂow sensitivity

Investment_{–cash flow sensitivity is a proxy for the degree of} un-certainty of using internal funds tofinance a firm's investments. In an imperfect market with information asymmetry, the cost of external capital exceeds that of internal funds. Without enough internal funds, afirm may have to resort to higher-cost external funds, which may lead to rejecting otherwise profitable investments. Hence, firms with favorable investment opportunities may want to rely on more internal sources of funds or try to reduce their cost of external funds.

The literature documenting the sensitivity ofﬁrms' investment to ﬂuctuations in their internal funds is substantial and has grown in recent years.Hubbard (1998)provides an extensive literature review and studies byAlmeida and Campello (2007)as well asDasgupta and

Sengupta (2007)present recent extensions on the topic. Beginning

withFazzari et al. (1988), these studies examine_{ﬁrms based on a}

priori measures offinancing constraints such as dividend payout, size, age, or debt ratings and compare the investment–cash flow sensitivity. For example, Fazzari et al. analyze differences in investment behavior by firms classified according to earnings retention. They find that investment offirms exhausting all of their internal financing is much more sensitive tofluctuations in cash flow than that of mature, high-dividendfirms. They attribute these results to a financing hierarchy in which internal funds have a cost advantage over new equity and debt. Subsequent studies byHoshi et al. (1991), Oliner and Rudenbusch

(1992), Whited (1992), Schaller (1993), Gilchrist and Himmelberg

(1995)and others provide support for Fazzari et al.

Early evidence suggests that investment decisions of more finan-cially-constrainedfirms are more sensitive to firm liquidity than those of less constrainedfirms. Other researchers have challenged the generality of this conclusion. For example,Kaplan and Zingales (1997)show that investment outlays of the least constrained_{firms are the most sensitive} to internal cashflow. Others obtain results consistent with those of Kaplan and Zinglaes includingCleary (1999),Kadapakkam et al. (1998),

andKaplan and Zingales (2002).

Allayannis and Mozumdar (2004)examine the notion that when

firms incur cash losses, investment cannot respond to cash flow. They find that while negative cash flow observations explainCleary's (1999) results, theKaplan and Zingales (1997)results are driven more by a few influential observations in a small sample. Allayannis and Mozumdar alsofind a decline in investment–cash flow sensitivity over 1977–1996, especially for the most constrained firms.

Moyen (2004) shows that the conﬂicting regression results

obtained by Fazzari et al. (1988) and Kaplan and Zingales (1997) depend on the criterion used to identify whether a_{firm experiences} financing constraints. She constructs an unconstrained model, in whichfirms have perfect access to external financial markets, and a constrained model, in which firms have no access. Using low dividends to identifyfirms with financing constrains leads to Fazzari et al.'s results that low-dividendfirm's investment is more sensitive to cash flow than high-dividend firms' investment. Using the con-strained model to identifyfirms with financing constraints supports the results of Kaplan and Zingales that constrainedfirms' investment is less sensitive to cashflow than unconstrained firms' investment. Thus, the results show that cashflow sensitivity of firms described by the constrained model is lower than that described by the uncon-strained model.

Froot et al. (1993)develop a model for analyzing corporate risk

management policies in which external funds are more costly than internal funds. Hedging helps ensure that ﬁrms have sufﬁcient internal funds for investment spending. Following their model,

Allayannis and Mozumdar (2000)andDeshmukh and Vogt (2005)

examine the difference in investment–cash ﬂow sensitivity between hedgers and non-hedgers. The differences between these two studies lie in the sample used and the target of hedging. Allayannis and Mozumdar use a sample of ﬁrms based on the S&P 500, while

2 C.-C. Yang et al. / Journal of Business Research xxx (2008) xxx-xxx

(3)

Deshmukh and Vogt use a sample of manufacturingfirms. The former focuses on currency hedging, while the latter examines both currency and interest rate hedging. Deshmukh and Vogt_{find that investment} spending is less sensitive to cashflow for hedgers than for non-hedgers. They alsofind that investment spending is less sensitive to cash_{flow among hedgers when the extent of hedging is higher. The} results of both studies are consistent with the implications of the model by Froot et al.

3. Cashﬂow models

Given the conflicting views on the direction and significance of the relation between listing switches and the investment–cash flow sensitivity of switching firms, the impact of listing switches on investment–cash flow sensitivity remains an empirical issue.

Follow-ingFazzari et al. (1988)andDeshmukh and Vogt (2005)among others,

we use a cashﬂow model to examine this relation. Their original cash ﬂow model follows:

Ii;t Ki;t−1¼ β0þ β1Qi;t−1þ β2 CF K i;t−1þei;t ð1Þ

where I is capital expenditures in the cashflow statement; K is the net fixed assets in the balance sheet; Q is Tobin's Q, which is the market value of total assets over the book value of total assets, and the market value of total assets is equal to market value of total equity plus the book value of total debt; CF is the cashflow from operating activities in the cashflow statement; ε is the error term; and i and t are the specific firm and period, respectively.

In this original cash ﬂow model, β1 measures how ﬁrms'

investment decisions react to their investment opportunities (Q). If β1is positive (negative),ﬁrms increase (decrease) their investments

when they expect better investment opportunities.β2measuresﬁrms'

investment sensitivity to their cashﬂows.

Our cashflow model uses the slope dummy approach to examine the difference of the regression coefficients between the before- and after-switching regression: Ii;t Ki;t−1¼ β0þ β1Qi;t−1þ β2DQi;t−1þ β3 CFi;t Ki;t−1 þ β4D CFi;t Ki;t−1 þ ei;t ð2Þ where I, K, Q, CF, andε are as defined before and D is the slope dummy variable (D = 0 and D = 1 denote the observations for NASDAQ firms before and after switching to the NYSE, respectively).

In our model, the slope dummy variable is 0 and 1 for the sample of observations before and after switching, respectively. β1 and β3

measure the regression coefﬁcients of investment on the Q and CF variables, respectively, for the sample before switching. _β2 and β4

measure the difference in regression coefﬁcients of investment on the Q and CF variables between the sample before and after switching. If β1is positive (negative), it indicates that positive (negative)

invest-ment–Q sensitivity exists before switching. We expect the sign of β1to

be positive, sinceﬁrms tend to increase their investments when they expect good opportunities (Q). Ifβ3is positive (negative), it indicates

that a positive (negative) investment–cash ﬂow sensitivity exists before switching. Empirically, the sign of β3 is ambiguous. Less

(greater) investment sensitivity of cashﬂow indicates a lower (higher) cost of external funds, as indicated by Allayannis and Mozumdar

(2000)andDeshmukh and Vogt (2005).

The signs of β2 and β4 are also an empirical issue. To our

knowledge, no previous literature exists about whether_{ﬁrms have a} greater investment–Q or investment–cash ﬂow sensitivity after switching. We conduct panel regressions for each year and pooled years, using four years of data (t− 4, t− 3, t− 2, t− 1and t+ 1, t+ 2, t+ 3, t+ 4).

For comparison purposes, we also use two (t− 2, t− 1and t+ 1, t+ 2) and

three years of data (t− 3, t− 2, t− 1and t+ 1, t+ 2, t+ 3).

4. Data and methodology

4.1. Overview of NASDAQ and the NYSE

In addition to dissimilarities in visibility, liquidity, and reputation, NASDAQ and the NYSE differ in other respects such as size, market microstructure, and risk–return characteristics. NASDAQ is the largest U.S. electronic stock market with about 3200 companies and, on average, trades more shares per day than any other U.S. market. The NYSE is the largest stock exchange in the world by dollar volume and, with about 2800 listed companies, has the second most securities of all stock exchanges. As of December 31, 2007, the combined capitalization of all NYSE-listed companies was about $27.1 trillion, which is much greater than NASDAQ. Another distinguishing characteristic is that NASDAQ is a dealer market, while the NYSE is primarily an auction market. Compared to the NASDAQ composite index, the NYSE composite index has a lower daily mean return and volatility as measured by the standard deviation of daily mean returns. 4.2. Data sources and sample

We identify NASDAQ-to-NYSE switches between 1992 and 2005 using the NYSE Factbook and use COMPUSTAT to obtainfinancial data between 1988 and 2006.Table 1indicates the number of NASDAQ-to-NYSE switches during the period 1992–2005. Of the 573 listing switches, we exclude 149firms not in the database (missing firms), 105financial firms (banking, insurance and security firms), and 15 firms with missing data on variables used for testing. Thus, we obtain an initial sample of 304firms between 1992 and 2005. Financial data on missing firms are unavailable due largely to mergers and acquisitions. We dropfinancial firms from our sample because they differ substantially from industrial _{firms in terms of fixed asset} investments and financing. Our sample excludes IPOs, and bank-ruptcies play a minor role involving the missing data. Our testing sample of 304 consists of 200firms with December year ends and 104 firms with non-December fiscal year ends.

The number of switches trends upward between 1992 and 1997, which coincides with a period of economic growth in which some firms could more easily meet NYSE listing requirements. The number of listing switches decreases in 1998 and 1999, a super-bullish period in which NASDAQfirms generally experienced outstanding stock price performance and had ready access to financing. Not surprisingly, fewer NASDAQfirms transferred to the NYSE during this period of a stock market bubble. The relatively low number of switches during

Table 1

Companies switching from NASDAQ to the NYSE between 1992 and 2005 Year Number switching Firms not in the database Financial firms Firms with missing data Testing sample size 1992 45 19 7 1 18 1993 45 13 11 1 20 1994 41 19 5 0 17 1995 61 19 12 1 29 1996 84 26 20 3 35 1997 92 28 10 4 50 1998 66 12 12 2 40 1999 18 5 5 0 8 2000 24 5 1 0 18 2001 27 1 7 1 18 2002 36 1 8 1 26 2003 20 1 4 1 14 2004 9 0 3 0 6 2005 5 0 0 0 5 Total 573 149 105 15 304 This table reports the distribution offirms switching from NASDAQ to the NYSE between 1992 and 2005. The test sample excludesfirms not in the database, financial firms, and firms with missing data. The data source is the NYSE Factbook 2007 and COMPUSTAT North America.

(4)

2000 and 2001 coincides with the economic downturn. During such a period, NASDAQﬁrms were less motivated to switch to the NYSE and were less likely to meet NYSE listing requirements. In recent years, both NASDAQ and NYSE have engaged in actively promoting the beneﬁts of their respective markets. These factors may partially explain the decline in listing switches between 2002 and 2005.

Table 2 provides summary statistics of our sample companies

before and after switching. We calculate averages for the investment-to-capital ratio, cashflow-to-capital ratio, and Tobin's Q ratio for the pooled years in Panel A. We use capital to stand for netfixed assets in the balance sheet and define the ratios calculated here in Section 3. Panel B presents the results of a paired t-test and a Wilcoxon matched-pairs signed-rank test for pooled to investigate whether these three ratios differ significantly before and after switching data for 1992–2002 using four years of data (t− 4, t− 3, t− 2, t− 1and t+ 1, t+ 2,

t+ 3, t+ 4), 1992–2003 using three years of data (t− 3, t− 2, t− 1and t+ 1, t+ 2,

t+ 3), and 1992–2004 using two years of data (t− 2, t− 1, and t+ 1, t+ 2).

As Panel A ofTable 2shows, the average investment-to-capital ratio using pooled data decreases from for each period examined. For the period 1992–2002 using four years of data before and after switching, the average investment-to-capital ratio decreases from 0.433 to 0.331. This decrease suggests a reduction in investment risk, which could result in lowerﬁnancial constraints. As Panel B ofTable 2indicates, both the paired t-test and the Wilcoxon test are statistically signiﬁcant at the 0.01 level. Panels C and D show similar results using both three years and two years of data before and after switching, respectively.

Panel A ofTable 2also shows that the average cashﬂow-to-capital ratio decreases from 0.786 to 0.763 over the period 1992–2002 using four years of pre- and post-switching data. This result suggests that ﬁrms have less internal funds and fewer investment opportunities

after switching. That is, thesefirms do not appear to achieve gains by switching to a larger exchange because the average cash flow-to-capital ratio decreases. Having less internal funds may increase financial constraints because of reduced solvency. Panel B ofTable 2 presents the results of the paired t and Wilcoxon tests, both of which are statistically signi_{ficant at the 0.01 level. As Panels C and D show,} the statistical results generally become weaker when using three and two year data before and after switching.

Panel A ofTable 2indicates that the average Q ratio decreases after switching. For example, using four year data for 1992–2002, the average Q ratio decreases from 2.478 before switching to 1.945 after switching. The lower average Q ratio after switching may reflect reduced stock performance, which, in turn, may lead to higher financial constraints. As Panels B, C, and D show, none of the paired t-tests or Wilcoxon t-tests are statistically significant at normal levels for the average Q ratio.

For testing purposes, we use panel regressions with two to four years of data before and after listing. We focus on the results using four years of data before and after switching because the investment_– cashﬂow sensitivity estimates are likely to be more reliable using longer samples with more observations. To estimate our model for 1992–2002 using four years of data (t− 4, t− 3, t− 2, t− 1and t+ 1, t+ 2, t+ 3,

t+ 4), we use 279 companies and 2232 observations (8 × 279). For

example, for NASDAQ-to-NYSE switchers in 2002, we include data between 1998–2001 (before listing) and 2003–2006 (after listing). Estimating our cashﬂow model for 1992–2003 using three years of data (t− 3, t− 2, t− 1and t+ 1, t+ 2,t+ 3) involves 293 companies and 1758

observations (6 × 293). When applying two years of data (t− 2, t− 1, and

t+ 1, t+ 2) to estimate our model for 1992–2004, we have 299 companies

and 1196 observations (4 × 299) for our analysis. 5. Empirical evidence

InTable 3, we present the empirical results of the question of

whetherﬁrms increase their investments when they expect better investment opportunities as well as whether switching from NASDAQ to the NYSE in years affects investment–cash ﬂow sensitivity. Panel A contains the regression results for individual years using four-year panel regressions. Panels B, C, and D report the results of the pooled regressions using four, three, and two years of data, respectively. In interpreting the results ofTable 3, we believe the pooled results and four-year panel regressions have greater reliability than the single year results because of the larger sample size.

5.1. Investment decisions and opportunities

In the regression model,β1measures howﬁrms' investment decisions

react to their investment opportunities (Q) before switching. Ifβ1 is

positive (negative),ﬁrms increase (decrease) their investments when they expect better investment opportunities. β2 shows how these ﬁrms'

investment decisions when facing investment opportunities may differ before and after switching. If the coefﬁcient (β2) of the investment dummy

DQ is positive (negative),ﬁrms have more (less) investment sensitivity to investment opportunities after switching than before switching.

As Panel A ofTable 3shows,β1for Q is positive in 10 of the 11 years

and signiﬁcantly positive in 6 of the 11 years (1992, 1996, 1997, 2000, 2001, and 2002).β1for Q is signiﬁcantly negative in 1994. Since β1is

generally positive, this suggests that _{ﬁrms typically increase their} investments when they expect better investment opportunities. Panel B indicates that β1 for the four-year pooled data is positive and

signiﬁcant at the 0.01 level. As Panels C and D show, similar results occur when using three-year and two-year pooled data in thatβ1for

Q is positive and statistically signiﬁcant at normal levels. These results support our expectation and provide evidence that ﬁrms tend to increase their investments when they expect better investment opportunities before switching.

Table 2

Summary statistics and tests of companies switching from NASDAQ to the NYSE partitioned by before-switching (D = 0) and after-switching (D = 1)

Panel A. Pooled data

Switching years 1992–2002 1992–2003 1992–2004 Dummy D = 0 D = 1 D = 0 D = 1 D = 1 D = 1 Number ofﬁrm years 1116 1116 879 879 598 598 Average investment-to-capital ratio 0.433 0.331 0.534 0.334 0.476 0.361 Average cash

ﬂow-to-capital ratio

0.786 0.763 0.689 0.760 0.763 0.606 Average Q ratio 2.478 1.945 2.268 1.966 2.191 2.008 Panel B. Tests using pooled data

Tests using pooled data

Average investment-to-capital ratio

Average cash ﬂow-to-capital ratio

Average Q ratio Panel B. 1992–2002 four years before and after switching (t−4, t−3, t−2, t−1and t+ 1, t+ 2, t+3, t+4)

Paired t-test (t-value) 4.248⁎⁎ 3.602⁎⁎ 0.248 Wilcoxon test (z-value) 9.761⁎⁎ 3.440⁎⁎ 0.416

Panel C. 1992–2003 three years before and after switching (t−3, t−2, t− 1and t+ 1, t+ 2, t+ 3)

Paired t-test (t-value) 2.633⁎⁎ 3.564⁎⁎ 0.330 Wilcoxon test (z-value) 7.635⁎⁎ 1.995⁎ 0.620

Panel D. 1992–2004 two years before and after switching (t− 2, t− 1and t+ 1, t+ 2)

Paired t-test (t-value) 2.890⁎⁎ 2.042⁎ 0.587 Wilcoxon test (z-value) 4.778⁎⁎ 0.144 0.787

⁎, ⁎⁎ Signiﬁcant at the 0.05 and 0.01 levels, respectively.

This table reports summary statistics of the sample companies before and after switching. Panel A provides the average investment-to-capital ratio, average cashflow-to-capital ratio, and average Q ratio for the pooled years. Panel B presents the t-values for the paired t-test and the z-values for the Wilcoxon matched-pairs signed-rank test for pooled data to test whether the average investment-to-capital ratio, average cashflow-to-capital ratio, and average Q ratio differ significantly between before-switching (D=0) and after-switching (D=1).

(5)

Our evidence is generally consistent with thefindings in previous studies such asFazzari et al. (1988)andKaplan and Zingales (1997). In their study on the effects of Q on investment, Fazzari et al.find all positive regression coefficients for each of the three dividend-retention classes for the period 1970–1979 and 1970–1984, and one negative and two positives for the period 1970–1975. Kaplan and Zingales also report all positive Q regression coefficients for all six financially-constrained classes for the period 1970–1977 and 4 of the 6 positive Q regression coefficients for the period 1978–1984.

β2 shows how these ﬁrms' investment decisions, when facing

investment opportunities, may differ before and after switching. For the four-year regressions, Panel A of Table 3 shows that β2 is

significantly negative in 3 of the 11 years (1999, 2001 and 2002) and significantly positive in one year (1994). As Panel B reveals for the pooled data,β2is negative but not statistically significant at normal

levels. As Panels C and D show,β2using pooled data for three and two

years before and after listing, respectively, is also negative but not statistically signiﬁcant. Thus, based on the pooled data, ﬁrms do not appear to have more or less investment sensitivity to investment opportunities after switching than before switching.

5.2. Investment–cash ﬂow sensitivity

β3measuresﬁrms' sensitivity of investment to cash ﬂow before

switching. As β3 increases, investment–cash ﬂow sensitivity also

increases._β4 measures the difference in regression coefﬁcients of

investment on cashﬂow variables between the sample before and after switching.

Fazzari et al. (1988)indicate that high (low) dividend payoutﬁrms

have low (high) sensitivity of investment to cashflow. The significance of the cashflow regression coefficient is evidence of the degree of financing constraints facing firms. Allayannis and Mozumdar (2000) and

Deshmukh and Vogt (2005) use this coefﬁcient to indicate that as

ﬁrms' volatility of cash ﬂow increases (decreases), their sensitivity of

investment to cashflow will be higher (lower). Both studies indicate that hedging against interest or exchange rate risk lowers the volatility of firms' cash flows and hence reduces the firms' cost of external capital.

For the four-year regressions in Panel A ofTable 3,β3is signiﬁcantly

positive in 7 of 11 years (1992, 1994, 1995, 1996, 1998, 2000, and 2002), but never signiﬁcantly negative during the period 1992–2002. As Panel B shows,β3for the pooled data is signiﬁcantly positive at the 0.01 level. As

Panels C and D show, the results for the pooled data using three-year and two-year regressions, respectively, are also positive and statistically signiﬁcant at normal levels. A positive β3 indicates that positive

investment–cash ﬂow sensitivity exists before switching.

As previously discussed, conflicting views exist on whether a firm's sensitivity of investment to cash flow is higher or lower after switching. If the coefficient (β4) of the cashflow dummy D(CF/Ki,t− 1)

is negative, this indicates that investment–cash flow sensitivity decreases after switching. A negative coefficient means that firms rely less heavily on internalfinancing after switching, which suggests they may find accessing external financing easier. We then have evidence thatfirms may benefit from switching in terms of a lower cost of external capital due to such factors as increasing visibility, liquidity, and reputation. However, ifβ4of the cashflow dummy D(CF/

Ki,t− 1) is positive, this indicates that investment–cash ﬂow sensitivity

increases after switching, whichDharan and Ikenberry (1995)indicate is evidence thatﬁrms may time the switching.

As shown in Panel A ofTable 3, the four-year regressions for the individual years reveal mixed results involving investment_{–cash ﬂow} sensitivity before and after switching. That is, over the period 1992– 2002, the sign ofβ4is signiﬁcantly positive in 2 of the 11 years (1997

and 2002), and statistically negative in 2 of the 11 years (1994 and 2000). Using the pooled data regression with four-year data to examine the impact of the switch,β4is signiﬁcantly negative at the

0.01 level. A negative coefﬁcient on D(CF/Ki,t− 1) may indicate that

ﬁrms have less investment–cash ﬂow sensitivity due to a lower cost of external capital after switching. This result supports the view that

Table 3

Tests on whether switching from NASDAQ to the NYSE affects investment–cash ﬂow sensitivity

Year Intercept Q DQ (CF/K) D(CF/K) Adj. R2 _{Number of observations}

Total D = 0 D = 1 Panel A. Individual years

1992 0.193 (2.57)⁎ 0.150 (3.83)⁎⁎ −0.042 (−1.09) 0.056 (1.77)⁎ −0.06 (−1.17) 0.117 144 72 72 1993 0.364 (5.74)⁎⁎ 0.012 (1.03) −0.045 (−0.88) 0.050 (1.52) −0.057 (−1.04) 0.011 160 80 80 1994 4.77 (9.68)⁎⁎ −3.530 (−18.10)⁎⁎ 1.381 (4.50)⁎⁎ 2.062 (58.20)⁎⁎ −1.79 (−2.45)⁎ 0.961 136 68 68 1995 0.768 (1.89)⁎ 0.004 (0.03) −1.122 (−0.66) 0.203 (3.51)⁎⁎ −0.237 (−1.22) 0.037 232 116 116 1996 0.234 (6.39)⁎⁎ 0.065 (6.15)⁎⁎ −0.023 (−0.98) 0.062 (2.63)⁎⁎ −0.047 (−1.48) 0.184 280 140 140 1997 0.278 (11.30)⁎⁎ 0.051 (6.93)⁎⁎ −0.009 (−0.778) −0.004 (−0.69) 0.038 (2.52)⁎ 0.127 400 200 200 1998 0.361 (10.90)⁎⁎ 0.002 (0.60) −0.022 (−1.07) 0.019 (1.68)⁎ −0.009 (−0.62) 0.003 320 160 160 1999 0.319 (4.04)⁎⁎ 0.052 (1.20) −0.153 (−2.53)⁎ −0.067 (−1.24) 0.267 (1.39) 0.119 64 32 32 2000 0.238 (7.03)⁎⁎ 0.025 (2.36)⁎ −0.007 (−0.42) 0.021 (2.27)⁎ −0.043 (−3.32)⁎⁎ 0.101 144 72 72 2001 0.254 (7.43)⁎⁎ 0.117 (7.59)⁎⁎ −0.080 (−5.02)⁎⁎ −0.018 (−0.98) 0.029 (1.45) 0.283 144 72 72 2002 0.260 (8.93)⁎⁎ 0.054 (5.61)⁎⁎ −0.070 (−3.48)⁎⁎ 0.048 (4.90)⁎⁎ 0.072 (2.97)⁎⁎ 0.626 208 104 104 Panel B. 1992–2002 four years before and after switching (t− 4, t− 3, t− 2, t− 1and t+ 1, t+ 2, t+ 3, t+ 4)

1992–2002 pooled 0.336 (19.50)⁎⁎ 0.015 (4.35)⁎⁎ −0.009 (−0.91) 0.043 (5.99)⁎⁎ −0.032 (−2.67)⁎⁎ 0.031 2232 1116 1116 Panel C. 1992–2003 three years before and after switching (t− 3, t− 2, t− 1and t+ 1, t+ 2, t+ 3)

1992–2003 pooled 0.335 (5.64)⁎⁎ 0.043 (2.16)⁎ −0.023 (−0.74) 0.089 (6.22)⁎⁎ −0.088 (−4.90)⁎⁎ 0.029 1758 879 879 Panel D. 1992–2004 two years before and after switching (t− 2, t− 1and t+ 1, t+ 2)

1992–2004 pooled 0.268 (8.12)⁎⁎ 0.077 (6.40)⁎⁎ −0.020 (−1.22) 0.021 (2.39)⁎ −0.020 (−1.98)⁎ 0.047 1196 598 598 ⁎, ⁎⁎ Signiﬁcant at the 0.05 and 0.01 levels, respectively.

The following regression model produces the results reported in this table: Ii;t

Ki;t−1¼ β0þ β1Qi;t−1þ β2DQi;t−1þ β3

CFi;t Ki;t−1 þ β4D CFi;t Ki;t−1 þ ei;t

For the slope dummy variable, D = 0 denotes the observations for NASDAQfirms before switching to the NYSE and D=1 denotes the observations for these firms after switching to the NYSE. The variable I represents thefirm's capital expenditure and K is net fixed assets. Q is Tobin's Q defined as the market value of total assets over the book value of total assets. Market value of total assets is equal to market value of total equity plus the book value of total debt. CF equals the cashflow from operating activities. The term ε is the error term. Each estimator's t-value and F-value is in parenthesis. The regression coefficient on D(CF/K) can be interpreted as the difference of investment–cash flow sensitivity between D=0 and D=1.

(6)

positive effects exist for NASDAQ-to-NYSE switches. Panels C and D also show thatβ4is signiﬁcantly negative at normal levels when using

pooled data with three- and two-year regressions, respectively. 6. Conclusion and directions for future work

We investigate the relation between listing switches and invest-ment–cash flow sensitivity by analyzing a sample of NASDAQ-to-NYSE switches mainly over the period 1992–2002. Contrary views exist on the effect of listing switches on investment sensitivity to cashflow. We conduct panel regressions to determine whetherfirms increase their investments when they expect better investment opportunities and whether switching from NASDAQ to the NYSE affects investment–cash flow sensitivity. We focus on using four-year data before and after switching to test the impact of switching.

Our results using pooled data show that the sensitivity to in-vestment opportunities does not differ significantly before versus after switching. Based on pooled data over 1992–2002, our evidence is consistent with the view that NASDAQ-to-NYSE switchers experience significantly lower investment–cash flow sensitivity. This finding means thatfirms rely less heavily on internal financing after switching andfind accessing external financing easier. Consequently, firms may benefit from switching in terms of a lower cost of external capital due to such factors as increasing visibility, liquidity, and reputation.

In this study, we have taken an initial step to examine the relation between switching and investment–cash flow sensitivity. Future avenues exist for more work on theoretical and empirical modeling related to this topic. For example, one avenue would be to investigate whether investment sensitivity to cashflow significantly changes for switches between other exchanges, such as moves from the AMEX to the NYSE, or when dual or cross-listing occurs. Another direction for further work would be to investigate whether the investment sensitivity to cash flow significantly changes when a firm delists from an exchange. References

Allayannis G, Mozumdar A. Cashﬂow, investment, and hedging. Working paperUni-versity of Virginia; 2000.

Allayannis G, Mozumdar A. The impact of negative cash flow and influential observations on investment–cash flow sensitivity estimates. J Bank Financ 2004;28(5):901–30.

Almeida H, Campello M. Financial constraints, assets tangibility and corporate investment. Rev Financ Stud 2007;20(5):1429–60.

Baker HK, Edelman RB. The effects on spread and volume of switching to the NASDAQ National Market System. Financ Anal J 1992a;48(1):83–8.

Baker HK, Edelman RB. AMEX-to-NYSE transfers, market microstructure, and share-holder wealth. Financ Manage 1992b;21(4):60–72.

Baker HK, Powell GE, Weaver DG. Does NYSE listing affectﬁrm visibility? Financ Manage 1999a;28(2):46–54.

Baker HK, Powell GE, Weaver DG. Listing changes and visibility gains. Q J Bus Econ 1999b;38(1):46–63.

Baker HK, Nofsinger JR, Weaver DG. International cross-listing and visibility. J Financ Quant Anal 2002;37(3):495–511.

Barry CB, Brown SJ. Limited information as a source of risk. J Portf Manage 1986;12 (2):66–73.

Cleary S. The relationship betweenﬁrm investment and ﬁnancial status. J. Finance 1999;54(2):673–92.

Dasgupta S, Sengupta K. Corporate liquidity, investment andﬁnancial constraints: implications for a multi-period model. J Financ Intermed 2007;16(2):151–74. Deshmukh S, Vogt SC. Investment, cashﬂow, and corporate hedging. J Corp Finance

2005;11(4):628–44.

Dharan BG, Ikenberry DL. The long-run negative drift of post-listing stock returns. J Finance 1995;50(5):1547–74.

Edelman RB, Baker HK. Liquidity and stock exchange listing. Financ Rev 1990;25 (2):231–50.

Fazzari SM, Hubbard RG, Petersen BC. Financing constraints and corporate investment. Brookings Pap Econ Act 1988;1:141–95.

Froot KA, Scharfstein DS, Stein JC. Risk management: coordinating corporate investment andﬁnancing policies. J Finance 1993;48(5):1629–58.

Gilchrist S, Himmelberg C. Evidence on the role of cash ﬂow in reduced-form investment equations. J Monet Econ 1995;36(3):541–72.

Grammatikos T, Papaioannou GJ. Market reaction to NYSE listings: tests of the marketability gains hypothesis. J Financ Res 1986;9(3):215–28.

Hoshi T, Kashyap AK, Scharfstein D. Corporate structure, liquidity, and investment: evidence from Japanese panel data. Q J Econ 1991;106(1):33–60.

Hubbard RG. Capital market imperfections and investment. J Econ Lit 1998;36 (1):193–225.

Kadapakkam PR, Kumar C, Riddick LA. The impact of cashflows and firm size on investment: the international evidence. J Bank Finance 1998;22(3):293–320. Kaplan SN, Zingales L. Do investment–cash flow sensitivities provide useful measures of

ﬁnancing constraints? Q J Econ 1997;112(1):169–215.

Kaplan SN, Zingales L. Investment–cash ﬂow sensitivities are not valid measures of ﬁnancing constraints. Q J Econ 2002;115(2):707–12.

Merton RC. Presidential address: a simple model of capital market equilibrium with incomplete information. J Finance 1987;42(3):483–510.

Moyen N. Investment–cash ﬂow sensitivities: constrained versus unconstrained ﬁrms. J Finance 2004;59(5):2061–92.

Oliner SD, Rudenbusch GD. Source of theﬁnancing hierarchy for business investment. Rev Econ Stat 1992;74(4):643–54.

Schaller H. Asymmetric information, liquidity constraints, and Canadian investment. Can J Econ 1993;26(3):552–74.

Whited T. Debt, liquidity constraints, and corporate investment: evidence from panel data. J Finance 1992;47(4):1425–60.